w It dZ POLICY RESEARCH WORKING PAPER 2941 Income Convergence during the Disintegration of the World Economy, 1919-39 Branko Milanovic The World Bank Development Research Group u Poverty Team January 2003 POLICY RESEARCH WORKING PAPER 2941 Abstract Some economists have argued that the process of continued and even accelerated. Since the mid-19th disintegration of the world economy between the two century, incomes of rich countries tended to converge in world wars led to income divergence between the peacetime regardless of whether their economies were countries. This is in keeping with the view that economic more or less integrated. This, in turn, implies that it may integration leads to income convergence. The paper not be trade and capital and labor flows that matter for shows that the view that the period 1919-39 was income convergence but some other, less easily associated with divergence of incomes among the rich observable, forces like diffusion of information and countries is wrong. On the contrary, income convergence technology. This paper-a product of the Poverty Team, Development Research Group-is part of a larger effort in the group to study global inequality. Copies of the paper are available free from the World Bank, 1818 H Street NW, Washington, DC 20433. Please contact Patricia Sader, room MC3-556, telephone 202-473-3902, fax 202-522-1153, email address psader@worldbank.org. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The author may be contacted at bmilanovic@worldbank.org. January 2003. (35 pages) The Policy Research Working Paper Series dissemrnates the findings of work in progress to encourage the exchange of ideas about development issues. An obhecttve of the series is to get the findings out quickly, even if the presentations are less than fully polished The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors They do not necessanly represent the view of the World Bank, its Executive Directors, or the countries they represent. Produced by the Research Advisory Staff INCOME CONVERGENCE DURING THE DISINTEGRATION OF THE WORLD ECONOMY 1919-39 Branko Milanovict Development Research Group World Bank JEL classification: F02, N30 Keywords: globalization, inequality, world, inter-war history I am grateful to Prem Sangraula for excellent research assistance. I am grateful to Bernard Wasow, Mansoob Murshed, Martin Ravallion and Thomas Pogge for very helpful comments. 1. Economic integration and income convergence One of the main arguments in favor of economic integration is that, in addition to the fact that it raises incomes of all the participants, it helps proportionately more the poorer one. This is the view that has informed much of the recent literature on income convergence-whether of the conditional or unconditional variety. It is a view that has a long and distinguished pedigree in economic theory, and is supported by a fair amount of contemporary evidence. In theory, increased trade raises real incomes of all participants. But access of the poor country to superior technology embodied in goods or capital or simply through intellectual exchange, allows greater productivity gains in a poor country that is further away from the production possibility frontier. Free capital flows will also help the poor country more, by bringing in new technology and by allowing it to tap into larger savings pool of a rich country. Finally, migration too should contribute to convergence in incomes, as people from poor countries migrate to the rich. Thus, greater integration reflected in closer sharing of information and technology (knowledge spillover), more trade, greater capital flows, and labor migration should help reduce the gap between the poor and the rich. This view is behind a score of empirical papers on income convergence. The earliest papers on the convergence among industrialized countries over the period of a century beginmng in lb'10 were by Baumol (1986), and Baumol and Wolff (1988). The convergence literature continued with papers by Barro and Sala-i-Martin (1992) on convergence among OECD countries, and then among European Community members (Ben-David, 1993), individual US states (Barro and Sala-i-Martin, 1992), European regions (e.g. Cannon and Duck, 2000, p. 418), Spanish provinces (Goerlich and Mas, 2001), and so forth. 2 In all such cases, greater economic integration among units (countries or regions or states) was shown to have resulted in income convergence-as we would expect from economic theory. 2 See also the review of findings in Barro and Sala-i-Martin (1995). 2 More recently somewhat greater attention was paid to the historical process of income divergence (Maddison 1995 and 2001, Pritchett 1997) but that fact did not detract from the mainstream belief in strong causal link between economic integration and income convergence. This is because the "Great Divergence" (as named by Kenneth Pomeranz) was due to the technological breakthroughs of the Industrial Revolution, while the divergence in GDPs per capita between the countries over the last 20 years was explained away by the fact that the slow-growing (or declining) countries were precisely those that did not integrate. 3The only possible shadow was cast by those who regarded the "Great Divergence" as not something that occurred-for whatever institutional or geographical reasons-in one part the world (the "North") and then (slowly) spread to the rest, but who held that the growth and industrialization in the North were linked with the decline and disindustrialization in the South. Under the latter hypothesis, it is clearly integration that is the cause of the South's decline and divergence of incomes.4 The view is expressed in Krugman (1991), and was recently summarized by Baldwin and Martin (1999, p.7 ): At a time before the Industrial Revolution, "..regions are initially identical, so the question which region takes off is a matter of happenstance. Whichever region edges ahead initially, call it North finds itself in a virtuous circle. Higher incomes lead to a larger local market in the North and this in turn attracts relatively more investment to the North. Of course, the higher investment rate leads to a growing market-size gap and the cycle restarts.. .As the North experiences this stylized Industrial revolution, Southern industry rapidly disappears in the face of competition from northern exports. In a self- generating process, the North specializes in industry and the South in primary goods." So, at least we see that there is a possibility of economic integration leading to a decline in incomes in a part of the world and/or to divergence. The introduction of increasing retums to scale in the context of neoclassical or endogenous growth model (for a review see Easterly and Levine 2001) makes this a more realistic possibility. A similar view is forcefully made by Rodriguez and Rodrik (2000) who, based on numerous 3 For the most recent manifestation of such a view see World Bank's report on globalization (2002). 4Even if the South's decline (see Bairoch, 1997, vol. 2, pp. 549, 576, 648; also 1989, p.238) may not be viewed as the cause of the Northern success. On a more radical note, Frank (1998) arguers that the South's decline helped North's advance (Frank, 1998). 