Report No. 32553-AR Argentina Seeking Sustained Growth and Social Equity Observations on Growth, Inequality and Poverty October 21, 2005 Argentina, Chile, Paraguay, and Urguay Country Management Unit Poverty Reduction and Economic Management Unit Latin America and the Caribbean Region Document of the World Bank ARGENTINA SEEKING SUSTAINED GROWTHAND SOCIALEQUITY TABLE CONTENTS OF ACKNOWLEDGMENT ................................................................................................................................ i PREFACE ...................................................................................................................................................... ii EXECUTIVE SUMMARY ......................................................................................................................... ... 111 A. ECONOMIC GROWTH, POVERTY AND INEQUALITY PERFORMANCE INARGENTINA .............................. ... 111 B. SEARCHING FORTHE CAUSESOF SLOW GROWTH AND POVERTY INCREASES ........................ C. THERECENT ECONOMICCRISISAND RECOVERY .................................................................................. vi D. GOVERNMENT POLICIES ....................................................................................................................... ix Infrastructure Investment......................... ................................ ix Fiscal Policies ........................................................................................................ Trade Policies........................................................................... Education Policies Private Sector Dev .............................................................................. 1. INTRODUCTION ............................................................................................................................... 1 2. ECONOMIC GROWTH, POVERTY AND INEQUALITYINARGENTINA ............................ 5 A. ACCOUNTING FORECONOMIC GROWTH ................................................................................................ 7 The Growth Record: Lessonsfrom Histoly ................................................. 7 The Growth Record: Volatility ................... Aggregate Growth Accounting.. ....................................... B. DETERMINANTS OFECONOMIC C. POVERTYINARGENTINA..................................................................................................................... GROWTH..,. 17 D. RECENT TRENDS ININEQUALITY .......................................................................................................... 22 E. ONTHE LINKSAMONG POVERTY, INEQUALITY AND ECONOMIC GROWTH .......................................... 26 TheEffect of Inequality and Poverty on Economic Growth: Theoretical Research TheEffect of Inequality and Poverty on Economic Growth: Empirical Research... TheEffect of Economic Growth on Inequality............................................................. TheEffect of Economic Growth on Poverty ..................................................... 30 3. THE RECENT ECONOMIC CRISISAND RECOVERY ............................................................ 32 A. RECENT TRENDS POVERTY AND GROWTH IN ....................................................................................... 33 B. PRO-POORGROWTH EFFECTS .............................................................................................................. 36 c. SECTOR GROWTH PATTERNS THROUGH THE RECOVERY FROM THE CRISIS ......................................... 42 D. SECTORAL DECOMPOSITIONPOVERTY OF VARIATIONS ....................................................................... 47 4. OPTIONSFORSELECTED GOVERNMENT POLICIES ......................................................... 51 A. INFRASTRUCTURE INVESTMENT ........................................................................................................... 52 Infrastructure, Economic Growth and Poverty ........................................ 52 Infrastructure Investment in Argentina................................................................................................ 53 Infrastructure Bottlenecks .................................................. Increasing Infrastructure Investment ....................................................... 57 B. FISCALPOLICIES ............. Macroeconomic Stability ...................................... 59 Tax Policies ......................................................................................................................................... 60 Expenditure Policies,.............................. C. TRADEPOLICIES ..................................... TradePolicies during the 1990s............................................... Further TradeLiberalization Policies ................................................................................................. 70 D. EDUCATIONPOLICIES.......................................................................................................................... 74 Recent Trends in Education......... Returns to Education ........................................................................................................................... 76 Policy Implicationsfor Education ... E. PRWATESECTORDE Standards and Quality of Argentine Goods .................................................. 81 Innovation System in the Argentina ................................... Investment Climate .. Policy Implications .. .......................... 84 5. CONCLUSIONS ................................................................................................................................ 87 REFERENCES ............................................................................................................................................ 89 TABLES TABLE2.1: ANNUALGROWTH ............................................................................. 8 TABLE2.2: TRENDGROWTH OUTPUTVOLATILITY, SELECTEDCOUNTRIES, 1960-99............................ RATES OF GDP PERCAPITA AND 11 TABLE2.3: RECENT STUDIES OF TOTALFACTOR PRODUCTIVITY(TFP) GROWTH ........................................ 13 TABLE2.4: GROWTH ACCOUNTING ADJUSTING HUMAN FOR CAPITAL, 1961-2000 ....................................... 14 TABLE2.5: DETERMINANTSARGENTINA'S OF GROWTH ...............16 TABLE2.6: DECOMPOSITION THEPOVERTY-INDEX VARIATIONS.............................................................. RATEOF GDP PERCAPITA, BY DECADES OF 21 TABLE3.1: POVERTYAND INDIGENCERATES, 2001-04 ................................................................................. 34 TABLE3.2: DECOMPOSITIONPOVERTYAND EXTREME ................................... TABLE3.3: AVERAGEPERCENTAGECHANGESININCOMEOFTHEPOOR, 2001-04 ........................................ OF POVERTYRATES, 2001-04 36 37 TABLE3.4: MOBILITY PROFILESBY INITIALPOSITION, 1996-2003 ................................................................ 40 TABLE3.5: GROWTH DECOMPOSITION SECTOR, 1994-2004..................................................................... BY 43 TABLE3.6: EMPLOYMENT DISTRIBUTION ECONOMIC BY 45 TABLE3.7: SECTORLEVEL GROWTH ACCOUNTING SECTOR, 1995-2004................................................. SECTORS, 2001-04 ................................................ BY 46 TABLE3.8: EDUCATIONAL ATTAINMENTBY SECTOROFEMPLOYMENT ....................................................... 47 TABLE3.9: SECTORALDECOMPOSITION OF POVERTY, 2001-04 .................................................................... 48 TABLE4.1: POTENTIALGROWTH IMPACT FROMEXPANDED IMPROVEDINFRASTRUCTURE AND ...................52 TABLE4.2: PUBLICINVESTMENT ININFRASTRUCTURE 55 .................................................................................. TABLE4.3: ARGENTINA'S ENDOWMENT PRODUCTIVEINFRASTRUCTURE................................................. OF 55 TABLE4.4: LOWERBOUND TABLE4.5: INCIDENCEOFTAXESBY QUINTILE, 1998 ................................................................................... 62 ANNUALINFRASTRUCTUREINVESTMENTNEEDS ARGENTINA.................... FOR 58 TABLE4.6: INCIDENCEOFSOCIALEXPENDITURES 65 SPENDINGTO CHANGES INGDP, 1980-2004 ............................ ........................................................................................ TABLE4.7: ELASTICITIESOFGOVERNMENT 66 TABLE4.8: TRADEOPENNESS INDICATOR .................................................... 68 TABLE4.9: EMPLOYMENT TRADE.......................................................................................................... AND MERCOSUR, 1991-2000 AND 71 TABLE4.10: THEIMPACTOFWTODOHA TABLE4.11:EDUCATIONATTAINMENTOFTHETOTALPOPULATIONAGED15AND OVER, 1960-2000.........73 NEGOTIATIONS, CHANGESINPOVERTY LINES ............................. 74 TABLE4.12: PISA 2000AVERAGE 76 TABLE4.13: RETURNS EDUCATIONINARGENTINA,1992-2002 ............................................................... SCORES BY COUNTRY............................................................................. TO 78 TABLE4.14: DESTINATIONSNEW ............................................... 81 TABLE4.15: INDICATORSOFINVESTMENTCLIMATE..................................................................................... 84 OF INVESTMENTS INARGENTINA,2001-04 FIGURES FIGURE2.2: MAPPING RELATIONSHIPOFINEQUALITYAND...................................................................... FIGURE PERCAPITA GDP.POVERTY AND INEQUALITY. ...................................................................... 5 2.1: THE 6 FIGURE2.3: RATIOOFARGENTINA'S CAPITA GDP .................................................................................. PER 7 FIGURE2.4: POVERTYHEADCOUNT INGREATERBUENOS 1980-2004 .................................... RATIO AIRES, 18 FIGURE2.5: POVERTYHEADCOUNT INLACCOUNTRIES.................................................................... RATIO 20 FIGURE 2.6: GIN1COEFFICIENT OFHOUSEHOLDCAPITA INCOME ........................................................... 23 FIGURE2.7: GINI COEFFICIENTINLACCOUNTRIES...................................................................................... PER 24 FIGURE GROWTH 3.1: INCIDENCE CURVES ................................................................................................. 38 FIGURE3.2: GROWTH INCIDENCE CURVES .................................................................................................... 39 FIGURE 3.3: REALWAGES INDEX, 2002-2005 ............................................................................................... 42 FIGURE3.4: SECTORECONOMIC GROWTH AND CONTRIsUTIONTO POVERTY CHANGES ............................... 49 FIGURE3.5: SECTORECONOMIC GROWTH AND CONTRBUTIONTO POVERTYCHANGES ............................... 50 FIGURE4.1: INVESTMENTININFRASTRUCTUREINARGENTINAAND LATINAMERICA................................... 54 FIGURE4.2: BREAKDOWNFEDERAL FIGURE4.3: EVOLUTIONOFTHEGIN1INDICATOR OFCONCENTRATION......................................................... OF GOVERNMENT REVENUES, TAX 2004 .............................................. 61 69 FIGURE4.4: AVERAGE RETURN SCHOOLINGOVERT ~ E TO .......................................................................... 77 ACKNOWLEDGMENT This report has been prepared by a team led by Carlos G. Fernhndez Valdovinos (Task Manager, LCSPE), under the overall guidance o f Mauricio Carrizosa (Sector Manager, LCSPE) andJames Parks (Lead Economist, LCC7A). The team was comprised by William Tyler (Consultant) on Argentina's growth record, Gary Fields (Cornel1 University) and Maria Laura Shchez (Consultant) on individual's income mobility, Evelina Bertranou (LCSHS) and Melanie Khamis (Consultant) on labor markets, Carlos G. Femandez (LCSPE), Alvaro Vivanco (LCSPE) and Alejandro Guerson (LCSPE) on sectoral growth, Marianne Fay (LCSFP) on infrastructure, Emily Sinnott (YPP) on fiscal policies, Guido Port0 (DECRG) and Pablo Sanguinetti (Consultant) on trade policies, Mauricio Santamaria (LCSPP) and Ana Maria Diaz (LCSPP) on education policies, and JosC Luis Guasch (LCSFP) on private sector development. During several missions to Argentina, the team met with and received support, guidance and encouragement from the main counterparts including: Oscar Tangelson (Secretary o f Economic Policy), Gerard0 Hita (National Director o f Projects with International Credit Organizations), Sebastian Katz (Sub-secretary o f Economic Programming), Carlos Bolo Bolafios (Financial Director, Direction o f Projects with International Credit Organizations), and Roxana Maurizio (Advisor, Sub-secretary o f Economic Programming). The report was enhanced by the valuable comments by peer reviewers responsible for the official review o f the document -Emanuela Galasso (DECRG), Paul Levy (SFRCR) and Josk Seligmann (IDB)-, as well as by the following participants at discussion meetings and/or readers of the document at its various stages: Axel van Trotsenburg (Director, LCC7), Juan Gaviria (Sector Leader, LCSFP), Jesko Hentschel (Sector Leader, LCSHD), Carter Brandon (Sector Leader, LCSES), Pablo Fajnzylber (LCSFR), Harry Patrinos (LCSHE), Humberto Lopez (LCRCE), and Jaime Saavedra (LCSPP). Useful editorial suggestions were received from Emily Evershed(Consultant). Maritza Bojorge (LCC7A) was responsible for formatting and compiling the final document. The team benefited from the suggestions o f government officials and other participants o f the private sector, think tanks, and academia in a seminar organized in December 2004 in Buenos Aires. Claudia Nin (LCC7C) and Maria Emilia Sparks (LCC7C) helped to organize the seminar. The main conclusions o f this report were presented to the Government in June. In addition, the team received comments from government officials inAugust. These comments were incorporated to the report. PREFACE Selected technical background papers prepared for this report were edited in a separate volume and are available upon request from the task manager; they can also be foundon the regional website (wuwbancomundial.org.nr). The following background papers were included: "Labor Markets and Pro-Poor Growth" by Evelina Bertranou and Melanie Khamis; "Poverty, Inequality and Economic Growth" by Gary Fields; "How i s Convergent Mobility Consistent with Rising Inequality? A Reconciliation in the Case o f Argentina" by Gary Fields and Maria L. Shnchez Puerta; "The Argentinean Competitiveness: Seeking Growth Sustainability - Issues that Merit Attention" by Josk L. Guasch; "Trade Policy, Trade Flows, Poverty and Income Inequality in Argentina" by Guido Port0 and Pablo Sanguinetti; "Fiscal Policy Options for Equity and Growth: A Reading of the Economic Literature Relevant to Argentina" by Emily Sinnott; and "Argentina: The Record of Past Growth" by William Tyler. .. 11 EXECUTIVESUMMARY 1. This report seeks to contribute to the ongoing dialogue with the Argentine authorities on economic policies to support sustained growth with social equity. It has beenprepared in close consultation with the authorities inthe Ministryo f Economy and Production and reflects the Government's priorities in exploring certain topics. As a result, the report is not a comprehensive treatment o f the sources of growth inArgentina. Inparticular, it does not cover the important andcomplex issues related to the financial sector and institutional development which are being addressed in other studies. Rather, the report attempts to deepen the analysis on selected topics, chosen incollaboration with the authorities, regarding the inter-linkages between economic growth, income distribution and poverty, as well as the respective roles o f these factors in explaining the historical underperformance o f the Argentine economy. The objective o f the study i s to identify relevant issues for policy formulation and further economic work. Its emphasis is on longer-term structural factors which are thought to determine productivity and income distribution, not the short-term conjunctural challenges facing the country. The conclusions presented in the study are not definitive, given the selectivity of the topics covered andthe limitations on the data and analytical techniques employed. Nevertheless, they hopefully can contribute to the policy debate and help set the agenda for future research. 2. Improving economic growth to reduce poverty is one o f Argentina's most difficult development challenges. This study reviews Argentina's long term growth performance and the relationship between growth, poverty and income distribution. It also discusses selected key policies that could help improve the country's economic performance. The report argues that Argentina's past growth performance suffered from a mix o f extemal shocks and inappropriate economic policies; that increased poverty over the last three decades resulted fi-om stagnant growth coupled with increased inequality; that the pattern o f economic growth failed to reduce poverty even in periods displaying high growth rates; that economic volatility, as painfully observed during the last economic cycle, has particularly affected the incomes o f the poor; and -looking forward- that well conceived policies could help improve growth and poverty reduction performance. These arguments are summarizedinwhat follows. A. ECONOMIC GROWTH, POVERTY AND INEQUALITYPERFORMANCEINARGENTINA 3. Argentina's slow economic growth over the past ninety years has been extensively discussed by economic historians. Argentina was among the world's highest income countries in 1913. Historical data show that economic growth was clearly disruptedin all countries by the adverse shocks o f the Great War, the interwar economic strains, the Great Depression and World War 11. Yet other countries with comparable living standards and growth records before these events did manage to continue subsequently their trend growth. This was not the case for Argentina, where growth did not resume at anywhere near previous observed rates and economic slowdown set in beginning roughly around the time o f World War I.Over the period 1950-2000, Latin America's per capita income as a whole diverged downward vis-&vis the industrialized ... 111 countries as a group, and Argentina's per capita income converged downwards towards Latin America's per capita income. In contrast, there was convergence within industrial countries, as the initially poorer ones (e.g., Japan, Ireland, Portugal) made progress in catching up with the early leaders. A number o f Asian countries also converged towards the richer countries. Within Latin American countries as a group, Argentina stood out as a particularly slow growth country, with per capita income having grown only 1.1percent annually over the 50 year period. Only in two Latin American countries (Venezuela and Bolivia) was growth lower. As a result, Argentina, once an unquestionably advanced country, became relatively poor. 4. This is one great puzzle o f Argentine economic history. Unfortunately, there are no totally convincing simple answers to the question o f what accounted for the languishing performance o f Argentine growth since World War I.In any event, the country is a unique case o f unresumed growth, or arrested development. In Rostovian terms, Argentina is the only country that clearly reached the "take off' phase for economic growth, but then aborted its flight into sustained growth. 5, The last several decades have uncovered another important puzzle: inequality increased persistently, resulting in increased poverty in the face o f stagnant average - albeit widely fluctuating- per capita output. After peaking in 1974-80, per capita GDP declined by 22 percent during the 1980s, increased by 42 percent through 1998, declined by 22 percent through 2002, and recovered to the 1998 level by 2005. All in all, per capita GDP in 2004 was at about the same level as in 1974. Nevertheless, poverty was much higher in 2004, reflecting an increasingly unequal income distribution over the last three decades. Remarkably, the rise in inequality was observed inperiods o fboth growth and recession. Similarly to poverty measures, the inequality measures for Argentina in the last 15 years deviated from the trend o f the rest o f the countries in Latin America. Duringthis period, the changes ininequality have beenmuch larger inArgentina thanin other countries. The country used to be one of the three most equal in Latin America, along with Venezuela and Uruguay. Today, this is not the case anymore and other countries like Costa Rica, Peru and Jamaica are showing a lower degree of income inequality. Although the region has not been very successful in reducing poverty, the recent record o f most Latin American countries was better than the Argentine performance. As a result Argentina today i s not the low-poverty and equitable country (by LatinAmerican standards) that it was inthe early 1990s. 6. Not only has growth been slow in Argentina but there has also been substantial volatility in economic performance. Business cycles are common in all economies, but the Argentine case seems to be set apart when compared to the cycles o f other countries. This high volatility may have contributed to slower economic growth over time. Recent empirical analysis, based upon cross-country regressions, provides strong evidence that greater volatility in output growth consistently reduces the average rate o f growth. Data show that Argentina possessed greater output volatility than any other Latin American country or any major region during the 1960-99 period. Argentine volatility was greatest duringthe chaotic (for Argentina) decade o f the 1980s. For that period only Peru, with a still more chaotic decade, surpassed Argentina in terms o f volatility. With stabilization iv and greater macroeconomic balance in the 1990s, Argentina's volatility was reduced but it still remained higherthan inmost other countries. 7. Poverty and inequality indicators substantially worsened during the last, and very deep, business cycle. The economic collapse in Argentina culminating in 2002 was dramatic, resulting in the deepest political and economic crisis in generations. Few countries have experienced such economically catastrophic events: over the four-year period from 1999 to 2002, GDP fell by over 20 percent. The human costs o f the precipitous decline in economic activity were substantial, with the crisis resulting in severe social dislocations and a large reduction o f welfare for Argentina's population. The economy began to rebound strongly in 2003. Sparked initially by increased exports and by a gradual expansion o f consumption spending, a demand-led recovery emerged. The recovery has begun to have a positive impact on the living standards o f the population, as shown by the steady improvement in key social indicators since 2002. Nevertheless, the economic and social advances were from a very low base, and large segments o f the Argentine population continue to face great hardship. B. SEARCHINGFORTHE CAUSESOFSLOW GROWTH AND POVERTY INCREASES 8. Why has economic growth been so low inthe country? Under one approach, the explanation i s the lack o f productivity growth. Productivity growth has long been recognized as a driving force for sustained economic growth. Abundant empirical work (so-called growth accounting) indicates that total factor productivity (TFP) growth i s seen to dominate factor accumulation as a "source" o f economic growth inmost countries. In the case o f Argentina, there have been a large number o f recent growth accounting studies which have sought to measure TFP growth since the 1940s. There is a certain commonality in the major results o f these analyses. First, for the longer periods o f time for which TFP i s estimated (three to five decades), average TFP growth i s generally seen to be low, hardly more than 0.5 percent per year. Second, TFP growth rates are quite erratic, reflecting Argentina's highvolatility noted above. Third, the average TPF growth estimates are typically negative for the chaotic 1980s and most positive for the 1990s. Fourth, as expected, when attempts are made to incorporate human capital accumulation, the estimates o f TFP growth decrease appreciably. 9. The "informal empiricism" o f economic historians highlights other key developments to be considered in understanding Argentina's poor growth performance: the closing o f the economy in the 1930s, the reduced access to international capital markets that followed, inadequate investment incentives throughout much o f the post-war period, macroeconomic instability, the increased burden o f the state after the 1930s, and weaker property rights from 1935 to about 1952. The possible negative impact o f these developments is validated by intemational experience and recent cross-country econometric studies. These studies on the determinants o f growth support the conventional wisdom that has been emerging with respect to growth economics over the past 30 years. More specifically, improving macroeconomic stability, facilitating private sector investment, strengthening the legal and institutional framework, having a well functioning system for financial intermediation, promoting trade expansion and greater V openness o f goods markets, and extending human capital formation would appear to be crucial actions for achieving sustained highrates o f economic growth inArgentina. 10. Together with low average growth, Argentina has experienced increasingpoverty. Considering the last 30 years, an upward trend o f poverty can be seen in the data. The data also show two sharp peaks, one in 1989 and the other in 2002, both crisis years. Moreover, poverty rates were higher in 2002 than ever before. Between 1992 and 2002, 11 million individuals became poor: 3.5 million entered poverty during the economic growth period o f the 1990s, another 3.5 millionjoined that group inthe first phase o f the recession (1998-2001), while about 7.5 million crossed the poverty line during the crisis 2001-02. The ongoing economic recovery has substantially reduced the number o f poor by around 3.3 million individuals. The most striking fact in the data is that poverty increased even in periods o f economic expansion. The fact that poverty sometimes increased when growth was taking place is exceptional inintemational experience. 11. As suggested before, low growth does not tell the whole story, increased poverty was also linked to a more unequal income distribution. The question then arises as to why income became more unequal. Recent research has uncovered a number o f factors that help explain the change in income inequality in the country during the 1990s. Key factors include: (i) an increase inthe relative retums to higher education; (ii)rise inthe a retums to unobservable factors and the relative decline in hours o f work for unskilled workers, which were particularly important from 1992 to 1998; (iii)an increase in unemployment during the 1990s and changes in labor force participation from 1996 onwards; and (iv) increased integration into the global economy which encouraged production and employment in the sectors using natural resources, helped reduce the relative price o f capital, and introduced new "skilled labor" intensive technologies (both o f which raised the skill premium). A fall in the minimum wage in real terms and increased family size also played a role. 12. The high degree o f income inequality in the country is a factor that could have impaired growth. There is a substantial theoretical literature that identifies channels through which income inequality negatively affects economic growth. Also, several empirical studies have found a statistically significant relationship whereby highincome inequality reduces economic growth. As stated in one o f these studies: "There i s no evidence in the data that increases in inequality are good for growth. Infact, the bulk o f the evidence goes inthe opposite direction." However, further tests are needed before an unquestionable conclusion can be reached. Given the rising trend in inequality in Argentina, it will be worthwhile to fkrther explore this issue infuture research. c. THERECENTECONOMICCRISISAND RECOVERY 13. As indicated before, Argentina's volatility stands out when compared to other countries. According to empirical studies, a high volatility on macroeconomic performance usually contributes to increase the level o f inequality and poverty in different countries. A review o f the last economic cycle offers an eloquent glimpse on the link of volatility to poverty in Argentina. After three years o f continuing recession following the peak inoutput achieved in 1998, the economic and financial situation o fthe vi country worsened during 2001. Various attempts to spur growth and improve public finances andthe debt profile failed. The year culminated with the resignation o f President de la Rua, followed by a quick succession o f appointed presidents, a formal announcement of default on public sector debt, and, in early 2002, the abandonment of the convertibility plan. The social situation in the country, which was not good in 1998, significantly worsened in2002 after GDP declined by 10.8 percent duringthat year. 14. As expected, poverty and extreme poverty rates rapidly increased during the economic crisis. The official poverty rate shows that some 53.0 percent o f the population was poor inMay 2002, up from 35.9 percent a year earlier. This sharp increase inpoverty also appeared at the household level: the percentage o f poor households went up from 26.2 percent in May 2001 to 41.4 percent in May 2002. Additionally, extreme poverty (indigence) more than doubled during the same period. The percentage o f households in extreme poverty increased from 8.3 percent inMay 2001 to 18.0 percent inMay 2002. At the individual level, the indigence rate rose from 11.6 percent to 24.8 percent during the same period. 15. The economy rebounded quite strongly in 2003, with real GDP growing at 8.8 percent. Nevertheless, duringthe early phase o f the recovery, poverty and indigence rates continued to increase, albeit only slightly. For example, the percentage o f poor individuals (households) increased from 53.0 percent (41.4 percent) to 54.7 percent (42.6 percent) between M a y 2002 and May 2003. The economy continued to show a strong performance the next year, growing 9.0 percent during 2004. Importantly, economic recovery finally began to have a positive impact on living standards during that year, as shown by the steady improvement in key social indicators. Poverty and indigence rates fell between the second semester of 2003 and the second semester 2004. The percentage o f individuals (households) in poverty declined to 40.2 percent (29.8 percent) at the end of 2004. At the same time, the indigence rate for individuals (households) also dropped to 15.0 percent (10.7 percent). 16. A decomposition analysis o f poverty changes during 2001-04 shows that, as observed when analyzing longer periods, poverty changes were largely determined by economic growth. Nevertheless, during 2003-04 the "distribution effect" turned out to be as important as the "growth effect" in explaining changes inpoverty and extreme poverty inthe country. To better understand the social impact of the recent crisis and upturn, this report also investigated how the poor shared the losses (benefits) o f the recession (recovery) calculating percentage changes in the income o f the poor over the different sub-periods. It was found, not surprisingly, that economic activity during the cycle was closely, and inversely, linked to income changes. Even when adjusted by government transfers, the income o f the poor showed a positive correlation with the growth rate o f the economy: it fell significantly during the crisis and experienced a revival with the upturn o f the economy. That is, the data indicate that while the crisis hit more the poorer than the richer population groups, the subsequent recovery was clearly a pro-poor growth episode (inthe absolute sense). 17. Additionally, growth incidence curves indicate that the poor benefited relatively more from the recent recovery. Incidence curves look at patterns of income changes vii across the different percentiles o f the entire income distribution curve. Overall, they show that the incomes o f the poor grew (fell) faster than the average income o f the entire population during the upturn (crisis). That is, the analysis using growth incidence curves found that the recent phase o f economic recovery has clearly been a pro-poor growth process, not only inthe absolute sense but also inthe relative one. 18. This report also presents preliminary evidence that, as a result o f changes in relative input prices, changes in sector growth patterns during the recent recovery could have helped increase its positive effects on the poor. The reason is that changes inrelative input prices probably led to changing patterns o f input and sector contributions in aggregate production. A brief comparison with the 1990s could illustrate the point. Duringthe 1990sinterest rates decreasedrelativeto wages, andthe most dynamic sectors tended to be capital intensive. In contrast, real currency depreciation after the crisis substantially reduced the cost o f labor in the country during the recovery. For example, real wages in the private and public sector dropped sharply after the crisis. As a consequence, labor intensive sectors became more competitive after the currency depreciation. 19. A overall review o f the sectoral growth pattern during the recovery period that followed the 2001-02 crisis reveals that: 0 sectors that contributed most to aggregate GDP growth inthe recovery period are different from those that explained growth in the 1990s. The recovery shows the relatively higher importance o f the contribution o f the goods- producing sectors, mainly manufacturing, while in the 1990s the services sectors were most important; 0 the dynamic sectors during the recovery have been relatively more labor- absorbing; 0 the most dynamic sectors have shown higher increases ininvestment and total factor productivity; 0 the most dynamic sectors during the recovery demand relatively more intensively low-skilled workers compared to the sectors that led growth inthe 1990s. 20. Since labor income i s more important among the poor as a share o f total income, and since the poor tend to be less-skilled, it follows that the pro-poor growth pattern observed during the recovery may have been linked to the sectoral composition o f the recovery. Infact, this study examined whether those labor-absorbing sectors that grew the most during the recovery also accounted for most o f the changes in poverty during the period. It was found that the correlation between sector growth and contribution to poverty reduction was positive during the most recent part o f the recovery. That is, the higher the economic growth o f the sector (weighted by employment share o f the sector), the higherwas its contribution to poverty reduction. 21. The main policy implications of the analysis o f the recent crisis and recovery cycle are the following. First, macroeconomic stability i s important not only for long-run growth but also for poverty reduction. Poverty levels usually increase sharply after an economic crisis, and Argentina was no exception. After the recent crisis, poverty rose V l l l ... dramatically inthe country: the number o f poor grew by roughly 20 percentage points to more than half o f the total population, while the number of extreme poor nearly doubled. Second, it was found that the recovery was pro-poor and that one important determinant o f this outcome could have been the change inrelative prices after the devaluation. It was found that sectors that contributed most to GDP growth duringthe recovery were also the sectors that were relatively more labor-absorbing and that usually demand less-skilled workers. These characteristics o f the recent recovery were probably crucial to finding the result of pro-poor growth patterns. This hypothesis deserves further study and analysis, particularly as the recovery continues to unfold. Third, the role o f government transfers was critical to alleviating poverty, especially extreme poverty. It was found that poverty and indigencerates might have beenhigher hadthose government transfers not existed. D. GOVERNMENTPOLICIES 22. Sustaining the recent economic recovery and ensuring that it will benefit the poorest i s a matter o f top priority for the government. In fact, one o f the difficult tasks that Argentina faces today i s to implement policies needed to ensure that the country embarks on a path o f sustained and equitable economic growth. The sudden impoverishment experienced after the crisis has called (and still calls) for interventions that support the economic recovery, assist employment creation, and enhance the living conditions o fthe population. 23. The final part o fthis study discusses a selected set o f government policies that the theoretical and empirical literature has found effective in delivering high economic growth. O f course, an analysis o f all the options currently available to the government i s well beyond the scope o f this report. Thus, this work focuses on a limited number o f options in the areas of: infrastructure, fiscal policy, trade policies, human capital formation (education) and private sector development. An extensive number o f studies have concluded that these policies are major determinants o f economic growth. But in addition to considering how these policies could help foster growth in Argentina, this report also examines their likely effects on povertyand inequality. InfrastructureInvestment 24. Empirical studies have found a substantial growth pay-off in increasing the quality and quantity o f infrastructure. According to one o f these studies, improving infrastructure stocks in Argentina to the level o f the regional leader (Costa Rica) could result, ceteris paribus, in a rise in the GDP growth rate o f 1.3 percentage points per annum. In addition, improving the quality o f infrastructure in the country could add another 0.4 percentage points. This adds up to an overall improvement in growth performance of 1.7 percentage points per annum. Similarly, these studies also find important effects o f infrastructure on reducing inequality. For example, the impact for Argentina o f improving its infrastructure performance to the level o f Costa Ricawould be a reduction o f its Gini coefficient by 0.04, while increasing it to the level o f Korea would decrease the Gini index by 0.06. ix 25. Historical data suggest that in the 1980s Argentina's public investment in infrastructure was around 3 percent o f GDP. However, it fell to less than one-half o f a percent o f GDP in the 1990s and went down further to about one-quarter o f a percent by the beginning of the new millennium. This negative trend o f public investment in infrastructure was only partially offset by private investment in the sector, which amounted to around 1.2 percent o f GDP on average during this period. Therefore, average total investment in infrastructure during the 1990s remained under 2 percent o f GDP in the country. Compared to other Latin American countries, these rates o f investment would place Argentina close to average for the 1980s but much below it for most o f the 1990s. Similarly, cross-country regression analyses suggest that Argentina has significantly less infrastructure than would be expected given its level o f income and its socio-economic characteristics. 