WPS4681 Policy ReseaRch WoRking PaPeR 4681 Assessing Interactions among Education, Social Insurance, and Labor Market Policies in a General Equilibrium Framework: An Application to Morocco Mohamed A. Marouani David A. Robalino The World Bank Middle East and North Africa Region Human Development Department & Human Development Hub July 2008 Policy ReseaRch WoRking PaPeR 4681 Abstract This paper develops a general equilibrium model to projects the revenues and expenditures of the pension analyze the marginal and joint impacts that alternative system. The model is applied to the case of Morocco. macroeconomic, education, and social protection Simulations show that even under positive assumptions policies have on the dynamics of employment and regarding economic growth, unemployment rates are unemployment by skill level. The model introduces likely to remain close to current levels in the next decade. a disaggregated treatment of the labor market that The paper argues that only an integrated package of incorporates an informal sub-sector in every sector of the policies that affect the macro-economy, the investment economy. The analysis explicitly models the distribution climate, and the education and social protection systems of skills in the labor force by following over time sex- would allow sustainable creation of enough "good age cohorts across various levels of the education system quality" jobs. and in the labor market. And it integrates a module that This paper--a product of the Human Development Department of the Middle East and North Africa Region and the Human Development Hub--is part of a larger effort in the department to better integrate/coordinate macroeconomic, social protection and education policies within a strategy for employment creation. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The author may be contacted at drobalino@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective 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 necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Assessing Interactions among Education, Social Insurance, and Labor Market Policies in a General Equilibrium Framework: An Application to Morocco Mohamed A. Marouani Université Paris1-Panthéon-Sorbonne, DIAL and ERF David A. Robalino World Bank We thank Carlos Silva, Milan Vodopevic, Mohamed Benrida, Michal Rutkowski, Robert Holzmann, David Steel and the participants at the conference on "Skills Development, Social Protection, and Employment in Morocco" held in Rabat in April 2007, for comments on a preliminary version of the paper. Special thanks go to: Abdelkalek Touhami, who contributed to the design of the model used in the analysis and was responsible for preparing the Social Accounting Matrix; Andras Bodor, who conceived and programmed the Human Capital accumulation module; Montserrat Pallares who was responsible for calibrating and conducting simulations with the pensions module; and Anca Mataoanu who was instrumental in the compilation of some of the data used in the analysis. The views expressed in the paper are those of the authors and do not necessarily reflect those of the institutions they represent or those of the Government of Morocco. Introduction One of the main preoccupations of governments around the world is how to improve labor market outcomes, both generating sufficient jobs for all members of the labor force and increasing the productivity, and therefore the earnings, associated with these jobs. The problem is that there are multiple policies operating at various levels of the economy that affect labor market outcomes. These include macroeconomic and institutional policies that influence investments decisions, economic growth, and labor demand; education and training policies that affect the distribution of skills in the labor force; and, social protection policies that affect the bargaining power of workers, their mobility, as well as labor costs and the tax-wedge.1 Policymakers face several challenges when it comes to designing and implementing policies at these various levels because they need to accurately predict the joint effects that alternative policies have on labor market outcomes. It would also be interesting to transform/aggregate the changes in labor market outcomes into a measure of social well being. In terms of the link between policies and outcomes, two general approaches can be considered here. One is to conduct ex-post evaluations of a given policy intervention. These are powerful tools for guiding policymaking and are becoming more and more popular. They usually refer, however, to specific, narrow, and well targeted interventions. Their use is more limited when a comprehensive strategy ­ such as one to promote employment creation ­ is going to be conducted. In the latter case, general equilibrium and macroeconomic effects are important and there is less room for "piloting" before implementation. Ex-ante evaluations in this case remain important. These involve the existence of an economic model linking policy levers to outcomes of interest. Unfortunately, the economics profession is still lacking an integrated framework that can be used to analyze labor market policies (see Fields, 2007). There are still several questions to address in terms of what type of models should be used to link policies to outcomes; to complicate matters these tend to depend on the country context. 1See for instance World Bank (2007) for a discussion on framework that integrates these policies. 2 Coming up with a welfare function has also remained elusive. Some of the possibilities have been proposed by Lambert (1989), Squire (1981), Somavia (1999) and are reviewed in Fields (2001, 2005 and 2007). When doing applied policy analysis, however, the form of the welfare function is likely to be very difficult to determine -- since it reflects social and cultural preferences. This paper is concerned with the first problem: linking, ex-ante, policies and outcomes. The specific objectives are twofold. At the methodological level, the paper develops a general equilibrium model for the evaluation of integrated policy packages to promote employment creation. The model has been named SSELMA which stands for Social Security, Education, Labor, and the Macro- economy.2 The model presents innovations in the disaggregation of the labor market, the modeling of the supply of different skills, and the representation of the social insurance system. From a policy perspective, this paper uses the Moroccan case to address two questions. First, how alternative education, SP, and macroeconomic policies affect key labor market outcomes including unemployment rates for workers with different levels of education. Second, how the effects of integrated policy packages compare to the effects of partial policy interventions. In terms of education policies, the focus is on changes in internal efficiency: repetitions and dropout rates at various level of the education system. Regarding social protection policies, the focus is on reforms to the pension system that affect the tax wedge. Finally, macroeconomic policies focus on incentives to promote investment in high value added sectors, which is a recent policy in the country (see below). The paper starts by discussing the unemployment problem in the Moroccan case, the need for an integrated approach, and thus the rationale for a model such as SSELMA. It then introduces the main features of the model and describes, in general terms, the methodology used for calibration purposes. The last two sections present the results of policy simulations and offer general conclusions regarding the main results of the analysis. I. Unemployment in Morocco and the Need for an Integrated Policy Framework Morocco is a country that has introduced several pro-growth structural reforms since the mid- 1990s. The country has achieved macroeconomic stability with low inflation, a strong external position and an ongoing fiscal consolidation. Following a decade of per capita GDP growth close to 2Selma means secure in Arabic and is also the name of the Queen of Morocco but this is only a coincidence. 