Morocco - Country economic memorandum : towards higher growth and employment (Vol. 2 of 2) : Annexes

Report No. 14155-MOR Kingdom of Morocco Country Economic Memorandum.:Towards Higher Growth and Employment (In Two Volumes) Volume II Annexes September 15, 1995 Country Operations Division Country Department I Middle East and North Africa Region *lo~m.nt * ' Wo-' ''*-'  CURRENCY AND EXCHANGE RATE Currency Unit = Dirham (DH) 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 DH per US$, End of Period 9.55 9.62 8.71 7.80 8.21 8.12 8.04 8.15 9.05 9.65 DH per US$ Period Average 8.81 10.06 9.10 8.36 8.21 8.49 8.24 8.71 8.54 9.30 FISCAL YEAR January Ist - December 31st This report was supervised and drafted by Roumeen Islam. It is based on the contributions of a team consisting of Jean-Paul Azam (The Labor Market, consultant), Jacques Coudol and John Nellis (Privatization and Private Sector Development), Luc De Wulf (Financial Sector Development), Kouassi Soman (The Macroeconomic Framework, Statistical Annex), and Sethaput Suthiwart-Narueput (International Trade). Khadija Khoudari (Gender Issues in the Labor Force, consultant), Benoit Millot (Gender Issues in Education), and Anita Schwarz (Financial Sector Development - Pension Systems) also contributed to the report. Ishac Diwan, Ross Levine and Klaus Schmidt-Hebbel were peer reviewers for the report and provided guidance. Additional advice and support was provided by Rend Vaurs, John Underwood and Miria Pigato. Carine Bolou-Mansilla and Magalie Pradel desk-topped the report.  LIST OF ACRONYMS AND ABBREVIATIONS BNDE Banque Nationale pour le D6veloppement Economique BRPM Bureau de Recherches et des Participations Minieres CDT Confederation D6mocratique du Travail CDVM Comite Deontologique des Valeurs Mobilieres CGE Calibrated General Equilibrium CHELCO A Private Company CIH Credit Immobilier et H6telier CIMR Caisse Interprofessionnelle Marocaine de Retraite CIOPE Centre d'Information et d'Orientation pour l'Emploi CIOR Ciment de l'Oriental CMH A Private Company CMM Caisse Marocaine des Marches CNCA Caisse Nationale de Credit Agricole CNRA Caisse Nationale de Retraite et d'Assurance CNRE Caisse Nationale de Retraite et d'Epargne COMANAV A Private Company CPI Consumer Price Index CREDIT-EQDOM A Household Furnishing Credit Agency CTM-LN A Bus Company DW Durbin-Watson EIR Effective Incentive Rate ENPS National Survey on Population and Health EPR Effective Protection Rate FDI Foreign Direct Investment FEC Fonds d'Equipement Communal FERTIMA A Private Company FF Fully Funded FRAM A French Company GATT General Agreement on Tariffs and Trade GAVW Growth of Average Wage GPBM Confeddration Generale des Banques Marocaines HS Harmonized System IBRD International Bank for Reconstruction and Development IGR Impot General sur le Revenu (Income Tax) IMF International Monetary Fund IS Corporate Business Tax ITCB International Textiles and Clothing Bureau MENA Middle East and North Africa MOBIL A Private Company MODULEC A Private Company MUV Manufacturing Unit Value OECD Organization for Economic Cooperation and Development ONT Office National des Transports P/E Price/Earning PEP Plancher des Effets Publics PERL Public Enterprise Restructuring Loan  PETROM A Private Company QR Quantitative Restriction RAD A Private Company RCAR Regime Collectif d'Allocation de Retraite ROW Rest of the World SAMIR An Oil Refinery SCP An Oil Refinery SFI A Financial Institution SITC A Public Enterprise SMIF Societe Marocaine d'Intermediation Financiere SNI A Public Enterprise SOCHEPRESS A Newspaper and Magazine Distributor SODERS A Private Company SOFAC A Private Company SOMAS A Gas Storage Company SONASID A Private Company SUBM A Sugar Company SUNAT A Sugar Company SUTA A Sugar Company UGTM Union G&nerale des Travailleurs Marocains UMA Arab Maghreb Union UMT Union Marocaine du Travail UNDP United Nations Development Program UNICEF United Nations Children Fund USAID United States Agency for International Development VAT Value-added Tax  KINGDOM OF MOROCCO COUNTRY ECONOMIC MEMORANDUM TOWARDS HIGHER GROWTH AND EMPLOYMENT VOLUME II: ANNEXES CONTENTS ANNEX 1: THE MACROECONOMIC FRAMEWORK ........................... I A. Background ........................................... I B. Recent Economic Developments .............................. I C. Medium-Term Prospects - Low Case Scenario ..................... 6 D. Medium-Term Prospects - High Case Scenario .................... 10 ANNEX II: INTERNATIONAL TRADE .............. ..1.................... A. Additional Notes on the Uruguay Round ......................... 1 B. Notes on Morocco's Trade Regime ............................ 2 C. Preliminary Agricultural Agreement with the EU ................... 6 D. Fiscal Impact of Trade Liberalization ........................... 7 E. Trade Reform and Labor Market Rigidities ....................... 8 F. Free Trade Agreement with the EU ........................... 10 ANNEX III-A: THE LABOR MARKET .................... I 1. An Overview and Main Conclusions ................................. 1 2. A Statistical Analysis of the Functioning of the Urban Labor Market in Morocco .... 10 A. Earnings in the Urban Private Sector .......................... 12 B. Impact of the Minimum Wage and Social Charges on Urban Private Sector Wages .................................. 15 C. Earnings in the Government Sector and the Public and Semi-Public Urban Sectors ...................................... 20 D. Urban Unemployment ................. I ................. 23 E. Unemployment Duration and Matching ......................... 29 ANNEX III-B: GENDER ISSUES IN THE LABOR MARKET ...................... 1 1. Labor Force Education ......................................... 1 A. Enrollments ........................................... 1 B. Determinants of Gender Disparities ............................ 3 C. Government and Donor Strategy .............................. 5 D. Family Planning ........................................ 6 II. Labor Force Participation ....................................... 6 A. Women in the Labor Market ................................ 6 B. The Rural Labor Market .................................. 12  Table of Contents (cont'd) C. Conclusion .. ............. 13 ANNEX IV-A: FINANCIAL SECTOR REFORM .1 A. Introduction and Background .1 B. The Reform Agenda .. 7 1. Interest Rate Liberalization ............................. 7 2. Legal and Regulatory Reform of Securities, Bond and Money Markets .10 3. Reform of the Postal and Contractual Savings System .15 4. Eliminating the Government's Preferential Access to Credit .20 5. Accelerating the Use of Indirect Instruments for the Conduct of Monetary Policy .22 6. Integrating Morocco into World Financial Markets .23 ANNEX IV-B: PENSION SYSTEMS IN MOROCCO. I A. Introduction . B. Brief Description of Major Pension Systems. 2 C. Demographic Outlook for Morocco. 5 D. Simulated Results for CNSS and CMR. 6 E. Reform Proposals .11 ANNEX V: PUBLIC ENTERPRISES AND PRIVATIZATION . . A. The Public Enterprise Sector: Size, Significance and Performance . B. The Privatization Program .8 C. Towards a New Public-Private Mix for Morocco .18 D. Conclusion .26 ANNEX VI: STATISTICAL APPENDIX  ANNEX I THE MACROECONOMIC FRAMEWORK This annex is subdivided into three parts. First, it provides a brief background of the Moroccan economy by describing the main achievements of the last decade of stabilization and recent economic developments. Second, it provides more detail on a "business as usual" low case and a dynamic high case scenario as discussed in the main text of the Country Economic Memorandum. A. Background 1. Morocco's experience with stabilization and structural reforms has been successful. It has emerged from a country with an economic andfinancial crisis in the early 1980s to one with a relatively more stable macroeconomic environment over the past fewyears. A steady pursuit of stringent monetary and fiscal policies reduced the budget deficit from an average of 12 percent of GDP in 1980-83 to 3.2 percent in 1993; domestic inflation is under control and has been kept within single digits. Better external debt management and creditors' assistance through debt relief have combined to improve foreign debt indicators consistently since the mid-1980s. Gradual but persistent implementation of stabilization and structural reforms has resulted in a shift from Government-administered to more market-determined pricing in economic, financial and trade areas; this has increasingly exposed the economy to domestic and external competition. 2. Nonetheless, performance indicators have lagged in key social areas such as literacy and primary school enrollment rates (especially among women), nutritional status and access to safe water, and more importantly labor absorption. The latter has resulted in one of the most challenging issues for this decade, i.e. the problem of a high and rising urban unemployment that averaged some 15 percent over the past five years. Furthermore, macroeconomic developments especially in the area of public finance in the past two years indicate an emergence of small but increasing imbalances. Over the 1992-1994 period, the fiscal deficit has increased from 2.2 percent to 3.8 percent of GDP despite a real increase of 11.5 percent in the latter in 1994, resulting from the doubling of central government domestic borrowing relative to GDP. Exports have fallen by two percentage points of GDP due partly to the appreciation of the real effective exchange rate of 7.9 percent since from 1990-1994. The end of external debt relief since 1992 has translated into an increase of 10 percentage points in the debt service ratio over the period. In addition, both investment and savings have continuously declined as a share of GDP. All these negative trends are projected to worsen in 1995 with fiscal and current account deficits reaching 4.2 and 3.4 percent of GDP, respectively, and with inflation rising. These developments indicate the need for further macroeconomic stabilization. B. Recent Economic Developments 3. The drought-related slumps of recent years indicate that the Moroccan economy is still strongly dependent on agriculture despite efforts to industrialize and diversify the sources of economic growth over the past decade. After a drought-induced 4 percent real decline in GDP in 1992, followed by a further I percent drop in 1993, the outcome for economic growth in 1994 was excellent (key economic indicators are summarized in table 1). Abundant rainfall boosted overall GDP growth by 11.5 percent, with the agriculture sector leading at 63 percent. Total cereal output is expected to reach a record 9 million tons for the 1993-94 crop campaign, from 2.1 million tons the previous season. Moreover, growth in non- agricultural value-added reached 3.6 percent in 1994 with tourism-related services leading at 4.3 percent. Annex I  Table 1. Historical Macroeconomic Indicators Period Averages Yearly Prel. 1980-83 1984-87 1988-92 1992 1993 1994 Growth Rates (% p.a.) GDP at Market prices 3.6 4.7 3.9 -4.4 -1.1 11.5 - Agriculture -0.7 10.3 1.8 -36.0 -6.2 63.0 - Non-agriculture 4.5 3.5 4.2 4.0 -0.2 3.6 Per-capita GDP 0.3 2.2 1.6 -6.4 -3.1 9.4 Inflation (GDP deflator) 8.0 8.1 5.1 4.5 3.8 2.7 Share of GDP at m.p. A. National Accounts Gross Investment 25.6 24.1 23.1 22.7 21.2 21.0 Fixed Capital Formation 25.0 21.9 22.5 22.8 22.4 20.2 - Central Governrent 7.7 4.9 4.4 4.7 6.2 5.5 - Other Public Sector 5.2 5.3 4.9 5.0 3.7 3.0 -Private Sector 12.1 11.8 13.2 13.1 12.5 11.7 Five-Year ICOR 5.2 6.1 4.7 5.8 14.7 6.5 Domestic Savings 13.5 16.5 18.3 16.2 15.7 15.9 National Savings 16.1 20.3 22.0 21. 1 19.1 18.7 - Central Government -1.5 -1.8 3.0 4.9 4.6 2.9 - Non-Central Government 17.6 22.1 19.0 16.2 14.5 15.8 Foreign Savings 9.6 3.7 1.1 1.6 2.0 2.3 B. Central Goverfment Budget Current Revenue 21.6 20.3 23.7 26.4 26.7 23.4 Governnent Expenditure 33.8 29.2 27.6 28.6 30.0 27.2 Fisc. Deficit, com. basis -12.2 -8.8 -3.9 -2.2 -3.3 -3.8 - Privatization proc. 0.0 0.0 0.0 0.0 0.9 0.7 - Domestic borrow, net 5.6 5.8 1.6 2.1 2.8 4.8 - External borrow., net 6.6 3.1 2.3 0.1 -0.4 -1.7 C. Balance of Payments Exports of GNFS 19.6 23.1 23.2 22.8 23.3 21.7 Imports of GNFS 31.7 30.7 28.0 29.3 28.8 26.8 Resource Balance -12.1 -7.6 -4.8 -6.5 -5.5 -5.1 Current Act. Deficit -9.6 -3.7 -1.1 -1.6 -2.0 -2.3 Foreign Direct Invest. 0.7 0.5 1.1 1.8 2.0 2.5 Reserves (months of imp.) 1.3 1.2 3.5 5.5 6.2 6.4 D. External Debt Burden Debt Outstanding & Disbursement 74.6 113.0 85.7 74.6 80.5 68.5 Debt Service/XGS 37.7 31.3 25.7 23.3 30.4 32.6 Sources: Official data; Statistics Office and Ministry of Finance. 4. Gross investment levels have declined by four to five percentage points of GDP since the economic crisis of the early 1980s as smaller but more productive private investments have gradually replaced inefficient and high government capital outlays. Central Government investment declined to merely 5.5 percent of GDP in 1994 compared to an average of 8 percent in the 1980-83 period. Private sector investment averaged 13 percent of GDP per year in the 1985-1992 period, or 1.4 times public sector investment; total investment relative to GDP over the same period averaged 22 percent per year. Annex I  - 3 - However, foreign investment in industrial activities increased modestly, with its share in total industrial outlays increasing by one percentage point to 21 percent over the 1988-92 period; but the industrial investment performance of the domestic private sector was better. In spite of the gradual diversification of investment in manufacturing over the past several years, the three traditional sectors of agro-industry, textiles and leather, and chemicals still hold a major share at some 70 percent in 1992. 5. With a relatively stable share of gross investment in GDP in recent years, developments in consumption have closely followed that of income estimated at US$1040 in GNP per-capita terms in 1993"; domestic savings rates are lower than comparable middle income countries. Except for the 1992- 93 drought years, per-capita private consumption has grown steadily at 3-4 percent per annum since 1984 with a sharp increase in 1994 due to the rise in agricultural incomes. Domestic savings have increased over the past five years due, in part, to stronger fiscal performance; they remain relatively low nonetheless. National savings rates, averaging some 20 percent of GDP in the past decade, have suffered from large outflows of factor service income, especially since the end of external debt relief in 1992; however, savings rates have continuously improved on the basis of substantial inflows of private transfers from Moroccans living abroad (7-8 percent of GDP). 6. Inflation is under control after the sharp increase in the 1990-91 period when changes in the urban consumer price index (CPI) averaged 8 percent. Over the past three years, inflation based on the CPI averaged 5 percent, while that based on the implicit GDP deflator was even lower at about 4 percent. Nonetheless, preliminary projections for 1995 indicate a slightly higher inflation rate measured both ways at around 6 percent. Since the world inflation rate as measured by the G-V Manufacture Unit Value index (MUV) has averaged about 3-4 percent over the past three years, the country is bound to lose some of its competitiveness on external markets if current developments are not corrected. 7. In spite of transitory difficulties, the balance of payments has improved significantly; the current account deficit declined steadily from 12 percent of GDP at the height of the economic crisis in 1983 to about 2 percent since the early 1990s. Recent developments in the external environment have been mixed. The worldwide recession, the increasing appreciation of the real exchange rate, and the drought- induced agricultural downturn worsened the trade balance in 1992-93 through their negative impact on exports and food import requirements; on the other hand, international prices for phosphate-based fertilizers have recovered slightly from their downward trend since the late 1980s. 8. For exports, the most severe impact was on manufactured goods which declined by 4 percent in the 1992-93 period, after steady growth averaging more than 10 percent per annum since 1985. Actions consisting of selected export promotion policies in terms of institutional support and tax concessions to exporters, combined with increased demand for clothing and agro-industrial products, supported this performance. In addition, phosphates-based products and tourism-related services grew steadily, resulting in an overall export growth averaging some 4 percent per year since the late 1970s. 9. On the imports side, capital-related imports decelerated due to cuts in the public investment program; overall imports grew at about 3.5 percent per annum over the 1980-93 period. Large 1/ Estimated by the Bank's International Economic Department (IECSE) based on the conversion method used for the World Bank Atlas. Annex I  - 4 - fluctuations in imports of food items and other consumer goods (around 20 percent of goods and non- factor services imports) are due to fluctuations in incomes which are, in turn, caused by external shocks and weather-induced drops in agricultural output. By way of example, the severe drought of 1992 resulted in close to 50 percent increase in imports of foodstuffs, mainly cereals. 10. Gains from external debt rescheduling ended in 1992; mobilizing much needed foreign capital inflows is one of the major challenges facing the authorities. Nonetheless, direct foreign investment flows have more than doubled since 1990 from US$227 million to US$522 million in 1993 (2 percent of GDP); they are estimated at US$776 million in 1994 (2.5 percent of GDP). A large proportion of FDI accrued directly to the state's Treasury as proceeds from the on-going privatization program concerning some 112 public enterprises. During the economic reform period of the 1980s, many foreign exchange restrictions were dismantled; this led to full current account convertibility in 1993. The dirham rate, which is pegged to a basket of currencieS2', has appreciated in the past three years since inflation has been higher than in partner countries. Morocco witnessed a sharp increase in its foreign exchange reserves over the past several years, reaching the equivalent of six months of imports of goods and nonfactor services at end- 1994; however, preliminary trends for 1995, if unchecked, could lead to a loss of a month's equivalent of foreign reserves. 11. Taking into account exceptional financing in terms of debt relief and cancellation, net long-term external debt disbursements to Morocco have declined constantly in the past two years, going from a yearly average of US$940 million in the 1980-85 period to about US$700 million over 1986-90, and to only US$80 million in the past two years; estimates of overallforeign net disbursements stood at negative US$253 million in 1994. This poor result in 1994 is partly explained by low gross disbursements from bilateral creditors and selected multilateral donors such as the Bank Group which managed to disburse less than US$300 million for the first time in ten years. However, net disbursements from all multilateral creditors have remained largely positive, averaging about US$300 million since the mid-1980s. From a high 113 percent of GDP on average in the 1984-87 period, the stock of external debt declined steadily relative to GDP to 75 percent in 1992, and the debt service effectively paid declined from 31 to 23 percent of exports over the same period; but with the end of debt relief, external debt indicators have slightly worsened since 1993. 12. During the economic reforms era, fiscal performance was hampered by repeated shocks such as droughts, adverse movements in international interest rates and losses in the external terms of trade; delays in preparing and implementing tax reforms occurred, as did minor overshooting in budgetary expenditures because of domestic political and social considerations. Nonetheless, the authorities consistently relied on stringent monetary and fiscal policies to reduce the budget deficit from an average of 12 percent of GDP in 1980-83 to 3.3 percent in 1993; however it is estimated that the fiscal deficit has worsened to 3.8 percent of GDP in 1994 due partly to various subsidies, price support mechanisms and transfer expenditures associated with the protected agriculture sector which experienced a boom that year. 2/ For more details on the most recent exchange arrangements and exchange restrictions, see the International Monetary Fund's Annual Report 1994, pp. 337-342. Annex I  13. The Government's fiscal effort paid off in the sense that, after averaging 19.7 percent of GDP over 1985-90, totalfiscal revenue increased to 23 percent of GDP on average in the 1991-93 period, with improved performance in the collection of both direct and indirect taxes. A track record for good fiscal performance has been established in the past several years. Alongside the introduction of a value-added tax in 1986 when the revenue to GDP ratio was only 18.8 percent, the authorities implemented a sweeping tax reform that effectively increased the tax buoyancy and decreased fiscal distortions. The continuous decline of the share of nontax revenue (transfers from OCP and other public monopolies) is an indication of the relatively successful implementation of the tax reform. However this decline is also partly due to the poor performance of some key public monopolies in recent years; in addition, an import surcharge (PFI) at rates ranging from 10 percent (products under the investment code) to 15 percent (normal rate) was introduced in 1988 to replace a special 5 percent import tax. Conversely, Morocco's central government revenue and expenditure when compared to those of I I developing countries show that the sample averages for budgetary revenue and expenditure were 21 and 20 percent of GDP, respectively, over the 1991-92 period; while the same ratios for Morocco were significantly higher at 26 and 25 percent of GDP, respectively3'. Nonetheless, the loss of 1.5 percent of GDP in direct tax revenues (mainly general income and corporate taxes' is worrisome. 14. Progress was also achieved in reducing total budgetary expenditures to 27percent of GDP in the 1988-92 period from an unsustainable level of 34 percent in 1980-83 which resulted from high public sector investments; however, the current level of expenditure is still 5-6 percentage points of GDP higher than the average for the sample of countries discussed above. The Central Government wage bill remained stable as a share of GDP at around 10 percent; however this ratio is high when compared to other middle income countries. There has been a modest increase in expenditures on goods and services which reached 3 percent of GDP in 1992 for the first time since 1988. Efforts to constraining the growth of external public debt, as well as mandatory placement of Treasury paper at below market rates has helped stabilize interest expenditures to about 6 percent of GDP over the past three years. Capital expenditures, including capital transfers to local collectivities (about half of total CG transfers over the past three years), have amounted to 6-7 percent of GDP since 1984. 15. In terms of treasury financing, both domestic and external sources offunds (including debt relief averaging 3-4 percent of GDP per annum since 1983) had been relied upon almost equally. Over the past 2-3 years however, external financing has declined in favor of greater financing from the domestic banking system and other special financial entities such as the CDG (Caisse de Dep6t et de Gestion). Having averaged 4-5 percent of GDP over the 1980-87 period, external sources of financing dropped to 2.3 percent in 1988-92, and turned negative at -1.1 percent of GDP in 1993-94. Substantial accumulation of domestic arrears amounting to some DH3.2 billion at end-1993 occurred; these reflect the lack of Treasury liquidity. Traditionally, treasury borrowing from the monetary system takes three major forms: i) auction market bond issues, ii) direct and indirect advances from the central bank( Al-Maghrib), and iii) mandatory placements with banks and other financial institutions. 3/ The Kingdom of Morocco, "Public Expenditure: Overview and Issues (Report No. 13413-MOR)", August 1994 (draft). Countries used in the sample include Thailand, Malaysia, Tunisia, Turkey, South Korea, Bolivia, Iran, Philippines, Uruguay, Hungary, and Romania. Annex I  - 6 - 16. The size of the consolidated public sector deficit, together with the way it is financed, constitute an important indicator for gauging the inflationary pressures, and crowding out of the private sector. To get an estimate of this, deficits for the local government and public enterprise sectors need to be estimated and added to the Treasury deficit. Several recent studies4' have estimated the consolidated public sector deficit at 4.5 percent of GDP in 1992; early indications for 1993 point to a higher consolidated public sector deficit of at least 5-6 percent of GDP, with the fiscal deficit for the budgetary Central Government reaching 3.3 percent; similar deterioration appears to have continued through 1994, with the CG deficit estimated at 3.8 percent of GDP. 17. At the peak of the economic crisis in the early 1980s, monetary policy was heavily constrained by the need to finance large fiscal deficits averaging 12 percent of GDP in the 1980-83 period. The sharp improvement in the fiscal position brought about by reforms since the late 1980s allowed the authorities to start implementing a more consistent set of monetary policies aimed at redirecting more financial resources to the private sector for low inflation growth, developing a modern financial sector, and re- establishing a sustainable balance of payments position. Broad money with an increasing share of quasi- money grew at 9.3 percent in 1992 and 8.7 percent in 1993, compared to an average of 15 percent per annum over the 1986-91 period, and private sector credit increased modestly. 18. As indicated earlier, prospects for 1995 are influenced by the anticipated drought that would cause agricultural value-added to decline by almost 37 percent (following the exceptionally good bumper crop in 1994); this will lead to yet another recession with overall GDP expected to decline by 5.1 percent. At 2.5 percent, growth in non-agricultural GDP is projected to be less than the previous year's performance; and both the fiscal and current account deficits are forecasted to be higher at 4.2 and 3.4 percent of GDP, respectively. C. Medium-Term Prospects - Lowcase Scenario 19. Key Assumptions and Results. The low case outlook assumes a 'business as usual approach by extrapolating recent growth performance; as illustrated by key economic indicators summarized in table 2, this scenario would not fulfill the country's medium-term development agenda. Under this low case outlook, it is assumed that further economic and financial reforms necessary to boost both the volume and efficiency of private sector investment will not be pursued; on-going privatization and private concessions will continue on a piecemeal basis but will not accelerate beyond the current pace and efficiency-enhancing foreign direct investment will lag. Furthermore, new fiscal measures needed to recoup potential revenue losses due to developments in international trade will not be undertaken, and finally restrictive expenditures policies necessary to reduce the deficit and put more resources in the hands of the private sector will not be implemented. All these would result in stagnant economic growth averaging 3-4 percent per year with even greater dependence on the agriculture sector, and an unsustainable fiscal deficit that would absorb most of the domestic resources. Faster growth in the manufacturing and service sectors would not materialize. 4/ The Kingdom of Morocco, "Issues and Prospects in the Public Sector (Report No. 10157-MOR)" of June 1992; and "Public Expenditure: Issues and Outlook (Report No. 13413-MOR)" of August 1994 (draft). Annex I  - 7 - Table 2. Key Macroecononmic Indicators Medium-Term Outlook and Resource Allocation Low Case Scenario 1995 1996-2000 2000-03 2004-05 Growth Rates(%) GDP at market prices' -5.1 4.2 3.8 3.8 - Agriculture -36.5 8.9 3.5 3.5 - Non-agriculture 2.5 3.4 3.8 3.9 Per-capita GDP -6.9 2.4 2.1 2.3 Domestic Inflation 7.0 5.7 6.2 7.0 Share of GDP at m.p. A. National Accounts Gross Investment 20.6 21.8 22.5 22.1 Fixed Investment 20.9 21.8 22.5 22.1 - Central Government 4.4 4.4 3.9 3.7 - Other Public Sector 3.7 4.0 4.1 4.2 - Private Sector 12.8 13.4 14.5 14.2 Five-Year ICOR 16.8 6.6 5.5 5.7 Domestic Savings 14.1 15.8 17.5 18.1 National Savings 17.2 18.4 19.4 19.3 - Cent. Government 2.5 2.8 3.0 3.2 - Non-central Government 14.7 15.5 16.4 16.1 Foreign Savings 3.4 3.5 3.1 2.8 B. Central Government Budget Current Revenue 23.6 22.9 22.9 22.9 Total Expenditure 28.4 27.5 27.4 27.2 Fiscal deficit, commit, basis -4.7 -4.7 -4.5 -4.3 - Privatization proceeds 1.2 0.4 0.0 0.0 - Domestic borrowing, net 3.4 3.5 4.2 4.4 - External borrowing, net 0.2 0.7 0.3 0.0 C. Balance of Payments Exports of GNFS 23.0 22.0 21.4 19.8 Imports of GNFS 29.5 28.0 26.4 23.8 Resource Balance -6.5 -6.0 -5.0 -4.0 Curmnt Act. deficit -3.4 -3.5 -3.1 -2.8 Foreign Direct Invest. 2.1 1.7 1.6 1.6 Reserves(months of imports) 5.4 4.6 3.2 1.6 D. Ext. Debt Burden Debt Outstanding & Disbursed 69.8 54.4 44.0 36.3 Debt ServiceZXGS 29.9 25.5 28.7 29.6 20. In addition, the unemployment problem will grow worse. Simulations of the impact of various growth scenarios on employment were conducted using recent projections of the active population and an estimated elasticity of labor absorption of 0.7. The annual increase in the rate of labor force participation was assumed at its historical average for the past decade, i.e. about half a percent. Based on the country's key demographic characteristics such as the age structure, the increase in industrial Annex I  activities and urbanization that impact on the labor market, and the rising labor market participation of women, it is expected that this sensitive parameter of overall labor participation rate will increase in the future. Therefore, another scenario using an increase of one quarter of a percent to 0.75 percent per year in the labor force participation rate was performed for each of the growth scenarios. 21. Under this low case scenario, results of the above simulations were as follows: The rate of urban unemployment increases steadily from 16 percent in 1994 to 21 percent by 1997-98, and then further to 22 percent from 2000 onwards. Results from the analysis based on higher labor force participation rate are more explosive, since they lead to an urban unemployment rate close to 23-24 percent of the active population by 2000. 22. The share of gross investment in GDP is projected to remain constant at about 22 percent over the projections, or only one percentage point above its current rate. However, a shift towards more productive private investment takes place at the margin and in line with current trends. The investment share of the overall public sector holds basically flat at 8-9 percent of GDP. As compared with 1993, the private sector would increase its share in total investment from 12 to 14 percent of GDP, or 1.7 times public sector investment. Failure to attract more efficiency enhancing private investment would lead to a relatively unproductive capital stock; the marginal efficiency of capital as measured by the ICOR is projected at its historical average of 5-6 units of capital per unit of output. 23. In spite of the steady decline in population growth, moderate gains in income due to poor investment perfonnance would limit increases in per-capita private consumption to less than 3 percent per year on average over the next ten years. In line with past trends, domestic savings would increase slightly to about 17 percent of GDP by the end of the projections after the current slump. At 18 percent of GDP, national savings would perform a little better in light of the moderate inflows of private transfers of Moroccans living abroad but would fail to reach its ten year average of 20.6 percent of GDP in the 1984-1993 period. This low level of savings would constrain economic growth prospects. 24. As the authorities are faced with increasing fiscal disequilibria, deficit financing would become problematic. First, printing money could cause the domestic inflation rate to increase beyond its current rate of 4-5 percent per annum. Such an inflation rate would likely be higher than that of Morocco's main trading partners; thus, in conjunction with worsening fiscal and foreign reserves positions, this development would threaten the stability of the real exchange rate that has helped the country remain competitive in the past. Second, high domestic borrowing and accumulation of arrears would crowd out the private sector and depress growth prospects further. Finally, borrowing externally would increase the debt burden which absorbs more than a third of foreign earnings already. 25. As illustrated over the last two years, Morocco's balance of payments remains vulnerable to the external environment, as well as to the weather pattern through its ill-effect on the trade balanceforfood items. Among adverse shocks, external terms of trade losses could be the most severe, although the growing diversity of exports has reduced their potential impact somewhat. Over 60 percent of the country's merchandise trade is with the European Union which is also the source of most tourism receipts and foreign investment flows; no changes would be foreseen under this scenario. Trade flows with other Maghreb countries, despite the creation of the Arab Maghreb Union (UMA), remain negligible; this trend would also continue. Annex I  - 9 - 26. Both imports and exports are projected to achieve their historical growth performance at 4-S percent on average over the projections period, but with declining shares of GDP; merchandise trade would remain structurally stable and the resource balance would worsen marginally. For imports, the share of food and other consumer goods would continue to claim a high 22 percent while for exports, the share of agricultural and phosphate related products would amount to some 60 percent of merchandise exports by the turn of the decade. To the extent that consumption can be constricted, deficit in the trade balance would stabilize at around 10 percent of GDP, and tourism related services would also help contain the deterioration of the resource balance at 5-6 percent of GDP. Sluggish but steady inflows of private transfers notwithstanding, the current account deficit is projected to worsen in line with recent trends; it would average 3-4 percent of GDP over the medium-term. 27. In the capital account, the most promising element has been the steady increase in foreign direct investment (FDI) inflows over the past five years. Under conservative assumptions, FDI flows are expected to grow at 3-4 percent per annum for the next five years, reaching the level of US$800 million by the year 2000 (1.7 percent of GDP). Gross public and publicly-guaranteed debt disbursements are projected to more than double their current level of about US$2 billion, with an increasing share of official bilateral creditors. The external debt burden would remain high, claiming more than a third of export earnings. In the absence of sizeable external grants or exceptional financing commitments, forecasted imports and consumption needs would not be fulfilled without recourse to foreign reserves and further external borrowing; any major shocks such as drought or large terms of trade losses would weaken the external position further. External reserves are projected to decline drastically, averaging merely 1-2 months of imports of goods and nonfactor services in the medium-term. 28. Under the low case scenario, a steady deterioration of the fiscal deficit is projected over the medium-term, reaching 4-5 percent of GDP by the end of the projections. Without any major additional fiscal reforms, it is assumed that the authorities would merely build upon earlier policies that led to the recent success in terms of buoyancy of the tax system. On the one hand, mere collection efforts would not suffice to maintain revenue relative to GDP at its current level; revenues are projected to drop at 22- 23 percent, i.e. a steep decline since 1993 due to steady decline in tax rates and failure to broaden the tax base to compensate for the decline in trade taxes associated with the implementation of a Free Trade Accord (FTA) with the EU. On the other hand, the attempt to curtail expenditures would encounter implementation difficulties, leading to a flat expenditures to GDP ratio of 28-29 percent on average. Over the medium-term, the country would witness an increasingly hard to finance fiscal deficit, with the non-interest primary surplus of less than 1 percent of GDP. With the piecemeal but necessary liberalization of domestic interest rates, the total interest burden on the budget would increase from 20 percent of total expenditures in 1993 to 26 percent by the end of the projections, with an increasing share of over 70 percent on account of domestic debt. 29. On the expenditure side, the wage bill (IO percent of GDP in 1993) represents about half of total current expenditures and transfers. Results from the recent public expenditure report indicate that the above-average level of current expenditures relative to GDP in Morocco, as compared to other middle income countries, is due to expansionary personnel policy rather than high unit pay among its civil servants. Under the "business as usual" approach, wage expenditures are assumed to increase with nominal GDP, i.e. 34 percent in real terms. Moreover, fiscal measures aimed at gradually phasing out most consumer subsidies and current transfers, and allocating more resources to other non-salary recurrent expenditures will not be implemented; thus budgetary savings amounts to only 1-2 percent of Annex I  - 10 - GDP per year over the projections. Capital expenditures would remain flat or increase marginally in relation to GDP as both more relatively inefficient CG investment and capital transfers to public enterprises continue to be made. Other extra-budgetary capital expenditures, comprising VAT transfers to local governments and the balance of other Treasury special accounts and annex budgets would most likely be curtailed for lack of budgetary resources. 30. 