26. Reversing the declining trend in infrastructure investment is necessary for a sustained recovery that could create employment, generate wealth and improve the quality o f life o f Argentineans, including those in the lowest deciles o f income. Argentina's competitive position has improved owing to the recent devaluation. However, infrastructure bottlenecks are seen to be impeding economic activity in a number o f areas, especially exports. Energy, transport and logistics costs are among the key impediments to growth. Currently, the country has high logistics costs that are equivalent to 29 percent o f GDP, much higher than the OECD average and second only to those for Peru in Latin America. Moreover, following the 2001 devaluation, logistic costs in peso terms have been rising faster than the consumer price index and have also outstrippedthe producer price index. As a consequence, the port o f Buenos Aires, which i s the maingateway, i s losing competitiveness as a regional gateway. 27. Fiscal constraints are one reason why there is a need for a continued role o f the private sector ininfrastructure investment. Given projectedGDP growth for 2005 to 2010 o f 3.7 percent per annum on average, investment requirements in Argentina would be about 1.4 percent o f GDP per annum. This is probably a lower bound figure, given that it excludes ports, airports, urban transport and gas as well as rehabilitation or upgrading costs. Adding a minimum maintenance expenditure o f about 0.8 percent o f GDP yields an estimate for total infrastructure investment o f 2.3 percent o f GDP per year, an equivalent o f about AR$10 billion (in current AR$)per annum for the period. However, macroeconomic assumptions consistent with both the government's program and the authorities' expectations imply that the federal government i s projected to be able to allocate only around 1.1 percent o f GDP per year to capital expenditure. Therefore, the percentage o f GDP that could be allocated to public infrastructure investment under the government's macroeconomic program would not be sufficient to meet the country's requirement to support a GDP growth rate o f 3.7 percent inthe medium term. Fiscal Policies 28. This report highlights the notion that fiscal, monetary and financial policies that contribute to a stable macroeconomic environment and avoid financial and balance o f payments crises are crucial for long-run growth. Moreover, it emphasizes the idea that these sound policies will have also the greatest success in reducing poverty. These X conclusions are based on a large number o f country case studies that found inequality increasing along with absolute poverty levels during macroeconomic crises. Therefore, maintenance o f government policies aimed at macroeconomic stability will be key to a lastingreduction inpoverty andinequality inArgentina. 29. Given the low importance o f personal income taxes and property taxes in Latin American countries, the direct redistributive leverage o f the tax system inmost countries should be expected to be very small. Argentina i s no exception andtax policy, initself, is not a good tool for redistribution under those conditions. Its primary importance lies instead in generating an adequate amount o f revenues with which the government may tackle inequality and poverty (through expenditure policies). A fair distribution o f the tax burdenacross the income deciles is obviously important. However, evidence points to tax policy as not being a very effective way o f reducing inequality in developing countries. Empirical simulations for some developing countries suggest that even a moderately progressive tax system is unlikely to have much o f an impact on the post-tax income distribution. An additional problem o f using tax policy for redistribution purposes i s that the major taxes that contribute to revenues in developing countries, in general, and in Argentina, in particular, are flat taxes. Of course, the progressiveness o f the tax system could be increased, for example, through the elimination o f VAT on basic necessities. However, a high cost would be incurred in terms o f administrative effort, and the installation o f a system o f exemptions for certain goods would likely increase tax evasion. 30. Expenditure policies would be the most effective mechanism to diminish inequality and fight poverty under the conditions prevailing in Argentina today. To this end, a government typically wants to ensure that the composition o f expenditures promotes the reduction o f inequality and poverty through policies that increase opportunities for the poor and create well-targeted social safety nets. Nevertheless, public expenditures are not always efficiently targeted towards the poorest. In the case o f Argentina, an analysis o f the benefit incidence o f expenditures i s revealing. Inthe 1990s there was a shift in expenditure towards the social sectors. Yet this higher level o f social expenditures did not translate into lower income inequality. This occurred, first, because social spending was not always progressive. Second, a large component o f social spending was in-kind, such as the provision o f public services to individuals, so it did not have an immediate impact on incomes. For Argentina, the results o f incidence analysis studies for the country indicates that some important programs, such as pensions systems, tertiary education and health, are regressive. This illustrates the importance for the government of conducting on a regular basis an incidence analysis o f expenditures to gauge who benefits from access to public services. The government can then modify spendingto make it more progressive. Trade Policies 31. Argentina has taken major strides to liberalize trade since the end o f the 1980s, with a significant reduction in average tariffs and inthe percentage o f items covered by import licenses and other quantitative restrictions. These reforms are usually thought to have stimulated trade and economic growth. However, trade policy developments in xi Argentina during this period could hypothetically have reduced employment in some industries (and thus, increased poverty and inequality in the country). Data show that between 1993 and 1998, a fall in employment took place in most manufacturing activities. It seems potentially possible to associate at least part o f this phenomenon with trade openness, as the sector faced stiff competition from foreign producers as a result o f the liberalization policy and the overvaluation o f the national currency. However, results presented in this report show no significant (unconditional) correlation between employment changes and import penetration. 32. The effects o f trade liberalization policies on poverty can be also be gauged by analyzing the dynamics o f relative wages for different skill levels. In Argentina the service sector i s intensive in skilled labor while the manufacturing sector i s intensive in low-skilled labor (Le., unskilled and semi-skilled workers). Thus, if industries experiencing larger reductions inprotection levels employed a greater proportion o f low- skilledworkers, trade liberalization could hypothetically have the effect o f decreasing the wages o f less-skilled workers relative to those o f high-skilled workers. This study analyzed whether deepening trade liberalization during the 1990s had any identifiable impact on the distribution o f wages. More specifically, it tested whether those manufacturing sectors inwhich import penetrationrelative to gross value added deepened are, ceteris paribus, the sectors in which a higher increase in wage inequality by skill group occurred. The results show that indeed the increase in import penetration in the manufacturing sector contributed to the increased wage inequality in Argentina, harming the less-skilled (unskilled and semi-skilled) workers. Nevertheless, the identified effect does not seem to be the main cause o f growing wage inequality duringthe period. 33. Finally, the report also provided key inputs for analyzing the potential impact o f various trade initiatives on employment opportunities for relatively unskilled workers. This would, o f course, have direct implications for poverty and income distribution indicators. The general lesson is that trade liberalization, especially for agricultural products, can potentially be beneficial for the country: a liberalized international trade, mainly following the WTO agricultural reforms, would provide Argentina with new trading opportunities that would be likely to raise national welfare. It should be noted, however, that inany episode of economic reform, there will be adjustment periods and as a result there will be winners and losers. It i s the role o f complementary government policies to act as an engine to propel the new trading opportunities and, at the same time, address the more important transition costs. EducationPolicies 34. One o f the key challenges that Argentina faces today i s buildingthe skilled labor force needed to ensure that the country embarks on a path o f sustained and equitable economic growth. Recent data show that Argentina has one o f the most developed education systems in Latin America. Despite the recent economic crisis, school enrollment rates remain high. For example, the average years o f schooling for the population were 8.8 years in 2000, significantly higher than the regional average. Argentina also compares well with Eastern and Central Europe and also East Asia, where average educational attainments in 2000 were 8.4 years and 7.6 years, respectively. xii Moreover, the trend in the educational attainment o f the labor force has been positive in recent years. 35. Regarding the quality o f education in Argentina, the results o f the 1997 Latin American Laboratory for Assessment o f Quality inEducationindicates that students from Argentina were among the top performers, along with students from Cuba, Brazil, and Chile. Nevertheless, when compared to students from the rest o f the world, Argentine students are lagging well behind. In two recent international assessments including countries in Latin America and the rest o f the world -the Progress in International Reading Literacy Study (PIRLS) and the Program for International Student Assessment (PISA) -students from Argentina and other Latin American countries were among the lowest performers. 36. This report also discussed the results o f a recent study on the returns to education for workers belonging to different income groups in urban Argentina. In particular, the study asked the question o f whether there are systematic differences (heterogeneity) betweenthe poor and the non-poor inthe compensation they receive inthe labor market for their education. The main results are as follows: (i)relative returns to higher education (university level) compared to primary complete grew significantly over the last 12 years across the entire income distribution; (ii) relative returns to intermediate levels o f education, on the other hand, declined throughout the period or at best remained constant; (iii)retums appear to be higher for women than for men within each income group; (iv) for women, it i s very clear that those belonging to the highest segments of the income distribution enjoy larger returns; (v) for men, the analysis based on per capita income produced similar results to those for women, but the BNI-based analysis did not provide evidence o f significant differences in returns across groups, except for the last years and for the highest education level only; and (vi) both analyses were very clear in pointing out that for both men and women heterogeneity in the returns to education is only present for the higher income groups. 37. The heterogeneity o f educational returns is crucial for the design o f public policies. For example, evidence o f increasingreturns favoring the richer segments o f the population calls for policies that will improve the quality o f education and enhance the job .search process for the poor. Indeed, if such a pattern o f increasing returns remains, the possibilities o f mobility across the income distribution for lower income groups will be systematically limited and the value o f education as an anti-poverty engine will be constrained. A vicious circle could arise if the poor did not find it profitable to invest in education because o f low expected retums. According to endogenous growth theory, this shortfall inhuman capital investment would intum lower the country's growth prospects. Private Sector Development 38. Argentina's economic recovery in 2003-04 has been remarkable. The new post devaluation environment provided ample opportunities for economic growth. Yet those opportunities granted by the new terms o f trade are seldom lasting, as illustrated by the experience o f other countries. It i s quite likely that the current (real) exchange rate, which made both export and import substitution very attractive, will not be maintained in the ... X l l l mediumterm. Therefore, there is a real danger that firms that entered or expanded under these market conditions, despite doing well initially, might be unable to compete in the future. Such a situation could affect the sustainability o f current growth and induce intense pressures for protection as the exchange rate advantage subsides and firms drop out o f the export market or find it difficult to compete with cheaper imports. 39. In order to sustain the strong path of economic growth, it is important for Argentina to use the window o f opportunity created by the recovery to address a number o f important structural issues. These policies would help ensure a more favorable environment for private sector development and allow the country to attain a more competitive position for facing extemal competition over the mediumterm. For example, the government would be advised to undertake a comprehensive program o f action to improve the investment climate by simplifying and/or reviewing unnecessary (or inefficient) procedures and regulations that affect the entry, operation and exit o f businesses; revamp the measuring, standards, testing and quality systems in order to increase compliance with international quality standards and regulations; increase productive innovations by strengthening the national innovation system and scaling up the efforts to promote private investments inR&D; overhaul the SME support programs in order to increase their impact, help supply chain integration and reduce barriers to contracting between large and small and medium enterprises; and put in place an export friendly environment through the elimination o f anti-export biases and the development o f export promotion policies. While many o f these issues are not easily resolved in the short run, addressing them over the medium run will enhance the prospects for sustainable growth andreduce poverty and inequality. xiv 1. INTRODUCTION 1.1 The economic collapse in Argentina culminating in 2002 was dramatic, resulting in the deepest political and economic crisis in generations. Few countries have experienced such economically catastrophic events: over the four-year period from 1999 to 2002, GDP fell by over 20 percent. The human costs o f the precipitous decline in economic activity have also been substantial, with the crisis resulting in severe social dislocations and a large reduction o f welfare for Argentina's population. The rate o f unemployment reached 21.5 percent o f the economically active population in 2002, up from 14.0 percent three years before. Simultaneously, the urban poverty rate shot up to 57.5 percent in October 2002 as opposed to 36.0 percent in May 2001, while extreme poverty increased from 11.6 percent to a high o f 27.5 percent during the same period. Income inequality, which had been rising steadily since the mid-1970s, peaked in the aftermath o f the crisis. Additionally, 160,000 people migrated from the country over the period 2000-02, compared to an average o f 12,000 people per year over the period since 1950. 1.2 Argentina's economy began to rebound strongly in 2003. Sparked initially by increased exports and by a gradual expansion o f consumption spending, a demand-led recovery emerged. The GDP growth rate reached 8.8 percent inthat year. Economic conditions remained favorable during 2004, with real GDP growing by 9.0 percent. Remarkably, the recovery o f the economy was supported by prudent macroeconomic policies. Inflation remained subdued with annual inflation attaining only 3.7 percent and 6.1 percent in2003 and 2004, respectively. The primary surplus for the federal government reached 2.3 percent and 3.9 percent o f GDP, respectively, in the same two years. The provinces also enjoyed a comfortable fiscal situation. In 2003 the primary surplus o f the provinces reached 0.9 percent o f GDP, while for 2004 the estimated primarysurplus was around 1.4percent of GDP. 1.3 The recovery has begun to have a positive impact on the living standards o f the population, as shown by the steady improvement in key social indicators since 2002. Nevertheless, the economic and social advances were from a very low base, and large segments o f the Argentine population continue to face great hardship. Unemployment rates dropped from the peak levels reached during the crisis, but remained at 12.1 percent by the end o f 2004 (16.2 percent if the beneficiaries o f the Heads o f Households Program -Plan Jefes y Jefas de Hogar- are classified as unemployed). In parallel, real wages in the private sector have been trending upward, although they are still, on average, below pre-crisis levels. The combined impact o f higher employment and real wages generated a decline in the poverty rate, which reached 40.2 percent in the last quarter of 2004 (down from the 47.8 percent observed a year ago). Inequality, even though it was reduced between 2002 and 2004, remains high and Argentina i s no longer among the three countries showing the lowest degree of inequality inthe region. The significant percentage o f the population still living inpoverty andthe high inequality in the distribution o f income underscore the extent of the social challenge confronting the country. 1.4 Sustaining the economic recovery and ensuring that it will benefit the poorest i s a matter o f top priority for the government. Infact, one o f the difficult tasks that Argentina faces today i s to implement policies needed to ensure that the country embarks on a path o f sustained and equitable economic growth. The sudden impoverishment experienced after the crisis has called (and still calls) for interventions that support the economic recovery, assist employment creation, and enhance the living conditions o f the population. The purpose o f this report is to contribute to the ongoing dialogue in the country on how to achieve these goals through efficient policy interventions. It i s expected that the study will provide useful analytical inputs for the Argentine government and that its results will help indesigning a strategy that delivers highgrowth within a framework o f social equity. 1.5 The report has been prepared inclose consultation with the authorities inthe Ministry o f Economy and Production and reflects the Government's priorities inexploring certaintopics. As a result, the study i s not a comprehensive treatment o f the sources o f growth in Argentina. In particular, it does not cover the important and complex issues related to the financial sector and institutional development which are being addressed in other studies. Rather, it attempts to deepen the analysis on selected topics, chosen in collaboration with the authorities, regardingthe inter-linkages between economic growth, income distribution and poverty, as well as the respective roles o f these factors in explaining the historical underperformance o f the Argentine economy. The objective o f the report i s to identify relevant issues for policy formulation and further economic work. Its emphasis is on longer-term structural factors which are thought to determine productivity and income distribution, not the short-term conjunctural challenges facing the country. The conclusions presented are not definitive, given the selectivity o f the topics covered and the limitations on the data and analytical techniques employed. Nevertheless, they hopefully can contribute to the policy debate andhelp set the agenda for future research. 1.6 The report is organized as follows. Chapter 2 briefly reviews the time-dynamics o f economic growth, poverty and inequality in Argentina. The first section highlights the fact that the country's economic growth was very low during the second halfo fthe twentieth century and that, within Latin American countries, only in Venezuela and Bolivia was growth lower on average. Moreover, evidence i s presentedshowing not only that growth has been slow but that there has also been substantial volatility in economic performance. The section notes that this high volatility may well have prevented higher growth rates, as shown by recent empirical studies. Finally, the section presents the implications o f recent cross-country studies for Argentina. Overall, these studies support the conventional wisdom that has been emerging with respect to growth economics over the past 30 years. More specifically, improving macroeconomic stability, facilitating private sector investment, strengthening the legal and institutional framework, having a well functioning system for financial intermediation, promoting trade expansion and greater openness o f goods markets, and extending human capital formation are crucial actions for achieving highrates o f economic growth. 1.7 The next two sections o f the chapter study the evolution o f poverty and inequality. Inequality rose persistently in the country during the last several years, resulting in steady increases in poverty in the face o f stagnant average - albeit widely fluctuating - per capita * output. Data presented in the second section show that poverty was higher in 2002 than ever before and, to reach this point, it followed a W-shaped path over time. The analysis presented in this section shows that poverty rates have always increased inperiods o f recession but that they have sometimes went up inperiods o f growth as well. The fact that poverty sometimes increased 2 when growth was taking place is exceptional in international experience. The major findings regarding inequality inArgentina are presented inthe third section. It i s noted that inequality has increased dramatically, though not monotonically, since 1990 and, notably, increases in inequality were observed inperiods o fboth growth and recession. 1.8 Chapter 3 reviews the latest episode o f economic volatility in Argentina, a period that started with the sharp decline o f economic activity in 2002 and continued with a recovery since 2003. Special attention i s paid to the impact o f this recent volatility on economic growth, poverty and inequality. The first section o f the chapter studies the social impact o f the recent crisis and recovery. The second analyzes how the poor shared the losses (benefits) o f the most recent recession (recovery). More specifically, it employs different methodologies to test whether the recent recovery was pro-poor. The third section presents an analysis o f the recession and recovery at the sector level, while the final section identifies the economic sectors that contributed most to poverty reduction duringthe latest upturno f the economy. 1.9 The key findings o f this chapter are as follows. First, and not surprisingly, aggregate growth during the cycle was closely, and inversely, linked to poverty changes. Second, growth had a larger impact on the average incomes o f the poor than on the average incomes o f the rich. That is, data indicate that the recent episode o f economic recovery was a pro-poor growth episode, not only inthe absolute sense but also inthe relative one. Third, the pattern o f growth in different economic sectors probably responded to changes in relative input prices after the devaluation, with the sectors that were more dynamic being also relatively more labor-absorbing. Fourth, those economic sectors that experienced the greatest growth (weighted by employment share o f the sector) also showed the highest contribution to poverty reduction during the recovery. 1.10 Chapter 4 then turns to a selected set o f government policies that could help to deliver high economic growth. In addition to considering how these policies could be effective in fostering growth, the chapter examines their likely effects on poverty and inequality. Since an analysis o f all the options currently available to the government is well beyond the scope o f this study, the chapter focuses on a limited number o f options in selected areas, namely: infrastructure, fiscal policy, trade policies, human capital formation (education) and private sector development. 1.11 The first section o f the chapter observes that, compared to other Latin American countries, rates o f infrastructure investment in Argentina are now much lower than the average. Therefore, a government policy that promotes increased infrastructure investment would be an effective mean of addressing potential bottlenecks in the economy. Such a policy would help sustain the current economic recovery, foster productivity and competitiveness, and reduce inequality. The second section notes that, in the face o f widespread poverty, there is an expectation that fiscal policy will play a major role in combating poverty and reducing inequality. The section highlights the idea that macroeconomic stability i s crucial to reducing poverty rates since the empirical evidence indicates that economic crises disproportionately affect the poor. Another major lesson fiom this section is that tax policy is not a good tool for redistribution in developing countries. Instead, expenditure policies should be used as the effective mechanism for confronting inequity and fighting poverty. The third section analyzes 3 the potential impact o f various trade initiatives on employment opportunities for relatively unskilled workers inArgentina. It i s expected that higher rates o f employment for these workers will have direct implications for poverty and income distribution indicators. The general lessonis that trade liberalization can potentially be beneficial for the country, particularly trade liberalization o f agricultural products under the terms o fthe WTO reforms. 1.12 The fourth section presents the results o f a recent study on the returns to education for workers across income groups. The heterogeneity between the poor and the non-poor in the compensation they receive for their education in the labor market is revealed, especially among women. The evidence o f increasing returns favoring the richer segments o f the population calls for policies that will improve the quality o f education and will enhance thejob search process for the poor. Finally, the fifth section o f the chapter reviews some key issues involved in ensuring a favorable environment for private sector development. For example, it i s found that the government needs to take actions that will facilitate a better business and investment climate; improve the productivity, the quality and product mix o f the country's goods; increase productive innovations; help supply chain integration; reduce barriers to contracting between large enterprises and small and medium enterprises (SMEs); and increase linkages to sources o f knowledge, technology and innovation. While manyo f these issues are not easily resolvedinthe short run, addressing them over the medium run will enhance the prospects for sustainable growth andreduce poverty and inequality. 1.13 The Argentine government has adopted a strategy which emphasizes rebuilding the economy with a view to delivering sustained growth with social inclusion, an appropriate goal in view o f the highdegree o fpoverty and inequality inthe country. It is important to recognize that growth i s not only essential, but that it i s needed over sufficiently long periods and with sufficiently pro-poor impact so that all segments o f the population can expect a notable improvement in their lives and their livelihoods. The text that follows presents an in-depth analysis o f topics that may shed light on how policymakers could build a stronger and more equitable economy inArgentina. 4 2. ECONOMIC GROWTH,POVERTYAND INEQUALITYIN ARGENTINA 2.1 Argentina was among the world's highest income countries in 1913.Historical data show that economic growth was clearly disrupted in all countries by the adverse shocks o f the Great War, the interwar economic strains, the Great Depression and World War 11. Yet other countries with comparable living standards and growth records before these events did manage subsequently to continue their trend growth. This was not the case for Argentina, where growth did not resume at anywhere near previous observed rates and economic slowdown set in beginning roughly around the time o f World War I.Displaying a low and volatile economic growth rate since then, Argentina, once an unquestionably advanced country, unexpectedly became relatively poor.' Figure 2.1: Per Capita GDP, Poverty and Inequality, 1974-2004 I 9000 8000 7000 6000 5000 4000 3000 2000 1000 0 1974 1978 1876 1880 1882 19&1 1888 1888 1980 1892 1894 1988 1896 2000 2002 2004 ...... Poverty(OficialGBA Percent) -Inequality (GiNI Coefficient) -Per Capita GDP (1995 USD) Source: World Bank staff. 'Argentine per capita income grew at 2.5 percent annually over the period 1870-1913, with GDP growing inexcess of 6 percent per year. Such high growth i s truly phenomenal, and was possibly unprecedented inworld history until then. However, the rate o f per capita income growth since the early years o f the twentieth century dropped to less than 1percent annually. 5 2.2 This is one great puzzle o f Argentine economic history. Unfortunately, there are no totally convincing simple answers to the question of what accounted for the languishing performance o f Argentine growth since World War I. Why the economy stagnated over the past 90 years i s a question for continued analysis and debate among economic historians. A conclusive answer to this question goes well beyond the scope o f the present study. Inany event, the country is a unique case o f unresumed growth, or arrested development. InRostovianterms, Argentina i s the only country that clearly reached the "take off' phase for economic growth, but then aborted its flight into sustained growth. 2.3 The last several decades have uncovered another important puzzle: inequality increased persistently, resulting in increased poverty in the face o f stagnant average -albeit widely fluctuating- per capita output. After peaking in 1974-80, per capita GDP declined by 22 percent during the 1980s, increased by 42 percent through 1998, declined by 22 percent through 2002, and recovered to the 1998 level by 2005 (Figure 2.1). All in all, per capita GDP in2004 was at about the same level as in 1974. Nevertheless, poverty was much higher in 2004, reflecting a long term increase ininequality. Available data since around 1980 show that Argentina's income inequality has increased over the years. Figure 2.2 maps the inequality and poverty data o f to show more clearly how increased poverty has been linkedto increased inequality. 2.4 This chapter briefly reviews the time-dynamics o f economic growth, poverty and inequality inArgentina. First, it presents a discussion o f the economic sources o f growth over the recent (and more readily measurable) past and the contributing factors to that growth. Next, the chapter reviews some o f the existing literature on the evolution of inequality and poverty in Argentina during the last decade and a half. The final part of the chapter deals with the international evidence on the relationship between poverty, inequality and economic growth and on the lessons that could be drawn for the country. Figure2.2: Mappingthe Relationshipof Inequalityand Povertv 55 50 .-E w s 45 a3 0 40 0 .-.-S 0 (3 3J 30 25 0 10 20 30 40 50 50 Poverty Headcount Source: World Bank staff. 6 A. ACCOUNTINGFORECONOMIC GROWTH The GrowthRecord:Lessons from History 2.5 Argentina's long-tenn growth record has been the subject o f a considerable amount o f research. Much o f this research has focused on the first puzzle mentioned above, namely, Argentina's decline from its economic prominence at the beginning o f the twentieth century. Several hypotheses have been advanced with regard to the timindreasons for the decline. Guido di Tella (1986) advanced the "early-retardation" hypothesis that growth declined after 1914 because the frontier had been closed by that time with the full occupation o f the Pampas. Di Tella and Zymelman (1967) further argue that poor policies also accounted for poor performance from 1914 to 1933. Carlos Diaz-Alejandro (1970) advanced the "late-retardation" hypothesis that Argentina's growth decline began with the onset o f the Great Depression in 1929. Inreviewing these hypotheses, della Paolera and Taylor (2003) note how recession during World War I triggered inward-looking economic policies, which Diaz-Alejandro regarded as discouraging export-led growth. Finally, Sanz Villarroya (2002), using unit root and structural break methods and with Canada as the benchmark, identifies 1913, 1929 and 1975 as years with structural breaks2. Figure2.3: Ratio ofArgentina's Per Capita GDP to Per Capita GDP*, 1900-2001 I 3 2 5 2 1 5 1 0 5 0 I 1-aaylacy -awesty --ayoffshotsy/ * OfWestern Europe (aywesty), FourWestern Offshoots (ayoffshotsy), and LatinAmerica (aylacy). Source: World Bank staff. As noted above, growth since the mid-seventies has been less than lackluster, allowing little ifany improvement in average economic conditions and with income dispersion widening. 7 2.6 Taylor (1992) makes the further point that the retardation view depends on the countries used as benchmarks. And indeed it does. Figure 2.3 shows comparisons with 12 Western European economies, four Western "offshoots" (the United States, Canada, Australia, and New Zealand), and Latin America. Comparison with Western European economies suggests an earlier relative decline (the early 1920s) than comparison with the four Western offshoots (the early 1930s). Comparison with Latin American economies shows a continued decline from the beginningo fthe twentieth century, at the least. 2.7 Argentina's low average growth continued into the second half o f the twentieth century. Table 2.1 presents per capita income growth data for the period 1950-2000 for a number o f countries. Over this period, Latin America's per capita income as a whole diverged downward vis-A-vis the industrialized countries as a group, and Argentina's per capita income converged downwards towards Latin America's per capita income. Within industrial countries there was a convergence trend, as the initially poorer countries (e.g., Japan, Ireland, Portugal) have made progress in catching up with the early leaders. Convergence with the richer countries was also observed for a number of Asian countries. Within Latin American countries as a group, Argentina stands out as a particularly slow growth country, with per capita income having grown at only 1.1 percent annually over the 50 year period. In only two Latin American countries (Venezuela and Bolivia) was growth lower. As a consequence, Argentina's relative position vis- A-vis the rest o f Latin America (and the industrialized countries) has deteriorated in the last 50 years. Table 2.1: AnnualGrowthRatesof GDP per Capita for Selected Countries, 1950-2000 (exmessed in Iercentages) DevelopingCountries IndustrializedCountries Argentina 1.1 Australia 2.1 Bolivia -0.02 Austria 3.5 Brazil 3.0 Belgium 2.8 Chile 2.2 Canada 2.2 Colombia 1.8 Denmark 2.3 Costa Rica 1.7 France 2.9 Ecuador 1.5 Ireland 3.7 Mexico 2.2 Italy 3.4 Paraguay 1.4 Japan 4.9 Peru 1.2 Portugal 4.0 Uruguay 1.2 Norway 2.9 Venezuela 0.2 Spain 3.8 South Korea 5.4 Sweden 2.3 Taiwan 6.3 United Kingdom 2.2 Thailand 3.9 UnitedStates 2.3 burce: Tyler (2005). 8 2.8 It is likely that scholars will continue to debate Argentina's growth puzzle for years to come. However, the debate so far, even if cast in terms o f "informal empiricism," already suggestsuseful growth policy lessons from Argentina's own history: 0 Trade Openness: Argentina benefits from trade openness and access to international markets. The work o f di Tella and Zymelman conflicts with that o f Diaz-Alejandro on the merits o f trade openness. The first source condemns trade openness as helping explain Argentina's "early retardation" (i.e., after 1913)' while Diaz-Alejandro's identifies the import substitutionpolicies that started in the 1930s as the explanation for "late retardation" (Le., after 1929). Taylor reconciles to some degree these opposing views by addingto the picture Argentina's exogenous loss o f capital market access after 1914, a serious shock to an open economy which i s critical to explaining "early retardation" in an environment o f low domestic savings. Chapter 4 discusses in detail trade policies inArgentina's current context. 0 Investment: There is a need to ensure an efficient allocation o f investment. To quote Taylor (2003)' "(t)he great tragedy o f the postwar era is that most o f the very generous supply o f savings was misspent on investment activities that were either overpriced or unnecessary or both," and "to look forward we can now hope that Argentine investment i s now entering a fourth phase when domestic and foreign savings are abundant and market allocation mechanisms once again operate effectively to channel precious resources for growth to their most efficient uses." Chapter 4 discusses this issue in Argentina's current context with reference to investment ininfrastructure and humancapital. 0 Macroeconomic Stability: Argentina's history shows, after World War I1and until the early 1990s: (i)high volatility o f nominal variables (money supply, inflation); (ii) a a higher volatility o f the fiscal balance; and (iii) a high volatility o f real monetary aggregates and o f the real exchange rate.3 This macroeconomic volatility may well help explain Argentina's lower growth performance, as suggested by cross-country econometric growth equations discussed later in this chapter. Chapter 4 covers Argentina's present macroeconomic stability challenges. 0 Public Sector Efficiency: Argentina's public expenditures increased sharply from 1934 (they represented 6.6 percent o f GDP) to 1949 (13.1 percent) and then from 1971 (12.7 percent) to 1981 (16.4 per~ent).~The negative effects o f the increasing burden o f the state in Argentina are discussed, for example by Berensztein and Spector (2003). Chapter 4 discusses Argentina's current public sector issues from the perspective o f bothtaxes and expenditures. 0 Property Rights: Prados and Sanz Villarroya (2003) uses broad money excluding currency as a proxy o f property rights in Argentina since the late nineteenth century, under the assumption that asset holders reallocate to currency when their property See Sturzenegger and Moya (2003). Time series data on government expenditure or consumption are particularly fragmentary, with a gap inthe early 1980s. 9 rights are under threat. Relating this proxy to historical developments, they find a weakening o f property rights from 1935 to about 1952 and a recovery through 1990, and verify the positive effects on growth o f stronger property rights inthe context o f a time series growth model. 2.9 Importantly, the focus o f these lessons is on variables that have been tested in cross- country macroeconometric work using evidence since the 1960s, including work discussed hrtherbelow. The Growth Record: Volatility 2.10 Another lesson from history i s the possible role o f volatility in discouraging growth and affecting p~verty.