76758 Choosing a System of Unemployment Income Support: Guidelines for Developing and Transition Countries Milan Vodopivec Mounting evidence suggests that excessive job protection reduces employment and labor market flows, hinders technological innovations, pushes workers into the informal sector, and hurts vulnerable groups by depriving them of job opportunities. Flexible labor markets stimulate job creation, investment, and growth, but they create job insecurity and displace some workers. How can the costs of such insecurity and displacements be minimized while ensuring that the labor market remains flexible? Each of the main unemployment income support systems (unemployment insurance, unemployment assistance, unemployment insur- ance savings accounts, severance pay, and public works) has strengths and weaknesses. Country-specific conditions—chief among them labor market and other institutions, the capacity to administer each type of system, and the size of the informal sector—determine which system is best suited to developing and transition countries. This article evaluates five systems of unemployment income support, identifying the key strengths and weaknesses of each system, assessing the applicability of each sys- tem to developing and transition counties, and drawing lessons about the design of income support systems. Two sets of criteria are used to judge the desirability of each type of system in a particular country. Performance criteria assess how well a system typically works—how well it protects workers’ incomes and how it affects efficiency and income distribution. Design and implementation criteria assess how suitable a system is in a particular country given the country-specific conditions. To evaluate the desirability of a particular income support system, the article examines theoretical and empirical insights about its performance and examines how the prevailing conditions in developing and transition countries are likely to affect its functioning. This pragmatic approach is used to supplement the findings of economic models, which can provide only partial guidance for policymakers in © The Author 2006. Published by Oxford University Press on behalf of the International Bank for Reconstruction and Development / THE WORLD BANK. All rights reserved. For permissions, please e-mail: journals.permissions@oxfordjournals.org. doi:10.1093/wbro/lkj003 Advance Access publication 14 February 2006 21:49–89 developing and transition countries, because implicit assumptions of theoretical modeling (such as the existence of suitable institutional capacity and strict enforce- ment of law and order) often do not apply. What Are the Options? Policymakers can choose among various income support systems for the unemployed. Five systems are examined here: unemployment insurance, unemployment assis- tance, unemployment insurance savings accounts, severance pay, and public works. Unemployment Insurance Unemployment insurance (together with pension and health insurance) is provided as part of social insurance. The system typically requires that workers and their employers pay earnings-related contributions that, upon separation, entitle workers to unemployment benefits according to predetermined eligibility conditions. While it mimics market insurance, the system deviates from actuarial principles by charging premiums that do not reflect individual risks. To qualify for benefits, a worker must satisfy the minimum covered employment or contribution requirement. The cause of dismissal may affect the individual’s eligibility for benefits, with workers who quit their jobs often disqualified. Continuing eligibility requires that applicants be avail- able for jobs, actively search for them, and be willing to accept them. Benefits are typically determined by the replacement rate—the proportion of the individual’s pay in the most recent employment spell. The initial replacement rate is usually 40–75 percent. If provided, unemployment insurance is mandatory in developed countries, where it is the most widely used income protection system for the unemployed, typi- cally covering all workers (table 1). Why is unemployment insurance provided by the public and not the private sec- tor? The market fails to provide unemployment insurance primarily because of sig- nificant information asymmetries, which give rise to “moral hazard� and adverse selection problems that cannot be handled by private providers (moral hazard arises because unemployment insurance reduces self-protection; adverse selection arises because information problems prevent insurers from charging higher premiums to poor risks). Correcting for market failures calls not only for regulation—obligatory membership to avoid adverse selection—but also for its public provision, to improve monitoring capacity. Another advantage of publicly providing unemployment insurance is the enhanced ability to pool resources across large groups, which reduces strains on the system arising from the covariant nature of unemployment risk. Moreover, financial backing by the government is often provided when the sys- tem faces financial difficulties. 50 The World Bank Research Observer, vol. 21, no. 1 (Spring 2006) Table 1. Unemployment Support Systems, by Country Unemployment Unemployment Insurancea Assistancea Unemployment Insurance Savings Accounts Public Worksc Reference Variables Replacement Rate (Percent Severance of Individual’s Replacement Contribution Payb Pre-unemployment Rate (Percent Rate (Benefits Wage where Maximum Maximum of Individual’s (Employer + After Expenditure Share of not Indicated Duration Replacement Duration Pre-unemployment Potential Worker 20 Years Participation as Percent Unemployment Informal d Country Otherwise) (Months) Rate (Months) Wage) Duration Obligation) of Service) Rates of GDP Rate Employmente Low-income countries Afghanistan n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — — — n.a. Angola n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — — — n.a. Armenia $4–$7 12 n.a. n.a. n.a. n.a. n.a. 2 — — 22.9 26.7 Azerbaijan 70% 6 n.a. n.a. n.a. n.a. n.a. 4 — — 1 — Bangladesh n.a. n.a. n.a. n.a. n.a. n.a. n.a. 9.3 0.2 — 2.5 — Benin n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — — — — Burkina Faso n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — — — 65 Burundi n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — — — n.a. Cambodia n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — — 2.2 n.a. Cameroon n.a. n.a. n.a. n.a. n.a. n.a. n.a. 6.75 — — — — Central n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — n.a. African Rep. Chad n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — n.a. China Flat (set below 12–24 n.a. n.a. n.a. n.a. n.a. 20 — — 3 21.9 minimum wage) Congo, Dem. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — n.a. Rep. of Congo, Rep. of n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — n.a. Côte D’Ivoire n.a. n.a. n.a. n.a. n.a. n.a. n.a. 6.8 n.a. n.a. — — Ethiopia n.a. n.a. n.a. n.a. n.a. n.a. n.a. 9.3 1 — — — Gambia, The n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 11.9 (Continued) Table 1. (Continued) Unemployment Unemployment Insurancea Assistancea Unemployment Insurance Savings Accounts Public Worksc Reference Variables Replacement Rate (Percent Severance of Individual’s Replacement Contribution Payb Pre-unemployment Rate (Percent Rate (Benefits Wage where Maximum Maximum of Individual’s (Employer + After Expenditure Share of not Indicated Duration Replacement Duration Pre-unemployment Potential Worker 20 Years Participation as Percent Unemployment Informal Country Otherwise) (Months) Rate (Months) Wage) Duration Obligation) of Service) Rates of GDP Rated Employmente Ghana n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 0.1 — — 72.3 Guinea n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — n.a. Haiti n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — — — n.a. Honduras n.a. n.a. n.a. n.a. n.a. n.a. n.a. 8 1.8 — 3.7 — India n.a. n.a. n.a. n.a. n.a. n.a. n.a. 6 2.1 — — 50.4 Indonesia n.a. n.a. n.a. n.a. n.a. n.a. n.a. 9 2.2 — 5.2 37.4 Kenya n.a. n.a. n.a. n.a. n.a. n.a. n.a. 10 0 — — 40.8 Korea, Dem. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 20 — — — — People’s Rep. of Kyrgyz Rep. 100%–150% of 6 n.a. n.a. n.a. n.a. n.a. 1 — — — 40 minimum wage Lao PDR n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — — — — Liberia n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — — Madagascar n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — n.a. — 57.5 Malawi n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — n.a. — 51.7 Mali n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 0.1 — — 36 Mauritania n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — — 28.9 — Moldova 50%–60% of — n.a. n.a. n.a. n.a. n.a. 2 — — 6.1 — average wage Mongolia n.a. n.a. n.a. n.a. n.a. n.a. n.a. 1 — — 5.7 11.5 Mozambique n.a. n.a. n.a. n.a. n.a. n.a. n.a. 28.5 — — — 72.5 Myanmar n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — — Table 1. (Continued) Unemployment Unemployment Insurancea Assistancea Unemployment Insurance Savings Accounts Public Worksc Reference Variables Replacement Rate (Percent Severance of Individual’s Replacement Contribution Payb Pre-unemployment Rate (Percent Rate (Benefits Wage where Maximum Maximum of Individual’s (Employer + After Expenditure Share of not Indicated Duration Replacement Duration Pre-unemployment Potential Worker 20 Years Participation as Percent Unemployment Informal Country Otherwise) (Months) Rate (Months) Wage) Duration Obligation) of Service) Rates of GDP Rated Employmente Nepal n.a. n.a. n.a. n.a. n.a. n.a. n.a. 20 — — 1.1 — Nicaragua n.a. n.a. n.a. n.a. n.a. n.a. n.a. 20 1.1 — 14.6 — Niger n.a. n.a. n.a. n.a. n.a. n.a. n.a. 6.8 0.2 — — — Nigeria n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — — 7.8 48.9 Pakistan n.a. n.a. n.a. n.a. n.a. n.a. n.a. 20 0.1 — 5.7 60 Rwanda n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — — — — Senegal n.a. n.a. n.a. n.a. n.a. n.a. n.a. 1 0.5 — — 62.4 Sierra Leone n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — — — — Somalia n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — — Sudan n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — — — — Tajikistan