3 zero, between 2001 and 2006, GDP per capita grew at an average of 2.5 percent per year. In 2006, GDP grew at a rate close to 7 percent. Faster growth allowed the national unemployment rate to decline from 13 percent in 2000 to 9.7 percent in 2006. In urban areas the unemployment rate was reduced from 21.2 percent to 15.5 percent. In general, the employment-creation intensity of the economy has been falling over the last few years. While between 1984 and 1990 a 1 percent increase in urban GDP was associated with a 1.44 percent increase in employment, during the last years the same growth rate increased employment by only 0.35 percent. On one hand this is welcome as it reflects gains in labor productivity. On the other hand, it implies that faster growth will be necessary to absorb new entrants to the labor market. One challenge facing the government is that pressures on the labor market remain strong and economic growth might not remain at current levels. Today there are approximately 11.1 million individuals working or looking for a job in Morocco. Although the country is ending the second phase of the demographic transition, the labor force is expected to grow at an annual 1.8 percent over the next 10 years. This implies that there will be around 260,000 net new entrants to the labor force each year between now and 2020. If the current unemployed are taken into account, Morocco would need to create around 3.85 million jobs by 2020 to eliminate unemployment. That is an average of 295,00 jobs per year; whereas in the years before 2006 the number of jobs created each year did not surpass 190,000 . If as a result of better access to education and lower fertility rates the change in participation rates accelerates (say converging to OECD levels in three decades), labor force growth could approach 2.5 percent per year. Yet, recent projections suggest that the GDP growth rate is actually going to slow down to 3.5 percent in 2007, a rate closer to the more modest historical average. Medium-term projections building upon very favorable conditions suggest that Morocco would barely reach an average growth rate of 5.3 percent per year during the period 2007-2011 (World Bank, 2006). The other problem is that the majority of the new jobs created are low-productivity and low- skilled, many in the informal sector. For instance, 90 percent of the jobs created since the year 2000 went to individuals without higher education diplomas. Also, the large majority (67 percent) of new 4 jobs created since 2002 were in commerce and construction/public works sectors, which often tend to be informal jobs. The informal sector supplies 39 percent of non-agricultural jobs and 90 percent of jobs in the commerce sector. Another 10 percent of the jobs created where in agriculture. The agricultural sector still supplies 44 percent of the jobs and around half of those jobs are really unpaid on-farm family support. The other issue is the very high unemployment rates among young educated individuals, which co-exist with low levels of education in the labor force and skills shortages in strategic sectors. In 2005, 59.8 percent and 39.8 percent of those with post secondary education diplomas in the 15-24 and the 25-34 age categories were unemployed. This group will constitute more than one third of the new entrants to the labor market over the coming years. At the same time, 35 percent of the labor force does not have formal education and 75 percent have less than secondary education. Only 7.5 percent of individuals in the labor force have higher education. The problem is even more severe in the case of rural areas where 55 percent of the labor force does not have formal education and close to 90 percent have less than secondary education. To deal with the unemployment problem, the government has recently launched many initiatives including the Program for the Promotion of Strategic Sectors3. The program is part of a strategy to diversify the economy and promote competitiveness through investments in high value added and export oriented sectors. Around half a million jobs are expected to be created through this program by year 2020. In this paper we argue that this program alone is not the solution to the unemployment problem. The success of employment strategies will depend on the ability of the government to coordinate, implement and evaluate policies across sectors and through various agencies. The reminder of this paper develops and uses an applied general equilibrium model to illustrate how a sub- set of these policies would interact and affect employment dynamics and the future evolution of unemployment rates. The focus is on policies that improve the internal efficiency of the education, promote productivity growth, control the tax wedge and contain the accumulation of unfunded 3Programme d'Emergence. 5 pension liabilities. We also look at the interaction of these policies with the government Emergence Program. II. An Introduction to SSELMA A first attempt to develop an integrated framework dealing with labor market issues in Morocco was provided by Agenor and El Aynaoui (2003). Based on a modified version of IMMPA4, the authors introduce Harris-Todaro type mechanism to determine the supply of unskilled labor in the formal sector and, a monopoly union approach to determine skilled workers' wage in the private formal sector. The model takes into account internal and international migration. However IMMPA is quite stylized in terms of sectors and factors modeling. It is thus not suited for simulating the impact of a policy on a particular sector (Financial Sector for example) and does not allow to analyze the impact on a particular category of labor (technicians or university graduates for example). Moreover, it does not deal with social security reform (transfers to households are exogenous), does not take into account training and does not capture the demographic determinants of the labor market. Given our objectives, this tool was not fitted for analyzing the interactions between education, training, social security and employment policies. We thus had to develop a new framework for that purpose. SSELMA is a multi-sector, sequential dynamic general equilibrium model. The main innovative features of SSELMA are related to the modeling of the labor market, the supply of skills, and the representation of the social insurance system. Regarding the labor market, the model integrates an informal sub-sector in each sector of the economy (except agriculture). The model also takes into account labor market segmentation along the following dimensions: urban/rural, formal/informal, public/private, and permanent/temporary employment. In terms of the types of skills and their supply, the model includes a module that follows individuals over time across various levels of the education system and through several diplomas/specializations in post secondary education.5 In this version of the model the various skills supplied by the education system are aggregated into six skills categories (see Table 1). The model is also connected to a template that projects expenditures in the pension system by following over time the age/sex cohorts of contributors and retirees, their 4See Agenor, Izquierdo and Jensen (2007) for a detailed presentation of the IMMPA framework. 5For a description of the HC accumulation model the reader is referred to Bodor, Robalino, and Rutkowski (2007). 6 wages, and their pensions (the letter calculated according to specific policy rules or through replacement rates).6 The model also opens the possibility to introduce unemployment insurance but the option is not used in the Moroccan case. In this section we focus mainly on the innovations of the model while describing very briefly the standard applied general equilibrium specifications. Table 1: Typology of Skills in the Model SELMA Typology of Labor Categories Diplomas and Specializations Included No formal education or only Initial Training (i.e., Unskilled workers (un) VT diploma obtained where the only condition is to read and write). Low skilled workers type A (lsa) Primary education + VT (specialilsation) or 9th + VT.(qualification) or Secondary education only Low skilled workers type B (lsb) Only primary education or 9th grade Skilled workers (sk) Secondary education + VT (Techniciens specialisés) Highly skilled workers type A (hsa) Graduates from Grandes Ecoles and faculties with technical specializations. Highly skilled workers type B (hsb) University graduates (except physicians) without a VT diploma Source: Authors' design. The number of workers in each skill category by economic sector was provided by the High Council for Planning on the basis of the labor force survey. The production and factors demand block The economy is disaggregated into 20 sectors. Within each sector, production factors are subdivided in 7 items: capital plus the six labor categories described above. The production function is a nested one. At the highest level we assume that the production of each sector is a Leontief function of value added and total intermediate consumption. The demand for capital and the 6 skills levels is modeled through a nested CES function at 5 levels, which allows to have differentiated elasticities of substitution between the different factors. Capital and highly skilled labor have been modeled relatively complementary, following the Fallon-Layard hypothesis (1975) which has been confirmed by various empirical studies. Concerning the other bundles we have aggregated factors by pairs in CES bundles so as to have higher substitution elasticities between similar factors (for example 6The basis of the template is the World Bank model PROST (see World Bank. 2005). 