~Not only has growth been slow in Argentina but there has also been substantial volatility in economic performance and output. Business cycles are common in all economies, but the Argentine economy seems to be set apart when compared to those of other countries. This high volatility may indeed have contributed to slower growth. For example, volatility could hamper growth through its dampening effect on investment inboth physical and human capital. Recent empirical analysis, based upon country cross-section regressions, provides evidence that greater volatility inoutput growth consistently reduces the average rate o f growth.6 Similarly, empirical work focusing on LatinAmerica, employing paneldata for awide variety o f countries, has provided evidence that volatility has been a significant variable in explainingboth the time trend and the cross-country variation o f economic gr~wth.~ 2.11 Argentina's high output volatility relative to other countries and regions is demonstrated inTable 2.2. The first data column shows trend GDP per capita growth for the period 1960-99, with the trend computed using the Baxter-King bandpass filter detrending technique. The remaining data columns depict output volatility @e., the standard deviation o f the cyclical component o f output) for different time periods. For the entire 1960-99 period, Argentina possessed greater output volatility than any other Latin American country or any major region. Argentine volatility was greatest duringthe chaotic (for Argentina) decade o f the 1980s. For that period only Peru, with a still more chaotic decade, surpassed Argentina in terms of volatility. With stabilization and greater macroeconomic balance in the 1990s' Argentina's volatility was reduced but it still remained higher thaninmost other countries.* Chapter 3 analyzes growth and poverty during Argentina's last and very deep cycle. See Ramey and Ramey (1995), Hausmann and Gavin(1996) and IADB (1995). * Calderon,time Fajnzylber and Loayza (2002). . The data series end in 1999. Had the years 2001 and 2002 been included, the pattern of high volatility would appear accentuated. 10 Table 2.2: Trend Growth andOutputVolatility, SelectedCountries,1960-99 I Trend # Growth' Volatility of Output per Capita RegionsKountries Obs 1960-99 1960-99 1960-70 1971-80 1981-90 1991-99 IAll Countries 106 2.35 0.0224 0.0195 0.0263 0.0213 0.0149 Industrial Countries 22 2.52 0.0140 0.0133 0.0163 0.0138 0.0102 DevelopingCountries 84 2.18 0.0304 0.0254 0.0363 0.0285 0.0200 East Asia 12 4.33 0.0296 0.0232 0.0286 0.0257 0.0275 EastemEurope 3 2.17 0.0235 0.0122 0.0171 0.0198 0.0300 Middle East 9 2.83 0.0345 0.0435 0.0426 0.0270 0.0133 South Asia 5 2.03 0.0185 0.0219 0.0232 0.0119 0.0083 Sub-Saharan Africa 29 1.13 0.0350 0.0300 0.0464 0.0308 0.0224 Latin America & Car. 26 1.42 0.0290 0.0208 0.0356 0.03 14 0.0162 Argentina 0.88 0.0369 0.0350 0.0319 0.0475 0.0307 Brazil 2.29 0.0228 0.0163 0.0244 0.0283 0.0172 Chile 2.39 0.0328 0.0161 0.0456 0.0418 0.0205 Colombia 1.88 0.0127 0.0076 0.0137 0.0092 0.0188 Mexico 1.97 0.0215 0.0158 0.0174 0.0288 0.0252 Paraguay 1.59 0.0206 0.0109 0.0215 0.0324 0.0112 Peru 0.33 0.0360 0.0151 0.0185 0.0635 0.0307 Uruguay 1.14 0.0294 0.0190 0.0244 0.0465 0.0230 a. Average o f annual change (in%) o f the permanent component (trend) o f log output per capita, as estimated using the Baxter and Kingbandpassfilter detrending technique. b. As measuredby the standard deviation of the cyclical component o f the log o foutput per capita. Source: Tyler (2005). 2.12 The highvolatility ineconomic performance is not an exclusive characteristic o fthe most recent decades. Evidence from a longer time series analysis o f cyclical behavior (covering 1887- 1990) also indicates that Argentine economic volatility was generally higher than that of other countries. Applying the same Baxter-King bandpass detrending technique, Sturznegger and Moya (2003) compare output volatility with that o f a number o f other countries, including the United States, the United Kingdom, Canada, Australia and B r a ~ i lFor the 1945-90 sub-period . ~ Argentina witnessed substantially more output volatility than the United Kingdom, Canada, Australia and Brazil. Only in the case o f the United States did the levels o f volatility approach Argentine levels." Curiously, Argentine cycles do not appear to be correlated with those ofthese other countries, except with the United Kingdom during the period 1887-1914 and during the period 1919-39 with all five countries. AggregateGrowthAccounting 2.13 Productivity growth has long been recognized as a driving force in a country's sustained economic growth. The simplest measure of productivity -the value of output divided by labor inputs- enables easy cross-country comparisons, which clearly show that developing countries, including Argentina, have much lower productivity levels than the more advanced industrial countries. The problem with these simple labor productivity measurements i s that they do not capture the contribution o f the non-labor factors o f production and, therefore, overstate the 9Sturzenegger and Moya (2003). loDuring the World War I1period (1939-45) the opposite was the case -i.e., volatility was greater inthe comparison countries (all combatants) than inArgentina, except for Brazil. Butthis period must be considered atypical. 11 contribution o f labor alone. More complete work that includes non-labor factors, especially capital, focuses on the concept o f total factor productivity (TFP). 2.14 The analysis o f factor accumulation and productivity changes over time can provide insightsto a better understanding o f economic growth. Ingeneral terms, the growth o f output per capita can be expressed as a function o f three major determinants: (i) physical capital deepening; (ii) capitaldeepening; and(iii) human productivity growth. A qualification, however, relates to causality. While the common assumption i s that factor accumulation and productivity growth drive output growth, it can be argued that economic growth may itself push accumulation and productivity increases. Thus, factor accumulation and productivity have an important element o f endogeneity. A broader understanding o f this type o f endogeneity i s central to the analysis o f economic growth. 2.15 The common analytical framework for undertakingthe analysis o f the sources o f growth i s a neoclassical production function, most frequently a Cobb-Douglas. Underlying assumptions include constant returns to scale (whereby the income o f the productive factors exhausts output) and competitive conditions (with the relative prices of the productive factors determined by their marginal products). Since factors o f production possess decreasing returns in the neoclassical framework, technological progress plays a key role in output expansion. While Hicks neutral technological progress i s generally assumed, endogenous technological progress can be incorporated in the model. In any case, abundant empirical work (so-called growth accounting) conducted over the past several decades -stemming from the seminal Solow work- has generally emphasized the importance of productivity growth. Indeed, in most instances TFP growth is seen to dominate factor accumulation as a "source" o f economic growth. 2.16 In the case of Argentina, there have been a large number of recent growth accounting studies which have sought to measure TFP growth since the 1940s. Table 2.3 summarizes the results found by some o f them. There is a certain commonality in the major results o f these analyses. First, for the longer periods o f time for which TFP i s estimated (three to five decades), TFP growth i s generally seen to be low, hardly more than 0.5 percent per year. For example, Fajnzylber and Lederman (1999) estimate its contribution to GDP growth at the average for Latin America (0.2 o f a percentage point)." Second, TFP growth rates are quite erratic, reflecting Argentina's high volatility noted above. Third, the average TPF growth estimates are typically negative for the chaotic 1980s and most positive for the 199Os.l2 Fourth, and as expected, when attempts are made to incorporate human capital accumulation, the estimates o f TFP growth decrease appreciably. Indeed, the average for 1960-2000(-0.3 percent) calculated by Calderon, Fajnzylber and Loayza (2002), suggests that there may not have been any net TFP growth duringthis period. " Inan analysis basedoninternational comparison, estimates byHalland Jones (1999) place Argentina lagging Brazil and Mexico in productivity levels, although Argentina's levels still exceed those o f most other Latin American countries. l2 One recent study (Coremberg, 2004) estimates significantly lower -and even slightly negative- rates o f TFP growth for the 1990s. The difference lies in Coremberg's concept and measurement o f capital services, which he argues grew more rapidly inthe 1990sthan other studies have indicated, 12 Table 2.3: RecentStudies of TotalFactorProductivity(TFP) Growth HumanCaDital Adjusteh Authors 1 Period I I TFP Growth (Annual%) I TFPGrowth (Annual `YO) Gomes, Pessoaand Veloso (2003) I 1950-2000 1 -0.1 Elias (1992) 1940-90 0.5 1940-50 3.1 1950-60 0.8 1960-70 0.2 1970-80 -0.3 1980-85 -2.6 De Gregorio and Lee (1999) 1960-90 -0.5 1960-69 0.7 1970-79 0.6 1980-89 -2.6 Kydland and Zarazaga (2002) 1950-70 1.4 1970-79 -0.3 1979-90 -2.1 1991-97 5.8 HopenhaynandNeumeyer (2001) 1949-70 -1.1 1971-80 -1.8 1981-90 -1.8 1991-97 3.1 Maia andNicholson (2001) 1962-64173-7'5 2.2 1973-75189-91 -1.3 1989-91196-98 3.5 Ministerio de Economia (2004) 1960-74 1.8 1975-90 -1.6 1991-2000 2.4 Guerson and Vasquez (2002) 1960-2000 0.4 -1.5 1960-80 0.8 -1.9 1981-90a -1.9 -3.4 1991-2000a 3.1 2.6 Calderh, Fajnzylber and Loayza (2002) 1961-70 0.96 0.21 1971-80 0.24 -0.46 1981-90 -2.43 -3.31 1991-2000 3.05 2.49 a. Estimates using employed labor, rather than total labor, cover the years 1981-2000. 2.17 The recent estimates o f capital, labor and TFP growth in Argentina for 1960-2000 indicate the close relationship o f growth performance to TFP performance. Table 2.4 reproduces estimates by Calderh, Fajnzylber and Loayza (2002). While economic growth during the 1960s was reasonably robust with GDP growing at 3.9 percent, factor accumulation accounted for most o f this growth. Capital accumulation was particularly important, accounting for 52 percent o f the observed growth. TFP growth -after adjustment for human capital increases- was insignificant. The 1970s reveal a similar story. Physical capital accumulation was responsible for 65 percent o f observed growth, with the TFP contribution to growth being negative during the period. The 1980s were characterized by negative GDP growth attributed to a number o f adverse economic shocks, both external and domestic in origin, including the debt crisis, the Malvinas war and growing macroeconomic imbalances. Net investment was practically nil, and TFP growth was negative. For the high growth 199Os, capital accumulation i s seen as less important than previously &e., the 1960s and 1 9 7 0 ~ )while TFP growth seems to be dominant in ~ accounting for the decade's strong economic expansion, even when adjustments are made for 13 human capital accumulation. A pertinent question i s why TFP growth was so high in some periods. From the policy perspective, it should be noted that cross-country research links TFP growthto policy reform. Fajnzylber and Lederman (1999), by comparing periods o f reform with periods o f no reform in Latin American economies, conclude that TFP growth was more rapid during periods of reform than during periods o f no reform, with period classification based on barriers to imports, the black market exchange rate premium, and the presence or absence o f state export monopo~ies.'~ Table 2.4: Growth AccountingAdjustingfor Human Capital, 1961-2000 (in percentages) 1961-70 1971-80 1981-90 1991-2000 GDP 3.88 2.95 -1.50 4.57 Labor 1.66 1.51 1.78 1.64 (% contribution) (43%) (51%) (-119%) (36%) Capital 2.00 1.91 0.03 0.44 (% contribution) (52%) (65%) (-2%) (10%) TFP 0.21 -0.46 -3.3 1 2.49 (% contribution) (5%) (-16%) (221%) (54%) ~ Source: Calder6n et al. (2002). B. DETERMINANTSECONOMICGROWTH OF 2.18 TFP growth is of course only one o f the components of GDP growth, with investment and labor growth also contributing from the growth accounting perspective. Investment is likely to respond to economic growth and growth prospects, as suggested by the rise in private investment during the 1990s and its decline during the late 199Os.l4 Similarly, labor is likely to respond to economic growth, as suggested by Solimano (2002), who concludes that "the two most important explanatory variables for net international migration to Argentina in the 20th century (are) the income per capita differential o f Argentina vis-&vis other source economies for migrants and Argentina's political regimes." 2.19 Hence, the important question from the policy perspective is on the factors that determine growth: whether it takes place through increases in capital (human or physical) or through productivity. The late 1980s and 1990s witnessed a revival inthe interest in and analytical work on economic growth. This resurgence covered both theoretical work and empirical analysis. Advances in both areas have been inter-related and have helped improve communication l3 See also Tyler (2005) for a discussion o f the issue. The author presents some suggestive evidence for Argentina that may be offered by the Brazilian case, which also experienced highrates o f TFP growth inthe 1990s. l4As argued originally by Diaz-Alejandro (1970), the integrity o fpast investment data inArgentina is compromised by price distortions which created a large premium of domestic investment goods prices over international prices, Tyler (2005) argues that this premium was reduced with reforms during the 199Os, to the effect that investment prices have now converged back to international levels. 14 between theorists and empiricists employing modern econometric methods. As a result o f these parallel but related advances, a better understanding of the sources o f economic growth and growth process'eshas emerged. 2.20 The empirical framework in these studies i s mainly derived from the underlying neoclassical growth model, as extended and complemented by more recent endogenous growth models. The principal questionposed inthese cross-country analyses i s simple enough. Why do growth rates vary across countries? In other words, what explains growth? Related to this question are others related to economic policies. What policies are likely to be effective in increasing growth? H o w do institutions have an impact on the growth process? 2.21 A recent cross-country empirical study by Calderh, Fajnzylber and Loayza (2002) has produced useful results in the tradition o f the recent work referred to above. This study has important implications for Argentina, which i s included among the studies covered. The study focuses on the determinants o f growth, employing cross-country analysis and regressions. The estimated basic regression, commonly used by other cross-country growth studies, takes the following form: where y i s the log o f output per capita, yTis the trend component o f output per capita, the expression (Yi,t-I -YTi,t-l)represents the output gap at the beginning o f the period, Z i s a set o f other initial state variables, X i s a set o f control and environment variables postulated as growth determinants, and E i san error term inthe form o f the regression residual. The authors introduce a time specific effect, pt,to allow for the introduction and control o f changing intemational conditions which affect the growth performance o f countries for the time period. They also include a country specific variable, vi,to capture factors that affect country growth and are possibly correlated with the explanatory variables. With normalization o f the time period to 1, the expression on the left-hand side o f the equation is the growth rate o fper capita income for the period under ana1y~is.l~ 2.22 Two major categories o f explanatory variables are analyzed. They are: (i) state initial variables, and (ii) control and environmental variables. Initial state variables include the log o f per capita income at the beginning o f the period, the initial output gap and the stock ofphysical and human capital at the beginningo f the period under analysis. The control and environmental variables comprise those actions chosen andpursued by governments or private agents. They can reflect government policies, institutions or private preferences. These control and environmental variables contain subcategories o f (i) structural variables, (ii) macroeconomic variables, and (iii) external environment conditions variables.I6 2.23 The estimates obtained by Calderh, Fajnzylber and Loayza (2002) can be used to examine the Argentine experience. Specifically, the parameter estimates from the empirical analysis can be used to simulate changes in GDP per capita for the country for the different l5Estimation o f Equation (1) i s undertaken over a wide range o f countries with averages of discrete time periods (fenerally five or ten years) employing OLS andor GMM(General Method of Moments) procedures. In order to deal with the likely endogeneity of the explanatory variables, lagged values of those variables have been used as instruments. 15 periods covered by the study. For Argentina, as shown in Table 2.4, the projected changes between decades rather closely matched actual changes, although the projected changes understated the swings in observed perfonnance for all three time period comparisons, a reflection o fArgentina's highdegree o f volatility notedpreviously. 2.24 The 1980s versus the 1970s. Comparisons between the 1980s and the 1970s (the last column inTable 2.5) suggest that a deterioration inmacroeconomic policies was a major force in negatively affecting growth. The inflation rate increased -reflecting macroeconomic imbalances- and there was volatility in policy performance and output. A deterioration o f the commercial banking sector, as reflected by bankingcrises and related strains, also had a negative effect on growth. The only macroeconomic policy variable having a positive effect on growth was a reductioninthe chronic exchange rate overvaluation o fthe 1970s. Table 2.5: Determinantsof Argentina's GrowthRateof GDP Per Capita, by Decades 1990svs. 1990svs. 1980svs. GrowthDeterminants 1980s 1970s 1970s Initial State Variables 1.85 1.08 -0.38 InitialGDP per capita (inlogs) 0.15 0.16 0.01 Initialoutput gap (log[actual GDP/potential GDP]) 1.70 0.92 -0.39 Structural Policies and Institutions 1.07 1.67 0.60 Secondary enrollment (inlogs) 0.24 0.64 0.40 Private domestic credit/GDP (inlogs) 0.07 -0.04 -0.11 Structure adjustedtrade volume/GDP (inlogs) 0.56 0.80 0.24 GovernmentconsumptiodGDP (inlogs) -0.17 -0.35 -0.19 Maintelephonelines per capita(inlogs) 0.37 0.62 0.26 Stabilization and Macroeconomic Policies 1.73 0.12 -1.60 Inflationrate (inlog[ l+inflation rate]) 0.90 0.31 -0.59 Standarddeviation of output gap 0.37 -0.07 -0.44 Index of real exchange rate overvaluation (inlogs) -0.40 -0.09 0.30 Frequencyof years underbanking crisis 0.86 -0.03 -0.87 External Conditions -0.17 -1.67 -1.51 Growthrate ofterms of trade 0.3 1 0.05 -0.26 Period shifts -0.48 -1.72 -1.25 Projected Change: 4.45 1.18 -3.27 Actual Change: 6.71 2.40 -4.32 Source: Tyler (2005). 2.25 Structural policies and institutions had a positive effect on growth during the 1980s as compared to the 1970s, although the experience with the individual variables is mixed. Positive influences were exercised by human capital improvements (secondary school enrollments), expanded trade openness, and infrastructure improvements (telephone line coverage), which more than offset the negative effects coming from problems in the banking sector (credit reduction and crisis). Finally, a worsening inthe external terms o f trade had a negative impact on growth, although this was not as detrimental as the worsening o f overall world economic conditions for Argentina, with the 1980s being characterized by the Third World debt crisis and the accompanying reductions ininternational capital flows. 2.26 The 1990s versus the 1980s. As was the case for most Latin American countries, the 1980s was a lost decade for Argentina, with an acute external debt crisis, negative growth, 16 political turmoil andmacroeconomic instability. The economy suffered hyperinflationand severe recession in the late 1980s. Therefore, coming out of the economic downturn contributed in an important way to the positive growth rate in the 1990s (as seen in the initial output gap variable in the first data column in Table 2.5). Moreover, stabilization policies were successful in reducing inflation, which in turn was estimated to have made a significant contribution to increasing per capita growth relative to the 1980s. The more consistent macroeconomic policies also reduced output volatility, which in turn had a beneficial effect on growth. Improvements in the banking sector were also apparent. The only macroeconomic variable estimated as having had a negative effect was that reflecting exchange rate policy. Currency overvaluation increased, not only discouraging growth inthe 1990sbut also setting the stage for the 2001-02 crisis. 2.27 Structural policies and institutional improvements also positively influenced growth in the 1990s as compared to the 1980s. More open trade policies, infrastructure and human capital formation also contributedto higher growth. Financial sector deepening, as measuredby the ratio o f private domestic credit to GDP, exercised a positive, though small, contributory effect on growth. Under the structural policies variable grouping in Table 2.5, the only variable consistently exercising a negative influence on growth was government current expenditures - generally indicative o f the burden o f the public sector, In addition to the generally favorable influence o f macroeconomic policies and structural reforms, improved terms o f trade also helped to increase growth. 2.28 The 1990s versus the 1970s. The comparison o f the 1990s with the 1970s is similar to the comparison o f the 1990s with the 1980s. The growth impact o f structural reforms, however, i s greater and these reforms have had a longer period to demonstrate their effects. This indeed appears to be the case with trade policies, infrastructure improvements and human capital formation. 2.29 Chapter 4 discusses Argentina's options with regardto policies with a bearing on growth, with possible options covering both structural and macroeconomic policies. The chapter will focus chiefly on infrastructure, fiscal, trade, human capital, and private sector development policies. c. POVERTY ARGENTINA IN 2.30 The growth performance and policies discussed above are both key considerations o f poverty reduction performance. With disappointing growth over the past 90 years, a question arises as to its incidence. In other words, what happened to the distribution o f income and to poverty levels during this period? Some partial evidence is available in income distribution estimates by Galiani and Gerchunoff (2003) based on employee and household data from 1953 to 1999. The data show that the distribution o f wage income was almost invariant until the mid- 1970s, when it beganto become more uneven. For example, the income share o f the tenth decile was 25.6 percent o f the total in 1953 and 25.2 percent in 1974. It then rose to 34.2 percent in 1989, and dropped to 31.O percent in 1999. Regardingpoverty, Figure 2.4 shows the evolution o f the poverty headcount ratio for Greater Buenos Aires for a more recent period. An increasing trend o fpoverty can be seen, with two sharp peaks, one in 1989 and the other in2002, both, not surprisingly, crisis years. 17 2.31 Background work for this report17 analyzes the evolution o f the poverty rate and inequality in Argentina in the last 25 years." Duringthis period, Argentina underwent a major economic restructuring. The economy moved from having a large public sector, high inflation, andhighprotection o f trade and labor, to privatizing state-owned enterprises inoil, transport and telecommunications and achieving low inflation and low levels o f trade protection. After 1999, increased financial distress led the economy into recession. Accordingly, it i s appropriate to divide the record o f the past decade and a halfinto two periods: 1989-99 and 1999-2002. 2.32 The first period started with the Menem Administration in 1989. Growth was negative and unemployment peaked in 1990 in the aftermath o f the hyperinflation. The poverty rate in Greater Buenos Aires grew from 8 percent in 1985 to 41.2 percent in 1990. After that, efforts to restructure the economy were successful in restoring economic growth until 1998." The stabilization program was accompanied by a decrease in the poverty rate to 22 percent in 1994, but the one-digit poverty rates observed in the previous decade were not reached again. "SeeShnchezPuerta(2005). Inthiswork, povertyismeasuredusingthePermanentHouseholdSurvey (EPH), a (initially) semi-annual household survey covering about 30,000 urban households. The survey i s representative o f 71 percent o f urban areas. Since the share o f urban areas inArgentina is 87.1 percent, the sample o f the EPH represents around 62 percent o f the total population o f the country. The survey leaves out all o f the rural population, much o f which probably has a higher than average poverty rate. 18 Official poverty lines are set by INDEC. They consist o f the basic food basket and non-food consumptionbundle whose combined values are just sufficient to allow a typical individual to achieve a minimum level o f material welfare. The basic food basket was constructed based on food consumption patterns observed in the Survey o f Income and Expenditure administered for Greater Buenos Aires in 1985-86. l 9Between 1991 and 1998, the average annual growth o f real GDP was 6.4 percent. A negative annual growth rate of GDP was only observed in 1995: -2.8 percent. 18 Unemployment and poverty increased further afier the onset o f the Tequila Crisis in 1995, with poverty amountingto 29 percent in 199S2' despite the economic recovery o f 1996 to 2.33 The second period started in 1999, a year inwhich the economy entered into a prolonged recession, coupled with three years o f deflation. The period ended with the collapse o f the Convertibility Plan, a default on extemal debts, and a major exchange rate devaluation inJanuary 2002. As the economy stagnated and the financial and economic crisis worsened, poverty climbed to 37 percent inOctober 2001. However, the biggest increase inpoverty took place with the collapse o f the economy at the end o f 2001. The poverty rate inOctober 2002 reached a high o f 57.5 percent. Few countries in the world have seen such a rapid rise inpoverty -an increase inthe poverty rate ofmore than 50percent inonly one year. 2.34 The most striking fact is that poverty increased even in periods o f economic expansion. In a report documenting the socioeconomic situation in Argentina, CEDLAS (2004) presents a picture o f the evolution inthe number o f poor people inyears o f growth and years o f recession. It notes that between 1992 and 2003, 11million individuals became poor: "3.5 millions entered poverty during the economic growth period o f the 1990s, another 3.5 millionjoined that group in the first phase o f the recession (1998-2001), while about 7.5 million crossed the poverty line during the crisis 2001-02. The last economic recovery substantially reduced the number of poor inaround 3.3 million individual^."^^ 2.35 There are few cases where poverty has not risen during recessions and fallen with growth. In a study using evidence from Latin America, Morley (1998) finds that poverty is highlycorrelated with income and that poverty reductiondepends inlarge measure on economic growth. For example, inthe 1980s during the debt crisis, total poverty inthe region rose by about 17 percent in response to a 10 percent reduction in per capita income. The author remarks that recovery in the 1990s has reduced poverty though not by enough to compensate for the losses during the previous decade. CEDLAS (2004) also notes that although the region has not been very successful in fighting poverty, the record of most o f the Latin American countries is much better than the Argentine performance. While Argentina was a low-poverty country by Latin American standards (as measured by an internationally comparable poverty line in the early 1 9 9 0 ~ )it~ does not currently belong to this group (Figure 2.5). Moreover, even though the economy has shown signs o f recovery inthe last two years, per capita disposable income i s still at lower levels thaninthe 1990s. 20Real GDP growth rates during those years were 5.5 percent, 8.1 percent and 3.9 percent, respectively. 2' Iti s interesting to note the regional dimension to poverty changes inthe 1990s. According to Damill et al. (2002), inthe period 1990-94 poverty fell by48 percent, butit fell by only halfofthat percentage inthe two poorest regions, the Northeast and Northwest. In contrast, in the period 1994-98 the increase in poverty occurred mostly in the Greater Buenos Aires, Cuyo and Pampeana Regions. 22 Cruces and Wodon (2003) also find that from 1995 to 2002 poverty increased substantially inArgentina. In an attempt to measurethe chronic and transient components o fpoverty, they incorporate dynamics into the analysis and find a higher share of the poor being inchronic as opposed to transient poverty. 19 Figure2.5: Povertv Headcount Ratio inLAC Countries 80 7 0 60 5 0 4 0 30 20 10 0 A r o u n d 1990 1 0 0 8 0 6 0 4 0 2 0 0 ource: CEDLAS (2004) 2.36 Changes in poverty are related to economic growth and to changing inequality. First, if the dispersion o f income is held the same, the faster the rate o f economic growth is, the larger will be the reduction in poverty. Second, for any given growth rate, the more dispersed the distribution is, the smaller will be the reduction in poverty.23The relative importance o f these factors for poverty reduction can be gauged using the methodology laid out inKakwani (1994). The analysis is conducted as follows. There are two income distributions, say, for periods t and t+l. These will differ, in general, in their means and in their inequality. They will also, in general, have different values of any given poverty index, P(t) and P(t+I). The exercise then constructs a synthetic intermediate distribution which has the mean o f the t+l distribution but the spread o f the period t distribution. This synthetic distribution has a poverty number denoted by P(*) associated with it. Then, the change from P(t) to P(*) is referred to as the "growth component" o f the overall poverty change between the two periods, and the change from P(*) to P(t+l) as the "redistribution component." 23For empirical evidence on the effects of growth on poverty, see Deininger and Squire (1998), Dollar and Kraay (2002), Bourguignon (2003) and A d a m (2003). For the consequences o f higher inequality o n poverty, see Fields (2005). 20 2.37 Table 2.6 presents the results reported by Damill et al. (2002). These authors decompose the changes in poverty by separating the "growth effect" and the "distribution effect" for Argentina. They divide the 1990s into two periods. The first period i s from 1991 to 1994, in which they argue that the poverty index o f income per capita decreased 10 percentage points, almost only because o f the "growth effect.'' The second period, from 1994 to 2000, i s characterized by an increase inpoverty due more to the "distribution effect" than to the "growth effect," even though both effects were significant. In another study focusing on the case o f Argentina during the 1990s, Busso et al. (2004) decompose the change inpoverty into two parts following an altemative decomposition proposed by Mahmoudi (2001). These authors analyze the changes inpoverty for the periods 1992-98 and 1998-2002. Higher income inequality proves to be more important than changes in income in explaining the increase in poverty inthe period 1992-98. However, the opposite i s observed from 1998 to 2002 -Le., the "growth effect" is larger than the "distribution effect." Table 2.6: Decompositionof the Poverty-Index Variations (in percentages) I Period Total change Growth effect Distribution effect Residual I 91-94 -10 -10 -1.3 1.3 94-00 8.9 3.2 6.3 -0.6 91-00 -1.1 -5 4.5 -0.6 Source: Damill et al. (2002). 2.38 The main conclusions o f a decomposition analysis o f this kind, which has also been applied to other developing countries inAfrica, Latin America, and East and Southeast Asia, are that poverty variations are largely determined by economic growth andthat changes ininequality are o f secondary importance in the great majority o f cases. Argentina is one case that differs from the generality o f countries in that, in some periods, the "distribution effect" turns out to be more important thanthe "growth effect" inexplaining changes inpoverty.24 2.39 However, caution needs to be exercised regarding the policy implications o f these decompositions. Lustig and Kanbur (1999) wam against jumping to the conclusion that there is an implicit separability between the two components and that, somehow, the growth component couldbe accomplished independently o f the redistribution component. Ravallion (1997) provides a good example o f the complementarity between the two. He argues that inequality can be high enough in a country to result inrisingpoverty, despite good underlyinggrowth prospects at low inequality. Along the same lines are the conclusions reached by D e Janvry and Sadoulet (2001). They use data for 12 Latin American countries for the period 1970-94 to analyze the role o f aggregate income growth on changes in urban and rural poverty and inequality. These authors conclude that income growthturns out to be more effective inreducing poverty and inequality if 24Incountries' actual experiences, it has proved far easier to generate economic growth than to change inequality. For the case o f Argentina, Busso et al. (2004) try to answer how poverty can be reduced by various combinations of growth and reductions in inequality. They found that decreasing inequality was fairly easy to achieve. However, their simulation results also show that, for some provinces in Argentina, no change in inequality alone would be sufficient to achieve the goal o f reducing poverty by half. 21 the initial levels o f inequality and poverty are not too high and if educational levels are sufficiently high. 2.40 Insummary, the major findings about poverty inArgentina since 1990 are as follows: 0 Poverty in2002 was the highest on record. 0 Poverty's path up to this point has been markedby sharp peaks linkedto crises. Poverty has always risen in periods of recession, but it has also increased in some periods o f growth. The fact that when growth took place poverty sometimes increased i s exceptional in international experience. The evidence is mixed regarding whether the "growth effect" or the "distribution effect" i s more important in explaining changes in poverty in the 1990s. Sometimes the "distribution effect" has been more influential than the "growth effect" in determiningchanges inpoverty, a fact that is also unusual ininternational experience. D. RECENTTRENDSININEQUALITY 2.41 The previous section presented evidence that poverty has substantially increased in Argentina in recent years. Poverty is a concept that refers to the income distribution below a certain threshold. As such, poverty rates can increase after a shift o f the entire distribution to the left, and/or after an increase inthe dispersion o f the income distribution. However, inArgentina, mean income has fluctuated around a constant trend in the last 30 years. With no changes in income distribution, constant mean income would have implied stable poverty. However, the income distribution became substantially more unequal over the last 30 years, driving poverty up. Figure2.6 shows the evolution o f the Gini index for Argentina over time. It can be seen that inequalityincreased dramatically, thoughnot monotonically, inthe country. Remarkably, the rise in inequality was observed in periods of both growth and recession during the 1 9 9 0 ~As ~ . ~ indicated before, this rise ininequality i s another puzzle o f Argentina's economic history. 25Inequality can be calculated by different measures: shares o f each decile intotal income, income ratios or indices, such as Gini, Theil, Atkinson and Generalized Entropy indices. Inthis section, the analysis will be based on income shares and Gini indices o f household per capita income, 22 I Figure2.6: GiniCoefficientofHouseholdPer Capita Income -+- Gran BuenosAires +Total Pais cSource: Sanchez Puerta (2005). 2.42 Damill et al. (2002) review the changes in inequality among the active population. They report that in the period 1991 to 1994 there was a decrease in inequality among households. In contrast, they find that inequality increased from 1994 onwards, a result that they mainly explained by the observed rise in the unemployment rate. This finding regarding the worsening inthe distributionofwage income duringthe second part ofthe 1990s is also confirmed by the World Bank (2000): "looking at the data only since 1994, we find virtually substantial declines inthe real wages of skilled andunskilledworkers, while professional workers have continued to show gains." 2.43 Similar results are also found in a research paper by CEDLAS (2004). This study notes that the share of income received by the lowest 20 percent has declined since 1990 when they received 4.6 percent o f total GDP. By 2002 they obtained less than 3 percent o f total GDP. Conversely, the upper 20 percent increased their share in that period from 51 percent to 57 percent o f total GDP. This study shows as well that, between 1990 and 2002, the Gini coefficient increased from 0.46 to 0.53. According to this work, there were only few years in which this coefficient decreased: 1992-93, 1998-99 and 2002-03. In conclusion, inequality in Argentina, measured by the Gini coefficient or, alternatively, with the lower and higher 20 percent income shares, has increased steadily, although not monotonically, since 1990. 2.44 As inthe case o f poverty measures, inequality measures for Argentina inthe last decade deviated from the trend o f the rest o f the countries inLatinAmerica. The changes ininequality in Argentinahave beenmuch larger than inother countries. The country used to be one ofthe three most equal countries in Latin America, along with Venezuela and Uruguay (Figure 2.7). The existence o f a large middle-class was a distinctive feature o f the economy. Unfortunately, Argentina i s now displaying a degree o f inequality that i s more similar to the average in the region. 23 Figure2.7: GiniCoefficientinLAC Countries Early 1990s 60 55 50 45 40 3 2 g 8 2 p 8 8 9 -$ g z f a g Z !3 z ; : E l ? 7g $ g i ;2 ~ 5~ 8 n g m 0 0 2 I I Early 2000s 60 I Source: CEDLAS (2004). 2.45 But, what has produced the positive trend in inequality over time? In order to better understand the nature o f income inequality in Argentina, Sanchez Puerta (2005) applied a decomposition methodology to assess the relevance (distributional effect) o f the following factors: (i) changes inthe relative returns to higher education and experience; (ii) changes inthe endowments o f unobservable factors and their remunerations; (iii) changes inhours o f work and employment status; and (iv) the transformation o f the educational structure o f the population. As indicatedbefore, income inequality inArgentina has fluctuated around an increasingtrendwhich began in the mid-1970s. Focusing on two very different periods -1986-1992 and 1992-1998- and using data from the Permanent Household Survey (EPH) o f Greater Buenos Aires, the analysis indicates that the first period shows almost no change inthe Gini coefficient. However, in the second period there was a dramatic change in inequality, even though Argentina was experiencing macroeconomic and political stability, economic growth, and an absence o f natural disasters. 