7he recent trend towards increased domesticfinancing of the treasury deficit is likely to continue over the medium-term; despite the expected steady flows of foreign disbursements, a larger share is accruing to the non-government public sector, leading to a declining net foreign disbursement to the budget; in addition, the authorities would face difficulties in developing more attractive treasury borrowing instruments for the domestic nonbank. Such developments make the deepening of the domestic financial system more difficult. As for the consolidated public sector, the continuation of recent patterns point to an added deficit on the order of 3-4 percent of GDP for the other public sector. Combining this with the deficit of the CG at 4-5 percent of GDP, the overall deficit for the public sector could rise to an unsustainable level of 8-9 percent of GDP in the medium-term (see para. 47). Financial repression and crowding out of the private sector would occur, depressing growth prospects further. D. Medium-Term Prospects - High Growth Scenario 31. Introduction. Despite relatively good economic andfinancial management over the past decade, Morocco could do substantially better in terms of achieving higher economic growth, and reducing unemployment and poverty. Notwithstanding social progress in areas of education and health, per-capita income growth has averaged merely 2 percent per year since the mid-1980s, and urban unemployment is increasing. Conversely, recent developments including a thriving private sector, relatively stable macro-fundamentals, and keen interest among foreign investors indicate that the potential exists for rapid economic growth of the type experienced by the East Asian economies. 32. Key Assumptions. Underlying a dynamic and sustained growth scenario is the thorough implementation of a complete package of reforms (see main report) which will lead to increases in the efficiency and volume of investment. Only under such a high growth scenario, summarized in table 3 below, can widespread reductions in poverty and unemployment be achieved. Some elements of this scenario are increasing outward orientation and closer integration with world markets, higher technologically-rich foreign direct investment flows, a substantial quality improvement in the country's physical infrastructure, social and human capital, and its ability to attract and quickly adopt the foreign technology required by a private sector-led economy and accelerated financial sector deepening; all of these elements will raise the productivity of investment. 33. Under the high growth scenario, results of the simulations of the impact of growth on urban unemployment rate are as follows. The rate of urban unemployment first increases from 16 percent in 1994 to 18 percent by 1996-97 due essentially to the forecasted drought in 1995, before declining to perhaps 15 percent by 2000 and several points more by 2005. If labor force participation rates increase as before (para. 20), unemployment would still decline several points by 2005. Annex I  - 11 - Table 3. Key Macroeconomic Indicators Medium-Term Outlook and Resource Allocation High Case Scenario 1995 1996-2000 2000-03 2004-05 Growth Rates(%) GDP at market prices -5.1 5.2 6.1 7.5 - Agriculture -36.5 9.1 4.0 4.0 - Non-agriculture 2.5 4.5 6.6 8.2 Per-capita GDP -6.9 3.3 4.4 5.9 Domestic Inflation 5.6 2.6 2.0 2.0 Share of GDP at m.p. A. National Accounts Gross Investment 21.6 24.5 28.3 31.8 Fixed Investment 21.9 24.5 28.3 31.8 - Cent. Government 5.0 5.4 5.5 5.6 - Other Public Sector 3.1 3.0 2.5 2.3 - Private Sector 13.8 16.1 20.3 23.9 Five-Year ICOR 16.8 6.0 4.2 3.8 Domestic Savings 15.0 18.1 23.2 28.0 National Savings 18.3 22.2 27.1 30.9 - Cent. Government 3.0 5.3 7.7 9.6 - Non-central Government 15.2 17.0 19.5 21.3 Foreign Savings 3.4 2.3 1.2 0.9 B. Central Government Budget Current Revenue 24.4 23.6 24.2 24.1 Total Expenditure 28.6 25.8 24.3 23.6 Fiscal deficit, commit. basis -4.2 -2.2 -0.1 0.6 - Privatization proceeds 1.2 1.0 0.0 0.0 - Domestic borrowing, net 2.9 1.1 0.7 0.3 - External borrowing, net 0.1 0.2 -0.6 -0.8 C. Balance of Payments Exports of GNFS 24.2 27.2 31.9 34.2 Imports of GNFS 30.9 33.7 37.1 38.0 Resource Balance -6.7 -6.5 -5.1 -3.8 Current Act. deficit -3.4 -2.3 -1.2 -0.9 Foreign Direct Invest. 2.7 2.7 3.0 3.2 Reserves(months of imports) 5.6 5.1 5.1 5.1 D. Ext. Debt Burden Debt Outstanding & Disbursed 69.8 58.6 46.6 37.2 Debt Service/XGS 29.9 23.6 21.2 18.2 34. Under this dynamic outlook, Morocco would deepen macroeconomic reforms to promote steady growth in efficient private sector investment from 12 percent of GDP in 1993 to 24 percent in the next decade to reach over three times the level of total public sector investment. This would boost overall investment to over 31 percent of GDP with FDI flows reaching 3 percent of GDP by the turn of the decade; comparative economic growth and investment indicators over the 1970-92 period for the fast Annex I  - 12 - growing East Asian countries (except Taiwan) are summarized in tables 4.a and 4.b below. Moreover, foreign expertise and the embodied technological content in machinery and equipment would improve the marginal efficiency of capital with a five-year ICOR falling to under 4 units of capital per unit of output over the projections.5' 35. As a share of GDP, para-public sector investment would be cut by a third while complementary CG investment in core human resource development (health and educational supplies), and physical infrastructures (rural feeder roads, rural dispensaries, and low-income housing would increase from about 5 percent of GDP in 1995 to 5.6 percent, maintaining consolidated public sector investment at about 8 percent. 36. In addition, sources of growth would be diversified away from the drought-plagued agriculture sector and into light industry and tourism-related services where productivity has declined in recent years. With an average growth rate of for agriculture of about 9.1 percent p.a during 1996-2000 and 4 percent thereafter, the primary sector's share in total GDP would decline from 20 to 15 percent over the projections, while that of manufacturing would gain 6-7 points to 23 percent. The services sector would also decline somewhat in importance relative to the manufacturing sector to remain just below 50 percent of total value-added. 37. However, the gradual increase offorecasted GDP growth is based on international evidence of an investment-growth lag (illustrated in tables 4. a and 4. b) that requires, inter-alia, a rising investment rate well before growth follows. Before their strong growth performance in the 1980s, the selected fast growing East Asian economies achieved and sustained an investment to GDP ratio of 28-29 percent on average over the entire 1970-80 period, as compared to Morocco's 20 percent over the same period. Furthermore, efficient private sector investment was close to three times that of the public sector in those East Asian countries, in comparison to the case of Morocco where inefficient public sector investment has been twice the level of private investment. 38. High investment rates require higher savings rates; empirically, high savings rates of more than 30 percent of GDP on average have characterized fast growing economies of East Asia such as Malaysia, T7hailand, Hong Kong and South Korea for the past ten years, while negligible and declining saving rates have persisted in Africa, depriving that region of required investments in socio-infrastructures. Table 5 below summarizes savings, and capital efficiency indicators for Japan and selected middle income countries, including the fast growing East Asian economies. With rising wealth and incomes, as well as increasing private and public savings, private consumption would gradually increase, with growth in consumption averaging 4-5 percent in per-capita terms per annum in the medium term. 5/ This level of capital efficiency is in line with the performance of the following six fast growing East Asian economies over the 1980-90 period: Japan, Hong Kong, South Korea, Malaysia, Singapore and Thailand. Annex I  - 13 - Table 4.a Investment to GDP Ratios For Selected Comparator Countries, 197092 Period Averages Yearly 1970-75 1975-80 1980-85 1985-90 1991 1992 Chile 18.1 18.4 16.5 20.2 21.7 23.7 O/w: Private 7.3 10.3 9.8 11.2 15.5 16.8 Private/Public 0.7 1.3 1.5 1.2 2.5 2.4 Hong Kong 23.0 27.0 28.3 25.1 27.4 27.8 Indonesia 19.9 21.9 23.6 25.6 28.4 27.3 O/w: Private 11.2 12.3 12.6 11.9 13.4 12.4 Private/Public 1.3 1.3 1.1 0.9 0.9 0.8 Japan 34.6 31.2 29.1 29.4 31.8 30.8 Korea, Rep. of 23.5 28.6 29.1 30.6 38.4 36.6 O/w: Private 17.2 22.1 22.2 23.7 28.5 25.0 Private/Public 2.7 3.4 3.2 3.4 2.9 2.2 Malaysia 23.7 26.1 33.6 27.5 35.6 34.3 Olw: Private 16.0 16.3 18.0 16.5 24.1 23.4 Private/Public 2.1 1.6 1.2 1.5 2.1 2.1 Mexico 19.5 22.0 21.4 18.8 19.4 20.8 O/w: Private 12.6 12.9 12.6 13.4 14.9 16.9 Private/Public 1.8 1.4 1.4 2.5 3.3 4.3 Morocco 16.1 26.3 24.4 21.9 22.3 23.0 Olw: Private 5.7 12.4 12.0 12.6 13.2 13.1 Private/Public 0.5 0.9 1.0 1.3 1.5 1.3 Philippines 19.2 26.0 25.7 18.8 20.1 21.8 Olw: Private 14.4 18.6 18.7 15.2 14.9 17.5 Private/Public 3.0 2.5 2.7 4.2 2.9 4.1 Singapore 36.0 36.5 44.9 37.1 39.4 40.4 O/w: Private 28.2 26.0 31.4 26.6 31.3 32.5 Private/Public 3.6 2.5 2.3 2.5 3.9 4.1 23.1 25.1 27.8 31.0 41.2 39.1 Thailand 17.3 17.8 19.4 24.6 34.0 31.6 O/w: Private 3.0 2.5 2.3 3.9 4.7 4.2 Private/Public Memo items: 1/ 27.3 29.1 32.1 30.1 35.6 34.8 HG Sample Average 16.1 26.3 24.4 21.9 22.3 23.0 Morocco 1/ The high-growth countries (HG) sample includes: Japan, Hong Kong, S. Korea, Malaysia, Singapore, and Thailand; Source: World Bank Economic and Social Database. Annex I  - 14 - Table 4.b GDP and Exports Growth For Selected Comparator Countries, 1970-92 Period Averages Yearly 1970-75 1975-80 1980-85 1985-90 1991 1992 GDP Growth (% p.a.) Chile -1.3 7.3 0.9 6.5 6.1 10.3 Hong Kong 6.6 12.0 5.6 7.9 4.2 5.0 Indonesia 6.9 7.5 5.0 6.2 6.7 6.3 Japan 4.5 4.6 3.7 4.5 4.3 1.2 Korea, Rep. of 9.5 7.7 8.5 17.2 9.1 5.1 Malaysia 7.1 8.5 5.1 6.8 8.7 7.8 Mexico 6.5 6.6 2.0 1.4 3.6 3.0 Morocco 4.7 6.5 3.3 4.3 6.2 -4.4 Philippines 5.7 6.0 -1.3 4.7 -0.5 0.1 Singapore 9.4 8.5 6.1 7.9 6.7 5.8 Thailand 5.8 7.8 5.4 10.4 8.1 7.6 Memo items: 1/ HG Sample Average 7.1 8.2 5.7 9.1 6.8 5.4 Morocco 4.7 6.5 3.3 4.3 6.2 -4.4 Exports Growth (% pa) Chile 5.6 15.1 1.8 10.6 10.0 16.7 Hong Kong 3.8 15.2 10.0 17.3 14.9 18.1 Indonesia 11.6 6.9 -5.7 8.2 24.2 16.4 Japan 9.2 9.7 7.6 3.6 5.2 5.2 Korea, Rep. of 24.4 16.4 10.1 13.9 11.8 11.0 Malaysia 5.9 9.8 7.1 14.7 15.3 5.0 Mexico 4.7 12.2 9.3 5.3 5.4 0.8 Morocco 1.6 5.2 4.1 7.4 -2.8 0.9 Philippines 4.4 15.2 -2.4 9.9 5.8 3.9 Singapore --- 16.2 5.3 15.4 9.0 5.7 Thailand 6.2 13.0 8.1 19.3 15.2 12.0 Memo items: 1/ HG Sample Average --- 13.4 8.0 14.0 11.9 9.5 Morocco 1.6 5.2 4.1 7.4 -2.8 0.9 1/ The high-growth countries (HG) sample includes: Japan, Hong Kong, S. Korea, Malaysia, Singapore, and Thailand; Source: World Bank Economic and Social Database. Annex I  - 15 - Table 5. Savings to GDP Ratios and ICORs For Selected Comparator Countries, 1970-92 Period Averages Yearly 1970-75 1975-80 1980-85 1985-90 1991 1992 National Savings/GDP: Chile 12.8 15.7 7.2 17.1 21.4 21.3 Hong Kong 25.9 29.7 29.1 32.6 31.0 30.2 Indonesia 20.3 26.9 27.0 28.5 30.8 36.5 Japan 37.2 32.2 31.1 33.0 35.0 34.5 Korea, Rep. of 18.3 24.1 24.2 34.4 35.7 35.0 Malaysia 22.7 28.4 26.3 28.5 26.5 30.5 Mexico 16.8 18.5 21.9 19.0 17.6 15.5 Morocco 17.5 15.6 16.3 20.9 20.4 20.5 Philippines 23.5 26.9 21.2 18.5 21.4 20.8 Singapore 23.7 32.4 40.3 42.4 46.9 47.6 Thailand 23.0 21.7 23.0 28.4 29.0 33.3 Memo items: 1/ HG Sample Average 25.1 28.1 29.0 33.2 34.0 30.2 Morocco 17.5 15.6 16.3 20.9 20.4 20.5 Period Averages (10-Year ICOR) 1970-80 1980-90 Chile 6.5 5.0 Hong Kong 2.9 4.1 Indonesia 0.4 4.5 Japan 6.8 6.7 Korea, Rep. of 3.0 2.1 Malaysia 3.1 4.9 Mexico 3.7 13.0 Morocco 4.4 5.8 Philippines 3.3 12.6 Singapore 4.1 5.7 Thailand 4.1 3.4 Memo items: 1/ HG Sample Average 4.0 4.5 Morocco 4.4 5.8 1/ The high-growth countries (HG) aample includes: Japan, Hong Kong, S. Korea, Malaysia, Singapore, and Thailand; Source: World Bank Economic and Social Database. 39. A sound domestic financial system will be instrumental in mobilizing domestic savings expected to reach more than a quarter of GDP. At about 31 percent of GDP, national savings would be almost sufficient to finance the needed investment. Relative to GDP, this level of savings would be around the sample average for the six fast growing Asian economies over the 1980-90 period; thus the country would achieve its growth targets with a little higher external indebtedness; the external debt stock relative to Annex I  - 16 - GDP is projected to decline to under 40 percent, a sharp decline from its current level of 70 percent; and debt service relative to exports is forecasted to decline to under 20 percent, a reduction of more than ten percentage points over the projections. 40. Under the high case scenario, increased national savings would be explained by relatively higher domestic savings, a lower external debt burden, and moderate private transfers from abroad. The reason for the latter is that relatively younger and single workers with lower propensity to save and less close family links (hence lower transfers) would be emigrating abroad; in addition, external debt indicators improves significantly although better creditworthiness will result in relatively more private financing on commercial terms. 41. The pursuit of sound monetary and credit policies to accommodate economic growth and overall macroeconomic stability would maintain domestic inflation at around the international rate of 2-3 percent per year. 42. Morocco is assumed to liberalize its trade regime further and to increase labor and capital mobility across all sectors through an efficient incentive and regulatory framework, its efforts in developing new external and regional markets will be vigorously pursued. The share of agricultural products in exports is projected to decrease gradually in favor of light industry and tourism-related services. Growth in manufacturing exports is expected to average 10 percent per annum in the medium- term, boosting its share in merchandise exports to over 40 percent by the end of the projections. Overall, the trade balance would deteriorate marginally due to high import requirements for capital and non-food consumer goods; but the current account deficit to GDP ratio would improve substantially reaching near balance at the end of the projections. Without major droughts, food imports would decline to around 10 percent of merchandise imports, while the share of intermediate and investment goods would increase sharply. The tourism sector is expected to remain competitive. 43. Despite the sharp rise in the level offoreign direct investmnent flows (FDI) over the past decade, their proportion to GDP stood at 2 percent in 1993, compared to 3-4 percent for 7hailand and Makaysia in the pastfive years. Under an accelerated privatization program of core public enterprises, increased private concessions and prudent macroeconomic policies, FDI flows and other portfolio investments would increase further and reach about US$2 billion per year (i.e. 3 percent of GDP) by the turn of this decade. Renewed private creditors' interest and non-publicly guaranteed commitments would push gross external debt disbursements to around US$2-3 billion per year. As the proportion of borrowing on commercial terms increases over time, however, so will the interest burden in the medium-term. 44. With the basic thrust of ensuring that the private sector can gradually expand, the overallfiscal policy of low and predictable budget deficits will be pursued vigorously under this high growth case; only such macroeconomic policies are compatible with expansionary private sector investment and credit, low domestic inflation, and a viable balance ofpayments. Among other things, additional revenue-enhancing fiscal reforms (perhaps into protected sectors such as agriculture and land--politically sensitive areas) would be undertaken in light of the continuous erosion of the fiscal base due to further trade liberalization, financial sector reforms, and the hard-to-tax informal sector. Implementing such policies would help maintain the baseline revenue to GDP ratio in the medium-term by making up for potential revenue losses from further trade liberalization, and more importantly by increasing the formal sector's proportion in economic activity. In the long-run, revenues are projected to average 24 percent of GDP, Annex I  - 17 - i.e. roughly at their current share. Nonetheless, this average revenue to GDP ratio would remain above that of Thailand [19 percent], Turkey [20 percent], the Philippines [18 percent], and South Korea [17 percent] over the 1991-92 period; but below that of Malaysia [25 percent]. Thus, further progress in expenditure reallocation and reduction should be pursued to allow corresponding relief of the overall tax burden; such policy would preserve the fiscal balance and put more resources into private hands. 45. Concerning expenditures, a comprehensive wage policy is the core ingredient for a sustainable overall expenditure policy; and a more radical approach of selectivity, disengagement and greater cost effectiveness and recovery is needed to support a thriving private sector. This approach will necessarily include a more restrictive civil service personnel policy given the weight of the wage bill in recurrent expenditures; it is therefore assumed that wage expenditures would increase with domestic inflation, i.e. to remain constant in real terms. Overall, additional investment needs in basic physical and social infrastructures and costs of fiscal reforms are expected to be undertaken, leading to a more gradual reduction in the budget deficit. By the turn of the decade, a balanced CG budget and a small surplus towards the end of the projections will demonstrate the Government's commitment to economic and financial reforms. Furthermore, a comprehensive privatization program coupled with more management and financial responsibilities of local governments will lead to a negligible overall public sector deficit of one percent of GDP or less by the end of the projection period. 46. The sustainability offiscal policy is a key issue when assessing the effectiveness ofmacroeconomic policies in support of higher growth; a given policy is sustainable if the government can continue to pursue it indefinitely. Applying strictly the above criterion to total public debt (external and domestic) involves analysis of the time path of the debt stock resulting from a specific fiscal policy. But before proceeding, it is equally important to assess constraints on the economy's resources imposed by the initial debt burden. The rationale for a low budget deficit for Morocco is that resources must be released to support an expansion of private investment, which in turn is needed to promote higher growth. Morocco's total public debt currently stands at about 100 percent of GNP with 70 percent owed to external creditors at end-1994; in per-capita terms, external debt alone amounted to 80 percent of GNP over the 1991-93 period, and its service claimed 31 percent of foreign earnings in 1993. By contrast, Tunisia's external debt accounted for 60 percent of GNP in per-capita terms over 1991-93, and claimed only 20 percent of export earnings in 1993; as for Thailand, external debt stock reached 38 percent of GNP per-capita, and its service claimed a mere 19 percent of export earnings in 1993 (see table 6 below for more comparator country data). Thus Morocco remains a highly indebted country and should design and implement a public deficit-cum-debt strategy aimed at reducing sharply its public debt burden in order to enhance the country's economic growth prospects and creditworthiness. Some elements of such a strategy include low public sector deficits and a stronger program of public enterprise divestitures which would generate additional budgetary revenues to retire public debt. 47. When addressing the sustainability issue, one criterion is that the deficit profile should not entail an increase in either the stock of public debt relative to GDP or the stock of (high-powered) money as a ratio of GDP6". A simple "back of the envelope" calculation can be done on the basis that only four ways exist to finance the budget deficit: i) extension of free central bank loans to the Government 6/ See Fisher and Easterly, "The Economics of the Government Budget Constraint", The World Bank Research Observer, July 1990. Annex I  - 18 - (seignorage); ii) net borrowing from abroad; iii) running down foreign reserves; and iv) borrowing domestically from commercial banks and the private sector. Table 6. External Debt Indicators For Selected Comparator Countries, 1970-93 Period Averages Yearly 1970-80 1980-90 1991 1992 1993 Per-Capita GNP (US$) Chile 1,279 1,810 2,400 2,893 Indonesia 262 516 609 679 Korea, Republic of 857 2,967 6,752 7,003 Mexico 1,516 2,355 3,358 3,763 Morocco 550 764 1,039 1,044 Philippines 387 637 731 840 Thailand 406 921 1,613 1,879 Tunisia 759 1,212 1,529 1,809 Debt Stock/GNP (%) Chile 46.9 91.2 55.9 48.6 49.0 Indonesia 38.1 47.7 68.9 66.9 65.9 Korea, Republic of 38.2 38.6 13.6 14.4 14.4 Mexico 24.8 55.7 41.2 35.5 35.5 Morocco 33.3 98.6 81.0 78.9 81.7 Philippines 37.3 74.9 70.6 59.4 63.7 Thailand 16.8 36.0 39.0 36.4 37.6 Tunisia 38.0 59.2 65.5 56.1 59.9 Debt Service/Exports (%) Chile 35.1 35.5 23.2 21.1 23.4 Indonesia 17.0 26.9 32.0 30.6 32.6 Korea, Republic of 15.2 20.9 7.1 7.6 9.2 Mexico 42.0 42.0 24.2 34.4 32.7 Morocco 15.0 32.3 25.2 23.7 30.8 Philippines 21.2 32.5 23.1 24.5 24.9 Thailand 14.2 22.4 13.0 14.1 18.6 Tunisia 11.5 21.5 23.9 20.1 20.2 Debt Per-Capita (USS) Chile 581.2 1,514.1 1,340.7 1,407.0 Indonesia 91.9 241.1 419.7 454.5 Korea, Republic of 333.9 937.9 918.3 1,011.3 Mexico 400.6 1,233.1 1,384.8 1,334.9 Morocco 210.0 734.8 840.9 824.6 Philippines 156.6 469.3 516.2 499.6 Thailand 74.6 329.4 629.1 683.1 Tunisia 300.0 718.0 1,001.6 1,014.8 Sources: World Bank Databases (World Debt Tables and Social Indicators of Development, 1994). Annex I  - 19 - 48. In simplified terms, a given fiscal policy is sustainable if the total debt stock (relative to GDP) of the public sector does not increase over time. The starting point of the analysis is the primary deficitlsurplus"' which defines the availability of resources to service existing debt without new commitments. Next, there are two contributing factors that work in opposite directions: i) non- inflationary seignorage earnings tend to lower the debt stock, ceteris paribus and ii) the excess cost of borrowing (average real interest rate) over real growth in the economy increases the debt stock; an adjusted debt stock is obtained by accounting for the net effect of these two factors. Finally, a declining adjusted debt stock over time indicates that the fiscal situation is sustainable, while an increasing stock indicates that the fiscal situation is unsustainable. Estimated results and projections for Morocco under both scenarios are as follows. Under the low case scenario, the relatively high stock of public debt grows steadily by 1-2 percent per year to reach about 110 percent of GDP by the end of the projection period; thus the underlying fiscal policy is unsustainable. Under the high case scenario, on the other hand, the stock of public debt declines constantly by 5-6 percent per year on average throughout the projections to around 60 percent of GDP by 2005, with external debt service claiming about 18 percent of export earnings. 7/ Defined as non-transitory revenue (i.e. excluding proceeds from privatization, tax amnesty, etc..) less non-interest expenditures. Annex I  - 20 - BIBLIOGRAPHICAL REFERENCES CERED - Centre d'Etudes et de Recherches Demographiques, "Fecondite, Infecondite et Nouvelles Tendances Demographiques au Maroc", Direction de la Statistique, Ministere des Affaires Economiques et Sociales, Royaume du Maroc, 1993. Fischer, S. and W. Easterly, "The Economics of the Government Budget Constraint", The World Bank Research Observer, July 1990. International Monetary Fund, Morocco - A Review of Adjustment Experience, IMF Mimeo, June 1994. Miller, R. and M. Sumlinski, "Trends in Private Investment in Developing Countries 1994: Statistics for 1970-92", IFC Discussion Paper No. 20, February 1994. Parker, K. and S. Kastner, A Framework for Assessing Fiscal Sustainability and External Viability, with an Application to India, IMF Working Paper No. WP/93/78, October 1993. Schmidt-Hebbel, K., L. Serven and A. Solimano, Savings, Investment, and Growth in Developing Countries: An Overview, World Bank Policy Research Working Paper No. 1382, November 1994. Spaventa, L., The Growth of Public Debt: Sustainability, Fiscal Rules, and Monetary Rules, IMF Working Paper No. WP/86/8, October 1986. World Bank, Kingdom of Morocco: Issues and Prospects in the Public Sector, Gray Cover Report No. 10157-MOR, June 1992. World Bank, "The East Asian Miracle: Economic Growth and Public Policy", World Bank Policy Research Series, Oxford University Press, September 1993. World Bank, World Development Report 1994 - Infrastructure for Development, Oxford University Press, June 1994. World Bank, Kingdom of Morocco: Preparing for the 21st Century - Strengthening the Private Sector in Morocco, Gray Cover Report No. 11894-MOR, July 1994. World Bank, Kingdom of Morocco: Poverty, Adjustment, and Growth, Gray Cover Report No. 11918- MOR, January 1994. World Bank, Kingdom of Morocco: Public Expenditure - Issues and Outlook, Green Cover Report No. 13413-MOR, August 1994. Annex I  ANNEX II INTERNATIONAL TRADE 1. This annex develops some of the issues discussed in Chapter 2 of the text in greater detail, and provides some additional information on the principal topics covered in the main report. A. Additional Notes on the Uruguay Round (UR) 2. As a result of the dismantling of the MFA, preliminary estimates indicate large gains for the competitive and highly constrained exporters in China, South Asia, and ASEAN while less constrained exporters such as Latin American lose. Large welfare gains accrue to the EU and North America who no longer pay quota rents to foreigners." Since the MFA functions like a voluntary export restraint, all countries exporting to the EU suffer a terms of trade loss. However, for the highly restrained exporters (ASEAN, South Asia, China), this loss is offset by increased volume (see Table 1). Since Morocco previously faced few constraints in exporting to the EU, there is no compensating volume increase to offset the terms of trade loss. Its loss of quota rents is increasing in the magnitude of the price decline and the elasticity of Moroccan export supply. 3. While minimum import prices are ostensibly illegal under Article 4 of the Uruguay Round-GATT agreement, exceptions are allowed under the Special Safeguard Provisions in Article 5. Not only does the replacement of reference prices (in the EU offer to the GAIT) by higher and more binding minimum entry prices have consequences for Morocco, but the timing of these price restrictions is also critical. For tomatoes for example, while proposed entry prices are lower than previous reference prices during some periods, these are largely during those months where Morocco has few exports (April through September). By contrast, whereas there was previously no reference price imposed during Morocco's heavy export period (from end December to end March) there is now a binding entry price. 4. Outstanding import quota restrictions for MFA-restricted goods, will be expanded by the prevailing bilaterally negotiated quota growth rates plus 16 percent in the first three years, by 25% in the subsequent four years, and by 27 percent in the final three years. 5. The UR secures bindings on agricultural tariffs and average tariff cuts of 36 percent over six years for industrial countries and 24 percent over 10 years for developing countries. Other provisions include reductions in domestic and export support and minimum access requirements. 6. As mentioned in the main report, the Moroccans estimate that under the original terms of the EU offer to the GATT, they would lose approximately 90 percent of their tomatoes and clementines exports and 40 percent of their orange exports.' This estimate was based on a comparison of the proposed minimum import prices with the internal Moroccan selling prices in the EU. For periods where the latter 1/ See Youngzheng, W. Martin and K. Yanagishima, 1994, 'Evaluating the Benefits of Abolishing the MFA in the Uruguay Round Package". 2/ See Nouvel Accord Maroc-Union Europ&enne: Mbmorandun Marocain, 14/2/94. Annex II  - 2 - is lower, Moroccan exports are assumed to be priced out of the market and zero. Other things equal, the method of analysis used will tend to overstate the impact because the internal selling price would have been higher. In any given period, exports would only go to zero if the minimum import price were set above the counterfactual internal price with EU supply alone (the autarkic price). Table 1: EU Textile and Clothing Quota Enlargement by Growth Factors (Tons) Quota Initial increase Increase with growth factors Country 1994 2004/1994 2004/1994 Argenina 31,486 56.8% 93.1% Brazil 114,098 38.4% 61.1% China* 255,093 42.4% 42.4% Hong Kong 152,963 13.5% 20.5% India 176,625 37.0% 58.8% Indonesia 65,976 54.0% 88.1% Korea 127,010 32.8% 51.8% Macau 24,005 16.2% 24.7% Pakistan 142,795 54.3% 88.6% Peru 15,490 74.1% 140.9% Sri Lanka 9,221 102.3% 204.2% Malaysia 35,973 47.3% 76.4% Philippines 24,711 67.0% 111.5% Singapore 19,205 47.3% 76.4% Thailand 74,609 46.0% 74.2% Total 1,269,260 40.4% 60.2% C China does not benefit from expanded quota growth because it is not a GATT member. Source: International Textiles and Clothing Bureau (ITCB). B. Notes on Morocco's Trade Regime 7. Table 2 presents tariff, and PFI rates. The PFI is intended as a revenue collection rather than an protectionist device. However evaluated solely from the standpoint of revenue generation, the PFI should not tax goods uniformly but should tax goods with inelastic demand more heavily. Table 3 reports overall nominal protection levels from 1990-93. Table 4 shows tariff and PFI rates by sector for 1993. Import weighted tariffs are a poor indicator of (even nominal) protection levels. Low import weighted averages could imply tariff levels that are so high that imports were significantly restricted. Production weighted tariffs would be a better measure, but are difficult to calculate for Morocco in light of data availability and the lack of a clear correspondence between trade and production data. 8. Year to year comparisons of total protection rates are problematic because average PFI rates are unavailable for previous years. Annex II  - 3 - Table 2: Total Rates at 15% PFI, 1994 (Percent) Ta&ffP PFI* Total 0.00 15.00 15.00 2.50 15.00 17.50 7.50 15.00 22.50 12.50 15.00 27.50 17.50 15.00 32.50 25.00 15.00 40.00 30.00 15.00 45.00 35.00 15.00 50.00 *Excludes some rates cited in circular, but not found in other data sources. Does not include the rates of 40% and 45% which still exist for some agricultural goods. The PFI was 12.5 percent for medical products and 10 percent for products which fall under the investment code; however, the L.oi de Finances 1995 eliminates the PFI on investment goods. Source: Administration des Douanes et Imp6ts Indirects; Circulaire No. 4320 du 17/3/1994. Table 3: Evolution of Nominal Protection Percent Tariffs PFI** Total Unweighted Weighted* Unweighted Weighted* Unweighted Weighted 1990 24.8 14.3 12.5 37.3 26.8 1991 25.1 15.3 12.5 37.6 27.8 1992 24.5 14.5 12.5 37.0 27.0 1993 23.0 14.2 14.8 14.4 37.8 27.0 Note: Quota coverage (percent of imports subject to lists B (subject to ministerial authorization) and C (prohibited)) was 11 percent in 1990 and 15 percent in 1992. * Weighted by total imports. ** Formal PFI averages (weighted and unweighted) unavailable for 1990-1992. Source: World Bank estimates based on data from the Ministere du Commerce Exterieur. 1990-1992 figures taken from World Bank (1994), Report No. 11894-MOR. 9. Table 5 shows the overall tariff collection rate, i.e., excluding exemptions for the first six months of 1994. 10. Table 6 shows that actual tariff collection rates tend to be significantly higher on final consumption goods than on other products. 11. Additional protection is provided by the system of reference prices for industrial goods. Though these prices in principle are supposed to cover no more than 10 percent of industrial production, it is unclear how this can be verified in practice since there is no obvious way to link the trade and industrial production data. Reference prices are specified at the product level according to the harmonized system (HS), while Moroccan industrial production is only tracked at a fairly aggregated subsector (sous-branche) level. Annex II  Table 4: Tariff and PFI Rates by Sector, 1993 (Percent) Tariff rate P F I Category Sector Unweighted Weighted* Unweighted Weighted* 1 Agriculture, Forestry, Fishing 20.72 8.46 14.60 14.39 4 Non Metal Minerals 10.03 8.62 15.00 15.00 5 Metal Minerals 8.97 0.96 15.00 15.00 6 Hard Fuel & Crude Petroleum 2.50 2.50 15.00 15.00 7 Refined Petroleum & Derivatives 17.17 21.28 15.00 15.00 8 Other Sources of Energy 2.50 2.50 15.00 15.00 10 Produce from Food Industries 28.19 12.24 14.56 14.49 11 Other Produce from Food Industry 30.41 15.41 14.94 14.31 12 Drinks & Tobacco 31.14 10.58 15.00 15.00 13 Textiles & Hosiery 30.04 12.70 15.00 15.00 14 Dressing Except Shoes 34.88 34.89 15.00 15.00 15 Leather & Leather Shoes 24.11 15.85 15.00 15.00 16 Woods & Wood Articles 26.30 10.15 15.00 15.00 17 Paper, Cardboard & Printing 29.03 18.79 14.87 13.48 18 Ore Minerals Goods 24.70 23.45 15.00 15.00 19 Basic Metal Industry 10.80 7.75 15.00 15.00 20 Metal Works (excluding Transport Machinery) 29.82 29.07 14.94 14.99 21 Mechanic Equipment Material 18.12 15.16 14.16 14.06 22 Transport Material 23.87 15.80 14.73 12.64 23 Electric & Electronic Material 23.83 25.20 15.00 15.00 24 Office Machines, Precision Instruments, Optic 22.59 17.49 15.00 15.00 25 Chemicals & Parachemicals 14.72 8.83 14.64 13.58 26 Rubber or Plastic Articles 27.45 25.47 15.00 15.00 27 Other Manufacturing Produce 31.36 29.11 15.00 15.00 t Weighted by a modified measure of imports excluding untaxed categories. Source: Minist?re du Commerce Exr:trieur. 12. Though anti-dumping measures exist in principle, their use as a protectionist device appears limited. To date, there have been no anti-dumping measures taken. There are currently two anti- dumping cases under consideration. The competition law in process appears to contain few egregious protectionist measures and to conform largely with international norms. 13. The removal of quantitative restrictions (QRs) under the UR will mean greater transmission of world price signals, though the dual (base plus marginal) tariff structure under consideration for some agricultural products will dampen these transmissions.3' Some products, e.g. milk and meat, will also benefit from the UR minimum access requirements. 14. Table 7 reports the principal findings from a recent study on effective protection. Effective incentive rates (EIR) indicate total protection of value added (VA) taking into account the value of incentives (from e.g. the Code d 'Investissement) and are defined as EPR + 1, where EPR is the effective protection rate and I is the value of incentives divided by VA at world prices. 3/ See MinistWre de I Agriculture, 'Note sur les Syst,?mes de Protection. Cas du Blc Tendre Annex II  Table 5: Imports by Tariff Rate, January - June 1994 (NMillions of Dirham) Tariff Rate Imports Excluding Percent of Estimated Tariff (percent) Untaxed Categories v Imports Revenue 0 5,118 21.63 0 2.5 4,385 18.54 109.50 7.5 1,825 7.71 136.80 12.5 3,541 14.97 442.50 17.5 2,919 12.34 510.80 25 1,475 6.09 368.52 30 1,441 6.09 432.10 35 2,910 12.30 1,018.60 40 5 0.02 2.00 45 37 0.16 16.80 Total 23,657 100.00 3,038.32 Estimated tariff collection rate 12.8% 1/ The higher (40% and 45%) tariff rates apply to agricultural imports. 2/ Untaxed imports include the admission temporaire and investments code categories. Source: Administration des Douanes et Imp6ts Indirects. 15. Morocco grants export incentives: these incentives apply in proportion of production exported and include a five year tax holiday with a 50 percent rate reduction afterwards; duty free imports and PFI reimbursement; and limited VAT exonerations.4' Table 6: Imports by Usage (Groupement D'Utilisation), January - June 1994 (Millions of Dirham) Usage Category Imports Excluding Estimated Tariff Weighted Tariff Untaxed Categories 1/ Revenue Rate Food, drink and tobacco 3,969 493 12.42 Energy and lubricants 2,681 253 9.44 Animal and vegetable based goods 2,123 189 8.9 Mineral based goods 302 5 1.66 Agricultural material 5,313 696 13.1 Industrial material 20 4 20.0 Consumption goods 7,100 900 12.68 2,149 500 23.27 Total 23,657 3,039 12.85 1/ Excluding untaxed (adnission tmporaire and investment code) categories. 4/ For further detail, see World Bank (1994), Report No. 12947- MNA. Annex II  -6- C. Temporary Agricultural Agreement With The EU 16. Morocco reached a temporary agreement with the EU regarding its tomato and zucchini exports for the 1994/95 season.5' 6] Future growth prospects for agricultural exports to the EU appear limited. 17. While the agreement appears to mitigate the full effect of the EU GATT offer, there will still be a negative impact on Moroccan exports. Tomatoes will be subject to a duty free quota of 130,000 tons. The entry prices for exports to the EU are higher than the previous reference prices during certain months of the year. Previously, Morocco benefited from a duty free quota of 212,800 tons.7' Exports during the 1993-94 season were 160,000 tons.8' The entry price for clementines at 645 ECU/ton is very high relative to EU market prices. The quota for oranges is slightly lower than export levels in previous years.9' Duty free quotas for clementines and oranges were 114,000 tons and 285,000 tons respectively. Another significant point of contention is the imposition of import licenses on EU importers of Moroccan fruits and vegetables."0' 18. Furthermore, the ability to export at or near previous levels alone does not mean that there is no welfare impact for Morocco since there could be unfavorable terms of trade effects as well. For example, even if the entry (minimum import) prices were to be set equal to the previous reference prices, the fact that Morocco (i) has lost its preferential quota access and (ii) has to pay import duty will lead to a diminution of its previous quota rents. It is not clear how much of the EU offer to the GATT fails to apply to Morocco, nor how long the above access levels will be maintained. Table 7: Effective Protection and Incentive Rates, 1991 (Percent) Sector Nominal Local Export Local Export Anti- Protection Effective Effective Effective Effective Export Protection Protection Incentive Incentive Bias Agro-industry 0.23 1.10 -0.25 0.25 -0.22 0.47 Textiles and leather 0.28 0.49 -0.13 0.55 0.05 0.5 Chemical & remated industry 0.14 0.04 -0.26 0.10 -0.05 0.15 Metals and machinery 0.37 0.67 -0.26 0.78 -0.10 0.88 Total 0.22 0.22 -0.23 0.30 -0.05 0.35 51 A temporary agreement on squash exports was also reached. 6/ The following information is from the December 8, 1994 issue of L'Economzste. 