7 the elasticity of substitution between university graduates and engineering schools graduates is higher than between these two categories and other labor categories). Value added is a Constant Elasticity of Substitution (CES) function of a capital-highly skilled labor bundle (KHS) and a low skilled labor-unskilled labor bundle (ULS). Through this function the relative demand of the bundles depends on the evolution of their relative prices (PKHS and PULS) given the elasticity of substitution (1). The value added and related demands are thus characterized by: 1 (1-1) (1-1)(1-1 ) VA = A11KHS 1 + (1-1)ULS 1 (1) KHS = ULS1- 1 PULS 1 1 PKHS (2) At the second stage, KHS and ULS are two CES functions of respectively capital (K) and a skilled-highly skilled labor demand bundle (HSK), and unskilled (un) and a low skilled labor demand bundle (LS). The first order conditions give us the derived demand of respectively K, HSK, un, and LS. We have: 2 (2-1) (2-1)(2-1 ) KHS = A2 2K 2 +(1-2)HSK 2 (3) 3 (3-1)(3-1 ) ULS = A33LDun (3-1) 3 + (1-3)LS 3 (4) K = HSK1- 2 2 PHSK 2 (5) r LDun = LS1- 3 3 awcun PLS 3 (6) At the third stage, HSK and LS are respectively two CES functions of a composite highly skilled labor bundle (HS) and skilled labor (sk) and two low skilled categories lsa and lsb. The 8 demand for these bundles and worker categories is derived through the same optimization process as above. 4 (4-1) (4-1) HSK = A4 4HS 4 + (1-4)LDsk (4-1) 4 (7) 5 LS = A55LDlsa (5-1) (5-1)(5-1) 5 + (1-5)LDlsb 5 (8) HS = LDsk 4 awcsk 4 1- 4 PHS (9) LDlsa = LDlsb 1- 5 awclsb 5 5 awclsa (10) At the fourth stage, HS is a CES function of two highly skilled workers categories, namely hsa and hsb. We have: 6 HS = A6 6LDhsb (6-1) (6-1)(6-1 ) 6 + (1-6)LDhsa 6 (11) LDhsb = LDhsa 1- 6 6 awchsb awchsa 6 (12) Finally, in the last stage, labor demand is disaggregated into two categories: permanent workers covered by the social security system and temporary workers excluded from it. For the former labor demand by skill is a CES function of permanent and temporary labor demand (PLD and TLD). lf LDlf = Alf lf PLDlf (lf -1) (lf -1)(lf-1) lf + (1-lf )TLDlf lf (13) PLDlf = TLDlf lf 1 lf 1 -lf (1+scelf +uilf ) (14) Wages of permanent workers include employers' social security contributions (sce) and unemployment insurance (ui). We distinguish legal (lsce) and effective rates of contribution to social 9 security (the difference is due to exemptions, evasion, etc.). The model also takes into account the cost of the implicit cost of severance pay, using as proxy an implicit unemployment insurance rate. We have: awclf = wlf ((1+ scelf + uilf )* PLDlf + TLDlf ) (15) LDlf scelf = lf *lscelf (16) The labor supply block As mentioned above, one innovation in SELMA is the detailed projection of various skills levels based on repetition and drop-out rates at various levels of the education system, including 85 combinations of diplomas and specialisations in higher education and vocational training. There are 6 skills levels associated with primary education depending on the grade an individual completed. Similarly 3 skill levels are associated with low secondary education and another 3 skill levels with upper secondary education. There are 9 skill levels associated with vocational training depending on the level of vocational training7 (and the length of the vocational training (1 or 2 years). In higher education the various areas of specialization,8 combine with the length of study (varying between 2 to 7 years depending on the specialization) and whether the education is provided by public or private institutions of higher education, thus generating 64 distinct skill levels within the model. At each time t, these skills are aggregated into 6 types of labor. Here we describe how we allocate the total supply of labour across economic sectors and formal/informal markets. In essence, there are three stages. In the first stage we model the migration process between the rural and urban labor markets. At the second stage, the urban labor market is divided into a formal and an informal segment. The informal labor supply is also then subdivided into independent workers and wage earners. At the first level, following Cogneau et al. (1996), the migration process from rural to urban areas is modelled through an extended Harris-Todaro function. Thus, the relative supply of labor in 7Certificat d'apprentissage, Spécialisation, Qualification, Technicien and Technicien Spécialisé 810 specializations altogether: Médecine Dentaire; Médecine et Pharmacie; Enseignement Originel; Sciences Juridiques, Economiques et Sociales; Lettres et Sciences Humaines; Sciences; Sciences et Techniques; Technologie; Sciences de l'Ingénieur; Commerce et Gestion 10 the urban (LSU) and rural (LSR) sectors is a function of the expected wage in the urban sector (auw) and the observed wage in the rural sector (nwr). This function incorporates an elasticity of mobility (1) of labor from rural to urban areas. The expected urban average wage is a weighted average of the wages in the urban sector (informal, formal private and public) multiplied by the probability of finding a job in the urban sector (1-u). We have: LSU E(auw)1 (17) LSR = c1 nwr (1- u) i nwii(TLDi + PLDi) + nwf i E(auw) = (TLDi + PLDi) + i= pformu (18) LSU nwg =inf o (TLDgov + PLDgov) where TLDi and PLDi are the total labor demand of temporary workers and permanent workers in sector i, as defined above; nwii are wages in the informal sector, nwfi wages in the formal sector, and nwg the wage in the government sector. Regarding the split between formal and informal sector, we assume that the supply of labor in the latter results from a cost-benefit analysis made by workers and is not simply a residual. This view is consistent with the recent empirical evidence that shows that informality is both the result of exclusion (the traditional view) and strategic decisions that respond to perceive benefits (e.g., flexibility, no taxation) and perceived costs (lower wages, lack of access to formal institutions). In this framework, an increase in labor productivity in the formal sector or stronger public institutions would reduce informality by increasing the opportunity cost of not participation in the formal sector of the economy (see World Bank, 2007). Thus, we model labor mobility between the formal and informal sector in the same way as we model mobility between rural and urban sectors: the ratio of labor supply in the formal and informal sectors is a function of the expected wage in the formal sector and the observed wage in the informal sector. The mobility elasticity (2 ), however is assumed to be lower than 1 given that the cost of migration from the rural areas to the urban areas are likely higher than those related to the movement between informal and formal sectors. We have: 11 LSF E(afw)2 (19) LSI = c2 aiw nwf i(TLDi + PLDi) + nwg(TLDgov + PLDgov) E(afw) = (1-u) i=pformu (20) LSF In terms of wages we consider a disequilibrium framework: wages do not clear the urban labor market. Following Cogneau et al. (1996), they are modelled as an extended wage curve9. Thus, we assume that the average private wage by is a function of the unemployment rate, the public wage and the minimum wage10. The following equation is thus estimated econometrically: ln(afw) = 1 + 2 ln(gw) + 3 ln(minw) + 4u (21) Sectoral wages by skill are then assumed to be equal to the average wages by skill defined above multiplied by an exogenous wage differential by sector and skill. wfi = afw* fwdisti i PFORMU (22) wii = aiw*iwdisti i INFO (23) The formal labor force is split between independent workers and wage earners. We assume that the initial proportions of these two categories in the informal labor force are fixed. The social security block A social security account is modeled separated from the government budget. The account receives employers and employees social security contributions (their total amount being endogenous) and pays benefits to households. The benefits are split into three components: old-age pensions, disability and survivorship pensions, and "other benefits". 9See Blanchflower and Oswald (1994) for a comprehensive presentation on wage curves. 10Minimum wages are set as a binding constraint. 