2.46 A second finding is that the small change in inequality between 1986 and 1992 was the result o f mild forces that offset one another. In contrast, for the rise o f inequality between 1992 and 1998 nearly all effects listed above acted in the same direction. The increase in the relative retums to higher education, in the retums to unobservable factors, and the relative decline in 24 hours o f work for unskilled workers were particularly important. Surprisingly, the dramatic changes inthe educational structure,26the decreases inthe gender wage gap and the increases in the unemployment rate appear to have hadonly a mildeffect on householdincome inequality. 2.47 Menendez and Rozada (2002) also analyze the sources o f changes in income inequality for the period 1991-2001. Their paper aims to understand how the behavior o f the labor market affected income inequality. For this purpose, these authors try to assess the impact on income inequality o f changes in such variables as the rates o f return to individual socio-demographic characteristics, the labor force participation rate, the rate o f unemployment, and the formal education o f the participants inthe labor force. They conclude that unemployment accounted for a large part o f the increasing inequality during the 1990s. In addition, they find that changes in labor force participation worsened income distribution from 1996 onwards. 2.48 Galiani and Sanguinetti (2003) study whether the rapid and deep process o f trade liberalization played a role in shaping the Argentine wage structure during the period. They find some evidence suggesting that those sectors inwhich import penetration deepened were also the sectors in which, ceteris paribus, a higher increase in wage inequality was observed. However, the same authors acknowledge that trade deepening can only explain a relatively small portion of the observed rise inwage inequality within sectors. 2.49 The links between the integration o f Argentina into the world economy and the growth o f income inequality were also studied by Gasparini and Bebczuk (2001). These authors find that increasing integration into the global economy was a significant determinant o f the increase in inequality through two channels. The first channel was the shift towards production and employment in the sectors using more intensively such factors as natural resources and skilled workers, whose relative returns increased as a consequence. The second channel was the fall in the relative price o f capital and the introduction of new "skilled labor" intensive technologies, both o fwhich raisedthe skill premium. Gasparini and Bebczuk also identify other factors behind the increase in inequality: the decline o f labor union power, the fall in the minimumwage, the increase in unemployment, the increase in the relative size o f the poorest families, and the increase inthe dispersion o f pensionpayments. 2.50 In summary, the major findings regarding inequality in Argentina since 1990 are the following: Inequality has increased dramatically, though not monotonically, inthe period. There were increases ininequality inperiods o fboth growth and recession. 0 Different factors that help explain the change in income inequality during the 1990s are: o the increase inthe relative returns to higher education o the relative decline inhours o fwork for unskilledworkers o the increase inunemployment rate o trade liberalization o the fall inthe relative price o f capital o the introduction o f new "skilled labor" intensive technologies 26For example, there was a large reduction in the proportion of youths and adults with only a primary education and a significant increase inthose with a college education 25 o the fall inthe minimumwage o familysize o the dispersiono fpensionpayments. 2.5 1 Once again, the second key puzzle covered by this chapter -the reasons for the increase in inequality that led to more poverty- i s likely to continue receiving attention. Some of the factors that have been identified as causing inequality, including increased relative returns to higher education, trade liberalization, and the fall in the relative price o f capital, may also have contributed to growth. This brings us back to the cautionary words regarding the limited separability o f growth and distribution. It should be recognized that growth enhancing measures may have distributional effects with an impact on poverty that may well require complementary policy responses. More specifically, two key policy lessons are: (i) the need to respond to skills needs, a response that could help support growth and improve distribution; (ii) need to the improve the social safety net to address transitional increases in poverty that may result from growth enhancing measures. Chapter 4 discusses current human capital policy issues and makes reference to expenditure policies insupport ofthe social safety net. E. ONTHE LINKS AMONGPOVERTY,INEQUALITY AND ECONOMIC GROWTH 2.52 This chapter concludes with an examination of the relationship among the three key variables discussed above, a topic that i s the subject o f much unresolved controversy. The discussion is organized around four questions. First, what does theory have to say about the effect o f inequality and poverty on economic growth? Second, what i s the empirical evidence on the effect o f inequality and poverty on economic growth? Third, looking inthe reverse direction, what is the effect o f economic growth on inequality? Fourth, what is the effect o f economic growth on poverty? The Effect of Inequality and Poverty on Economic Growth: Theoretical Research 2.53 There i s now a substantial literature on how income inequality and poverty affect economic growth. Early growth models posed the possibility -indeed, the likelihood- that beyond some point growth and equality are conflicting objectives. This early literature used savings-based arguments to develop the theme.27 One strand o f the literature employed as a framework the well-known Harrod-Domar growth model. The key equation o f this model is that inequilibrium: where s i s the economy's savings rate, B i s the capital-output ratio, g is the growth rate, and 6 i s the rate o f depreciation o f the capital stock. It follows that a higher savings rate would raise the equilibrium growthrate o fthe economy. This result raises the question o f what inturn would increase the savings rate. 2.54 In a series o f papers, Kaldor (1956, 1957 and 1958) provided one answer. Inhis class- based model, capitalists have a higher marginal propensity to save than workers do. Thus, the 27See Basu (1997) and Ray (1998) for textbook-level treatments o f these models. 26 larger the capitalist's share o f national income is, the higher the economy's savings rate will be. Inturn, these higher savings translate into greater investment, increased capital formation and, finally, faster economic growth. Inthis way, a higher income share o f capital may be presumed to leadto higher subsequent growth. 2.55 Complementarily, Solow (1956) made a related argument. In his neoclassical growth model, for which he won the Nobel Prize, the equilibrium level o f income in an economy (but not the equilibrium growth rate) increases with the savings rate in the economy. If the marginal propensity to save rises with income, greater inequality in the distribution o f income will raise the economy's equilibriumlevel o fincome. Thus, this model also leads to the same conclusion as before: via the savings mechanism, higher inequality is good for the growth rate (or level) o f national income. 2.56 Conversely, other authors2* have proposed many channels through which inequality mightbeharmful to growth. RentSeeking 0 High inequality increases the ability o f high-income people to use their wealth to secure political outcomes favorable to them. Such actions include political contributions, lobbying, cronyism, bribery and other forms o f "rent-seeking" behavior. To the extent that resources are allocatedto these activities, less is available for productive investment, and this lowers growth. 0 High inequality strengthens the economic and political bargaining power o f the rich, both directly and through the implied threat o f capital flight. 0 Highinequality may leadthe poor to engage inrent seeking or predatory behavior at the expense o fthe middle and upper classes. PoliticalEffects 0 Highinequality contributes to political andmacroeconomic instability. 0 High inequality causes the median voter to favor populist social programs. To help finance such redistributive programs, taxes on capital may be raised, which acts as a disincentive to investment. 0 High inequality gives the poor the ability to limit, via the political system, the private appropriability of investments inskills or technology. Productivity 0 Highinequality stifles agricultural output. This is because landproductivity tends to be higher on smaller farms than on larger ones, 0 Highinequality leads to investments inhigher educationandadvanced healthcare for the few, rather than primary and secondary education and primary health services for the many. The resulting low levels o fbasic human capital hamper economic growth. 28See Barro (2000), BCnabou (2000), Fields (2001), Thorbecke and Charumilind (2002), and Banerjee and Duflo (2003). 27 0 High inequality o f income, or assets that could be used as collateral such as land, limits the ability of the poor to acquire the resources that they need (land, draft animals, machinery, equipment, and fertilizer) in order to operate small farms efficiently. In this way, high inequality worsens the growth-impeding effects o f capital market imperfections. CapitalIntensity 0 High inequality increases the demand for luxury goods relative to basic goods. If luxury goods are producedwith more capital-intensive technologies than basic goods are, then the demand for factors o f production i s skewed in favor o f capital and against labor, perpetuating the initial inequality. 2.57 Inaddition to the role these channels play inleading from highinequality to low growth, there are also mechanisms through which a highpoverty rate could lower economic growth.*' Productivity 0 The poor maybe so badly nourishedthat they are too weak to perform up to their full physical potential. L o w nutrition thus engenders low productivity and continued low incomes. 0 The poor may be unable to undertake economically worthwhile investments owing to the underdevelopment o f capital markets. HumanCapitalAccumulation 0 The poor may be unable to afford to forgo the labor of their children, so that they do not send them to school, thereby passing on the heritage o f low human capital to the next generation. 0 Ifthe poor constitutea majority ofa country's people, the country maybetoo poor to be able to afford to buildsufficient schools, health clinics, andphysical infrastructure. 2.58 Through such mechanisms, poverty begets low growth which begets continued poverty. We have, inshort, a "poverty trap."30 The Effectof Inequalityand Povertyon EconomicGrowth:EmpiricalResearch 2.59 Turning now to the specific question o f how initial economic inequality affects subsequent economic growth, the first empirical evidence was presented in Fields (1991). That study found a negative relationshipbetween initial income inequality and subsequent growth, but this relationshipwas statistically insignificant owingto avery small sample size. 29See Bliss and Stem (1978), Behrmanand Deolalikar (1988), Drbze and Sen (1990 and 1995), Dasgupta, (1993), Ray (1993), Basu (1997), and Ravallion(2004). 30The idea of a poverty trap dates back to Myrdal (1944), Nurkse (1953), Nelson (1956) and Leibenstein (1957). These authors used such terms as "cumulative causation," "circular constellation o f forces" and "low-level equilibriumtrap." Theoretical models of suchtraps are reviewedinBanerjee andNewman(1994). 28 2.60 Working with larger samples in a multiple regression framework, Persson and Tabellini (1992), Alesina and Rodrik (1992, 1994), and Perotti (1992) all found a statistically significant relationship whereby high income inequality reduces economic growth. This result was reaffirmed in many other studies.31 More recently, some studies have appeared in which inequality is found to be positively relatedto economic growth: see, for example, Brandolini and Rossi (1998) for a sample o f developed countries, and Li and Zou (1998) and Forbes (2000) for developing countries. Both sets o f results -the negative relationship found by some and the positive relationship found by others- have been challenged by Banerjee and Duflo (2003). From a careful study, they conclude: "There is no evidence in the data that increases in inequality are good for growth. Infact, the bulk o f the evidence goes inthe opposite direction.'' 2.61 All the preceding studies were basedon income inequality. However, asset inequality has also been shown to be a determinant o f growth. Alesina and Rodrik (1994) and Deininger and Squire (1998) found that higher inequality o f land slows subsequent growth. This finding was reaffirmed by Birdsall and Londoiio (1997) who reached two additional conclusions: that higher educational inequality also depresses subsequent growth and that inequality of assets causes inequalityo fincome to become statistically insignificant. 2.62 The effect o f poverty on economic growth has also been examined, but no predominant result has emerged. The hypothesis o f "unconditional convergence" -i.e., that poorer countries tend to catch up with richer ones- has been tested and rejected.32However, the evidence does not support the opposite hypothesis either. An examination of the data reveals no pattern in the averages: countries that were richer in the beginning grew at the same average rate as initially poorer countries. 2.63 Why some countries grow at faster rates than others is a crucial question in macroeconomics. The evidence presented here shows that inequality and poverty have no systematic effect on economic growth. An understanding o f why these effects go one way in some countries and the other way in others lies at the research frontier o f the economics profession, and further tests are needed before an unquestionable conclusion can be reached. The Effectof EconomicGrowth on Inequality 2.64 Fields (2005) presents a detailed review o f the available evidence on the Kuznets hypothesis (i.e., the idea that income inequality tends to increase inthe early stages o f economic development and to decrease in the later stages). It was not the empirical information presented by Kuznets (1955) that made his work a classic -he had data for only five countries at a single point in time. Rather, it was that he first articulatedwhat he thought was the primary mechanism by which growth affects income inequality: namely, intersectoral shifts. According to Kuznets, economic growth brings about the gradual reallocation o f economic activity from relatively low inequality "traditional" activities to relatively highinequality "modern" activities. 2.65 Incross-section work, researchers have generally found higher inequalityinthe middle- income countries than in either the richer or the poorer countries. This produces a cross-country 3'See Fields (2001) for a survey. 32See Barro (1991), Ray (1998) andDurlaufandQuah(1999). 29 inverted U. But although the inverted U fits the cross-sectional data better than any other shape curve does, variation innational income explains only a small fraction o f the variation inincome inequality. Furthermore, the cross-section inverted U arises because o f the particular econometric method employed (ordinary least squares) and because the highest inequality countries are all middle-income Latin American countries. When fixed effects estimation i s used instead, the inverted U pattern disappears and inequality i s shown to fall over a wide range o f incomes inthe developing countries. 2.66 Looking at the question over time, the invertedU has been found inthe economic history o f some countries that are both developed and developing, but this happens in only a small minority o f cases. Most o f the evidence i s to the contrary -the dominant pattern being a fall in inequality over time during the twentieth century. Looking at shorter periods o f time for a much greater number o f countries, research by Deininger and Squire (1996) has also demonstrated that the Kuznets curve is not an empirical regularity. About 10 percent o f the country cases are consistent with Kuznets' inverted U, another 10percent support an ordinary U, andthe remaining 80 percent exhibit no statistically significant tendency at all. 2.67 Finally, the Kuznets hypothesis leads to the prediction that inequality would decrease more often in the "high-income developing countries" (i.e., those richer than the cross-sectional turning point) than in the "low-income developing countries." But the evidence shows that inequality rises halfthe time and falls half the time inboth groups o f countries. Andno empirical tendency i s found for the rise or fall ininequality to be linked to the rate o f economic growth ina country or to the initial level o f inequality. 2.68 One view often attributed (mistakenly) to Kuznets i s that inequality must increase before it decreases. What Kuznets in fact hypothesized is that inequality tends to' increase before it decreases. But from the findings reviewed here, even this hypothesis i s open to question. The Effect of Economic Growth on Poverty 2.69 Most economists accept without question that economic growth reduces absolute poverty.33Some o f the phrases in economics reflect this: "trickle-down,'' "a rising tide lifts all boats," "the flying geese," and so on. This view can be referred to as the "pro-poor growth" position, inthat when economic growth takes place, the poor and others share the fruits o f it, to a greater or lesser degree.34 2.70 In contrast to the "pro-poor growth" position, the literature also offers a distinctly less popular view according to which economic growth might be thought to make the poor poorer. To take one example, Nobel Prize-winning economist Arthur Lewis (1983) gave six reasons why the development o f enclaves may lower incomes inthe traditional sector: (i) development o f the the enclave may be predatory on the traditional sectors; (ii) products o f the enclaves may compete with and destroy traditional trades; (iii) wage level o f the enclave may be so high that it the ~~ 33 The evidence reviewed in this section i s on "absolute poverty" -i.e., the extent o f poverty when a fixed real poverty line is used. 34 Whether it i s to a greater or lesser degree is, of course, what inequality i s all about. The preceding section reviewed the growth-inequality evidence. 30 destroys employment in other sectors; (iv) the development o f the enclave may result in geographical polarization; (v) the development o f the enclave may lead to generalized improvements in public health and therefore lower death rates; and (vi) the development o f the enclave may stimulate excessive migration from the countryside. 2.71 The present incarnation o fthis view is inthe work of immiserizing growth theorists, who have proven rigorously that it i s possible that economic growth might make the poor poorer.35 Whether growth does make the poor poorer i s an empirical question, to which we now turn. 2.72 Thirty years of research, beginningwith Ahluwalia (1976), have shown convincingly that in a cross-section of countries those with higher per capita income or consumption have less poverty. The cross-sectional version o f the absolute impoverishment hypothesis has been thoroughly discredited. Turning to the time-series evidence within countries, the results can be summed up simply. Usually, but not always, economic growthreduces absolute poverty. On the other hand, when poverty has not fallen it is generally because economic growth has not taken place. Consequently, the pro-poor growth view i s clearly a better general description o f the growth-poverty relationship than i s the immiserizing growth position. 2.73 One final point should be made. While on one hand it is difficult to argue that poverty reduction can be achieved through redistribution policies alongside economic stagnation (let alone economic decline), growth that is associated with progressive distributional changes will have a greater impact on reducing poverty than growth that leaves the distribution unchanged. For example, Ravallion (1997), Bourguignon (2002) and Son and Kakwani (2003) review the poverty-growth-inequality relationship and note that the impact o f growth on poverty is a decreasing function of the degree o f inequality. Poverty will therefore be more responsive to growth the more equal the income distribution is. Intuitively, if the poor have a low share in existing income, they will be likely to have a low share innewly created income. 2.74 The findings presented in this section reaffirm the presumption in favor o f pursuing economic growth in order to reduce poverty. As discussed in the previous section, there may be growth policies that increase inequality and that may bringthe incomes o f some groups o fpeople downto poverty levels. The example givenwas one inwhich trade opening displaces low-skilled workers. While this is not a necessary outcome, it is a possibility that needs to be considered in the design o f safety nets or compensation mechanisms. 35A good summary may be found in Bhagwati and Srinivasan (1983). Bhagwati himself now dismisses these theoretical possibilities as the work o f "ingenious economists properly [making] their mark by proving the improbable." See Bhagwati (1991). 31 3. THE RECENT ECONOMICCRISIS AND RECOVERY 3.1 The previous chapter showed not only that growth has been slow in Argentina, but also that there has been substantial volatility in its economic performance. This high volatility may have contributed to slower output growth and higher poverty rates. Indeed, recent cross-country empirical analysis provides strong evidence that greater volatility in output growth consistently reduces average growth rates o f the economies. This chapter reviews the latest episode of economic volatility inArgentina, a period that started with the sharp decline of economic activity in 2002 and continued with recovery since 2003. More specifically, the chapter analyzes the impact o f recession and growth on poverty levels duringthe 2002-04 cycle, including the role o f relative price changes triggered by the crisis. 3.2 As discussed in Chapter 2, Argentina stands out as a particularly slow growth country. Nevertheless, more rapid economic growth was observed in the 1990s. A number o f empirical studies, based on aggregateddata, show that the highrates o f growth observed duringthis period were mainly driven by increases inproductivity. Regrettably, the convertibility regime ended in a deep political and economic crisis: output collapsed and productivity went down in 2002. The crisis carried with it severe consequences for the poor. The poverty rate, as traditionally measured, rose from 35.9 percent in2001 to a peak of 57.5 percent inOctober 2002. The number of people living below the indigence line doubled. Unemployment rose, but not nearly as much as expected, perhaps because o f the mitigatingeffects of increasedgovernment employment. 3.3 Growth recovered starting in 2003, reaching an average 8.9 percent during the last two years. For 2005, projections indicate another year o f strong growth, possibly around 7 percent. Inflation i s broadly in check (albeit increasing inrecent months) and the exchange rate has been stable. This recovery has improved social conditions inthe country. Although they are still high, unemployment levels declined to 12.1 percent at the end o f 2004 (when the beneficiaries o f the Heads o f Households Program were included as employed). Real wages in the formal private sector have largely rebounded to pre-crisis levels. Wages inthe informal and public sectors still remain about 25 percent below those levels, but they are now 18 percent higher than in December 2001. As a consequence, by the end o f 2004 the percentage o f the Argentine people living inpoverty had fallen by slightly more than one-third since October 2002. The percentage o f people living below the indigence line dropped from 27.5 percent in the same month to 15.0 percent in the last semester o f 2004. Clearly, the latest episode o f economic growth was beneficial to the poor. 3.4 The economy has now recovered to its pre-crisis level and this year will probably reach its 1998 level, the highest reference level in the GDP series. Nevertheless, an important qualitative change has taken place during the upturn of the economy. When analyzed by economic sectors, it i s evident that the sectoral pattern o f recent growth has been quite different from the one observed during the 1990s. As then, growth has been uneven across economic sectors. But the winners and losers across sectors have been quite different. The sectors displaying above average growth rates during the convertibility plan were usually capital- absorbing and employedhighskilled workers. Incontrast, duringthe recent recovery, growthhas 32 been stronger insectors that were more labor-absorbing and that employed low skilledworkers.36 This diverse pattem o f (sectoral) economic growth has probably been a key factor driving the decline o fpoverty and indigence inthe country. 3.5 This chapter is structured as follows. The first section reviews the social impact of the recent crisis and recovery. The second one analyzes how the poor shared the losses (benefits) o f the most recent recession (recovery). More specifically, it employs different methodologies to test whether the recent recovery was pro-poor. The third section presents an examination o f the recession and recovery at the sector level. This analysis provides a better understanding o f what lies behind the aggregate performance o f the Argentine economy and helps distinguish the economic sectors that were responsible for job creation during the recent economic recovery. The final section identifies which factors and sectors contributed most to poverty reduction duringthe latest uptumofthe economy. 3.6 The key findings o f this chapter are as follows. First, and not surprisingly, aggregate growth during the cycle was closely, and inversely, linked to poverty changes. Second, growth had a larger impact on the average incomes o f the poor than on the average incomes o f the rich. That is, data indicate that the recent episode o f economic recovery was a pro-poor growth episode, not only inthe absolute sensebut also inthe relative one. Third, the pattem o f growth in different economic sectors probably responded to changes in relative input prices after the devaluation, with the sectors that were more dynamic being also relatively more labor-absorbing. Fourth, those economic sectors that experienced the greatest growth (weighted by employment share o f the sector) also showed the highest contribution to poverty reduction during the recovery. A. RECENTTRENDS INPOVERTYAND GROWTH 3.7 After three years o f continuing recession, Argentina's economic and financial situation worsened during 2001. Various attempts to spur growth and improve public finances and the debt profile failed. The year culminatedwith the resignation o f the President de la Rua, followed by a quick succession o f appointedpresidents, a formal announcement o f default onpublic sector debt, and, in early 2002, the abandonment o f the convertibility plan. The social situation inthe country, which was not good in 1998, worsened in2002 as GDP declined by 10.8 percent during that year. 3.8 Table 3.1 presents trends inpoverty and indigence inArgentina for the period o f interest. It can be seen that poverty and extreme poverty rapidly increased during (and short after) the economic crisis. The official poverty rate shows that some 53.0 percent of the population was poor in May 2002, up from 35.9 percent a year earlier. This sharp increase in poverty also appeared at the household level: the percentage o f poor households went up from 26.2 percent in May 2001 to 41.4 percent in M a y 2002. Additionally, extreme poverty (indigence) more than doubled during the same period. The percentage o f households in extreme poverty increased from 8.3 percent inM a y 2001 to 18.0 percent inMay 2002. At the individual level, the indigence rate rose from 11.6 percent to 24.8 percent duringthe same period. 36That is, economic sectors displaying higher growthrates also contributed most to job creation inthe economy. 33 Table 3.1: Poverty and Indigence Rates, 2001-04 May-01 May-02 May-03 2S-03 2S-04 Total Urban Household 26.2 41.4 42.6 36.5 29.8 Individuals 35.9 53.0 54.7 47.8 40.2 Excluding income FromPlanJefes Household 43.2 37.1 30.6 Individuals 55.3 48.5 40.9 May-01 May-02 May-03 2S-03 2S-04 Total Urban Household 8.3 18.0 17.9 15.1 10.7 Individuals 11.6 24.8 26.3 20.5 15.0 Excludingincome FromPlanJefes Household 20.5 17.2 13.0 Individuals - 29.7 23.5 18.2 3.9 The economy rebounded quite strongly in 2003, with real GDP growing at 8.8 percent. Nevertheless, during the early phase o f the recovery, poverty and indigence rates continued to increase, albeit only slightly. For example, the percentage o f poor individuals (households) increased from 53.0 percent (41.4 percent) to 54.7 percent (42.6 percent) between May 2002 and May 2003.37The economy continued to show a strong performance the next year, growing 9.0 percent during 2004. Importantly, economic recovery finally began to have a positive impact on living standards during that year, as shown by the steady improvement in key social indicators. Poverty and indigence rates fell between the second semester o f 2003 and the second semester 2004. The percentage o f individuals (households) in poverty declined to 40.2 percent (29.8 percent) at the end o f 2004. At the same time, the indigence rate for individuals (households) also dropped to 15.0 percent (10.7 percent). 3.10 These poverty and indigence rates are based on households' income. However, during these period government transfers accounted for a large part o f households' income in many cases. This fact "biases" poverty and indigence indicators downwards. In other words, poverty and indigence rates might have been higher had those government transfers not existed. The lower parts o f Table 3.1 present poverty and indigence rates, once they are adjusted for the effects o f the Heads o f Households Program. It can be seen that the negative trend inpoverty and indigencerates remains in the adjusted figures. Also, poverty rates are only slightly higher when 37Poverty ratespeaked inOctober 2002, reaching 57.5 percent o f individuals and45.7 percent ofhouseholds. 34 the effects o f the social program are excluded.38Nevertheless, the impact on the indigence rates i s quite important. The adjusted numbers are around 3 percentage points higher when government transfers are excluded. 3.11 An illustrative exercise i s to decompose changes inpoverty (and extreme poverty) during the recent period o f economic volatility by separating the "growth effect" and "the distribution effect."39 Table 3.2 shows the results o f this analysis. Three different periods are considered: Crisis (May 2001 to May 2002), EarlyRecovery (May 2002 to May 2003), andRecovery (Q4-03 to Q3-04).40 As previously noted, the first period (2001-02) shows an important increase in poverty and extreme poverty. When the increase in poverty i s decomposed, it can be seen that the "growth effect" accounts for most o f the change. The "distribution effect" plays some role, but it is not the major contributor to poverty changes. Nevertheless, in explaining increases in extreme poverty rates, both effects are equally important. 3.12 The relative importance o f "explanatory" variables in the early phase o f the recovery (from2002 to 2003) is quite similar. The period is characterized by a small increase inpoverty rates, which i s explained mostly by the "growth effect." Extreme poverty also rose during this time, but in this latter case the distribution component had a minor dampening effect on the poverty increase. Finally, for the last period (2003-04) it is found that "growth effects" and "distribution effects" contributed equally to reducing poverty. In the case of indigence rates, "growth effects" inaddition to a small "distribution effect" explain the drop inextreme poverty. 38 The average increaseinthe povertyrate for individuals andhouseholds is less than 1percentage point. 39 For a description of the methodology employed inthis part, see Bertranou and Khamis (2005). 40 The mainsources ofinformationfor the analysis were the Permanent Household Survey (Encuesta Permanente de Hogares, EPH) and the Continuous Permanent Household Survey (Encuesta Permanente de Hogares Continua, EPH-C). The EPH i s available until May 2003, while the EPH-C i s available since the last quarter o f 2003. Therefore, for the period 2001 to 2003 the old EPH surveys corresponding to May were used. For the most recent period, the analysis i s based on the quarterly EPH-C surveys for the last quarter o f 2003 and the third quarter o f 2004. 35 Table 3.2: Decomposition of Poverty and Extreme Poverty Rates, 2001-04 (in percentages) Period Total Growth Distribution Change Effect Effect Poverty 2001-2002 16.9 12.9 4.0 2002-2003 1.3 0.9 0.4 2003-2004 -6.7 -3.4 -3.3 Indigence 2001-2002 13.1 6.8 6.3 2002-2003 0.6 0.7 -0.1 2003-2004 -2.9 -2.3 -0.6 Source: Bertranou and Khamis (2005). 3.13 The main conclusions of the decomposition analysis on poverty changes for the most recent period are mostly the same as those presented inChapter 2, when a longer time period was considered. In other words, poverty changes in recent years were largely determined by economic growth. This same result holds when a decomposition analysis o f this kind i s applied to other developing countries inAfrica, Latin America, and East and Southeast Asia.41However, Argentina differs from the generality o f countries inthat insome periods the "distribution effect" turns out to be as important as the "growth effect" inexplaining changes inpoverty and extreme poverty. 3.14 Overall, the most recent episode o f economic volatility had a significant impact on social indicators. Whether or not incomes are adjusted for government transfers, poverty and indigence rates sharply increased during the crisis. Conversely, the combined impact o f higher levels o f economic activity and employment reduced those rates in 2004. Moreover, the growth- redistribution analysis presented in this section indicates that economic growth was the main determinant o f poverty changes inrecent years. Whether volatility has had a larger impact on the average incomes o f the poor is a question that needs to be assessed. Thus, the next section calculates how pro-poor recent growth rates have been. B. PRO-POOR GROWTH EFFECTS 3.15 In order to understand the social impact o f the recent crisis and upturn, it is crucial to investigate how the poor shared the losses (benefits) o f the recession (recovery). This section reports on changes in the income o f the poor for the following three sub-periods: Crisis (May 2001 to May 2002), Early Recovery (May 2002 to May 2003), and Recovery (44-03 to 43- In this section the absolute definition o f pro-poor growth is initially used. By this definition, growth is considered pro-poor if, on average, the incomes o f the poor are rising. In other words, the poor benefit inabsolute terms, as indicated by a falling poverty measure.43 41For example, see Datt and Ravallion (1992) for the cases o f Brazil and India. 42See the previous section for a discussion on the main sources o f information for the analysis. 43See Ravallion and Chen (2003) or Kraay (2004). 36 3.16 Table 3.3 presents the average percentage changes in the income o f the poor during the different sub- period^.^^ It can be seen that the poor experienced a serious decline in their household income in the first sub-period (-36.7 percent). However, they show on average a 7.3 percent growth in income duringthe first part o f the recovery. The same dynamic was observed during the most recent period o f the recovery. Between 2003 and 2004 pro-poor growth effects were even stronger, with the average income o fpoor households growing at 8.5 percent. Table 3.3: Average Percentage Changes inIncome of the Poor, 2001-04 2001-2002 2002-2003 2003-2004 Total Household Income -36.7 7.3 8.5 Excluding Income FromTransfers -37.3 -10.3 8.6 Source: Bertranou and Khamis (2005). 3.17 As before, these numbers are "biased" since household income includes government transfers under social programs. The lower part o f Table 3.3 shows the results when household income excludes government transfers. Clearly, the number for the first period does not change much, as the drop in the average income o f poor households is only slightly higher than it was before (-37.3 percent). However, the pro-poor growth effect previously observed in the second period disappears: the growth rate in mean income i s now negative (-10.3 percent). Thus, the pro-poor growth effect in this period can be explained entirely by government transfers under social programs. Finally, the average income o f the poor went up by 8.6 percent duringthe last period when government transfers are excluded in the calculations. This rate i s only slightly higher thanthat based on incomes including government transfers. 3.18 As a conclusion, while the crisis significantly hit the poorer population groups, the subsequent recovery was clearly a pro-poor growth episode. Evenwhen adjusted by government transfers, the income o f the poor showed a positive correlation with the growth rate o f the economy: it fell considerably during the crisis and experienced a revival with the upturn o f the economy. Only for the EarlyRecovery phase is the evidence mixed. 3.19 The previous analysis on average changes in the income o f the poor can be complemented by a discussion o f growth incidence curves. This exercise looks at patterns o f income changes over the different percentiles o fthe entire income distribution curve.45The same three periods are considered in the analysis, namely: Crisis, Early Recovery and Recovery. Again, two different income measures are employed: total per capita household income and total per capita household income excluding government transfers.46 As was previously mentioned, government transfers under social programs, such as the Heads of Households Program, could 44That is, individuals whose incomes are below the official poverty line. 45 The vertical axis in the growth incidence curves figures represents the growth rate inhousehold incomes, while the horizontal axis represents the percentiles o f the distribution o f income inthe population. The vertical lines inside the figures are the poverty lines for a given year. 46 In addition, Bertranou and Khamis (2005) present estimates o f growth incidence curves considering different economic sectors (manufacturing, construction and retail) and sector o f employment (formal and informal). Results are similar to those presented inthis section. 