7/ 92,800 tons to the EU and a special quota to France of 120,000 tons. 8/ 1993 Moroccan exports to the EU for its three main agricultural products were: 1448,9 million DH (452 million DH tomatoes, 305.3 clementines, 691.6 oranges). 9/ Alternatively, Morocco could apparently accept to export at lower than proposed entry prices, but these would not decline in subsequent years. 10/ Both import licenses and minimum import prices are ostensibly illegal under Article 4 of the UR GAFT agreement, but could be allowed under the Special Safeguard Provisions in Article 5. Annex II  D. Fiscal Impact Of Trade Liberalization 19. Tariff and PFI revenue accounted for 25 percent of total tax revenue in 1993. Total direct and indirect taxes accounted for 30.6 percent and 44.4 percent of total tax revenue."' 20. PFI revenue proportion is almost as high as total VAT revenue (29 percent), a high proportion of which was from imports (52.7 percent)."2' 21. The formation of a FTA with the EU would lead to a significant fiscal loss for Morocco. Table 8 presents tariff and PFI revenue by sector. This is multiplied by the share of imports coming from the EU within that sector to yield an estimate of tariff and PFI revenue loss of approximately 7,137 million dirhams. This figure should be inflated by 19 percent to capture the (foregone) cascade effect resulting from application of the VAT on the tariff and PFI-inclusive import price."3' This yields a figure of about 8,500 million DH, which represents 68 percent of total 1993 customs revenue."4' 22. Other things equal, the method used generates a lower bound estimate of the likely revenue loss because it implicitly assumes no substitutability between imports from the EU and the rest of the world (ROW). Since there is likely to be some substitutability in reality, lowering tariffs on the EU to zero while maintaining tariffs on ROW while lead to an even higher share of EU imports than currently and would increase the likely revenue loss. 23. The estimate also assumes that Morocco does not allow duty-free access for agricultural imports from the EU under the FTA. In an earlier Moroccan proposal regarding the FTA, agriculture was included."5' Morocco would provide phased duty free access for both agricultural and non-agricultural imports from the EU in return for free access of Moroccan agricultural exports. However, since a 11/ There appears to be a discrepancy in the BAM report (6,312 million DH tariff revenue; 6,235 million DH PFI revenue) and the 1993 Loi de Finances (7,243 million DH tariff revenue, 8,486 million DH PFI revenue). The former figures are used since they also appear in the tax and budgetary receipt numbers. PFI revenue accounts for roughly as much revenue as tariffs because of fewer exemptions. 12/ Since agriculture is exempt from VAT, we would expect the ratio of VAT on imports total VAT 52.7 percent) to be comparable to the ratio of imports to total value added (VA) excluding agricultures (agriculture, sylviculture, etp&he), which was 42.3 percent for 1993. This would appear to indicate poorer internal collection. However, recall that import VAT has a cascading effect. Applying a 19 percent VAT rate on total 1993 tariff and PFI revenue was 12,547 million DH (Bank-Al-Maghrib figures) yields DH 2,383 million. This figure should be excluded from the above VAT ratio for comparison with the benchmark imports: VA ratio. With this adjustment, the VAT ratio becomes 43.5 percent, very close to the 42.3 percent expected benchmark. VAT exonerations (e.g., for imports used in exports) included in the above figure would only reinforce this result. 13/ Note that since imports from the EU would presumably still pay the VAT under the FTA, we are only estimating that portion of VAT revenue which would be foregone from an elimination of tariffs and the PFI on EU imports. This is why we multiply the 19 % normal VAT rate by tariff and PFI revenue, and not by the actual value of imports themselves. Tax revenue before the FTA = (t+v + tv)PFr where t = tariff plus PFI rate, v = value added tax rate. Tax revenue after the FTA = vP,. P., is the tax exclusive price of imports from the EU. Thus revenue falls by (t+tv)PF5. 14/ Morocco's 1993 total imports and exports were 61,908 and 34,366 million DH respectively, of which 54.5 and 62.4 percent were with the EU. The customs revenue loss from a FTA is therefore larger than the overall trade share would suggest. 15/ Nouvel Accord Maroc-Union Europeenne: Memorandum Marocain, 14/2/94 Annex II  separate preliminary agreement on agriculture has been reached which fails to provide the latter, Morocco may not provide EU agricultural imports with duty free access under a FTA.16' Were Morocco to provide EU agricultural imports with duty free access, this would increase the revenue loss to 8,972 million DH. 24. In the absence of any reliable import demand elasticities it is not possible to estimate the revenue effects of non-discriminatory tariff liberalizafion with any precision. Some benchmark estimates are done using constant import levels. To do this non-exempt imports, equal to 23,65 million DH from January- June 1994 as shown in Table 5 (equal to 49,630 million DH in 1993) are multiplied by the proposed tariff rate. Based on this figure, at a uniform tariff of 15 percent,17' tariff revenue equals 3,549 million DH (23,657 x 15 percent) and at a rate of 20 percent, 4731 million DH. This contrasts with current tariff revenue (excluding PFI) of 3,039 million DH. This increase is not surprising since the current estimated tariff collection rate (= tariff revenue/non-exempt imports) is only 12.8 percent. 25. Since this 15 percent or 20 percent uniform tariff is "all-inclusive", current PFI revenue has to be subtracted. PFI revenues for January-June 1994 were 3,555 million DH.15'. Therefore the net impact of a uniform tariff is -3,044 million DH (=3,549 - 3,038 - 3555) for a 15 percent tariff and -1,862 million DH (= 4,731 - 3,038 - 3555) for a 20% tariff. Incorporating the foregone revenue due to the VAT cascade effect would raise these figures by about 19 percent to -3,622 and -2,216 million respectively). 26. The above analysis is based on 1994 data when many agricultural imports were under quota. The fiscal impact of these conversions will not be known for some time. In addition, agricultural products will very likely have very high tariff rates; proposed rates are on the order of 200 percent. E. Trade Reform And Labor Market Rigidities 27. The existence of labor market rigidities in Morocco is evidenced by the high levels of urban unemployment. They could be induced by regulations on minimum wages or firing restrictions which lead firms to only hire at wages above the market clearing level. A binding and enforced minimum wage such as the SMIG set above market clearing levels is an obvious candidate. Firing restrictions could also lead to a similar outcome.'9' 16/ Of course, this depends upon the details of the FrA negotiations which remain to be determined. 17/ The 6 month 1994 figure is from the douane itself and likely to be more reliable. The latter is estimated by the Ministry of Commerce. 18/ This figure is from the customs department. 19/ For example, in a Stiglitz-Shapiro-type efficiency wage model, firing restrictions (lower probability or higher severance pay) lower the threat of layoffs and require a higher wage to be paid in order to satisfy the no-shirking condition. In a Dixit-type model with hysteresis, the existence of exit costs from firing restrictions causes firms to hire orly at a higher marginal value product than otherwise (akin to setting a higher strike price in a financial option). Annex II  -9 - 28. Figure 1 depicts a specific-factors model with two sectors, one import-competing (M) WE W sector the other export- competing (X).20' MVPi, Wo \ . / wi, and Li respectively denote the wage, marginal value product of labor, and labor employed in MVP a- P(l. t).MPM sector i, where i =X (i=M) is the export MVi i (import) competing Pml sector. The demand curve for labor is given V LL. La L. os by the marginal value I product of labor, which in the import-competing sector is MVPM = PM (1 +t) MPM, where t is the tariff on imports. Demand for labor in X is given by MVPX = Px MPx. Px and PM are the (given) world prices expressed in dirhams. Labor market rigidities cause the wage at which firms hire (w) to be higher than the market clearing wage w* . OLM and OLx workers are hired in the import and export-competing sectors, respectively, and unemployment is given by LMLX. Lowering tariffs from t to t, lowers the demand curve for labor in the import competing sector to MVPM'. The import competing sector shrinks while the export sector does not expand. Unemployment increases to LM'Lx.2"' 29. In the presence of such labor market rigidities, expansion of exports is better accomplished by an exchange rate depreciation.22' Exchange rate depreciation raises the domestic price of the export- competing good Px and shifts out the demand curve for labor in the export sector. It would be best, however, to target and address these rigidities directly. In a labor market without rigidities, reducing tariffs from t to t, would shrink the import sector but expand the export sector to OXLMl. 30. The above analysis considers only with the urban sector. How is it affected by a consideration of the rural sector? A reduction in urban unemployment could increase the expected payoff from urban employment and lead to greater rural-urban migration.3' Since the rural areas also export, it could be argued that an increase in industrial exports is being obtained simply at the expense of agricultural 20/ Capital is sector-specific: equipment used to produce (export-competing) textiles cannot be used to produce (import- competing) appliances. Labor is not sector-specific and can move across sectors. Since the amount of capital in a sector is fixed at any given moment, the marginal value product of labor in each sector is decreasing in the amount of labor employed in that sector. 21/ In a labor market without rigidities, tariff reduction would have shrunk the import sector but expanded the export sector to OXLM1. 22/ Of course, tariff reductions themselves will also engender some exchange rate depreciation since the demand for imports and foreign currency increase. 23/ As in the usual Harris-Todaro setup. Graphically, we can depict rural-urban migration in Figure 2 as an outward shift of the vertical axes with the demand curves for labor intersecting at the same height on the expanded axes (since the MVP has not changed for the first worker hired in each sector). The rest of the analysis then follows tirough. Annex II  - 10 - exports. However, since Morocco is effectively quota-constrained in many of its agricultural exports (e.g., fruits and vegetables to the EU), such items effectively become non-tradables at the margin. F. Free Trade Agreement With The EU 31. In the context of a free trade area, the gains for each country come largely from preferential liberalization by the others."' Morocco already has preferential access to the EU for its industrial exports. A Free Trade Accord will give the EU free access to Moroccan markets.25' 26/ 32. Consider Figure 2 which shows the worst possible outcome re ...e 2 for any good. MD denotes import Af ld grecmerrtwrh the EU Pdrr Fm. Trod. Agree.entwith the EUdemand for a product by Morocco. XS denotes export supply by the EU, which is assumed to be less xsi </ elastic than P,, supply from the rest G v tI r \ t;XSI ptof the world (ROW). Under a non- H /U discriminatory tariff t, export R MD supplies as perceived by buyers in Morocco are given by XS' and P"j. Imports come from both the EU F 1 a3 nIity and ROW and equal OQ, and Q,Q3, respectively.27/ Moroccan tariff revenue is given by the rectangle GHNS. 33. Morocco now forms a FTA with the EU while maintaining tariffs on ROW. Imports from the EU rise to 0Q2 while those from ROW decline to Q2Q3. The EU (Morocco) experiences an improvement (loss) in its terms of trade from P. to Pwt and gains a surplus of HUFG. Morocco loses GFLH which 24/ This is true in the Vinerian setup which underlies much of the standard discussion of trade creation and diversion. This section draws upon A. Panagariya, "Rethinking the New Regionalism," 1994. 25/ The fact that the EU is Morocco's largest trading partner does not change any of the results regarding gains and losses. There is no sound theoretical reason why trade shares in and of themselves matter directly for assessing the impact of a FTA. The net welfare impact depends largely upon (i) elasticities and (ii) whether or not the partner is the most efficient supplier over the relevant range. As shown below, larger shares could indicate that the partner is more efficient over a certain range, but it could also mean larger losses from a FTA. 26/ Some goods would be granted immediate free access; others over five and twelve years. The agreement would also cover issues other than trade, e.g. foreign direct investment and competition law. 271 This is borne out by the data. Morocco imports from both EU and the ROW even at the very disaggregated, harmonized system (HS) product level. This is the relevant level of disaggregation since tariffs are assessed on the basis of the HS. Annex 11  - 11 - it previously earned as tariff revenue.28' Since imports from ROW do not disappear, Morocco's internal price--and its expenditure and resource allocation--remain unchanged. There is no trade creation. 34. Trade creation could occur if imports of a given good from the EU completely displaced imports from ROW following the FTA. If EU export supply is given by dashed curves XS1 in Figure 2, then Morocco's internal price drops and one gets trade creation corresponding to the shaded triangle. Note that one is only likely to get trade creation when EU imports account for a very large share of total imports prior to the formation of the FTA (i.e., when XS,' intersects Pw' only slightly to the left of S).2' However, this is also precisely the situation in which foregone tariff revenues are likely to be large. The analysis in Annex 2 and Table 8 presents a rough estimate of 8,500 million DH in foregone customs- related revenue from forming a FTA with the EU. 35. In actual practice, the EU is quite large relative to Morocco, and may also be the lowest cost producer for many goods. In this case, for EU goods (the homogenous goods case) that are currently imported, the supply curve may coincide with P (sub w). For these goods, the gains to Morocco are unambiguously positive. The price falls to P (sub w) and tariff revenues are transferred to Moroccan consumers. The net gain to Morocco is given by triangle SNR. 36. It has been argued that the FTA is necessary for Morocco to maintain its existing preferential access to the EU. In that case, the benefits of the latter would have to be included in the calculation. The gains to Morocco from duty free access to the EU are conceptually similar to those that the EU would gain from having duty free access to the Moroccan market. We can therefore re-revisit Figure 2 with some change in notation. Let MD now represent import demand for a product by the EU. Since Morocco benefits from duty free access, Moroccan and ROW export supply as perceived by EU buyers is given by XS and P.,% respectively. The latter is also the EU domestic price. Imports from Morocco equal OQ2. 37. Suppose that Morocco's duty free access to the EU is revoked and it faces a non-discriminatory tariff t. The net price that its exporters receives falls to Pu, its exports to the EU decline to OQ1. Compared to the previous situation, Moroccan exporters have lost surplus rent given by area HUFG. For a given product, this is the value of Morocco's duty free access to the EU. 38. We calculate an indicative upper bound benefit to Morocco's duty free access by estimating rectangle HLFG, which is strictly larger than Morocco's surplus HUFG. HLFG is given by the value of Morocco's existing exports multiplied by the EU tariff rate in percentage terms. Table 9 presents post- Uruguay Round average tariff rates for the EU. These are multiplied by Moroccan exports to the EU 28/ The difference, triangle ULF, is lost due to trade diversion of Q1Q2 from the more efficient ROW supplier to the EU. 29/ We also get trade creation if XS,' intersects P.' to the right of S; this would imply that Morocco imports the good solely from the EU. Annex II  - 12 - on a sector-by-sector basis in Table 10 to yield an estimated (producer) surplus to Morocco from duty free access to the EU between 611 to 1,026 million DH.0' 3" Table 8: Estimated Tariff and PFI Revenue Loss from FTA, 1993 (Dirhams) Category Sector Tariff & PFI Share of Estimated Revenue Imports Revenue Loss from EU I Agriculture, Forestry, Fishing 1,397,690,478 28.8% - 4 Non Metal Minerals 32,342,468 12.8% 4,128,001 5 Metal Minerals 19,623,203 51.0% 9,998,432 6 Hard Fuel & Crude Petroleum 1,169,291,199 0.7% 8,128,508 7 Refined Petroleum & Derivatives 450,070,382 59.5% 267,750,450 8 Other Sources of Energy 75,884,793 0.0% - 10 Produce from Food Industries 342,230,153 11.3% 38,579,783 11 Other Produce from Food Industry 705,845,241 39.3% 277,435,039 12 Drinks & Tobacco 216,460,956 45.35 98,012,066 13 Textiles & Hosiery 382,674,579 74.3% 284,378,161 14 Dressing except Shoes 15,475,828 90.0% 13,929,547 15 Leather & Leather Shoes 18,701,902 87.95 16,439,523 16 Woods & Wood Article 309,161,916 29.1% 90,115,394 17 Paper, Cardboard & Printing 389,793,823 52.0% 202,674,780 18 Ores and Minerals 297,175,918 71.0% 211,116,118 19 Basic Metal Industry 843,694,463 80.0% 674,689,430 20 Metal Works (excluding Transport Machinery) 554,298,609 78.8% 436,941,825 21 Mechanic Equipment Material 1,097,523,439 80.8% 887,155,485 22 Transport Material 1,573,061,456 71.7% 1,127,367,917 23 Electric & Electronic Material 1,378,013,341 73.6% 1,014,479,497 24 Office Machines, Precision Instruments, Optic 440,951,660 66.2% 291,855,678 25 Chemicals & Parachemicals 1,304,744,138 71.5% 932,520,618 26 Rubber or Plastic Articles 272,133,662 71.1% 193,381,461 27 Other Manufacturing Produce 79,495,193 69.7% 55,442,706 Total Tariff and PFI Loss 7,136,520,418 VAT Loss from Cascade Effect (@ normal 19% Rate) 1,355,938,879 Total Revenue Loss 8,492,459,298 Source: Revenue and import share data from Ministere du Commerce extrrieur. 30/ The latter figure (1,028 million DH) is likely to be an overstatement since it applies the "overall merchandise trade' average tariff rate to a large portion of Moroccan exports ("remaining exports" in Table 8). This rate includes agricultural goods, which has the highest EU tariffs (22.5 percent) by far, and we know that the "remaining exports" for Morocco do not include any agricultural goods. 31/ The weighted EU agricultural tariff rate is 22.5 percent. Applying this to total 1993 Moroccan agricultural exports to the EU of 3,995 million DH yields an estimated surplus of 899 million DH. This increases the estimated surplus from duty free access to between 1,510 and 1,927 million DH, which is still far below the revenue loss estimate. If Moroccan agricultural exports to the EU remain subject to import duties (per the EU offer to the GATT) under the proposed preliminary agreement, then the gains are even less. Annex 11  - 13 - Table 9: Average EU Tariff Rates - Post Uruguay Round (Percent) Average Tariff Reduction Average Tariff Level Unweighted Weighted by Post Uruguay Weighted by Summary Category (1 digit) Imports from Round Imports from around the Unweighted the World ** World Agriculture (0+1 +4+27-28) * 3.0 2.7 22.5 21.2 Fertilizers 2.4 1.9 1.7 2.0 Mineral fuels, etc. (3) 1.4 1.2 2.1 2.0 Chemicals (5) 3.1 3.6 4.7 4.3 Basic manufactures (6) 3.6 4.4 3.6 3.0 Machines, Transport Equipment (7) 2.6 3.0 2.7 3.0 Miscellaneous Manufactured Goods (8) 2.6 2.6 6.0 6.2 Goods not classified by kind (9) 2.4 1.7 1.8 1.5 All Merchandise Trade 3.0 3.1 7.0 5.9 * Includes estimated tariff equivalents of tariffied NTBa from M. Ingco, 'How much agricultural trade liberalization was achieved in the Uruguay Round?' * Import valuea for year 1986. Table 10: Estimated Surplus from Duty Free Access to EU Market Excluding Agriculture, 1993 ()irhams) Sector Moroccan Exports Average EU Estimated to EU Tariff * Surplus Agriculture (sector 1) 3,995,293,500 Fertilizer, minerals, ores, scrap (sectors 4,5,18; SITC 27-8) 1,903,221,446 1.7 32,354,765 Mineral Fuels (sectors 6,7,8; SITC 3) 653,432,976 2.1 13,722,092 Chemicals (sector 25;' SITC 5) 2,628,667,250 4.7 123,547,361 Subtotal 9,207,615,172 170,083,218 Remaining Exports 12,249,926,239 Basic Manufactures (SITC 6) 3.6 440,997,345 Miscellaneous Manufactured Goods (SITC 8) 6.0 734,995,574 All Merchandise Trade 7.0 857,494,837 Trade (range) 611,080,563 905,078,792 1,027,578,055 * Average unweighted EU tariff levels following reductions at the Uruguay Round. See Table 9. Annex II  - 14 - 39. Several caveats are worth noting. First, there is no official correspondence between the EU tariff data which are on a (one-digit) SITC basis and the Moroccan export data which are on a sectoral or branche basis (Moroccan nomenclature). Some judgement regarding correspondence was therefore required. The issue of which are the appropriate tariff rates to apply accounts for the range of estimates. This is reported in the Table 10. Second, surplus from agricultural exports to the EU are excluded. This is not because they are trivial, but because the terms of the agricultural agreement are separately determined prior to negotiations regarding the establishment of the FTA. It therefore should not enter into the decision regarding the desirability of a FTA. 40. The net direct gains for Morocco from the FTA with the EU are difficult to calculate, given the wide variety of goods imported by Morocco. Here we have approximated two possible elements. The benefits of maintaining the existing Moroccan access to EU markets were approximated above at roughly 1 billion DH. Table 8 presents a rough estimate of foregone tariff revenue of about 8.5 billion DH. If the arguments above concerning the infinite elasticity of supply of most goods currently exported by the EU to Morocco are correct, the net gains, even excluding the maintenance of access to the EU market, will be positive but not necessarily extremely large.2' Most proponents of these types of FTA arrangements, paring a developing country with an industrial region, cite indirect gains as the most important. These indirect gains -- notably increased confidence by investors in the Moroccan economy and in the stability of Moroccan economic policies -- are undoubtedly important but are even more difficult to quantify. In any case, liberalization with respect to the rest of the world as well would enhance Morocco's welfare gains. 32/ A study by T.F. Rutherford, E.E. Rutstrom and David Tarr, (1993) using a computable general equilibrium model for Morocco with 39 production sectors (and based on a 1980 input-output table), estimates net welfare gains to Morocco of 1-2% of GDP. This study assumes, among other things that there is an accompanying real depreciation of the exchange rate (to ensure current account sustainability) and that there are no terms of trade effects associated with the FTA.  ANNEX III-A THE LABOR MARKET Section I: An Overview of the Labor Market and Main Conclusions 1. This annex provides the background information and analysis for the issues discussed in Chapter III-A in the main report of the CEM. It first presents the main results related to the functioning of the labor market and then presents the analysis on which these results are based. 2. Introduction. The population of the Kingdom of Morocco is growing rapidly, with an estimated demographic growth rate at just above 2 percent in 1992." This figure represents a fast reduction compared to 1987, when the rate of demographic growth was estimated at 2.4 percent. The process of urbanization has also been rapid. In 1994, the total population was estimated at 26.1 million, divided almost equally between the rural and the urban regions, with 12.66 million living in the former and 13.41 in the latter. In 1987, the corresponding figures were 12.32 million and 10.56 million, respectively. These figures suggest that the rural population has been nearly stagnant, with an average annual growth rate over 1987-1994 of 0.4 percent, while the urban population has had fast growth, at an average annual rate of 3.4 percent over the same period. This implies that the labor market must absorb an increasing number of job seekers every year, with most of the growth in the labor force taking place in the urban sector. The rural labor market is not as well known as the urban one. Most of the information available on the former comes from the Rural Labor Force Survey 1986-87, and from the LUving Standards Survey 1990-91. In contrast, the latter has been surveyed every year since 1984, and the analysis below is largely based on information from the Urban Labor Force Survey 1993. 3. The Labor Force. The Living Standards Survey of 1990-91 puts the gross participation rate of the labor force at 39 percent of the total population on average, with a noticeable difference between the rural and urban sectors, whose rates were 44 percent and 33 percent, respectively. The difference results mainly from the participation rates for women, which are estimated at 33 percent in the rural sector and 17 percent in the urban sector. The average participation rate in the urban sector is 47.9 percent of the urban population above the age of 15, divided between 72.1 percent for males and 22.1 percent for females.2/ 4. Employment type differs markedly between the urban and the rural sectors. The frequency of wage employment varies between the two.3' In the former, 62 percent of the labor force is classified as wage earners, while only 17 percent of the rural labor force work for a wage. Wage employment is particularly rare among rural women, at a rate of 5.2 percent. However, an increasing number of women are said to accept wage employment in the irrigated areas, where farms are run in a more modern fashion, while poor areas are still governed by traditional rules. For men, the fraction of the labor force working for a wage is 25.4 percent, a non negligible figure. Moreover, this may be regarded as an 1/ Publication of the results of the recent census, due in 1995, will provide a more precise estimate. 2/ From the Urban Labor Force Survey, 1993. 3/ Using the data from the Living Standards Survey 1990-91. Annex III-A  underestimate, as people are classified according to their main activity, while many rural workers not only own their own farm but also work seasonally for a wage (see Pascon and Ennaji, 1987). These individuals are probably classified as self employed, while wage employment is a crucial complement of their income. Hence, labor market conditions in the rural sector affect more than one male worker out of four.' 5. The Rezulatory Framework. The regulatory framework governing the labor market in Morocco is not the most conducive to the fast growth of employment that is required. There is a worldwide emerging consensus that a free-working labor market is an important facilitating element for fast growth, as exemplified by several East-Asian countries." For example, labor market institutions in Hong Kong, Korea, Singapore and Taiwan have kept this market free from most of the distortions that can be found in many developing countries.6' As a result, these countries which have successfully pursued a policy of export-oriented labor-intensive growth, have seen both low levels of unemployment and fast growth of real wages. The Moroccan labor market on the other hand seems to be overly restrictive, preventing the efficient reallocation of labor required for high growth. 6. Three main elements of the labor market deserve particular attention (in Morocco). These are: first (a) minimum wage legislation, which affects the cost of labor for the lower-paid jobs; second, (b) social charges and other levies bearing directly on the payroll; third, (c) restrictions on worker dismissals. These elements play a large role in shaping the way this market functions. Unlike many other developing countries, Morocco has well enforced labor regulations. This is due both to the strength of the administrative framework, where the Ministry of Employment and Social Affairs has a very active enforcement department (Inspection du travail), and to the importance of trade unions. There are three main trade unions in Morocco: the Union marocaine du travail (UMT), the Union gengrale des travailleurs marocains (UGTM), and the Confdfration democratique du travail (CDT). To some extent, the strength of the trade unions results from their close links to political parties. These unions are very active in helping to enforce labor regulations. 7. Minimum Wage Policy. While there is a negligible level of unemployment in the rural sector, the rate of urban unemployment is alarmingly large, reaching 15.9 percent in 1993 (14.2 percent for men and 21.7 percent for women). The minimum wage legislation appears to be one of the main reasons for the high rate of unemployment in the urban sector. In the fast growing East-Asian NICs for example, there is either no minimum wage legislation, or the minimum wage is fixed so low that it is irrelevant (Fields and Wan, 1989). As shown in section B of this annex, the minimum non-agricultural wage, called the SMIG (Salaire minimum interprofessionnel garanti), has a very clear impact on the wage distribution in the urban private sector in Morocco. A very large share of the labor time employed in this sector is paid at or just above the SMIG7' as shown by the wage distribution (chart Al). 4/ Landowners as employers are of course, also affected by labor market conditions. 5/ World Development Report, forthcoming 1995. 6/ Fields and Wan (1989), and Fields (1994). 7/ Unless the SMIG happens to reflect labor productivity for all these workers it is distortionary Annex III-A  - 3 - 8. Econometric evidence suggests that the SMIG is revised periodically to catch up with the average wage rate. The process that determines SMIG revisions is not crystal clear. Although the government seems to play the major role, trade unions and employers associations, like the Chambre de Commerce et d'Industrie, have a non-negligible influence. 9. The empirical evidence suggests that the government over-compensates for changes in the average wage, so that the margin between the SMIG and the average wage has been shrinking quite rapidly since 1978. Given the average wage, this implies that the whole wage distribution has become more compressed. Hence, while increases in the minimum wage may not have caused the average wage rate paid in the private sector to increase (so far), its fast growth has resulted in a compression of the lower tail of the wage distribution. It is the wage rates of the lowest paid workers, and thus their employment opportunities, that are directly affected by the SMIG. This is supported by equation (3) in Section B, which shows that the margin between the minimum wage and the average wage is an important determinant of the urban rate of unemployment. Hence, in order for private sector urban employment to increase faster, the government must temporarily arrest revisions of the SMIG, in order to restore the margin with the average wage rate to a more appropriate level. In addition to the unemployment effect of the too high SMIG, one may expect a disincentive effect on workers at the upper end of the distribution, as wage differentials become smaller." 10. The data also indicate that there are individuals that are paid below the SMIG. This does not necessarily imply that the SMIG is not enforced in these firms or for certain types of workers because the law allows employers to pay less (in some cases much less) than the SMIG to workers under 18 years of age. The legal working age is 12 years. Table 1: The SMIG and the SMAG SMIG(I) SMAG(2) Ratio (1)/(2) 1/1/1988 34.56 22.35 1.546 1/5/1989 38.00 24.58 1.546 1/5/1990 41.76 27.03 1.545 1/1/1991 48.00 31.08 1.544 1/5/1992 53.00 34.18 1.551 Source: Direction de la Statistique (1993). 11. In addition, there is a large gap between the SMIG and the SMAG (Salaire minimum agricole garanti), the minimum wage applicable in the agricultural sector. Table 1 shows the daily rates for the SMIG and the SMAG, and their ratio, at the dates of their recent changes between 1988 and 1993. This 8/ It is noteworthy that in most countries of the developing world during the 1980s, the gap between the minimum and average wage has been increasing, and this seems to have occurred as a result of falling real minimum wages. See "The impact of Labor Market Regulations, Lyn Squire and Sethaput Suthiwart-Narueput, World Bank-PRD, September 1994, draft paper. Annex III-A  - 4 - gap, which is equal to about 55 percent of the SMAG is probably unwarranted for cost-of-living reasons9'. It also creates an incentive for rural labor to seek employment in the non-agricultural sector, and hence, to a large extent, to migrate to cities where non-agricultural jobs are most likely to be found. The SMAG is well enforced in the rural sector, and rural trade unions are said to be even more active than urban ones in getting regulations enforced."0' 12. The econometric results suggest that the real growth of the SMAG had a positive effect on the fast growth of agricultural output that occurred in Morocco in the 1980s.1"' The SMAG policy pursued in the 1980s probably also played a useful role in reducing the incentive to migrate from the countryside, by drastically reducing the incidence of poverty. As such, it probably slowed down the growth of urban unemployment. 13. However, the positive effect of the SMAG described above was made possible by historical circumstances, which have changed in the 1990s. First of all, the SMAG was very low in real terms in 1978, so that its subsequent fast growth was to some extent making up for lost ground. Second, the agricultural sector has been exempt from most taxes since 1984, and this boosted the incentive to expand production and investment in this sector. However, it is also true that there is a point beyond which any increase in the SMAG will entail an increase in unemployment in the agricultural sector (Azam, 1994.b). Although it is not possible to numerically identify this level of the SMAG, the results from reaching it would be increasing rural unemployment due to policy-induced labor market distortions or increasing migration to towns. In terms of policy conclusions, we can say the following: in the future, the SALAG must not grow faster than agricultural productivity, which is itself determined by investment and modernization. 14. To sum up, the foregoing analysis suggests that the future course of action regarding the SMIG and the SMAG should aim at progressively reducing the (unwarranted) gap between the two, by arresting growth of the SMIG, and by increasing the SMAG no faster than agricultural productivity. 15. Non Wage Costs of Labor. The various levies on labor drive an important wedge between the cost of labor and the net take home pay enjoyed by the workers. The social charges levied by the CNSS (Caisse nationale de securite sociale) in the urban private sector amount to 18.6 percent of the take home pay. The CNSS has an elaborate inspection system, that ensures a high degree of compliance. Firms must also pay a 1.6 percent levy in favor of vocational training. In addition, firms pay various types of insurance (accident, etc.), which are not compulsory. Thus, the wedge driven by these different levies between the cost of labor to the firm and the take home pay may vary between 20.2 percent and 35 9/ The Poverty Assessment indicates that there is no systematic price differential between urban and rural areas. 10/ Fields and Wan (1989) show that the gap between agricultural and manufacturing wages is very narrow in the fast growing East-Asian NICs. II/ As shown in World Bank (1994), the incidence of poverty was fell between 1984-85 and 1990-91, with the number of people with average consumption below the poverty line falling from 21 to 13 percent. Annex III-A  percent, depending on the firm."2' Assuming, for example, that the elasticity of the demand for labor with respect to the cost of labor is 60 percent, and that the representative rate of charges and other levies is 30 percent, while the wage rate is fixed by the SMIG, the wedge between the cost of labor and the take home pay results in a level of employment that is 18 percent lower than would be the case were these funds raised by other means (e.g. lump-sum taxes). In section B a sensitivity analysis is provided for this simulation, with respect to the rate of the charges, and to the elasticity of the demand for labor. In most cases, the deviation of employment from the level it would reach without the wedge is significant. 16. It is noteworthy that among the fast-growing East-Asian NICs, Singapore is an example where social charges are in fact extremely high, falling in the late 1980s from 50 percent to 35 percent of the wage rate."3' However, it is also a country where there was wage repression during the same period to keep wages rather low. It is the total cost associated with hiring an additional worker that matters for labor demand. In situations where social charges are a high percentage of wages, lower wages will compensate for the higher non wage costs. However, in Morocco, this is not possible for all wage classes because of the added restrictions implied by the binding SMIG. 17. Though non-wage costs are high in Morocco, a well functioning social security system is an essential ingredient of a well-functioning labor market. Lump sum non-distortionary taxes to finance social security are often not available. Thus, it is best to reduce the distortionary labor market effects by social security reform,'4' to institute some form of unemployment compensation, and to reduce other costs/rigidities in the labor market. 18. Worker Dismissals. The Moroccan system penalizes worker dismissal. Instead of leaving the decision to fire workers fairly free for the firm, while offering the workers unemployment benefits, the labor code emphasizes severance pay, and judicial recourse when employment is terminated. Trade unions are said to be very active in making sure that these regulations are enforced. Individualfiring for economic reasons (eg mismatch, technical changes, dynamic changes) is forbidden by the law. '"' Only collective layoffs for economic reasons are allowed, subject to a preliminary authorization by the regional authority. Any such decision must go to the court, unless an expensive compromise is found. The new labor code, due to be published in 1995, provides some more flexibility than the old one in this respect, by mandating a deadline beyond which the authorization is regarded as having been granted. 