3 empirical evidence and reruns of a number of equations originally estimated by various authors, argue that economic integration and convergence are orthogonal. However, this possibility is not very seriously contemplated by many economists. The finding of income convergence among the club of the rich countries (Western Europe and its offshoots-to use Maddison's terminology) during the earlier period of globalization 1870-1913 provides empirical support for the mainstream view. 5 Then, following these results and theoretical predictions, the next period, 1919-1939-the period of retreat from globalization-is to be characterized by increasing income gaps between the countries. And indeed Lindert and Williamson (2001, p.13) write: "Real wages and living standards converged among the currently-industrialized countries between 1850 and World War r' and then for the inter-war period, "..there was no period when divergence between countries was more 'big time'. We do not yet know how much of this should be attributed to the great depression, two world wars, anti-global policies and other forces" (p. 19). Lindert and Williamson neatly summarize their results in a table where the period 1914-1950 is described as the period of retreat from globalization which widened (notice the causality) the gaps between nations. We have shown elsewhere (Milanovic, 2002) that Lindert and Williamson's claim of income convergence during the globalization phase (1870-1913) is true only for a narrow subset of countries (the Western Europe and its offshoots) and not for the world as a whole. However, this finding does not necessarily invalidate the claim that globalization did lead to income convergence because one can argue that other countries were not really integrated in the world economy. Here however we shall address the second part of their statement, namely that the process of de-globalization during the inter-war period was accompanied by (or caused) income divergence between the rich countries. Contrary to what they argue, we shall show that income convergence accelerated in the period between the two wars and did so precisely amongst the select 5See Williamson (1998, Figure 1), Lindert and Williamson (2001), O'Rourke and Williamson (1999). 4 club of rich countries that were the main leaders in globalization before the First World War, and were the main leaders in deglobalization between the two wars. 6 In the next section, we review briefly, the relatively well known facts that show that the period 1918-1939 was indeed characterized by economic disintegration. In Section 3-the main part of the paper-we show that this period witnessed fast income convergence particularly among the "most important" subset of rich economies, and in Section 4, we discuss what this finding implies for our views on the relationship between economic integration and income convergence. 2. Disintegration of the world economy 1919-1939: some facts There is little doubt that the inter-war period was characterized by disintegration of the world economy. While the disintegration movement was not entirely clear until the mid-1920, as economies recovered from the War, and would, even under the best circumstances, have taken some time to regain the levels of financial or trade integration achieved before the outbreak of the conflict, the trend is unmistakable from the mid- 1920's. There are several simultaneous developments which very clearly underline the trend. First, ideologically, protectionism was in the ascendant in Western Europe and the US. Its extreme form was achieved, of course, in autarkic systems set in place, first in the Soviet Russia, then in Italy, Germany, Spain, and gradually throughout most of Europe. It is important to stress that autarky was not viewed by the new Fascist ideologies as a reaction to other countries' unfriendly policies as it was regarded by the democracies when they engaged in competitive devaluations and tariff rises or even by the Soviet planners who faced the enmity of the capitalist powers. Autarky was viewed as a desirable attribute of a nation-the best economic policy one could pursue. 7 6 Rodriguez and Rodrik (2000, p. 47) is, to my knowledge, the only paper that shows, using Maddison's data, that a convergence among the future European Community countries continued during the inter-war period. 7The two German Four-year plans had as their objectives an increase in self-sufficiency and the development of synthetic products replacing the raw materials Germany did not produce. 5' Second, the responsive-non-ideological protectionism-was espoused by democracies during and after the Great Depression. The famous cobweb graph of world trade (Kindleberger, 1986) shows that the volume of world trade diminished for 49 consecutive months from January 1929 to February 1933-lots of it due to "beggar-thy- neighbor" policies. As a League of Nations document (1936, p. 186-7) puts it "in order to trade with countries of highly developed protectionism, it is often necessary to adopt methods complementary to their systems." In consequence, by the early 1930's, there was little doubt that the world was engaged into a period of disintegration, reflected in all the statistics (trade, capital flows, migration), but also driven by a changed ideology and by the experience of the Great Depression. The. changed ideological climate is well captured in the words of Arthur Lewis (1949, p. 155): "it was then [after the Depression] that the international system seemed finally to break down; that currency controls multiplied; that tariffs reached enormous proportions and licenses became diminutive; and that the free multilateral flow of trade was constrained into bilateral channels. All these obstacles existed in the 1920, as an aftermath of war. But while in 1920 men regarded them as temporary, looked forward to their speedy removal... .in the 1930's the obstacles came to be regarded by a much larger circle as desirable in themselves, and not just as temporary weapons for coping with a slump, but as a necessary part of national economic system." Let us consider some facts. Stagnation of trade volumes Figure 1 shows world trade in manufactured and primary products (in constant prices and annual averages) for the period 1876-1938. After a fast real increase during the heyday of globalization (late 19'h century and up to 1913), the volume of trade in manufactured products stagnated and then declined during the Great Depression. By 1938, it was still 8 percent below the 1913 level. Trade in primary products, however, continued to rise throughout the inter-war period reaching on the eve of the Second 6 World War a level some 16 percent above the one on the eve of the Great War. Basically, overall world trade, after rising steadily prior to 1913, stagnated afterwards.8 Figure 1. World trade in manufactured and primary products, 1876-1938 (armual averages, in 1913 prices, $ million) 35000 __ _____ … ___ ____ __ ___ _______ _ __ ___ Primary products 3 0 0 00 -- - - - -- - - - -- - - - -- - - - -- - - - -- - - - -- - - - -- - - - -- - - - -- - - - -- - - - -- - - - 25000 . . 2 0 0 0 0 - -- - - - - - - - - - -- - - - - - - - - - -- - - - - - - -- - - - - - - -- - - - - - - - - - - - - - -- - - - - - / ~~~~~~~~~~~~~Manufacturing products 1 5000 **------------- /77-~ - - *- S=oX>- 1 0000 -- - - - - -- - - - - -- - -- - - -- - -- --- - - - - - - - - - -- -- -- ---- - - - - - - - - - 0 1876-00 1881-85 1888-90 1891-95 1898-00 1901-05 1906-10 1911-13 1913 1921-25 1928-30 1930 1931-35 1936-38 Source: League of Nations (1945, p.157). The share of trade in GDP either stagnated or declined. For the US, and especially, the UK, the decrease is quite clear. On the eve of the Great War, trade/GDP ratio amounted to 45 percent for the UK, and 12 percent for the US. The UK numbers steadily declined and by the mid-1930's were below 30 percent. In the US, the share dropped to 8 percent (see Baldwin and Martin, 1999, p. 15). Increasing barriers to trade Generally speaking, barriers to trade increased compared to the period before the War. While there is some debate regarding the extent of protectionism in the 1920's- with Paul Bairoch (1993) holding that the period saw a declining or steady level 8 In nominal terns, trade declined quite significantly as prices of both manufactured and primary products declined. By 1938, manufactured unit price was 23 percent below the 1913 level while prices of primary products declined by more than 40 percent (League of Nations, 1945, p.157). 