n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — — — — Tanzania, n.a. n.a. n.a. n.a. n.a. n.a. n.a. 12 — — — 42.2 United Rep. of Togo n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — — — — Turkmenistan Limit of three n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — — monthly wages Uganda n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — — 7.4 56.4 Vietnam n.a. n.a. n.a. n.a. n.a. n.a. n.a. 3 — — — — Yemen, Rep. of n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — — 11.5 — Zambia n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — — 15 — Zimbabwe n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — — 6.5 33.9 (Continued) Table 1. (Continued) Unemployment Unemployment Insurancea Assistancea Unemployment Insurance Savings Accounts Public Worksc Reference Variables Replacement Rate (Percent Severance of Individual’s Replacement Contribution Payb Pre-unemployment Rate (Percent Rate (Benefits Wage where Maximum Maximum of Individual’s (Employer + After Expenditure Share of not Indicated Duration Replacement Duration Pre-unemployment Potential Worker 20 Years Participation as Percent Unemployment Informal Country Otherwise) (Months) Rate (Months) Wage) Duration Obligation) of Service) Rates of GDP Rated Employmente Lower-middle-income countries Albania Flat 12–18 n.a. n.a. n.a. n.a. n.a. 10 — — 15.8 — Algeria 75%–300% of 12–36 n.a. n.a. n.a. n.a. n.a. n.a. — n.a. 28.3 — base Belarus 50%–70% 6.5 n.a. n.a. n.a. n.a. n.a. 3 — — 2.7 — Bolivia n.a. n.a. n.a. n.a. n.a. n.a. n.a. 20 6.8 — 3.9 51.3 Bosnia and 30%–40% 3–12 n.a. n.a. n.a. n.a. n.a. 6.7 — — — — Herzegovina Bulgaria 60% 4–12 Flat 3 n.a. n.a. n.a. 1 — 0 14.8 63 Colombia n.a. n.a. n.a. n.a. Lump sum 9.3% + 0 26.8 — — 13.6 53.8 Costa Rica n.a. n.a. n.a. n.a. n.a. n.a. n.a. 8 6.2 — 5.7 — Cuba n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — — Dominican n.a. n.a. n.a. n.a. n.a. n.a. n.a. 15.3 n.a. n.a. 16.1 52 Rep. Ecuador n.a. n.a. n.a. n.a. Lump sum 8% + 0 20 — — 12.8 58.8 Egypt, Arab 60% 4–7 n.a. n.a. n.a. n.a. n.a. n.a. 1.3 — 21.3 64.2 Rep. of El Salvador n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 0.6 — 7.7 — Georgia Flat 6 n.a. n.a. n.a. n.a. n.a. 1 — — 11.9 36.7 Guatemala n.a. n.a. n.a. n.a. n.a. n.a. n.a. 20 — — — — Iran, Islamic 55% of average 6–50 n.a. n.a. n.a. n.a. n.a. 20 n.a. n.a. — — Rep. of wage Iraq n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — Table 1. (Continued) Unemployment Unemployment Insurancea Assistancea Unemployment Insurance Savings Accounts Public Worksc Reference Variables Replacement Rate (Percent Severance of Individual’s Replacement Contribution Payb Pre-unemployment Rate (Percent Rate (Benefits Wage where Maximum Maximum of Individual’s (Employer + After Expenditure Share of not Indicated Duration Replacement Duration Pre-unemployment Potential Worker 20 Years Participation as Percent Unemployment Informal Country Otherwise) (Months) Rate (Months) Wage) Duration Obligation) of Service) Rates of GDP Rated Employmente Jamaica n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — — 16 — Jordan n.a. n.a. n.a. n.a. n.a. n.a. n.a. 20 n.a. n.a. 15 — Kazakhstan n.a. n.a. n.a. n.a. n.a. n.a. n.a. 1 n.a. n.a. 12.7 40 Latvia 90% 6 n.a. n.a. n.a. n.a. n.a. 1 — — 16 — Lithuania 50% 6 n.a. n.a. n.a. n.a. n.a. 24 — — 15 — Macedonia, 40%–50% 24 n.a. n.a. n.a. n.a. n.a. 10 — — 33.4 — FYR Morocco n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — — 19.8 — Papua New n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — — Guinea Paraguay n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 8.2 — Peru n.a. n.a. n.a. n.a. Lump sum 8% + 0 12 0.8 0 7.6 54.6 Philippines n.a. n.a. n.a. n.a. n.a. n.a. n.a. 20 0 — 8.6 30.6 Romania 50%–55% 9 30% 18 n.a. n.a. n.a. n.a. — — 6.8 42.7 Russian 45%–75% n.a. n.a. n.a. n.a. n.a. n.a. 1 — — 11.5 42.2 Federation South Africa 45% 6 n.a. n.a. n.a. n.a. n.a. 5 — — 21.2 — Sri Lanka n.a. n.a. n.a. n.a. n.a. n.a. n.a. 20 — — 11.3 31.3 Swaziland n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — — (Continued) Table 1. (Continued) Unemployment Unemployment Insurancea Assistancea Unemployment Insurance Savings Accounts Public Worksc Reference Variables Replacement Rate (Percent Severance of Individual’s Replacement Contribution Payb Pre-unemployment Rate (Percent Rate (Benefits Wage where Maximum Maximum of Individual’s (Employer + After Expenditure Share of not Indicated Duration Replacement Duration Pre-unemployment Potential Worker 20 Years Participation as Percent Unemployment Informal Country Otherwise) (Months) Rate (Months) Wage) Duration Obligation) of Service) Rates of GDP Rated Employmente Syrian n.a. n.a. n.a. n.a. n.a. n.a. n.a. 17.5 n.a. n.a. — — Arab Rep. Thailand n.a. n.a. n.a. n.a. n.a. n.a. n.a. 10 — — 1.9 — Tunisia n.a. n.a. Minimum 3 n.a. n.a. n.a. 3 n.a. n.a. — 57.1 wage Ukraine 35%–60% 12 75% of 6 n.a. n.a. n.a. 1 — — 9.1 — minimum wage Uzbekistan 50% 6 n.a. n.a. n.a. n.a. n.a. 2 — — 0.4 — Yugoslavia, n.a. n.a. 70% 3–30 n.a. n.a. n.a. 4 — — — — Fed. Rep. (Serbia/ Montenegro) Upper-middle-income countries Argentina 60% 4–12 n.a. n.a. Lump sum 12% (8% 10 6.4 — 15.9 — (available only after first in construction) year) + 0 Botswana n.a. n.a. n.a. n.a. 3 10.2 — 21.5 — Brazil 50% 6–11 n.a. n.a. Lump sum, 8% + 0 6 — — 7.9 49.2 plus 40% of individual’s pre- unemployment wage if not dismissed for “just cause�) Table 1. (Continued) Unemployment Unemployment Insurancea Assistancea Unemployment Insurance Savings Accounts Public Worksc Reference Variables Replacement Rate (Percent Severance of Individual’s Replacement Contribution Payb Pre-unemployment Rate (Percent Rate (Benefits Wage where Maximum Maximum of Individual’s (Employer + After Expenditure Share of not Indicated Duration Replacement Duration Pre-unemployment Potential Worker 20 Years Participation as Percent Unemployment Informal Country Otherwise) (Months) Rate (Months) Wage) Duration Obligation) of Service) Rates of GDP Rated Employmente Chile n.a. n.a. n.a. n.a. 30–50% 5 2.4% + 11 7.7 — 6.5 40 months 0.6% Croatia 60%–80% 2.5–10 16% of — n.a. n.a. n.a. 10 — — 11.2 70 average wage (in 2001) Czech Republic 40%–50% 6 n.a. n.a. n.a. n.a. n.a. 2 0.2 0 5.6 — Estonia n.a. n.a. $23 6 n.a. n.a. n.a. 4 — 0 10.2 — Gabon n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — — Hungary 65% 9–12 Flat 9 n.a. n.a. n.a. 5 2.2 0.2 8.7 — Korea, Rep. of 50% 3–8 n.a. n.a. n.a. n.a. n.a. 20 5.2 — 3.9 — Lebanon n.a. n.a. n.a. n.a. n.a. n.a. n.a. 10 — — 8.6 — Libya n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — — Malaysia n.a. n.a. n.a. n.a. n.a. n.a. n.a. 12 — — 2.9 — Mauritius n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — — 9.8 — Mexico n.a. n.a. n.a. n.a. n.a. n.a. n.a. 17.4 4.4 0.5 3.7 — Oman n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — — Panama n.a. n.a. n.a. n.a. Lump sum — 5 0.7 — 13.4 — Poland Flat, 23% of 6–18 n.a. n.a. n.a. n.a. n.a. n.a. 0.7 0.1 12 — average wage (in 2001) Saudi Arabia n.a. n.a. n.a. n.a. n.a. 17.5 n.a. n.a. — — (Continued) Table 1. (Continued) Unemployment Unemployment Insurancea Assistancea Unemployment Insurance Savings Accounts Public Worksc Reference Variables Replacement Rate (Percent Severance of Individual’s Replacement Contribution Payb Pre-unemployment Rate (Percent Rate (Benefits Wage where Maximum Maximum of Individual’s (Employer + After Expenditure Share of not Indicated Duration Replacement Duration Pre-unemployment Potential Worker 20 Years Participation as Percent Unemployment Informal Country Otherwise) (Months) Rate (Months) Wage) Duration Obligation) of Service) Rates of GDP Rated Employmente Slovak Rep. 45%–50% 6–9 Flat 6 n.a. n.a. 2 — 0 13 — Trinidad and n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 15.1 Tobago Turkey 50% 6–10 n.a. n.a. n.a. n.a. n.a. 20 — — 6.9 — Uruguay n.a. n.a. n.a. n.a. 50% 6 15% + 6 n.a. n.a. 10.5 — months 12.5% (compre- hensive insurance) Venezuela, R.B. 60% 3–6 n.a. n.a. Lump sum — 9 n.a. n.a. 11.9 — de High-income countries Australia n.a. n.a. 41%–99% No limit n.a. n.a. n.a. 2 0.3 0.1 7.9 6.4 of average wage (in 2001) Austria 55% of net 5–18 51% of net 6 Lump sum 1.53 9 — 0 4.9 16 earnings earnings (extend- able) Belgium 44%–60% No limit n.a. n.a. n.a. n.a. n.a. n.a. 3.2 0.5 9.1 — Canada 55% 11 n.a. n.a. n.a. n.a. n.a. 1.3 0.1 0.1 8.8 4.4 Denmark 90% 12–48 n.a. n.a. n.a. n.a. n.a. 3 0.8 0.2 5.9 15.4 Finland Up to 90% 16.5 Flat No limit n.a. n.a. n.a. 1 2 0.3 12.9 — France 57%–75% 4–60 — No limit n.a. n.a. n.a. 2 1.5 0.2 11.9 9 Table 1. (Continued) Unemployment Unemployment Insurancea Assistancea Unemployment Insurance Savings Accounts Public Worksc Reference Variables Replacement Rate (Percent Severance of Individual’s Replacement Contribution Payb Pre-unemployment Rate (Percent Rate (Benefits Wage where Maximum Maximum of Individual’s (Employer + After Expenditure Share of not Indicated Duration Replacement Duration Pre-unemployment Potential Worker 20 Years Participation as Percent Unemployment Informal Country Otherwise) (Months) Rate (Months) Wage) Duration Obligation) of Service) Rates of GDP Rated Employmente Germany 60% of net 6–32 53% of net No limit n.a. n.a. n.a. n.a. 1.6 0.3 9 22 earnings earnings Greece 40%–50% 4–9 n.a. n.a. n.a. n.a. n.a. 5.8 n.a. n.a. 9.8 — Hong Kong, n.a. n.a. $165– No limit n.a. n.a. n.a. 2 n.a. n.a. 3.8 — China $231 per month (2002) Ireland Flat, 13% of 15 Flat, 13% No limit n.a. n.a. n.a. 2.5 3.7 0.6 9.6 — average wage (in of average 2001) wage (in 2001) Israel 40%–80% of 4.5/year n.a. n.a. n.a. n.a. n.a. 20 — n.a. 7.7 — average wage Italy 40%–80% 6–27 n.a. n.a. n.a. n.a. n.a. 18 — 0 11.5 39 Japan 60%–80% 10 n.a. n.a. n.a. n.a. n.a. 20 n.a. n.a. 3.8 — Kuwait $250–$680 (in 12 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. — — 2002) Netherlands 70% 6–60 Flat, 24 n.a. n.a. n.a. n.a. 0.3 0.4 5.4 — 70% of minimum wage (Continued) Table 1. (Continued) Unemployment Unemployment Insurancea Assistancea Unemployment Insurance Savings Accounts Public Worksc Reference Variables Replacement Rate (Percent Severance of Individual’s Replacement Contribution Payb Pre-unemployment Rate (Percent Rate (Benefits Wage where Maximum Maximum of Individual’s (Employer + After Expenditure Share of not Indicated Duration Replacement Duration Pre-unemployment Potential Worker 20 Years