37 have major impacts in the lower tail of the income distribution and therefore could lead to an underestimation (overestimation) o f the effects o fthe slump (re~overy).~~ 3.20 Figure 3.1 and Figure 3.2 show the results. Inthe first figure, growth incidence curves for the crisis and the early recovery period are presented. It i s very clear that the period o f the slump i s characterized by negative growth rates in household incomes across the entire distribution. Therefore, the average growth rate i s also negative (around -31.O percent). However, the figure shows that the crisis had a larger negative effect on the lower percentiles than on the higher percentiles o f the income distribution. This shows up in an upward-sloping growth incidence curve. The exclusion o f government transfers for the calculations does not change the results very much. Figure 3.1: Growth Incidence Curves I 2001-02 and 2002-03 Household income ier capit, - total & without transfers 0 0 - 7 poverty rate 2001 -- poverty rate 2002 *"`c, &w=' , 0 I 9 - , I 0 Source: Bertranou and Khamis (2005). 47 For the analysis presented in this section, it i s assumed that the cash transfer i s the "net gain" from the program. That is, it ignores the opportunity cost o f participation inthe program and, thus, could overestimate its impact. See Galasso and Ravallion (2003) for a discussion on this topic. 38 Figure 3.2: GrowthIncidence Curves 2003-04 '"I Household inco r capita- total & \ ~ \\ I -5 1 I I I I 0 20 40 percentiles60 80 100 Source: Bertranou and Khamis (2005). 3.21 For the early phase o f the recovery, a downward-sloping growth-incidence curve can be observed. This curve shows positive growth rates in household incomes at the lower percentiles o f the income distribution but negative growth rates at the higher percentiles. Thus, the average growth rate is around zero. The downward-sloping curve and positive growth rates at the low percentiles o f the income distribution make it possible to initially classify economic growth during this period as a process o f pro-poor growth. Nevertheless, the exclusion o f government transfers from the calculations significantly changes the results. The growth incidence curve for the period now looks almost h~rizontal.~' The pro-poor growth effect, with higher and positive growth rates at the lower end o f the income distribution, seems to disappear when per capita household incomes are corrected by excluding government transfers. Thus, it can safely be concluded that the pro-poor growth effects observed during the early stages o f the recovery are almost entirely due to the impact o f government transfer programs such as the Heads o f Households Program. 3.22 For the 2003-04 growth episode, the available data restricted the analysis to the time interval betweenthe second semester o f 2003 and the first semester o f 2004. 3.23 Figure 3.2 depicts a downward-sloping growth incidence curve for this period. Moreover, the curve exhibits positive growth rates in household incomes across almost the entire income distribution (the exception being the very high percentiles). The average growth rate o f household incomes i s positive and i s around 4.9 percent. Moreover, when government transfers are excluded from the calculations the growth incidence curve follows a very similar trend. ~~ 48Moreover, the average growth rate of income for the period i s -8.6 percent inthis case. 39 3.24 Overall, the conclusions o f the growth incidence curves are similar to those found when calculating average changes in the income o f the poor. Specifically, the recent upswing o f the economy was clearly a pro-poor growth process. Moreover, the incidence curves showedthat the incomes o f the poor grew (fell) faster than the average income o f the entire population duringthe upturn(crisis). That is, growth was pro-poor inboththe absolute andthe relative sense. 3.25 The calculations of the average changes in the income o f the poor and the growth incidence analyses previouslypresented look at changes inthe income distribution o f anonymous individuals or households from comparable cross-sectional surveys or censuses. These studies compare one set o f poor households in the base year to another set o f poor households in the final year. They say nothing about the income changes o f initially poor or rich households. To see how those initial positions in the income distribution have fared, longitudinal data are required.These are available from the Encuesta Permanente de Hogares (EPH) and can be used to examine earnings mobility inArgentina.49 Table 3.4: Mobility Profiles by InitialPosition, 1996-2003 Changes in Reported Initial Earnings 1996-97 * 1997-98 * 1998-99 * Mean Std.Dev Mean Std.Dev Mean Std.Dev PoorestQuintile 207.2 494.7 264.9 501.7 243.4 479.2 2"dQuintile 22.2 207.6 33.3 207.0 20.8 200.2 3'd Quintile 7.7 262.5 13.8 301.1 -18.1 238.1 4thQuintile -25.4 376.0 -39.1 368.0 -36.1 345.7 RichestQuintile -230.1 784.4 -181.8 1016.0 -319.7 1055.3 1999-00 * 2000-01 * 2001-02 * Mean Std.Dev Mean Std.Dev Mean Std.Dev PoorestQuintile 249.3 536.3 238.1 552.8 163.6 410.4 2"dQuintile 2.2 184.8 2.7 183.4 -65.5 139.3 3'd Quintile -18.4 232.4 -23.8 196.4 -126.9 204.1 4'h Quintile -61.2 322.0 -49.4 315.0 -213.7 287.7 RichestQuintile -314.2 882.0 -331.7 980.0 -606.3 846.3 2002-03 * Mean Std.Dev Poorest Quintile 216.8 401.0 2"dQuintile 28.2 204.5 3rdQuintile -3 1.4 165.6 4thQuintile -72.2 212.1 RichestQuintile -301.3 591.7 ' The hypothesis o f equality of means by groups i s rejected at 99 percent significance level Source: Fields and Sanchez Puerta (2005). 49The survey i s a rotating panel, with a quarter o f the households rotated out each period, so that a given household can be followed up to three periods. 40 3.26 In the empirical work presented in Table 3.4, the dependent variable is the change in labor market earnings for the same individual from May of one year to May of the next.50The reason for the choice o f change in eamings as the dependent variable o f interest i s that in developing countries earnings have been shown to constitute the single most important source of variation o f change in income -more so than all the other income sources ~ombined.~' The paramount role o f changes in labor earnings in explaining changes in incomes points to the importance o f understanding earnings dynamics and employment transitions more fully. Therefore, the focus o f this part i s on analyzing the way in which labor markets distribute rewards among the population. 3.27 Table 3.4 presents one o f the key findings from this analysis. The table relates the changes in an individual's labor market eamings to the quintiles o f reported base year earnings for Argentina.52 The table considers seven annual panels, from 1996-97 to 2002-03. It can be seen that the relationship between the quintile o f reported earnings and average eamings changes is statistically significant and negative. This result means that households that reported lower initial income experienced larger average gains the following year, not just in pesos but in percentage^.^^ This conclusion is robust across all the panels considered. That is, the result holds also for the recent crisis and recovery. Another major conclusion o f the panel data analysis is that those who gain the most when the economy i s growing are also those who gain the most or lose the least when the economy is contracting. This means that earning changes are determined by structural variables rather than being symmetrical to growth.54 Finally, the findings also strengthen the case for economic growth. The panel perspective shows that in urban Argentina higher economic growth raises the meanincome increase ofidentified individuals. 3.28 To better understand the dynamics o f poverty through the recovery period, it i s revealing to analyze the growth pattern o f the different economic sectors. The hypothesis is that the pro- poor effects o f the post-crisis economy may have been linked to a sector growth pattern different from that prevailing before the collapse. This could be the case if sectors that have shown the fastest recovery rate are more labor-absorbing than those that have shown slower recovery. In addition, the pro-poor impacts would have been higher if labor-absorbing sectors demanded more low-skilled workers than the average. The next two sections discuss these hypotheses. 50 For the exercises, changes in earnings are calculated over a one year period (May-May) to avoid capturing changes ineamings due to seasonality. 5 'See Fields et al. (2003a). 52 The analysis i s conducted using eamings changes inpesos, which measures absolute earnings gains. All earnings are expressed in 1999 pesos per month. Nominal earnings are deflated by the April Consumer Price Indices for Greater Buenos Aires to obtain real earnings. Eamings include wage or salary, self-employment income, and earnings as an owner or employer. 53 Moreover, changes ineamings for the lowest quintiles are usually higher during positive growth years than during negative growth years. 54 Nevertheless, a relative mobility analysis reveals that, even though individuals that reported the lowest base year income enjoy the most favorable income changes, for the majority o f them the gain is not big enough to put them in the highest quintiles o f the income distribution. That is, the richest and the poorest individuals are likely to remain in the same income quintile. The greatest positional movement is found to be inthe middle income quintiles. 41 c. SECTOR GROWTH PATTERNSTHROUGH THE RECOVERY FROMTHE CRISIS 3.29 Changes in relative input prices could lead to changing patterns o f input contributions in production. For example, during the 1990s interest rates decreased relative to wages, and the most dynamic sectors tendedto be capital intensive. Incontrast, real currency depreciation after the crisis substantially reduced the cost o f labor in the country. As shown in Figure 3.3, real wages in the private and public sector dropped sharply after the crisis.55 As a result, labor intensive sectors became more competitive after the currency depreciation. In order to better understand the adjustment o f the economy to the crisis, this section examines sector growth patterns duringthe recent recovery and compares them to those observed duringthe 1990s. /+Overall - -Registered Private -Unregistered Private *- -PublicI I Source: World Bank staff. 3.30 The following key observations interms o f the growth pattern inthe recovery period that followed the 2001-02 crisis are noted: Sectors that contributed most to aggregate GDP growth in the recovery period are different from those that explained growth in the 1990s. The recovery shows the relatively higher importance o f the contribution of the goods-producing sectors, mainly manufacturing, while inthe 1990sthe services sectors were most important. 0 The dynamic sectors during the recovery have been also relatively more labor- absorbing. 55 More recently, real wages in the formal private sector have largely rebounded to pre-crisis levels,butwages inthe informal and public sectors still remain about 25 percent below those levels. 42 0 These sectors have shown higher increases ininvestment andtotal factor productivity. 0 The most dynamic sectors during the recovery demand relatively more intensively low-skilled workers compared to the sectors that led growth inthe 1990s. 3.31 These observations could be helpful in explaining how the pattern o f pro-poor growth found inthe previous section occurred. Since labor income is more important among the poor as a share o f total income, and since the poor tend to be less-skilled, it follows that pro-poor growth duringthe recovery could be linkedto the sectoral compositiono fthe recovery. Table 3.5: Growth Decomposition by Sector, 1994-2004 Growth I RecessionI Growth I1 RecessionI1 Crisis Recovery 14-94 to 24-95 to 24-96 to 34-98 to 44-01 to 14-03 to 1Q-95 14-96 34-98 24-01 44-02 44-04 Total 5.14 -3.42 7.12 -2.00 -10.71 8.87 Total Goods 1.80 -1.77 2.45 -1.15 -3.56 3.89 Agriculture 0.44 0.03 0.21 0.02 -0.13 0.19 Fishing 0.00 0.00 0.01 0.00 -0.01 -0.01 Mines 0.21 0.26 0.00 0.03 -0.04 0.04 Industry 0.66 -1.37 1.33 -0.90 -1.82 2.21 Electricity 0.21 0.12 0.16 0.10 -0.07 0.21 Construction 0.27 -0.81 0.75 -0.41 -1.50 1.26 Total Services 3.39 -1.43 4.32 -0.48 -5.88 3.49 Retail 0.73 -1.27 1.28 -0.67 -2.3 1 1.56 Hotels 0.18 -0.07 0.27 -0.06 -0.25 0.17 Transport 0.67 0.01 0.77 -0.01 -0.70 0.94 Financial 0.67 0.00 0.80 0.04 -1.20 -0.55 Real Estate 0.85 -0.19 0.65 -0.03 -0.83 0.64 Public Adm. -0.13 -0.02 -0.02 0.02 -0.05 0.08 Schooling& Health 0.19 0.18 0.25 0.20 -0.02 0.26 Other Services 0.24 -0.07 0.31 0.04 -0.52 0.39 ource: Fernhndez Valdovinos et al. (2005). 3.32 The first key observation i s that sectors that contributed most to the recovery o f output after the crisis are different from those that explained growth prior to the crisis. Table 3.5 presents a growth accounting analysis, where the average annual growth rate o f GDP i s decomposed into the contribution o f the different economic sectors. Inthe table, the sum o f the growth contributions o f the different sectors i s equal to the growth rate o f GDP.56The table shows that during the 1990s the service sectors made the largest contribution to aggregate growth, notably retail, transport, financial services and real estate. Services explain about two- thirds ofthe growth rate registeredinthis period. This pattern also holds inthe recovery from the recession that began in late 1995. In contrast, in the recovery period that followed the 2002 56The aggregate economy growth rate is an average, over a given period, of the growth rates of quarterly GDP with respect to the same quarter o f the previous year. 43 crisis, the growth pattern was reversed: growth was driven relatively more by the production o f goods than by the production o f services. 3.33 About one-fourth o f total economic growth during the recent recovery is explained by growth in the industrial sector. This sector accounted for 2.2 percentage points out o f the 8.9 percent growth rate o f the aggregate economy. The retail sector ranked second, followed by the construction sector inthe third place and transport inthe fourth one. This growth pattem i s quite different from that observed in the recovery periods of the 1990s. During that decade, industry accounted for much less of the aggregate growth during the two growth periods. This result also holds true for the retail, construction and transport sectors, which have proved more dynamic duringthe recovery from the recent crisis thanduringthe 1990s. 3.34 Several o f the service sectors that pushed aggregate growth in a significant way during the 1990s have proved to be much less important incontributing to the post-crisis recovery. For example, financial services have shown a negative contribution to growth during the recent recovery, while in the 1990s the contribution o f this sector was very important during periods o f positive growth. Moreover, during this decade this sector has also helped cushion recessions by showing positive growth in times o f aggregate GDP decline. Health, education and other social services also made a slightly higher contribution (in percentage terms) to total growth in the 1990s than during the recent recovery, showing strong resilience during recessions in the past decade as well. 3.35 The second key observation in this section is that the sectors that contributed most to GDP growth duringthe recovery were also relatively more labor-absorbing. Table 3.6 shows the amount o f labor employed by sector according to the sample o f the Permanent Household Survey. The indicators for May 2001 and May 2002 can be regarded as representative o f the crisis period, and the indicators for May 2003, the fourth quarter o f 2003 and the third quarter o f 2004 as representative o f the recovery period. Two o f the sectors that were identified as main contributors to the post-crisis recovery are among the ones that have shown the largest increases in the amount of labor employed: manufacturing and retail.57 Moreover, the three main contributors to growth in the recovery (manufacturing, retail and construction) accounted for almost 52 percent o fjob creation in the post-crisis recovery. The same three sectors accounted for about half o f the total employment on average duringthis period.58 ''Only '*The education and health created more jobs during the recovery. main sources o f information for the table were the Permanent Household Survey (Encuesta Permanente de Hogares, EPH) and the Continuous Permanent Household Survey (Encuesta Permanente de Hogares Continua, EPH-C). The EPH i s available until May 2003, while the EPH-C is available since the last quarter o f 2003. For the period 2001 to 2003 the old EPH surveys corresponding to May were used. For the most recent period, the analysis i s based on the quarterly EPH-C surveys for the last quarter o f 2003 and the third quarter of 2004. Therefore, the figures of employment by economic sector for May 2003 and fourth quarter 2003 (upper part o f the table) are not comparable. 44 Table 3.6: Em lo men Distribution by Economic Sectors, 2001-04 InThousands May-01 May-02 May-03 4Q-03 34-04 ~ 1Primary 91 79 108 120 125 Manufacturing 1,180 1,073 1,101 1,182 1,273 1Construction 625 461 550 624 689 Retail, Rest. and Hotel 2,000 1,784 1,888 2,093 2,188 Utilities and Transport 748 654 666 639 677 Finance and Real Estate 803 711 826 826 767 Public Adm. and Defense 585 607 617 662 680 Education and Health 1,055 1,104 1,252 1,108 1,185 Other Services 1,177 1,147 1,148 1,078 1,122 Shares May-01 May-02 May-03 4Q-03 3Q-04 Primary 1.1 1.o 1.3 1.4 1.4 Manufacturing 14.3 14.1 13.5 14.4 14.2 Construction 7.6 6.1 6.7 7.3 7.5 Retail, Rest. and Hotel 24.2 23.4 23.2 24.7 25.1 Utilities and Transport 9.0 8.6 8.2 7.7 7.7 Finance and Real Estate 9.7 9.3 10.1 9.7 9.9 Public Adm. and Defense 7.1 8.0 7.6 7.9 7.9 Education and Health 12.8 14.5 15.4 13.8 13.3 Other Services 14.2 15.1 14.1 13.2 12.9 Changes inThousands May-01 May-02 4Q-03 ~ May-02 May-03 34-04 Primary -12 29 5 Manufacturing -107 28 90 Construction -163 88 65 Retail, Rest. and Hotel -216 104 95 Utilities and Transport -93 11 38 Finance and Real Estate -92 115 -59 Public Adm. and Defense 22 10 18 IEducation and Health 49 148 78 Other Services -29 1 44 Source: Bertranouand Khamis (2005). 3.36 The third key observation is that sectors that have contributed more to growth duringthe recovery period have also exhibited the highest contribution o f capital and total factor productivity. Table 3.7 shows the results o f a growth accounting exercise at the sector level. The table shows that the four sectors o f the most rapid growth and higherjob creation inthe recovery (industry, construction, retail and transport) have also been the sectors with the highest contribution o f capital and total factor productivity (TFP) to sector growth. Infact, according to this analysis, most o f the growth observed in those sectors appears to be driven more by the combination o f investment and TFP thanby increases inthe labor input. 45 v) e, m X Y u 0 3.37 The fourth observation i s that the most dynamic sectors through the recovery demanded relatively less-skilled workers, whereas the sectors driving growth in the 1990s demanded higher-skilled workers. This observation can be supported by analyzing the percentage o f workers with different levels o f education hired by each economic sector, taking education as a proxy for skill or human capital. Table 3.8 shows that the four most dynamic sectors inthe recent recovery period (manufacturing, construction, retail and transport) are intensive in low skilled labor. In these sectors, the large majority o f workers (85 percent in average) at most complete secondary studies. On the other hand, financial services and education and health services, two sectors with strong economic perfonnance inthe past, are intensive inhighskill labor. More than halfo fworkers have a university degree inthese sectors. EconomicSector Percentageof workers with a Complete Complete Complete PrimaryStudies Secondary Tertiary (at most) Studies Studies Primary 46 36 18 Manufacturing 36 49 16 Construction 56 39 6 Retail,Rest. and Hotel 38 46 16 Utilities and Transport 29 55 16 Finance and Real Estate 9 34 56 PublicAdm. and Defense 21 46 33 Education and Health 11 22 67 Other Services 50 36 14 3.38 In sum, available evidence indicates that the economic sectors that have been more dynamic inthe recovery from the 2001-02 crisis have been relatively more labor-absorbing, have shown the highest contribution o f capital and TFP to growth accompanying the utilization o f labor, and have demanded workers with relatively lower human capital. These characteristics of the recent recovery were probably crucial to finding the result o f pro-poor growth patterns. This hypothesis deserves further study and analysis, particularly as the recovery continues to unfold. The next and final section checks the factors and sectors that have contributed most to the reduction o fpoverty duringthe recovery. D. SECTORAL DECOMPOSITION POVERTYVARIATIONS OF 3.39 Poverty changes can also be looked at from a sectoral perspective. In the analysis presented below, the contribution o f a particular economic sector to total poverty changes in a given period is mea~ured.~' this section, poverty changes are decomposed into intra-sectoral, In 59 In this section, sector means economic sector, as defied innational accounts. However, it could also imply type o f job status (formal and informal), labor force status (employed, unemployed and inactive) or labor market status (employer, self-employed, employee and non-paid), 47 inter-sectoral (population-shift) and interaction effects.60 The results o f these alternative decompositions o f poverty are presented in Table 3.9. Obviously, the general poverty trend indicated previously -a rise inpoverty during 2001-02 and 2002-03 and a decline during 2003- 04- remain. 3.40 During the first period, all economic sectors contributed to the observed change (increase) in poverty. Manufacturing, construction, retail-restaurants-hotels, utilities-transport and other services were the sectors that accounted most for the poverty increase. In contrast, inter-sectoral and population shift effects slightly diminished poverty rates. The early recovery period presented only a small increase in poverty. Contrary to what was observed during the crisis, while some economic sectors contributed to a rise in poverty during the period (public administration, education, social and health services, other services), others had dampening effects (industry, construction, utilities-transport), Finally, during the most recent period of the recovery, overall poverty was declining. All the economic sectors helped to reduce poverty rates during this phase, but the main contributors were industry, public administration, finance and real estate, and transport. I Table 3.9: Sectoral DecomDositionof Povertv. 2001-04 s!~;iO1 t; May-02 EconomicSectors' Cont. Primary 1.14 1.13 1.04 3.43 1.77 -2.45 Manufacturing 14.38 17.11 13.97 -39.86 13.61 -22.39 Construction 9.57 13.44 9.15 -6.22 7.62 -5.01 Retail,Rest. and Hotel 23.29 25.49 22.65 9.08 21.87 -8.00 Utilitiesand Transp. 8.75 10.73 8.27 -26.46 7.11 -11.22 Financeand RealEstate 9.08 4.05 8.43 1.37 8.39 -15.55 PublicAdm. and Def. 7.24 6.27 8.13 77.07 10.13 -23.44 Education and Health 11.75 9.03 12.79 74.41 15.83 -10.12 Other Services 14.80 14.24 15.57 40.07 13.68 -10.91 Total Intra-sectoral 101.48 132.89 -109.09 -1.04 -52.63 6.02 I Population shiji effect Interaction effect -n-.44 . . 19 74 . 7 n i 3.41 One may ask, what is the relationship between economic growth, employment and poverty changes? Are those economic sectors that grew (shrank) most also the ones that contributed more to reducing (increasing) poverty in a given period? H o w i s employment in the sector affecting this relationship? The previous section showed that those economic sectors that have shown the fastest uptum during the recent recovery were also relatively more labor- absorbing, helping to create the majority o f newjobs. It was shown also that these same sectors usually demand workers with a relatively low level o f human capital, an educational 6oThe sample includes only those individuals who are employed. For a full description o f the methodology used in this section, see Bertranou and Khamis (2005). 48 characteristic o f the poor. Thus, this final part o f the section examines whether those labor- absorbing sectors also accounted for most of the changes inpoverty over time. 3.42 Figure 3.4 and Figure 3.5 help to answer this question. Ineach figure, the horizontal axis measures the contribution to changes in poverty o f each economic sector. The vertical axis measures economic growth by sector, weighted by the employment share o f the sector at the beginningo fthe period.6' Thejustification for this is as follows. As is found ina largenumber o f studies, economic growth i s important to reduce poverty. Countries that have historically experienced the greatest reduction in poverty are those that have experienced prolonged periods o f sustained economic growth. However, the type or quality o f economic growth i s also important. If growth i s concentrated more in the labor-absorbing sectors, it i s more likely to create a larger number o fjob positions as the economy expands. As a consequence, this probably will have a larger impact onpoverty reduction through greaterjob creation. Figure 3.4: Sector Economic Growth and Contribution to Poverty Changes May2002-May2003 Utiiiliesand Transport Finance and Real Estate B - I a, - .,," v) Construction U I B Educationand Health j s.- c m PublicAdm. and Defenst Ii -6 ) 0 -40.0 -20.0 . Primary 20,0 40.0 60.0 80.0 10 .o OtherServices i ~ 7% Retail,Rest. and Hotel y 3.43 The correlation between sector growth and contribution to poverty reduction m not be r very strong for the early phase o f the recovery (Figure 3.4). It is important to remember that during this period poverty rates had actually increased. However, the relationship is clearer during the most recent part of the recovery (Figure 3.5). That is, the higher the (weighted) economic growth o f the sector was, the higher was its contribution to poverty reduction. In summary, economic growth is a valuable means o f achieving poverty reduction. However, this growth could have greater pro-poor effects if it occurs in those sectors that are relatively more Employment data are from the May 2002 and last quarter o f 2003 householdsurveys. 49 labor-absorbing and relatively more demanding o f unskilled labor. As suggested before, this patterni s likely to be observed ifit i s encouraged byrelative inputprices. Figure3.5: Sector EconomicGrowthandContributionto PovertyChanges March2003 -March2004 Financeand RealEstate C '? re d B n .-m p! C s -25 0 -20 0 -15 0 .IO0 -5 0 5 0 Contributionto PovertyChanges Source: World Bank staff. 3.44 Overall, the main policy implications o f the analysis o f the recent crisis and recovery cycle are the following. First, macroeconomic stability is important not only for long-run growth but also for poverty reduction. Poverty levels usually increase sharply after an economic crisis, and Argentina was no exception. AAer the recent crisis, poverty rose dramatically inthe country: the number o f poor grew by roughly 20 percentage points to more than half o f the total population, while the number o f extreme poor nearly doubled. Second, it was found that the recovery was pro-poor and that one important determinant o f this outcome may have been the change inrelative prices after the devaluation. It was found that sectors that contributed most to GDP growth during the recovery were also the sectors that were relatively more labor-absorbing and that usually demand less-skilled workers. These characteristics of the recent recovery were probably crucial to finding the result o f pro-poor growth patterns. This hypothesis deserves further study and analysis, particularly as the recovery continues to unfold. Third, the role o f government transfers was critical to alleviating poverty, especially extreme poverty. It was found that poverty and indigence rates might have been higher had those government transfers not existed. 50 4. OPTIONS FOR SELECTEDGOVERNMENT POLICIES 4.1 The previous chapter showed that the recent economic crisis in Argentina resulted in severe social dislocations and a major reduction o f welfare for its population. The Convertibility Plan collapsed in 2002, resulting in a default on government foreign debts, increased inflation, falling output and exchange rate devaluation. This sharply aggravated the already difficult social conditions in the country. The economy started to bounce back in 2003. Since then unemployment levels had declined, reaching 12.1 percent at the end o f 2004 (14.5 percent if beneficiaries o f the Heads o f Households Program are considered unemployed). Additionally, real wages have largely rebounded, although wages in the informal and public sectors still remain about 25 percent below pre-crisis levels. As a consequence, a steady improvement in social conditions was observed in the country. Moreover, as the empirical evidence presented in Chapter 3 indicated, the latest recovery was clearly a case o fpro-poor growth. 4.2 Nevertheless, the economic and social improvements were from a very low base, and large segments o f the Argentine.population continue to face great hardship. The poverty rate remains high, affecting 40.2 percent o f the population in the last quarter o f 2004, while 15 percent o f the population was still below the extreme poverty line at that date. Consequently, one o f the key challenges that the country faces today i s to implement appropriate policies so that the country embarks ina path o f sustained and equitable economic growth. Strong economic growth, based on sound macroeconomic management, will be crucial for a further decline inpoverty. 4.3 A large number of recent cross-country empirical studies have attempted to identify the factors or policies that could deliver strong economic growth. The empirical framework inthese studies is mainly derived from the underlying neoclassical growth model, as extended and complemented by more recent endogenous growth models. As seen in Chapter 2, two major categories o f explanatory variables are usually employed in these studies: initial state variables, and control and environmental variables. Initial state variables include the log o f per capita income at the beginning of the period, the initial output gap andthe stock ofphysical and human capital o f the population. The control and environmental variables comprise those actions chosen and pursued by governments or private agents. These control and environmental variables contain subcategories o f (i) structural variables; (ii) macroeconomic variables; and (iii) extemal environment conditions variables. 4.4 The aim o f this final chapter i s to discuss a selected set o f government policies that the theoretical and empirical literature has found effective in delivering high economic growth. Of course, an analysis o f all the options currently available to the government i s well beyond the scope o f this study. Thus, the chapter focuses on a limited number o f these options in selected areas, namely: infrastructure, fiscal policy, trade policies, human capital formation (education) and private sector development. As presented in Chapter 2, a large number o f studies concluded that these policies are major determinants o f economic growth. Besides considering how these policies could help foster growth inArgentina, this chapter also examines their likely effects on poverty and inequality. 51 A. INFRASTRUCTURE INVESTMENT 4.5 It has long been recognized that an adequate supply o f infrastructure services is an essential ingredient for productivity and growth. In recent years, however, the role o f infrastructure has received increased attention. From the academic perspective, a rapidly growing literature has sought to quantify the contribution o f infrastructure to income and growth. From the policy perspective, the renewed concern can be traced to two worldwide developments. The first was the retrenchment o f the public sector from its dominant position in the provision o f infrastructure. The second was the opening up o f infrastructure industries to private participation, which has been reflected inthe widespread privatization o f public utilities and the multiplication o f concessions and other forms o f public-private partnership. This section reviews the current stance o f infrastructure investment in Argentina. It first examines the implications o f empirical studies for increasinginfrastructure inthe country. Next, it reviews recent trends ininfrastructure investment, both public and private. The third part analyzes some bottlenecks that could be impedingeconomic activity. Finally, the section discusses investment needs inArgentina. Infrastructure, Economic Growth and Poverty 4.6 Empirical studies have investigatedthe impact o f infrastructure on economic growth.62In one recent and detailed study, Calderbn and Servkn (2004a) find a substantial growth pay-off in increasing the quality and quantity o f infrastructure. Their results are presented in Table 4.1.63 According to them, improving infrastructure stocks in Argentina to the level o f the regional leader (Costa Rica) could result, ceteris paribus, in an increase in the GDP growth rate o f 1.3 percentage points per annum. Inaddition, improving the quality o f infrastructure in the country could add another 0.4 percentage points. This adds up to an overall improvement in growth performance o f 1.7 percentage points per annum. Similarly, improving infrastructure stocks and performance to the level o f the median country in East Asia and the Pacific (Korea) could add over 3 percentage points to annual growth. Table 4.1: Potential Growth Impact from Expandedand Improved Infrastructure (in percent of GDP) Improvement to levels of LAC Improvement to levels of EAP Leader* Median* Stocks Oualitv - Total Stocks Oualitv Total Argentina 1.3 0.4 1.7 2.2 0.9 3.2 Brazil 1.5 1.4 2.9 2.4 1.9 4.4 Chile 1.3 0.0 1.3 2.3 0.6 2.8 Colombia 1.9 1.2 3.1 2.9 1.7 4.6 Costa Rica Regional Target 1.o 0.5 1.5 Mexico 1.4 0.2 1.7 2.4 0.8 3.2 *Regional leader is Costa Rica. East Asia and the Pacific (EN) median is Korea. Source: Calder6n and Servtn (2004a). 62 See Calderon and Servtn (2004a) for a full review o f the literature on infrastructure and growth. Country-specific differences may not be totally captured in cross-country analyses, and therefore the results presented inthis part should be interpreted cautiously. 52 4.7 Calder6n and ServCn also find an important effect o f infrastructure on poverty. They suggest that the effect i s twofold, first by promoting overall income growth and second by reducing inequality. This means that economic growth based on infrastructure investment would be especially beneficial to the poor. According to these authors, the impact for Argentina o f improving its infrastructure performance to the level of that o f Costa Rica would be a reduction o f its Gini coefficient by 0.04, while increasing it to the level o f Korea would decrease the index by 0.06. InfrastructureInvestmentinArgentina 4.8 What is Argentina's recent history o f investment in infrastructure? Historical data available from Calder6n and ServCn (2004b) suggest that in the 1980s Argentina's public investment in infrastructure (defined here as water and sanitation, transport, energy and telecommunications) was around 3 percent o f GDP. Nevertheless, it fell to less than one-half of a percent o f GDP in the 1990s and went down further to about one-quarter o f a percent by the beginning o f the new millennium (Table 4.1). This negative trend o f public investment in infrastructure was only partially offset by private investment in the sector, which amounted to around 1.2 percent o f GDP on average duringthis period. Therefore, average total investment in infrastructure during the 1990s remainedunder 2 percent o f GDP in the country. Compared to other,LatinAmerican countries, these rates o f investmentwould place Argentina close to average for the 1980s but much below it for most o f the 1990s. Moreover, private investment flows into infrastructure declined further with the recent economic crisis and appear to have continued doing so in the following years. O f the private investment that occurred, over half went to the telecommunications sector so that other sectors suffered disproportionately from under- investment . 53 Figure4.1: InvestmentinInfrastructureinArgentina andLatinAmerica 1980-2000 (a) Total investment: Argentina is below LAC average 5%, 5%, 4% 4% -$3% g3% c? sO2% s 2% 1% 1% 0% 0% -1-9 countryLACa m g e +Argentina +9 wuntry LACawage +Argentina Source: Calder6n and ServCn(2004b). 4.9 Although there are no readily available consolidated government data on infrastructure investment, investment plans do provide a breakdown that i s roughly compatible with the above definition o f infrastructure. Based on the 2004-06 investment plans, it can be concluded that less than half of Argentina's public investment is destined for infrastructure (Table 4.2). Of that destined for infrastructure, the lion's share is taken up by transport which accounts for nearly one-third o f public investment, while energy and water take up about 5 percent each. Telecommunications i s a negligible0.1 percent o f overall public i n v e ~ t m e n t . ~ ~ 64Social investments inworkfare, the housing and urban sectors and education make up most o f the balance. 54 II Table 4.2: PublicInvestment inInfrastructure Share of federally financedpublicinvestment allocatedto infrastructure(YO) Transport 32.7 Energy, combustible and mining 5.8 Water and Sanitation 5.3 Communications 0.1 Total infrastructure 43.9 Yource:Fay (2005). 4.10 By benchmarking the stock o f Argentina's productive infrastructure services against those o f other countries, it becomes clear that Argentina has under-invested in electricity generation, telephony and paved roads. This comparison can be made using the results o f a cross-country regression analysis performed by Foster and Yepes (2004). For a sample o f some 80 developing countries, they consider two alternative model specifications: one which controls only for variations inper capita income, and a second one which attempts to capture other factors that contribute to or hamper service e~pansion.~'Both models suggest that Argentina has significantly less infrastructure than would be expected given its level of income and its socio- economic characteristics. Their results are presented in Table 4.3. The table reports percentage deviations between observed and predicted coverage statistics. Thus, a negative value indicates that Argentina has achieved coverage levels inferior to those predicted by the model, while a positive value indicates the opposite. Table4.3: Argentina'sEndowmentof ProductiveInfrastructure Relativeto Other Countries Electricity Fixed Cellular Overall Paved road generationcapacity teledensity teledensity teledensity density Simple model -38.4 -40.2 -30.1 -36.1 -97.9 I Full model -33.4 -38.5 -47.7 -42.8 -92.2 I 1 Source: Foster and Yepes (2004). I InfrastructureBottlenecks 4.11 Reversing the declining trend in infrastructure investment is necessary for a sustained recovery that will create employment, generate wealth and improve the quality o f life of Argentineans, particularly in the lowest deciles o f income. Argentina's competitive position has improved owing to the recent devaluation. However, infrastructure bottlenecks are seen to be impeding economic activity in a number o f areas, especially exports. Energy, transport and logistics are among the key impediments to growth. 