7 protection and Kindleberger (1989) arguing that protection was by then already higher than before the Great War-there is no dispute about the 1930's. With the Hawley-Smoot tariff act in the US, and then the Great Depression, tariffs increased worldwide. More important, and more pernicious, was the erection of numerous non-tariff barriers (NTB), and following them the advent of bilaterism in trade with individual pairs of countries negotiating tariff rates, doing barter deals and using special currencies (the most famous of which was German ASKI mark) to pay each other for exports and imports. 9 Tables 1 and 2 show different calculations of average tariff rates (trade-weighted and unweighted) in the 1920's and 1930's as compared to the pre-War period. There is little doubt that by the mid-1930's, both tariff rates and non-tariff barriers have risen manifold compared to the situation before the World War I. "' As the League of Nations (1936, p. 188) eloquently put it, "whenever trade crosses these [restricted trade] areas, and even within the area of freer trade, the present tendency seems to be for the new form of organization [protectionism] to gain ground, as if by a species of Gresham's Law." 9 Perusal of the League documents from the 1930's is indeed a melancholy exercise as both the author's and the reader's patience is taxed by a monotony enumeration of many restrictions, complicated bilateral arrangements and multitudes of exchange rates. The League of Nations continued providing very mformative annual economic Surveys until 1944. '0 While prior to World War I, quantitative restrictions were negligible, during the 1930's between 50 and 70 percent of world trade was estirated to have been subject to NTB (Crafts, 2000, p.29). 8 Table 1. Bariers to trade, 1875-1930's Tariff rates 1875 1913 1930s France 12-15 20 30 Germany 4-6 17 21 Italy 8-10 18 46 Spain 15-20 41 63 United Kingdom 0 0 na USA 40-50 44 48 Non-tariff barriers (% of all imports) France na 0 58 Germany na 0 100 Italy na 0 100 United Kingdom na 0 8 USA na 0 5 Source: Bairoch (1993), Schoot (1994), Gordon (1941), Kuwahara (1998) as reported in Crafts (2000, p. 28). Tariff rates are average tariff rates on manufactured goods. Table 2. Average unweighted tariff rate, 1913 and 1925 1913 1925 Change Argentina 26 26 0 Australia 17 25 +8 Austria 18 12 -6 Belgium 6 8 +2 Canada 18 16 -2 Czechoslovakia 18 19 +1 Denmark 9 6 -3 France 18 12 -6 Germany 12 12 0 Hungary 18 23 +5 India 4 14 +10 Italy 17 17 0 Netherlands 3 4 +1 Poland -- 23 Spain 33 44 +11 Sweden 16 13 -3 Switzerland 7 11 +4 Yugoslavia -- 23 UK -- 4 +4 USA 16 29 +13 Source: League of Nations (1927, p.15). 9 Moreover, by the mid-1930's, the world had broken into regional trading blocs. Germany established its dominance and signed bilateral treaties with a number of South- East European countries. Italy tried to do the same within its fledgling Empire. Britain introduced the system of Imperial preferences, and Japan created the East Asian Co- Prosperity zone. In addition, the United States withdrew behind the high protective wall (in 1931, the average tariff on dutiable imports was 55 percent as against 38 percent on the eve of the World War I)" and the Soviet Union, first out of necessity, and then out of ideology, led an explicit autarkic policy. Autarky, not entertained as a serious idea even by the early mercantilists, became an explicit goal and part of the ideology of the most authoritarian right-wing and Fascist movements that increasingly held sway in Europe (Italy, Germany, Spain, Poland, Baltic republics, the Balkans), and Asia (Japan). Table 3 illustrates the increasing importance of economic blocs. Table 3. Trade with 'economic blocs' as percentage of total country's trade Economic bloc Imports Exports 1920 1938 1920 1938 United Commonwealth, Kingdom colonies, protectorates 30 42 44 50 France French colonies, protectorates and 12 13 7 12 mandated territories Netherlands Dutch colonies 5.5 9 9 11 Italy Italian colonies and Ethiopia 0.5 2 2 23 Japan Korea, Formosa, Kwantung, Manchukuo 20 41 24 55 Germany South-East Europe, Latin America 16.5 28 13 24.5 Source: League of Nations (1939, p.186). Bairoch (1993) quoted in Baldwin and Martin (1999, p.14). 10 Abandonment of convertible currencies Coupled with protectionism and regional blocs was, quite naturally, the end of the Gold Standard. Table 4, taken from a League of Nations documents, charts the abandomnent of the Gold Standard and the introduction of capital controls with, in almost all cases, multiple exchange rates. As the Table shows, between December 1929 and April 1933, thirty countries, including the two most important, the UK and the US, went off the Gold Standard. Thus, the entire mechanism of freely convertible currencies and fixed exchange rates that underpinned massive increase of trade and capital flows from the mid-1850's to 1914, came to an end. Table 4. End of the Gold Standard and of convertible currencies Official Official Official Official abandonment of control of abandonment of control of the gold foreign the gold foreign standard exchange standard exchange South Africa Dec. 1932 January 1933 Greece April 1932 Sept. 1931 Germany July 1931 Hungary July 1931 Argentina Dec. 1929 October 1931 India Sept. 1931 Australia Dec. 1929 Ireland Sept. 1931 Austria April 1933 October 1931 Japan Dec. 1931 July 1932 Bolivia October 1931 October 1931 Latvia Oct. 1931 Brazil May 1931 Malaysia Sept. 1931 Bulgaria October 1931 Mexico July 1931 Canada October 1931 Nicaragua Novem. 1931 Chile April 1932 July 1932 Norway Sept. 1931 Colombia Sept. 1931 Sept. 1931 New Zealand Jan. 1932 Costa Rica January 1932 Palestine Nov. 1931 Denmark Sept. 1931 November Paraguay August 1932 1931 Egypt Sept. 1931 Peru May 1932 Ecuador February 1932 April 1934 Iran May 1932 Spain May 1931 Portugal Dec. 1932 Oct. 1922 Estonia November Romania May 1932 1931 USA March 1933 March 1933 UK Sept. 1931 Finland October 1931 El Salvador Oct. 1931 Thailand May 1932 Yugoslavia October 1931 Sweden September 1931 Turkey February 1930 Czechoslovakia September Uruguay Dec. 1929 Sept. 1931 1931 1 1 Source: Statistical Yearbook of the League of Nations 1932/33, Geneva 1933, p. 265. 11 Declining capital and laborflows As trade protectionism and nationalism were on the rise, and capital controls became the norm, intemational capital flows dried out. Before World War I, most of capital flows took the form of purchases of railway and govermment bonds. According to the data quoted by Bordo, Eichengreen and Irwin (1999, p. 30) the UK, the largest creditor nation, held 40 percent of its overseas investments in railway, and 30 percent in govemment bonds. The taste for both declined as investors faced increased political and economic hurdles and risks.12 The devastation of France and Belgium, and weakened financial position of the Great Britain, combined with huge reparations imposed on Germany, cut the potential supply of funds in the largest capitalist countries (other than the US). As Table 5 shows, foreign-held assets as a share of world GDP halved. Average (unweighted) current account deficit (or surplus) as percentage of GDP-the obverse side of capital transactions-decreased from about 4 percent in the 1870-1914 period, to as little as just over 1 percent in the 1930's (Baldwin and Martin, 1999, Figure 2, p. 9). Table 5. Estimated foreign assets/world GDP (in percent) 11870 1900 1914 1930 l 6.9 18.6 17.5 8.4 Source: Crafts (2000, Table 2.3, p. 27). The original sources given there. Labor migration which according to Williamson (1996) and O'Rourke and Williamson (1999) helped wage convergence within the Atlantic economy'3 in the late 19*' century, driving wages up in the out-migrant countries of Northern and Western Europe, and wages down in the in-migrant countries (US, Australia, New Zealand, Canada) all but stopped as the policies of the largest recipient country became much more restrictive in the early 1920's. The US immigration rate fell by almost 2/3 (see Table 6). 