Participation as Percent Unemployment Informal Country Otherwise) (Months) Rate (Months) Wage) Duration Obligation) of Service) Rates of GDP Rated Employmente New Zealand n.a. n.a. 15%–40% No limit n.a. n.a. n.a. 11 0.9 0 6.7 — of average wage (in 2001) Norway 73% 18–36 n.a. n.a. n.a. n.a. n.a. n.a. 0.1 0 4 — Portugal 65% 12–30 Flat, 80% 12–30 n.a. n.a. n.a. 20 0.8 0.1 6.2 — of mini- mum wage Singapore n.a. n.a. n.a. n.a. n.a. 20 n.a. n.a. 3.2 — Slovenia 60%–70% 3–24 Flat, 80% 15 n.a. n.a. n.a. 10 1.1 0.1 7.4 31 of mini- mum wage Spain 60%–70% 24 Flat 6–18 n.a. n.a. n.a. 12 1.4 0.1 20 21.9 Sweden 75% 10 Flat 5 n.a. n.a. n.a. n.a. 2.7 0.4 8.8 19.8 Switzerland 70% 5–17 n.a. n.a. n.a. n.a. n.a. n.a. 0 0.2 3.4 — Taiwan, China 60% 6 n.a. n.a. n.a. n.a. n.a. 7 n.a. n.a. 2.8 — Table 1. (Continued) Unemployment Unemployment Insurancea Assistancea Unemployment Insurance Savings Accounts Public Worksc Reference Variables Replacement Rate (Percent Severance of Individual’s Replacement Contribution Payb Pre-unemployment Rate (Percent Rate (Benefits Wage where Maximum Maximum of Individual’s (Employer + After Expenditure Share of not Indicated Duration Replacement Duration Pre-unemployment Potential Worker 20 Years Participation as Percent Unemployment Informal Country Otherwise) (Months) Rate (Months) Wage) Duration Obligation) of Service) Rates of GDP Rated Employmente United Kingdom Flat, 41% of 6 Flat No limit n.a. n.a. n.a. 5 0.1 — 6.5 — average wage (in 2001) United States 50% 6.5–10 n.a. n.a. n.a. n.a. n.a. n.a. 0.1 0 5 — —, not available; n.a., not applicable. Note: Included are countries covered in U.S. Social Security Administration (1999) with populations of 1 million or more. a Reference period for countries in Europe and East Asia and the Pacific is 2002; for all other countries reference period is 1999. Replacement rates are percentage of gross wages, unless specified otherwise. Benefit levels in dollars are benefits in local currency converted into dollars at average yearly exchange rates. For Algeria base is average monthly salary plus monthly national guaranteed minimum salary divided b Reference period is 1999. Benefits refer to months of salary paid to dismissed employee with 20 years of service. Figure for Sri Lanka is an estimate (box 1). c For developing and transition countries, includes social funds. Reference period is the 1990s and 2000–02 (yearly average or single year). Participation rate is defined as 100 times the number of participants divided by the labor force. d Five-year averages for 1995–99 (if single year, numbers appear in boldface and refer to one of the years during 1995–2000). e Estimates of the share of the labor force employed in the unofficial economy in the capital city as a percent of the official labor. Reference period is late 1990s. Figures are based on surveys and, for some countries, econometric estimates. Source: Unemployment insurance, unemployment assistance: U.S. Social Security Administration (1999, 2002a, 2002b); OECD (1999); Vodopivec, Wörgötter, and Raju (2005). Unemployment insurance savings accounts: Koman, Schuh, and Weber (forthcoming); Acevedo, Eskenazi, and Pages (forthcoming); Heckman and Pages (2000); Lipsett (1999); Mazza (1999). Severance payments: Botero and others (2002); OECD (1999). Public works: World Bank databases on public works and social funds projects, available at www1.worldbank.org/sp/safetynets; Betcherman and Islam (2001); Beegle, Frankenberg, and Thomas (1999); Marquez (1999); OECD (1999); Subbarao (1997, 2003). Unemployment: ILO (2001); World Bank (2003). Share of informal employment: Botero and others (2002). It should be emphasized that social insurance, including unemployment insur- ance, is a response of modern industrial society to the changing nature of the labor market, above all to the development of a modern employment relationship (Atkinson 1995a). In an industrialized and urbanized society, employment becomes discretion- ary: workers either work or do not work. This development has strong implications for unemployment—if they cannot find a job in wage employment (working for others), workers are unable to resort to self- or home-production, because they are divorced from ownership of the means of production. Similarly, in industrial societies, older and less productive workers stop working altogether once their productivity drops substantially. In contrast, in traditional societies they continue to be economically active as long as they produce something of value. The “invention� of unemploy- ment and the changing nature of retirement created the need to insure against non- employment contingencies. Social insurance is thus based on the concepts of unemployment and retirement as specific social constructs. Unemployment insurance has four key strengths.1 First, because it pools resources across a wide base, it provides good protection (that is, enables strong con- sumption smoothing) for all covered workers. Studies on the United States find that the average welfare of benefit recipient households is only 3–8 percent lower than that of households whose head is employed (Hamermesh and Slesnick 1995); in the absence of unemployment insurance, average consumption expenditures would fall about 20 percent (Gruber 1997). Second, the system performs well under all types of shocks. During the 1990s, when transition countries experienced massive increases in unemployment, the system provided effective insurance against idiosyncratic, sectoral, regional, and, somewhat less successfully, national shocks (there was some scaling back of benefits when unemployment escalated economywide; see Vodopivec, Wörgötter, and Raju 2005). Third, by automatically injecting additional resources and reducing taxes during recessions, unemployment insurance acts as an auto- matic stabilizer and thus moderates the magnitude of the economic downturn. Empirical evidence for Canada and the United States shows that unemployment insurance reduces GDP losses during downturns by 10–15 percent (Dungan and Murphy 1995; Chimerine, Black, and Coffey 1999; Hamermesh 1992). Fourth, unemployment insurance encourages the emergence or expansion of more risky jobs and industries, which may increase efficiency (Acemoglu and Shimer 1999, 2000). These strengths have to be weighed against four weaknesses. First, the system creates reemployment disincentives and wage pressures, which increase the equilib- rium unemployment rate of the economy (see, for example, Holmlund 1998).2 Sec- ond, by interacting with adverse shocks, the system contributes to the persistence of unemployment (Blanchard and Wolfers 2000). Third, unemployment insurance may create large unfunded liabilities, which put pressure on the government to sub- sidize the system.3 Fourth, protection is limited to formal sector workers. 62 The World Bank Research Observer, vol. 21, no. 1 (Spring 2006) Unemployment Assistance Unemployment assistance is a means-tested form of income support granted to working-age individuals who are unemployed and lack the necessary financial resources to maintain a minimum standard of living for themselves and their families. Self-standing unemployment assistance exists in few countries, notably Australia; Hong Kong, China; and New Zealand, among high-income countries, and Tunisia and Serbia and Montenegro among middle-income countries (table 1). In all other countries in which it is offered, unemployment assistance exists in tandem with unemployment insurance and is available upon expiration of eligibility to unem- ployment insurance to those unemployed who pass the means test. Like unemployment insurance, unemployment assistance systems require appli- cants to be capable of, looking for, and available for work. In some countries, eligibil- ity for unemployment assistance is not conditional on previous employment or contribution history. However, in most countries, particularly those with dual unemployment insurance/assistance systems, unemployment assistance is an extension of the unemployment insurance offered to the long-term unemployed who have satisfied some minimum length of employment and do not have the economic means to support their households. Means- or income-testing is conducted not only on the financial resources of the applicant but also on that of the applicant’s spouse and other adult members within the household. The system is typically financed by governments through general tax revenues. In countries with dual unemployment insurance/assistance systems, the source of financing is usually the same as for unemployment insurance. Unlike unemployment insurance, which grants benefits to all workers with suffi- cient paid contributions, unemployment assistance screens potential benefit recipients with an income/means test. This feature generates outcomes that are different from those of unemployment insurance in terms of both income protection and efficiency. Unemployment assistance has two main advantages over unemployment insur- ance. First, it is better able than unemployment insurance system to redistribute income to the poor (although, interestingly, it does not necessarily generate savings, as shown below). Second, the system can offer benefits to workers with weak labor force attachment, including informal sector workers, who are typically outside the reach of unemployment insurance. Unemployment assistance also has six disadvantages compared with unemploy- ment insurance. First, it offers lower level of protection to high-income workers. Second, it imposes larger administrative costs (the costs of monitoring income and assets of initial and ongoing claims). Third, it may reduce the system’s take- up. As Atkinson (1995b) shows, in Western countries a third or more of potential claimants never receive means-tested benefits (because of information problems, administrative complexities, and stigmatization of recipients). Fourth, because Milan Vodopivec 63 unemployment assistance benefits are contingent on the household income (and assets) of the unemployed individual, they may reduce the labor supply of other family members (for evidence from transition countries, see Vodopivec, Wörgötter, and Raju 2005). Fifth, unemployment assistance may create perverse entitlement effects. Young