19. As a result of the above rigidity, hiring decisions are to some extent made irreversible. As is well known by now from the investment literature,'6' this irreversibility (when combined with the intrinsic 12/ A study "International Comparisons of the Wage and Non Wage Costs of Labor", WPS 188, 1989, shows that in per capita terms non wage costs in Morocco are among the highest in a group of middle income countries. 13/ Fields and Wan, 1989. 14/ See Chapter Ill-A of the main report and Annex IV-B. 15/ Any such decision must go to the court unless an expensive compromise is found. Only individual dismissal for disciplinarv reasons are allowed. 16/ See e.g. Dixit, 1992. Annex III-A  - 6 - uncertainty of the future) raises the payoff from delaying hiring. As information becomes available with the passage of time, the firm has an incentive to wait until uncertainty is reduced before making an irreversible decision. Large severance pay and other firing costs result in this type of effect regarding hiring decisions, and thus reduce the rate of job creation. Hence, the Moroccan system, which results in the firm having to pay on average about two months of wage per year of seniority to a separated worker, protects the incumbent workers at the expense of the new entrants." An interesting contrast is the situation prevailing in Hong Kong, where firms are required to pay a flat week of wages as severance pay.'8' More generally, the fast growing countries of East Asia have relied less on rigidities in the labor market to protect workers and more on job creation in a flexible environment. The Moroccan system is not an efficient policy in a country with a rapidly increasing urban population and laborforce. A better course of action would be to rely more on well designed and targeted safety nets to protect dismissed workers from destitution, while leaving firms to adapt their workforce to changing conditions as efficiently as possible. 20. Severance payments have been signalled as possible causes of the increase in structural unemployment in OECD countries."9' High and uncertain severance payments make it hard for firms (especially the financially distressed ones) to adapt to change. They also induce conflicts between employer and employee since a worker with many years of service has an incentive to "induce" his retirement. In addition, others may delay job changes in order to receive higher future severance pay. In the long-run it would be advisable for Morocco to move away from this sort of compensation for unemployment towards unemployment insurance designed to minimize the negative incentive effects, while allowing mobility.?" 21. One result of the rigidities in the law regarding firing (which only applies to "permanent" workers, i.e. those who have worked in the firm for over a year) seems to have been an increase in temporary employment.21' While it is true that the use of temporary workers allows firms to circumvent labor market rigidities to a certain extent, they are not the first best solution.2' a) The use of temporary workers may lead to the under provision of training or the accumulation of firm specific human capital. This has negative effects on firm productivity. It may bias this type of hiring towards the unskilled who are generally the poorer and the less protected individuals 17/ 'Te amount of severance pay actually mandated is far below the average amount actually paid. It varies from 5 weeks for up to 5 years of service to 38 weeks for 15 years of service. 18/ Fields and Wan, 1989. 19/ "Labor Policy in Democratic Chile", Ren6 Corazar, IADB, OECD Development Center. International Forum on Latin American Perspectives. 20/ Chile's experience in this regard is useful. 21/ See Poverty Assessment-MOR. 22/ There has been an unprecedented use in temporary unemployment in Spain, which has one of the worst urban unemployment records in the developed world and also one of the most rigid labor codes related to worker dismissal. Annex III-A  in society. In Morocco for example, the share of unskilled workers in total permanent employment fell from 75 percent in 1984 to 42 percent in 1990. b) The development of temporary contracts may lead to an increase in overall unemployment. The reasoning is as follows: when the proportion of temporary workers is high in a firm or a sector, it strengthens the bargaining power of the permanent workers. This may result in higher wages than would otherwise be the case for the permanent workers and could have negative effects on the level of employment. This also has negative income distributional consequences."' c) The disincentive effects of temporary employment on the worker may lower long-run productivity. 22. Another probable consequence of the large firing costs is the especially archaic matching technology between job searchers and firms in Morocco. Labor market intermediation by private firms was forbidden by the old labor code, but is to be liberalized in the new one. However, firms had circumvented this problem, by relying on "consulting firms" for getting "advice" on their recruitment decisions. More importantly, the data show that job seekers rely mainly on personal contacts with potential employers and on friends and relatives for finding employment (see tables 10 and 11). This is what would be expected in more traditional economies rather than swiftly modernizing ones. This institutional distortion may in fact be endogenous, resulting not from, as some may believe, the cultural background, but from the high firing costs imposed on firms. When linked to his employee by a non- anonymous social relationship like that created by "quasi-clannish" ties, the employer can collect a lot of information on the would-be employee from the group before hiring. In addition, he can find various means of exercising some social pressure on incumbent employees to quit "voluntarily" when the need occurs, without resorting to the legal system. 23. As a result of the high cost of labor, entailed by the minimum wage, the high social charges, the high and uncertain severance pay, legal and other related costs, the incidence of long-term unemployment is extremely high in the Moroccan urban sector. Table 9 shows that more than 50 percent of the unemployed remain in this position for more than a year, irrespective of educational achievement. Those with no education at all seem to find employment slightly faster than the others, while more than 70 percent of the unemployed having some education stay on the dole more than a year. About 85 percent of those with the certificate of secondary education or the baccalaureat stay unemployed more than a year. However, educational achievement does have an impact on earnings and the probability of being unemployed. 24. New Labor Code. A new draft labor code has been prepared to replace the old one. This code aims to make labor market legislation more flexible. For example, under the previous code relating to collective dismissals, employers were required to wait for three months for approval which had to come from the governor. Under the new code, the waiting time has been reduced to two months after which, in the case of no response, approval can be assumed. In addition, approval is not required from the governor but from a delegated regional official. In practice, it remains to be seen how this will affect the dismissal regime. The draft labor code introduces the possibility of establishing more than one 23/ Spain's experience in this regard is a good case in point. Annex III-A  - 8 - minimum wage, the level of which would depend on the nature of the economic sector. This development would render labor market legislation more complex and could introduce new rigidities. 25. Impact of Educational Achievement on Earnings and Unemployment Probability. Statistical analysis of the relationship between educational achievement and (a) earnings and (b) the probability of being unemployed was carried out; the results are presented here. However, due to some data problems, these results remain preliminary and more detailed work should be done in these areas in the future. (a) Unemployment Probability. The probability of being unemployed as a function of various worker characteristics is analyzed with the aid of a probit equation shown in Table 7. Among other sets of variables, this equation describes the impact of educational achievement on the probability of being unemployed, while controlling for many other effects. The results indicate that the Certificates of Primary or Secondary Education do not make a significant contribution to reducing the probability of being unemployed, compared with no diploma at all. Even the Baccalaureat and Vocational Degrees do not have a significant impact. Higher education is what really makes unemployment less likely, partly because it provides access to jobs which are unaffected by the minimum wage. The most useful diplomas for reducing the probability of being unemployed are the middle range technical diplomas that open the way to jobs classified as Cadre Moyen. These degrees require a shorter course of study than University degrees and provide some basic training in either administrative and management skills, or some technical production skills. After these, the diplomas offered by Universities, Grandes Ecoles and Medical Schools are significant reducers of the risk of unemployment. However, the category of Technical Degree (Cadre Moyen) is by far the top performer, suggesting that there is a shortage of middle rank management in Morocco. (b) Earnings. An analysis of the determinants of earnings in the urban private sector shows that middle ranking technical diplomas between the upper secondary and the University ones do not only reduce the probability of being unemployed compared to most other diplomas, but affects earnings positively. Table 3 presents the results of an estimated earnings function for the private sector, using the data from the 1993 Urban Labor Force Survey. The dependent variable is the monthly wage income of the interviewed employed person. Besides other sets of variables that control for many characteristics of the individuals whose monthly wage incomes are analyzed, this equation describes the effects of educational achievement, here captured by the highest degree held by the person. The reference category is no diploma at all. The Certificate of Primary Education does not have a significant effect on earnings."4 (This may be partly due to the wage compression induced by minimum wage legislation). What is more surprising is the fact that degrees from Grandes Ecoles and Medical School are not significant. However, it is highly probable that the failure of these variables to have a significant impact on the equation is entirely due to their quasi-collinearity with another included variable, namnely Scientific and Liberal Professions, that comprises medical doctors, qualified engineers, etc. It is shown that the Certificate of Secondary Education, which is acquired after four years of secondary education, makes a slight contribution to average monthly earnings (DH 157.36), whereas the Baccalaurelat, which is acquired after seven years of secondary education, has a more noticeable impact (DH 557.22). The technical diplomas described above, that usually correspond to two or three years of post-secondary 241 Note that consideration of private benefits and costs alone are not sufficient to guide private investments. It is the social costs and benefits that should really be considered. Annex III-A  - 9 - education, have a significant effect while University diplomas (medical excepted), lead to the top paying jobs. 26. To summarize, the analyses in section C show, as described above, that higher education is an important contributor to high salaries and low unemployment risk. Secondary education diplomas do not significantly reduce the probability of being unemployed, but do have an impact on expected earnings. 27. Aware of the problems associated with labor market rigidities, the government is directing special efforts in enhancing employment opportunities for people with diplomas. First, it is developing a network of placement bureaus especially for them, called the C.I.O.P.E. (Centres d'information et d'orientation pour 1'emploi), which allows for a special type of labor contract (Stages de formation-insertion), exempting firms hiring young people with diplomas from many obligations of the Labor Code for a period of 18 months. It is not clear what type of arguments can be put forward for defending this type of special treatment, beside political economy reasons.25' From a purely static microeconomic point of view, this type of endeavor resembles the "Concorde Syndrome": having produced (and subsidized the production of) a lot of individuals with diplomas that are regarded for some reasons as being not well enough employed in the economy, the government carries on spending money on them in order to make sure that their employment prospects improve. From a microeconomic point of view, this course of action seems unwarranted. From a more dynamic point of view, one can regard the accumulation of human capital as a source of dynamic externalities, justifying the government intervention. However, more empirical analysis is required to support this assertion and to present an estimate of the value of such an effect. Moreover, the development of the CIOPE system is explicitly based on the diagnosis of an inadequate matching technology in the labor market, described to above. Hence, it is more important to reduce rigidities that result from the Labor Code directly.26' 28. Sectoral Effects on Earnings and Unemployment Risk. Among the various characteristics whose impacts are analyzed with the regression equations mentioned above, the sector of activity plays an important role in determining earnings and unemployment risk. Individuals working in Gas, Electricity and Water, or in General Administration, are significantly less exposed to the risk of unemployment than the others. In these sectors, public and semi-public ownership are the dominant types of organization, and this may explain the relative low risk of unemployment in these sectors. Urban workers involved in Agriculture, Forestry and Fishing, and Food, Drinks and Tobacco, are those with the highest risk of unemployment. 29. Sectoral effects are similarly present in the earnings function. For the private sector equation, the reference sectors are Mining and Quarrying, Electricity, Gas, and Water, and General Administration, where earnings seem to be homogeneously highest. Then, Other Manufacturing Industries does not fall far behind, paying on average (and ceteris paribus) DH 244.73 less. Then, a group of sectors including Repair, Wholesale and Retail Trade, and Transport and Telecommunication, pay respectively DH 325.11, 358.73, and 301.30 less than the top sectors. Not surprisingly, Agriculture, Forestry and Fishing, and 25/ In view of the rigidities in the Labor Code, apprenticeships could have been extended to the entire labor force, rather than being limited to this group. 26/ See para. 21 for discussion on temporary workers. Annex III-A  - 10 - Personal and Domestic Services are the bottom sectors, with an average pay that is respectively DH 636.86 and DH 573.87 less per month than the top sectors. Notice that Restaurant and Hotels and Collective Services are not much above this bottom group, paying on average (and ceteris paribus) DH 465.23 and DH 486.89 less than the top sectors, respectively. 30. Hence, Gas, Electricity and Water, and General Administration, which are dominated by public or semi-public ownership, are the sectors that offer both the highest earnings, and the lowest probability of being unemployed. It is interesting to note that in recent years the public sector wage bill has been increasing as a result of an increase in the number of workers employed, rather than higher wages as the authorities struggle with the high unemployment rate. 31. In the public and semi-public sector, the top paying sector is bank, Insurance, and other Services to Firms, which pays about DH 700 more than other sectors, controlling for many variables (see table 6). Other Manufacturing Industries pay DH 292 more than the reference sectors, while Collective Services pay significantly less (- DH 433). All the other sectors can be bunched together as reference sectors (see table 6). 32. The Informal Sector. While there is some qualitative information on this sector, it is difficult to come up with any quantitative information on it.27' The Direction de la Statistique conducted a survey on localized unstructured firms in 1988: however, it is not certain that this corresponds to what is usually called the informal sector. Their definition in this survey is based on whether or not firms have a formal accounting system. These firms were estimated to employ about 480,000 persons. Note that even in the non-structured sector workers on average were paid about the minimum wage. More importantly, this definitional issue cannot be made operational for the data used here, which are based on household surveys. No question has been posed about whether people were employed in a formal or an informal firm.28' Section II: A Statistical Analysis of the Functioning of the Urban Labor Market in Morocco 33. Wage employment is the dominant mode of employment in the urban sector of Morocco, affecting about 60 percent of its working population and factors affecting it are of extreme importance for the development of the urban sector. Table 2 presents a classification of the working population according to their employment status in Morocco. 27/ See Salahdine (1988, 1991). 28/ Another estimate implies that of the 600,000 new jobs created between 1986 and 1990, only 30 percent were for salaried workers and that the difference can largely be attributed to the expansion of the informal sector. Report No. 1 1918-MOR. Annex III-A  - 11 - Table 2: Number of Working Persons by Employment Status Status 1992 1993 % Growth Independent 687,285 708,700 4.2 Wage Earner 2,089,831 2,170,038 3.8 At Home 159,300 179,758 12.8 Employer 150,146 169,987 13.2 Family Aid and Apprentice 279,133 285,273 2.2 Member of a Cooperative 125,572 140,081 11.6 Undeclared 3,004 5,482 Total 3,494,271 3,659,319 4.7 Source: Direction de la Statistique (1994). 34. Table 3 presents the numbers of wage earners classified by the status of their employers. This table shows that private enterprises are the main type of employers of wage labor in Morocco. However, the government (central and local) and the public and semi-public sectors, taken together, are not far behind in terms of numbers of employees. It is noteworthy that employment in private enterprises has shown healthy growth between 1992 and 1993, though 1992 and 1993 have been recession years, with GDP decreasing by 4.1 percent in 1992 and by 1.1 percent in 1993. The large increase in the number of jobs in the central and local government sectors is probably an attempt to mitigate its effects on global employment.29' Table 3: Number of Employees by Institutional Sector Sector of Employment 1992 1993 % Growth Central and Local Government 629,516 676,731 7.5 Public and Semi-Public Sectors 241,792 238,855 -1.2 Private Enterprises 1,114,509 1,153,040 3.5 of which: < 10 Employees 477.512 505,46 5.8 2 10 Employees 636.997 647.571 1.6 Household Employees 81,013 74,809 -7.6 Undeclared 23,001 26,603 . 29/ Also, since the recession was caused by the drought, it is possible that rural wages and employment were affected more adversely. Annex III-A  - 12 - A. Earnings in the Urban Private Sector 35. Table 4 presents the results of an estimated earnings function for the private sector, using the data from the 1993 Urban Labor Force Survey. This equation analyses the different characteristics that determine the level of this wage income. The dependent variable is the monthly wage income of the interviewed employed person. The sample comprises 12,670 employed persons, and the equation explains 25 percent of the variance of monthly earnings. The F-test shows that this equation is significantly different from a constant mean value. 36. The first set of characteristics included are personal data. The age variable shows that seniority matters in the Moroccan private sector, as both age and its square are significant in this equation. However, the quadratic term has a negative sign, showing that the impact of seniority decreases as the person gets older. The estimated parameters imply that ceteris paribus, a wage earner gets an additional DH 24.94 per year of age when he (or she) is twenty, 16.74 when he (or she) is thirty, and 8.54 when he (or she) is forty. The maximum effect of cumulated seniority is found when the person is 50.41, corresponding to the age at which the derivative of this equation with respect to age is nil. After that point, earnings fall at the margin for every additional year of age. 37. Household size seems to have a slightly negative effect, as each additional person entails on average a drop of DH 10.44 in the wage rate. This is probably due to the additional pressure that household size exerts on individuals to find employment, so that members of large households tend to be less selective. Another potential explanation could be that more educated people who presumably earn more, tend to have less children. However, the absolute size of the effect is not very large. Notice that gender is not significant in this equation. This is probably the result of multicollinearity with other included variables. For example, gender effects might be captured through the sectoral variables, the professional variables, and the education variables, as men and women have distinctly different characteristics with respect to these variables. 38. Lastly in this set of variables, it is shown that non-permanent workers (including only seasonal and occasional workers) are paid significantly less than permanent workers. However, the latter category is defined in such a way that very few workers are regarded as non permanent. The criterion is: having worked regularly in the present job for at least six months. Only less than 6 percent of the employed people were classified as non-permanent in 1991. 39. Regarding the regional characteristics, we find that the factors affecting earnings do not vary a great deal by region. The reference regions are North-Centre (including Fes), South-Centre (including Meknes), and Oriental (including Oujda), among which the wage distribution seems to be quite homogeneous. Wages seems to be lower than this (on average and ceteris panibus) in the Tensift region (including Marrakech), by DH 58.07 per month. But the attached t-ratio shows that this effect is borderline significant (less than 10 percent). Wage rates are higher on average (ceteris paribus) in the North-West region (including Rabat and Kenitra), by DH 83.23 per month, and in the Centre (including Casablanca), by DH 122.09 per month. The South region (including Agadir and Laayoune) seems to pay the highest wages (ceteris paribus), with an extra DH 172.34 per month. However these results seem slightly spurious, as the cities involved are fairly small compared to the urban centres of the other regions. On the other hand, this might reveal that firms have to pay a premium in order to attract workers in this peripheral region. Annex III-A  - 13 - 40. The next set of variables is concerned with sectoral effects. The reference sectors are (2) Mining and Quarrying, (7) Electricity, Gas, and Water, and (16) General Administration, where earnings seem to be homogeneously highest. Then, (5) Other Manufacturing Industries does not fall far behind, paying on average (and ceteris paribus) DH 244.73 less. Then, we find a group of sectors including (6) Repair, (9) Wholesale and Retail Trade, and (11) Transport and Telecommunication, which pay respectively 325.11, 358.73, and DH 301.30 less than the top sectors. Not surprisingly, (1) Agriculture, Forestry and Fishing, and (13) Personal and Domestic Services are the bottom sectors, with an average pay that is respectively DH 636.86 and DH 573.87 less per month than the top sectors. Notice that (10) Restaurant and Hotels and (14) Collective Services are not much above this bottom group, paying on average (and ceteris paribus) DH 465.23 and DH 486.89 less than the top sectors, respectively. 41. It is striking to see that the classification by profession, that is captured by the next set of variables, seems to play less of a differentiating role than the other set of variables included in the equation. In other words, once personal characteristics, the economic sector, and the educational achievement variables are controlled for, we find that (2) Commercial Personnel, (4) Services Specialized Workers, (5) Agriculture, Forestry and Fishing, and (6) Non Agricultural Workers, are not treated differently by the labor market. Only (1) Scientific and Liberal Professions (that include Lawyers, Medical Doctors, etc.) and (3) Administrative Personnel earn significantly more (on average and cetenis paribus), with an extra DH 1150.55 and DH 1048.89 per month, respectively. This suggests that there is some mobility between the professional categories, except for the last two, which seem to benefit from significant barriers to entry, probably related to educational requirements. 42. Lastly, we turn our attention to the effects of educational achievement, here captured by the highest degree held by the person. The reference category is no diploma at all, but it turns out that the Certificate of Primary Education does not make much of a difference (given the existing wage regulations). What is more surprising is the fact that (6) Grandes Ecoles and Medical School are not significant. It is highly probable that the failure of this variable to make a significant impact on the equation is due to its quasi-collinearity with the variable (1) Scientific and Liberal Professions, which includes medical doctors and qualified engineers, etc. Then, it is shown that the Certificate of Secondary Education, that is acquired after four years of secondary education, makes a slight contribution to the average monthly earning (DH 157.36), whereas the Baccalaureat, which is passed after seven years of secondary education, has a more noticeable impact (DH 557.22). Technical Degrees, that usually correspond to two or three years of post-secondary education, have a significant effect, while University diplomas (medical excepted), are the road to the top paying jobs. Annex III-A  - 14 - Table 4: Urban Sector Earnings Function: Private Sector (1993) Dependent Variable: Monthly Wage Earnings Coefficient 1-Ratio Constant 430.49 (2.80) Age 41.34 (8.77) (Age)2 -0.41 (7.13) Married 330.42 (12.98) Household Price -10.44 (3.09) Number of Days Worked in the Month 5.72 (2.91) Non Permanent -213.13 (4.39) South 172.34 (3.70) Tensift -58.07 (1.61) Center 122.09 (4.26) North-Center n.s. North-West 83.23 (2.39) (1) Agriculture, Forestry and Fishing -636.86 (5.23) (2) Mining and Quarrying n.s. (3) Food, Drink and Tobacco -418.68 (3.50) (4) Textile, Clothing, Shoes and Leather -406.23 (3.54) (5) Other Manufacturing Industries -244.73 (2.13) (6) Repair -325.11 (2.62) (7) Electricity, Gas and Water n.s. (8) Building and Public Works -416.83 (3.60) (9) Wholesale and Retail Trade -358.73 (3.08) (10) Restaurants and Hotels -465.23 (3.91) (11) Transport and Communications -301.30 (2.48) (12) Banks, Insurance, and Other Services to Firms -448.62 (3.51) (13) Personal and Domestic Services -573.87 (4.80) (14) Collective Services -486.89 (3.77) (Health, Cult, Teaching, Promotion Nationale) (1) Scientific and Liberal Profession 1,150.55 (16.92) (2) Commercial Personnel n.s. (3) Administrative Personnel 1,048.89 (26.20) (4) Services Specialized Workers n.s. (5) Agriculture, Forestry, Fishing, Hunting, etc. n.s. (6) Non Agricultural Worker n.s. (2) Certificate of Primary Education n.s. (3) Certificate of Secondary Education 157.36 (3.65) (4) Baccalaur&at 557.22 (6.01) (5) University (Except Medical) 1,463.62 (12.67) (6) Grandes Ecoles and Medical School n.s. (7) Technical Degree (Cadre Moyen) 767.23 (12.93) Number of Observations = 12,670; R2 = 0.25; F(29; 12,640) = 146.34 Annex III-A  - 15 - B. Impact of the Minimum Wage and Social Charges on Urban Private Sector Wages 43. In Morocco, all firms are required to apply the Minimum Wage Law, that was enacted by the Dahir dated 18 June 1936, in the wake of the election of the "Front Populaire" in France.3'' There is a different minimum for agricultural wages, called the SMAG (Salaire minimum agricole garanti), and for other wages, called the SMIG (Salaire minimum interprofessionnel garanti). There are provisions for paying reduced rates to younger workers, according to the following schedule: Percentage cut: SMIG: SMAG: 14-15 years: 50 %; 12-15 years: 50 %; 15-16 years: 40 %; 15-18 years: 30 %. 16-17 years: 30 %; 17-18 years: 20 %. 44. Assessing the actual impact of the minimum wage on the distribution of urban wage rates, and on urban employment is not straightforward. Charts A. 1 through A.5 show the relevant part of the wage distribution for the private urban sector, for 1989 through 1993.31' People working for the government or for local administration, for the public sector or the semi-public sector, are not covered by the CNSS. In 1993, 820,866 persons were registered with the CNSS, out of an estimated number of about 1.2 million wage earners in the relevant sectors, as estimated from the Labor Force Survey.32' The representativeness of these data may be an issue. Firms have an obvious incentive to under-report the wages they pay, as the CNSS collects a payroll tax on the basis of these declared amounts. However, regular inspections are performed on the basis of a scientifically determined sample frame, and it is believed that fraud is not wide-spread at least in the formal sector. However, no quantitative data allows for an evaluation of the weight of the "informal sector", where this type of regulation might be evaded. The gap between the number of wage earners in the CNSS files and the number estimated in the Labor Force Survey provides an upper bound on this number. The rate of the payroll tax or "social charge" is 18.6 percent for 1995, to which 1.6 percent must be added as a levy in favor of vocational training, and various types of insurance (accident, etc.), which are not compulsory. Thus, the wedge between the cost of labor and the take home pay may vary between 20.2 percent and 35 percent. 45. Charts A-1 through A-5 show the wage distribution for these workers in various years. Monthly wage classes are distinguished on the horizontal axis of Charts A. 1 through A.5, that show the impact of the Minimum Wage on the private sector wage distribution. Each class represents a wage bracket of 30/ See Benhayoun and Bazen, 1991. 31/ The data come from the CNSS (Caisse nationale de sicurite sociale), the main private sector social security organization and not from the Direction de la statistique. 32/ The Labor Force Survey is conducted by the Ministry of Plan, Direction de la Statistique. Annex III-A  - 16 - DH 500 per month. For ChanAl: PrwaeStkisthWDEtMubs example, class I refers to ,n,>d,d,^w of Fpeople earning between GMOlt ~ \ Tdal 1989 DH l and 500 per month, ,m < \while class 5 refers to am / / \\people earning between gmce //u., DH 2,001 and 2,500 per month. In order to correct for the different number of hours worked on average ________________ per month by the different 2 M s,c a 4 5 a 7 a 9 10 individuals, we do not 1S0O 2OE 000 aooo00o 50oo present their number in each wage class. The vertical axis of these graphs measures the number of hours worked by individuals in the different wage classes. We restrict the analysis to the people earning less than DH 5,001 per month on average, because the wage classes above that threshold do not contain any significant share of the hours worked. 46. The distributions are quite evidently uni-modal, in the neighborhood of the SMIG. The latter has evolved during this period according to the following schedule: 1-1-1988: DH 898.56/Month, 1-5-1989: DH 988.00/Month, 1-5-1990: DH 1,085.76/Month, 1-1-1991: DH 1,248.00/Month, 1-5-1992: DH 1,373.00/Month, 1-5-1993: DH 1,510.00/Month. 47. Hence, the SMIG is at the upper end of wage ChartA2 UrbanPivaeSetoWaDis5ibutior class 2 in 1989, and keeps Nanbo ofHowotl moving within class 3 in the following years. It ED \ Tota 1990 reaches the top of class 3 A \ in 1994. The series clearly / show that most wage earners cluster at, or just above, the level of the 2o / Women minimum wage. In 1 -, particular, it is interesting to notice how wage class 2 I 02 S.IG O 2 e O (DH 501-1000), which contains the minimum wage in 1989, and up to May 1990, sees a significant reduction in the number of hours worked as the SMIG jumps into the next class. It is close to the modal class in 1989, whereas very few people fall into this wage class after 1990. Moreover, by observing how the distribution changes during the period under study, it can be seen that the SMIG seems to push to the right-hand side not only the wage level of the "Smigards" (those earning just the SMIG), but also the wage rates of the people Annex rn-A  - 17 - belonging to the two wage classes immediately above. This is especially visible in the case of class 4, which contains very few people in 1989, and gets larger as time passes. In 1993, it is very close to the mode of the distribution. 48. Changes in the level of the minimum wage are determined by the government. A special Committee, the Commission centrale des prix et des salaires, was created in 1959 in order to advise the government on this matter. However, it has not ChartkA UrbanPrrvate Sector Weg)i:twibution convened since 1961. Chart [NLmber ofHoLaeJ A.6 shows how the SMIG has evolved over 1961-1993, ota1 in relation to the average wage, using CNSS data. As 4>- / X\ the SMIG does not __ /9en \ Xnecessarily change on the first of January each year, 2eoo j/ 4 the series represents in fact / Z Wmen the mean value of the SMIG e________________________________________________LE over the year. This graph 1 2 sa 4 5 7 a S 1C showsthatthe SMIG did not change at all during the 1960s. It started to grow exponentially like the average wage only in 1971. A more informative comparison of the two series can be done from chart A.7, where both wage series have been deflated by the CPI. It can be seen that the real SMIG fell consistently up to 1978, after which it started to increase fast. These curves suggest that there has not been much correlation between the two series except over the last few years, starting in ChartA4 UrbanPFivoteSectorWageDistribution 1988. The gap between 70 (Ntimib of Hotzsi the two seems to be shrinking in the 1980s and Om Total 1952 the early 1990s. mm 40M~~~~e .me WAGE e- : 4 I I iC 1 2 sun a 4 b 6 7 A S 10 10ee 2mo MMo *00 5mm Annex 111-A  - 18 - 49. This impression is confirmed by looking at Chart A.8, which represents the percentage gap (difference in log) between the average wage and the SMIG. The series clearly shows a steeply rising section up to 1978, and a steeply falling ChtA& UrbanivaSedorWageDisbibLtior one afterwards. The (Numbe ol Howrsi percentage gap has e / \ t h u s f a I l e n dramatically. Hence, we may expect that the e - / /r M \\ \ impact of the minimum 2Ke- /,' '\ wage on the average Neee / // \sX wage has differed , 7/' ----- --.'.. significantly in the ee . < / Wemen -. ~ ifslE three periods: during .......... ..........---LE/EL the 1960s when it did SMG 2 4 6 a* 7 * * 10 not change, during the 1970s and early 1980s when it was far away from the mean of the wage distribution, and during the late 1980s-early 1990s, when its value became closer to the average wage. 50. To investigate the relationship between the SMIG and the average wage further, some econometric analysis was conducted. The preferred equations, estimated over 1971-1993, are: Growth Rate of the SMIG: GSMIG = - 0.28 GSMIG(-I) + 2.26 GAVW(-2) - 0.80 GCPI(-2), (1) (1.69) (4.44) (1.61) N = 23, R2 = 0.39, F(4,19) = 6.49, D.W. = 2.09, LMF(6,17) = 0.78, ChowF(5,18) = 1.54. Growth Rate of the Average Wage: GAVW = 0.46 + 0.47 GAVW(-I) + 0.41 GAVW(-2) (1.89) (2.04) (2.31) + 0.55 GCPI - 0.45 GCPI(-I) - 0.28 RAVW(-I) (2) (3.84) (2.50) (1.81) N = 23, R2 = 0.66, F(6,17) = 6.53, D.W. = 2.30, LMF(8,15) = 1.24, ChowF(5,18) = 1.40.33' 331 Most of the usual tests are presented after the equations. But, as the lagged endogenous variable is included, the standard D. W. test might be biased. This is taken care of by using the LM test of residuals autocorrelation up to the second order, which confirms that no autocorrelaon is present. Then, the Chow test of parameters constancy has been used for checking whether the structure of the equations have changed since 1989. No structural break has been found. Hence, no particular econometric problem seems to be present, so that we can comment on the results. Annex III-A  - 19 - 51. I n t h e s e Chast A&Morthy esectSorAtwt equations, GSMIG Om and SMIG represents the growth rate of the SMIG, 2/ while GAVW is the 2W -Av-%%G growth rate of the is average wage rate. The rate of inflation is low represented by GCPI, SW the growth rate of the a . . . . .. CPI (Cost of Living * a *, *7 'A 73 n7s n 79 73 aK a tr n #tK Index). RAVW is the cNn real value of the average wage rate, which enters the GAVWequation (A.4) lagged once; It thus plays the part of an error- correction term in the relationship between the average wage rate and the CPI. 52. The first equation (1) suggests that the SMIG is revised with a view to catching up with rises in the average wage, with a two year lag. Moreover, the growth rate of the SMIG lagged once enters with a negative sign, suggesting that the government tends to avoid granting two strong SMIG increases in a row, alternating years of fast and slow growth of the minimum wage. It is noteworthy that the estimated elasticity of the SMIG with respect to the average wage is very high, in accordance with the fact that the SMIG has been catching up very quickly with the average wage since 1978. This suggests that the growth of the SMIG should be slowed down for a while, as the current course of action would result in the SMIG catching up "too fast" on the average wage, thus crushing the lower tail of the wage distribution, and creating more unemployment.' 53. To sum up, Ch.tA.7Fv-"i ....... "~ ..... S^ ,;(..