12 Direct foreign investments were, compared to portfolio investments, much less important before 1914 than they are today (Bordo, Eichengreen and Irwin, 1999, p. 35). 13 "Atlantic economy" is defined to include other countries of European settlement like Austrialia and New Zealand. 12 Similarly, large trans-Oceanic or continental migrations (from India to the West Indies and South-East Asia; from China to the United States and South Asia, and from Africa to North and South America) also diminished as slavery and indentured serfdom were abolished. 14 In keeping with the fact that the inter-war period was a "political" period par excellence, many of the new migrants were political, escaping first from the Bolshevik revolution, then Hitler's tyranny and Francoist revanchism. America's closed doors, and European countries' unwillingness to absorb political refugees from neighboring countries led to the burgeoning of "stateless" persons. They lived in countries of refuge unprotected, and resented. In words of the president of International Red Cross, "it is impossible that in the twentieth century, there could be 800,000 men in Europe unprotected by any legal organisation recognized by international law" (quoted in Mazower, 2000, p.63). Table 6. Immigration to the United States 1870 1890 1910 1930 Imnigration rate 6.4 9.2 10.4 3.5 (per 1000 population) Foreign bom as % 13.9 14.6 14.6 11.5 of population I I I Source: Crafts (2000, p. 30). The original source is the US Bureau of the Census. Conclusion All these data-decline in capital flows and migration, stagnation in trade, increased obstacles to free trade-show a clear pattern of economic disintegration during the period between the two World Wars. Trade as a share of GDP decreased, as did capital flows and the importance of foreign-held assets. Trade barriers increased in all major countries, and in some of them-which strived for autarky-they became practically impassable. Bilateral trade, championed by Germany, increased all around. And, of course, underlining all of this was the fact that currencies ceased to be convertible, and migration virtually stopped-except for the trickle driven by political conflicts. 14 Between 1811 and 1870, about 2 million slaves from Africa were sent to the Americas (Bairoch, 1997, vol. 2, p. 691). 13 3. What happened to income convergence? Using Maddison 's data Such a violent process of disintegration of the world economy is, according to neoclassical economic theory, expected to lead to a slowdown in growth, and-what is important for our purposes-to affect disproportionately poorer countries. As the world economy disintegrates and trade and capital and labor flows dry out, the poorer countries would lose many advantages associated with greater economic integration: ability to use foreign technology, to receive capital, to export people and goods. All but the last are supposed to be greater for the poor members, since in a neoclassical world they benefit from easy application of the already known technology, and are supposed to be recipients of capital, and exporters of people. Did then incomes really diverge during the inter-war years? We calculate Gini, Theil and coefficient of variation for unweighted GDPs per capita of twenty major Westem counties (the WENAO: Westem Europe, North America and Oceania)'5 and also for a more restricted sample of the countries of the Atlantic economy (as called by Lindert and Williamson).16 Gini and Theil are, of course, measures of inequality, closely related to the coefficient of variation (a) which is often used in convergence discussions (so called "sigma" convergence). We prefer Gini or Theil because they are better established measures of inequality and also allow us to move easily between measurement of inter-national,17 domestic, and global inequality (the latter is inequality among all individuals in the world). It was the Gini coefficient that was originally used to measure convergence (Summers, Kravis and Heston, 1984) until it was overtaken by 1B and ca convergence. We calculate however all three measures of inequality as shown in '5 Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Netherlands, New Zealand, Norway, Portugal, Spain, Sweden, Switzerland, United Kingdom and the US. 16 That is, the same WENAO countries from the previous footnote minus Greece, Portugal and Spain. This is the group called "Western Europe and its offshoots" by Maddison (2001). 17 Inter-national inequality comes in two "fonns": Concept I inequality which we calculate here where each country's GDP per capita is assigned the same weight, or Concept 2 inequality, where each GDP per capita is weighted by country's population (see Milanovic, 2002a). 14 Figures 2 and 3. The data on GDP per capita (expressed in 1990 Geary-Khamis dollars) are taken from Maddison (2001). Figure 2 shows that using either Gini, Theil or the coefficient of variation, we find that WENAO incomes (GDPs per capita) did not diverge during the inter-war period. If anything, there was a mild convergence. In 1919, the Gini, Theil and coefficient of variation were respectively 19, 0.06 and 0.34; in 1939, they were 15, 0.04 and 0.27. 18 It is the Second World War which wrought a massive disruption of economic activity in a number of continental European countries (between 1939 and 1945, Germany's GDP per capita decreased by 23 percent, France's by almost 50 percent, Greece's by two-thirds etc.). On the other hand, the US, Canada, and Australia surged ahead (by respectively 78, 50 and 18 percent), thus widening differences in GDPs per capita and "creating" the divergence. The Gini went up from 15 just before the outbreak of the War to 31 at its end; the coefficient of variation from 0.27 to 0.58. Of course, income divergence is not unique to the Second World War. The same divergence in incomes, albeit of a smaller size, occurred during the First World War (see Figures 2 and 3). 1 The Gini coefficient is expressed in percents. 15 Figure 2. Gini, Theil, coefficient of variation of GDP per capita (WENAO countries, 1820-950) 070- _ . Vorid World O 60 'i Varld I . War 11 0560 0450 040O A;4oefoeto aiabont 030-A A O s0 ~~~ ~ ~ ~~~~Gini 020 - 9 7e 010 le I'l 11 -.,AP 1, Rleele ,* le , e 'I, , le e@ lse ,s le lsee 'IOe 146 1.1 ,se2> 'IO ,e ,s'I Figure 3. Gini, Theil, coefficient of variation of GDP per capita (Atlantic economy, 1820-1950) o eoo _ . . . Wodd World War I War 11 0 300 -- - - - - - - - - - - - - - - - - - - - -- - - - - - -- - - - - - 0.500 ---- --- .. .......... .. .. ... ........... / ~~~~~~~Theil 0 000 e b'° 4' 4P e P ee ee ee ef ,# ,16 e e e e le 8 lM g w l 16 But much more dramatic and telling is the example of the Atlantic economy. Figure 3 shows that the inter-war period witnessed thefastest income convergence ever recorded up to then. The negative slope of the Gini, Theil and coefficient of variation line is much steeper than during the heyday of the first globalization. The Gini coefficient in 1918 was 20; on the eve of the Second World War, it has almost halved: it was only 11. The coefficient of variation decreased from 0.38 in 1918 to 0.2 twenty years later. This is all the more interesting since it is with respect to this group that Lindert and Williamson claim that disintegration of world economy led to income divergence. As can be easily checked, their mistake stems from a comparison of 1913 and 1945. Indeed, income differences in 1945 were greater than in 1913, but that was entirely due to the huge difference in fortune during the Second World War. Ascribing the increase in between- country inequality to the developments during the inter-war period is entirely wrong. The Gini coefficient of GDPs per capita of the Atlantic economy countries declined almost uninterruptedly between 1919 and 1939. Their incomes have never after 1870 been more similar than on the eve of the Second World War. As Table 7 illustrates, income differences between the countries of the Atlantic economy in 1939 were about the same as in 1973. It is only during the last thirty years that the differences between these countries have shrunk below the level that obtained in 1939. The Gini coefficient of GDPs per capita was (as already mentioned) only 11 on the eve of the Second World War, 11.2 on the eve of the first oil crisis, and 6.4 today. 