people may claim to be unemployed in order to collect benefits, for example, even though they may not be seriously searching for work or engaged in training. Sixth, unemployment assistance may discourage savings, because capital enters the income and asset tests (Atkinson and Micklewright 1991). Unemployment Insurance Savings Accounts Unemployment insurance savings accounts are a relatively new instrument, used only in Latin America and Austria (table 1). The accounts try to weave self-insurance into social insurance. Employers and workers deposit contributions into the savings accounts of individual workers; upon becoming unemployed, workers finance their unemployment benefits by withdrawing from these accounts. Upon retirement, pos- itive account balances are added to workers’ pensions. Under some proposals, unem- ployed workers are able to draw benefits under rules similar to traditional unemployment insurance, switching to funding from the solidarity fund when they exhaust resources in their individual accounts (the solidarity fund is financed by contributions from workers and employers and may be subsidized by the govern- ment; access to the fund is limited). The most important advantage of the system is arguably the reduction in the moral hazard inherent in traditional unemployment insurance, because the sys- tem internalizes the costs of unemployment benefits.4 Because the benefit is pay- able in cases of voluntary separations, the system encourages labor reallocation and cuts down on the litigation costs incurred under severance pay. If backed by government subsidies, the system has the potential of attracting informal sector workers. These strengths have to be weighed against four shortcomings. First, the sys- tem—in its pure form—does not pool risk among individuals and may thus be less efficient than social insurance systems. (For example, young workers may not be able to accumulate enough savings at the time of separation to be able to self-finance their unemployment.) Second, the version of the system that allows individuals to borrow from their accounts may suffer from a moral hazard problem of its own: it may generate incentives to withdraw from the formal sector and find a job in the informal sector, thereby avoiding the repayment of the debt upon reemployment in the formal sector (Parsons forthcoming). Third, the system requires a relatively well- functioning financial sector. Fourth, the system has higher administrative costs than other systems. 64 The World Bank Research Observer, vol. 21, no. 1 (Spring 2006) Severance Pay Severance pay, typically either mandated or required by collective agreements, pro- vides a lump-sum benefit to qualifying laid-off workers. Its coverage in developing countries is often limited to sectors, industries, or firms above a certain size. MacIsaac and Rama (2000) estimate that only about 20 percent of private sector workers, most of them wealthier workers, are legally entitled to severance pay in Peru. Benefits usually depend on years of service and individual’s earnings. They are paid by employers, sometimes with government assistance (for a review of the ori- gin, economic rationale, and current attempts to reform severance pay systems around the world, see Holzmann, Iyer, and Vodopivec forthcoming). Except in low- income countries, severance pay is the most prevalent income support system in developing countries. Because it often offers inadequate income protection and may impose large effi- ciency costs, severance pay is often considered one of the least appropriate options of income support. It does have three strengths, however. First, because the amount of severance pay is not contingent on the duration of subsequent unemployment, the system minimizes disincentives for subsequent job search. Second, by offering risk- pooling at the level of the firm, it may, if the scheme is generous enough, offer solid income protection. Third, administration of the system is simple. Severance pay suffers from several weaknesses, related to both income protection and efficiency. On the income protection front, there are two main weaknesses. First, the system suffers from the so-called “nonperformance problem� (the fact that despite legal entitlement, many workers fail to obtain benefits). MacIsaac and Rama (2000) estimate that only about half of all workers legally entitled to severance pay are likely to receive the benefit if dismissed (the payment is more likely if workers have a written contract and if they work in a larger, unionized firm that pays social security contributions). In Malaysia, employers disbursed only 83 percent of the sev- erance pay claims of workers laid off in 1998 (Mansor and others 2001). Nonperfor- mance is related to the limited risk-pooling ability of the system, coupled with the nonfunded nature of the system and the fact that the liabilities often arise when the firm is least capable of paying them. Second, severance pay benefits are inefficient, as the same amount is paid regardless of the duration of the unemployment spell. MacIssac and Rama (2000) find that severance pay was overly generous in Peru, where per capita consumption of those unemployed who received severance pay was higher than the consumption of otherwise similar workers who were employed. On the efficiency front, severance pay’s scorecard is also rather negative. First, several studies show that strict employment protection reduces employment. Lazear (1990) finds that severance pay reduces both employment and labor force participa- tion. Fallon and Lucas (1991) show that strengthening job security regulations led to a strong decline in employment in India and Zimbabwe. More recent studies Milan Vodopivec 65 confirming the link between job security and lower employment include Haffner and others (2001) for Organization for Economic Cooperation and Development (OECD) countries, Heckman and Pages (2000) for OECD and Latin-American countries, Besley and Burgess (2004) for India, and Haltiwanger, Scarpetta, and Vodopivec (2003) for OECD and transition countries. Heckman and Pages (2000) attribute a reduction in employment of 5 percentage points to job security provisions in Latin America. OECD (1999) finds insignificant effects on overall employment rates but notes that negative effects are concentrated among prime-age women, young peo- ple, and older workers. Studies also show that severance pay increases part-time employment and self-employment. Second, mounting evidence suggests that severance pay reduces inflows to and outflows from unemployment. By doing so, it contributes to longer unemployment spells; flows through employment may not be affected as strongly (for a recent survey, see OECD 1999; for evidence on transition countries, see Haltiwanger, Scarpetta, and Vodopivec 2003). Reduced labor market flows may hinder labor force adjustment and the reallocation of jobs, thereby slowing aggregate productiv- ity growth (for a survey of the effects of job reallocation on aggregate productivity growth, see Davis and Haltiwanger 1999). Third, de Ferranti and others (2000) report large litigation costs arising from disputes over the cause of separation in Latin America. Public Works Among systems that do not, strictly speaking, provide insurance, public works are worth singling out. Public works provide publicly financed, low-wage work to poor people. These programs are geared toward labor-intensive projects, with wages often set below the prevailing market rate. Public works have multiple objectives, with income support and the provision of goods and services the most important ones. In low-income countries, public works are the most prevalent income support program, available in more than 80 percent of countries. Interestingly, the inci- dence of public works is also relatively high in middle- and high-income countries (table 1). Public works programs have several strengths. They are effective in reaching the poor, have good targeting properties, and have substantial capacity to redistribute income from the rich to the poor. These programs also have strong capacity to attract informal sector workers, and they allow flexible and rapid responses to shocks (Subbarao 2003). Moreover, public works are administratively less demand- ing than other public income support systems for the unemployed. There are also four weaknesses of the program. First, high nonwage costs reduce the effectiveness of public works in reaching the poor. Ravallion (1999) estimates that $5 of public transfers are needed to increase the earnings of the poor by $1, 66 The World Bank Research Observer, vol. 21, no. 1 (Spring 2006) partly because of the leakage of spending on the nonpoor. Second, the countercy- clical pattern of funding shows that it is difficult to raise funding during crises, when support is needed most (Wodon 2000). Third, because of highly redistribu- tive character of public works, it may be difficult to gain political support; some leakage to the nonpoor may be necessary to gain such support. Ravallion (1991), for example, finds that the leakage of benefits to nonpoor participants in the Maharashtra Employment Guarantee Scheme in India may have been instrumen- tal in obtaining sustained budgetary support. Fourth, public works may stigmatize participants (for the experience in transition countries, see Betcherman, Olivas, and Dar 2004). Distributional Effects of Income Support Systems Different income support systems have different effects on income redistribution. Vodopivec (2004) shows that among the income support systems discussed here, public works are the most progressive. Subbarao (2003) also documents the strong potential of public works for redistribution. He reports that nearly 100 percent of public works participants in Chile and 60–70 percent in India (both in the nation- wide program and the Maharashtra Employment Guarantee Scheme) were poor. Based on their analysis of 101 national public works projects in South Africa, Haddad and Adato (2001) conclude that under fairly robust assumptions, the vast majority of public works programs considerably outperform hypothetical bench- marks consisting of untargeted transfers. This does not mean that the targeting method under public works—self-selection