12 To determine the level of these benefits we run the PROST (Pensions Reform Options Simulation Tool) model11 which feeds SSELMA with the evolution of dependency ratios and average replacement rates (the ratio between the average pension and the average wage in the economy). SSELMA uses these two inputs to compute total pensions and other social security expenses, which are proportional to the total wage bill. The income and expenditures block Households earn their income from wages, returns to capital, interests on bonds and transfers (mainly government and social security transfers and migrants' remittances). Their expenditures are composed of consumption of goods and services, social security contributions, interest payments and transfers. The composition of their basket of goods and services is determined through the maximization of a Linear Expenditure System (LES) function under their budget constraint. The government earns income from various taxes (income taxes, corporate taxes, tariffs, production and consumption taxes) and from foreign transfers. Its expenditures consist of government consumption (mainly civil servants wages), social transfers and interest payments on public debt. The foreign trade block The allocation of output between domestic and foreign markets is modeled as a Constant Elasticity of Transformation (CET) function. On the demand side, the Armington assumption is adopted to describe imperfect substitution between domestic products and imports. The small country assumption holds for imports, which implies that world import prices are exogenous, however, an export demand function is modeled, which means that Moroccan exporters can reduce their prices if they want to increase their market share on international markets. The closures of the model SSELMA has 5 closures: a macro closure, a government closure, an external balance closure, a closure of the social security system and a labor market closure. Concerning the macro closure, this version of SELMA is savings driven (households' marginal propensity to save is exogenous). The government closure chosen consists in fixing government 11Developed by the World Bank Human Development Network.. 13 expenditures (their growth rates) and tax rates and leaving the government budget balance endogenous to allow an assessment of the fiscal sustainability of the various scenarios. The foreign balance closure consists in fixing the exchange rate and leaving the current account balance endogenous to allow for the possibility of simulating the impact of exchange rate variation on the relevant variables. The social security balance is also endogenous. Finally the labor market closure of the SELMA model consists in a joint determination of unemployment and average formal wage through the wage curve described above. The dynamics of the model The model dynamics are of the sequential type. Capital accumulation is sectoral. Each year the stock of capital of each sector corresponds to the sum of the stock of last year and new investment minus the depreciation of capital. Following Bchir et al. (2002), sectoral investment (INVi) has been modeled as a function of the sectoral stocks of capital (KDi), sectoral rates of return to capital (rki) and capital acquisition costs (PK) net of subsidies (subv): lambdai*rki INVi = gamma* KDi *e PKi -subvi (22) gamma is an endogenous adjustment variable. It permits to ensure that total investment is equal to total savings. Sectoral investment increases with sectoral rates of return to capital, government subsidies and decreases with sectoral acquisition costs of capital. Yearly labor supply growth by skill is determined using the Human Capital projection model12. The HC model projects labor force by age, gender, and skill type based on mortality and fertility rates, labor force participation rates, and enrollment, repetition, and dropout rates at various levels of the education system. Government wages increase at a fixed rate set by the government. The other exogenous variables (mainly transfers) are supposed to vary each year at the same rate as GDP. III. Calibration of SSELMA in the Moroccan Context Due to the large number of parameters, calibrating a general equilibrium model remains an ¨art.¨ Part of the calibration involves setting the initial value of macro-economic aggregates (e.g., 12For a description of the model see Bodor, Robalino, and Rutkowski (2007). 14 relative wages and the distribution of the labor force by sector) and this is usually straightforward when national accounts and labor force surveys are available. The other part, however, refers to unobservable parameters such as the elasticities of substitution in the production functions. One would like, for instance, to estimate the distributions of these parameters to replicate the statistical properties of known time series predicted by the model. Such exercises have actually been conducted in the case of small ¨toy models¨ (see Abdelkhalek and Dufour, 2006), but are difficult to implement in the case of large scale models such as SSELMA. The model, however, is not intended to be used for prediction purposes. The more modest objective is to help the analyst understand the direction and order of magnitude of changes in output variables resulting from policy interventions, and how these policies interact, based on a set of reasonable assumptions regarding the value of the various parameters. Initial conditions The model was calibrated from a 2003 database for the Moroccan economy. The social accounting matrix (SAM) has been built on the basis of the national accounts provided by the Moroccan "Direction de la Comptabilité Nationale" and of the data from the labor force survey and the informal sector survey provided by the "Direction de la Statistique." The SAM distinguishes 32 activities (13 of which are informal) and 19 commodities. It is composed of 5 agents (Households, Firms, Government, Social Security and Rest of the World) and 7 factors (described above). The investment matrix has also been provided by the "Direction de la Statistique". In terms of wages, the 1998 household survey was used (unfortunately, the most recent 2002 household survey did not include information on earnings). First, relative wages by skill level were estimated. To this end, the various skills available in the survey were collapsed into the six skills used in the model and outliers drop from the survey. The relative wages where then used to calibrate initial average wages by sector, in order to replicate the total wage bill of the sector given the distribution of skills provided by the labor force survey. 15 The parameters The elasticities of substitution and transformation of the CES and CET functions have been fixed a priori at "reasonable levels."13 For the production function elasticities we have for example a much lower elasticity between capital and highly skilled labor than between the other factors14. The derivation of the scale and share parameters of the functions follows the usual procedure. Based on the initial values of the variables and exogenous parameters the scales/shares are computed endogenously. The parameters of the Human Capital accumulation model were estimated on the basis of data on student flows (enrollment, dropout, repetition and advancement rates) across different levels of the education system provided by the Ministry of Education. The flows are subject to change as the demographic composition of the population changes even if the transition variables of the education system remain constant. Finally, the parameters of the wage curve were estimated using time-series on wages provided by the "Caisse Nationale de Sécurité Sociale" (CNSS). IV. Policy Simulations This section starts by presenting the dynamics of key macroeconomic aggregates under the status-quo or baseline scenario. It then discusses the impact of alternative policy interventions in the education and social protection system, implemented individually or as a package (see Table 2). The focus is on the following aggregates: (i) the growth rate of GDP; (ii) the level of investments; (iii) the fiscal deficit; (iv) the aggregate unemployment rate in urban areas ; (v) the size of the informal sector in terms of its contribution to total urban employment; and (vi) unemployment rates by skills level. It is important to mention that the focus of the analysis is more on the direction of changes observed among the various output indicators than about the magnitude of the changes. The tables with the results of the various simulations are presented in the Appendix. 13See table A1 in the appendix 14Fallon and Layard (1975) hypothesis. 16 Baseline dynamics In the baseline scenario the model is calibrated to project a real GDP growth rate that fluctuates between 5 and 6 percent per year ­ which basically replicates the official short-term projections. Hence, GDP grows at 4.5 percent in 2007, increasing to 6 percent in 2010 and then fluctuating between 5 and 6 percent until year 2015 ­ the end of the simulation period. The projected fiscal deficit, investment levels and the current account balance are also consistent with official projections (see Table A2). The results show that despite a positive outlook in terms of economic growth unemployment rates in urban areas would go down only very gradually. Hence, the model projects that the urban unemployment rate would first increase and remain constant at around 13 percent.15 Only in year 2010 it would start to gradually decline converging to 9.3 percent in year 2015. These results are consistent with the aggregate projections that one would obtain assuming that the aggregate employment-output elasticity remains unchanged (see World Bank 2007, Chapter 2). The dynamics of unemployment rates are very different between skills categories; prospects are especially worrisome for Specialized Technicians. For unskilled workers, unemployment rates could rapidly converge to "natural levels." In part this occurs as their share among new entrants in