~", ..this analysis of the 1.73 dynamic impact of the minimum wage on the average wage rate has .US revealed some _.Ce interesting aspects of wage determination in 0.7ei - <I SMIG ~ ~ ~ ~ ~ t Morocco. Whereas 0.E observation of the latter over 1989-1993 merely indicated a el 81 Ks 67 as 71 79 75 77 73 e1 a K6 87 Ko o 91 I_______________________________________________________ strong positive correlation between the SMIG and the average wage, this analysis has revealed that it is the SMIG which is adjusted in accordance with changes in the average wage. 34/ See para. 69 on the effects of the SMIG on unemployment. Annex IH-A  - 20 - 54. In addition, minimum wage ChItAt StdeWagDis tibuijonlS 13 adjustments seem to NbdP) over-compensate to changes in the average wage: the gap between the two has - been shrinking consistently since so/ 1978. There is some 20 evidence that the lo' minimum wage has O1 2 SM: S 4 6 a 7 10 started to have _ significant effects on the wage distribution at the end of our period of analysis suggesting that it may be desirable to arrest the continuous growth of the SMIG that occurred since 1978. 55. In order to get a better understanding of the relationship between these two series, some econometric analysis was done. The estimation was restricted to the period 1971-1993, and a strong positive relationship was found between the average wage and the SMIG. Though an error-correction representation was tried it was not significant.35' This might be due to a structural break, as seems likely in view of the change in the behavior of the SMIG series after 1978. However, the sample size is probably too small for this diagnosis to be made very confidently. C. Earnings in the Government Sector and the Public and Semi-Public Urban Sectors 56. The pay policy pursued in the public sector seems quite different from that of the private sector. This comes out clearly from the distribution of the state employees by wage class presented in Chart A. 8. Here there is no correction for the number of hours worked over the relevant period, so that the height of the distribution measures the number of people in each wage class. The wage classes refer to annual earnings in 1993, and each class groups people within a DH 5,000 per year bracket. In particular, class 3 comprises all those individuals earning between DH 15,000 and 20,000 per year. This is the wage class that contains the SMIG. We have truncated the distribution like we did for the private sector, at a level corresponding to DH 5,000 per month. However, contrary to the case of the private sector, this leaves out of the picture a non-negligible share of the relevant employees. For example, there are 27,613 people earning between DH 60,000 and 70,000 per year, and 8,807 earning between DH 70,000 and DH 80,000 per year, compared to the total number of state employees: 372,906. 57. It appears quite clearly that the distribution is much less concentrated in the neighborhood of the SMIG than in the private sector. On the contrary, the distribution is quite flat, and is bimodal. We find a first mode at class 3, which contains the SMIG, suggesting that the latter is enforced in the government 35/ However, as these two series are non stationary, this regression line might be spurious. Over such a small sample, a full blown co-integration analysis was not suitable (Barnerjee, et al., 1992) Annex III-A  - 21 - sector, and constrains the lower tail of the distribution from below, producing a small "spike". However, there is a second mode, comprised of nearly twice as many people, in class 7 (between DH 35,000 and 40,000 per year). This wage bracket corresponds to more than twice the SMIG. 58. Table 5 presents the results of an estimated earnings function for the entire public and semi-public urban sector, similar to the one presented above for the private sector. This equation helps to analyze the different characteristics that affect the level of this wage income. It is based on the data from the 1993 Urban Labor Force Survey. The dependent variable is again the monthly wage income of the employed person interviewed. The sample consists of 9,706 employees of the public and semi-public sectors. The equation explains about 25 percent of the variance of monthly earnings. The F-test shows that this equation is significantly different from a constant mean value. 59. The first set of characteristics included are personal data. The age variable shows that seniority matters in the Moroccan public and semi-public sectors, even more than it does in the private sector, as both age and its square are more significant in this equation. The quadratic term has a negative sign, showing that the impact of seniority decreases as the person gets older. However the impact of age is stronger than for the private sector equation, and does not become negative at any age. The estimated parameters imply that, ceteris paribus, a wage earner earns additional DH 77.09 per year of age when he (or she) is twenty, 69.19 when he (or she) is thirty, 61.29 when he (or she) is forty, and 53.39 when he (or she) is fifty.6' 60. Household size has a slightly negative effect, as it does in the private sector, as each additional person in the household entails on average a drop in the wage rate by DH 31.21. The size of the effect is not very large, but larger than that estimated for the private sector. The number of active household members seems to have a marginally significant positive effect, which offsets to some extent the impact of household size, suggesting that it is really the fact of having children that entails a negative impact. This might be correlated with the cultural background of the person, as long-term urban families have less children than newly migrating families.37' Notice that gender is significant in this equation, contrary to the result found in the private sector equation, as males earn cetenis panibus DH 413.52 more per month. This could be due to the fact the fact that sectoral and professional variables are not as significant here as in the private sector equation, supporting the view that the result found in the private sector equation was most probably due to multicollinearity with other included variables. 61. The results show that regional dispersion in earnings is not extremely large. The reference regions are North-Centre (including Fes) and Oriental (including Oujda), among which the wage distribution seems to be quite homogenous. Wages seems to be lower than this (on average and ceteris paribus) in the South-Centre region (including Meknes), by DH 73.39 per month, in the Tensift region (including Marrakech), by DH 63.69 per month, and as well, quite surprisingly, in the North-West region (including Rabat and Kenitra), by DH 55.38 per month. However, the attached t-ratio shows that this effect is borderline significant (less than 10 percent). Wage rates are higher on average (ceteris panibus) in the Centre (including Casablanca), by DH 126.01 per month. The South region (including Agadir and 36/ The age at which the derivative of this equation with respect to age becomes zero is 117.58, outside the relevant range. 37/ This may also reflect the fact that poorer (as measured by wages) people tend to have larger households. Annex III-A  - 22 - Laayoune) also pays the highest wages (ceteris paribus), with an extra DH 249.18 per month. As suggested when discussing the private sector wage equation, this result may be spurious, as the cities involved are fairly small compared to the urban centres of the other regions. Nevertheless, this confirms that employers probably have to pay a premium in order to attract workers in this peripheral region. 62. The next set of variables is concerned with sectoral effects. They are in general much less significant than for the private sector, showing that pay policy is much more homogenous in the public and semi-public sectors. Only three sectors are significantly different from the others. We first find two sectors that belong more to the semi-public sector than to the public sector proper, namely (5) Other Manufacturing Industries and (12) Banks, Insurance, and Other Services to Firms, which pay respectively DH 292.62 per month and DH 699.93 per month more than the reference sectors. On the contrary, people employed in the (14) Collective Services sector (Health, Education, Promotion nationale) are paid much less, with DH 432.50 per month less than the reference sectors. 63. Employees in the (5) Agriculture, Forestry and Fishing sector are paid DH 252.48 per month less than employees in the reference sectors. Those employed as (3) Administrative Personnel earn significantly more (on average and ceteris paribus), with an extra DH 151.44 per month. 64. Lastly, the effects of educational achievement, captured again by the highest degree held by the individual, has an even more significant effect on earnings than in the private sector. The reference category is no diploma at all. Contrary to what was found for the private sector the Certificate of Primary Education does make a difference in the public and semi-public sector. The Certificate of Secondary Education, that is acquired after four years of secondary education, makes an important contribution to average monthly earnings (DH 582.89), and the Baccalaureat, has a sizeable impact (DH 1078.23). (5) University (Except Medical) and (6) Grandes Ecoles and Medical School are the top set of diplomas in terms of monthly pay, earning respectively DH 2,591.10 and DH 1, 891.60 more than the reference group. (7) Technical Degrees, that usually correspond to two or three years of post- secondary education, have a significant effect, with an extra DH 1221.06 per month, while (8) Vocational Degrees, which are not significant in the private sector equation, entail a significant extra pay here (DH 531.68 per month). Annex Ill-A  - 23 - Table 5: Urban Sector Earnings Function: Public and Semi-Public Sectors (1993) Dependent Variable: Monthly Wage Earnings Coefficient T-Ratio Constant 1,243.32 (5.29) Age 92.89 (8.53) (Age)2 -0.79 (6.28) Male 413.52 (10.55) Married 274.18 (6.73) Household Size -31.21 (4.70) Number of Active Household Members 20.74 (1.17) Number of Days Worked in the Month 1.79 (0.72) South 249.18 (3.95) Tensift -63.69 (1. 11) Centre 126.01 (2.77) Centre-South -73.39 (1.17) North-West -55.38 (1.17) (5) Other Manufacturing Industries 292.62 (2.29) (12) Banks, Insurance, and Other Services to Firms 699.93 (7.35) (14) Collective Services -432.50 (10.45) (Health, Cult, Teaching, Promotion nationale) (3) Administrative Personnel 151.44 (4.06) (5) Agriculture, Forestry, Fishing, Hunting, etc. -252.48 (2.18) (2) Certificate of Primary Education 239.22 (4.85) (3) Certificate of Secondary Education 582.89 (11.00) (4) Baccalaureat 1,078.23 (13.83) (5) University (Except Medical) 2,591.10 (37.46) (6) Grandes Ecoles and Medical School 1,891.60 (35.40) (7) Technical Degree (cadre moyen) 1,221.06 (21.45) (8) Vocational Degree (C.A.P.) 531.68 (5.29) Number of Observations = 9,706; R2 = 0.25; F(24; 9,681) = 134.10 D. Urban Unemployment 65. Chart A.9 represents the rate of unemployment in the urban sector over the period 1979-1993. The Urban Labor Force Survey was only started on a yearly basis in 1984, so that the data points for the previous dates do not correspond perfectly with the post 1984 data. The year 1983 was missing, and replaced by the sample mean. The figure for 1982 comes from the census that took place that year.35" 66. Due to the lack of homogeneity of the sample noted above, it is difficult to compare the results regarding the difference in the average rate of urban unemployment before and after 1983. However, restricting our comments to the period after 1983, the rate of unemployment shows a slight upward trend (despite the increasing employment in the public sector). The empirical evidence suggests a positive 38/ All these data are published by the Direction de la Statistique. Annex III-A  - 24 - relationship between the rate of urban unemployment and the real value of the SMIG (deflated by the CPI). This consists of a scatter diagram, with upward sloping simple regression straight line. This suggests that the real value of the SMIG positively affects the rate of unemployment.39' Despite the possible heterogeneity of the data before and after 1983, and despite the number of data points, we have attempted to use econometric analysis to shed some light on the determination of this rate of unemployment. After an extensive specification search "a la Hendry", the preferred equation is: U = 0.28 - 0.16 GAP + 0.30 Dinfl, (3) (11.10) (5.43) (2.38) N = 15, R2 = 0.73, F(3,12) = 16.35, D.W. = 1.47, LMF(4,11) = 0.27, B.J. = 1.14, WVrte-F(6, 9) = 1.71, RESET(1)F(4,11) = 3.95, Chow[Split 1984] = 3.15. 67. The tests presented after equation (3) are fairly standard. N is the number of data points; R2 is the usual coefficient of determination; F(3, 12) is the joint test that both coefficients of the explanatory variables are nil, which rejects this assumption; D. W. is the standard test for serial independence of the residuals, which falls just below the d, critical value at 5 percent (1.54). However auto-correlation is rejected by the lMF(4,11) test, which is the Lagrange Multiplier test for first-order auto-correlation. B.J. is the Bera-Jarque normality test, which does not reject the assumption that the residuals are drawn from a normal distribution. Mite- CJIARTA9: U,bP. .n Un.pIyn.nt(%) F(6, 9) is White's heteroscedasticity test 018 A (White, 1980), which 776 / \ ~ / does not reject the 08/ /' assumption of 06 /- homoscedasticity, 019 / while 012 RESET(1)F(4,11) is 010 the RESET test for 0 09 . . . . . functional form 79 so 1I 82 J3 S- S6 go 87 88 89 9g 91 92 97 s 9gg ndI9,dreD58sbq-{lca3mgIvb) ~which amounts to including the square of the fitted value in the equation. It falls very close to the 5 percent threshold, and would in fact reject if we accepted a 10 percent type 1 error. Finally, Chow [Split 19841 is Chow's test of parameter constancy, performed here by splitting the sample in 1984, in order to test whether the data generation process changed at that date, 39/ However, the rate of unemployment is probably affected as well by other variables, so that the simple regression analysis might lead to a biased result. Annex III-A  - 25 - which corresponds, as stated above, to the starting date of the yearly labor force survey. Quite luckily, it falls between the 5 percent and the 10 percent thresholds, thus failing to reject the assumption of parameter constancy. 68. These tests did not detect any particular econometric problem, so the specification was retained. U represents the rate of urban unemployment, that is explained by this equation. After experimenting with a short unrestricted distributed lag effect of inflation, somewhat in the spirit of a Phillips curve, the best specification appeared to be the inclusion of Dinfl, the first difference in the rate of inflation. This might be interpreted as a measure of inflation surprise. Notice that it affects the rate of unemployment in a positive way, a bit like in Nugent and Glezakos (1982). These authors explain such a positive impact by assuming that a positive inflation surprise means that employers have under-estimated the price of their product, when deciding on the output level on the basis of their expectations. This assumes that there is a positive production lag between the decision to produce and the delivery of the goods. Had they made no expectation error, the employers would have planned for a higher output level, and thus a higher level of employment. Another potential explanation is that a positive inflation surprise, in a predominantly agrarian economy like the Moroccan one, is generally the result of a below-normal supply of food, entailing a supply-driven positive inflation shock. Then, crop failure is an incentive for more rural workers to migrate, maybe temporarily, to look for a job in the urban sector. Like most modem Phillips-curves, equation (3) does not leave any scope for a systematic Keynesian policy towards the reduction of urban unemployment: only inflation surprises have any impact on the rate of unemployment, and this impact turns out to be positive. This is a very strong argument in favor of macroeconomic stability. 69. GAP is the percentage difference between the average wage rate in the private urban sector and the SMIG. As discussed above, (see Chart A 10) this is a measure of how much the latter is binding. The smaller the gap, the more is the SMIG compressing the lower tail of the wage distribution. We thus expect a negative relationship between unemployment and this CHARTA10 P.-nIa.G.p.lw-nth.A-...g. W.9. ..ndth. SMIC gap, with a larger gap meaning that the SMIG is 1.10 less binding, leaving IMl ,-\ . . \.- , more room for cLoa zemployment in the lower a X ' .8 tail of the wage distribution. This equation suggests that this a. ..,.,.,. .,, ,, ,, ....'EAR effect is very significant in 5 67 71 3 m 78 m 9 a inthe case of Morocco. Hence, this equation suggests that the level of the SMIG, in relation to the level of the average wage, is a major determinant of the rate of unemployment in the urban sector. As seen above, this percentage gap has been decreasing steadily since the late seventies, reaching in 1993 a low level comparable to that of the early 1960s. It can thus safely be recommended that the increase in the SMIG relative to the average wage rate should be halted, and Annex III-A  - 26 - reversed to some extent, in order to restore the appropriate level of this gap. 7his would be a major impulse to urban employment and to the growth of private sector output. 70. Notice that this equation points out to a very strong "structural" rate of urban unemployment. To see this, assume a steady rate of inflation (Dinfl = 0), and assume a 100 percent gap between the two wage rates (GAP = 1), as was prevailing in the late seventies; then, the high level of the intercept implies that a 12 percent rate of urban unemployment would follow. Though the number of data points are small, notice that the standard error of the estimated intercept is very small in this equation, suggesting that this estimate is not as fragile as the small number of data points might suggest. 71. We can now go one step further in the analysis of urban unemployment, looking into the microeconomic characteristics of the unemployed people. Table 7 presents the estimated coefficients of a probit equation of the incidence of unemployment. The latent variable is the propensity of the person to be unemployed, and the observed dependent variable is a dummy taking the value 1 when the person interviewed is unemployed, and zero otherwise. The data are extracted from the Urban Labor Force Survey 1993 done by the Direction de la Statistique. 49 000 observations have been picked up randomly from the survey data, after clustering by sector of employment. Each coefficient is proportional to the impact of the corresponding variable on the probability of the person being unemployed. 72. The first set of variables concerns the age of the person. We can see that except for the under 15 and the above 60, age is not an important sorting variable. Between 15 and 60, all people have (ceteis paribus), about the same probability of being unemployed. 7hen, we find that male workers are less likely than female ones to be unemployed. Household size a factor that is positively related to the probability of being unemployed. 73. Then come the variables concerning the sector of present or previous employment (for the unemployed). Here, the reference is (16) Undefined Activity and Unemployed Seeking a First Job. It is thus not so surprising to find that all people with some work experience have a lower probability of being unemployed. However, there are differences among the sectors. We can rank them according to the (absolute) size of the estimated coefficient; the higher this size, the lower the probability of being unemployed. Table 8 reproduces the relevant information. We find for example that working in (7) Electricity, Gas and Water reduces the probability of being unemployed by 52.8 percent more than working in (3) Food, Drink and Tobacco. Some of the rankings are a bit surprising, like that of (12) Banks, Insurance, and Other Services to Firms, which is close to the bottom of the list, while (4) Textile, Clothing, Shoes and Leather is in the top three group. 74. We then have the group of variables representing professions. The reference is: (8) Other Persons Not Classified in Other Categories. We can see that only two professions seem to make a significant difference, namely (4) Services Specialized Workers and (6) Non Agricultural Worker, which do not reduce the probability of being unemployed as much as the other four. Notice that (1) Scientific and Liberal Profession is the best profession in terms of the reduction in the probability of being unemployed. 75. Total Labor Cost and Unemployment. In order to get an order of magnitude of the percentage reduction in the level of employment that is implied by the wedge between the cost of labor and the level of take home pay, a simple simulation is performed. Assume that the demand for labor is given by a Annex IE-A  - 27 - downward sloping function L(w), such that L' (w) < 0, where w is the cost of labor to the firm. Assume in addition that the minimum wage is fixed, and determines the take home pay for workers. Denote as "s" this wage rate. Then we can compare the outcome when there is a wedge, x, between w and s; i.e. w=(I +x)s, so that when x is zero w=s. For that purpose, a first order Taylor expansion of L(s) is used: L(1 +x)s = L(s) + x s L'(s). (4) Now, it is straightforward to rearrange (4) as: L(1 +x)s) - L(s) = -xeL, (5) L(s) where El = - s L' (s)/L(s) is the elasticity of L(s) with respect to s, taken positively. If eL were known, it would be very simple to calculate the percentage loss of employment entailed by different levels of the wedge x, using equation (5). However, it is not known and a range of reasonable values are presented in Table 6. This table shows the percentage loss of employment resulting for a range of percentage wedges between the cost of labor and the level of the take home pay, assumed fixed by the minimum wage, for three plausible values of the elasticity of the demand for labor with respect to the cost of labor. For example, if the level of the wedge is 25 percent, a fairly realistic figure for the private sector in urban Morocco, and if the elasticity of the demand for labor is 0.6, then the resulting percentage of employment in the urban private sector that is lost due to the wedge between take home pay and the cost of labor resulting from social charges and other levies on employment in Morocco, compared with a situation without any such wedge, falls somewhere between 8 percent and 28 percent. For lack of precise estimates of the average wedge in the sector and of the elasticity of labor demand, it seems that a reasonable number would lie somewhere about 15 percent (+ or - 3 percent). Table 6: Simulated Percentage Loss of Employment for Different Percentage Wedges Elasticity eL , 4.0 0.6 0.8 Wedge x: 0.1 4 6 8 0.15 6 9 12 0.20 : 8 12 16 0.25 : 10 15 20 0.30 : 12 18 24 0.35 : 14 21 28 Source: Author's calculation. Annex 11l-A  - 28 - 76. The next set of variables in the equation presented in table 7 is educational achievement, a crucial piece of information regarding unemployment. The reference here is initially no diploma at all. But it turns out that the Certificates of Primary or Secondary Education do not make any significant contribution to reducing the probability of being unemployed given the current regulatory environment. Even the Baccalaurdat and Vocational Degrees do not have a significant impact, although the latter come very close to having a statistically significant effect (at the 15 percent level). Higher education is what really makes unemployment less likely, listed in order of effectiveness: (7) Technical Degree (Cadre Moyen), (5) University (Except Medical), and (6) Grandes Ecoles and Medical School; these are significant reducers of the risk of unemployment. Technical Degree (Cadre Moyen) is by far the top performer, suggesting that there is a shortage of middle rank management and hierarchical employees in Morocco. Table 7: Urban Unemployment - Probit Function (1993) Dependent Variable: Being Unemployed Coefficient T-Ratio Constant 1.91 18.07 Less than 15 Years of Age 0.25 4.06 15-24 Years 0.54 10.50 25-55 Years 0.60 12.38 45-59 Years 0.55 11.70 60 Years and Above 0.26 5.27 Male -0.23 9.17 Married -0.26 7.17 Widowed -0.20 2.15 Divorced -0.80 1.12 Household Size 0.01 2.91 (1) Agriculture, Forestry and Fishing -2.51 14.30 (2) Mining and Quarrying -2.87 17.67 (3) Food, Drink and Tobacco -2.46 22.45 (4) Textile, Clothing, Shoes and Leather -3.06 31.88 (5) Other Manufacturing Industries -2.84 28.95 (6) Repair -3.05 26.39 (7) Electricity, Gas and Water -3.76 10.12 (8) Building and Public Works -2.67 26.86 (9) Wholesale and Retail Trade -2.99 26.93 (10) Restaurants and Hotels -2.66 23.13 (11) Transport and Communications -2.99 26.98 (12) Banks, Insurance, and Other Services to Firms -2.67 22.02 (13) Personal and Domestic Services -2.87 27.94 (14) Collective Services -2.87 25.93 (Wealth, Cult, Teaching, Promotion Nationale) (15) General Administration -3.13 30.09 (1) Scientific and Liberal Profession -0.72 5.56 (2) Commercial Staff -0.60 4.76 (3) Administrative Staff -0.59 5.44 (4) Services Specialized Workers -0.44 4.01 (5) Agriculture, Forestry, Fishing, Hunting, etc. -0.64 3.42 (6) Non Agricultural Worker -0.45 4.31 (4) Baccalaureat -0.21 0.39 (5) University (Except Medical) -0.47 3.81 (6) Grandes Ecoles and Medical School -0.23 2.02 (7) Technical Degree (Cadre Moyen) -1.25 5.86 -0.11 1.43 Number of Observations = 23160; Pseudo R2 = 0.45; Chi2(36) = 8399.47. Annex III-A  - 29 - Table 8: Ranking of Sectors by Probability of Being Unemployed (Reverse Ranking, Cetenis Panbus) (7) Electricity, Gas and Water -3.76 (10.12) (15) General Administration -3.13 (30.09) (4) Textile, Clothing, Shoes and Leather -3.06 (31.88) (6) Repair -3.05 (26.39) (9) Wholesale and Retail Trade -2.99 (26.93) (11) Transport and Communication -2.99 (26.98) (2) Mining and Quarrying -2.87 (17.67) (13) Personal and Domestic Services -2.87 (27.94) (14) Collective Services -2.87 (25.93) (Health, Cult, Teaching, Promotion Nationale) (5) Other Manufacturing Industries -2.84 (28.95) (8) Building and Public Works -2.67 (26.86) (12) Banks, Insurance, and Other Services to Firms -2.67 (22.02) (10) Restaurants and Hotels -2.66 (23.13) (1) Agriculture, Forestry and Fishing -2.51 (14.30) (3) Food, Drink and Tobacco -2.46 (22.45) Source: Table 7 E. Unemployment Duration and Matching 77. The incidence of long term unemployment is very noticeable in Morocco. The share of the unemployed for 12 months or more in the urban unemployed labor force is 67.6 percent, being somewhat lower for men (66 percent) than for women (71.1 percent). Long-term unemployment affects mainly people aged between 15 and 44 (15-24: 67.7 percent, 24-44: 69.3 percent), while workers aged 45 and more have much shorter spells of unemployment, with 39.6 percent of them staying unemployed for more than a year, while 31.2 percent of them stay unemployed less than three months. Having diplomas does not make long term unemployment less likely, among the unemployed. Table 9 shows the distribution of unemployment spells according to the highest diploma held. Only the individuals with no diploma at all seem to avoid, to some extent, the long-term unemployment trap. Table 9: Unemployment Duration by Diploma Level (1993) (Percentage of Class) Less than 3 months More than a Year No Diploma 20.1 53.8 Primary 9.7 69.5 Secondary 4.1 83.3 General Certificate of Education (Baccalaur&at) 2.9 86.3 Higher Education Diploma 6.6 74.5 "Cadres Moyens" 6.4 74.8 Vocational 6.6 74.7 Source: Direction de la Statistique. Annex HI-A  - 30 - Note: The percentage of each class with duration between 3 months and a year has been skipped, so that the rowa do not add up to 100. 78. Tables 10 and 11 present some information on the methods used by unemployed individuals to search for a job. The relative importance of the different methods differs between table 10, which refers to people who have already worked before, and table 11, which refers to people looking for a first job. The latter are (obviously) younger, and use Newspapers, Exams and Written Applications more than the former. This is especially true for the 25-34 age group, which probably contains a large number of students who have just completed their studies. However in both cases, Personal Contact with Employers, Friends and Relatives play a major part in the matching technology. Some information on *the efficiency of this search technology can be obtained from table 12, which presents the duration of unemployment according to the method used for searching. 79. Among the methods of search that concern a significant number of people, it is evident that Personal Contact with Employers is the most efficient way offinding ajob, while Newspapers, Exams and Written Applications is the worst performer. Nevertheless, the latter is only used by 41 percent less persons than the former. The Mokaf seems very efficient, but it only concerns a small fraction of the job seekers; so do Other Means. These results underline the fact that the labor market in Morocco is far away from the textbook image of a market for an anonymous and homogenous good called "labor-time". It is really personal contact that matters in this market for getting a job. Table 10: Number of Unemployed by Method of Search (Experienced Workers) Age 15-24 25-34 35-44 45-59 60+ Total Placement Bureau 63 903 87 127 0 1,180 Friends and Relatives 14,341 24,589 7,455 3,215 371 49,971 Personal Contact with Employers 71,034 103,382 29,450 14,159 1,408 219,433 Newspapers, Exams and Written 7,540 23,941 6,296 1,018 0 38,795 Applications Mokaf (Street Corner Labor Market) 3,119 3,373 2,296 1,866 313 10,967 Other Means Not Declared 708 2,340 2,030 1,667 296 7,041 883 1,150 596 106 79 2,814 Total 97,688 159,678 48,210 22,158 2,467 330,201 Source: Direction de la Statistique Annex HI-A  - 31 - Table 11: Number of Unemployed by Method of Search (Unexperienced Workers) Age 15-24 25-34 35-44 45-59 60+ Total Placement Bureau 893 2,341 191 0 0 3,425 Friends and Relatives 39,348 16,034 1,026 67 0 36,475 Personal Contact with Employers 79,413 40,455 1,497 0 0 121,365 Newspapers, Exams and Written 60,592 99,151 2,446 0 0 162,189 Appl icationsl Mokqf (Street Corner Labor Market) 673 1,373 174 0 0 2,220 Other Means 688 1,213 0 0 0 1,901 Not Declared 1,827 1,056 142 0 0 3,025 Total 183,434 161,623 5,476 67 0 350,600 Table 12: Number of Unemployed by Method of Search and Unemployment Duration (Months) Age <3 3 to S 6 to I1 12+ Not Total Declared Placement Bureau 520 298 319 3,520 (0) 4,605 (11.3) (6.5) (6.9) (76.4) 0 (100) Friends and Relatives 9,998 9,792 8,973 77,454 229 106,446 (9.4) (9.2) (8.4) (72.8) (0.2) (100) Personal Contact 51,795 42,974 37,922 207,368 739 340,798 with Employers (15.2) (12.6) (11.1) (60.8) (0.2) (100) Newspapers, Exams 7,747 14,086 15,005 164,066 80 200,984 and Written Applications (3.8) (7.0) (7.5) (81.6) (0.04) (100) Mokaf 7,629 1,269 769 3,520 0 13,187 (Streetcorner Labor Market) (57.9) (9.6) (7.7) (26.7) (0) (100) Other Means 1,947 3,255 695 3,045 0 8,942 (21.8) (36.4) (7.8) (34.0) (0) (100) Not Declared 360 781 219 1,369 3,110 5,839 (6.2) (13.4) (3.8) (23.4) (53.3) Total 79,996 72,455 63,902 460,290 4,158 680,801 Source: Dirction de la Statistique Annex III-A  - 32 - References: Azam, J.-P. (1992): "The Agricultural Minimum Wage and Wheat Production in Morocco", Journal of African Economies, 1, 171-191. Azam, J.-P. (1993): "Employeurs dominants et salaire minimum dans l'agriculture marocaine", Revue gconomique, 44, 1151-1168. Azam, J.-P. (1994.a): "Salaire minimum et production agricole au Maroc", in M. Benoit-Cattin, M. Griffon and P. Guillaumont (eds.): Economie des politiques agricoles dans les pays en developpement, Vol.3: les fondements microeconomiques, Editions de la Revue fran raise d'economie, 331-344. Azam, J. -P. (1994.b): Tax Incidence on Agriculture in Morocco (1985-1989), Middle East and North Africa Discussion Paper No.11, The World Bank: Washington, D.C. Banerjee, A., J. Dolado, J.W. Galbraith and D.F. Hendry (1993): Co-Integration, Error-Correction, and the Econometric Anatysis of Non-Stationary Data, Oxford University Press: Oxford. Benhayoun, G. and S. Bazen (1991): "Salaire minimum et structure des salaires au Maroc", in 0. Bahraoui and G.Benhayoun (eds.): Restructuration economique et developpement regional et urbain au Maroc, 187-216, Ministere du Plan: Rabat. Bentolila, S. and J. J. Dolado (1994): "Labour Flexibility and Wages: Lessons from Spain", Economic Policy, 18, 55-99. Direction de la Statistique (1992): Niveaux de vie des mOnages 1990/91, premiers resultats, Vol. 1: Rapport de synthese, Direction de la Statistique: Rabat. Direction de la Statistique (1994): Annuaire statistique du Maroc 1994, Direction de la Statistique: Rabat. Direction de la Statistique (1994): Population active urbaine 1993, Royaume du Maroc, Ministere Charge de l'Incitation de l'Economie: Rabat. Dixit, A.K. (1992): "Investment and Hysteresis", Journal of Economic Perspectives, 6, 107-132. Fields, G. S. (1994): "Changing Labor Market Conditions and Economic Development in Hong Kong, the Republic of Korea, Singapore, and Taiwan, China", World Bank Economic Review, 8, 395-414. Fields, G. S. and H. Wan Jr. (1989): "Wage-Setting Institutions and Economic Growth", World Development, 17, 1471-1483. Annex III-A  - 33 - Layard, R., S. Nickell and R. Jackman (1994): The Unemployment Crisis, Oxford University Press: Oxford. Mazumdar, D. (1983): "Segmented Labor Markets in LDCs", American Economic Review (Papers & Proceedings) 73, 896-908. Nugent, J.B. and C. Glezakos (1982): "Phillips Curves in Developing Countries: The Latin American Case", Economic Development and Cultural Change, 30, 321-334. Pascon, P. and M. Ennaji (1987): Les paysans sans terre au Maroc, Editions Toubkal: Casablanca. Saint-Paul, G. (1993): "On the Political Economy of Labor Market Flexibility", NBER Macroeconomics Annual 1993, 151-187. Salahdine, M. (1988): Les petits metiers clandestins, EDDIF Maroc: Casablanca. Salahdine, M. (1991): "The Informal Sector in Morocco: Failure of Legal Systems?, in A. L. Chickering and M. Salahdine (eds.): 7he Silent Revolution, 15-38, International Center for Economic Growth, ICS Press: San Francisco. World Bank (1994): Kingdom of Morocco: Poverty, Adjustment & Growth, Report No. 11918-MOR, The World Bank: Washington, D.C. Annex III-A   ANNEX III-B GENDER ISSUES IN THE LABOR MARKET I. LABOR FORCE EDUCATION A. Enrollments 1. The overall performance of the Moroccan education sector is lagging behind those of comparable countries, whether the latter are defined in terms of economic development or in terms of geographic proximity. Not only is the literacy rate the lowest of all LMI/MENA countries," but Moroccan schools enroll fewer proportions of each respective cohort at primary, secondary and tertiary education levels than these comparable countries (Table 1). Table 1: Education Indicators in Morocco and Comparable Countries (1990) All All LMI MENA Morocco Jordan Syria Tunisia Algeria Iran Countries Countries Literacy Rate Total 49 80 64 65 57 54 74 55 Females 38 70 51 56 45 43 68 42 Primary Net Enrollment Ratio 55 98 95 88 94 87 Primary Gross Enrollment Ratio Total 68 97 108 115 95 112 100 97 Females 55 98 102 107 88 106 98 89 Secondary Gross Enrollment Ratio Total Females 36 56 54 44 60 47 56 56 30 45 39 53 37 51 Source: World Bank, WDR 1993 and 1994; Social Indicators of Development, 1993. 2. The crudest measure of the national stock of human capital, literacy rates, indicates a significant gender gap, with the female literacy rate at 38 percent, against an average of 49 percent for the entire country (1992). Although all LMI/MENA countries share a similar pattern, the gap in Morocco ranks amongst the highest. 11 'Comparable countries" refer either to countries belonging to the same income group or to the same geographical area; the intersect of these two groups is an even better basis of comparison. The first cluster is composed of the lower- middle-income countries (LMI), the second one is the Mid-East and North-Africa region (MENA) countries, and the intersect (LMI/MENA) is the subset of 6 countries belonging to both clusters. Annex III-B  -2- 3. Access to the formal education system, which translates into the current accumulation of human capital, also reflects substantial gender differentiation, especially at the critical, i.e. basic education level. (a) Under-representation of girls at the basic-i level (enseignementfondamental, ler cycle) is much more marked in rural areas: net schooling rates for girls in urban areas hovered around 71 percent, but did not exceed 25 percent in rural areas (respective figures for boys were 77 percent and 54 percent). (b) The gap widens dramatically at the basic 2 level (enseignementfondamental, 2eme cycle), and at the secondary level, where enrollment rates stay almost unchanged in urban settings, but drop sharply in rural areas (down to 5 percent in basic-2, and to 1 percent at the secondary level), and much Chart 1: Girls' Participation Rates in Primary Education by Age and Grade (In %) (1993-1994) 50 _40 ..-30 z 20 n~10 0 6 7 8 9 10 ]1 12 13 14 15-17 Age | st gr 3 rd gr 0 4th gr 0 6th gr more so than for boys. (c) Gross enrollment ratios2' among girls for primary education (ler cycle de l'enseignement fondamental)3' reach a low 55 percent, a 26-point lag compared to boys. No other country of the LMI/MENA countries shows such a low score and such a gap. The share (participation rate) of girls in the cohort of newly enrolled students in first grade was 43 percent in 1993/94, which suggests that changes are underway, but at a slow pace. Indeed, gender differentiation starts at an even earlier stage, i.e. at the pre-school level, where the participation rate of girls in 1993/94 was only 30 percent (Koranic and modern schools included). There are numerous positive externalities of girls' basic education (on 2/ The gross enrollment ratio is defined as: (number of girls enrolled at level x) / (cohort of all school aged girls in the population). 3/ The Moroccan education system is structured in 3 cycles: (i) a two-cycle basic education system totalling 9 years, with a 6-year-lst cycle and a 3-year-second cycle, (ii) a 3-year secondary education slice, and (iii) a third education level with either specialized institutions or universities with cycles of various durations. Annex Ill-B  -3- health, fertility, hygiene, nutrition, etc), thus its contribution to sustainable economic growth could be substantial. (d) What happens at post-basic education levels is mainly the legacy of the situation prevailing in primary education. Gross enrollment ratios for females in secondary education level off at 30 percent, while the national average is 36 percent, both figures being under the LMI/MENA comparators. However, girls' share in total enrollments at this stage of the pyramid stays virtually at their primary education level (42 percent), as the transition to secondary education is also lower for boys. In higher education, gross enrollments rates in 1990/91 were slightly higher in Morocco than in other Maghreb countries, with a share of 38 percent, women were in a relatively good position compared to that in neighboring countries. 