19So, if we look at how wide was the dispersal of the Atlantic economies' incomes in the past compared to today, it is only with greatest difficulty that we can discern some difference until rather recently. And note that we are comparing the two situations (in 1939 and 1973) where trade and capital flows and even more so, trade ratios were vastly different. Incomes dispersion in a heavily integrated and a heavily disintegrated economy are thus shown to be fairly similar. 19 In 1939, income ratio between the richest and poorest country was about 2 to 1. It widened to about 3 to 1 in 1973, and then shrunk to less than 2 to 1 today. 17 Table 7. Incomes in the Atlantic economy in 1939, 1973 and 1999 GDP per Relative GDP per Relative GDP per Relative capita in GDP per capita in GDP per capita in GDP per 1939 capita 1973 capita 1999 capita (1990 PPP (richest (1990 PPP (richest (1995 PPP (richest dollars) country-l) dollars) country=i) dollars) country=1) USA 6568 1.00 20106 0.89 30610 1.00 New Zealand 6492 0.99 13653 0.60 16660 0.54 Switzerland 6273 0.96 22674 1.00 26760 0.87 UK 5979 0.91 13469 0.59 20983 0.69 Denmark 5766 0.88 14688 0.65 23252 0.76 Australia 5631 0.86 13343 0.59 21173 0.69 Germany 5549 0.84 15199 0.67 21340 0.70 Netherlands 5409 0.82 14085 0.62 22469 0.73 Belgium 5040 0.77 14778 0.65 23668 0.77 Sweden 5029 0.77 14248 0.63 20339 0.66 France 4748 0.72 14671 0.65 22848 0.75 Canada 4518 0.69 15461 0.68 23162 0.76 Austria 4123 0.63 13414 0.59 23229 0.76 Norway 4108 0.63 11459 0.51 24074 0.79 Italy 3444 0.52 12360 0.55 20720 0.68 Finland 3310 0.50 12290 0.54 20985 0.69 Ireland 1/ 3116 0.47 7036 0.31 22271 0.73 Gini 11.0 11.2 6.4 Theil 0.019 0.025 0.008 Richest- poorest ratio 2.1 3.2 1.9 Sources: Data for 1939 from Maddison (1995). Data for 1999 from World Bank SIMA database. Countries ranked by GDP per capita in 1939. The approximate conversion between 1995 and 1990 PPP dollars is 1.16 to 1. "Atlantic" economy includes New Zealand and Australia. 1/ Data for 1938. We can test for convergence also using more standard regressions test. As is conventionally done, we regress growth rate of GDP per capita (change in income logs) on initial level of income (yi, t-1) where i indicates country subscript, and t time, 20 20 This formulation is rife with problems. Other than the most obvious econpometric problems of omitted variable and endogeneity (which we also address below), formulation such as (1) suffers from Galton's fallacy (see Quah, 1993, Bliss 1999), weakness of empirical tests used to test for the 3 convergence so that beta convergence can be observed both when one moves forward and backward in time and can exist whether the underlying distribution diverges, converges or stays the same (see Wodon and Yitzhaki, 2002), and even interpreatation of the obtained results (Quah, 1996). We use it because it is the simplest and the most commonly used formulation in the (immense) literature on convergence. As explained above, our view is that direct tests of unconditional convergence (as implied in the calculation of the Gini coefficients) are far superior to the regression analysis. 18 (1) lnya -lny,,t -i =lo+,llnyz,,t-I+ 2hnZe+ui+v:+eet and hn Zit = In (n,t+g+6) where n1t= population growth rate, g=rate of labor augmenting technological progress and 6=depreciation rate (all derived from the textbook Solow model of economic growth), and ui, vt, and e,t, country- time- and both-dependent error term. All GDP per capita values are taken at 5 year intervals, and thus the growth rate (the dependent variable) is the average growth rate over a five year period. Equations such as (1) potentially suffer from a number of econometric problems. The most obvious are the omitted variable bias where relevant country-specific information is not included, 21 and endogeneity where the dependent and independent variables are jointly determined. We thus run three formulations of (1). The results are shown in Table 8. In the first formulation, we simply run a pooled regression using indiscriminately cross-section and time-series data. For the pre-1914 period, we see no evidence of convergence; for the inter-war period, we obtain a statistically significant and negative coefficient on initial income. In the second formulation, we address potential endogeneity by instrumenting the right-hand side variables by their lagged values. No variable is still significant for the pre-1914 period; for the inter-war period, the coefficient on the initial income declines (as we would expect) and becomes significant at 1 percent level. Finally, in the third formulation, we adjust for country-specific effects by estimating a fixed-effects model. The results are again the same: the pre-1914 period exhibits no convergence, the inter-war period does, and a very strong one. 21 Since our model is by necessity (since other relevant variables like investment rate, education levels etc. are unavailable) very stripped-down, there are strong grounds to believe that relevant country-specific features are ornitted. 19 Table 8. Convergence in the two periods (1870-1913 and 1918-1938) (dependent variable: annualized GDP per capita growth over the five-year interval) Pooled regression IV regression Fixed effects 1870-1913 1918-1938 1870-1913 1919-1938 1870-1913 1919-1938 In y, -0.001 -0.018* -0.016 -0.013** -0.0009 -0.057* (0.78) (0.02) (0.68) (0.01) (0.91) (0.02) ln(n+8+X) -0.008 0.025 -0.002 0.024 -0.011 0.073 (0.32) (0.38) (0.89) (0.12) (0.46) (0.12) Constant -0.0007 0.235* 0.02 0.19 -0.01 0.69** (0.99) (0.02) (0.70) (0.10) (0.88) (0.008) No. of 127 90 109 69 127 90 observations R2 0.01 0.05 0.00 0.13 0.01 0.07 Note: For most of the pre- 1900 years, the annualized growth rate is calculated over a ten-year interval smce GDP per capita data are available at such intervals only. Growth rate of population is calculated in exactly the same fashion, and over the same period, as that of GDP per capita. In conclusion, the standard convergence regressions confirm that the evidence of convergence is stronger during the inter-war period, while it is entirely absent for the pre- 1914 period. Using Bairoch and Prados de la Escosura 's data In addition to Maddison's data which are the most complete, we have two other GDP per capita series that cover the period 1870-1939. They are Bairoch's (1997) data, and those produced by Prados de la Escosura (2000).22 Figures 4 and 5 show the Gini and Theil coefficients using these alternative sources, and covering the same set of countries. For the period 1870-1938, the country coverage in the three databases (Maddison, Bairoch and Prados de la Escosura) is practically the same (see Annex). For the period before 1870, Prados de la Escosura's coverage is more limited (13 or 15 countries vs. 19 for Bairoch and Maddison). Bairoch and Prados de la Escosura data are available only for selected years. Using Bairoch's series, we find that both Gini and Theil indexes are stable between 1890 22 Bairoch's GDPs per capita are given in 1960 mternational dollars. Prados de la Escosura's are expressed in current dollars of equal purchasing power panty, so that between-country comparisons for a given year are possible, but not comparisons between the years. The data base is scaled (for each year) in such a way that the US GDP per capita is equal to 1. 