based on working in the program— outperforms other social safety net programs that use different targeting methods. After comparing the effectiveness of various targeting methods in 67 programs, Coady, Grosch, and Hoddinott (2004) conclude that no method provides clear advantages.5 Although a public works program (Argentina’s Trabajar) received the highest score for its targeting performance, three public works programs ranked among the worst 10 programs. Coady, Grosch, and Hoddinott also find considerable evidence that it is the implementation of the program, rather than the method of tar- geting, that is key to successful targeting.6 Unemployment assistance also has a strong redistributive potential. In a study of 13 OECD countries, Vroman (2002) shows that the share of unemployment benefits received by the bottom three deciles ranged from 20.8 percent in Italy to 58.0 per- cent in Australia, the only country in the study with a self-standing unemployment assistance system. The top three deciles in Australia received 7.4 percent of trans- fers, the lowest percentage across the 13 countries. The income distribution effects of unemployment insurance and unemploy- ment assistance in transition countries are also strongly progressive. Vodopivec, Milan Vodopivec 67 Wörgötter, and Raju (2005) show that in the mid-1990s, the bottom 40 percent of households in seven European transition countries received 70–80 percent of the benefits. They report that when considering distributions of households with at least one unemployed member, the incidence of benefits was progressive when distribution was based on pre-benefit income and regressive when based on post- benefit income. The exceptions are Estonia and the Slovak Republic, the coun- tries with the smallest share of unemployment benefits in household incomes, where the small scale of transfers did not make a difference in the post-benefit distribution. The strong redistributive performance of unemployment benefits in transition countries may be the result of circumstances—such as nearly universal unemploy- ment benefit coverage—that are not present in developing countries. Brazil’s unem- ployment benefit system has a neutral effect on the redistribution of income (de Ferranti and others 2000). Unemployment benefits have not been an important tool for income redistribution in industrial countries: unemployment benefits are pro- gressive in about half of OECD countries and neutral in the other half (Forster 2000). Nonetheless, the evidence on transition countries suggests that under specific cir- cumstances, unemployment insurance and assistance systems offer substantial scope for income redistribution. De Ferranti and others (2000) show that most participants in both the Colombian unemployment insurance savings accounts and the Peruvian severance pay sys- tem belong to the richest segments of the population (this pattern is not an inher- ent property of these systems, but it is probably typical for low-income countries). In addition to affecting income distribution directly, severance pay also has an indirect effect, by hindering access to jobs by disadvantaged groups. Blanchard (2000) shows that an increase in firing costs leads to higher unemployment of marginal groups of workers because of their inferior access to jobs. Because the productivity of these workers before hiring is not easily revealed, their probability of being hired in the presence of increased firing costs is lower. Indeed, an OECD study (1999) finds that stricter employment protection legislation reduces employment among prime-age women and young people, rendering them more susceptible to unemployment risk. By contributing to labor market dualism, sever- ance pay thus increases the advantage of already privileged formal sector workers, increasing inequalities in society. In line with these findings, Besley and Burgess (2004) show that “pro-worker� legislation may work against the poor in develop- ing countries. What are the likely distributive effects of introducing unemployment insur- ance savings accounts? While in principle the system can provide the same income protection as traditional unemployment insurance system, switching from unemployment insurance to unemployment insurance savings accounts changes income distribution because the benefits are financed differently. 68 The World Bank Research Observer, vol. 21, no. 1 (Spring 2006) Simulations suggest that the distributive effects for the United States are likely to be small and regressive. Feldstein and Altman (1998) find that individuals in all quintiles gain slightly and individuals in the bottom quintile lose slightly (the fact that these estimates do not take account of the behavioral responses to the changed system most likely means that they understate the distributive effects). Similar simulations for Estonia show that the income redistribution achieved by unemployment insurance savings accounts is substantially less than that achieved under unemployment insurance and that unemployment insurance savings accounts allow for more redistribution when unemployment is higher and unemployment benefits more generous (Vodopivec and Rejec forthcoming). Choosing a System In choosing among income support systems for the unemployed, policymakers should base their decisions on the strengths and weaknesses of the systems as well as on country-specific circumstances. Depending on institutional capacity and the development of the labor market, some countries may find social insurance an attractive option. Other countries may try to improve their severance pay systems, including by converting them into unemployment accounts. To provide protection to informal sector workers, policymakers should also consider public works and innovative approaches. This section discusses how alternative income support systems fit institu- tional and other circumstances typically found in developing and transition countries. It draws policy implications about choosing among them (see Vodopivec 2004). What Conditions Are Conducive for Unemployment Insurance? Several institutional and labor market features are conducive for unemployment insurance. The first is strong administrative capacity to monitor initial and particu- larly continuing eligibility. Recent studies suggest that effective monitoring and the use of sanctions can strongly reduce the average duration of unemployment benefit payments and increase the transition rate to employment (OECD 2000; Boone and others 2001. Thus the stricter the monitoring of the behavior of recipients, the milder the disincentives of unemployment insurance. A second feature is an informal sector that is not too large. The higher the infor- mality of the economy, the more abundant are opportunities for undeclared paid work and thus the higher the costs of monitoring. Milan Vodopivec 69 A third feature is the low incidence of private transfers. If the introduction of pub- lic insurance breaks down the habit of self-help among extended families in commu- nities, replacing private transfers by social insurance may be welfare reducing (Attanasio and Rios-Rull 2000). Absent these conditions, unemployment insurance does not perform well, increas- ing inefficiencies, reducing welfare gains, or both. For example, reemployment incentives depend crucially on a country’s monitoring capacity. This capacity deter- mines how strictly the conditions of initial and, perhaps even more important, con- tinuing eligibility are imposed. As the experience with Argentinean unemployment insurance suggests, the capacity to screen initial eligibility has not been a problem— the system draws on the capacity of other social protection systems—but the coun- try still has to monitor continuing eligibility (Mazza 1999). Inadequate monitoring creates leakages that add to the costs of the system. In view of the above, how suitable is unemployment insurance in developing countries? Prompted by increased exposure to foreign markets and fearing future international crises, some developing countries (including the Philippines and Sri Lanka) are contemplating introducing unemployment insurance. Immediate introduction in a country like the Philippines may be premature before certain con- ditions are fulfilled (Yoo 2001). Gill and Ilahi (2000) argue that many Latin-American countries lack the capacity to run an efficient unemployment insurance system. They argue that although introducing unemployment insurance should be a long- term goal of these countries, it is either infeasible or too costly a strategy in the medium term. They propose that Latin-American governments augment other instruments, such as self-insurance, to overcome the lack of market insurance in the medium term. In sum, because successful performance of unemployment insurance relies on strong administrative capacity to monitor system eligibility and an informal sector that is not too large—conditions typically lacking in developing and transition countries—the case for its introduction in these countries is less compelling than it is in developed countries. The presence of unemployment insurance may also reduce incentives for self-protection and break down the habit of self-help within communities. Introducing full-scale, general unemployment insurance is thus viewed as a longer-term goal for most developing countries. Under suitable cir- cumstances, countries with the most pressing needs could consider introducing a scheme with modest benefits and limited coverage (a scheme offered by larger employers, for example). Of course, to ensure affordability, minimize adverse effects, and enable a smooth start, a new unemployment insurance system has to provide modest benefits and be introduced on a small scale. Moreover, it can be introduced only after the needed support infrastructure is put in place. Specific his- toric and political conditions may be instrumental in introducing such a system (box 1). 