4. Flows. Participation rates4' of girls in public primary education decrease steadily from 43.6 percent in first grade to 39.5 percent in 6th grade (1993/94). However, girls tend to go through the system more rapidly than boys, and cross-sectional data show (i) lower repetition rates for girls (11.4 percent vs 14.6 percent), and (ii) no special gender bias for drop-out rates. Girls's distribution by age in primary education (Chart 1) clearly shows that within each grade, girls are concentrated in younger age groups, and their share tends to dwindle systematically as they enter older age groups. These figure suggest that, at an equal (or even better) performing level, girls do not remain in school, and that, in case of failure, their family take them back home for domestic work, while boys will be given second or third chances.5' (a) Trends. Literacy rates for women have almost doubled between 1960 and 1992 (while increasing by 50 percent overall); however, gains have benefitted urban areas almost exclusively and, as a result, the literacy gap between the two milieux has worsened, to the detriment of rural women; (b) Enrollment ratios for women at the primary level have gained 20 points since their 35 percent level of 1965, and are picking up again after the long plateau which characterized the mid-80's, gaining more than ratios for men (who obviously had a shorter gap to fill); (c) likewise, female enrollment ratios in secondary education have been multiplied by 5 since 1970 (3 times across genders), a time where they were still well under 10 percent; the same trend is noticeable in higher education, where the participation rate of women has doubled between 1974 and 1991. B. Determinants of Gender Disparities 5. Identifying the precise reasons for the unfavorable situation of girls in the education system is difficult because of the combined occurrence of factors with time-lagged effects. Gender disparities vis-a- 4/ The participation rate is defined as: (number of girls enrolled at level x) / (the total enrollment at level x). 51 A longitudinal analysis would be needed to confirm these hypotheses, and in particular, to shed more light on the drop- out patterns, which reflect parents' strategies and choices between school and non-school activities). Annex III-B  -4- vis schooling often result from other kinds of gender-biased factors, such as the milieu of residency, the level and source of the household's income, and other household's characteristics. Two sets of factors for the under-representation of women can be identified, and a recent study performed in rural primary schools helps to separate them empirically.6' 6. Supply-side factors. School-based elements are more powerful in explaining girls' access to school than their survival in the system. (a) Access to school. A distance greater than 500 meters hinders attendance by girls, while there is no significant effect on boys up to a distance of 2,000 meters. It affects both girls' access and retention, and their chances of transition to higher levels of schooling. (b) Equipment. The presence of some amenities such as water, electricity and latrines have a slight, but significant impact on girls's enrollment (almost none for boys). Likewise, the effect of a completely equipped canteen is stronger for enrollment of girls than boys. (c) Parents's preference for female teachers for girls is stronger than their preference for male teachers for boys. However, while female teachers outnumber male teachers in urban areas, they represent only 21 percent of the total teacher population in rural areas where, precisely where this attitude is more prevalent. 7. Demand-side factors. They influence both girls' access to and retention in primary schools; they also influence their likelihood of continuing to secondary schools. (a) Girls are still considered a second-priority for investment in education. Girls' school attendance is more sensitive to external obstacles, and this is why supply-side factors are usually more strongly associated with enrollments for girls than for boys (see para. 9, above). To separate out the respective role of culture, religion, and sheer poverty would need more analysis and data.' (b) A mother's literacy level has a strong effect on girls' access to school (and to a lesser extent, on her survival in the system), while it does not affect boys' chances of attending primary school. Girls' probability of attending primary school are lessened when their fathers are farmers (unfortunately a frequent occurrence in rural areas), and their chances of dropping out earlier increase sharply when their fathers are absent. (c) Household size, when combined with household composition exerts a significant influence on girls' attendance: the existence of siblings younger than six-years-old is negatively associated with girls' access to school, while boys are unaffected by this factor. This correlation is mainly the result of girls' responsibility for watching over younger brothers and sisters. 6/ See in particular Kingdom of Morocco- Literacy and Schooling in Rural Areas . Mingat A., Jarousse J.P. The World Bank, December 1993 7/ For an illustration of the difficulties linked to this exercise, see for instance: Heyneman, Steve, and Esim, Simel. 1994. " Female Educational Enrollment in the Middle East and North Africa: A Question of Poverty or Culture ? ". Educational Horizons, (Summer): 166-169 1994. Annex Ill-B  - 5 - (d) Foregone earnings are more important for girls than for boys, because the former are more often in charge of either domestic activities (wood and water fetching) or even working outside to bring home earned wages. (e) Although family characteristics do not play as significant a role in girls' chances to enter post-primary education (2?me cycle fondamental & secondaire), a mother's literacy and child-minding responsibilities remain important predictors of girls's access to this level of education, while boys' access is not sensitive to these factors. 8. Data regarding the gender distribution of performance is lacking, and it is not possible to know if this aggravates or mitigates biases occurring earlier in the education process. To the extent that literacy and formal education are more and more generally perceived as prerequisites for entry into the modern labor market, gender biases in education are poised to translate into biases in access to this market. C. Government and Donor Strategy 9. The Moroccan Government currently channels 22 percent of the public budget to the education sector, a proportion higher than what is observed in many LMI/MENA countries (16 percent on average). 10. However, these significant budgetary efforts have mainly been focused on secondary education (which absorbs up to 47 percent of the sector's recurrent budget and half of its investment budget) and, to a lesser extent, on higher education. The lack of efficiency and the concentration of Government resources on post-basic education levels contributes to the scarcity of resources available to primary education. At this level, which receives only a third of the sector's recurrent budget (and one tenth of its investment budget), most endeavors have been concentrated around the expansion of schooling capacities as well as on increasing the teaching staff. Much less has been done in terms of trying to address the factors which deter parents from sending their daughters to school, even when the facilities exist. Likewise, the outcomes of the literacy programs launched in the 70s and 80s have fallen short of expectations mainly because of (i) the lack of resources and (ii) the lack of follow-up actions to increase employment activities in post-literacy programs, particularly important in the case of women. 11. Under these circumstances, the Government's objective of reaching universal primary education by the year 2000 will not be reached before another decade unless major improvements are made in reducing the marginal cost of reaching--and retaining--each additional student. Nevertheless, even if the pace of universalization of education is slower than scheduled, and even without any special targeting, improvement in total enrollments should benefit girls most. 12. In addition, the Ministry of Education, with the assistance of UNICEF, UNDP and other agencies, has devised a comprehensive strategy focused on rural areas and aimed at correcting gender imbalances, especially in low density population zones. The thrust of it consists in encouraging families to send their children to school by alleviating schooling costs and involving local communities in education activities. The ministry also favors a multiplication of "micro-schools", to make schools closer to more families. This approach might prove (i) expensive, because of the scattered patterns of dwellings in the 10 percent tranche of the population not yet served, (ii) inefficient, because the quality Annex Ill-B  of the service is likely to be low, and (iii) uncertain, as households from remote or inaccessible areas are also more likely to move to other (mostly urban) centers, making these costly extension centers useless. D. Family Planning! 13. After a rapid increase in the 60s, 70s and 80s (population has doubled between 1960 and 1993) population increase is now on a slower growth path, with an annual growth rate approaching 2 percent. Morocco stands at an intermediary rank, between MENA countries, where the average growth rate is still above 3 percent, and East Asian countries (Korea, Indonesia and Thailand), where it has dipped under 2 percent . 14. This positive trend is the result of the decline of the total fertility rate (down to 4.2 in 1993 from 7.0 a generation ago), which more than offset the impact of the decrease in mortality. Fertility rates are still twice as high in rural areas than in urban areas (5.7 vs 2.8), and its downward trend is much less steep in rural areas. The National Survey on Population and Health (ENPS-II) of 1992 sheds light on the changes in behavior which have led to the national trend towards lower fertility and shows that the desire for limited size families is emerging in rural areas and growing in cities, as education, urbanization and higher incomes are becoming more widespread. 15. The lower fertility rate is attributable first to the dramatic change in the marriage age of women, which has gone up substantially from its average of 17 years in the 1960's. Indeed, the level of education strongly influences the age of marriage, which is still 18 for illiterate women, against 23 for women with secondary education. 16. This factor has been amplified by the more frequent use of modern contraceptive methods. Family planning campaigns started in the mid-60's, and were reinforced in the early 1980's. Their impact can be measured by a 63 percent increase in the use of contraceptives between 1983 and 1992, which currently reaches 42 percent, placing Morocco above most other Arab countries (lagging behind only Egypt and Tunisia). As expected, the cultural environment and the education level strongly determine the use (and method) of contraception: only a third of rural women and the same proportion of uneducated women use contraceptives (vs two thirds for urban women and educated women). Oral contraception is the most popular method used in Morocco (68 percent), and the main source of supply are the public health centers. Condoms are a marginal method of contraception, and non-governmental organizations do not play a major role in their supply. These are probably potential avenues for further progress in a population control process which, until now, seems to have been mostly borne by women, mainly under public auspices. II. LABOR FORCE PARTICIPATION A. Women In the Labor Market 17. All over Morocco, there has been an appreciable increase in the relative share of female employment during the last thirteen years. The aggregate percentage of the female labor force in the Annex III-B  country averages 32.9 percent, which is well in line with the ratio of the female labor force in other growing developing economies, such as Indonesia (23.2 percent), Malaysia (25.3 percent), and the Philippines (25.6 percent). 18. This paper uses recent survey data to investigate gender differences in labor market participation. The most recent information on the rural population dates back to 1986-1987 statistics since recent information on the rural labor force is lacking. This section discusses in particular the conditions of women in the urban labor force, and briefly examines available data for rural areas. 19. Urban Areas. In 1993, the urban population totalled 13,332,000, of which the female population was 6,627,147 (or 49.7 percent). Women represent 21.4 percent of the active labor force in urban areas (23.4 percent in 1985), while this percentage increases to about 44.5 percent in rural areas. In urban areas, only 15 percent of the female population are considered "active", while the balance (85 percent) includes women with no professional activity, students, retired, etc. 20. The following table compares the urban population according to gender, age and activity status: Table 2: Urban Population by Gender, Age and Activity % Active % Active Employed Unemployed % Inactive Age M F M F M F Less than 15 2.1 0.9 - - 97.9 99.1 15 to 24 32.8 12.6 13.7 5.9 53.5 81.5 25 to 34 75.3 24.6 16.4 8.5 8.3 66.9 35 to 44 92.8 23.0 4.7 1.9 2.5 75.1 45 to 59 85.8 13.4 2.9 1.0 11.3 85.6 60 and more 34.3 5.7 0.5 0.1 65.2 94.2 Total 42.9 11.8 7.0 3.2 50.1 85.0 Source: Annuaire Statistique du Maroc 1994, Direction de la SSatistique. * Percentages refer to the proportion of the activity status of each age group, by gender, compared to the total age group. 21. The above table shows that the rate of female labor force participation in age groups 0-15 is similar to that expected in a middle income developing country, while that of the age group 15-24 is significantly lower (32 percent in Indonesia, 38 percent in the Philippines, 43 percent in Malaysia and 76.6 percent in Thailand). 22. The following table compares the ratio of female participation by sector in the urban labor market in the years 1985, 1991, 1992 and 1993. Annex Ill-B  -8 - Table 3: Ratio of Female Participation in Urban Labor Force by Selected Economic Sectors (Female Percentage of Active Employed Urban Population) Economic Sector 1985 1991 1992 1993 Agriculture, forest, fishery 14.0 17.5 16.8 14.5 Mining industry 3.3 4.5 3.2 2.8 Manufacturing industry 41.7 42.3 38.3 37.5 Maintenance 0.9 1.0 0.8 1.6 Electricity, gas, water 5.0 8.6 7.6 8.7 Public works 1.1 1.1 1.3 1.6 Trade (wholesale and retail) 6.7 8.5 7.4 7.3 Restaurants/hotels 11.8 11.6 13.5 12.7 Transport/communication 4.5 5.0 5.1 5.1 Banks, insurance, real 26.5 30.3 30.5 31.5 estate, services Household services 55.3 52.3 49.7 48.4 Social services 32.1 35.3 35.1 35.6 General administration 13.9 15.6 15.7 15.1 Miscellaneous 23.1 16.8 18.3 19.2 Total 23.4 23.8 21.7 21.4 Source: Annuaire Statistique du Maroc: 1987, 1992, 1993 and 1994, Direction de la Statistique. 23. The above figures provide insights into the general structure as well as the direction of changes in the gender distribution of employment in various skill categories in urban areas. A gradual shift can be noted from manual occupations (agriculture, mining, household services) towards activities with higher professional content (services, banking, etc.). It is interesting to note that female participation in manufacturing has declined somewhat between 1985 and 1993. In addition, except for the manufacturing sector (37.50 percent) and social services (35.6 percent), women have not been able to make significant strides in the labor market. 24. The evolution between 1980 and 1993 of female participation in the urban labor force is summarized in the table below: Annex Ill-B  Table 4: Evolution of Female Participation in the Urban Labor Market Year 1980 1981 1982 1984 1985 1991 1992 1993 Female Participation 25.0 27.8 28.9 24.5 23.4 23.8 21.7 21.4 Source: Annuaire Stanstique du Maroc, 1987,1994. 25. The total ratio of "active" urban women, which peaked in 1982 (28.9 percent) has been declining in the last eight years (23.4 percent in 1985 and 21.4 percent in 1993). This may in part be related to the general trend towards the attainment of higher degrees by women, which in urban areas has outpaced the corresponding trend in male education (see para. 17). A significant portion of urban female students tend to stay longer in school, therefore delaying entry into the labor market. 7his may be a result of the higher investments required by women prior to entering the labor market. 26. The participation of Moroccan women workers in household services has been declining but is still very important (personal and household services: 48.4 percent in 1993). This is true even given the relatively higher educational level of women. 49.3 percent of the active female urban population holds no educational degree or diploma as against 57.8 percent of the male active urban population (see Table 5). 27. There are conspicuous regional differences in women's participation in the urban labor force. The higher rates are found in the industrialized regions: centre 27.8 percent and centre-sud 25.3 percent; the lowest in the oriental region 7.6 percent (70.1 percent of men are active in these areas). 28. Employment and Wages - Legal Framework. The Moroccan constitution offers equality of opportunity for all citizens with respect to employment. Unfortunately, it is not certain that this principle is consistently applied across the whole spectrum of the labor market.8' 29. Data for state employees show that female employees in public administration increased from 92,229 in 1986 to 115,504 in 1993 (+25.2 percent). By comparison, male employees increased from 230,998 to 254,016 (+ 10 percent). An analysis of the wage distribution for state employees is shown in the following table: 8/ An estimated earnings function for the private sector did not show that gender was a significant determinant of the level of earnings (see Annex III-A). Annex Ill-B  - 10 - Table 5: Comparison of Gross Salary Earning in the Public Administration by Gender and Wage Segment (*) Wage segment (DH) Male % M/Mt Female % % F/Ft % M/Mt in l _____________________ ___________ _________ F/Ft in 1986 1986 Less than 20,000 22,437 8.8 16,575 14.3 45.3 33.2 20,000 to 40,000 144,209 56.8 77,990 67.5 51.3 58.1 40,000 to 60,000 44,414 17.5 13127 11.4 1.3 5.2 60,000 to 100,000 29,593 11.6 5,398 4.7 1.1 2.3 More than 100,000 13,363 5.3 2,414 2.1 1.0 1.2 Total 254,016 100 115,504 1 00 I 100 100 Source: Annuaire Starisnique du Maroc 1987 and 1993, Direction de la Statistique. (*) Include regular and temporary staff. Table 6: Repartition of Active Urban Population by Gender and Educational Level Educational Level M M/Mt F F/Ft RF/RM No Diploma 1,908,107 57.9 481,272 49.3 0.85 (1,769,287) (59.0) (535,658) (52.7) (0.89) Certificates of Primary Education or 518,627 15.7 100,617 10.3 0.66 equivalent (444,226) (14.8) (95,038) (9.3) (0.63) Certificate of Secondary Education 309,577 9.4 67,581 6.9 0.73 1st cycle (263,875) (8.8) (63.039) (6.2) (0.70) Baccalaureate or equivalent diplomas 84,659 2.6 29,780 3.0 1.15 (78,649) (2.6) (33,001) (3.2) (1.23) Degrees of higher education except 88,409 2.7 47,240 4.8 1.78 Medicine (74,115) (2.4) (32,999) (3.2) (1.35) Degrees of higher education 133,773 4.0 51,797 5.3 1.32 (124,655) (4.1) (49,952) (4.9) (1.20) Technical diplomas 144,735 4.4 67,680 6.9 1.57 (131,996) (4.4) (49,560) (4.9) (1. 1 1) Other professional diplomas and 107,744 3.2 130,024 13.3 4.15 certificates (110,490) (3.7) (156,289) (15.3) (4.16) Not declared 212 93 (733) (-) Total 3,295,843 976,084 (2,998,026) (1,015,536) Source: An,iuaire Staristique du Maroc 1993 (data for 1991 are in parenthesis). Annex Ill-B  - 11 - 30. The table shows that in urban areas the educational level of women in the labor force is slightly superior to that of men. 49.3 percent of women have no educational degree, compared to 57.9 percent of men. Also, in the active urban population, apart from the first two stages of the education cycle, women outperform men in educational achievement. (This is in general even more significant in the 1993 figures compared to the 1991 figures). This superior achievement is particularly striking for higher degree diplomas. These numbers imply that women are much more likely to pursue higher education. 31. The situation depicted in Table 5 is not mirrored in the employment statistics. In 1991 there were in urban areas 236,171 unemployed women, of which 104,099 between 15 and 24 years and 108,099 between 25 and 34 years. Table 6 provides a comparison of unemployment structure by gender. Table 7: Comparison of Unemployment Structure by Gender M % F % Un-employed without previous 209,783 45.7 141,613 60.0 working experience Un-employed with previous working 249,526 54.3 94,558 40,0 experience Total 459,309 100 236,171 100 Source: Population Active Urbaine 1991, Direction de la Statistique. 32. The percentage of unemployed females without previous working experience is much higher than for males indicating the greater difficulties for female employees in accessing first employment than for men.9' Discriminatory selection criteria and practices are reported, preventing women from accessing non-traditional jobs."0' There is widespread evidence that for the same managerial (or highly technical position), a woman is asked for more educational credits than a man. 33. The female unemployment rate, compared to the male rate, appears to have worsened between 1985 and 1991 in selected age groups, as shown in the following table: Table 8: Unemployment Rate by Gender and Age Group Age Group Year M F Ratio F/M 15 to 24 1985 27.0 28.8 1.066 1991 31.3 35.3 1.128 25 to 34 1985 11.4 18.2 1.596 1991 19.8 25.4 1.282 35 to 44 1985 4.8 11.9 2.479 . ________________ 1991 4.7 9.5 2.021 Source: Annuaire Statistique du Maroc: 1992, 1993, 1994 and Population Active Urbaine 1991: Direction de la Statstique. 9/ An analysis of the probability of being unemployed as a function of various characteristics suggests that male workers are, ceteris paribus, less likely to be unemployed than female ones. 10/ When applying for a position requiring a certain degree of skill and responsibility, women's availability in termns of working hours are typically questioned. Annex Ill-B  - 12 - 34. The table shows in particular a worsening of female unemployment for the age groups 15/24 and 25/34. The female unemployment rate appears to have decreased in the age group 35/44 and much more so than the male rate. In relative gender terms, the ratio female/male unemployment rate has improved for the age groups 25/34 and 35/44, while it has worsened for the age group 15/24. This comparison may be affected by many factors, however it reflects in general terms the increasing difficulty for urban female workers to access first employment. A detailed analysis would be required to detect structural trends and specific issues for the age groups 15/24 and 25/34. B. The Rural Labor Market 35. There are no recent studies on the structure and organization of Moroccan women in the rural labor market. The last survey was conducted in 1986-1987, and this section is based on this survey. Eight years after this survey, the labor force in the rural areas has grown considerably, educational levels may have changed, and the socio-economic situation also may have changed considerably. Therefore the following considerations may need to be revised in light of updated information. 36. The available data show that a high proportion of Moroccan women are employed in rural areas. According to the 1986-1987 survey, about 2,348,196 women (36.0 percent of rural female population) are described as employed, as compared with about 2,923,492 men (47.0 percent of the rural male population). Of these 36.0 percent, 3.5 percent are involved in wage and salary employment compared to 22.9 percent of men. 84.5 percent of these women are engaged in family and parental farm work compared to 36.2 percent of men. Table 9: Repartition of Active Rural Population by Gender and Activity Activity Male % Female % Employer 2.1 0.0 Independent 37.1 11.6 Employee 22.9 3.6 Family /apprenticeship 36.2 84.5 No occupation 1.7 0.3 Source: Population Active Rurale 1986 / 1987, Direction de la Statishique. 37. As already indicated (para 4) , the participation rate of rural women in the labor market amounted to 44.5 percent (1986-1987) and is thus much higher than the same rate in urban areas (21.4 percent). Most of the active rural women are engaged in family work (84.5 percent). The significant presence of women in the rural labor market is mainly concentrated in agriculture and agriculture-related activities (85 percent). Only 15 percent are active in non-agriculture activities. Annex Ill-B  - 13 - 38. The large female involvement in agriculture is reflected in heavy farm (typically family farm) work, which is inadequately rewarded, as it consists largely of mobilization into family farm activities. Women's contribution is often not formally recognized and the constraints they face in coping with both farm and family work, not always understood. 39. The available data (1986-1987), show a striking disparity between the educational levels of active women in rural areas versus men, as seen in the following table: Table 10: Rural Population by Gender, Educational Level and Activity Act. Employed Unem loved Inactive M/Mt F/Ft M/Mt F/Ft M/Mt F/Ft Population less than 10 years 1.70 4.6 - - 98.3 95.4 No degree 71.4 50.6 5.4 0.6 23.1 48.7 Primary school diploma 34.5 13.0 8.4 1.8 56.1 85.2 Secondary school, high school 33.4 10.9 12.4 8.2 54.1 80.1 and vocational training degree _ Non declared 81.7 69.7 5.1 4.8 13.6 25.5 Source: Population Active Rurale - Enqu&tes Statistiques 1986/87. 40. The difference in educational level between men and women increases from primary to secondary school. C. Conclusion 41. Female participation in the labor market is much higher in rural rather than in urban areas. High female participation in rural economic activity is usually under-remunerated. Low female educational levels do not ostensibly affect female participation in the rural labor market, while improved female educational levels in urban areas have not to date, produced a noticeable improvement in terms of job access, and career prospects. Therefore Moroccan women face relatively accessible, under-remunerated and unskilled occupation possibilities in rural areas, against a situation of better remunerated but hardly accessible urban jobs, where they face increased competition vis-a-vis the male labor force. 42. Gender problems reflected in the labor market are linked to other aspects of Moroccan life. One cannot address one of them, such as employment, disregarding the others, such as issues related to education, health, family relations and responsibilities. Comprehensive solutions are needed. The steps needed to improve the status of women, reduce poverty and to increase labor productivity, are: i) better access to education; ii) growth facilitating policies: while higher growth will improve labor market access for women, releasing women from constraints to the development of their human capital will raise growth; Annex III-B  - 14 - iii) proactive measures including raising the awareness of society of the economic cost of continued discrimination, iv) more progress in reducing the population growth rate through direct efforts. 43. In addition, it is important to increase the scope of research and data collection on gender factors in employment, identifying patterns of job segregation and detecting specific groups with particular needs and dynamics. Annex Ill-B  ANNEX IV-A FINANCIAL SECTOR REFORM 1. The text of the CEM argues that in order to accelerate economic growth, a precondition to reducing the high rate of unemployment, Morocco would benefit from implementing financial sector reforms in six major areas: (1) interest rate liberalization, (2) regulatory and legal reform of securities, bond, and money markets, (3) reform of the postal and contractual savings system, (4) elimination of the Government's preferential access to credit, (5) accelerating the use of indirect monetary policy instruments, and (6) greater integration with world financial markets. Reform in these areas would promote the provision of financial services needed to support higher growth in Morocco. This Annex discusses these recommendations in greater detail. The first Section briefly describes the structure of Morocco's financial system as well as the importance of developing the financial sector, based on the experience of other countries and on empirical work. The second six-sections discusses action plans for each of the above recommendations. A. Introduction and Background 2. Developing the financial system in Morocco is important because it will (a) increase investment and (b) raise productivity by improving the allocation of financial, and other resources in the economy. The financial system performs this role by providing the following principal functions: mobilizing resources, lowering transactions costs, facilitating risk management, evaluating managers/monitoring managers, and providing information'/. Financial sector development is also closely tied to reform in other sectors of the economy. For example, reform of the PE sector including privatization2' is more effective when financial markets are well developed since newly privatized firms must go to financial markets and institutions to raise capital. In fact, the larger the public enterprise reform program the larger will be the required development of the financial system. It also enhances the resource reallocation effects of trade reform: movement of capital to its most productive use will encourage the movement of labor (and other inputs) to their most productive use. Cross-country work shows the strong correlation between financial sector development and growth3". In fact, the level of financial development has been found to be a robust predictor of economic growth4". 3. At present, the services that a developed financial sector can provide are not provided in Morocco to the extent required for higher sustained growth and there is need for substantial reform in this sector. 1/ Ross Levine, Does the Financial System matter? World Bank May 1994 (mimeo). 2/ "The Financial System and Public Enterprise Reform: Concepts and Cases.", Asli Demirgue-Kunt and Ross Levine, Draft, NB., August 1994. 3/ Gurley, John and Shaw, E.S., "Financial Structure and Economic Development", Economic Development and Cultural change, Vol 15. April 1987, pp. 257-68; Alan H. Gelb, Financial Policies. Growth and Efficiency. World Bank, Policy, Planning, and Research Working Papers, WPS 202, 1989: Gerther, Mark and Rose, Andrew, Finance, Growth and Public Policy World Bank, Policy Research Working Papers, WPS 814, 1991. 4/ See Levine and Renelt (1992). No other macroeconomic indicator has this predictive content. Annex IV-A  -2 - 4. The empirical evidence indicates that there is no unique structure of the financial system that is necessary for the efficient provision of the above services. Different mixes of institutions and markets can work together to achieve the same goals. Therefore, the goal of financial policy should be to ensure that the incentive and regulatory framework allows the efficient provision of these services. Reform in the six areas mentioned above should facilitate and encourage the development of a sophisticated financial sector in Morocco, capable of performing the necessary functions to support accelerated growth by increasing the efficiency and magnitude of investment. 5. Though the size and structure of financial systems vary across countries, some generalizations can be made. As countries become richer, (a) financial systems get larger; (b) banks grow in importance relative to the central bank in allocating credit; (c) credit to the private sector increases in proportion; (d) non-bank financial institutions become more important (savers and investors have many more options); (e) stock markets become more developed.5' 6. One example of the different structures of financial systems that can produce similar services is provided by comparing France, Italy, and Japan with the United States. In the former group of countries bank credit provides about 30 percent of the financial resources of enterprises against only 9 percent provided by bonds and stock; the remaining 61 percent are from internally generated resources. In the latter, on the other hand, 45 percent of the financial resources of the enterprise sector are provided through corporate bonds. 7. The Financial Sector in Morocco. Morocco's financial sector includes a central bank, thirteen commercial banks, a financial holding company which was recently privatized (SNI), six Government controlled specialized credit institutions, the Treasury, the postal checking system, the National Savings Bank, the Caisse Centrale de Garantie, 22 insurance companies, pension funds, 6 leasing companies, a few recently created brokerage houses and the Casablanca Stock Exchange. (a) 7he Central Bank, Bank Al-Maghrib, is responsible for executing monetary policy. The main policy making body on monetary and credit policy is the Credit and Financial Market Committee, which includes the Minister of Finance, the Governor of the central bank, and representatives from Morocco's banking and financial community. (b) All commercial banks have significant, though minority, foreign bank participation in their capital, except for the Banque Marocaine du Commerce Exterieur which was 50.01 percent publicly owned until early-1995 when the Government sold 43.01 percent of the 5/ Ross Levine, Financial Functions, Institutions, and Growth, Draft, World Bank 1994. Annex IV-A  - 3 - capital it owned to the private sector. These banks together have a network of about 900 branches. There are also three offshore banks in Tangiers' offshore banking zone. Conventional methods of calculating competition in the banking sector suggest that the level of competition is lower than in modem developed banking systems such as those of Turkey, the Philippines and Chile, similar to that in Tunisia, but greater than those of Pakistan or Columbia.6' Satisfactory prudential regulations have recently been put in place. (c) The sector of specializedfinancial institutions is made up of the Banque Nationale pour le Developpement Economique (BNDE), which lends for industrial projects, the Caisse Nationale de Credit Agricole (CNCA), which lends to the agricultural sector, the Caisse de Depot et de Gestion (CDG) a Government-owned bank used as a conduit to channel savings towards the public sector, the Credit Immobilier et Hotelier (CIH), which lends for the housing and tourism sectors, the Fonds d'Equipement Communal (FEC), which lends to local Governments, and the Caisse Marocaine des Marches (CMM), which lends to private companies engaged in public works. (d) The Treasury performs certain banking operations, and accepts deposits from non- financial agents, particularly in the form of deposits in the postal checking system and treasury deposits. (e) The Postal Checkdng System receives deposits from the private sector throughout the country, while the Caisse d'Epargne Nationale (CEN) operates out of the post offices and mobilizes private sector savings. (f) There are 22 insurance companies, which provide the usual range of coverage, including management of some private sector pension funds. (g) The pension system is composed of the Caisse Marocaine de Retraite (CMR) for civil servants, the Caisse Nationale de Securite Sociale (CNSS) for private sector employees, the Regime Collectif d' Allocation de Retraite (RCAR) for contractual employees in the public sector, the Caisse Interprofessionelle Marocaine de Retraite (CIMR) that provides voluntary and supplemental pensions to private sector employees, plus a number of pension funds run by professional associations and enterprises. (h) The Casablanca Stock Exchange provides a market for Government and Government- guaranteed bonds and equity. 8. The financial system in Morocco is dominated by the deposit money banks which provide substantial resources to finance the public deficit and extend mainly short term credit to the private sector. Medium- and long-term credit to the private sector is provided largely by the specialized financial institutions (See Table 2), which obtain their resources from bond issues and from borrowing abroad, 6/ See Kingdom of Morocco: Sustained Investment and Growdt in the Nineties, Annex, World Bank, November 15, 1990, pp. 32-33. Data are for the 1980's. Annex IV-A  - 4 - both of which are granted Government guarantees. These institutions lend to their own privileged clientele. The share of medium- and long-term credit in total credit extended by the deposit money banks rose from 4 percent in 1980 to 10 percent in 1993; while rising, this share is still rather modest. Including the medium-and long-term credit granted by the other financial institutions, such credit rose from 24 percent to 30 percent of total outstanding credit during 1980-1993. Enterprises finance much of their medium- and long-term needs by rolling over short-term credit. The dependence of enterprises on bank credit to satisfy their financing needs fell slightly during the period under review. It stood at 54 percent in 1991-1993 against 65 percent in 1980-83. Private enterprises have not yet issued domestic bonds. The capitalization of the stock market has increased in recent years, because of recent privatizations, the distribution of stock dividends and the appreciation of the outstanding shares. However, its size is still relatively small as compared with similar middle-income countries. Pension funds and insurance companies have accumulated substantial reserves, most of which are invested in public debt and in bonds issued with Government guarantees; some is invested in the equity market. The CNSS has also accumulated substantial reserves, all of which is deposited with the CDG. Similarly, all deposits collected by the Caisse d'Epargne Nationale (CEN) is deposited with the CDG. 9. One measure of the size of the financial system that is often used is called depth. Size is thought to be positively correlated with financial development7'. To get a measure of financial depth, Table 1 compares M2/GDP in Morocco with some other countries. Morocco seems to compare Chart 1: Morocco - Financial Deepening, 1980-93 favorably. It also shows the 0o7 efficiency of investment in Morocco, (column 3), which compares poorly with the same 0.5 _ countries. M2/GDP rose from 0.4 Cs/D 30 percent in 1960 to 42 0.3 2/GDP1percent in 1980 and further to 0.un64 percent in 1993,ae (see Chart 1) suggesting a steady Tunisia, for instance, here savers maeasilyplaceheirsavingsi and substantial progress in 0 n c financial deepening. This ratio bearing iiabiiities af banias and aon banic financiai incompares favorably with the p~~o~~,id~~~d by Mo~~~~co~~, ~~ 1960 financial depth of countries that managed to grow rapidly in the succeeding 30 years. However, the high level of M2/GDP in Morocco in 1993 also reflects the fact that savers have no alternatives to bank deposits in which to place their savings. In Tunisia, for instance, where savers may easily place their savings in Treasury paper, the M2/GDP ratio 7/ Depth is normally defined to equal currency held outside the financial institutions plus demand deposits and interests bearing liabilities of banks and non bank financial intermnediaries divided by GDP. 8/ M2 is defined as the sum of currency in circulation, demand and time deposits at money banks, the postal system and the Treasury. Bank deposits (i.e. minus currency in circulation) over GDP rose from 28 percent in 1980 to 50 percent in 1993, and suggests a rapid rise in the use made of banks by the public at large. In Tunisia this ratio stood at 32 percent in 1993. Annex IV-A  is only 42 percent; this however should not be taken to mean that Tunisia's financial sector is less developed. Table 1: Financial Development and Growth, 195847 Countries grouped by GDP Gross national Change in GDP/ M2/GDP Growth Rate savings/GDP investmnent High Growth rate(7%) Seven Countries 28.0 26.3 43.0 Excluding China 23.2 33.1 Medium Growith (3-7%) 51 countries 18.5 23.6 31.2 Low Growth countries (less than 3%) 19.0 10.1 23.8 Morocco 1980-93 19.7 a/ 16.7 49.4 a/ Average domestic savings over GDP for 1980-93 was 16.3 percent Source: World Develooment Report 1989, World Bank, 1989, (Washington, DC), Table 2.1 p. 27; Morocco data arefrom World Bank fikes. 