the labor force diminishes. For low-skilled workers unemployment rates are also on a downward trend although it would take them longer (around year 2015) to reach natural levels. For highly skilled workers initial unemployment rates are higher and could increase over the short term, but would start declining after year 2010. Still, by year 2015, 26 percent of university graduates in non-technical fields and 9.3 percent of graduates in technical fields would remain unemployed. The situation is more critical, however, for Specialized Technicians. They start with the highest unemployment rates (41.4 percent) but because of a rapid increase in their number, unemployment rates could continue to climb stabilizing at around 47.2 percent in year 2015. This result is driven of course by the 15For consistency reasons, labor data in the model are in full-time equivalents, which means that the baseline unemployment rate is higher than the official one. Indeed, employment and unemployment figures are presented in number of people mixing full time and partial time workers. In models we always transform the data in full-time equivalents because we cannot add someone who works one month and someone who works 9 months. This means that our unemployment figures are also in full-time equivalents and thus higher than official ones. 17 assumptions of the downward rigidity of wages that are not able to fall to stimulate labor demand sufficiently. The main message from the baseline scenario is that additional policy interventions would be necessary to reduce unemployment rates, particularly among high skilled workers. First there is a need to control student flows. The current education/training system does not respond to market signals and therefore can oversupply certain skills. The simulations indicate, for instance, that the current flows of Specialized Technician are simply too high for current demands. But demand side interventions are also necessary to create more jobs at the aggregate level. The next sections discuss the impact of some of these policies. Dynamics under alternative policy interventions The four sets of policies considered in the analysis are presented in Table 2. The first set consists of policies that improve the technical efficiency of the education system. In essence, reductions in drop-out and repetitions rates at various levels of the education system. These affect the supply of various skills of labor through the Human Capital projection module. Second, policies related to the social security system. One is the absence of reforms in the pensions system that implies a continuous increase in the tax-wedge to cover expenditures. The other, is a reform program that gradually reduces replacement rates, which permits to reduce pension expenditures and thus to avoid an increase in workers' contributions. The third set of policies is assumed to affect total factor productivity growth. These policies could include improvements in the design of the continuous training system, better quality and relevance in higher education and vocational training, and/or those that support the adoption and diffusion of new technologies. Of course, we are not able to model the direct impact of this type of policies, but explore what would happen if, through their implementation, TFP increases. The fourth set of policies is related to the implementation of the current government programs to promote the development of high-value added economic sectors. To this end we simulate the effects of higher investments in agro-industry, the financial sector, the mechanical industry and electronics. 18 The simulations look at the marginal impacts of each of the policy packages, as well as their joint effects. Table 2: Description of Policy Simulations Policy Simulation Mechanism A. Improvement in internal efficiency Repetition, dropouts, and enrollment rates improve by 50 of education system. percent at all levels of the education system. B Policy A with higher total factor the growth rate of TFP increases by 20 percent (e.g., from 1 productivity growth. percent per year to 1.2 percent per year). C. Failure to reform social security A 10 percent increase of the average effective contribution rate per year is assumed (e.g., from 10 percent to 11 percent) D. A Ssustainable Social security It assumes that the pension system is reformed and that reform16 average replacement rates are reduced over time The contribution rate to the social security remains constant. E. Support to emergent sectors (Agroindustry) 10 percent subsidy to the acquisition costs of capital. F. Support to emergent sectors (Financial sector) Same as before G. Support to emergent sectors (Mechanical Industry and Same as before Electronics) H. Support to emergent sectors (D+E+F) The subsidy is applied to the four sectors simultaneously I. Integrated package (B+D+H) All "shocks" are introduced simultaneously Source: Authors' design. Education and training policies. An important insight from the simulations is that simply improving internal efficiency at various levels of the education system might not contribute much to growth or employment creation. In theory, better internal efficiency would imply a higher stock of human capital, lower labor costs for skilled workers and faster GDP growth over the long term. However, the simulation of a policy that increases internal efficiency across education levels shows that the impacts on GDP growth would be negligible over the short and medium term (between 0.1 and 0.3 percentage points). The effects on the aggregate unemployment rate would also be small at least until year 2011 when it could be reduced by 0.5 percentage points (see Table A3). On the contrary, the simulations show that unemployment rates could increase for some individuals, depending on their skills. The relative effects are complex and depend on demographics 16The reform is composed of the following elements: review benefit formulas and eligibility conditions in order to ensure financial sustainability; and eliminate incentives for evasion and early retirement; make redistribution transparent and progressive; and ensure the full portability of benefits across pension funds and, to the extent possible, identical provisions. 19 (which affect the supply) and on the structure of the production function (which affect the demand). The model shows that internal efficiency alone could hurt unskilled and highly skilled workers, and be beneficial for those in the middle. Unemployment rates could increase among unskilled workers at the benefit of low skilled workers without VT diplomas. Indeed, higher internal efficiency at all levels of the education sector reduces the supply of unskilled workers and therefore drives up their wages. Because they become more expensive relative to low skilled-workers, particularly those without VT diplomas, a substitution effect takes place. Hence, unskilled workers (un) loose jobs while low skilled workers without VT diploma (lsb) gain jobs. Unemployment rates for the latter could be reduced significantly over the medium and long term. A similar story can be told about workers with only secondary education or 9th grade plus VT (lsa). Indeed, lower dropout rates reduce the number of individuals who drop out of primary education and/or before completing 9th grade and obtain a VT diploma, or enter the labor market with only secondary education. This increases their wages and therefore firms have incentives to substitute them for skilled workers (secondary + VT). In this scenario, employment prospects for the latter could improve -- even if their salaries also go up as a result of lower supply reflecting a lower dropout rate at the Baccalauréat. For highly skilled workers (hsa and hsb) the situation is different. Higher internal efficiency implies a rapid increase in the number of university graduates. Even if their salaries initially go down and they can be substituted for skilled workers, downward rigidities put a limit to the demand given a level of output. Hence, without faster economic growth new highly skilled workers entering the market simply can not be absorbed. Total factor productivity. What would happen if total factor productivity increases as a result of: (i) better quality of education and more relevant diplomas/specializations in higher education and VT; (ii) better incentives to invest in training within enterprises; and (iii) other policies that affect innovation. In the simulations, the combined effect of these policies is assumed to increase the growth rate of total factor productivity growth rate by 1/5th of its baseline level. This increases investments and the demand for labor at a given wage. The results show important improvements 20 both in terms of economic growth and employment creation. GDP growth could increase by 1 percentage point over the short term and 1.3 percentage point over the medium term. The aggregate unemployment rate could be reduced by up to 2.6 percentage points (see Table A4). Unemployment rates by types of