20 and 1929, and then display a very strong income convergence between 1929 and 1939. Using Prados de la Escosura's data, there is a convergence between 1860 and 1913, and then divergence during the inter-war years.23 As a glance at Figures 2-5 reveals, original income divergence, according to Bairoch, is much sharper and seems to have lasted longer than the one obtained from Maddison's data. According to Bairoch, divergence starts around 1800 and goes on, almost without interruption, until 1890. After that, inequality is stable until the Great Depression, and only during the last decade before the World War II, there is convergence. If we look at Maddison's data, however, the divergence begins in 1820 (when his series originate) and reaches its peak around 1880. After that, there is at first a slow, and then a faster convergence until the First World War. The inter-war period is characterized by a mild convergence. 24 Table 9 summarizes the findings regarding convergence and divergence using the data from the three authors. 23 Prados de la Escosura data are obtained by the so-called "short-cut" method, that is from a regression between the price level (purchasing power exchange rate over market exchange rate) on the LHS, and GDP per capita at current exchange rate and several other controls (openness, current account balance) on the RHS. The regression is run, of course, only for the countries for which the data are available. The estimated parameters from such an equation together with values for each independent variable are then used to predict the price level (that is, PPP) for the missing years and countries (see Prados de la Escosura, 2000, pp. 8-1 1). The fact that Prados de la Escosura data show income divergence while both Bairoch and Maddison show income convergence may be explained by the use of current PPPs by Prados de la Escosura. The implication is that prices of non-tradables have increased more in rich than in poor countries. 24 The increase in inequality following the Industrial Revolution is much greater if one uses Bairoch's rather than Maddison's data. According to Maddison, the Gini in 1820 was 12 (see Figure 1). According to Bairoch, it was (for the same set of countries) only 6 in 1800 and 9 in 1830. This is due to the fact that Bairoch's data show poor WENAO countries with (relatively) higher GDPs per capita than Maddison's. For example, in 1820, the ratio between the richest and poorest WENAO country (UK and Finland) is 2.3 to 1 in Maddison's data, but in Bairoch's, it is only 1.9 to 1 in 1830 and 1.3 to 1 in 1800 (in both cases, UK vs. Finland). In general, Bairoch's estimates of (relative) income per capita of the future less developed countries at the time of the Industrial Revolution are generally higher than Maddison's. 21 Table 9. Income convergence and divergence according to different authors Beginning of the 19th 1870 to 1913 Between the two wars century to 1870 (Heyday of (Modern era) globalization) (Deglobalization) Maddison Strong divergence Convergence Mild convergence (strong convergence for Atlantic economy) Bairoch Very strong Divergence and Convergence divergence then stability after 1890 Prados de la No data Convergence Divergence Escosura 22 Figure 4. Gini coefficients, 1800-1938 (calculated using Bairoch and Prados de la Escosura data) 20 20 Prados 15 .5 10 1800 1820 1830 1850 1860 1870 1880 1890 1900 1913 1929 1938 Figure 5. Theil coefficients, 1800-1938 (calculated using Bairoch and Prados de la Escosura data) 0 07 - - -- 0 06 00 IL111 0 04 0 03 0 02 S~~~~~~~~~airoch 0 02 0001 1800 1820 1830 1850 1880 1870 1880 1890 1900 1913 1929 1938 23 Growth rates Finally, we may ask whether the inter-war period was unique by having had very low growth rates, as it is sometimes believed. Table 10 gives the population-weighted average growth rate of GDP per capita (using Maddison's data) for WENAO countries. The end point of the first period is the peak before the crisis of 1890; the end point of the third period is the peak before the Great Depression. For the other two periods, the "natural" end-points are, of course, the two Wars. We see that the 1929-39 period displays a somewhat lower growth rate than the 40+ years of the first globalization. The differences however are not enormous. After a generation (twenty years) of growth at 0.61 percent per annum, a person's real income would be some 13 percent higher than in the beginning; if the growth rate were 0.45 percent per annum, his income would be 9 percent higher. Table 10. Average population-weighted growth rates of the WENAO region (per capita, per annum) 1870-1890 1890-1913 1919-1929 1929-1939 Growth rate 0.58 0.67 0.98 0.45 Note: growth rates calculated using a log-regression. These are average growth rates taling into account the entire population of WENAO. Source: Calculated from Maddison (2001). Table 11. Average unweighted GDP growth rates of the WENAO region (annualized five-year averages) 1870-1890 1890-1913 1919-1929 1929-1939 Growth rates 1.21 1.43 1.52 1.55 Observations 62 65 34 56 Average growth 1.4 1.5 (stand. (1.3) (2.9) deviation) I_I Source: calculated from Maddison ( 2001). The differences between the periods are even less if we look at countries' growth rates abstracting from the population size (see Table 11). Both the median and mean (annualized five-year) growth rates were almost the same in the pre- and post-1914 24 periods. It is only because of the Great Depression and due to the after-war rebound that the inter-war growth rates were more dispersed (see Figure 6). Figure 6. Distribution of countries' growth rates (annualized five-year averages) 40 - 30 - Pre-1914 30 - 20 - 10q_Q z< 0 -.1 0 .1 yy Source: Calculated from Maddison ( 2001). Finally, one may wonder to what an extent income convergence in the inter-war period was due to Italy and Germany, the two major Fascist powers whose growth was fairly fast during a part of the inter-war period. Both Italy and Germany were less rich in 1919 than the median WENAO country. 25 Italy's growth rate between 1922 and 1940 was only slightly higher than the unweighted WENAO average: 1.1 percent per person per annum vs. the mean of 0.93. Germany's rate between 1933 and 1940 was indeed much higher than the mean WENAO rate: 7.2 percent vs. 2.7.26 But if we drop Germany from the sample, there is practically no change in the Gini or Theil index. 25 Italy's rank was 1 Ih, Germany's 12' out of 17 WENAO countries. 26Part of it was certainly due to the catch-up effect following upon an extremely high decline in GDP during the Great Depression. That catch-up element would have been here with or without the Nazis. 