70 The World Bank Research Observer, vol. 21, no. 1 (Spring 2006) Box 1. Improving Income Support for the Unemployed in Sri Lanka By reforming its nontransparent, discretionary, and costly severance pay program, Sri Lanka could improve income protection of the unemployed while promoting employment and efficient reallocation of labor and ensuring more equitable access to formal sector jobs. The current severance pay system not only provides extremely high compensation to laid-off workers—and imposes correspondingly high costs on employers— its discretionary nature and lack of transparency impose additional costs by creating lengthy procedures and generating uncertainty about firms’ ability to lay off workers. Consequences of high separation costs include lower labor market flows, reduced aggregate employment, and repressed investment incentives, especially for foreign investors. Currently, the government sets the amount of compensation to laid-off workers and has the authority to reject employers’ demands. In 2000–01 the average compensation amounted to two to three monthly wages per year of service, and total payments in some cases exceeded 40 monthly wages. The time needed to process the request by the government has been unpredictable, averaging six months and in some cases taking much longer (the procedure involves hearings, in which employers explain their financial performance and busi- ness plans to the government to justify the layoffs). A self-standing, radical reform of the Sri Lankan severance pay system is one option. But because a political decision has been made that the reform of severance pay is contingent on the introduction of unemployment insurance, a difficult question has emerged: should Sri Lanka introduce unemployment insurance? If its introduction brings substantially more flexible employment protection legislation and the design of the sys- tem is adjusted to Sri Lanka’s circumstances, the answer is yes. But if the unemployment insurance system is costly, if it is not adapted to country circumstances (including the lack of institutional capacity), and if it leaves the existing severance pay essentially intact, the answer is probably no. Such a system would add another protection program to an already privileged group of formal sector workers, and it may reduce labor market performance, thus failing to produce desirable equity and efficiency gains. Source: Vodopivec (2005). Unemployment Assistance: How Attractive Is Means-Tested Targeting? Does targeting of benefits to the most “needy� improve incentives and produce sav- ings? Not necessarily. As Vroman (2002) shows, means testing per se does not ensure that an unemployment benefit system is inexpensive. Depending on the sys- tem’s parameters—primarily the income-threshold that triggers unemployment assistance—costs can be high, low, or in between. What are the implications for providing unemployment assistance in developing and transition countries? First, in many of these settings, a potential advantage of unemployment assistance—that eligibility does not require prior contributions— renders such a system nonviable. With large segments of the labor force either underemployed or unemployed, providing an income support system that fails to exclude people without prior work in the formal sector (that is, people who have not paid system contributions) would be fiscally unsustainable. It therefore seems unavoidable that unemployment assistance systems in developing countries base benefit eligibility on the prior payment of system contributions, as is done, for example, Milan Vodopivec 71 under unemployment insurance. Second, due to administrative constraints typi- cally faced by low-income countries, few, if any, of these countries could conduct the required level of monitoring (Vodopivec 2004). Third, abundant informal sec- tor employment opportunities mean that the problem of employment disincentives for other members of the household is more pronounced than in developed coun- tries. If informal sector opportunities abound, ineffective monitoring produces large leakages. Effective monitoring would reduce overall expenditures, but it would impose large administrative costs and perhaps reduce employment and earnings in the informal sector. The applicability of unemployment assistance thus seems lim- ited to countries with relatively developed administrative capacity, a small informal sector, and large fiscal pressures, perhaps as a transition system to unemployment insurance. Unemployment Insurance Savings Accounts: How Much Insurance Can They Offer? Among the new approaches to improved income protection of the unemployed, unemployment insurance savings accounts are the most radical, offering the poten- tial to alleviate the moral hazard inherent in traditional unemployment insurance. But, as critics maintain, these accounts do not create the ills of unemployment insurance because they do not provide such insurance in the first place. In its pure form, the system is just a form of forced savings; by its very design it does not pool risk among individuals.7 Under certain circumstances, the absence of pooling across individuals may not be critical. Under modest and frequent shocks, as the comprehensive insurance theory of Ehrlic and Becker (1972) suggests, self-insurance through savings may provide adequate smoothing of consumption. Moreover, because the absence of cross-sectional pooling seriously limits the provision of insurance, some unemploy- ment insurance savings account proposals combine individual accounts with public insurance in order to better address large and persistent shocks (Feldstein and Altman 1998; Guasch 1999). For example, under the proposal of Feldstein and Altman, unemployed workers are able to draw benefits monthly, as under the tradi- tional unemployment insurance; the government lends money to accounts when the balance falls below zero (the unemployment insurance savings accounts cum borrowing version). Under such a proposal, the consumption smoothing properties of the unemployment insurance savings accounts system would be no worse than under traditional unemployment insurance, because individuals with negative bal- ances would still receive benefits, as rules of withdrawal would be the same as under the unemployment insurance system. But the savings accounts system would reduce labor market disincentives for workers who end their working careers with positive account balances. This version of the system reduces the gains in terms of reemployment incentives but increases its insurance function. 72 The World Bank Research Observer, vol. 21, no. 1 (Spring 2006) A similar approach—combining self-insurance and social insurance—is applied in the Chilean unemployment insurance system introduced at the end of 2002. Under the system, employers and workers make contributions to individual savings accounts. Employers and the government contribute to a solidarity fund (employers’ contributions reduce their severance payments obligations, so the new unemploy- ment insurance system also partly replaces the severance pay system). To stimulate reemployment, the unemployed first draw benefits from their individual accounts; upon depletion, and subject to the usual unemployment insurance eligibility condi- tions, they draw from the solidarity fund (Acevedo, Eskenazi, and Pages forthcoming). What is the rationale for public intervention in unemployment insurance savings accounts? In addressing this question, it is necessary to distinguish between pure unemployment insurance savings accounts and versions that include solidarity funding. The rationale for pure unemployment insurance savings accounts is based largely on country-specific circumstances. These accounts are just a form of self- insurance or forced savings: they support consumption during spells of unemploy- ment by allowing individuals to draw on savings accumulated while working. While the system is likely to generate few inefficiencies, it offers income protection only under very restrictive conditions, such as frequent, modest shocks (short expected durations of unemployment), and it fails to protect some groups of workers (such as workers at the beginning of their working careers, whose accumulations are likely to be low). Such a system could be advocated on the basis that workers are irrational (that they have too high a discount rate or underestimate the risk of becoming unemployed), so that left by themselves they make inadequate provisions for the risk of becoming unemployed. Pure unemployment insurance savings accounts thus cannot be advocated on the grounds that they correct market failures. But it is possible that they may correct “gov- ernment failures�: such accounts may be justified if they are the only politically palat- able solution to reform another badly performing public income support system (such as severance pay) or if the government cannot commit itself not to bail out unprofit- able firms in order to avoid politically unacceptable increases in unemployment. Pure unemployment insurance savings accounts may be preferred to traditional unemploy- ment insurance if traditional unemployment insurance creates severe inefficiencies, particularly if the risk of unemployment is frequent and moderate. In contrast, the rationale for public intervention under the unemployment insur- ance savings accounts cum borrowing is twofold. First, the system helps smooth consumption by providing access to credit. It thus corrects for capital market imper- fections that prevent borrowing against future earnings—a limitation that is often binding in developing countries. Second, this system combines self-insurance with social insurance (because the debt of the workers who end their working lives with negative balances has to be financed through a solidarity fund). The rationale for public intervention is thus widened to include all considerations that underlie the Milan Vodopivec 73 public provision of unemployment insurance. The unemployment insurance sav- ings accounts cum solidarity fund version (such as that adopted in Chile’s new sys- tem) shares this feature, but its consumption-smoothing capacity is somewhat more limited, as this version does not allow borrowing. The borrowing version of the scheme potentially offers several advantages. It addresses capital market imperfections, enables widespread risk-pooling, and offers other advantages over the public provision of unemployment insurance (public agencies are better able than private insurers to monitor benefit eligibility, for exam- ple). At