labor would fall for most types of labor, bringing down the aggregate unemployment rate relative to the baseline. The exceptions are highly skilled workers. With the assumed increase in TFP, the unemployment for university graduates with technical specializations could be reduced in the short term but would still be higher than in the baseline over the medium term, but much lower than in the previous scenario. The increase in unemployment for this category means that the demographic and educational effects on labor supply are stronger than the productivity and growth effect on labor demand for this category of workers. In essence, even with the higher rate of output growth generated in this scenario the economy would not be able to absorb the flow of qualified individuals. Again, this result is driven by the assumption of wages downward rigidity. This would seem to contradict the observation in Morocco that in certain high value added sectors there are shortages of technical professionals. These shortages, however, could still exist for specific diplomas/specializations even with a high unemployment rate for the broader class of skills used here. Nonetheless, one of the policy interventions analyzed below looks at the effect of the expansion of some of these strategic sectors. Social security reform. The simulations show that failing to reform the social security system and allowing social security contributions to increase can aggravate the unemployment problem. The simulation discussed here focuses on the pension system. It considers a gradual increase in the contribution necessary to make the pension system solvent ­ around a 10 percent increase per year during 5 years. This is the contribution rate that is necessary to ensure that current liabilities are equal to expected assets (including the pay-as-you-go asset) 17 even if today it increases the surplus of the pension system. The results show that higher contribution rates could increase the aggregate unemployment rate by 2.5 percentage points over the medium term (see Table A5). 17The pay-as-you-go asset is equal to the present value of future contributions net of the pension rights accruing from those new contributions. 21 Contrary to common belief, higher pay-roll-taxes and/or workers contributions could mainly affect highly skilled workers with technical specializations and workers with secondary education or intermediate levels of vocational training. These groups would face the sharpest reductions in employment levels and therefore the highest increases in unemployment rates -- close to 4 percentage points by year 2010. On the contrary, unskilled workers would be the least affected. There are two factors that explain these results. First, a portion of the unskilled and low skilled workers is likely to enter the informal sector (see below). Second, due to the fact that unskilled workers reach their natural unemployment level in 2012 (due to demographic and educational dynamics), there is much less pressure on low skilled workers which are relatively more substitutable than skilled workers. The higher cost of labor would also induce a mild expansion of the informal sector. Indeed, unskilled workers who do not find a job in the formal sector fill the ranks of those working in the informal sector. However, the increase in the contribution of the informal sector to total urban employment would be modest, starting with 0.1 percentage point and converging to close to one percentage point. On the contrary, reforming the pension system by gradually aligning benefits with contributions does not have a negative impact on employment creation. In the simulation, this assumes that pay-roll taxes and workers contributions stay constant at current levels, the social security surplus increases while the effects on growth are insignificant (Table A6). However, given the issue of financial sustainability of the social security system, and the necessity of a reform, this option seems more desirable if we take into account the objective of unemployment reduction. Support to emergent sectors. The results of the analysis indicate that each sector on its own will have little effects on economic growth and unemployment. The stimulus subsidy in the agro-industrial sector would have no impact over the short-term and very little impact over the medium term. The growth rate of GDP could increase by 0.1 percentage point and unemployment rates for some skills categories could be reduced by a similar amount (Table A7). No sizable effects would take place on the aggregate unemployment rate. The stimulus in the mechanical and electronics industrial sector could even have negative effects. Basically, the cost of the subsidy, which reduces government savings and aggregate investments, would be higher than the benefits in terms of 22 more investment and growth in the sector (Table A8). More promising results can be observed in the financial sector. The subsidy there could increase the GDP growth rate by 0.4 percentage point in 2008 and about 0.8 percentage point in 2015. Reductions in the unemployment rate would be modest, between 0.2 and 0.3 percentage points (see Table A9). The combined effect of the investment subsidy in the three sectors considered here would only add around 200,000 new jobs between 2007 and 2015. The almost totality of these jobs would be created through the stimulus in the financial sector. The stimulus in agro-industry would generate less than 10,000 jobs and the net impact in the mechanics and electronics industrial sector would be nil. Yet, the stimulus package would increase the government deficit by 0.2 to 0.5 percent of GDP over the short term (see Table A10). The main message here is that the Government would need to be conservative in terms of the direct impact that the Emergence Program has on employment. Clearly, there is a need to analyze the impact of the stimulus package on other sectors, but in general, it seems that reaching the target of 500,000 new jobs would require considerable public resources. As discussed in World Bank 2007 (Chapter 1), the value added of the Emergence Program could be more in terms of its impact on economic diversification, and through this channel, innovation/self-discovery and productivity growth. These effects were not captured in the simulations. The important role of productivity growth, however, was illustrated above. An integrated package. Not surprisingly, an integrated package that combines the various policies discussed above can have important impact on economic growth and employment creation. More specifically, this policy package would include improved internal efficiency in the education system, higher growth in total factor productivity resulting from policies that promote innovation within firms, social security reforms that keep pay-roll taxes constant and increase domestic savings, and the support to high-value added emergent sectors. Under this package, GDP growth could accelerate above 6.5 percent per year. Its level is higher by 1 percentage point to its reference scenario level during the whole investigation period. The aggregate unemployment rate would be reduced by 3 percentage points in the medium run (Table A11). Highly skilled unemployment will continue to 23 increase due to the high pace of increase of the number of graduates (mainly under the scenarios A and B), but in the integrated scenario this increase is lower. V. Conclusions This paper has introduced the dynamic general equilibrium model SSELMA (Social Security, Education, Labor, and the Macro-economy). The model was designed to help analysts to better understand how macroeconomic, education, and social protection policies interact to affect the dynamics of employment. The strength of the model relative to other models of its kind is the high resolution of the labor market (including the inclusion of a formal and informal sub-sector in all sector of the economy but agriculture), the inclusion of a social security module to project pension expenditures, and a human capital accumulation model that predicts the distribution of skills in the labor force. Like any model, SSELMA remains open to refinements. One of the main challenges is to validate empirically the appropriateness of selected functional forms and model parameters. Two critical modules are those determining the dynamics of the labor force between urban and rural areas and the formal and informal sectors, as well as the demand for various skills levels. The model was used to illustrate the impact that alternative education and social protection policies have on selected macroeconomic aggregates. The ones considered here include: GDP growth, the fiscal balance, the aggregate unemployment rate, unemployment rates by skill level, and the size of the informal sector. The results presented provide several insights. The simulations first confirm that, even under optimistic scenarios regarding economic growth high unemployment rates in urban areas would persist, particularly among high skilled workers. The baseline scenario in the model projects GDP growth rates between 5 and 6 percent. These are barely enough to bring down the urban unemployment rate below the 10 percent barrier in year 2015. Setting aside the issue of skills mismatch,18 the current economy seems unable to absorb current flows of skilled and highly skilled workers. Both skilled and highly skilled workers face the highest 18These are not modeled within SELMA. 