25 4. The implications We can now see that the statement in Lindert and Williamson (2001) that the disintegration of world economy generated income divergence is quite misleading. After comparing inequality in incomes on the eve of World War I and at the end of World War 11-which in itself is a rather dubious way of making a comparison akin to comparing the peak to a trough of a business cycle-and finding that inequality in 1945 was much greater, they "assigned" the increasing income gap miscellaneously to "the great depression, two world wars, anti-global policies and other forces". Now we can readily see that all of the increased gap was due to the effects of the World War II, and none to "anti-global policies." Despite "anti-global" policies, income gap continued to shrink between 1919 and 1939. If both (i) greater integration of world economy, and more specifically closer links between the advanced capitalist economies (both before World War I and during the last 50 years), and (ii) disintegration of world economy, produce about the same effects on relative income gaps between the countries, then the trade-induced theory of convergence cannot be right. Our empirical findings would seem to suggest that convergence is a phenomenon independent of economic integration. In other words, greater trade, migration, or capital flows, have no discemable effect on the catch-up of the poorer countries. In effect, poorer countries catch up within the subset of rich (WENAO) countries, all the same whether there is economic integration or not. This could then, in turn, suggest several possibilities. The first possibility is that endogenous or neoclassical models that display increasing returns to scale may explain what we have observed during the inter-war years. If there are increasing returns to capital or labor, then for the poorer countries to cut off the links with the rest of the world, is a way to catch up. But this is doubtful as after World War II, the period of rising integration was also accompanied by income convergence. 26 A different explanation is as follows. Consider Figure 7 which shows the Gini coefficient of per capita incomes of WENAO countries for all years between 1870 and 1998. Over this long period of almost 130 years, income differences among the set of rich countries are either constant or decreasing during all peacetime periods. Only during wars do their incomes diverge. And it is only by the mid-1970's (there is a slight difference depending on whether we measure it using the Gini or Theil) that the level of similarity between their incomes had reached the values achieved before the Second World War or even before the First. The underlying policies-integration or disintegration, openness or autarky-changed during this long period of 130 years but did not seem to have had much of an effect on convergence of countries' incomes. If whether countries trade more or less, or invest more or less into each other's economies, does not seem to matter for convergence of their incomes-or in other words, does not affect the growth rate of poor economies vis-a-vis rich economies-then what other factors might explain such an outcome? We propose the following hypothesis. 27 Figure 7. Inequality among WENAO countries, 1820-1998 Gini and Theil index 0 30 0230 -& -- - - -- - - -- - - -- - - -- - - -- - - - -- - -- - - - -- - - - -- - -- - - - - - ------------- IX- - - - - Gini 0 15 ..X 0 205 <- -- -- I--- - C -. ... . . . .....................'- Neee I'll 111db ,#6 ,* lNee 1*'Op lKele lyt,e lKe lse lse l, ,ee 14; F 0 AR ,PI ,,4 ,se ,#+ l,# le 'IO "A", FE'4R l ee 1, ,g 10Be Note: After 1950, Israel and Turkey are added to the WENAO group. Since we deal here with a subgroup of rich Western economies that are well integrated, in a cultural sense, so that technological transfers (via books, private exchange of information, personal and business travel etc.) do take place almost as much whether there is a lot of trade and direct foreign investment or not, then convergence may be simply a reflection of that deeper integration. Transmission of information is what may drive modernization of the techniques of production, total factor productivity growth and ultimately income convergence (as implied by endogenous growth literature; see Jones, 1997, p. 25 or Easterly and Levine, 2001, p. 185). It is not irrelevant to this line of thought that even during the era of the 1930's, Italy's industrialization, for example, was decisively influenced by the American example. Giovanni Agnelli, FIAT's owner, after a visit to Ford, copied Ford's techniques of production. Olivetti, the office-equipment maker, and Pirelli, the tire company, were set on the path of becoming large multinationals in these years-again applying American techniques of mass production (i.e. what was later termed "Fordism"). 28 In conclusion, for economies similar in terms of their incomes, structure and cultural proximity, trade and direct investments may not matter as much (or at all). Within their "club", the poorer economies' growth rate relative to growth rate of the rich, may not be affected by greater or lesser integration. Whether for dissimilar countries, where the links between the economies and populations are few, the same is true, or whether in that case, economic integration needs to be "embodied" in goods and capital in order for the catch-up to take place, remains an open question. 29 REFERENCES Bairoch, Paul (1989), "The paradoxes of economic history: Economic laws and history", European Economic Review, vol. 33, pp. 225-249. Bairoch, Paul (1993), Economics and World History: Myths and Paradoxes, Chicago: University of Chicago Press. Bairoch, Paul (1997), Victoires et deboires, Histoire economique et sociale du monde du XVP' siecle a nosjours, Paris:Folio Histoire Gallimard. Baldwin, Richard and Philippe Martin (1999), "Two Waves of Globalisation: Superficial Similarities, Fundamental Differences", National Bureau of Economic Research Working Paper 6904, January 1999. Barro, Robert and Xavier Sala-i-Martin (1992), "Convergence", Journal of Political Economy, vol. 100 (2), pp. 223-25 1. Barro, Robert and Xavier Sala-i-Martin (1995), Economic Growth, New York: McGraw Hill. Baumol, William (1986), "Productivity Growth, Convergence, and Welfare: What the Long-run Data Show", American Economic Review, vol. 76, December, pp. 1072-1116. Baumol, William and Edward Wolff (1988), "Productivity Growth, Convergence, and Welfare: Reply", American Economic Review, vol. 78, December, pp. 1155-1159. Ben David, Dan (1993), "Equalizing Exchange: Trade Liberalization and Income Convergence", Quarterly Journal of Economics, vol. 108, No. 3. Bliss, Christopher (1999), "Galton's Fallacy and Economic Convergence", Oxford Economic Papers, vol. 51, pp. 4-14. Bordo, Michael D, Barry Eichengreen and Douglas A. Irwin (1999), "Is Globalization Today Really Different from Globalization a Hundred Years Ago?" National Bureau of Economic Research Working Paper No. 7195, June 1999. Cannon, Edmund S. and Nigel W. Duck (2000), "Galton's Fallacy and Economic Convergence", Oxford Economic Papers, vol. 52, pp. 415-419. Crafts, Nicholas (2000), "Globalization and Growth in the Twentieth Century", IMF Working Paper 2000/44, March 2000. 