65 In addition to per capita GDP, the expanded model includes other variables such as urbanization, income distribution, population density, growth ofurban areas, and average householdsize. 55 TheEnergy Sector66 4.12 As businesses and domestic households throughout the country depend upon a reliable supply o f electricity, approximately 50 percent o f Argentina's electricity generation in turn depends upon gas. Gas i s also a primary input for industrial producers and for an increasing proportion o f transport service providers, while families heat their homes and cook with gas. Shortages, outages, unpredictable prices, or expected capacity shortfalls ineither sub-sector thus affect every area o fthe economy. 4.13 The unprecedented growth in demand for energy has been met with increases in electricity generation capacity, gas production (which are not regulated) andgas transport (which i s regulated). The wholesale market operator (CAMMESA) has estimated risks o f supply interruption over the period 2004-06. CAMMESA's risk assessment was prepared under different scenarios o f demand growth, exports and gas availability. The projections suggested that the probability o f supply interruption were "limited" in 2004, "possible" in 2005 and "significant" in 2006. Also, it was estimated that there would be a supply deficit in 2007 ifnew generation and transmission projects were delayed.67. The government i s currently leading efforts to increase the electricity generation capacity with two 800 MW plants for which the finance i s still being finalized and may not be on stream before 2007. Gas transport capacity has been marginally increased to about 130 million m3per day from 120 m3per day but i s still a constraint. 4.14 The reason for these energy supply constraints is a convergence o f lower than (regional) average total public expenditures in infrastructure over the last decade, with a decreasing private investmenttrend noticeable during the last five years, and a suddenboom in demand due to the strong economic recovery. The demand for electricity has been driven by GDP growth and is not highly price elastic. This suggests that "normal" price adjustments alone will have only a small impact on the imbalance (Le., price adjustments will need to be very high to have an effect), and that rationing to industries without guaranteed contracts will probably increase. Thus, putting in place an environment that allows for further investment in gas production, electricity generation, and gas transport will be crucial. Transport and Logistics 4.15 Logistics costs remain very high inArgentina: they are estimated at nearly 29 percent o f GDP. InLatin America they are second only to those for Peru, and they are almost three times higher than the OECD average. Argentina's high transport and logistics costs are due to a combination of factors, including decreasing investment in transport infrastructure, poor access to provincial locations, record levels o f usage, and administrative barriers to the movement of goods. Prior to the crisis, a study o f logistics costs, excluding inventory, found that intermodal inefficiencies, customs clearance and pre-inspections were adding over US$800 in costs to a typical container moved through the port terminals o f Buenos Aires. Following the 2001 ~~ 66 Social policies directed at improving the access o f the poor to basic services and the affordability o f those services are addressed ina backgroundpaper for this report. See Fay (2005). 67 CAMMESA's projections assume average rainfall. The supply-demand gap would thus be further exacerbated by poor hydrology, as hydropower accounts for about 50 percent o f energy production. 56 devaluation, logistics costs in peso terms have been rising faster than the consumer price index and have even outstripped the producer price index. 4.16 An indication of the impact o f logistics costs on Argentina's competitiveness has been the decline of Buenos Aires' dominant role as a gateway to the Rio de la Plata region. Between 1996 and 2002, Buenos Aires lost about one-third o f its position in the container trade. In its place, the Brazilian ports o f Rio Grande and Santos have grown in importance. In 1998, after four years of significant private port investments, the combined terminals o f Buenos Aires had grown to cover 45 percent o fthe region's 2.7 million TEUtrade. By 2002, that level haddropped to 27 percent even though, in absolute numbers o f TEU, Buenos Aires was handling about the same amount as six years previously. While the traffic through Buenos Aires terminals increased by 20 percent in2003, the increase was not enough to turn around the trendinmarket share loss. 4.17 An efficient system o f logistics and freight transport will be crucial for sustainable GDP growth. Moreover, it will contribute to a more equitable distribution o f economic growth, as high logistics costs have a disproportionate impact on the competitiveness o f small-scale firms as well as firms based in the interior o f the country. For example, Guasch (2003) suggests that a 33 percent drop inArgentina's logistics burdenwould result in an increase inemployment across all firms of around 6.5 percent. IncreasingInfrastructureInvestment 4.18 Clearly, Argentina has been investing very little in infrastructure, but the question arises as to what i s an appropriate investment level. Investment "needs" depend on the objective pursued.For example, it would cost Argentina about 1percent o f GDP per year over 10 years to achieve the level o f water and sanitation, electricity and telephony (mainlines) coverage o f Korea, all o f which are reasonable objectives. However, reaching the road density o f Korea would cost the equivalent of 264 percent o f GDP, and it is clearly not a reasonable goal, given that Korea has 44 times the population density o f Argentina. 4.19 Fay and Yepes (2004) estimate the investment needed to satisfy the expected increase in demand (from firms and consumers) associated with a particular GDP growth path. Their model finds that, given a projected GDP growth for 2005 to 2010 o f 3.7 percent per annumon average, infrastructure investment requirements in Argentina would be about 1.4 percent of GDP per annum. This i s probably a lower bound figure, given that it excludes ports, airports, urban transport and gas as well as rehabilitation or upgrading costs (the estimation o f the last requires micro data that are unavailable). Adding a minimum maintenance expenditure o f about 0.8 percent o f GDP yields an estimate for total infrastructure investment o f 2.3 percent o f GDP per year, an equivalent o f about AR$10 billion (in current AR$)per annum for the period (Table 4.4). 57 Table 4.4: Lower BoundAnnual InfrastructureInvestmentNeeds for Argentina Share of GDP (%) Current AR$ billion I Investment 1.44 6.48 Maintenance 0.83 3.73 Total 2.28 10.21 Source: Fay and Yepes (2004). 4.20 How does this compare to what could be achieved under reasonable fiscal scenarios? Macroeconomic assumptions consistent with both the government's program and the authorities' expectations imply that the federal government could allocate around 1.1 percent of GDP per year to capital expenditure. This estimate is quite sensitive to key macroeconomic and public expenditure allocation assumptions. Inparticular: 0 Capital expenditure remains a small share o f total expenditures so that small reductions inthe relative share o f other important expenditure items would create the needed fiscal space to boost investment.68 a The estimates are sensitive to core macroeconomic assumptions that affect the evolution o fthe tax base as well as to changes inthe efficiency o ftax collection. 0 The estimates are sensitive to fiscal primary surplus commitments in the medium term, which are a function o ftargets consistent with debt sustainability. 4.21 Therefore, the percentage o f GDP that could be allocated to public infrastructure investment under the government's medium term macroeconomic framework would not be sufficient to meet the country's requirement to support a GDP growth rate o f 3.7 percent in the medium term. These constraints indicate that there is a need for a continued role of the private sector in infrastructure investment. Unfortunately, the private sector's participation in new capacity investments in infrastructure in Argentina i s currently limited. At present private investors inthe energy sector have mainly focused on selected gas transport pipeline investments that are crucial but are still not sufficient. At the same time, the population's perception o f private involvement ininfrastructure has become more negati~e.~' 4.22 An important factor in attracting private investment in infrastructure and protecting consumer interests is the quality o f the regulatory framework. Argentina currently faces the challenge o f strengthening its regulatory framework and renegotiating 63 federal level regulated concession contracts for infrastructure and public services which have been under stress after the 2001-02 crisis. This has proved to be a lengthy and difficult process and the emergency law enacted inresponse to the crisis has been extended through the end o f 2005. At present there are 8 renegotiated concessions sanctioned by Congress, 12 more with a tentative agreement between the Executive and the concessionaires pending Congress ratification, 13 concessions that reached For example, average capital expenditures during the period 1993-98 were only 6 percent o f total federal government expenditures, declining even further to an average4.4 percent inthe period 1999-2003. 69The results of a survey reported inEl Cronista (April 18, 2005) shows that the share of the population believing that public services should be brought back into state hands increased from 36 percent in November 2000 to 78 percent inMarch 2005. Similarly, 65 percent o f the population appears to think the state would offer better service than the privatized companies. 58 their expiry date and were re-tendered, 2 rescinded concessions, and 28 concessions which are still to be renegotiated. A full treatment o f the complex issues involved i s beyond the scope of this report. However, in constructing the medium-term framework for private participation in infrastructure, the key will be to meet the social concerns o f Argentina's citizens to have access to basic services while ensuring a transparent framework within which public and private investors can invest and develop public services and infrastructure. A later section inthis chapter will consider other policies that could improve Argentina's competitiveness and enhance the investment climate inthe country. B. FISCALPOLICIES 4.23 The role of the government in the provision o f public goods in developing countries is critical. In the face o f widespread poverty, there i s an expectation that fiscal policy will play a major role in combating poverty and reducing inequality. Yet this does not necessarily imply a large government: an economy suffers if faced with the burden o f a large, but inefficient, government. This i s likely to occur if, for example, a government imposes distortionary and inefficient taxation systems, assumes roles more appropriate for the private sector, and maintains ineffective public programs and a bloated bureaucracy. In this case, a reduction in government size would have a positive impact on economic growth. For example, substantial government cutbacks, under initial conditions o f large public debt and/or fiscal deficit, have been associated with a subsequentboom inthe economy for OECD c~untries.~'This section presents a review of the relevant literature on the role o f fiscal policy inpromoting growth and reducing poverty and inequality. The importance of macroeconomic stability for growth and poverty reduction i s discussed first. Next, the section looks at the efficacy of tax policies in reducing inequality. Finally, the role o f expenditure policies is examined. Macroeconomic Stability 4.24 As stated by Barro (1990), fiscal, monetary and financial policies that contribute to a stable macroeconomic environment and avoid financial and balance o f payments crises are crucial for long-run growth. At the same time, policies that promote high and sustained growth will have the greatest success in reducing poverty.71 Furthermore, macroeconomic stability in itself i s also likely to benefit the poor. For example, a large number o f country case studies find inequality increasing along with absolute poverty levels duringmacroeconomic crises.72 4.25 There i s international evidence indicating that inflation i s a crucial conduit through which an economic crisis affects the poor. Easterly and Fischer (2001) find that high inflation tends to worsen inequality and increase poverty. These authors use polling data for 31,869 households in 38 countries and find that the poor are more likely than the rich to mention inflation as a top national concern. They also observe that highinflation tends to lower the share intotal income of the bottom quintile, decrease the minimum wage and increase poverty. One argument that has been put forward for this "unequal" inflation effect i s that the rich are better able to protect themselves from the consequences o f higher inflation. For example, the rich may have access to financial instruments that they can use to hedge against high rates o f inflation. In contrast, the 70See Perotti (1999). 71See Dollar and Kraay (2002). 72See, for example, Lustig(2000), Baldacci et al. (2002) and World Bank (2001). 59 portfolios o f the poor are more likely to be mostly in cash. In addition, poorer households are more likely to depend on state-provided income that is not fully indexedto inflation. 4.26 Inmost cross-country paneldatastudies, inflationhasbeenfoundto increasepovertyand inequality. For a cross-section o f OECD countries in 1988, Romer and Romer (1998) find inflation to have a negative impact on the mean income o f the bottom quintile. Likewise, Lopez (2003) finds a positive association between inflation and inequality for 61 countries over the period 1960-2000. Evidence has been found for a non-linear relationship between inflation and poverty, with a reduction in hyperinflation causing a greater fall in inequality than moving from low to very low inflation.73InLatin America there i s some evidence that inflation i s likely to be detrimental to the poor. While Cardoso (1992) argues that those below the poverty line in Latin America are unlikely to be affected by an inflation tax because o f their negligible cash holdings, she does find that higher inflation i s associated with lower real wages for a panel o f seven Latin American countries. InArgentina, following the fall o f the convertibility regime inJanuary 2002, emerging inflation has been blamed in part for the dramatic increase in poverty. Rozada and Menendez (2002) use a micro-simulation approach to identify the sources o f changes inobserved inequality and poverty in Argentina. For the immediate post-convertibility period, they report that inflation had a non trivial impact on poverty. These findings suggest that, at least inthe short run,inflationhas significantly increasedpovertyinthe country.74 4.27 Argentina is now emerging from its deepest economic crisis inrecent history. Economic growth has rebounded, with real GDP growing by 8.8 percent in 2003 and 9 percent in 2004. Importantly, the economy's recovery was supported by prudent macroeconomic policies. The primary surplus o f the central government climbed from 2.3 percent o f GDP in 2003 to 3.9 percent in the next year. The fiscal situation o f the provinces also improved, with a primary surplus o f 1.4 percent o f GDP estimated for 2004. Annual inflation fell from 41.0 percent in 2002 to 3.7 percent in 2003. The maintenance o f government policies aimed at macroeconomic stability will be key to a lasting reduction inpoverty and inequality. For example, prudent fiscal policy, backed by sufficiently high primary surpluses, can reduce the risk o f crisis which disproportionately affects the poor. Tax Policies 4.28 While a tax system alone cannot be relied on to achieve redistribution, the current Argentine tax system is regressive and thereby contributes to a worsening o f the income distribution in the country. A fair distribution o f the tax burden across the income deciles i s important for equity and the incidence o f taxation should be closely studied to ensure that the burdendoes not fall heavily on the poorer strata o f society. However, the role o f tax policy, in itself, for distribution is limited inArgentina. Its primaryimportance lies rather in generating an adequate amount o f revenues with which the government may tackle inequality and poverty (through expenditure policies). In Latin America, evidence points to tax policy as not being a very effective way o f directly reducing inequality, since it is difficult to achieve large modifications in the income distribution via the tax system. Even a moderately progressive tax 73See Bulir (2001). l4These results underscore concerns with the recent upward trend in inflation, which climbed to 6.2 percent in2004 and has already reached 4.5 percent on a cumulative basis in2005. 60 systemis unlikely to havemucho f an impact on thepost-tax income distribution. This is because in Latin American countries, personal income taxes and property taxes, which have most potential for progressiveness, constitute a low share o f overall taxes compared to the more advanced economies. Moreover, the major taxes that contribute to revenues are often non- progressive flat taxes. Argentina i s not an exception. For example, VAT and the financial transactions taxes accounted for about 40 percent o f overall tax revenues in2004, compared with 23 percent for income taxes (Figure 4.2). The progressiveness o f the tax system could be increased, for example, through the elimination of VAT on basic necessities. However, a high cost would be incurred in terms o f administrative effort, and the introduction o f a system o f exemptions for certain goods would be likely to increase tax evasion. Over time, the structure o f taxation could also be changed to give more weight to income and property taxes, thereby allowing for a greater direct taxation role inredistribution. Figure 4.2: Breakdown of Federal Government Tax Revenues, 2004 Social security Others revenues ?n/ Income tax `$70 Other trade 17% taxes --. 3% m o r t taxes 10% Financial transaction tax' 32% 8% Excise and fuel tax 5% Source: Sinnott (2005). 4.29 As previously mentioned, findings from the empirical literature measuring tax incidence do not give much support to a tax policy that aims at redistribution. Several studies use household level data to estimate individual tax burdens, and then estimate the distributional impacts o f alternative policies. In one o f the first applications o f tax incidence theory, Pechman and Okner (1974) analyzed the U.S. tax system. They found that the system was not significantly progressive and did not have a large effect on the distribution o f income. The authors employed a range o f most progressive and least progressive assumptions to shift each tax and to see how the burden changed across individuals. The taxes considered were the personal income tax, the payroll tax, sales and excise taxes, corporate income taxes and property taxes. For each combination o f taxes, they calculated the effective tax rate on each household. Grouped by annual income, their results showed that the most progressive set o f assumptions did not yield a very different income distribution than the least progressive set o f assumptions. Thejustification for this outcome was that the progressive effects o f the personal income tax and the corporate income tax are largely offset by the regressive impacts o f payroll taxes, sales taxes and excise taxes. 61 4.30 I s this result particular to the United States? There i s one significant study that considers the case o f a middle-income Latin American country. Engel et al. (1999) analyze the incidence o f Chile's tax burden by income group using household data. This study could be o f interest to other developingcountries, as Chile's tax system i s something o f a model in the region inbeing efficient and subject to low corruption. The authors find that the tax system in Chile i s broadly proportional (ie., not regressive or progressive). They also undertake a simulation exercise to look at the impact of a number o f changes to the tax system on income distribution. The most progressive of these i s where they replace Chile's personal income tax by a 20 percent flat tax incorporating a high exemption level (over US$2,000 a month). The results confirm the relative insensitivity o f the income distribution to the tax system.75Moreover, the study finds that the targetingo f expenditures and the level o fthe average tax rate are more important determinants o f income distribution (after government transfers) for Chile. 4.31 For Argentina, a tax incidence study of the depth o f the Chile study is not available. Studies have generally focused on the incidence o f fiscal policy as a whole. Sabaini and Santiere (1993), AFIP (1998)' Gasparini (1998), Llach and Montoya (1999) and Rossignolo (2000) are some o f them. These studies differ intheir coverage o f taxes, the tax-incidence assumptions, and the welfare variable (income or consumption) considered. However, their findings are similar: social expenditures are found to be progressive, even when their financing i s factored in. Gasparini (2004) further argues that the distributional impact o f fiscal policy in Argentina has not changed significantly over time. The author presents a rough exercise to assess the distributional impact o f fiscal policies since 1980. The results are that the tax system has been increasingly regressive. While taxes became increasingly concentrated in the upper strata during the 1 9 9 0 this concentrationdidnot grow as rapidly as income concentrationinthe same period. ~ ~ Table 4.5: Incident ofTaxes by Quintile, 1998 Quintiles I I1 TI1 IV V Total ~ Taxes % of total tax collection 3.6 6.5 10.5 18.0 61.3 100.0 % of disposable income 43.0 36.7 35.4 36.0 35.5 36.5 Public social spending net of taxes O h of disposable income 122.1 40.6 17.8 3.2 -14.5 Distribution of disposableincome Before social policy 3.1 6.5 10.8 18.3 61.3 100.0 After public social spending 6.3 8.7 12.3 18.7 54.0 100.0 After taxes 2.9 6.5 10.9 18.4 61.3 100.0 After social policy 6.8 9.1 12.7 18.9 52.4 100.0 Source: Ministryo f Economy and Production. Direccidn de Gastos Sociales Consolidados. 4.32 The latest available estimates for tax incidence in Argentina are from the Direccidn de Gastos Sociales Consolidados (2002) of the Ministry o f Economy and Production. The results are presented in Table 4.5. The table shows that the poor pay a higher proportion o f their disposable income in taxes than the rich. However, public social expenditures offset the regressivity o f the tax system, resulting in a redistribution o f income towards the poorer quintiles. 75The authors examine the impact on the Ginicoefficient and the ratio of income of the top to the bottomdecile. 62 4.33 Finally, it i s important to mention that any government that i s pursuing an equity and growth agenda should design a tax system that i s efficient. An inefficient tax system could be very costly for an economy. For example, it could distort savings and investment decisions, thus hampering economic growth. While an in-depth analysis of the efficiency o f the Argentine tax system i s well beyond the scope o f this report, there is room for the government to improve efficiency and raise tax revenues. Tax efforts, as measured by the average tax to GDP ratio over 1990-2000, are lower in Argentina than in countries o f similar per capita income levels.76Also, the imposition o f distortionary taxes on exports and financial transactions accounts for most o f the recent increase in tax revenues, accounting for 23.1 percent o f overall tax revenues in 2004. Inthe longrun, it would be prudent to follow a strategy that replaces these taxes with revenues from, for example, an increased rate o ftax compliance. 4.34 The findings reported inthis section are central for any policy agenda that seeks to reduce poverty and inequity. What i s most important i s the amount o f revenues to be redistributed, so that a high yield proportional tax i s better than a low yield progressive tax. Accordingly, the targetingo f public spending and the relative efficiency o f the tax system are far more important than the progressivity o f the tax system. The distribution o fpublic spending i s the most effective channel through which fiscal policies could help the government fight poverty and reduce inequality. ExpenditurePolicies 4.35 It is widely acknowledged that expenditure policies are a very important tool with which to diminish inequality and fight poverty. To this end, a government wants to ensure that the composition o f expenditures promotes the reduction o f inequality and poverty through policies that increase opportunities for the poor and create well-targeted social safety nets. Currently, Argentina has a relatively highratio o f government expenditure to GDP when compared to other Latin America countries, and one o f the highest shares o f total public spending going to the social sectors.77 The spending share o f the government was 23 percent o f GDP for the consolidated national and provincial overnment in2004, with social spending as a share o ftotal public spending at over 60 percent?g While Argentina has increased social spending since the early 1990s and had some success in reaching the poor, there remains an ample margin for improving the distributional impact o f public expenditures. 4.36 Government spending to promote equity can bejustified not only on purely redistributive grounds but also in terms of its positive influence on economic growth. In neoclassical growth models, the share o f government expenditure in output or the composition o f expenditure and revenue does not influence the long-run growth rate. However, in endogenous growth models, investment in human and physical capital does affect the steady-state growth rate o f the e~onomy.~'Regarding the empirical literature, Kneller, Bleaney and Gemmell (1999), using 76See D e Ferranti et al. (2004). 77See D e Ferranti et al. (2004). 78Nevertheless, social expenditures inreal terms were still below their pre-crisis levels in2004. 79The public-policy endogenous growth models o f Barro (1990), Barro and Sala-i-Martin (1995) and Mendoza et al. (1997) give the mechanisms through which fiscal policy could influence the level o f output and the steady-state growth rate. In these models, government expenditures are differentiated according to whether or not they are 63 cross-country growth regressions, found that when financed by a combination o f taxation and reallocation away from non-productive expenditures, an increase in productive public expenditures significantly enhances economic growth." On distributional grounds, human capital build-up may offer a valuable way through which government policies could enhance equity while promoting growth. In developing countries the government has a crucial role in providing and encouraging investment in human capital. Tanzi and Chu (1992) explain how labor productivity may be enhanced by tackling the inadequate nutrition, health and education that constrain the work effort o f the poor. 4.37 A mechanism that governments generally employ to mitigate the negative effects o f poverty and inequality are social protection programs. These programs usually take the form o f direct transfers or more targeted conditional transfers. Direct transfers are more about tackling risk than inequality: typically they consist o f cash transfers to confront problems that arise due, for example, to old age poverty, disability, unemployment and/or poor health. Thus, they are a mix o f social insurance and social assistance. The regressive elements o f social welfare usually dominate the progressive elements.81 For example, those in the informal sector find it hard to benefit from many o f these transfers. This i s typically the case o f unemployment assistance (insurance), which the poorest in the informal sector find hard to access. Another important example o f exclusion is the family allowance system in Argentina, which only covers formal sector workers. These are relevant issues for the government since inArgentina informal workers account for a large share o f the working population. At the same time, the informally employed constitute the largest portion o f the bottom three deciles o f the income distribution. By contrast, Chile has put in place a non-contributory insurance system that provides a degree o f protection for informal workers. 4.38 A successful direct income transfer program was implemented by the government o f Argentina, with the support o f the World Bank, after the recent economic crisis. The Heads o f Households Program was the main social protection tool used to combat the effects o f the crisis. It aimed to give direct income support to the families who had suffered the loss o f their main source o f income due to the crisis. A cash transfer o f 150 pesos per month was given to each head o f household inthe scheme, an amount that represented approximately 50 percent o f mean household income in 2002. Galasso and Ravallion (2003) assessed the impact o f the program usinghousehold data that covered the period o fthe crisis. They found that while the programhad a substantially positive poverty impact as it effectively reached the poor (90 percent o f recipients were poor), it did not only reach those formally unemployed but about half of its beneficiaries (largely women) had not been active in the labor force before. Further, coverage rates o f the lowest income quintile reached 40 percent which i s large by international standards - but also left 60 percent o fthe poorest population quintile uncovered. included as arguments in the production function. If they are included, then they have a positive impact on growth. Ifnot, they areunproductive. "These authors define expenditures with a substantial (physical or human) capital component as productive. Social security and welfare expenditures are defined as unproductive. For taxation, income and property taxes are considered distortionary. Meanwhile, consumption (expenditure-based) taxes are defined as non-distortionary since they do not reduce the returns to investment. 81See D e Ferranti et al. (2004). 64 4.39 In addition to the Heads of Household Program, which is a workfare program as it requires recipients to work at least 4 hours per day in exchange for the benefit, Argentina also has a conditional cash transfer program. The main characteristic o f conditional cash transfers, which sets them apart from other direct cash transfers, i s that the beneficiaries in principle are required to undertake a series o f actions in order to receive the cash transfers. In many cases, these are primarily an education subsidy, with programs offering cash transfers that are conditional on regular school attendance. A foodnutrition transfer is also often built into these programs and made conditional on attending health and nutrition clinics. Evaluation has often been given much emphasis in many o f these programs, with built-in early evaluation systems. The performance o f PROGRESA in Mexico, the most widely studied conditional cash transfer program, has been encouraging, with targeting having worked and with impressive education impacts. The Familias Program in Argentina i s a type o f conditional transfer scheme, with participationbeing dependent on school/health program attendance. 4.40 Expenditure policies offer opportunities to target particular income groups, especially social spending.Nevertheless, public expenditures are not always efficiently targeted towards the poorest. Ln the case o f Argentina, an analysis o f the benefit incidence o f expenditures is revealing. Inthe 1990sthere was a shift inexpenditure towards the social sectors. Yet this higher level o f social expenditures did not translate into falling income inequality. This occurred, first, because social spending was not always progressive. Second, a large component o f social spending was in-kind, such as the provision o f public services to individuals, so it did not have an immediate impact on incomes. Third, the increase in potentially progressive social spending might have beenovershadowed by regressive spending increases inother public spending areas. For Argentina, the results of incidence analysis studies for the country are covered in detail in Bertranou (2005). She reports that some programs, such as pensions systems, tertiary education and health, have beenmostly found to be regressive. Table 4.6 provides a summary o f the results o fthese benefit incidence studies. Table 4.6: Incidence 'Social Expenditures Function Characteristic Education Elementary Progressive Secondary Progressive Tertiary Regressive Health Public assistance Progressive Other spending RegressiveiProportional Social assistance Progressive Employment Employment Programs Progressive Others Regressiveffroportional Social security Regressive Water and sanitation Proportional Other urban services Proportional Housing No pattern wrce: Bertranou (2005). 65 4.41 Finally, another recurrent problem indeveloping countries refers to the pro-cyclicality o f public expenditures. Pro-cyclical public expenditures may lead to anti-investment biases during economic downturns, limiting the role o f the government in reducing market imperfections and thereby restricting future growth. Highly pro-cyclical social expenditures also weaken the government's impact on poverty in a recession, with social spending cuts occurring at a time when there is the greatest need for this type o f expenditure. Braun and Di Gresia (2003) show that social spending. is highly pro-cyclical in Argentina, with the correlation o f the cyclical component o f GDP to the cyclical component o f social expenditures in the last three decades estimated at 0.42. This compares with an average o f 0.32 for Latin America and of -0.10 for OECD countries. 4.42 The pro-cyclicality o f total and social spending increased substantially over the recent crisis period (See Table 4.7).82 Hence, social spending was cut significantly during the recent economic contraction. However, there i s some evidence that this type of expenditure suffered relatively less than total public expenditure when compared to the past. Sinnott (2005) found that the elasticity o f social spendingto a fall in total public expenditure was lower during the recent crisis than inprevious periods of fiscal contraction. Table4.7: Elasticitiesof GovernmentSpendingto Changes inGDP, 1980-2004 1980-97 1980-89 1990-99 2000-2004 1980-2004 Total PublicSpending 0.911 0.908 0.535 2.194 1.270 (0.283)*** (0.495) (0.29) (0.868)* (0.443)*** Total Social Spending 1.243 1.886 0.740 1.854 1.412 (0.466)** (0.807)* (0.274)** (0.711)* (0.416)*** regressed on the change inGDP and a constant for each o f the time periods. *, Note: The dependent variable i s the log difference of each o f the spending categories. This is **, and *** indicate significant at 1%, 5%) and 10%. Source: Sinnott (2005). 4.43 The findings inthis section illustrate the importance for the government o f conducting on a regular basis an incidence analysis of expenditures to gauge who benefits from access to public services. The government can then target spending to make it more progressive. Here, the focus has been on social expenditures. Other government expenditures also may have substantial distributive implications. Therefore, in any examination o f fiscal incidence the government should focus additionally on the role o f the non-social expenditures, such as infrastructure, government salaries and infrastructure, in reducing inequality. Regarding the pro-cyclicality o f public expenditures, possible policy solutions include establishing appropriate institutions, such as a fiscal rule to protect social spending in bad times and/or creating automatic stabilizers for the social spending budget that would kick induringbad times. Universal coverage systems, like general child allowances, also would act as an automatic stabilizer. 82Gasparini (2004) explains the observed pro-cyclicality o f public expenditures in Argentina by the political constraints on public savings during expansions, the limited creditworthiness of the country during recessions, and the low share o f automatic stabilizers inthe budget. 66 C. TRADEPOLICIES 4.44 Argentina has taken major strides to liberalize trade since the end o f the 1980s, with a significant reduction in average tariffs and inthe percentage o f items coveredby import licenses and other quantitative restrictions. These reforms are thought to have stimulated trade and economic growth. However, some economists argue that they may also have adversely affected wages and the employment status o f unskilled workers, especially those working in industries that experienced an increase incompetition from goods produced abroad. Ifthis hypothesis were true, these policies could have contributed to rising income inequality and poverty. On the other hand, there i s a claim that further trade liberalization (in Argentina and in important partners such as the United States and the European Union) in key sectors such as agriculture and agro- manufactures could be beneficial for the country. According to this hypothesis, such policies would increase wages and employment opportunities among the poorest segments o f the population in the country. The purpose o f this section is to test the validity o f these two hypotheses. More specifically, the section investigates the relationship between trade policies and income inequality and poverty in Argentina. Inthis part o f the study not only i s the impact on poverty and inequality o f past govemment trade policies reviewed, but the hypothetical effects o f further trade liberalization plans are considered as well. TradePoliciesduringthe 1990s 4.45 Argentina's trade liberalization process has been accomplished by policies applied unilaterally, regionally and also at the multilateral level withinthe GATT/WTO fiamework. The liberalization started as a unilateral policy in 1988, when tariffs were reduced fiom an overall average o f 45 percent in 1987 to 29 percent at the end o f 1988. Regarding non-tariff measures, there was also a significant reduction in the number o f items subject to quantitative restrictions. In particular, import licenses (authorizations) for approximately 3,000 tariff items were eliminated. The new administration that took office in 1989 continued with trade liberalization initiatives. Between August 1989 and January 1991, nominal tariffs were further lowered, reaching a level o f 18 percent, and the remaining import licenses were eliminated. At the end of 1991, the average tariff fell further to an unprecedented level o f 9.7 percent. However, trade liberalization was partially reversed when, in October 1992, the government established an extraordinary and temporary non-tariff duty o f 10percent on almost all tariff items (the so-called tasa estadistica). This extraordinary levy was reduced to 3 percent at the end o f 1994. Overall, and as a consequence o f this unilateral process o f liberalization, the average tariff in Argentina was reducedfrom a level o f 45 percent in 1987 to around 12 percent by 1994.83 4.46 The profound and rapid process o f trade liberalization followed by Argentina has had a significant impact on trade flows. Inparticular, total trade increased almost four-fold between 1985-87 and 1996-98. The trade openness indicator (exports plus imports over twice-GDP) almost doubled between 1991 and 2000. This ratio was 5.3 percent in 1991 and rose to 9.0 percent in 2000 (Table 4.8). In addition, as might be expected, MERCOSUR has become an importantmarket for Argentine exports and also an important source o f imports. The proportion 83Trade liberalization was enhanced and, more important, made more credible and durable with the establishment o f the MERCOSUR Treaty between Argentina, Brazil, Paraguay and Uruguay which was signed in 1991. The objective o f the treaty was to reach free trade within the region, while imports from the rest o f the world would be subject to a common external tariff (CET) varying from 0 to a maximumof 20 percent. 