24 unemployment rates.19 The projections show that for skilled workers these are likely to increase approximating 50 percent over the medium term. For highly skilled workers unemployment rates could increase over the short term and eventually fall over the medium term, but would remain above 25 percent for non-technical specializations and 10 percent for technical specializations. The simulations also suggest that simply improving internal efficiency in the education system could make things worse. Higher internal efficiency in the education system is a pre-condition to sustain the accumulation of human capital over the medium and long term and support economic growth. But having more people graduating from university does not automatically mean higher aggregate productivity and faster growth. Over the short and medium term, the higher inflow of skilled workers would contribute to increased unemployment rates. Unskilled workers would also be affected as their number is reduced, their wages go up, and then get substituted by more skilled workers. In the scenarios considered the model predicts that not reforming the social security could cost much more jobs than the Emergence Program could create under its best performance. Not reforming the pension system and/or not controlling the finances of the new health insurance system and letting pay-roll taxes and social security contributions increase would have a negative effect on employment levels. The aggregate unemployment rate could increase by up 2.5 percentage points as a result of a 10 percent increase in the contribution rate over a period of 5 years. Effects are particularly important for highly skilled workers for whom unemployment rates could increase by up to 4 percentage points. The model also warns about having high expectations regarding the effects of the Emergence Program. The stimulus package simulated in three sectors (agro-industry, financial sector, and mechanical and electronic industry) generated only 200,000 new jobs between 2007 and 2015 at a cost of 0.2 to 0.5 percent of GDP per year. If these ratios are preserved, the 500,000 target that the government has set could cost each year between 0.5 percent and 1.25 percent of GDP. But probably the most important message from the analysis is that isolated interventions will be insufficient to reduce unemployment. Each of the policy interventions simulated here can affect 19In our application, skilled workers refer to individuals who have a diploma of Specialized Technician from the VT system. Highly skilled workers, on the other hand, refer to university graduates, which the model separates into technical and non-technical diplomas. 25 unemployment rates on its own, but the effects tend to be modest. A sustained reduction in unemployment rates requires an integrated and well coordinated package of policy interventions. We argue that an integrated package that can rapidly reduce unemployment rates for all categories of workers would seek the following general objectives: (i) foster total factor productivity growth through better quality and the improved relevance of the diplomas/specialization supplied by the higher education and VT system, as well as better incentives to invest in in-service training; (ii) increase investments and promote faster growth in sectors intensive in highly-skilled workers; and (iii) reform the social security system to control labor costs. Such a package could also reduce the size of the informal sector, which implies an improvement in the average quality of the jobs created. The challenge of course is to identify the specific policies that would achieve these objectives. Several are discussed in a recent report on skills development, social protection, and employment in Morocco (see World Bank, 2007). References Abdelkalek T. and J.M. Dufour. 2006. "Confidence Regions for Calibrated Parameters in Computable General Equilibrium Models, Annales d'Économie et de Statistique, N° 81, pp. 1-31. P.R. Agenor A. Izquierdo and H.T. Jensen. 2007. Adjustment Policies, Poverty, and Unemployment: The IMMPA Framework, Blackwell Publishing. Agénor, P.R. and K. El Aynaoui, 2003. "Labor Market Policies and Unemployment in Morocco: A Quantitative Analysis", Policy Research Working Paper N°3091, the World Bank, Washington D.C. Blanchflower D. and Oswald A. (1994). The Wage Curve, MIT Press. Cogneau D., M. Razafindrakoto et F. Roubaud. 1996. « Le secteur informel urbain et l'ajustement au Cameroun », Revue d'économie du développement, (3), pp. 27-63. Bchir M.H., Y. Decreux, J.L. Guerin and S. Jean. 2002., "MIRAGE, a Computable General Equilibrium Model for Trade Policy Analysis", CEPII Working Paper N°17. Bodor Andras, David Robalino, and Michal Rutkowski. 2007. "Assessing the Distortions of Mandatory Pensions on Labor Supply Decisions and Human Capital Accumulation: Options to Bridge the Gap between Economic Theory and Policy Analysis". Policy Research Working Paper Series. No. xx World Bank. Washington. DC. 26 Fallon, P. and P. R. G. Layard (1975). "Capital-skill complementarity, income distribution and output accounting." Journal of Political Economy 83(2): 279-301. Fields, Gary S. (2001). Distribution and Development: A New Look at the Developing World. (Cambridge, MA: MIT Press and the Russell Sage Foundation). Fields, Gary S. (2005). "A Welfare Economic Analysis of Labor Market Policies in the Harris-Todaro Model," Journal of Development Economics 76: 127-146. Fields Gary. 2007. "Modeling Labor Market Policy in Developing Countries: A Selective Review of the Literature and Needs for the Future". Working Paper. World Bank. Washington DC. Lambert, Peter (1989). The Distribution and Redistribution of Income. (Oxford: Blackwell). Somavía, Juan (1999). Decent Work. Report of the Director-General to the 87th Session of the International Labor Conference. (Geneva: ILO). World Bank 2006. "Morocco: Country Economic Memorandum." World Bank. Washington DC. Squire, Lyn (1981). Employment Policy in Developing Countries (New York: Oxford University Press for the World Bank). World Bank 2006. "Morocco: Country Economic Memorandum." World Bank. Washington DC. World Bank 2007. "Morocco: Skills Development And Social Protection Within An Integrated Strategy For Employment Creation." Green Cover Report. World Bank. Washington DC. 27 Appendix Table A1: main elasticities of the model Elasticity of substitution between imports and local products 1.5 Elasticity of transformation between local products and exports 1.5 Elasticities of substitution -between a capital-highly skilled labor bundle (KHS) and a low skilled labor- 0.5 unskilled labor bundle (ULS). -between capital (K) and a skilled-highly skilled labor demand bundle (HSK) 0.4 - between unskilled labor (un) and a low skilled labor demand bundle (LS) 0.7 - between a composite highly skilled labor bundle (HS) and skilled labor (sk) 0.8 -between two highly skilled workers categories, namely hsa and hsb. 0.85 -between two low skilled categories lsa and lsb. 0.9 -between temporary and permanent labor 1.5 Table A2: The Baseline Scenario Macro Results 2007 2008 2009 2010 2011 2012 2013 2014 2015 GDP Billion Dh 516.9 543.3 573.6 608.6 640.4 677.1 718.4 764.3 814.6 GDP Growth 4.6% 5.1% 5.6% 6.1% 5.2% 5.7% 6.1% 6.4% 6.6% Government Balance Billion Dh -0.8 -0.1 0.8 2.0 3.0 3.9 5.0 6.3 7.5 Social Security Balance Billion Dh 7.90 7.73 7.41 6.99 6.39 5.63 4.64 3.48 2.07 Total investment Billion Dh 147.7 161.0 177.4 197.3 215.0 234.8 257.4 282.7 310.0 Total Unemployment 12.7% 13.1% 12.8% 11.9% 11.5% 11.1% 10.6% 9.9% 9.3% Share informal labor supply 23.7% 23.5% 23.4% 23.2% 23.1% 23.1% 23.1% 23.1% 23.2% Total Urban labor supply (Million) 9.74 10.09 10.41 10.70 10.97 11.22 11.45 11.66 11.86 Unemployment un 6.0% 5.8% 5.0% 3.5% 3.0% 3.0% 3.0% 3.0% 3.0% lsb 12.5% 12.3% 11.8% 10.9% 10.0% 8.6% 6.9% 4.7% 3.0% lsa 12.2% 12.2% 11.6% 10.2% 8.9% 7.0% 4.7% 3.0% 3.0% sk 41.4% 44.1% 45.7% 46.5% 47.1% 47.5% 47.6% 47.5% 47.2% hsb 32.9% 33.4% 33.2% 32.7% 32.2% 31.3% 29.9% 28.1% 26.0% hsa 16.5% 18.7% 18.5% 17.6% 17.0% 15.7% 14.3% 12.0% 9.3% 28 Table A3: Improving Internal Efficiency in the Education System (Policy A) Variation in % compared to the reference scenario Macro Results 2007 2008 2009 2010 2011 2012 2013 2014 2015 GDP 0.1% 0.1% 0.1% 0.1% 0.2% 0.2% 0.3% 0.2% 0.0% GDP Growth (percentage points) 0.0% 0.0% 0.0% 0.0% 0.1% 0.0% 0.0% -0.1% -0.1% - Government Balance 2.6% -25.8% 4.1% 1.9% 5.2% 4.8% 3.6% 1.0% -1.2% Social Security Balance 0.1% 0.2% 0.2% 0.2% -0.3% -0.4% -0.3% 0.4% 1.7% Total investment 0.2% 0.2% 0.2% 0.2% 0.5% 0.6% 0.6% 0.3% -0.1% Total