10. Bank credit has fallen slightly as a share of total financial assets in the economy (from 65 percent in 1980 to 61 percent in 1993), indicating a somewhat greater financial diversification. The growth in capitalization of the Casablanca Stock Exchange accounts for most of this decline. Assets of deposit money banks grew somewhat faster than those of other non bank financial intermediaries. 11. The share of credit to Government rose in total outstanding credit granted by deposit money banks, as shown in Chart 2 (from 31 percent in 1980 to 34 percent in 1993), during a period when the overall budget deficit fell from more than 10 percent of GDP to 2-3 percent of GDP, suggesting that deficit financing continues to compete heavily with the private sector for financial resources. Substitution of domestic for foreign financing of the deficit accounts for much of this development. Net credit to the Government as a share of total credit to the economy (Monetary Survey which included credit from the BAM and the IMF) fell from 59 percent in 1980 to 51 percent in 1993, mainly because credit from the BAM did not rise as fast as credit from the deposit money banks. 12. The stock market has become more developed with capitalization rising to over 10 percent of GDP in 1993 and to 14.3 percent in 1994 (Table 2, Table 3). However, market capitalization is substantially lower than in most rapidly growing emerging markets. Table 2 presents a description of the Moroccan financial system. Annex IV-A  - 6 - Table 2: Morocco Financial System, 1990-93 (Millions of Dirhams) 1990 1991 1992 1993 1993 % GDP 1. Deposit Money Banks Total Assets = Total Liabilities 95,026 114,309 123,910 135,557 54.7% Deposits 78,314 92,140 103,061 112,720 45.5% Sight Deposits 54,171 61,757 66,636 70,033 28.3% Time Deposits 24,143 30,383 36,425 42,687 17.2% Domestic Credit 72,966 88,591 101,896 111,005 44.8% o/w Government 31,967 33,385 39,548 44,657 18.0% Private Sector 40,730 54,218 61,833 65,840 26.6% o/w Medium/Long-Term 9,248 11,700 14,132 15,771 6.4% 2. Specialized Financial Institutions Total Assets = Total Liabilities 40,305 45,546 50,216 53,938 21.8% Deposit 15,014 17,410 19,670 21,836 8.8% Domestic Borrowing 8,701 9,152 10,178 10,642 4.3% External Borrowing 11,643 12,917 13,779 13,997 5.7% Other Credit o/w Government 5,923 6,568 7,191 7,980 3.2% Private Sector 27,357 31,092 34,072 35,542 14.3% o/w Medium/Long-Term 22,601 25,720 28,083 30,265 12.2% 3. Capitalization of the Stock Market 7,768 12,449 16,975 25,623 10.3% 4. Investments of Insurance Co. 12,852 15,202 17,092 20,288 8.2% 5. Resources of the Pension Funds N/A N/A 11,749 N/A 4.7% 6. Reserves of the Caisse Nationale de Securit6 Socialev 5,946 6,599 7,624 N/A 3.1% Source: Banque Al-Maghrib, Government of Morocco. 1/ Data are for 1992 2/ Other than reserveo of the pemion system. Annex IV-A  -7- 13. Financial reform was started in earnest in the early 1980's in Morocco. Its main achievements so far are the partial liberalization of interest rates allowing for real rates (see Chart 3), the elimination of quantitative credit controls and the reduction of obligatory lending to Chart 2: Morocco - Credit to Government, 1980-93 selected ysetors by selected sectors by 0.45 cornnercial banks, the 0.4 _ _ 1l strengthening of banking 0.35 legislation and the establishment of a legal PS 5 framework for the 1&2 * operation of capital 0.1 ~~~~~~~~~~~placements of below 0.15 * - - 1 11l||l||ll markets. Mandatory 0.05 market rate Treasury , _ _ . ~~~~~~~~~~~~paper with the comrnercial M190 1911 1982 11 19984 1905 1986 19 87 1939 19S9 1990 1991 1992 1993 3~: Ddo.o,dbdbyl r ifi. banks were also lowered, while BAM's monetary policy became largely based on changes in reserve requirements. B. Reform Agenda 1. Interest Rate Liberalization 14. The Current Situation. Interest rates in Morocco are not market determined: (a) The BAM sets the maximum lending rate. The rate is based on the cost of resources to the banking sector plus an intermediation margin of 3 Chart 3: Real Interest rates hi Morocco and France percent. Up to 12 May 1995 this le set the prime::} rate or 'Taux d e b a s e bancaire ' .I]Mroc (TAB) at 94 percent (with the cost of4 resources 19 1 199 1992 193 estimated at 6 S- BAI-MAWdib, B9q- deF- percent), after which it was adjusted to 8.5 percent. The maximum lending rate (taux plafond) for short- and medium-loans is set at 3 percentage point over the TAB, while the maximum Annex IV-A  - 8 - rate for long-term loans is set at 4 percentage points over the TAB. Nearly all lending takes place close to the maximum rate.9' Since March 1994 the maximum lending rate for working capital loans is 12 percent and for medium- and long-term loans, 13 percent. The ceiling on interest rates discourages banks from lending to small-and medium-sized enterprises,where greater risks and administrative expenses cannot be covered within the fixed interest rate margin."0' Chart 4 shows reference, ceiling and interbank rates for the period 1990-1994. Recent reforms have led to the development of market determined interbank rates. (b) Deposit rates are set within the guidelines provided by the Confed&ration Generale des Banques Marocaines (GPBM). At the end of 1994 the GPBM suggested rates were 7.5 for 3-month deposits, 8 percent for 6-month deposits and 9 Chart 4: Morocco - Interest Rates, 1990-94 percent for 12- month deposits. 1' The observed rates 12 for these maturities 10 at that time were 7 h percent, 7.5 6 N ' Referjice rates percent and 8-8.5 Ceilihn rate p e r c e n t Interbank rates respectively."' 2 Legal restrictions exist to prevent 0 S BAIMaghnb remuneration of m o s t s i g h t deposits.'2' Interest rates on savings accounts (comptes sur carnet) are set at 7 percent for commercial banks and 9 percent for the Caisse d'Epargne Nationale (CEN). These are to be changed periodically. 9/ The ceiling rate was to be recalculated every six months, but this was not done. In the meantime, deposit rates were reduced and according to some fragmentary information obtained, more loans are being made at rates below the ceiling rate than in mid-1994. 10/ Lending of foreign resources at below ceiling rates and a preferential BAM rediscount facility is used to compensate somewhat for this situation. 11/ Certain banks have given the freedom to their branch managers to negotiate the deposit rates within a .25-1 percent range from the observed deposit rates quoted above. L'Economiste, nr. 160, Thursday 29 December, 1994, p. 38. 12/ Exceptions are "depots sur carnet" by individuals with a maximum of DH 150,000, deposits with the CNCA and with the Treasury. Annex IV-A  - 9 - (c) In the absence of a market determined yield curve, returns on Government bonds are set by fiat and changed infrequently.'3' Government holds weekly auctions at which rates are established for selected maturities. However, these auctions are very thin (often a single bid or no bids for particular maturities), issues are not homogeneous and there is no secondary market for this paper. As a result, these auctions do not provide a yield curve. In addition, CDG which is a dominant player at the auctions (at the end of August 1994 45 percent of the outstanding debt auctioned off was on its books) needs to compensate nearly three quarters of its resources (deposits of the CNSS and the CEN) at a fixed 9 percent rate."4' The fixed cost nature of its assets prevents CDG from bidding for large quantities of Government paper at less than 9 percent. Other participants in the auction tend to follow CDG's lead. (d) Commercial banks are required to maintain monetary reserves (equal to 10 percent of sight deposits) and hold Government debt at below market rates (20 percent of deposits with a maturity of less than I year). The combined impact of these requirements reduces banks revenues by up to I percent of all outstanding credits. 15. Why Reform is Needed. The system of controls on interest rates prevents their functioning as a signalling and resource allocative device. For example, ceilings on lending rates will prevent firms from lending to riskier, but higher expected-return activities. The rigid interest rate structure and resulting lack of competition prevents rates from falling to reflect excess liquidity in the economy. Neither do banks have an incentive to mobilize savings agressively. In essence, the current policy prevents the financial system from effectively performing at least three of its functions: providing information, mobilizing resources, and facilitating risk management. 16. Recommended Reforms. With a continuing stable and sustainable medium-term macroeconomic framework, interest rates should be liberalized. Specifically, maximum lending rates would need to be abolished, while the deposit rates at the CEN and those received by the CNSS from CDG should be tied to market rates."5' GPBM guidelines on the level of deposit rates should be prohibited as they only reinforce the lack of competition in credit markets, while the prohibition against providing returns on sight deposits should be removed. 13/ The most recent change intervened in December 1994 when the rates were reduced by one half percentage point, to 9.5 percent for one-year, 10.5 percent for three-years and l1 percent for five-years. 14/ In 1992-93 an average of 72 percent of CDG's resources came from the CNSS and the CEN, organizations who by statute have to deposit their assets with the CNSS for which they receive a flat 9 percent return. CNSS resources deposited before January 1990 receive a slightly lower rate of return; on average CNSS received a return of 8.8 percent in 1993. 15/ CNSS should be permitted to invest in other instruments than deposits in the CDG as discussed below. Annex IV-A  - 10 - 2. Legal and Regulatory Reform of Securities, Bond and Money Markets 17. The Current Situation. At present there is a stock market in Morocco, but no private sector bonds other than those issued by state owned financial intermediaries (CIH, CNCA, and BNDE) with Government guarantees, and no real money market. Own resources, reinvestment of profits, and bank credit are the most important sources of enterprise financing. Until 1990 new bank credit exceeded the amount of capital raised through profit retention and new stock issues; however since then, the trend has been reversed; new bank credit is only half as large as the increase in capital of enterprises.'6' 18. Private Sector Bonds. Legislation dating from 1946, and which was in force until 1993, required the issuers of corporate bonds to inform the Treasury ex ante. This gave the Treasury the opportunity to deny or to ask for modification of the issue (schedule and details), with the purpose of preventing the crowding out of public sector bonds. With fiscal advantages attached to Treasury bonds until these advantages were eliminated in 1995, private sector bonds had to pay higher before tax returns to compete. The Government guarantee attached to bonds issued by state owned financial intermediaries continues to put the private sector at a disadvantage. The absence of unified accounting standards and a rating agency does not help. To date, there has been no demand submitted to the Treasury is issuing non-Government guaranteed bonds. The Comite d6ontologique des valeurs mobiliWres (CDVM), created in September 1993, is to issue regulations permitting the use of private sector bonds. The elimination of fiscal advantages attached to Treasury bonds (begun in 1994 and ended in 1995), will also facilitate the issuance of private sector bonds. 19. Issues of commercial paper and certificates of deposits are nearly non-existent in Morocco, except for the Billets de Tresorerie. These were used extensively for a brief period as a means of circumventing the credit allocation system in place until a few years ago. Since then, their use has been substantially reduced (from DH 3 billion in 1990 to no more than DH 1.3 billion in 1993), in part because the Government limited participation in this market to private non-financial institutions and individuals. Such paper is not traded. The CDVM recently issued detailed instructions regulating the issuance of commercial paper, including the Billets de Tresorerie. 20. The capitalization of the stock market in Morocco has grown from 2.6 percent of GDP in 1989 to about 14 percent of GDP in mid-1994 (Table 3). Even though the stock market has contributed substantially to the success of the privatization process and was in turn invigorated by it, it is still very underdeveloped: (a) Market capitalization is substantially lower than in most rapidly growing emerging markets. In addition, the number of enterprises listed on the exchange fell from 81 in the early 1980's to 65 in 1994. As a proportion of GDP the number of enterprises appears high compared with Mexico or Turkey, but low when compared with Jordan and 16/ Intermediation margins for commercial banks for 1991 were 4.4 percent, considerably higher than the 2-3 percent prevailing in the OECD countries. Another way of calculating the intermediation margin is implied in the BAM calculations of the Taux de Base Bancaire. In early 1994 the BAM estimated the average cost of total resources of the commercial banks at 5.5 percent. Comparing with the average lending rates, which are very close to the ceiling of 12 percent for medium-term loans and 13 percent for long-term loans, suggests an intermediation margin of 6-7 percent. Annex IV-A  - 11 - Pakistan. An active market exists for the shares of only a few companies. Transactions in the ten most actively traded shares account for 94 percent of total transactions (figures are for 1993) against one third to one half in the four comparator countries. These ten companies account for more than 60 percent of total market capitalization. Table 3: Morocco Stock Exchange - Major Indications, 1989-93 Morocco Jordan Turkey Macico Pakistan 1989 1990 1991 12 1993 1993 Market capitalization (in % of GDP) 2.6 3.6 5.1 7.0 10.3' 750 127 40.0 17.0 Share in total transactions of 10 most traded stock -- - 86 -- 94 51 31 52 30 l Number of listed enterprises 71 69 68 62 65 98 141 192 640 per Slbn ofGDP 3.3 12.5 1.4 0.6 15.2 Share of transactions off-floor 92 98 94 24 78 -- -- -- -- Change in price index (in %) 12.3 29.0 14.8 10.8 27.4 4.5 46.3 0.7 -7.8 Dividend yield 6.9 5.8 4 4 4.2 3.4 3.1 3.5 2.0 2.0 Price eanmings ratio (PIE) 5.1 7.4 7 7 -- - 15.2 11.9 12.5 18.0 Per capita income (in $)2 880 950 1030 1030 1030 1720 1980 3470 796 Srve: oww des Vakun de Clanca and Wodd Ban k Ea.Wl. '/ The October 1994 estimate is 14 percent. I/ World Bank Atlas methodology. (b) The stock market has made only a limited contribution to the mobilization of capital for the enterprise sector. The value of stock sold by listed companies plus the issuance of stock dividends nearly equals the value of the total turnover of stock at the exchange, largely because of the tax preference given to stock dividends. In 1993, when the total turnover at the exchange was DH 4.6 billion, stock worth DH 600 million was issued by listed companies, in addition to stock dividend distributions of DH 3.4 billion."7' 17/ The most important recent share issues were in favor of the Wafabank in late 1992 and ONA in early 1994. Annex IV-A  - 12 - Distribution of stock dividends is attractive to investors as, in contrast with cash dividends, they are untaxed even if they are subsequently sold. With the onset of the privatization program in 1993, four public enterprises were privatized through public offerings for a total of DH 1,9 billion." Each issue was several times oversubscribed, reflecting the interest in equity investment in Morocco. (c) Off-floor transactions, which result from direct negotiation between buyers and sellers, are three times more important than on-floor transactions. In fact, trading on-floor is extremely thin on most days, and the stock market cannot handle large block transactions yet. In addition, commissions on off-floor transactions are lower than on on-flooi transactions, explaining in large part the proliferation of the latter."9' This lack of liquidity in the market has led potential purchasers to ask market intermediaries (mainly the banks) to locate and acquire blocks of stock, which are transacted off-floor, at prices that are not necessarily related to the prices of on-market transactions. Foreign investors, who have become increasingly important on the stock exchange, have accumulated their portfolios though off-floor transactions. Hence, their participation has increased the volume of transactions, but not their transparency. 21. The free repatriation of capital and dividends makes the Moroccan stock exchange attractive to foreign investors. The price earnings ratio (P/E) is also rather low compared with other countries and adds to the attraction of the Moroccan stock market. (However, in the absence of unified accounting standards, such information must be taken with great caution.) About 10-20 percent of the capitalization of the stock exchange may be in foreign hands. In 1993, foreigners purchased one half of the new share capital offered (about $460 million), and several foreign mutual funds are presently accumulating Moroccan equity. Regulations pertaining to the operations of mutual funds have already been issued. 22. There is newly functioning money market in Morocco (see paras. 44-46 for the role of money markets in monetary policy). While a secondary market for Treasury paper sold at the weekly Treasury auctions is officially permitted, no such market exists, in part because of the lack of information regarding the outstanding supply of Treasury paper and problems regarding the transfer of ownership. The BAM does not publicly announce the amounts of paper auctioned off and outstanding. Only a limited secondary market in long-term Government securities exists at the Casablanca Stock Exchange. 18/ Compagnie des Transports Marocains-Ligne Nanonale, Ciments de L 'Oriental, an automobile credit agency SOFAC- Credit and the Societ Nationale d'Investissement. 19/ As the stock market develops and the accompanying legislation recommended below is put into effect, off-floor transactions should decline in importance. Annex IV-A  - 13 - 23. The recent creation of a bill discounting institution, the Socigtg Marocaine d'Intermgdiation Financi&re (SMIF), with capital contributed by a number of banks and the IFC, is a step in the right direction. The SMIF will draw, endorse and accept trade paper and discount Government securities. The SMIF could also act as a primary dealer in Government debt instruments, and provide placement support to the Treasury. However, the SMIF should not be given a monopoly position so that others can enter the market at a later sta'ge. 24. Legislation which was approved in December by Parliament is an important step towards the creation of a money market in Morocco. It permits transactions in short term paper issued by banks, finance companies and enterprises. Emissions of such paper will need to satisfy a number of financial conditions to be issued by the CDVM, while short term paper floated by enterprises will require approval from the CDVM. 25. Venture capital in an institutionalized format is in the early stages of development. Two venture capital enterprises operate in Casablanca. They provide management advice and capital to a limited number of established firms which could accede to the stock exchange at a later stage. A very large share of all new investments are financed with own resources. 26. Why Reform is Needed. Development of the capital market in Morocco should lead to substantial benefits in terms of improving resource mobilization, allowing better risk-hedging (through portfolio diversification, maturity diversification, etc.), and by monitoring and evaluating managers of firms. TIhese effects should raise investment and also the productivity of investment. One of the ways that stock markets enhance the productivity of investment is by allowing ownership changes without disruptive productive activities. By allowing (creditors) portfolio diversification opportunities, stock market development can also promote innovative investments. Savers can benefit from the increase in choice and the higher returns that become available. A multiplicity of institutions and players in a market leads to increased competition and greater efficiency in financial intermediation. 27. Recommended Reforms. The Government of Morocco agrees that a developed capital market is essential to sustain faster growth and in September 1993 created the CDVM, the equivalent to the Securities and Exchange Commission in the United States. The CDVM is to issue regulations that will permit enterprises to issue commercial paper and bonds, has issued regulations regarding the establishment of mutual funds that will invest in debt and equity instruments, and will supervise the issuance of, and transactions in equity instruments. 28. To promote trading in and the development of the above markets some additional measures are required: Annex IV-A  - 14 - (a) The CDVM is not fully operational yet, one of the problems being that staffing has not been completed."' Ensuring that the ComitL becomes fully operational should be a high priority of the Government. (b) CDVM must issue the disclosure and accounting requirements for enterprise bonds issuance; CDVM regulations should also permit banks or other agencies to issue mortgage backed bonds, which should provide long-term financing for onlending in the housing sector. (c) Ensure that the new accounting legislation (which requires standardized accounting and auditing) be applied rigorously (this legislation is to be applied for the first time for accounts of fiscal 1994). (d) Government guarantees issued to bonds of specialized financial institutions distort the market and reduce competition. They should be eliminated. (e) Enterprises should have access to foreign borrowing; this requires that they are able to cover their foreign exchange risk. This will be possible when a domestic money market and a foreign exchange market exist in Morocco (see below). (f) The procedures of the CSE regarding quotation and settlement must be modernized. The pending legislation should be issued. (g) The regulations licensing stock brokers must be issued. (h) For further deepening of the money market, development of the institutional and payments system, the creation of a multi-lateral market which could benefit from a screen-basis dealing network, training of the staff of major financial institutions, and the legal framework for recourse in case of failure to honor contractual obligations will need to be put in place immediately. 29. In addition, the authorities must ensure that all the regulations that have been passed are implemented. 20/ This is partly because civil service salaries may not be high enough to attract the necessary qualified staff. Annex IV-A  - 15 - 3. Reform of the Postal and Contractual Savings System 30. The Current Situation. Domestic savings in Morocco have risen gradually over the last 20 years but are still about seven percentage points of GDP lower than those in many low--and middle--income countries. Domestic savings are supplemented by many large transfers from nationals residing abroad (about 3-4 percent of GDP). Yet, national savings--which include these additional resources--are still several percentage points of GDP below those of comparator countries. As a result, Morocco's domestic investment financing has relied heavily on foreign borrowing. Such borrowing and its resulting debt burden would have been considerably less had domestic savings been more in line with that of comparator countries. 31. The mobilization of savings by the banking sector is inhibited by the prohibition against the remuneration of sight deposits; and the lack of competition for loanable funds (given the existing restrictions on lending and portfolio diversification). Bank deposits are the most important financial savings instrument in both the developed and emerging markets, but are significantly more important in the latter than in the former. Collective and contractual savings together represent 20 percent of the financial assets in developed countries, but only 7 percent in developing countries. The importance of financial savings mobilized through the long-term debt and equity markets is also much more important in developed than in developing countries (see Chart 5). Chart 5: Morocco - Structure of Selected Financial Systems, 1985 De,iopMe Mwkeu C_-d &~~~~D. b EmetliW Mwkets Source: WDR, 1989 Table 2.4 Annex IV-A  - 16 - 32. Postal Savings' The Caisse d'Epargne Nationale's (CEN) share of total savings mobilization has fallen steadily from 25 percent in 1985 to only 15 percent in 1993, in which year the increase in deposits were negligible, even though the network of the CEN (1319 branches at the end of 1993-- as many as all the other financial establishments combined) is growing. The CEN pays a fixed return to its depositors and deposits all its resources at the Caisse de Depbt et de Gestion (CDG) from which it receives a 9 percent return. In 1993 the financial margin between interest received from the CDG and interest paid to depositors was a low 1 percent, apparently an insufficient incentive to invigorate the mobilization of savings through the CEN. The average size of the deposits at the CEN is only 3500 DH. 33. The Moroccan authorities have realized that the savings mobilization potential of CEN has not been fully utilized, and in 1992 modified the regulations pertaining to the operations of the CEN, permitting it to engage in a broader range of banking activities. Yet, these regulations have not been implemented, in part because of the lack of enthusiasm on the part of the supervisory Ministry (Postes et Telecommunications). A study group (Treasury, BAM, CDG and the Ministry itself) was instructed to recommend a strategy to better tap the potential of the CEN. The results of this study group was not available to the mission. 34. Collective Savings Instruments and Institutional Savers. Regulations governing the creation of Mutual Funds, open-ended and closed-ended, that would invest in public and private sector debt instruments and in equity, have already been issued. It is believed that they will create a market for debt and equity instruments, while providing investors with a greater choice of investment vehicles, permitting better risk management. 35. 7he Pension system in Morocco is run by four main organizations plus a few smaller pension programs run by the banking and insurance sectors. (Annex IV-B provides greater details on the two main pension systems in Morocco and some recommendations for reform). (a) The Caisse Marocaine de Retraite (CMR) is for civil servants who pay 7 percent each of their pensionable income. Membership is mandatory. In 1993 there were 776,577 civil servants contributing to the system. Even though Government is statutorily obligated to match this contribution, it pays only as much as is needed to permit payment of the pensions. In 1990-93 Government's contribution fell more than DH 4 billion short of its scheduled contribution; total arrears of the Government vis-a-vis the pension system is estimated at DH 12 billion (4.8 percent of GDP in 1993). The CMR has no reserves and given the demographics of the system serious problems in the medium- to long term are foreseen. Either the pensions will need to be scaled down,21' or the contributions will need to be increased. (b) Affiliation in the Caisse Nationale de Securite Sociale (CNSS) is also mandatory and covers all private sector employees (in 1993 there were 55,992 enterprises, of which one 21/ In 1990 the base for calculating new pensions was increased by inclusion of the housing allowance plus 50 percent of the "indemnites' while that part of the pension that is based on the salary part of the total compensation of the pensioner is indexed to the cost of living. Old pensions were raised by 15 percent. Annex IV-A  - 17 - quarter have no employees; 840,000 employees). It provides family allowances, short- term benefits in case of medical disability and death, and pensions. Each benefit category is in theory financed by separate contributions of employers and workers. The family allowance system was to be used to redistribute income while the pension system was to accumulate reserves. In fact, the various operations of the CNSS have been combined, hiding the fact that the surpluses are generated by the family allowance system, and the deficits by the pension system. As early as 1980, pension payments and prorated administrative expenses have exceeded contributions. Since 1985, the operational deficit has exceeded the interest received, and accumulated reserves have started to dwindle. Since 1989 the reserves of the pension system have stabilized; in 1992 they amounted to slightly more than six months of pension payments (DH 497 million versus DH 1,037 million). The 1993 reform of the pension system increased the contributions to the pension plan and reduced those for the family allowance component.22' At present, all the reserves of the CNSS are deposited with the CDG and are remunerated at an average rate of 8.8 percent.23' The new pensions contribution rates will allow the CNSS to generate surpluses in its pension system for only a few years to come, after which pension payments will commence to exceed contributions, a disequilibrium that will increase from then on. (c) The Caisse Interprofessionnelle Marocaine de Retraite (CIMR) is a non-profit organization which private sector employers (2,748 at end-1993) can join on behalf of their employees (212,176 at end-1993) to ensure a supplementary pension to that of the CNSS. Pension contributions are made both by the employer and the employee at a rate between 3 percent and 6 percent of the actual monthly salary, depending on the agreement signed by the enterprise. Benefits are calculated on the basis of a points system, which itself is a function of the contributions made. Employer's contributions are paid into the CIMR, while employee contributions are retained in their name in an insurance company chosen by the employer. The employee can recover his accumulated contributions, while the employer's contribution must stay with the CIMR. CIMR reserves amounted to DH 2.5 billion at the end of 1993, or 3.6 times the amount of pensions paid that year. This ratio has fallen from 4.9 in 1982, prompting a slight revision of the contribution and pension calculations in 1992, intended to strengthen the system. Reserves are invested in a portfolio of foreign (7 percent of total) and domestic assets (93 percent of total). The foreign share of the portfolio is an inheritance from pre- independence years. Since that time no new foreign investment by domestic entities have been permitted. The CIMR portfolio is comprised of Government debt (25 percent), fixed income instruments (21 percent), real estate investments (4 percent), and stocks (50 22/ The 1993 reform of the CNSS adjusted from DH 3500 to DH 5000 the salary ceiling on which contributions are calculated as well as the base salary on which the pensions are being calculated. The contributions rates for the different systems run by the CNSS were also adjusted from 15.7 percent to 18.7 percent, with the share of contributions going to the pension system increased from 5.04 percent of base salary to 9.12 percent; the increase in contribution rates were phased in over a number of years. 23/ Deposits prior to end-1989 earn 8.5 percent, while deposits made since earn 9 percent. These rates of compensation are to be reviewed in theory every three years. Annex IV-A  - 18 - percent). The investment experience of the CIMR has been positive. In 1993 the rate of return on its assets was 11.5 percent, while unrealized capital gains amounted to about DH 900 million. There are no Government regulations (excepting on new foreign investments, mentioned previously)--affecting portfolio composition. Insurance companies manage their pensions reserves according to rules set up for them (see below). (d) The Caisse Nationale de Retraite et d'Assurances (CNRA) provides pensions to contractual workers in Government service, and is managed by the CDG. Affiliation is mandatory and in 1993 there were 184,525 contributing workers. At end-1983, total reserves amounted to DH 8.3 billion, 35 percent of which was invested in Government paper, 18 percent in Government-guaranteed bonds and 29 percent in equity. Overall returns on investments were 24 percent, including unrealized capital gains. (e) A number of state enterprises run their own pension schemes, while private schemes are also run by insurance companies. Professional associations maintain supplementary pension schemes, some of which are managed by the CDG. Their reserves amount to at least DH 2.5 billion, with DH 2.3 billion accounted for by OCP. 36. Insurance companies are a fast growing sector in Morocco; premia and accumulated reserves have grown at an average rate of 15 percent in 1992-93. Car insurance and pension plans are the most important branches in this sector, with respectively 36 percent and 21 percent of the total premia paid in 1993.' As discussed above, most of the pension plans managed by the insurance companies are related to the CIMR. Automobile insurance has realized losses for the last few years, and a number of companies heavily involved in this sector have gone bankrupt. This situation is partly explained by the fact that automobile insurance premia, which are fixed by Govemnment decree, are insufficient to cover the ensuing liabilities. In September 1994 these premia were raised by 12 percent, whereas a 36 percent increase would have been required to restore health to this sector (1992 study by the Ministry of Finance). The investments of the insurance companies, which totaled DH 20 billion at end-1993, are subject to prudential guidelines set by the Ministry of Finance. These establish that 40 percent (reduced from 60 percent in 1992) of total reserves must be invested in Government or Government-guaranteed bonds. At end-1993, 47 percent of total reserves were so placed, reflecting the absence of marketable securities; only 14 percent of reserves were invested in the equity market. 37. Why Reform is Needed. Reform of the postal and contractual savings systems are very important for Morocco and will help in achieving the high savings rate needed to finance higher growth. The postal savings system can be particularly useful in areas where there are few bank branches. These institutions can (if allowed to) also provide a significant boost to capital market development and thus raise the efficiency of investment. They provide an efficient way of combining small amounts of savings in order to finance larger investments and raise activity in the capital market. Country experience has shown that these effects can be important. In addition, pension systems should ideally be designed so as to reduce 24/ Present accounting practices do not permiit us to separate the reserves of the various branches of activity of the sector. The new accounting framework that is under active review, would require these companies to report their reserves per branch of activity. Annex IV-A  - 19 - labor market distortions (related to both labor costs and mobility). Annex IV-B provides greater detail on pension system reform. 