30 Easterly, William and Ross Levine (2001), "It's Not Factor Accumulation: Stylized Facts and Growth Models", The World Bank Economic Review, vol. 15, No. 2, 2001, pp. 177- 220. Ellison, David L. (2001), "CEEC Prospects for Convergence: A Theoretical and Historical Overview", typescript. Frank, Andre Gunder (1998), Re-Orient, University of California Press. Goerlich, Francisco J. and Matilde Mas (2001), "Inequality in Spain, 1973-91: Contribution to a Regional Database", Review of Income and Wealth, Series 47, No.3, September. Jones, Charles (1997), "On the Evolution of World Income Distribution", Journal of Economic Perspectives, vol. 11, No. 3 (Summer), pp. 19-36. Kindleberger, Charles P. (1986), The World in Depression 1929-33, Berkeley: University of California Press. Kindleberger, Charles P. (1989), "Commercial policy between the wars" in P. Mathias and S. Pollard, Economic Cambridge History of Europe, vol. VII, Cambridge: Cambridge University Press. Krugman, Paul (1991), Geography and Trade, London: MIT Press. Krugman, Paul (1995), "Growing world trade: causes and consequences", Brookings Papers on Economic Activity, No.1, pp. 327-362. League of Nations (1927), TariffLevel Indices, Geneva: League of Nations, International Economic Conference, May 1927. League of Nations (1939), World Economic Survey 1938/1939, Geneva: League of Nations, 1939. League of Nations (1936), World Economic Survey 1935/1936, Geneva: League of Nations, 1936. League of Nations (1945), Industrialization and Foreign Trade, Geneva: League of Nations, Economic, Financial and Transit Department. Lewis, Arthur (1949), Economic Survey, 1919-1939, London: George Allen and Unwin. Lindert, Peter and Jeffrey Williamson (2001), "Does globalization make the world more unequal", National Bureau of Economic Research, Working paper 8228, April 2001. Maddison, Angus (1995), Monitoring the World Economy, 1820-1992, Paris:OECD. 31 Maddison, Angus (2001), The World Economy: A Millennial Perspective, Paris:OECD Development Centre Studies. Mazower, Mark (2000), Dark Continent, Vintage books. Milanovic, Branko (2002), "The Two Faces of Globalization: Against Globalization as We Know It", downloadable from www.worldbank.org/research/inequality. Milanovic, Branko (2002a), "Worlds apart: the twentieth century's promise that failed", manuscript, downloaded from www.worldbank.research/inequality. O'Rourke, Kevin H. and Jeffrey G. Williamson (1999). Globalization and History: The Evolution of the 19h Century Atlantic Economy, Cambridge: MIT Press. Prados de la Escosura, Leandro (2000), "International Comparison of Real Product, 1820-1990: An Altemative Data Set", Explorations in Economic History, vol. 37, pp. 1- 41. Pritchettt, Lant (1997), "Divergence, Big Time", Journal of Economic Perspectives, vol. 11, No. 3 (Summer), pp. 3-17. Quah, Danny (1993), "Galton's Fallacy and Tests of the Convergence Hypothesis",: Scandinavian Journal of Economics, December 1993, vol. 94, no. 4, pp. 427-443. Available at http://econ.lse.ac.uk/staff/dguah/I/dp-93 galfall.pdf. Quah, Denny (1996), "Empirics for Economic Growth and Convergence", European Economic Review, vol. 40, pp. 1353-1375. Rodriguez, Francisco and Dani Rodrik (2000), "Trade Policy and Economic Growth: A Skeptic's Guide to the Cross-National Evidence", Version May 2000, downloadable from httD://ksghome.harvard.edu/-.drodrik.academic.ksg/skepti 1299.pdf. Summers, Rober, Irving Kravis and Alan Heston (1984), "Changes in world income distribution", Journal of Policy Modeling, May 1986, pp. 237-269. Williamson, Jeffrey (1996), "Globalization and Inequality: Then and Now: The Late 19,h and Late 20th Centuries Compared", National Bureau of Economic Research Working Paper No. 5491, Cambridge, Mass. Williamson, Jeffrey (1999), "Globalization, Labor Markets and Policy Backlash in the Past", Journal of Economic Perspectives, vol. 12, No. 4 (Autumn), pp. 51-72. Wodon, Quentin and Shlomo Yitzhaki (2002), "Growth and Convergence: An Alternative Empirical Framework", mimeo, March 2002. 32 World Bank (2002), Globalization, Growth and Poverty: Building an Inclusive World Economy, Policy Research Report, December. 33 Annex. WENAO Countries included in Maddison's, Bairoch's and Prados de la Escosura's datasets Year 1870 Year 1890 Year 1900 Year 1913 Maddison Bairoch Prados Maddison Bairoch Prados Maddison Bairoch Prados Maddison Balroch Prados AUS AUS AUS AUS AUS AUS AUS AUS AUS AUS AUS AUS AUT AHU AUT AUT AHU AUT AUT AHU AUT AUT AUT AUT BEL BEL BEL BEL BEL BEL BEL BEL BEL BEL BEL BEL CAN CAN CAN CAN CAN CAN CAN CAN CAN CAN CAN CAN CHE CHE CHE CHE CHE CHE CHE CHE CHE CHE DEU DEU DEU DEU DEU DEU DEU DEU DEU DEU DEU DEU DNK DNK DNK DNK DNK DNK DNK DNK DNK DNK DNK DNK ESP ESP ESP ESP ESP ESP ESP ESP ESP ESP ESP ESP FIN FIN FIN FIN FIN FIN FIN FIN FIN FIN FIN FIN FRA FRA FRA FRA FRA FRA FRA FRA FRA FRA FRA FRA GBR GBR GBR GBR GBR GBR GBR GBR GBR GBR GBR GBR GRC GRC GRC GRC GRC GRC GRC GRC GRC 1RL IRL IRL IRL ITA ITA ITA ITA ITA ITA ITA ITA ITA 1TA ITA ITA NLD NLD NLD NLD NLD NLD NLD NLD NLD NLD NLD NLD NOR NOR NOR NOR NOR NOR NOR NOR NOR NOR NOR NOR NZL NZL NZL NZL NZL NZL NZL NZL NZL NZL NZL NZL PRT PRT PRT PRT PRT PRT PRT PRT PRT PRT PRT PRT SWE SWE SWE SWE SWE SWE SWE SWE SWE SWE SWE SWE TUR USA USA USA USA USA USA USA USA USA USA USA USA Note: AUS=Australia, AUT=Austna, AHU=Austria-Hungary, CHE=Switzerland. 34 Year 1929 Year 1938 Maddison Bairoch Prados Maddison Bairoch Prados AUS AUS AUS AUS AUS AUS AUT AUT AUT AUT AUT AUT BEL BEL BEL BEL BEL BEL CAN CAN CAN CAN CAN CAN CHE CHE CHE CHE CHE CHE DEU DEU DEU DEU DEU DEU DNK DNK DNK DNK DNK DNK ESP ESP ESP ESP ESP ESP FIN FIN FIN FIN FIN FIN FRA FRA FRA FRA FRA FRA GBR GBR GBR GBR GBR GBR GRC GRC GRC GRC GRC GRC IRL IRL IRL IRL ITA ITA ITA ITA ITA ITA NLD NLD NLD NLD NLD NLD NOR NOR NOR NOR NOR NOR NZL NZL NZL NZL NZL NZL PRT PRT PRT PRT PRT PRT SWE SWE SWE SWE SWE SWE TUR TUR TUR TUR USA USA USA USA USA USA 35 Policy Research Working Paper Series Contact Title Author Date for paper WPS2923 Does Foreign Direct Investment Beata K Smarzynska October 2002 P FIewitt Increase the Productivity of Domestic 32724 Firms? In Search of Spillovers through Backward Linkages WPS2924 Financial Development, Property Stujn Claessens November 2002 R Vo Rights, and Growth Luc Laeven 33722 WPS2925 Crime and Local Inequality in Gabriel Demombynes November 2002 P Sader South Africa Berk Ozler 33902 WPS2926 Distinguishing between Rashmi Shankar November 2002 P. Holt Observationally Equivalent Theories 37707 of Crises WPS2927 Military Expenditure Threats, Aid, Paul Collier November 2002 A Kitson-Walters and Arms Races Anke Hoeffler 33712 WPS2928 Growth without Governance Daniel Kaufmann November 2002 K Morgan Aart Kraay 37798 WPS2929 Assessing the Impact of Carsten Fink November 2002 P Flewitt Communication Costs on Aaditya Mattoo 32724 International Trade Ileana Cristina Neagu WPS2930 Land Rental Markets as an Klaus Deininger November 2002 M Fernandez Alternative to Government Songqing Jin 33766 Reallocation? Equity and Efficiency Considerations in the Chinese Land Tenure System WPS2931 The Impact of Property Rights on Klaus Deininger November 2002 M Fernandez Households' Investment, Risk Songqing Jin 33766 Coping, and Policy Preferences Evidence from China WPS2932 China's Accession to the World Aaditya Mattoo December 2002 P Flewitt Trade Organization The Services 32724 Dimension WPS2933 Small- and Medium-Size Enterprise Leora F Klapper December 2002 A Yaptenco Financing in Eastern Europe Virginia Sarria-Allende 31823 Victor Sulla WPS2934 After the Big Bang? Obstacles to the Karla Hoff December 2002 A Bonfield Emergence of the Rule of Law in Joseph E Stiglitz 31248 Post-Communist Societies WPS2935 Missed Opportunities. Innovation and William F Maloney December 2002 P Soto Resource-Based Growth in Latin 37892 America Policy Research Working Paper Series Contact Title Author Date for paper WPS2936 Industrial Ownership and Hua Wang December 2002 Y. D'Souza Environmental Performance: Yanhong Jin 31449 Evidence from China WPS2937 The Determinants of Government Hua Wang December 2002 Y. D'Souza Environmental Performance. An Wenhua Di 31449 Empirical Analysis of Chinese Townships WPS2938 Recurrent Expenditure Requirements Ron Hood December 2002 M Galatis of Capital Projects: Estimation for David Husband 31177 Budget Purposes Fei Yu WPS2939 School Attendance and Child Labor Gladys L6pez-Acevedo December 2002 M Geller in Ecuador 85155 WPS2940 The Potential Demand for an HIV/ Hillegonda Maria Dutilh December 2002 H Sladovich AIDS Vaccine in Brazil Novaes 37698 Expedito J. A. Luna Moises Goldbaum Samuel Kilsztajn Anaclaudia Rossbach Jose de la Rocha Carvalheiro