67 o f exports and importsoriginating inMERCOSURcountries was 18 percent of total imports and exports in 1991and went up to 30 percent by the endo f the decade. Source: Port0 and Sanguinetti(2005). 4.47 Not only did total trade flows increase significantly during the liberalization period, but the composition o f these flows changed considerably. The shares o f imports o f crude material, fuel and lubricants, and chemical products declined, while those o f manufactured goods and machinery and equipment increased. On the other hand, exports became diversified during the 1986-99 period. For example, 47 percent o f total exports were accounted for by food and live animals in 1986. This share decreased to 34 percent in 1999. Export items that increased as a share o f total exports were minerals, fuel and lubricants, chemicals and machinery and transport equipments. 4.48 A moreprecise conclusionregarding import and export diversification canbeobtained by estimating concentration indicators. Gini coefficients for ISIC-4 and SITC-3 import and export product categories for the period 1986-99 are reported in Figure 4.3.84Various conclusions can be drawn from this eviden~e.~'First, as one would have expected for a relatively small economy such as Argentina, exports have a more concentrated structure than imports. Second, imports and exports show a secular trend towards greater diversification. Third, while the export concentration indicator declines steadily and continuously, the import concentration indexes show a structural change around 1991 and 1992 when trade-liberalization-orientedpolicies were implemented. The Gini indicator of import concentration has remained more or less constant since then. 84The ISIC-4 and SITC-3 are product classification systems. The SITC-3 is the Standard Intemational Trade Classification, Revision 3, and it i s the current release maintained by the United Nations. This five-digit classification is usedto report aggregated national trade statistics to facilitate international comparability. The ISIC- 4 is the Intemational Standard Industrial Classification, Revision 4, and i t is a four-digit economic activity classificationpublished also by the UnitedNations. I t i s employed as well to compare production and trade data for multiple countries. *'The Gini coefficient measures concentration and ranges between 0 and 1. Larger values are associated in the income distribution literature with a high-income inequality. In this context, a high coefficient i s an indicator o f a highly concentrated trade structure. 68 Figure 4.3: Evolution of the GiniIndicator of Concentration Argentina 1986-1999 +-EXPORTSSITC -f+EXPORTS-ISIC -,+IMPORTSSITC ftIMPORTS-ISIC 0.9 0.85 E 0.1 28 0 8 0.75 0.7 0.6 1986 1911 1088 1989 19W lSS1 1992 1993 1SS4 1995 18% 1997 lS91 1SW %"e: Porto and Sanguinetti (2005). 4.49 Trade policy developments in Argentina during this period could have affected employment in some industries (and thus, poverty and inequality). Between 1993 and 1998, a fall in employment took place in most manufacturing activities. Cross-industry data reveal that not all activities faced the same level o f decline. In other words, there was an important variability around the industry mean level drop o f 11.7 percent. For example, there was only a very small decline in employment inthe metal products and plastics industries (inthe latter sub- sector, employment actually increased during the period). At the other extreme, sectors such as tobacco, petroleum distillery and audio, video and communication equipment experienced a fall in employment o f over 30 percent. It seems quite plausible to associate at least part o f this phenomenon with trade openness, as the sector faced stiff competition from foreign producers as a result o f the liberalization policy and the overvaluation o f the national currency. However, no significant (unconditional) correlation between employment changes and import penetration can be found.86 4.50 Another type o f data that i s relevant to analyze the effects o f trade liberalization policies on poverty i s relative wages according to skill levels. Thus, if industries experiencing larger reductions in protection levels employed a greater proportion o f low-skilled workers, trade liberalization may have had the effect o f decreasing the wages of less-skilled workers relative to those o f high-skilled workers.87This report explores this question. More specifically, it analyzes 86See Porto and Sanguinetti (2005) for more details. 87Duringthe 199Os, the wages ofthe semi-skilled group (highschool graduates) didnot fallrelative to the wages of the unskilled group (those who at most attended high school but did not graduate), while both unskilled and semi- 69 whether deepening trade liberalization during the 1990s had any identifiable impact on the distribution o fwages. Usingmicro data, it i s possible to test whether those manufacturing sectors in which import penetration relative to gross value added deepened are, ceteris paribus, the sectors inwhich a higher increase in wage inequality by skill group occurred. The results show that the increase in import penetration in the manufacturing sector contributed to the increased wage inequality in Argentina, harming the less-skilled (unskilled and semi-skilled) workers. Nevertheless, the identified effect does not seem to be the main cause o f growing wage inequality duringthe period.88 Further Trade Liberalization Policies 4.51 What would the impact be o f further trade liberalization? Like other countries in the region and elsewhere, Argentina is currently negotiating free trade agreements (FTAs) with other countries/regions. There are talks on the Free Trade Agreements for the Americas (FTAA), the agreement between MERCOSUR and the European Union, and some initiatives aimed at facilitating trade with China. On the multilateral front, there is the new Doha round. A key ingredient for evaluating the consequences o f these policies for poverty and income inequality i s to evaluate their effect on labor markets. Assuming full employment, the main impact will be on wages, while if a "development approach" i s taken the main effect will be on employment. The analysis presented below takes this latter approach. Inparticular, it assumes that trade reforms affect poverty and income inequality mainly by affecting labor opportunities for the less educated and poor households. 4.52 Using the results o f a partial equilibrium exercise, this report analyzes the consequences o f various trade negotiations on trade flows and employment in labor-intensive industries in Argentina.89 Most o f these labor-intensive industries employ low-skilled workers relative to the rest o f industry. Thus one can be confident that shocks to labor demand in these activities, inducedby trade liberalization, have the potential to affect poor household incomes. Table 4.9 reports the results. In each case, the value o f exports, imports and the trade balance for the aggregate of all labor-intensive industriesis presented. The table also exhibits the initial situation givenby the values correspondingto the average o f2003-04. skilledwages fell relative to the wages o fthe skilled group (those that finished a tertiary degree). Indeed, the skilled- unskilled wage premium increased substantially duringthe decade. 88Porto and Sanguinetti (2005) provide the full analysis. 89 These sectors include 36 4-digit ISIC industries. In all cases the input-output coefficient associated with the utilization of labor for these activities was above the average for all industry sectors. See Porto and Sanguinetti (2005) for a complete list o f the industries. 70 Table 4.9: Employment and Trade Impact of Various Liberalization Initiatives Inmillions % Change Change in ofUS$ base year Employment BaseYear Exports 1,959 0 0 Imports 1,268 0 0 Trade balance 692 0 0 FTAA Exports 2,028 3.5 6,9 15 Imports 1,270 0.2 12 Trade balance 758 6,927 MERCOSUR-El Exports 2,000 2.1 2,339 Imports 1,315 3.8 -5,116 Trade balance 685 -2,778 FullMultilateral Exports 2,131 8.8 13,044 Imports 1,351 6.6 -8,272 Trade balance 780 4,773 50% Multilateral Exports 2,034 3.8 5,542 Imports 1,351 2.2 -2,758 738 2,784 ERCOSUR-China xports 2,009 2.5 2,430 1,316 3.7 -5,059 prade balance 693 -2,629 Source: Port0 and Sanguinetti (2005). 4.53 Exportsandimports grow inalmost all cases. Also, the trade surplus rises inall scenarios with the exception o f the MERCOSUR-EU where imports increase more than exports. However, in almost all cases the resulting changes in exports and imports are relatively modest, with the probable exception o f the full multilateral liberalization, where exports o f labor-intensive products increase by almost 9 percent while imports rise around 7 percent. The reasons for the relatively small increase in trade flows i s in part explained by the fact that tariffs for these productsare already quite low inmost o f these countries/regions (for example, the UnitedStates, Canada and the EU). Also, as a consequence o f the trade liberalization initiatives, Argentine would lose preferential access to the MERCOSUR markets, especially Brazil, where tariffs on these labor-intensive products are relatively high(above 15 percent). 4.54 The last column of Table 4.9 presents estimates o f the impact of the above changes in trade flows on employment. Ineach case, the table shows the changes in employment due to the increase in exports, imports and the net effect. It can be seen that for the FTAA and the multilateral liberalization policies the model predicts net employment creation. On the other hand, the MERCOSUR-China and the MERCOSUR-EU agreements have negative impacts on 71 employment o f about equal magnitude (2,700 jobs lost). Thus, even though the previous section concluded that an agreement with China did not have much o f an impact on the trade balance, it i s clear that when translated into employment this i s not the case. 4.55 The above results are interesting as they provide a key input for analyzing the potential impact o f various trade initiatives on employment opportunities for relatively unskilled workers. This would o f course, have direct implications for poverty and income distribution indicators. However, it i s clear that the order o f magnitude o f the computed changes inemployment suggest that the final impact would be quite small. Inthis sense, even inthe most positive scenario of the FTAA where employment increases by 7,000 jobs, this number represents only about 1percent o f the overall industry workforce in 2003-04. Furthermore, assuming that all the new employment comes from the current unemployed or sub-employed part o f the workforce, this would clearly imply a very small change inthe overall employment situation, as jobless workers currently number around 3 million people. 4.56 Finally, the potential role o f World Trade Organization (WTO) reforms on the distribution of income and poverty in Argentina i s analyzed. This country has a clear comparative advantage in agricultural products. It i s one o f the major exporters o f agricultural commodities, and it is an important seller o f agricultural manufactures. This last part explores the role of the exports o f agricultural products on the distribution of income and poverty in Argentina. More concretely, it investigates the potential role o f WTO reforms on boosting Argentine agricultural exports and the likely effects on the welfare o fthe poorest population. 4.57 This report assesses the poverty impacts o f WTO agricultural reforms in three steps. First, it calculates the changes inthe prices o f exports inducedby foreign trade reforms. Second, it estimates the wage price-elasticities that measure the responses o f Argentine wages to the price changes. Last, a procedure to update the poverty line i s developed. Based on previous findings in the literature, an estimate o fthe aggregate price change for agricultural products o fbetween 15.4 percent (upper bound) and 7.5 percent (lower bound) is used. As expected, it i s found that the prices o f exportable agricultural manufactures have a positive impact on wages for workers o f every skill and in every region." In addition, the analysis finds that the estimated wage elasticities do not vary much by skill levels or by regions: with few exceptions, the elasticities are around 0.70. 90 To uncover the regional differences, the estimations consider six regions: Greater Buenos Aires, Pampa, Northeast, Northwest, Cuyo and Patagonia. 72 POVERTY LINE Before After After Region lower bound increase upper bound increase (1) (2) (in%) (3) (in%) Greater Buenos Aires 5.08 5.10 1.93 5.12 3.97 Pampa 4.98 5.00 1.94 5.02 3.99 Northeast 4.96 4.98 2.06 5.00 4.22 Northwest 4.94 4.96 2.16 4.98 4.44 cuyo 4.95 4.96 1.86 4.98 3.82 Patagonia 5.03 5.05 1.76 5.07 3.62 4.58 To carry out the poverty analysis, the fractions o f the population that lived in poverty before and after the reform can be compared. The results are listed in Table 4.10. The main finding is that agricultural market access would cause poverty to decline in Argentina. From a national head count o f 26.1 percent, the poverty rate would decline to 25.6 percent inthe lower bound or to 24.9 percent in the upper bound. This means that between 0.5 to 1.2 percent of the population would be moved out of poverty by a set o f foreign trade reforms that bring agricultural export prices up. Even though these look like relatively small changes, the actual number of individuals affected i s not trivial: from 175,000 individuals in the lower bound to 420,000 individuals in the upper bound would exit from poverty as a result of higher market access. Moreover, poverty would decline inall regions. 4.59 Overall, the results presented in this section are interesting, as they provide a key input for analyzing the potential impact of various trade initiatives on employment opportunities for relatively unskilled workers. This would, o f course, have direct implications for poverty and income distribution indicators. The general lesson i s that trade liberalization, especially o f agricultural products under the terms o f the WTO reforms, can potentially be beneficial for the country. As shown in the last part o f the section, a liberalized intemational trade would provide new trading opportunities that would be likely to raise the national welfareg'. It should be noted, however, that inany episode o f economic reform, there will be adjustment periods and as a result there will be winners and losers. It i s the role o f domestic complementary government policies to act as an engine to propel the new trading opportunities and, at the same time, address the more important transition costs. Of course, a discussion of trade policies should also consider the existence o f high logistic costs in the country. Any policy promoting exports is dampened without proper transport channels. For more details on logistic costs in Argentina see the first section inthis chapter. 73 D. EDUCATION POLICIES 4.60 One o f the key challenges that Argentina faces today is building the skilled labor force neededto ensure that the country embarks on a patho f sustained andequitable economic growth. This section analyzes some aspects related to building skills inArgentina's labor market. More specifically, it discusses the results o f recent studies on the returns to education for workers belonging to different income groups in urban Argentina. In particular, these studies ask the question o f whether there are systematic differences (heterogeneity) between the poor and the non-poor intheir compensation for their education inthe labor market. 4.61 The heterogeneity o f educational returns is important for the design o f public policies. For example, evidence o f increasingreturns favoring the richer segments of the population calls for policies that will improve the quality o f education and enhance the job search process for the poor. Indeed, if such a pattern o f increasing returns persists, the possibilities o f mobility across the income distribution for lower income groups will be systematically limited and the value of education as an anti-poverty engine will be constrained. A vicious circle could arise if the poor did not find it profitable to invest in education because o f low expected returns. According to endogenous growth theory, this shortfall in human capital investment would in turn lower the country's growth prospects. RecentTrends inEducation 4.62 Argentina has one o f the most developed education systems in Latin America. Despite the recent economic crisis, school enrollment rates are high. The average in years o f schooling for the population was 8.8 years in 2000 (Table 4.11), significantly higher than the regional average. Argentina also compares well with Eastern and Central Europe and also East Asia, where average educational attainments in2000 were 8.4 years and 7.6 years, re~pectively.~~ -Table4.11: EducationAttainmentof the TotalPopulationAged 15 and Over, 1960-2000 AverageYears of Schooling Year Argentina Brazil Chile Mexico Uruguay Korea Canada Greece Italy Portugal Spain united States 1960 5.3 2.9 5.2 2.8 5.4 4.3 9.1 4.8 4.7 1.9 3.7 8.5 1965 5.5 3.0 5.0 2.9 5.1 5.4 8.8 5.1 5.0 2.4 3.8 9.1 1970 6.2 3.3 5.7 3.7 5.7 4.9 9.1 5.4 5.5 2.6 4.8 9.5 1975 6.3 3.0 5.6 3.9 6.2 6.6 9.8 5.9 5.7 2.8 4.7 9.7 1980 7.0 3.1 6.4 4.8 6.2 7.9 10.3 7.0 5.9 3.8 6.0 11.9 1985 7.1 3.5 6.7 5.2 6.9 8.7 10.8 7.3 6.2 3.9 5.8 11.6 1990 8.1 4.0 7.0 6.7 7.1 9.9 11.0 8.0 6.5 4.9 6.4 11.7 -arro 1995 8.5 4.5 7.3 7.0 7.3 10.6 11.4 8.3 6.9 5.5 6.8 11.9 2000 8.8 4.9 7.6 7.2 7.6 10.8 11.6 8.7 7.2 5.9 7.3 12.1 iource: andLee (2001). 92See Barro and Lee (2001). 74 4.63 The trend in the educational attainment o f the labor force has been positive in recent years. In 1992,40 percent of the labor force had a complete primary education or less, 18 percent had completed secondary school, and 12 percent had completed university. By comparison, in 2002 the proportion o f the labor force with a primary education or less had declined to 30 percent, while that o f highly educated individuals rose to slightly more than 16 percentg3.The proportion o f individuals with an intermediate education (high school or less) remained unchanged. 4.64 If the educational attainment of the population is calculated by income groups the conclusions are not so po~itive.'~Between 1992 and 2002, the average educational attainment o f the lower income group decreased. At the same time the middle and high income groups, especially the latter, increased their educational attainment. For example, in the lower income group the percentage o f highly educated workers declined from 6.8 percent to 3.8 percent between 1992 and 2002. Meanwhile, this percentage grew from 8.5 percent to 11.5 percent for the middle income group, and from 23 percent to 37 percent for the highest group. Only the secondary incomplete and tertiary incomplete categories increased for the lowest income group, while inthe middleandhighincome groups all the higher educationlevels increasedtheir share. 4.65 An interesting point is that, for the entire period, women in Argentina appear to be considerably better educated than men. For example, in 1992, while more than 16 percent o f women had completed college, this figure was only 9 percent for men. Moreover, more than 42 percent o f men had only a primary education or less, while among women this percentage was less than 35 percent. In2002, while 22.5 percent o f men had an incomplete secondary and 19 percent had a complete secondary education, the corresponding figures for women were 16 percent and 21 percent. On the other hand, 12 percent o f men had completed university compared to 23 percent o f women. That is, the educational gap between women and men widenedover this period. 4.66 Regarding the quality o f education in Argentina it i s useful to look at some examples. In the 1997 Latin American Laboratory for Assessment o f Quality in Education, a comparative study o f language and mathematics skills, third and fourth graders from 13 Latin American countries were evaluated in these skills. Students from Argentina were among the top performers, along with students from Cuba, Brazil, and Chile. Nevertheless, when compared to students in the rest o f the world, Argentine students are lagging well behind. In two recent international assessments including countries in Latin America and the rest o f the world -the Progress in International Reading Literacy Study (PIRLS) and the Program for International Student Assessment (PISA) -students from Argentina and other LatinAmerican countries were among the worst performers. InPIRLS, which tested the reading performance o f fourth graders in2001, Argentina rankedthirty-first out of a total of 35 participating countries. InPISA, where 15-year olds from participating countries are tested in mathematics, reading and science, Argentina ranked thirty-fifthinreadingskills, out o f a total o f 41 participating countries in2000. 93The same positive trend in the educational attainment o f the labor force is observed when considering a longer time series (for example, since the seventies). 94The analysis divides the sample into three income groups (high, medium, and low), based on the household per capita income distribution. In the work, educational attainment i s divided into six categories: primary incomplete, primary complete, secondary incomplete, secondary complete, tertiary (university) incomplete and tertiary complete. See Diaz and Santamaria (2005) for more details. 75 Only 15-year olds from Chile, Brazil, Macedonia, Indonesia, Albania, and Peruhad lower PISA scores in reading than students from Argentina. Inmathematics, the results were similar (Table 4.12). Table 4.12: PISA 2000 Average Scores by Country (Selected C untries) Country Reading Mathematics Country Reading Mathematics Finland 546 536 Russia 462 478 Canada 534 533 Thailand 431 432 Netherlands 532 564 Bulgaria 430 430 Australia 528 533 Romania 428 426 U.Kingdom 523 529 Mexico 422 387 Japan 522 557 Argentina 418 388 Sweden 516 510 Chile 410 384 France 505 517 Brazil 396 334 United States 504 493 Macedonia 373 381 Spain 493 476 Indonesia 371 367 Italy 487 457 Albania 349 381 Germany 484 490 Peru 327 292 came: World Bank (2005b). Returns to Education 4.67 The average return to investment inschooling inArgentina is highand has increased over the last ten years (Figure 4.4). In 2002, the average rate o f return on an additional year of schooling was 11.5 percent, while the average for middle-income countries was 10.7 per~ent.'~ Moreover, changes inthe returns to schooling by level were significant during the period 1992 to 2002. As reported by the World Bank (2005b), primary education exhibited its highest returns in 1992. Five years later the returns to primary education decreased, only to recover again in 2002 to a level slightly lower than at the beginning of the period. At the same time, the returns to incomplete secondary education have decreased over time, while the returns to complete secondary have remained remarkably stable during the period. The returns to higher education, both complete and incomplete university education, increased substantially between 1992 and 2002, with universityeducation exhibitingthe highest returns o f all education levels in200Lg6 95See Psacharopoulos and Patrinos (2004). 96As discussed inthe previous section o n trade policies, during the 1990sthere was an increase inwage differentials across skills. The result i s inline with the findings presented inthis section. 76 Figure 4.4: Average Return to Schooling Over Time (Dercent) l4 1 12 - 10 - 8 - 6 - 4 - 2 - 0; 1992 1994 1996 1998 2000 2002 Source: World Bank (2005b). 4.68 However, the returns to education could be heterogeneous across the income distribution. Inarecent work, Diaz andSantamaria (2005) addressedthis issue. Table 4.13 reports theirpoint estimates on the retums to education in urban Argentina for different levels o f per capita household income.97 ''The analysis estimates earnings equations, correcting for labor market participation (sample selection), for different groups o f workers. These groups were constructed on the basis o f the per capita income o f the household to which they belong (high, intermediate and low). Inaddition, a Basic Needs Index (BNI) constructed especially for the study (non-poor, poor and extreme poor) was employed. See Diaz and Santamaria (2005) for these complementary results. 77 Table 4.13: : eturnsto Education inArgentina, 1992-2002 Low IncomeGroup Educational level Male Female 1992 1996 2002 1992 1996 2002 PrimaryIncomplete -0.10 -0.17 -0.25 0.05 0.02 0.07 Secondary Incomplete 0.06 0.02 0.06 0.10 -0.04 -0.12 Secondary Complete 0.11 0.09 0.01 0.26 0.15 -0.08 Superior Incomplete 0.24 0.18 0.42 0.51 0.34 0.37 Superior Complete 0.23 0.57 0.48 0.38 0.45 0.52 Male Female 1992 1996 2002 1992 1996 2002 Primary Incomplete -0.19 -0.19 -0.22 -0.01 -0.13 0.04 SecondaryIncomplete 0.05 0.01 0.11 0.03 0.02 0.12 SecondaryComplete 0.19 0.13 0.20 0.23 0.13 0.25 Superior Incomplete 0.29 0.36 0.29 0.46 0.37 0.51 Superior Complete 0.46 0.54 0.58 0.53 0.63 0.82 Male Female 1992 1996 2002 1992 1996 2002 Primary Incomplete -0.37 -0.06 -0.25 -0.23 0.04 0.17 SecondaryIncomplete 0.07 0.11 0.05 0.11 0.12 0.48 SecondaryComplete 0.40 0.40 0.35 0.35 0.41 0.60 Superior Incomplete 0.48 0.59 0.70 0.43 0.61 0.70 Superior Complete 0.77 0.97 0.90 0.58 0.80 1.10 ource: Diaz and Santamaria (: 4.69 The main results o f the study are as follows: (i)relative retums to higher education (university level) compared to primary complete grew significantly over the last 12 years across the entire income distribution; (ii) relative retums to intermediate levels o f education, on the other hand, declined throughout the period or at best remained constant; (iii) retums appear to be higher for women than for men within each income group; (iv) for women, it i s very clear that those belonging to the highest segments o f the income distribution (or considered non-poor) enjoy larger retums; (v) for men, the analysis based on per capita income produced similar results to those for women but the BNI-based analyses did not provide evidence of significant differences in returns across groups, except for the last years and for the highest education level only, and (vi) both analyses were very clear in pointing out that for both men and women heterogeneity in the retums to education is only present for the higher income group. Even though the 78 point estimates o f the returns to education are not equal for the low and middle income groups, they are not statistically different.98 4.70 Why would returns to education differ across income groups? The central argument o f recent empirical studies points to a non-trivial role for unobserved worker characteristics in wage determinati~n.~~Factors such as quality o f education, family background, individual ability and labor market connections, to name a few, are distributed unevenly among workers. These factors may grant easier access to high paying jobs to certain groups, usually the non-poor, affecting both human capital accumulation and its returns. As a result, heterogeneity in labor market performance, conditional on measured worker characteristics, tends to be the norm. This heterogeneity may imply that, since the poor tend to have more restricted access to high quality education or to labor market connections, their returns to investments ineducation may in fact be lower than those observed among the non-poor. PolicyImplicationsfor Education"' 4.71 Education policy could begin by favoring the access o f the poor to good quality primary and secondary education. This i s an efficient way o f ensuring that workers coming from disadvantaged households have access to tertiary education, which i s where the highest returns can be realized. In this sense, a recent report by the World Bank (2005b) on education and skills in Argentina stresses the fact that the country faces two important challenges: raising school quality and reducing its inequality. This report argues that for the country to achieve a sustainable and more equitable economic growth, school quality and student performance need to increase. 4.72 There is considerable inequity inschool quality inArgentina. While meanstudent scores in international assessments are higher than inother Latin American countries, the dispersion in scores between the top and bottom performers is greater than inmost other countries in the region. This fact surely contributes to the observed behavior o f returns across income groups. Furthermore, the analysis o f factors associated with student learning indicates substantial diversity ineducation systems across the country, so that no one solution can improve the quality o f education for all. It i s important to understandthe realities o f each province, o f each school, and o f each particular student. Inany case, the most important policy message is that student assessments are a key tool for measuring the learning that takes place in schools, and can thus provide valuable information for teachers, parents, and policymakers. The recent Federal Council o f Culture and Education's resolution to introduce mandatory tests in the last year o f lower and upper secondary i s a promising initiative. Information from these assessments, fully analyzed, could be used to provide feedback to schools, parents, and especially teachers in areas where improved teaching practices are most needed. 98These results are quite similar to those found by the World Bank (2005b), where individuals are classified according to wage distribution. 99For example, see Maloney and Ribeiro (2001), Tannuri and Pianto (2002) and Arias et al. (2003). looSome labor market policies (like enhancing the job search process for the poor) are also essential to reduce the observed heterogeneity inthe retums to education betweenthe poor and the non-poor. However, a full analysis o f these topics i s well beyondthe scope o f this report. 79 4.73 Inthe best-case scenario, information from educational assessments would be used to provide feedback to school leaders and teachers, who would then use it to improve teacher and school management practices. Policymakers, on the other hand, would use the results to tailor teacher professional development programs to fill the gaps in teacher knowledge and pedagogical practice that would be highlighted in the assessments. Parents and communities would use the assessment results to help their local schools improve. Student assessments may also be used to reward schools and teachers that are doing a good job and to help improve schools that are not performing adequately by means o f extra incentives. Within the proposed new Education Finance Law, the national government would be taking on an increased role o f combating inequalities in education outcomes between provinces. Increased national transfers - especially to the poorer provinces- could then go hand in hand with needs- and result- based financing, in which provinces that do progress in improving education enrollment and outcomes would have access to additional financing. 4.74 Regular evaluations o f teachers and instructors are key components o f any strategy directed at improving the quality o f education. However, ifthey are not linkedto salaries and incentives provided to teachers, they are o f little use. Thus, teacher performance could be evaluated regularly and these evaluations could carry some impact on the incentives provided to teachers during their tenure, including promotions, wages and cash awards. On this subject, the recent experiences o f educationreform inColombia and Chile may prove very helpful. E. PRIVATE SECTOR DEVELOPMENT 4.75 As shown in previous chapters, the Argentina's economic recovery in 2003-04 has been remarkable. The new post devaluation environment has provided for ample opportunities for economic growth. Yet those opportunities granted by the new terms o f trade are seldom lasting, as illustratedby the experience o f other countries. It seems quite likely that the current (real) exchange rate, which has made both export and import substitution very attractive, will not be maintained in the medium term. There is a real danger, therefore, that firms that entered or expanded under these market conditions, despite doing well initially, might be unable to compete in the future. Such a situation could affect the sustainability o f current growth and induce intense pressures for protection as the exchange rate advantage subsides and firms drop out o f the export market or find it difficult to compete with cheaper imports. 4.76 In order to sustain the strong path of economic growth, it is necessary for Argentina to use the window o f opportunity created by the recovery to address a number o f important structural issues. This would allow the country to attain a more competitive position for facing external competition over the medium term. For example, the government would be advised to undertake actions to improve the investment climate by simplifying and/or reviewing unnecessary or inefficient procedures and regulations that affect the entry, operation and exit o f businesses; revamp the measuring, standards, testing and quality systems in order to increase compliance with international quality standards and regulations; increase productive innovations by strengthening the national innovation system and scaling up the efforts to promote private investments in R&D; 80 overhaul the SMEs support programs inorder to increase their impact, help supply chain integration and reduce barriers to contracting between large and small and medium enterprises (SMEs); and put in place an export friendly environment through the elimination o f anti-export biases and the development o f export promotion policies. StandardsandQualityof Argentine Goods 4.77 It is important that the window o f competitiveness that followed the devaluation be used not only to expand production so as to substitute for foreign goods, but also to improve quality standards. This would help domestic products to compete with foreign products under a more appreciated exchange rate in the medium term. For example, Table 4.14 presents the results o f a survey undertaken by INDEC o f the objectives o f investments by Argentine firms. It can be seen that in 2001 around 44 percent o f the investments were aimed at improving the quality o f the produced goods, while the share o f this objective dropped to approximately 18 percent four years later. Consistent with this feature, the investment directed at improving international competition decreased slightly inrecent years. On the other hand, the share o f investments targeted at increasing production doubled in the last four years. This is consistent with the hypothesis that quality enhancement is usually driven by competitive pressures, which was the case in 2001butwas no longer so after the devaluation. Table 4.14: Destinationsof New InvestmentsinArgentina, 2001-04 2001" 2002 2003 2004 Improvethe quality 43.6 23.9 29.7 17.9 Optimize inputsuse 17.3 20.7 18.0 20.6 To improve internationalcompetition 6.4 12.8 11.5 9.8 Increaseproduction 16.7 17.6 27.9 38.3 Reductionof nonlabor costs 7.7 7.5 4.7 3.7 Increaseinternationalparticipation 3.2 5.7 8.0 9.7 Decreasehoursworked 0.0 1.9 0.2 0.0 Decreasenumber of employees 5.1 1.1 0.0 0.0 Note: * December of 2001 Source:Guasch(2005). 4.78 Where does Argentina currently stand in terms o f quality? To answer this question, there are no clear indicators or they are difficult to impute. Often what i s done i s to use I S 0 9000 certifications, stock and flows, as a proxy (albeit imperfect) o f investment in quality. Looking at the raw numbers, the stock o f firms with I S 0 9000 certification, Argentina appears reasonable among Latin American countries. Only Brazil and Colombia do better. But when compared with other economies (Korea, Singapore, China, Malaysia), the Argentina numbers are dwarfed. Similarly, when the number o f certifications i s adjusted by value added inmanufacturing or total labor force, Argentina considerably under-performs other countries. Moreover, data show that the number o f certifications increased significantly over time, reaching a peak in 2001. After that, the numberofcertifications decreasedslightly andremainedbelow this maximum. 81 4.79 The system o f norms and standards will be a key instrument for improving the quality o f domestic products and opening foreign markets. The government has taken steps to increase the awareness o f the productive sector in this respect, and has begunto focus on sectors whose export potential can be enhanced ifproducts are certified to meet international standards. As a consequence, several "National Quality Plans" have been launched for agricultural products such as wheat, as well as for some industrial products such as wood, furniture, leather, and textiles. Complementing this, the National Institute for Technological Innovation (INTI), the first R&D institution in Argentina to have accredited laboratories, has started an initiative called "INTI seal o f quality," where certifications are voluntary. To date, INTI has an agreement to work in the precious metals sector and i s evaluating the prospects o f extending the initiative to other sectors such as clothing (labeling), leather, and textiles. Nevertheless, several limitations remain inthe system, among them the weakness o f the infrastructurethat provides accreditation and certification, a weak standardization and accreditation system, an underdeveloped and highly centralized certification infrastructure, and slow accreditation and high patenting costs. InnovationSystemin the Argentina 4.80 Technological innovation i s a key source o f economic, income and employment growth. As shown inD e Ferranti et a1(2003), most o fthe fastest growing economies over the last two decades owe their growth in large part to their ability to foster highlevels o f innovation in their economies. According to these authors, this could be achieved by facilitating access to, and assimilating and adapting, foreign technology. The same report also shows that the Latin American region as a whole does not have a good track record in technological innovation, and that Argentina underperforms significantly relative to countries with similar economies. The critical factor is the lack o f a coherent and effective National Innovation System, and o f a coherent and systemic public policy to stimulate innovation in Argentina. The final result is an insufficient research and development (R&D) effort inthe country. 