Unemployment (percentage points) 0.0% 0.0% 0.0% 0.0% -0.5% -0.5% -0.4% 0.3% 0.9% Share informal labor supply (percentage points) 0.2% 0.2% 0.1% 0.1% 0.0% -0.2% -0.2% -0.3% -0.4% Unemployment (percentage points) Un 1.5% 1.7% 1.7% 1.6% 0.4% 0.0% 0.0% 0.0% 0.0% - - Lsb 0.1% -0.2% 0.5% 1.0% -1.6% -2.1% -2.8% -1.7% 0.0% Lsa 2.2% 3.1% 4.0% 4.9% 5.8% 6.8% 7.9% 7.6% 5.0% - - - Sk 8.1% -8.6% 8.8% 9.0% -9.3% -9.5% -9.6% -9.5% -9.3% 12.5 15.6 18.3 21.0 23.8 Hsb 2.6% 3.9% 5.8% 9.3% % % % % % 10.7 Hsa 1.8% 2.2% 2.5% 3.0% 4.0% 5.1% 6.9% 8.7% % Table A4: Internal Efficiency and Faster Total Productivity Growth (Policy B) Variation in % compared to the reference scenario Macro Results 2007 2008 2009 2010 2011 2012 2013 2014 2015 GDP 3.3% 4.3% 5.5% 6.7% 7.9% 9.2% 10.3% 11.4% 12.5% GDP Growth (percentage points) 0.9% 1.0% 1.2% 1.2% 1.1% 1.3% 1.1% 1.1% 1.0% 244.3 125.9 102.2 Government Balance -125.0% -1357.9% % % % 89.3% 77.7% 67.4% 60.1% Social Security Balance 1% 1% 2% 2% 3% 3% 2% 0% -9% 13.1 Total investment 8.2% 10.6% % 15.3% 17.2% 19.0% 20.5% 21.5% 22.3% Total Unemployment (percentage points) -1.4% -1.9% -2.6% -2.6% -2.5% -2.4% -1.8% -0.9% -0.2% Share informal labor supply (percentage points) 0.1% 0.0% -0.1% -0.2% -0.2% -0.3% -0.3% -0.2% 0.0% Unemployment (percentage points) Un 0.1% -0.3% -1.0% -0.5% 0.0% 0.0% 0.0% 0.0% 0.0% Lsb -1.2% -1.9% -2.8% -4.1% -5.4% -5.6% -3.9% -1.7% 0.0% Lsa 0.3% 0.6% 0.7% 0.6% 0.5% 0.2% -0.6% 0.0% 0.0% - 11.1 - - - - Sk -9.3% -10.3% % -11.7% -12.5% 13.2% 13.6% 13.9% 14.0% Hsb 1.4% 2.3% 3.7% 6.9% 9.8% 12.6% 15.1% 17.6% 20.1% Hsa 0.2% 0.1% -0.3% -0.6% -0.1% 0.3% 1.6% 3.0% 4.6% 29 Table A5: Failure to Reform the Social Security (Policy C) Variation in % compared to the reference scenario Macro Results 2007 2008 2009 2010 2011 2012 2013 2014 2015 GDP 0.5% 0.6% 0.8% 1.2% 1.3% 2.1% 2.8% 3.5% 4.4% GDP Growth (percentage points) 0.2% 0.2% 0.2% 0.4% 0.2% 0.8% 0.7% 0.8% 0.9% - 259.6 - 19.1 Government Balance 26.1% % 39.7% % -2.7% 3.0% 5.4% 7.5% 10.4% Social Security Balance 65% 109% 167% 245% 284% 345% 452% 652% 1191% 12.7 13.2 Total investment 4.9% 7.4% 10.0% % % 14.0% 14.8% 15.7% 16.9% Total Unemployment (percentage points) 0.9% 1.4% 2.0% 2.5% 1.9% 1.1% 1.0% 0.9% 0.4% Share informal labor supply (percentage points) 0.1% 0.2% 0.3% 0.4% 0.4% 0.5% 0.6% 0.7% 0.8% Unemployment (percentage points) Un 0.9% 1.4% 1.9% 2.4% 1.4% 0.0% 0.0% 0.0% 0.0% Lsb 0.8% 1.3% 1.8% 2.2% 2.2% 2.0% 1.7% 1.5% 0.4% Lsa 1.4% 2.3% 3.1% 4.0% 4.0% 3.9% 3.6% 2.3% 0.0% Sk 0.9% 1.4% 1.8% 2.2% 2.0% 1.8% 1.6% 1.5% 1.3% Hsb 1.1% 1.7% 2.3% 2.9% 2.8% 2.6% 2.4% 2.3% 2.1% Hsa 1.5% 2.2% 3.0% 3.9% 3.7% 3.5% 3.3% 3.2% 3.0% Table A6: Reforming the Social Security (Policy D) Variation in % compared to the reference scenario Macro Results 2007 2008 2009 2010 2011 2012 2013 2014 2015 GDP 0.0% -0.1% -0.1% -0.1% -0.1% -0.1% -0.1% -0.1% 0.0% GDP Growth (percentage points) 0.0% -0.1% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% Government Balance 5.1% 26.5% -3.7% -1.7% -0.9% -0.7% -0.6% -0.4% 5.1% Social Security Balance 3.8% 3.5% 4.6% 5.3% 5.6% 5.8% 6.0% 5.8% 3.8% Total investment 0.2% 0.0% 0.0% 0.0% 0.0% 0.0% -0.1% -0.1% 0.2% Total Unemployment (percentage points) 0.0% 0.1% 0.1% 0.1% 0.0% 0.0% 0.0% 0.0% 0.0% Share informal labor supply (percentage points) 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% Unemployment (percentage points) Un 0.0% 0.1% 0.1% 0.1% 0.0% 0.0% 0.0% 0.0% 0.0% Lsb 0.0% 0.1% 0.1% 0.1% 0.1% 0.1% 0.1% 0.1% 0.0% Lsa 0.1% 0.1% 0.1% 0.1% 0.1% 0.1% 0.1% 0.0% 0.0% Sk 0.1% 0.1% 0.1% 0.1% 0.1% 0.1% 0.1% 0.0% 0.0% Hsb 0.1% 0.1% 0.1% 0.1% 0.1% 0.1% 0.1% 0.1% 0.1% Hsa 0.1% 0.1% 0.1% 0.1% 0.1% 0.1% 0.1% 0.1% 0.1% 30 Table A7: Stimulus to the Agro-Industrial Sector (Policy E) Variation in % compared to the reference scenario Macro Results 2007 2008 2009 2010 2011 2012 2013 2014 2015 GDP 0.0% 0.0% 0.0% 0.1% 0.1% 0.1% 0.1% 0.1% 0.2% GDP Growth (percentage points) 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% - - - - - Government Balance 40.5% 286.9% 38.4% 16.9% 13.7% 11.8% 10.0% -8.8% -8.2% Social Security Balance 0% 0% 0% 0% 0% 0% 0% 0% 0% Total investment 0.0% 0.1% 0.2% 0.3% 0.3% 0.4% 0.4% 0.4% 0.4% Total Unemployment (percentage points) 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% -0.1% 0.0% Share informal labor supply (percentage points) 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% Unemployment (percentage points) Un 0.0% 0.0% 0.0% -0.1% 0.0% 0.0% 0.0% 0.0% 0.0% Lsb 0.0% 0.0% 0.0% 0.0% 0.0% -0.1% -0.1% -0.1% 0.0% Lsa 0.0% 0.0% 0.0% -0.1% -0.1% -0.1% -0.2% 0.0% 0.0% Sk 0.0% 0.0% 0.0% 0.0% 0.0% -0.1% -0.1% -0.1% -0.1% Hsb 0.0% 0.0% 0.0% 0.0% -0.1% -0.1% -0.1% -0.1% -0.1% Hsa 0.0% 0.0% 0.0% -0.1% -0.1% -0.1% -0.1% -0.1% -0.2% Table A8: Stimulus to the Financial Sector (Policy F) Variation in % compared to the reference scenario Macro Results 2007 2008 2009 2010 2011 2012 2013 2014 2015 GDP 0.0% 0.4% 0.4% 0.5% 0.5% 0.6% 0.7% 0.7% 0.8% GDP Growth (percentage points) 0.0% 0.4% 0.0% 0.1% 0.0% 0.1% 0.1% 0.1% 0.0% - - - - - - - Government Balance 101.6% 640.9% 84.0% 36.5% 28.1% 22.4% 18.7% 16.3% 15.2% Social Security Balance 0% 0% 0% 0% 0% 0% 0% 0% 0% Total investment 0.0% 1.2% 1.5% 1.6% 1.6% 1.7% 1.8% 1.8% 1.7% Total Unemployment (percentage points) 0.0% -0.2% -0.3% -0.3% -0.2% -0.2% -0.2% -0.2% -0.1% Share informal labor supply (percentage points) 0.0% 0.0% 0.0% 0.0% 0.0% -0.1% -0.1% -0.1% -0.1% Unemployment (percentage points) Un 0.0% -0.2% -0.3% -0.3% 0.0% 0.0% 0.0% 0.0% 0.0% Lsb 0.0% -0.2% -0.2% -0.3% -0.3% -0.4% -0.4% -0.6% 0.0% Lsa 0.0% -0.3% -0.3% -0.4% -0.5% -0.6% -0.7% 0.0% 0.0% Sk 0.0% -0.1% -0.2% -0.2% -0.2% -0.3% -0.3% -0.3% -0.3% Hsb 0.0% -0.2% -0.2% -0.3% -0.3% -0.4% -0.4% -0.5% -0.5% Hsa 0.0% -0.2% -0.3% -0.3% -0.4% -0.5% -0.5% -0.6% -0.6% 31 Table A9: Stimulus to the Mechanical and Electronic Industrial Sector (Policy G) Variation in % compared to the reference scenario Macro Results 2007 2008 2009 2010 2011 2012 2013 2014 2015 GDP 0.0% -0.1% -0.2% -0.2% -0.2% -0.3% -0.3% -0.3% -0.3% GDP Growth (percentage points) 0.0% -0.1% -0.1% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 98.5 730.5 - - - - - - Government Balance % % -107.5% 49.3% 37.3% 29.9% 25.1% 21.6% 19.1% Social Security Balance 0% 0% 0% 0% 0% 0% 0% 0% 1% Total investment 0.0% -0.4% -0.6% -0.8% -0.8% -0.9% -0.9% -0.9% -0.8% Total Unemployment (percentage points) 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% Share informal labor supply (percentage points) 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% -0.1% -0.1% Unemployment (percentage points) Un 0.0% 0.0% 0.0% 0.1% 0.0% 0.0% 0.0% 0.0% 0.0% Lsb 0.0% 0.0% 0.0% 0.0% 0.0% 0.1% 0.1% 0.1% 0.0% Lsa 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% Sk 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% Hsb 0.0% -0.1% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% Hsa 0.0% -0.1% 0.0% -0.1% -0.1% 0.0% 0.0% 0.0% 0.0% Table A10: Stimulus to Agro-Industrial Sector, Financial Sector, and Mechanical and Electronic Industrial Sector (Policy H) Variation in % compared to the reference scenario Macro Results 2007 2008 2009 2010 2011 2012 2013 2014 2015 GDP 0.0% 0.3% 0.3% 0.3% 0.4% 0.4% 0.5% 0.6% 0.6% GDP Growth (percentage points) 0.0% 0.3% 0.0% 0.0% 0.0% 0.1% 0.1% 0.1% 0.0% 1645.5 - - - - - - - Government Balance 234.9% % 231.2% 103.7% 80.4% 64.7% 54.7% 47.2% 43.0% Social Security Balance 0% 0% 0% 0% 0% 0% 1% 1% 1% Total investment 0.0% 0.9% 1.0% 1.1% 1.1% 1.2% 1.3% 1.4% 1.3% Total Unemployment (percentage points) 0.0% -0.2% -0.3% -0.3% -0.2% -0.2% -0.3% -0.3% -0.1% Share informal labor supply (percentage points) 0.0% 0.0% 0.0% -0.1% -0.1% -0.1% -0.1% -0.2% -0.2% Unemployment (percentage points) Un 0.0% -0.2% -0.3% -0.3% 0.0% 0.0% 0.0% 0.0% 0.0% Lsb 0.0% -0.2% -0.2% -0.3% -0.3% -0.4% -0.5% -0.6% 0.0% Lsa 0.0% -0.3% -0.4% -0.5% -0.6% -0.7% -0.8% 0.0% 0.0% Sk 0.0% -0.2% -0.2% -0.3% -0.3% -0.3% -0.4% -0.4% -0.4% Hsb 0.0% -0.2% -0.3% -0.3% -0.4% -0.5% -0.5% -0.6% -0.6% Hsa 0.0% -0.3% -0.3% -0.4% -0.5% -0.6% -0.7% -0.8% -0.8% Table A11: An Integrated Policy Package (Policy J) 32 Variation in % compared to the reference scenario Macro Results 2007 2008 2009 2010 2011 2012 2013 2014 2015 GDP 3.3% 4.5% 5.6% 6.9% 8.1% 9.4% 10.7% 11.9% 13.0% GDP Growth (percentage points) 1.0% 1.2% 1.2% 1.2% 1.2% 1.3% 1.2% 1.1% 1.1% 137.5 Government Balance % 507.4% -30.9% -2.2% 3.4% 9.5% 10.1% 10.2% 9.5% Social Security Balance 5% 5% 6% 8% 9% 10% 10% 8% -3% Total investment 8.3% 11.5% 14.2% 16.3% 18.3% 20.2% 21.7% 22.7% 23.5% Total Unemployment (percentage points) -1.3% -2.0% -2.7% -2.7% -2.7% -2.5% -1.9% -1.0% -0.2% Share informal labor supply (percentage points) 0.1% 0.0% -0.1% -0.2% -0.3% -0.4% -0.4% -0.3% -0.2% Unemployment (percentage points) Un 0.1% -0.5% -1.1% -0.5% 0.0% 0.0% 0.0% 0.0% 0.0% Lsb -1.1% -2.0% -3.0% -4.3% -5.7% -5.6% -3.9% -1.7% 0.0% Lsa 0.4% 0.3% 0.5% 0.3% 0.1% -0.6% -1.4% 0.0% 0.0% - - - - - Sk -9.3% -10.4% -11.2% -12.0% 12.7% 13.5% 14.0% 14.3% 14.4% Hsb 1.4% 2.2% 3.6% 6.7% 9.6% 12.3% 14.8% 17.2% 19.8% Hsa 0.3% -0.1% -0.5% -0.9% -0.4% -0.2% 1.1% 2.3% 3.9% 33