38. Utilizing the already established mail-delivery system for the mobilization of savings has paid off handsomely in many East Asian countries and was used with success in most European countries at an earlier stage of their financial development, largely because their intermediation costs could be reduced by sharing an already existing infrastructure. More than any other financial institution, the postal savings system has been successful in mobilizing savings from small savers and the rural population, which was otherwise deprived of means for financial savings. In Japan and Italy the postal systems still collect around 30 percent of total personal savings. Management of the postal savings system should evolve together with the financial sector as a whole, and should not compete unfairly with the banking sector (as the Japanese postal system is presently been accused of).2' Its management should be made independent from that of the postal system itself, as in many instances, postal employees are not enthusiastic promoters of savings. To improve CEN's performance, restrictions on its investments should be removed so that it can invest in a wider variety of instruments and achieve a better risk-return structure.26' 39. Recommended Reforms. To benefit from the advantages of well-developed savings institutions, the Government should: (a) Repeal the prohibition against the remuneration of sight deposits by banks, both for enterprises and individuals. (b) Reactivate the Study Group that is to review the future role of the CEN and aim at putting in place a strategy to empower the CEN to enhance its role in savings mobilization. Provide CEN with greater freedom in its investment policy. As the CNRE is a relatively weak organization, its capacity to fulfill these new tasks must be strengthened. (c) CDVM should swiftly implement adequate supervisory capacity to protect investors in mutual funds. (d) Review the public pension system. The medium-term financial viability of the CMR and the CNSS is at risk. A thorough analysis of their contribution rates and pension calculations are required. (See Annex IV-B for detailed recommendations). The creation of a Work Group to study the pension funds is recommended. Two short-run recommendations are: 251 'Japan's Ambitious Postal Bank", The Ecounomist, October 29, 1994. p. 89-90 26! Malaysia separated the management of the savings system from that of the post office in early 1970's, when the proportion of saving mobilized by the postal system declined. The postal system continued to use the post offices as a deposit- taking branch network. The East Asian Miracle, a World Bank Policy Research Paper, published for the World Bank by Oxford University Press, 1993, p. 219. Annex IV-A  - 20 - i) Government contributions to CMR should be made as scheduled even if this system is to be run as a redistributive instrument. Activities that belong more in the general budget than in a pension system, should be separately managed and financed. CMR should be set on a sound footing. ii) Broaden the investment options for the CNSS, providing for improved rates of return and enhancing its medium- to long-term viability. While depriving the Treasury from a captive source of financing for its deficit, it would provide the private sector with much needed longer term financing resources. Once broader investment options are granted, prudential regulations would need to be issued. (e) The prudential regulations, as they pertain to the investment policies of the insurance companies should be modified. These should permit insurance companies to include investments in top graded private bonds alongside with Treasury paper in their 40 percent share of reserves that at present must be invested in Treasury or Treasury guaranteed paper. Again, this will only become fully operational when a rating system exists in Morocco. In the meantime, such a policy change would serve as an additional message to the investment community that the Government is serious about developing the capital market. 4. Eliminating the Government's Preferential Access to Credit 40. The Current Situation. At present, the Treasury has preferential access to financial resources through the following mechanisms: (a) Plancher des Effets Publics (PEP); banks are required to hold 25 percent of their sight deposits in public paper that pays 4.25 percent, substantially below market rates.27' Paying market rates (e.g. 10.5 percent) on these resources (DH 14 billion at end-1993) would have required additional expenses of DH 875 million in 1993, or .35 percent of GDP. Hence, the PEP is a disguised tax of that amount on the banking sector, a tax that is most likely passed on to the borrower in terms of higher lending rates. When the monetary reserve was reduced from its temporary peak of 25 percent to 10 percent of sight deposits in 1992, banks were required to shift the difference into seven year Treasury paper earning 10.5 percent with four years grace on interest payments. As a result, a full 27 percent of bank deposits is in mandatory placements with the Treasury. Nevertheless, because of the excess liquidity in the economy and the lack of alternative investments, banks hold an additional 18 percent worth of their deposits in Treasury paper. (b) The Treasury has interest free access from the BAM of DH 5.5 billion under the "avances conventionnelles" plus the equivalent of 10 percent of the previous year's tax 27/ In 1993 the share was reduced from 32 percent to 25 percent, but the difference was rolled over into 10-year paper at 10.5 percent. Annex IV-A  - 21 - revenue ("avances statutaires"). These "avances" were not always used to their maximum in 1993, providing the Treasury with a safety valve averaging about DH I billion. At a rate of 10.5 percent these "avances" would have cost DH 1.1 in 1993, or .4 percent of GDP. (c) All the reserves of the CNSS and the CEN are required to be deposited with CDG at fixed interest rates (9 percent at the end of 1994), and at end-1993 comprised 72 percent of CDG's total resources. CDG held 8.4 percent of all domestic debt at the end of 1993, a share that is growing, particularly with respect to the paper sold at auctions, where its share of outstanding paper rose from 30 percent at end-1993 to 45 percent at end-August 1994. At the end of 1993, Treasury paper comprised more than 40 percent of all CDG assets. (d) Up to December 1993, interest earned on Treasury paper held by private individuals was exempt from income taxes; a considerable fiscal advantage compared with full taxation of interest earned on other debt instruments. This fiscal advantage was halved in the 1994 Finance law and abolished totally with the 1995 Finance Law, creating a level playing field between the different debt instruments. (e) Insurance companies are required to invest 40 percent of their reserves in Treasury or Treasury guaranteed paper (see above). Ostensibly done for prudential reasons, this regulation creates a captive market for Treasury paper. (f) The Banque Centrale Populaire (BCP), is a State-owned bank which mobilizes its resources from a broad domestic network and from Moroccan residents abroad (64 percent of its deposit base at end-1993). Its charter instructs it to give preference to Treasury paper, the sector of small-and medium-sized enterprises and the tourist sector. At the end of 1993 about 55 percent of its outstanding credit went to the Treasury, often acquired after being asked to participate in its weekly auctions. 41. Why Reform is Needed. The Treasury should abandon its privileged position in the capital market and finance its deficit at market terms. By obtaining preferential access to credit, the Government is preventing the financial system from performing its role of mobilizing and allocating financial resources to their most productive uses, and is crowding-out the private sector. 42. Recommended Reforms. The following reforms should be phased in: (a) The PEP should be eliminated. (b) The requirement for the CNSS and CEN to deposit all their resources with the CDG should be eliminated; they should be given the authorization, within prudential guidelines, to choose the investment vehicle they deem most appropriate for them. (c) Advances of the BAM should be remunerated at market rates, to put the public sector on the same footing as the private sector in its competition for financial resources. This Annex IV-A  - 22 - would increase the cost of borrowing for the Government but there would also be additional profit taxes and profit remittances from the BAM to the Treasury. (d) The mandate of the BCP should be broadened to permit greater portfolio diversification; such mandate would also enhance much needed competition in the banking sector. (e) The domestic debt management system should be strengthened. Standardizing the issues offered at the weekly auctions would help, as well as issuing new paper (e.g. lottery features, early repayment options). Advertizing the amounts outstanding, as well as promoting the secondary market in Treasury paper should be part of this action plan. (f) The investment guidelines for insurance companies should be broadened, specifying that highly rated bonds (AAA) could be accumulated in reserves alongside with the Treasury and Treasury guaranteed bonds. This would broaden the demand for private sector bonds. 5. Accelerating the Use of Indirect Instruments for the Conduct of Monetary Policy 43. The Current Situation. The Government has in recent years shifted from an excessive reliance on direct policy instruments to somewhat greater use of indirect policy instruments. Credit ceilings were abolished in early 1992 and changes in reserve requirements have been actively used since 1986 to control the overall level of credit in the economy. The monetary policy instruments available to the BAM consist of reserve requirements with the banks, the rediscounting of commercial paper, auctions and repurchase agreements. In addition, the BAM sets maximum lending rates and certain minimum deposit rates to influence the interest rate structure in the economy. More specifically, (a) Reserve requirements can be set between 10 percent and 25 percent of sight deposits and up to 10 percent of time deposits. At present, the reserve requirement is set at 10 percent of sight deposits. (b) Rediscount policies are mainly used to channel resources to the export sector and to small- and medium-enterprises; together these accounted for 90 percent of all red iscounts in 1993. They are granted at the preferential rate of 7 percent. The amount of outstanding discounts has fallen sharply in recent years and stood at DH 1 billion at the end of 1993, compared with an average of DH 3.4 billion in 1992, largely because of weak credit demand. Rediscount policy has so far not been used to implement monetary policy. (c) There is currently a newly active money market in Morocco. An interbank market has begun to play an important role in interest rate determination. 44. Why Reform is Needed. Further development of indirect policy instruments will increase the options available, probably allow financing of a larger budget deficit, and increase the efficiency of resource allocation while allowing more flexibility in the conduct of monetary policy. Annex IV-A  - 23 - 45. Recommended Reforms. The creation of an active secondary market for Treasury paper is on the agenda of the Government, and should permit the BAM gradually to adopt open market policies. This process should be accelerated.2"' Effective management of an open market policy will require that: (a) the capacity of the staff in the BAM to engage in risk analysis and to monitor the liquidity needs of the economy be strengthened; (b) the infrastructure to implement the open market policy (trading room computerized network with traders and bankers) be set up; 46. Consideration should be given to remunerating the monetary reserves that commercial banks are required to hold with the BAM. The absence of remuneration is equivalent to a tax on banking resources and pushes the interest rate structure up. Unless a clear rationale can be made why borrowers must bear this additional "tax" monetary reserves should be compensated at market terms. In any event, the Treasury, as the largest borrower in the economy, would be a major beneficiary of the measure. 6. Integrating Morocco into World Financial lMarkets 47. The Current Situation. In recent years, and particularly after the announcement by King Hassan ll in 1991 that Morocco would have current account convertibility of the dirham in 1993. Along with current account convertibility (formally achieved with Morocco accepting the obligations under Article VIII of the Articles of Agreement of the IMF in January 1993), Morocco has established virtually full capital account convertibility for foreign investors."' Furthermore, the relaxation of surrender requirements and foreign borrowing restrictions, as well as the authorization of external investments by certain export firms and banks, represented a first move towards capital mobility for resident Moroccans. 48. There is no foreign exchange market at present in Morocco. Repatriation of all foreign exchange proceeds is required. Exporters have been permitted to retain a small share of their exports earnings with their banks (10 percent for exporters of goods and 5 percent for exporters of services) since 1993; the rest must be sold to the BAM. Banks are allowed to trade the foreign exchange they do not remit to the BAM, amongst themselves, but not against dirhams. Therefore no foreign exchange market exists as yet in Morocco. This greatly limits the possibilities available to resident traders and investors for covering 28/ Repurchase agreements have been the major method of open market policy in France, and were very recently introduced as one method of conducting open market pohcy in the United Kingdom. 29/ For a background paper on this issue of convertibility see Saleh Nsouli, Peter Cornelius and Andrea Georgiou, "Striving for Currency Convertibility in North Afiica", Finance and Development, December 1992, Volume 29, no. 4, pp.44-47. The preconditions for implementing currency convertibilty are (i) maintenance if internal and external balance, which require low inflation and an exchange rate that sustains the competitivity of the economy; (ii) adequate level of reserves to permit the country to absorb domestic or foreign shocks, without reintroducing restrictions, while domestic policies and/or the exchange rate are adjusted; (iii) the incentive system must be hberalized to permit the positive effects of the convertibility on resources allocation to be transmitted to the rest of the economy. Annex IV-A  - 24 - their foreign exchange risk, in the absence of which foreign borrowing tends to be limited to economic agents that have their own foreign exchange earnings. 49. Why Reform is Needed. A consensus is emerging that full convertibility of the dirham is within reach and would be beneficial.' Such a move would help in further attracting investment from abroad, by providing a strong signal regarding the irreversibility of reform and the commitment of the Government towards maintaining macroeconomic stability.3" It would also enable resident economic agents to obtain higher risk-adjusted rates of return by allowing them to hold internationally diversified portfolios. 50. Recommended Reforms. The most important precondition for moving towards full convertibility without incurring the danger of destabilizing capital outflows, is to ensure that the investment community is satisfied that Government intends and is able to maintain sound macroeconomic policies. Hence the requirement that monetary and fiscal policies ensure an environment of low inflation and that exchange rate policy preserves the competitivity of the Moroccan economy. Full capital account convertibility would of course require that the repatriation requirement be abandoned. Reforms of the financial sector should be pursued, in particular the prudential regulations covering foreign exchange positions should be put in place before full convertibility is implemented. 51. For a foreign exchange market to be put in place, an action plan must be prepared. Other key elements of this reform package are the following: (a) BAM must set prudential ratios pertaining to the open foreign exchange position dealers can have (for instance as a ratio of a dealer's capital--in Tunisia this ratio is 5 percent); procedures for licensing dealers and withdrawing licenses must be established. (b) BAM sets accounting and reporting standards to permit it to follow the foreign exchange activities of the authorized dealers, (which in the early stages of the reform may be limited to banks, so as not to unduly stretch the supervisory capacity of the BAM) and strengthen its supervisory and monitoring capacity. (c) The share of foreign exchange receipts that exporters are allowed to retain should be gradually increased to 100 percent. 30/ See Saleh M.Nsouli, Sena Eken, Klaus Enders, Van-Can Thai., Jorge Decressin, and Filippo Cartiglia, Resilience and Growth Through Sustained Adjustment. The Moroccan Experience, International Monetary Fund, Washington, DC, January 1995. p. 39 This paper details the liberalization measures that have been taken in recent years. 31/ The level of foreign exchange appears sufficiently strong to withstand a severe term of trade shock. Assuming that the balance of payments were hit during the same year by (a) a 30 percent drop in the price of phosphates (similar to the reduction that took place over two years in 1982-83): (b) a 20 percent increase in petroleum prices, (c) a 20 percent drop in tourism (akin to the effect on tourism of the Gulf War), and (d) a drought that reduces agricultural exports by 16 percent (same as in 1992) and increases imports of cereals and other food iniports by 47 percent (same as in 1992), the current account would deteriorate by $1.4.billion. This equals to about one third of total external reserves at the end of 1994. Hence, these reserves are sufficiently high to withstand such a shock. However, if the shock were to be more than temporary, macroeconomic policies would need to be adjusted in a timely fashion to avoid a chronic loss of reserves. Annex IV-A  - 25 - (d) BAM must give dealers the authorization to trade foreign currencies amongst themselves and against dirhams in the domestic and the international market. This interbank market will provide BAM with feedback on the direction of market forces. (e) BAM must establish a desired target rate for the dirham, and through intervention in the market be prepared to defend this rate. BAM may choose to publish only central rates without specifying buy and sell margins, as is done in Tunisia, or may not publish these central rates and convey its view of the desired rate to the market through its intervention policy. (f) The regulations pertaining to the opening of foreign exchange accounts by residents should be liberalized. While this is not essential to the creation of a foreign exchange market, it wouid strengthen such a market, and is advisable. In addition, dealers need to strengthen their capacity (technical and other) to engage in the various aspects of foreign currency dealings (swaps, futures, advise to clients). 52. Much progress has been made in endowing Morocco with a financial sector that can sustain and assist in promoting faster economic growth. Credit markets have been liberalized, and bank supervision has been modernized. Current account convertibility has been achieved while institutions have been modernized. However, much remains to be done to fully modernize the financial sector, which will be a prerequisite for improving resource allocation and promoting private-sector led growth. Many of the remaining reform issues have been widely discussed in Morocco in recent years and a broad agreement is emerging on what needs to be done. Some progress towards implementing some of these reforms is under way. Of essence now is the timely implementation of the overall reform framework. As discussed in this annex, many of the reforms are interdependent, so that their beneficial effects will only be felt if the various elements are implemented in a mutually supportive fashion. For instance, a well functioning money market should provide market related interest rates, without which the foreign exchange risk coverage mechanisms cannot function. Announcing the path of reform in advance and adhering to this schedule, will enhance the transparency of the process and permit the various participants to prepare themselves to take full advantage of a new developed financial sector. Annex IV-A  I  ANNEX IV-B PENSION SYSTEMS IN MOROCCO A. Introduction 1. Pension systems can play an important role in promoting growth by mobilizing resources, and boosting capital market development. They also play an important role in the labor market by providing social security. Pension system reform is recommended in Morocco to (a) promote financial sector development and (b) reduce labor market distortions and the informalization of production and employment while providing well-designed, well targeted social security."1 2. The objectives of most pension systems are threefold: the redistribution of income, the enhancement of savings, and insurance. The role of government in providing each of these services is different. Pension systems can take various forms: they can be voluntary/mandatory, pay-as-you-go (PAYG) or fully-funded (FF), have defined benefits or defined contributions, be public or private, be single pillar (one-kind) or multi-pillar (many different kinds co-exist)2". 3. The Moroccan pension system is dominated by the public sector. Affiliation in the public system is mandatory and they are PAYG systems with defined contributions. There are also a number of other organizations providing a variety of services, some providing only retirement income, others covering short term disability, health, family allowances, and other services. However, there are four primary organizations, the Caisse Nationale de Securite Sociale (CNSS) which serves employees of private businesses and agricultural wage earners, the Caisse Marocaine de Retraite (CMR) which serves government employees, the Regime Collectif d'Allocation de Retraite (RCAR) which serves contractual public sector personnel, and the Caisse Interprofessionnelle Marocaine de Retraite (CIMR) which serves private employees on a voluntary business. In addition, there are a variety of employer-based social security funds, private and public mutual insurance funds, employer-based mutual insurance funds, and health insurance companies. This annex looks at the two main public sector pension systems discussed above, in greater detail. 4. CNSS is the largest of the funds with 840,000 contributors. Affiliation in the CNSS is mandatory. It is currently running large surpluses, requiring only 92 percent of current revenues to fund current expenditures. These surpluses are expected to increase through the year 2015 as the working age population increases relative to the retiree population. Any increase in unemployment rates, above their current levels, however, will reduce these surpluses considerably. The surpluses are due to a very high contribution for family allowances, not subject to a wage ceiling, the proceeds of which are being used to subsidize retirement pensions. In 1993, for example, 60 percent of the contribution revenue came from family allowance contributions, while only 38 percent of the expenditures went for family allowances. The retirement pension contribution, which is subject to a wage ceiling, provides only 36 percent of contribution revenue, but requires 47 percent of total CNSS expenditure. This imbalance is 1/ Pension Reform and Growth, Giancarlo Corsetti and Klaus Schmidt-Hebbel, World Bank, mimeo, May 1994. 2/ "Averting the Old Age Crisis", a World Bank Policy Research Report, 1994. Annex IV-B  - 2 - expected to worsen as the population ages. CNSS is attempting to correct this imbalance by raising the retirement contribution rates from 5.04 percent prior to 1993 to 9.12 percent by 1995 and reducing the family allowance contribution rate from 9.4 percent prior to 1993 to 8.87 percent. Nevertheless, by 2003, the ratio of the pension expenditure to total expenditure will exceed the ratio of the new retirement contribution rate to the overall contribution rate. And given that the retirement contribution rate is subject to a ceiling while the family contribution rate is not, suggests that the long-term pension fund will be in deficit long before 2003 even with the increase in rates. 5. CMR is the next largest fund with 746,216 contributors. Affiliation is mandatory. It is in much worse shape financially. This fund is currently just covering costs, since the State contributes only enough to avoid deficits from emerging; these payments are much less than the statutory obligations of the State. For example, in 1993, the state provided only 43 percent of its normal obligation, which is still high compared to the 19 percent it provided in 1990. As a result of the State's poor payment record, the CMR has not been able to accumulate funds like the other organizations. Pensions for the disabled from the military, resistors, and other non-contributors, are also paid out of these funds. Assuming that the State continues to pay 43 percent of its obligations and that the number of public sector employees grows in line with the growth in the working age population, the deficits will begin to disappear in the year 1998, but will reappear 20 years later. If the State pays less as it has in the past few years or the public sector does not grow substantially, the surpluses will appear later and disappear much faster. 6. CIMR, with 212,176 contributors, is a private voluntary pension fund. It is well run and pays pensions on a defined contribution basis. Being private, CIMR does not have access to the government budget in case of insufficient funds to pay pensions. Limited information was available on RCAR, with its 184,525 contributors, a public fund with fairly high reserves currently. 7. This annex is organized as follows: Section B contains information on the main pension funds. Section C presents the demographic profile of Morocco. Section D presents results from a simulation model, which looks at the financial health of the main pension funds, and Section E presents reform proposals based on international experience. B. Brief Description of Major Pension Funds 8. CNSS. The main pension fund, covering private sector employees, covered approximately 840,000 employees in 1993, and provides a variety of services. These include family allowances, family health assistance, short term pensions to cover short term sickness, maternity, birth, and death, and long term pensions to cover disability, old age, and survivor benefits. In addition, CNSS operates 13 clinics and provides limited additional health care. 9. Most of the CNSS revenues are derived from contributions, collected from employers and employees. Funds totaling DH 8,5 trillion have been accumulated by end-1993, equivalent to almost 3 years' expenditures, interest on these reserves provides 18 percent of CNSS revenue. Contributions are specifically divided between the three major functions: family allowances, short term pensions, and long term pensions. While the rates have been gradually changed since 1993, by March 1995, employers will pay a total of 15.39 percent, 8.87 percent for family allowances, 0.44 percent for short term pensions, and 6.08 percent for long term pensions. Employees pay a total of 3.26 percent, 0.22 percent for short Annex IV-B  term pensions, and 3.04 percent for long term pensions. For both employers and employees, contributions for all pensions are subject to a ceiling of DH 5,000 per month, while the family allowance is not subject to such a ceiling. The ceiling covers 76 percent of the average wage bill in 1993. 10. Old age pension benefits are 50 percent of average covered earnings during the last 3 or 5 years, whichever is higher, plus an additional I percent for every 216 days in excess of 3240, for a maximum benefit of 70 percent. After a period during which pensions were not adjusted (1978-93), an adjustment was phased in for older pensions, and a new calculation was introduced in 1993. The average old age pension is only 46.4 percent of the average covered salary, indicating a fairly sharp erosion in the pension. Surviving widows or widowers get 50 percent of the pension, if age 50 or disabled, while children get 25 percent each if under age 12 or a student under age 21. The maximum survivors' benefit is 100 percent of the insured's pension. No method for adjusting pension payments to inflation or average wages is foreseen, though in principle such adjustment is possible. 11. Sickness benefits are 66 percent of salary up to 52 weeks, with a minimum benefit of 66 percent of minimum wage. Maternity benefits are 100 percent of daily earnings payable for 6 weeks prior to birth and 6 weeks after birth. 12. Family allowances are DH 54 per month for each of the first three children under the age of 12 or 21 if the child is a student, and DH 36 per month for each additional child up to the sixth. Family medical benefits also pay for up to DH 300 of services for first child and up to DH 100 each for subsequent children through the sixth. 13. The cross-subsidization of Chart 1: Distribution of Revenues and Expenditures, the various benefit types, suggested CNSS in the introduction, is shown in Chart 1. While contributions for family allowances constitute almost half of total revenue (49.9 percent), expenditures for family allowances are a much smaller share (32.6 percent). The short term sickness ".A-_ F.RK. pensions almost break even with l l the contributions providing 2.9 percent of revenue and the expenditures comprising 3.9 percent of total expenditures, but the big subsidies go to long term pensions where the revenue generated (29.7 percent) is far below the expenditures (40.4 percent). 14. The new contribution schedule will move toward remedying some of this imbalance as the pension contribution rate rises from 5.04 percent to 9.12 percent and the family allowance contribution rate is being reduced slightly, from 10 percent to 8.87 percent. But given that the pension contributions are subject to a wage ceiling, while the family allowance is not, the revenue share of pensions may still be under 50 percent. Furthermore, as the population ages and the family size decreases, the family allowance expenditure will fall, while the long term pension expenditure will rise. Annex IV-B  - 4 - 15. CNSS has accumulated hefty reserves, amounting to DH 8,5 trillion at end-1993. These are deposited at the Caisse de Dep5t et de Gestion, a state owned bank. CNSS has been earning 8.81 percent on its deposits, which when compared with an average inflation rate of 6.9 percent, does provide a positive real rate of return. However, when compared with the rate of return received by other pension funds, like CIMR which earns 13 percent and has near total freedom in its investment policy, the rate of return is significantly lower. 16. CMR. CMR, the second largest pension fund in Morocco, covers government employees only, approximately 776,577 employees at end-1993. The employees receive old age pensions, disability pensions, survivor pensions, and family allowances. The majority of covered employees work for the State itself, 89 percent, with a small number of employees from local governments, 8 percent, and from public enterprises, 3 percent. 17. By law, the Fund receives contributions from the State as employer and from the employee. The contributions are supposed to be split, 7 percent from the employer and 7 percent from the employee, on a wage base which consists of the base salary, a housing allowance, and 50 percent of other bonuses. Since these bonuses are a larger share of a higher paid worker's salary, the covered wage is 83.5 percent of the lowest income worker's total wage, but 65 percent of the highest paid worker's wage. The State is also obligated to cover non-contributor benefits, such as those for resistors. However, the government chooses to contribute only enough to keep the fund liquid, not its full share. The government contribution has varied from 19 percent of what was due in 1990 to 43 percent of what was due in 1994. In 1990-93, the government contribution fell DH 4 billion short of its statutory obligations. Overall these arrears amount to DH 12 billion equivalent to 40 percent of 1994 GDP. 18. Generally, pension benefits are annuities worth 2.5 percent of the covered wage immediately prior to retirement per year of service, subject to a minimum pension of DH 7,962 per year after 20 years of service. Surviving spouses receive 50 percent of the insured's pension and orphans receive a share of the remaining 50 percent. However, as noted for CNSS pensions, inflationary erosion has decreased the average pension benefit to only 36 percent of average wage. 19. CMR also provides benefits to specific groups of non-contributors, such as military disabled and resistors. The total number of beneficiaries, 256,406 can be divided into contributors and non- contributors. Contributors include civil servants, military personnel, and members of the Auxiliary Forces. Beneficiaries in these categories, including the disabled, number 205,597 or 80.2 percent of the total, but receive 88.5 percent of the expenditures. Non-contributor beneficiaries number 50,809 and receive 11.5 percent of the benefits. Annex IV-B  C. Demographic Outlook for Morocco 20. Morocco is demographically a young country. Chart 2 compares the percentage of the population above age 60 in Morocco with the same percentage for other countries. Morocco is currently much younger than the OECD countries and some of the older developing countries, l ike Argentina, Chait 2: Percentage of Population Over Age 60, but about the same or a little 1990 younger than the bulk of the W_.- developing countries. 21. However, as Chart 3 shows, FroW.last i rTurkey6Vj the population is expected to age rapidly early in the next century. Morocco in 2050 will look much like France today, demographically, with the percentage of the population over age 60 more than tripling during that time period. Pension systems operated on a pay-as-you-go system or a partially funded system experience difficulties as the population ages, since the contributions from the workers, those aged 20-59, are being used to pay pensions for the retirees, those aged 60 and higher. In Morocco today, there are more than seven workers to support each pensioner. By the end of the period shown, there will be less than two workers available to support each pensioner. 22. As the retiree bar grows in relation to the worker bar, pension systems begin to experience financial difficulties which can be resolved by modifying the parameters of the pension system. Common solutions include raising the retirement age, reducing pension benefits, and raising contribution rates. However, the radical Chart 3: Coniparison of Workers to Retirees reforms required is often not 35000 politically feasible. For example, 30000 25000 in CNSS today, the ratio of 20000 average pension to average wage is 15000 46.4 percent with the current 10000 contribution rate rising to 9.12 5000 percent. When there are only two 1990 2000 2010 2020 2030 2040 2050 2060 2070 workers to support each retiree, Y, the contribution required to la Ag. 20-50 E Asy 001 support that same pension benefit would be 23.2 percent (ignoring for a moment the revenues obtained from accumulated reserves--or when reserves are exhausted). The enormous tax on labor that this implies would have a detrimental impact on labor costs and employment. 23. Alternatively, the current 9.12 percent contribution rate could be held constant and the pension benefit adjusted. In 2060 and 2070 when the ratio of workers to retirees is close to two, this contribution Annex IV-B  - 6 - rate would only support a pension equal to 18.24 percent of average wage. This pension is so low it might not even guarantee poverty relief for elderly workers. 24. This is precisely the predicament that the OECD countries find themselves in today. The third solution often proposed is to raise the retirement age so as to raise the number of potential workers and reduce the number of potential retirees. However, two caveats to this solution need to be noted. First, raising retirement ages does not necessarily mean that potential contributors increase. Employment opportunities for these workers must be available. Second, none of these are permanent solutions. As life expectancies continue to rise, the retirement ages will have to be raised in tandem. Similarly, contribution rates will have to continue rising and benefit rates will have to continue falling. 25. A more radical solution to this problem is to move away from the pay-as-you-go or partially funded defined benefit pension system toward a defined contribution system. In such a system, rather than being used to fund current retiree benefits, the contributions from each worker are invested and the pension is paid from the returns on that investment. Under normal conditions and based on international experience, the real rate of return on such investment should average about 4 percent. By contrast, the real rate of wage growth which constitutes the real rate of return on pay-as-you-go pensions in the absence of population growth has averaged only about 2 percent. Using these international norms, a 10 percent contribution over a working career of 40 years should yield a 40 percent of average wage benefit for a retirement period of 20 years. This benefit is far superior to what the defined benefit system can pay as suggested above. D. Simulated Results for CNSS and CMIR 26. CNSS. Before discussing reform proposals in more detail, it is worthwhile looking at some simulation results for fiscal balance for the two largest funds. Chart 4 shows the projected surpluses for CNSS from the present to the year 2075. Because Morocco still has a fairly young population and CNSS requires a hefty contribution, the fund will show a surplus for many years, until the year 2036. Once the fund begins to accumulate deficits, the deficits will increase quite sharply and will continue Chart 4: Annual Surpluses, CNSS 1993-2073 forever. The simulation takes l193 2143 2913 2023 M32 2943 2143 2943 2~3 CNSS revenues for 1993 and 2uesf.1 expands them with the growth in the working age population, under ... | | ........ the assumption that employment and evasion remain at current levels. Expenditures from 1993 ,,,,, are similarly expanded, by the . growth in the population over age 60 for the long term pension expenditure, by the growth in the working age population for the short term pension expenditure, and by the growth in the number of children for the family allowances. The projected population growth is taken from the World Bank Population Projections for Morocco. Annex IV-B  27. Despite the huge deficits accumulated beyond 2035, CNSS currently holds a healthy reserve which will be increasing during the surplus years. As a result, CNSS will not show losses until the year 2073 (Chart 5). The reserve fund balance is shown in Chart 5 and shows a steady build-up until 2045. Chart 5: Projected Reserve Fund, CNSS Fo hno,tersrei . ~~~From then on, the reserve is gradually depleted until it disappears in 2073. The reserve funds are 6DOMM000 eYele | 210 lW | l expected to continue earning an 8.81 percent rate of return as they have 2DDDM ~~~~~~~~~~done recently, in an environment 0 liiuuiliiliii ~~~where inflation averages 6.9 percent, yielding a real rate of return of 1.91 percent. Should the V_ funds be invested at higher rates of return, as the privately managed funds already receive in Morocco, the reserve accumulation would initially be larger, and the reserve would last longer. 28. However, this relatively positive picture for CNSS only occurs due to the massive cross subsidization that occurs from the family allowance contributions to the long-term pension __= expenditures. Charts 6 and 7 show z 0m the revenue and expenditure figures for the family allowances separated l_oom_ _ _ from the long term pensions. The 2C 2O7 20 275 surpluses generated by the family WU allowances grow substantially before RevU EieedItures leveling out around 2050, while the deficits generated by the long term Chart 7: Projected Revenues and Expenditures pensions grow continuously. It is not 14000000 for Long-Tenn Pensions clear that this cross subsidization is -i2oooWWl desirable, and employers may i'