4.8 1 The World Bank (2004a) country innovation brieffor Argentina, claims that firms and countries need to develop "absorptive" or "national learning" capacities which, in turn, are hypothesizedto be functions o f spending on research and development (R&D). The mainresults ofthe briefare the following: 0 Argentina i s underperforming in innovation outputs, and this applies to both commercial patents and scientific publications. Its patents and publications are below the average for countries with similar economies. 0 This underperformance in innovation is due to an insufficient R&D effort. In fact, Argentina i s underperforming relative to the typical country with a similar economic size and labor force, in terms o f R&D investments relative to GDPbut not inlicensing. 82 Its underperformance in R&D investment relative to GDP is perhaps even more severe than what appears in the aforementioned international comparisons, because the potential economic returns to R&D are quite high. Argentina also suffers from inefficiencies in its National Innovation System (NIS) which are reflected in a low rate o f transformation o f R&D into commercial applications. These inefficiencies can be fully explained by the same factors that afflict the rest o f the region, namely: poor collaboration between private firms and researchers at universities and the poor quality o f their research as perceivedby entrepreneurs. The structural roots o f Argentina's inefficient NIS need to be investigated further, but this should not prevent the use of policy experimentation to address the innovation deficits that currently limit the country's long-term prospects for technological modernization. 4.82 There are several determinants for the poor outcome o f innovation inthe country. For example, resources and human resources dedicated to R&D are scarce. In addition, participation and investment on the part o f the private sector on R&D activities is limited. Moreover, linkages and technological cooperation among firms to promote innovation are missing, and efficient technological dissemination i s lacking. Other problems are the non-existence o f links between science and technology (S&T) institutions and the productive sector, and the poor coordination o f government support for innovation. Overall, there i s considerable room for improvement in order to reach an effective innovation system in the country, a necessary element for attaining high and sustainable economic growth inthe future. 83 Table4.15: Indicatorsof InvestmentClimate Varlabie Arg Brazil Chile Col Mex Peru Ven USA LACAvg OECDAvg Startina a Business Numberof Procedures 15 17 9 14 8 10 13 5 I 1 6 Time (days) 32 152 27 43 58 98 116 5 70 25 Cost (% of inmine per capita) 15.7 117 10 27.4 16.7 36.4 15 0.6 60.4 8 Min. Capital (%of inmme per caoita) 8.1 0 0 0 15.5 0 0 0 28.9 44.1 44 67 17 72 67 44 78 0 44.4 26.2 Rhadityof Hoursindex 80 80 20 60 60 60 80 0 53.3 50 Dlcum/ of Firing Index 30 70 20 20 90 60 10 10 34.3 26.8 Rigidtyof Employment index 51 72 19 51 72 55 56 3 44 34.4 Firing Costs(week of wages) 94 165 51 49 83 56 83 8 70.8 40.4 Registering Prow Numberof Procedures 5 14 6 7 5 5 8 4 6 4 Time (days) 44 42 31 23 74 31 34 12 56 34 I Cod f% 01 wroerb oer caDital 8.8 2 1.4 3.6 5.3 3.2 1.8 0.5 5.6 4.9 Gettin Credit' .' ' ' c o z to Create aiiaterai 1%of inmme per capita) 21.3 21.4 5.3 38.9 25.7 16 7.7 0.1 19.4 5.2 Legal Rights index 3 2 4 4 2 2 4 7 3.8 6.3 Credit information index 6 6 6 4 6 6 4 6 4.7 5 Public CrednRegistry Coverage 201 76 290 0 0 143 286 0 85.7 76 2 (borrowers per 1000adults) Private BureauCowrage 733 425 220 300 382 271 0 I000 325.1 577.2 (borrowersper 1000aduits) Protecting investors DisciosureIndex 5 5 6 2 5 4 1 7 2.3 5.6 EnforcingContracts Numberof Procedures 33 25 28 37 37 35 41 17 35 19 Time (days) 520 566 305 363 421 441 445 250 462 229 Cost 1% of debt) 15 15.5 10.4 18.6 20 34.7 28.7 7.5 23.3 10.8 2.8 I O 5.6 3 1.8 3.1 4 3 3.7 1.7 Cost 1% ofestalel 18 8 18 1 18 8 38 8 15.8 6.8 RecoveryRate (centson the dollar) 235 0 2 193 546 645 313 4 9 682 266 I Source: The World Bank(2004b). InvestmentClimate 4.83 The evidence linking a good investment climate to economic performance is mounting. Countries that have improved their investment climate considerably have shown significant gains in economic activity and overall increases in competitiveness. The investment climate is defined as the set o f country and location-specific factors shaping the opportunities and incentives for firms to productively invest, to create jobs and to expand. The key factors affecting the investment climate are ease of entry and exit, enforcement o f contracts, corruption, taxation, the regulatory burden and the extent o f red tape, factor markets (labor and capital), the quality o f infrastructure, technological and innovation support, and the availability and cost o f finance. 4.84 Table 4.15 compares the Argentine indicators for the investment climate with those for the rest o f the largest countries inthe region, as well with the United States and the average for OECD. Overall, Argentina is somewhere in the middle v i s - h i s other Latin American countries, but significantly underperfoms relative to best practice benchmarks. Improvements across all indicators in Argentina will be fundamental in facilitating the desired growth sustainability. Policy Implications 4.85 To facilitate growth sustainability, the government faces the challenge o f both short term and longer term actions at the micro level. Inparticular, it might consider the following initiatives. 84 Simplification of Procedures and Deregulation of the Economy 4.86 There are a number o f unnecessary or inefficient procedures at the entry, operating and exit levels for SMEs. These procedures increase costs, both direct and indirect. To reduce the costs o f business registration and o f operating and exiting businesses, the government may wish to consider the following activities: (i)overhaul business registration procedures independently o f a broader deregulation effort, and eliminate unnecessary business formalities that represent little or no net benefit for the government; (ii) design more efficient and timely entry and exit (bankruptcy) procedures; (iii) advantage ofinformation technologies to reduce the costs andthetimeof take business formalities; (iv) create one-stop windows which may involve the private sector; (v) eliminate or simplify the requirement for labor registry (registration with social security could be sufficient); (vi) review the labor taxes, reducing those components that do not bring a direct benefit; (vii) simplify the tax regime and compliance procedures, at least for SMEs, for all levels o f government (for example, the results o f the monotributo have not been promising); and (viii) initiate customs reform and simplification o f foreign trade procedures. A Coherent and Integrated Quality System 4.87 As it has been mentioned, the systemic quality o f Argentinean products is an important issue. To address this issue, keypriorities include: Leadership: the government could consider the creation o f a National Quality Counsel, private sector led, to foster quality consciousness, to interpret the value and impact o f quality, and to be the ultimate quality advocate. Institutions: the government could revise the structure and institutions o f certification and accreditation, and the roles o f the private and public sectors inthese activities. 0 International value: the government could expand its reciprocity agreements and international recognition o f its accreditation system. Speed and client orientation in the provision o f the services. The government could also integrate and streamline the current framework for phyto-sanitary certification. 4.88 Finally, several additional measures could be put in place to facilitate the certification process. Among other things, the creation o f a register o f accredited laboratories and the publication o f a guide or catalogue to help organizations that need to hire the services of a laboratory. A Pushfor Innovation and Technological Development 4.89 Argentina needs to scale up its efforts to facilitate innovation and technological development through private investments inR&D. It also needs to improve coordination o f the jurisdiction, leadership and resource allocation for R&D activities. It would be 85 equally important to increase gradually public expenditures for R&D and develop a comprehensive program to foster innovation based on a coherent National Innovation System. Particular emphasis could be placed on the development o f a program o f fiscal and financial incentives, such as matching grants, to stimulate private investment in R&D. It would be important for the government to allocate scarce public funds for innovation resources through an incentive driven, results-based process. That allocation could mainly reward projects from consortiums with linkages between universities or scientific institutions and the private sector. To foster supply chain development and integration, resource allocation could also rewardconsortiums o f large anchor firms and SMEs firms, as well as consortiums with linkages or partners abroad that are technologically advanced. Incentive Driven Assistancefor SMEs 4.90 Argentina has a plethora of programs for SMEs and little is known o f their value and impact. The government could develop a standard set o f indicators to measure performance o f SMEs and SMEs programs. In light o f the results, it could eliminate, redesign and merge them as necessary. It could also introduce elements o f cost recovery into those programs. In addition, the government could use extensive follow-up techniques (benchmarking), facilitate and reward linkages within the supply chain, and leverage their effectiveness with use o f ICT. It would be important to involve the participation of clients inthe design o f the programs and to ensure that any governmental assistanceis linked to performance. An Integral Program of Support to FacilitateExport Growth 4.91 For exports to become a sustainable engine o f growth and employment, the government could establish a more export-friendly environment as follows: (i) couldit consider gradually eliminating the anti-export bias (for example, the taxes on exports); (ii) couldensurefaster reimbursement ofdrawbacks andVAT;(iii)couldassist, it it through the commercial attach& offices, in an aggressive program to open new markets for Argentine products, employing market intelligence, identifyingmarket products and distributors, and using country marketing; (iv) it could facilitate the development o f export financing instruments such as export insurance, pre-shipment and post-shipment finance, agreements for buyers to finance with corresponding financial institutions, warehousing and factoring; and (v) it could help in the formation o f export consortiums and export promotion schemes. 86 5. CONCLUSIONS 5.1 This report presented the main findings o f a study on economic growth, poverty and inequality inArgentina. The purpose o f the study and this report was to contribute to the ongoing dialogue in the country on how to achieve the government's goal o f sustainable and equitable economic growth. The report provided some useful analytical inputsthat could help inthe design o f a strategy that would deliver highgrowth within a framework o f social equity. This final section highlightsits main conclusions. 5.2 The second chapter emphasized the fact that economic growth in Argentina was very low duringthe second half o f the twentieth century and that, within Latin American countries, only in Venezuela and Bolivia was growth lower on average. Moreover, the chapter presented evidence that economic performance and output in the country were substantially volatile, a feature that may have prevented higher growth rates. As a result, Argentina's relative position vis-&vis the rest o f Latin America and the most developed countries severely deteriorated duringthe last 50 years. Overall, the analysis presented in the chapter supported the conventional wisdom that has been emerging with respect to growth economics over the past 30 years. More specifically, essential elements for achieving high growth rates are improving macroeconomic stability, facilitating private sector investment, strengthening the legal and institutional frameworks, having a well- functioning system for financial intermediation, promoting trade expansion and greater openness o f goods markets, and extendinghuman capital formation. 5.3 In addition, the chapter studied the evolution of poverty and inequality in Argentina. Inequality grew persistently in the country during the last several years, resulting in steadily increases inpoverty inthe face o f a stagnant and widely fluctuating per capita output. The chapter reported that poverty in 2002 was the highest on record, and that poverty's path to this point hadbeen marked by sharp peaks linked to crisis. One important conclusion was that poverty rates always increased inperiods o f recession, but they sometimes also went up inperiods o f growth. The fact that when growth took place poverty sometimes increased i s exceptional in international experience. The major findings regarding inequality in Argentina were that it augmented dramatically, though not monotonically, since 1990 and that increases in inequality were observed inperiods o fboth growth and recession. 5.4 Chapter 3 reviewed the latest episode o f economic volatility in Argentina, analyzing in detail its impact on poverty. The first section o f the chapter found, not surprisingly, that aggregate growth during the 2002-04 cycle was closely and negatively linked to poverty changes. The second section employed different methodologies to test whether the latest growth period was pro-poor, concluding that the recovery had a larger impact on the average income o f the poor than on the average income o f the rich. That is, data indicated that the recent economic recovery was truly pro-poor inboth the absolute and the relative sense. The third section presented an analysis at the sector level, while the final section identified which economic sectors contributed most to poverty reduction duringthe latest upturno f the economy. The key findings were that the pattern o f growth o f different economic sectors duringthe recent recovery was different from that observed 87 inthe 1 9 9 0 ~ ~ the new patternprobably responded to changes inrelative inputprices, that and that those economic sectors that grew most also showed the highest contribution to job creation and to poverty reduction. 5.5 Chapter 4 turned to a selected set o f policies that the government may implement to deliver stronger growth with greater social equity. The chapter observed that rates o f infrastructure investment in Argentina are lower than the average in the region. Therefore, a government policy that promotes increased infrastructure investment could help sustain the current economic recovery, foster productivity and competitiveness, and reduce inequality. The chapter also noted that macroeconomic stability i s crucial to reducing poverty rates and observed that expenditures policies are more efficient tools for redistribution than tax policies indeveloping countries. Trade liberalization policies were also found to be potentially beneficial for the country, especially a liberalization o f agricultural products under the terms o f the WTO reforms. These policies would lead to higher rates of employment, particularly for relatively unskilled workers, and this inturn would have direct implications for poverty indicators. Also, the results o f increasing returns favoring the richer segments o f the population called for policies to improve the quality o f education and to enhance the job search process for the poor. Finally, the chapter considered some government actions that are hndamental to ensure a friendly environment for private sector development. These included the following recommendations: facilitate a better investment climate; improve the quality and product mix o f the country's goods; increase productive innovations; foster supply chain integration; reduce barriers to contracting between large enterprises and SMEs; and increase linkages to sources o fknowledge, technology and innovation. 88 REFERENCES Adams, Richard H. 2003. "Economic Growth, Inequality, and Poverty: Findings from a New Data Set." World Bank PolicyResearchWorking Paper No. 2972. AFIP (Administracibn Federal de Ingresos Pfiblicos). 1998. "Estimacidn de la carga tributaria de asalariados s e g h niveles de ingreso. '' Ahluwalia, Montek S. 1976. "Inequality, Poverty and Development," Journal of Development Economics, 3: 307-342. Alesina, Alberto, and Dani Rodrik. 1992. "Distribution, Political Conflict, and Economic Growth: A Simple Theory and Some Empirical Evidence." InPolitical Economy, Growth, and Business Cycles, edited by Alex Cukierman, Zvi Hercowitz, and Leonard0 Leiderman. Cambridge and London: M I T Press. . 1994. "Distributive Politics and Economic Growth." Quarterly Journal of Economics, 109: 465-489. Arias, O., G. Yamada and L.Tejerina. 2002. "Education, Family Background and Racial Earnings InequalityinBrazil." Unpublishedmanuscript. Bhagwati, Jagdish. 1991. "The World Trading System at Risk." Princeton: Princeton UniversityPress. Baldacci, E., L. de Mello, and G. Inchauste. 2002. "Financial Crises, Poverty, and Income Distribution." Finance and Development, 39: 24-27. Banerjee, Abhijit V., and Esther Duflo. 2003. "Inequality and Growth: What Can the Data Say?" Journal of Economic Growth, 8: 267-299. Banerjee, A. and A. Newman. 1994. "Poverty, Incentives, and Development." American Economic Association Papers and Proceedings, 84: 211-215. Barro, Robert. 1990. "Government Spendingina Simple Model of EndogenousGrowth." Journal of Political Economy, 98: S103-S125. .1991. "Economic Growth ina Cross-section of Countries." QuarterlyJournal of Economics, 106: 407-443. . 2000. "Inequality and Growth in a Panel of Countries." Journal of Economic Growth, 5: 5-32. Barro, R., and J.-W. Lee. 2001. "International Data on Educational Attainment: Updates andImplications." Oxford Economics Papers 53: 541-563. 89 Barro, R., and X. Sala-i-Martin. 1995. "Economic Growth." Cambridge: MIT Press. Basu, Kaushik. 1997."Analytical Development Economics." Cambridge: MIT Press. Behrman, Jere, and Ani1 Deolalikar. 1988. "Health and Nutrition." In Handbook of Development Economics, edited by Hollis Chenery and T.N. Srinivasan. Amsterdam: Elsevier. Bknabou, Roland. 2000. "Unequal Societies: Income Distribution and the Social Contract." American Economic Review, 90: 96-129. Berensztein, Sergio, and Horacio Spector. 2003. "Business, Government, and Law." InA New Economic History of Argentina, edited by Gerard0 Della Paolera and Alan Taylor. Cambridge: Cambridge UniversityPress. Bertranou, Evelina. 2005. "Politicas Fiscales para el Desarrollo Humano: Anhlisis del Gasto Publico Social y su Impact0 sobre la Pobreza y la Desigualdad", Unpublishedmanuscript. Bertranou, Evelina and Melanie Khamis. 2005. "Labor Markets and Pro-Poor Growth.'' Background Paper for the Report "Argentina: Seeking Sustained Growth and Social Equity." Regional Office for Latin America and the Caribbean, World Bank. Bhagwati, Jagdish, and T.N. Srinivasan. 1983. "Lectures in International Trade." Cambridge: MIT Press. Birdsall, Nancy, and Juan Luis Londofio. 1997. "Asset Inequality Matters: An Assessment of the World Bank's Approach to Poverty Reduction." AEA Papers and Proceedings, 87: 32-37. Bliss, Christopher J., andNicholas H.Stern. 1978. "Productivity, Wages andNutrition: 1. The Theory. 2. Some Observations." Journal of Development Economics 5: 363- 398. Bourguignon, Franqois. 2003. "The Growth Elasticity o f Poverty Reduction: Explaining Heterogeneity across Countries and Time Periods." In Inequality and Growth. Theory and Policy Implications, edited by T.Eicher and S. Turnovsky. Cambridge: MIT Press. Brandolini, Andrea, and Nicola Rossi. 1998. Yncome Distribution and Growth in Industrial Countries." InIncome Distribution and High-Quality Growth, editedby Vito Tanzi andKe-Young Chu. Cambridge and London: M I T Press. Braun, M., and L. di Gresia. 2003. "Towards Effective Social Insurance in Latin America: The Importance o f Countercyclical Fiscal Policy." Mimeo, preparedfor the seminar, Dealing with Risk: Implementing Employment Policies Under Fiscal Constraints, Annual Meetings o f the Board o f Governors, Inter-American 90 Development Bank and Inter-American Investment Corporation, Milan, Italy, March 23,2003. Bulir, A. 2001. "The Impact of Macroeconomic Policies on the Distribution of Income." Annals of Public and CooperativeEconomics, 72: 253-270. Busso, Matias, Federico Cerimedo, and MartinCicowiez. 2004. "Pobreza, Crecimiento y Desigualdad. Descifrando la Ultima DCcada en Argentina." Unpublished manuscript. Calderh, C., P. Fajnzylber and N. Loayza. 2002. "Economic Growth in Latin America and the Caribbean: Stylized Facts, Explanations and Forecasts." Unpublished manuscript. Calderh, CCsar, and Luis ServCn. 2004a. "The Effects of Infrastructure Development on Growth and Income Distribution." The World Bank, Washington, D.C. Processed. . 2004b. "Trends in Infrastructure in Latin America, 1980-2001." The World Bank, Washington, D.C. Processed. Cardoso, E. 1992. "Inflation and Poverty." National Bureau o f Economic Research Working PaperNo. 4006. CEDLAS. 2004. "Monitoring the Socio-Economic Conditions in Argentina, Chile, Paraguay andUruguay." Unpublishedmanuscript. Coremberg, Ariel Alberto. 2004. "TFP GrowthinArgentina during the 1990s: A Stylized Fact?" Unpublishedmanuscript. Cruces, Guillemo, and Quentin T. Wodon. 2003. "Transient and chronic poverty in turbulent times: Argentina 1995-2002." Economics Bulletin, 9: 1-12. Damill, M., R. Frenkel and Roxana Maurizio. 2002. "Argentina: A Decade of Currency Board. An Analysis o f Growth, Employment, and Income Distribution." ILO Employment Paper 2002/42, Buenos Aires. Dasgupta, Partha. 1993. "An Inquiry into Well-Being and Destitution." Oxford: Oxford University Press. Datt, G., and M.Ravallion. 1992. "Growth andRedistribution Components of Changes in Poverty Measures: A Decomposition with Applications to Brazil and India inthe 1980s." Journal of Development Economics 38: 275-295. De Ferranti, D., G. Perry, I. J.L. Guasch, W.F. Maloney, C. Shnchez-Piiramo andN. Gill, Schady. 2003. "Closing the Gap in Education and Technology." World Bank, LatinAmerican and Caribbean Regional Office, Washington, D.C. 91 D e Ferranti, D., G. E. Perry, F. H. G. Ferreira and M. Walton. 2004. "Inequality inLatin America: Breaking with History?" World Bank, Latin America and Caribbean Regional Office. Washington D.C. De Gregorio, JosC, and J. Lee. 1999. "Economic Growth in Latin America: Source and Prospects." Unpublished manuscript. Deininger, K., and Lyn Squire. 1996. "A New Data Set Measuring Income Inequality." World Bank Economic Review, 10: 565-91. . 1998. "New Ways o f Looking at Old Issues: Inequality and Growth," Journal of Development Economics, 57: 259-287. De Janvry, Alain, and Elisabeth Sadoulet. 2001. "HOW Effective Has Aggregate Income Growth Been in Reducing Poverty and Inequality in Latin America?" In Social Protection and Poverty, edited by Nora Lustig. Washington D.C.: The Brookings Institution. Della Paolera, Gerardo, and Alan M. Taylor. 2003. "Introduction." InA New Economic History of Argentina, edited by Gerardo Della Paolera and Alan M. Taylor. Cambridge: Cambridge UniversityPress. Diaz Alejandro, Carlos F. 1970. "Essays on the Economic History o f the Argentine Republic." New Haven: Yale University Press. Diaz, Ana Maria and Mauricio Santamda. 2005. "Educational Wage Differentials in Argentina: Who Gets More?" Background Paper for the Report "Argentina: Seeking Sustained Growth and Social Equity." Regional Office for LatinAmerica andthe Caribbean, World Bank. Di Tella, Guido. 1986. Economic Controversies in Argentina from the 1920s to the 1940s. In The Political Economy of Argentina 1880-1946, edited by Guido Di Tella and D.C. M.Platt. New York: St. Martin's Press. Di Tella, Guido, and Manuel Zymelman. 1967. "Las etapas del desarrollo econ6mico argentino." Buenos Aires: Editorial Universitaria de Buenos Aires. Dollar, D., and A. Kraay. 2002. "Growth Is Good for the Poor." Journal of Economic Growth,7: 195-225. Drbze, Jean, and Amartya Sen. 1990. "The Political Economy o f Hunger." Oxford: ClarendonPress. . 1995. "India: Economic Development and Social Opportunity." Delhi and New York: Oxford University Press. 92 Durlauf, Steven N., and Danny T. Quah. 1999. "The New Empirics o f Economic Growth." In Handbook of Macroeconomics, edited by John Taylor and Michael Woodford. Amsterdam: North-Holland. Easterly, W., and S. Fischer. 2001. "Inflation and the Poor." Journal of Money, Credit and Banking, 33: 160-78. Elias, Victor. 1992. "Sources o f Growth: a Study of Seven Latin American Economies." San Francisco: I C s Press. Engel, E.M.R.A., A. Galetovic and C. E.Raddatz. 1999. "Taxes and IncomeDistribution in Chile: Some Unpleasant Redistributive Arithmetic." Journal of Development Studies, 59: 155-92. Fajnzylber, Pablo, and Daniel Lederman. 1999. "Economic Reforms and Total Factor Productivity Growth in Latin America and the Caribbean, 1950-95: An empirical Note."World Bank Policy Research Working Paper No. 214 Fay, Marianne. 2005. "Infrastructure, Growth and Poverty Alleviation in Argentina." Background Paper for the Report "Argentina: Seeking Sustained Growth and Social Equity.'' Regional Office for Latin America and the Caribbean, World Bank. Fay, Marianne, and Tito Yepes. 2004. "Investing inInfrastructure -What is needed from 2000 to 2010?" The World Bank, Washington, D.C. Processed. Fernhndez Valdovinos, C., A.Guerson and A. Vivanco. 2005. "Sectoral Growth in Argentina." Background Paper for the Report "Argentina: Seeking Sustained Growth and Social Equity." Regional Office for Latin America and the Caribbean, World Bank. Fields, Gary. 1991. "Growth andIncome Distribution." InEssays on Poverty, Equity, and Growth, edited by George Psacharopoulos. Oxford: PergamonPress. . 2001. "Distribution and Development: A New Look at the Developing World." Cambridge: MIT Press and Russell Sage Foundation. . 2005. "Poverty, Inequality and Economic Growth." Background Paper for the Report "Argentina: Seeking Sustained Growth and Social Equity." Regional Office for Latin America and the Caribbean, World Bank. Fields, Gary S., Paul Cichello, Samuel Freije, Marta Mentndez and David Newhouse. 2003a. "For Richer or For Poorer: Did Household Incomes Converge or Diverge inthe 1990s?" Journal of Economic Inequality, 1:67-99. . 2003b. "Household Income Dynamics: A Four Country Story." Journal of Development Studies, 40: 30-54. 93 Fields, Gary and Maria L. Shnchez Puerta. 2005. "How is Convergent Mobility Consistent with Rising Inequality? A Reconciliation in the Case o f Argentina." Background Paper for the Report "Argentina: Seeking Sustained Growth and Social Equity." Regional Office for Latin America and the Caribbean, World Bank. Forbes, Kristin. 2000. "A Reassessment o f the Relationship between Inequality and Growth." American Economic Review, 90: 869-887. Foster, Vivien, and Tito Yepes. 2004. "Where do countries stand? Benchmarking Infrastructure Performance." Background Paper for the Colombia Recent Economic Development in Infrastructure (REDI). World Bank, Washington, D.C. Processed. Galasso, Emanuela, and Martin Ravallion. 2003. "Social Protection in a Crisis: Argentina's Plan Jefes y Jefas." World Bank Policy Research Working Paper Series No. 3165. Galiani, Sebastihn, and Pablo Gerchunoff. 2003. "The Labor Market." In A New Economic History of Argentina, editedby Gerard0 DellaPaolera and Alan Taylor. Cambridge: Cambridge University Press. Galiani, Sebastian, andPablo Sanguinetti. 2003. "Wage Inequality and Trade Liberalization: Evidence from Argentina." Journal of Development Economics. Forthcoming. Gasparini, Leonardo. 1998. "Incidencia distributiva del sistema impositivo argentino. En L a Reforma Tributaria en Argentina." FIEL, Buenos Aires. . 2004. "Argentina's Distributional Failure. The Role o f Integration and Public Policies." Inter-AmericanDevelopment Bank, Working Papers Series No. 515. Gasparini, L., and Ricardo Bebczuk. 2001. "Globalisation and Inequality. The Case of Argentina." Documento de Trabajo No. 32, Departamento de Economia, Universidad Nacional de La Plata. Gasparini, L.,Manana Marchionni and Walter Sosa Escudero. 2004. "Characterization o f Inequality Changes through Microeconometric Decompositions: The Case o f Greater Buenos Aires." InTheMicroeconomics of Income Distribution Dynamics in East Asia and Latin America, edited by F.Bourguignon, F.Ferreira and N.Lustig. New York: Oxford University Press. Gomes, Victor, S. Pessoa and F. Veloso. 2003. "EvoluqZo de Produtividade Total dos Fatores na Economia Brasileira: Uma Anhlise Comparativa." Pesquisa e Planejamento EconGmico, 33: 389-434. Guasch, Josk L. 2003. "Impact o f logistic costs in Latin America". Unpublished manuscript. 94 Guasch, JosC L.2005. "The Argentinean Competitiveness: Seeking Growth Sustainability -Issues that Merit Attention." Background Paper for the Report "Argentina: Seeking Sustained Growthand Social Equity." Regional Office for Latin America and the Caribbean, World Bank. Guerson, Alejandro, and Francisco Vazquez. 2002. "Argentina: Sources of Growth inthe Nineties and Growth Prospects inthe Aftermath of the 2002 Crisis." Unpublished manuscript Hall, Robert E., and Charles I. 1999. "Why do some Countries Produce so much Jones. more Output per Worker thanOthers?" The Quarterly Journal of Economics, 114: 83-116. Hausmann, Ricardo, and Michael Gavin. 1996. "Securing Stability and Growth in a Shock-Prone Region: The Policy Challenges for Latin America." In Securing Stability and Growth in Latin America, editedby Ricardo Hausmann and Helmut Reisen. Paris: OECD. Hopenhayn, Hugo, and Pablo Neumeyer.2001. "Economic Growth inLatin America and the Caribbean: Country Study for Argentina," unpublished working paper, Universidad Torcuato DiTella. IADB (Inter-American Development Bank). 1995. "Overcoming Volatility in Latin America." In Report on Economic and Social Progress in Latin America, 1995 Washington, D.C.: IADB. Kakwani, Nanak.1994. "Poverty and Economic Growth, with Application to Cote d'1voire.I' Review of Income and Wealth, 39: 121-139. Kaldor, Nicholas. 1956. "Alternative Theories of Distribution." Review of Economic Studies,23: 83-100. , 1957. "A Model o f Economic Growth." TheEconomic Journal, 67: 591-624. . 1958. "Capital Accumulation and Economic Growth." Paper prepared for the Corfu meeting o f the Intemational Economic Association, August, 1958. Reprinted in Further Essays on Economic Theory edited by Nicholas Kaldor. New York: Holmes and Meier Publishers, 1978. Kneller, R., M. F. Bleaney and N. Gemmell. 1999. "Fiscal Policy and Growth: Evidence from OECD Countries." Journal of Public Economics, 74: 171-90. Kraay, A. 2004. "When i s Growth Pro-Poor? Evidence from a Panel o f Countries." World Bank Policy ResearchWorking No. 3225. Washington, D.C. Kuznets, Simon.1955. "Economic Growth and Income Inequality." American Economic Review, 49: 1-28. 95 Kydland, Finn, and Carlos Zarazaga. 2002. "Argentina's Lost Decade." Review of Economic Dynamics, 5: 152-165. Leibenstein, Harvey. 1957. "Economic Backwardness and Economic Growth." New York: Wiley. Lewis, W. Arthur. 1983. "Development and Distribution." Reprinted in Selected Economic Writings of Arthur Lewis, edited by Mark Gersovitz. New York: New York University Press. Li, Hongyi, and Heng-fu Zou. 1998. Yncome Inequality is Not Harmful for Growth: Theory andEvidence." Review of Development Economics, 2: 318-334. Llach, J., and Montoya, S. 1999. "En pos de la equidad. L a pobreza y la distribucibn del ingreso en el Area Metropolitana de Buenos Aires: diagnbstico y altemativas de politica." IERAL,Buenos Aires. Lopez, Humberto. 2003. "Macroeconomics and Inequality." Unpublishedmanuscript. Lustig, Nora. 2000. "Crises andthe Poor: Socially Responsible Macroeconomics." Inter- American Development Bank, Sustainable Development Department Technical Paper Series, POV-108. Lustig, N., and Ravi Kanbur. 1999. "Why is Inequality Back on the Agenda?" Paper prepared for the Annual Bank Conference on Development Economics, World Bank, Washington, D. C., April. Mahmoudi, Vahid. 2001. "Growth-Equity Decomposition o f a Change in Poverty: an Application to Iran." Unpublished manuscript. Maia, Jose Luis, and Pablo Nicholson. 2001. "El Stock de Capital y la Productividad Total de 10s Factores en la Argentina." Unpublished manuscript. Maloney, William, and Eduardo Ribeiro. 2001. "Employment and Wage Effects o f Mexican Unions. A Case o f Extreme Efficient Bargaining". The World Bank, Washington D.C. Processed. Mendoza, E., G. Milesi-Ferretti, and P.Asea. 1997. "On the effectiveness o f tax policy in altering long-run growth: Harberger's supemeutrality conjecture." Journal of Public Economics, 66: 99-126. Menendez, A., and Martin Rozada. 2002. "Why have poverty and income inequality increased so much? Argentina 1991 to 2002. Linking Poverty and Inequality with the Labour Market inArgentina." Global Development Network (GDN). Ministerio de Economia y Produccibn. 2004. "Anttlisis No. 1 - Crecimiento, Empleo y Precios." Buenos Aires: MEC. 96 Ministerio de Economia. Direcci6n de Gastos Sociales Consolidados. 2002. "El Impact0 Distributivo de la P o l k a Social en la Argentina: Analisis Basado de la Encuesta Nacional de Gastos de 10s Hogares." Documento de Trabajo No. GP/12. Morley, S. 1998. "The Impact of the Macroeconomic Environment on Urban Poverty." Seminar on Urban Poverty sponsored by ALOP and the World Bank, Rio de Janeiro. Myrdal, Gunnar. 1944."AnAmerican Dilemma". NewYork: Harper andBrothers. Nelson, Richard R. 1956. "A Theory of the Low-Level Equilibrium Trap in Underdeveloped Economies." American Economic Review, 46: 894-908. Nurkse, Ragnar. 1953. "Problems of Capital Formation in Underdeveloped Countries." New York: Oxford UniversityPress. Pechman, J. A., and B. A. Okner. 1974. "Who Bears the Tax Burden?" Washington, D.C.: Brookings Institution. Perotti, Roberto. 1992. "Income Distribution,Politics, and Growth." American Economic Review, 82: 311-316. . 1999. "Fiscal Policies in Good Times and Bad." Quarterly Journal of Economics, 114: 1399-1436. Persson, Torsten, and Guido Tabellini. 1992. "Growth, Distribution, and Politics." In Political Economy, Growth, and Business Cycles edited by A. Cukierman, 2. Hercowitz, and L.Leiderman. Cambridge and London: MIT Press. Porto, Guido and Pablo Sanguinetti. 2005. "Trade Policy, Trade Flows, Poverty and Income Inequality in Argentina." Background Paper for the Report "Argentina: Seeking Sustained Growth and Social Equity." Regional Office for LatinAmerica andthe Caribbean, World Bank. Prados, Leandro, and Isabel Sam Villarroya. 2004. "Property Rights and Argentine's Economic Backwardness inHistoricalPerspective.'' Unpublishedmanuscript. Psacharopoulos, G., and H.A. Patrinos. 1994. "Returns to Investments in Education: A FurtherUpdate." Education Economics 12: 111-134 Ramey, Garey, and Valerie Ramey. 1995."Cross-Country Evidence on the Linkbetween Volatility and Growth." American Economic Review, 85: 1138-1151. Ravallion, M. 1997. "Can High-Inequality Developing Countries Escape Absolute Poverty?" Economics Letters, Vol. 56 (1997): 51-57 .2004. "Pro-Poor Growth: A Primer." World Bank Policy Research Working PaperNo. 3242. 97 Ravallion, M., and S. Chen. 2003. "Measuring Pro-Poor Growth." World Bank Policy ResearchWorking PaperNo. 2666. Ray, Debraj. 1993. "Labor Markets, Adaptive Mechanisms, and Nutritional Status." In Development and Change: Essays in Honor of K.N.Raj, edited by P.K. Bardhan, M.DattaChoudhuri andT. Krishnan. London: OxfordUniversityPress. . 1998,"Development Economics." Princeton: Princeton University Press. Romer C. D., and D.H.Romer. 1998. "Monetary policy and the well-being o f the poor." NBERWorking PaperNo. 6793. Rossignolo, D. 2000. "Impact0 de 10s hpuestos Sobre la Distribucibn del Ingreso en la Argentina en 1997." Unpublishedmanuscript. Rozada, M., and A. Menendez. 2002. "Why Have Poverty and income inequality increasedso much? Argentina 1991-2002." Unpublishedmanuscript Sabaini, J., and J. Santiere. 1993. "Quien paga 10s impuestos en la Argentina." Centro de EstudiosTributarios (CIET), Working paper No 1073. Shchez Puerta, Maria Laura. 2005. "Stylized Facts on Poverty, Inequality andEconomic Growth in Argentina.'' Background Paper for the Report "Argentina: Seeking Sustained Growth and Social Equity." Regional Office for Latin America and the Caribbean, World Bank. Sanz Villarroya, Isabel. 2002. "La Economia Argentina en el Largo Plazo: 1875-1990." Ph.D. Dissertation. Universidad de Alcalh de Henares. Sinnott, Emily. 2005. "Fiscal Policy Options for Equity and Growth: A Reading of the Economic Literature Relevant to Argentina." Background Paper for the Report "Argentina: Seeking Sustained Growth and Social Equity." Regional Office for Latin America and the Caribbean, WorldBank. Solimano, AndrCs. 2002. "Development Cycles, Political Regimes and Intemational Migration: Argentina inthe 20thCentury." Unpublishedmanuscript. Solow, Robert. 1956. "A Contribution to the Theory o f Economic Growth." Quarterly Journal of Economics, 70, 65-94. Son, Hyun, and Nanak Kakwani. 2003. "Economic Growth and Poverty Reduction: Initial Conditions Matters." Intemational Poverty Centre Working Paper No. 2. UnitedNationsDevelopment Programme. Sturzenegger, Adolfo, and Ramiro Moya. 2003. "Economic Cycles." InA New Economic Histmy of Argentina, edited by Gerard0 Della Paolera and Alan Taylor. Cambridge: Cambridge UniversityPress. 98 Tannuri-Pianto, M., and D. Pianto. 2002. " Informal employment inBrazil - a choice at the top and segmentation at the bottom: a quantile regression approach." Working Paper No. 236, Department of Economics, University o f Brasilia. Tanzi, V., and Chu K. 1992. "La politica fiscal para un crecimiento estable y equitativo en Amkrica Latina." In Los Problemas del Desarollo en Ame`rica Latina - Homenaje a R a d Prebisch, edited by L.Montuschi and H. Singer (eds). Buenos Aires: Fondo de Cultura Econ6mica. Taylor, Alan M. 1992. "External Dependence, Demographic Burdens and Argentina Economic Decline After the Belle Epoque." Journal of Economic History, 52: 907-36. . 2003. "Capital Accumulation." InA New Economic History of Argentina, edited by Gerard0 Della Paolera and Alan Taylor. Cambridge: Cambridge University Press. Thorbecke, Erik, and C. Charumilind. 2002. "Economic Inequality and Its Socioeconomic Impact." WorldDevelopment, 30: 1477-1495. Tyler, William. 2005. "Argentina: The Record o f Past Growth." Background Paper for the Report "Argentina: Seeking Sustained Growth and Social Equity." Regional Office for Latin America andthe Caribbean, World Bank. World Bank. 2000. "Poor people in a rich country: A poverty report for Argentina." Report No. 19992-AR. Latin America and Caribbean Regional Office, Washington D.C. . 2001. "World Development Report 2000-2001: Attacking Poverty." New York: Oxford UniversityPress. . 2004a. "Country Innovation Brief: Argentina." Office of the Chief Economist for LainAmerica andthe Caribbean. . 2004b. "Doing Business in2005: Removing Obstacles to Growth." New York: OxfordUniversity Press. . 2005a. "SAL Implementation Completion Report." Report No. 32155-AR. Latin America and Caribbean Regional Office, Washington D.C. . 2005b. "Building a Skilled Labor Force for Sustained and Equitable Economic Growth: Education, Training and Labor Markets in Argentina." Report No. 31850-AR. Latin America andCaribbeanRegional Office, Washington D.C. ******** 99