the same time, the system addresses the moral hazard problem inherent in unemployment insurance by introducing self-policing. This scheme can improve incentives if the self-insurance imposed by individual accounts reduces moral haz- ard by more than methods applied under traditional unemployment insurance. The scheme can also increase income protection over that provided by pure unemploy- ment insurance savings accounts (or a prefunded severance pay system), because it allows widespread risk-pooling. The system can also increase access to credit, because it allows workers to borrow from their accounts. Simulations by Hopenhayn and Hatchondo (forthcoming) show that when its parameters are appropriately selected, the unemployment insurance savings accounts cum borrowing scheme comes very close to the welfare properties of the optimal unemployment insurance system. According to some proposals, the efficiency properties of the unemployment insurance savings accounts cum borrowing system can be further improved by com- bining several risks under one system. Orszag and others (1999) and Yun (2001) propose an integrated unemployment insurance system that would combine unem- ployment insurance not only with the pension system but also with other systems, such as health, disability, and life insurance. Such a system would integrate inter- temporal pooling of various risks of the individual with cross-sectional pooling. By doing so, the system is expected to provide better insurance and significantly reduce disincentives compared with traditional unemployment insurance. Some design and implementation considerations speak in favor of introducing this system in middle-income and upper-middle-income developing and transition countries. First, weak monitoring capacity in these countries exacerbates the moral hazard problem inherent in traditional unemployment insurance and encourages other misuses of the system. Hence the self-policing nature of the unemployment insurance savings accounts represents a greater advantage. Second, various income support systems exist in developing countries. Their conversion into an unemploy- ment insurance savings accounts-type system could greatly facilitate its introduc- tion. In the Philippines, for example, several mandatory forced savings schemes could, together with severance pay, be merged and transformed into unemployment insurance savings accounts (Esguerra, Ogawa, and Vodopivec 2001). Third, under traditional unemployment insurance, employers in developing countries sometimes 74 The World Bank Research Observer, vol. 21, no. 1 (Spring 2006) fail to contribute to the system. By introducing personal accounts, workers monitor such payments themselves. This feature also makes the accounts less susceptible to political risk. Finally, the administrative complexities of introducing unemployment insurance savings accounts are not prohibitive. The old age insurance systems introduced in many Latin-American countries require similar information systems.8 Introducing unemployment insurance savings accounts may create adverse incentive problems of their own, which may be difficult to deal with. First, the sys- tem may produce excessive turnover by encouraging workers to choose to become unemployed in order to “get their money back,� a problem experienced by the Brazilian severance fund, FGTS (Amadeo, Gill, and Neri 2002). Setting administrative limits on when workers can access funds might help, but doing so would limit labor mobil- ity. Second, unemployment insurance savings accounts cum borrowing may create incentives to “dive and run,� that is, it may stimulate workers to borrow and then withdraw from the formal sector and find a job in the informal sector in order to avoid repaying the debt upon reemployment in the formal sector. Conceivably, one could use pension contributions as “collateral� in an integrated unemployment insurance savings accounts–pensions system. But this integration assumes that workers have very long-time horizons, which may not be the case (indeed, if workers had long horizons, they would save for themselves and there would be no need for forced savings mechanisms such as the pure unemployment insurance savings accounts system). In sum, unemployment insurance savings accounts with solidarity funding—and their variant, the integrated unemployment insurance system—may be promising options, particularly in East Asia and Latin America, where the existence of sever- ance pay systems may ease the transition to unemployment insurance savings accounts. There is a need, however, for further investigation—and piloting—of the system. Too little is known about the working of the scheme to know the types of workers for which, and the conditions under which, the system is likely to work best. Important design parameters of the system (regarding contribution rates and rules for withdrawal, for example) also need to be examined. Reforming Severance Pay Severance pay is often considered one of the least appropriate options of income pro- tection: it not only provides inadequate protection, it also imposes efficiency costs by distorting the behavior of firms and workers. Yet except in low-income countries, severance pay is the most prevalent income support system in developing countries. The question of how best to reform existing severance pay systems is thus of obvious importance. Two possible reform directions are examined here. The first is stream- lining the system and converting it into a funded system. The second is linking it to unemployment benefit or pension reform. Milan Vodopivec 75 Streamlining current systems. Countries may have severance pay systems that pro- voke undesirable responses from firms and workers, impose large monetary and other costs, and are not synchronized with other income protection systems. To streamline their severance pay systems, some countries could simplify system rules. Latin-American countries, for example, could reconsider additional payments imposed for “dismissals without cause� (which almost invariably include economic dismissals). This would not only reduce the costs of the system but also reduce the costs of litigation. Moreover, to reduce unnecessary costs and uncertainty about the ability to lay off workers, the lack of transparency and discretionary nature of sever- ance pay systems in some countries should be addressed. Countries with both social insurance and severance pay systems can reduce costs without reducing insurance by better coordinating payments under the two systems. Unemployment insurance eligibility rules could be adjusted so that insurance benefits would start only after severance benefits “expire,� that is, after n months if the individual received n monthly wages as severance payment (such a system is in place in some developed countries, such as Canada). To address the nonperformance of the system, some countries may consider establishing public guarantee funds, thus introducing pool- ing at the above-firm level. Prefunding. A more radical reform consists of converting severance pay into a funded system. Such a reform converts conditional obligations to laid-off workers under the traditional severance pay into unconditional obligations to all workers paid regularly to their individual accounts. The reform aims at addressing the non- performance problem of severance pay, as well as at ameliorating some of the ineffi- ciencies severance pay creates (by, for example, removing obstacles to labor market flows and reducing litigation costs). Conversion would introduce a version of unem- ployment insurance savings accounts (without solidarity funding). Prefunding reforms were introduced in Colombia in 1990 and in Austria in 2003. Colombia’s system requires employers to deposit a percentage of their workers’ wages into guaranteed individual accounts available to workers in the event of job separation (limited access to funds while employed was also foreseen). Kugler (1999, 2002) finds that the reform reduced labor market distortions for two reasons. First, introduction of individual accounts reduced wages; as a result, employment, measured by weekly hours, increased. Wages were reduced as employers shifted 80 percent of the severance payments’ costs to workers in the form of lower wages; the total compensation of workers (wages plus deposits to their savings accounts) increased. Second, because the reform removed the discretionary nature of sever- ance payments, both separations and accessions under the new system increased. Kugler (2002) finds that by transforming uncertain, conditional payments to unconditional payments monitored by a third party (the government), the reform also enhanced the insurance function of the severance pay. Before the reform, 76 The World Bank Research Observer, vol. 21, no. 1 (Spring 2006) nonperformance was a big problem: firms about to go bankrupt could simply not pay severance or negotiate a package substantially below what was owed in sever- ance payments. Thanks to the prefunding requirement, the reform increased the likelihood that workers actually avail themselves of their legal entitlement to sever- ance pay. Among those who received severance payments, consumption smoothing was equally effective before and after the reform. Austria converted its severance pay to a fully funded contributory system akin to unemployment insurance savings accounts (Koman, Schuh, and Weber forthcom- ing). The reform extended the entitlement to workers with short tenures and removed obstacles to worker mobility, granting full portability and allowing the accumulation of benefits from the beginning of an employment spell. Employers pay 1.5 percent of workers’s salary to them, with resources held in a central account and invested in the capital market. Laid-off workers with job tenure of three years or more can withdraw accumulations in their accounts or keep them and claim them upon retirement. Workers who separate voluntary or have tenures of less than three years are denied the right of immediate withdrawal, a feature that may hinder worker mobility. Conversion to a funded scheme reduces nonperformance problems, but it only transforms severance pay into a forced savings scheme in which workers insure themselves (self-insurance). One possible further step to improve the insurance func- tion of the system is to add a social insurance component (as a back-up, to minimize disincentives), in order to increase the pooling of the employment risk to the national level. Advantages of Public Works Many conditions prevailing in developing countries make public works especially attractive. First, the informal sector is large and pervasive. Informal sector workers do not have access to public income support systems that require social security con- tributions and thus remain vulnerable to even small income shocks. Second, due to a strong seasonal farm workload, particularly in mono-crop areas, public works can be inexpensively deployed in nonfarm activities in nonpeak periods. The system thus provides an opportunity to productively engage temporary “surplus� labor while minimizing forgone earnings and maximizing poverty reduction effects. Third, the existence of large mono-crop areas makes large segments of the population vulnera- ble to cyclical and structural shocks. Similar exposure is caused by geographic and climatic shocks. In the absence of market insurance, public works can provide effec- tive insurance in such cases. Fourth, public works do not require complex adminis- tration, and they may be quickly set up in areas affected by various shocks. Fifth, obtaining support for public works can benefit from traditions and values that emphasize cooperation and collective support, particularly in rural areas. Milan Vodopivec 77 Because establishing a program quickly in response to a shock is difficult, public works could be run on a permanent basis (as employment guarantee schemes). Expe- rience shows that public works in transition countries do not increase employability and may stigmatize participants, so this option is less desirable in these countries. National Income Level and the Choice of Program One size does not fit all. An unemployment support program suitable for one group of countries may not fit the institutional or other characteristics of another group. How should a country with a certain set of characteristics choose an unemployment support program? The preferred choices of unemployment support by groups of countries are sum- marized in table 2.9 For low-income countries, public works is the program of choice, particularly because the program is effective in combating poverty. Other programs are much less suitable. Unemployment insurance is not appropriate in low-income countries, because they lack the necessary administrative capacity, their labor markets are largely informal, and funds set aside for this program may be subject to political risk. Reducing overly generous severance pay benefits sometimes found in low-income countries would help reduce labor market duality. Similar considerations suggest that public works are the preferred scheme in lower-middle-income countries. Depending on circumstances, however, a simplified social insurance program, offering modest benefits, may be appropriate. Such a scheme may be necessary to pave the way to overhauling an outdated severance pay system (see box 1). Given the weak administrative capacity of lower-middle- income countries, unemployment insurance savings account cum solidarity fund- ing seems particularly desirable. Other existing institutions may also affect the choice of the unemployment sup- port programs. For example, a country with a provident fund allowing preretire- ment withdrawals (for housing or education, for example) may introduce unemployment income support by simply expanding the set of contingencies financed by such a fund. With their greater administrative capacity, transition countries may opt for unemployment assistance, in order to improve the targeting of benefits, but unemployment assistance should base benefit eligibility on the pay- ment of contributions to ensure fiscal sustainability. The program of choice for upper-middle-income countries is unemployment insurance, combined with other labor market programs. These countries should introduce unemployment insurance cautiously, starting with modest benefits and simple rules. To increase their generosity and thus their consumption-smoothing function, unemployment insurance should gradually improve the monitoring of benefit eligibility (by, for example, interlinking administrative information systems) and help beneficiaries improve their employability. To avoid perverse incentives, 78 The World Bank Research Observer, vol. 21, no. 1 (Spring 2006) Table 2. Recommended Public Unemployment Support Programs, by Country Income Level Country Income Critical Conditions and Considerations Supporting Group Recommended Unemployment Support Program Objectives Served Recommendation Low • Public works • Reduces poverty Low administrative capacity, high labor market income informality, and high political risk make this group unsuitable for unemployment insurance programs • Provides goods and services to the poor • Spurs community development and empowerment • Reduced severance pay benefits (if overly generous) • Reduces labor market duality Lower • Public works • Reduces poverty • Introduction of a social insurance should be middle conditional on improving other programs (by income cutting back excessive, nontransparent severance pay systems, for example, thereby moving from job to income protection) • Provides goods and services to the • To avoid perverse incentives and given the weak poor administrative capacity of lower-middle-income countries, unemployment insurance savings account cum solidarity funding may be desirable • Spurs community development • Unemployment assistance may be suitable for and empowerment transition countries that have better administrative capacity and can thus target the poor more effectively; eligibility should be based on previous contributions to the program • Simplified, modest, adapted social insurance pro- • Paves the way to reduce labor gram (unemployment insurance, unemployment market duality assistance, or unemployment insurance savings account cum solidarity funding), depending on circumstances (Continued) Table 2. Continued Country Income Critical Conditions and Considerations Supporting Group Recommended Unemployment Support Program Objectives Served Recommendation Upper • Social insurance programs (unemployment • Smoothes consumption • Experience and better integration of public information middle insurance, unemployment assistance, unemployment systems may lead to the increased sophistication of income insurance savings accounts), combined with other social insurance programs programs • Makes job search more efficient • To avoid perverse incentives, unemployment insurance savings account cum solidarity funding may be desirable High • Unemployment insurance, supplemented by labor • Smoothes consumption • Unemployment insurance relies on strong income market and other programs (public works, training, administrative capacity to monitor benefit eligibility social assistance) • Makes job search more efficient • Particularly suitable for highly formalized labor markets • Helps marginalized workers retain labor force attachment and improve employability • Reduces income inequalities and counters social exclusion unemployment insurance savings account cum solidarity funding may also be desir- able for this group of countries. In addition to providing unemployment insurance with a high level of con- sumption smoothing (and extending assistance upon its expiration as a social assistance benefit), high-income countries offer a full range of active labor market programs to the unemployed. These programs include job search assistance, train- ing, public works, youth unemployment programs, and often also employment subsidy programs. Adapting the System and Offering a Mix of Options Institutional and other characteristics of developing and transition countries strongly affect not only the choice but also the design of income support system as well as the mix of options offered. This section describes desirable design adjustments of unemployment insurance and explains why the simultaneous provision of several systems may be desirable. Adjusting Unemployment Insurance to Developing Countries When introducing unemployment insurance outside developed countries, how should the system be adjusted to enhance its protection while minimizing its nega- tive effects on incentives and efficiency? Two key features of developing countries that influence the design of unemployment insurance need to be taken into accounted: the level of administrative capacity, which is lower than in developed countries, and the abundant opportunities to work in the informal sector. The exam- ple of Sri Lanka is instructive. Sri Lanka lacks the administration capacity to effec- tively implement a Western-style unemployment insurance. Employment offices were established in Sri Lanka only in 2003, when 17 regional centers providing job- matching services were opened (Vodopivec 2005). Many activities, particularly effective monitoring of continuing eligibility, could be carried out properly only in the medium to long term.10 Moreover, the informal sector, which employs more than half the country’s workforce, offers abundant opportunities for work that are hard to monitor and record.11 The following design features are especially desirable to accommodate these country-specific circumstances. First, to minimize disincentives, both the level and duration of benefits should be limited. Flat-rate benefits are particularly appealing, because they are simpler to administer and offer fewer possibilities for misuse and fraud. Another option is declining benefits during individual unemployment spells. Second, continuing eligibility criteria need to be adjusted. Argentina’s unemployment insurance system has completely forgone the monitoring of continuing eligibility—a Milan Vodopivec 81 decision Argentina may reconsider once it develops sufficient monitoring capacity (the absence of monitoring continuing eligibility reinforces the need to offer modest benefits). Third, social insurance could be combined with self-insurance, following the approach of the new Chilean unemployment insurance system in order to reduce labor market disincentives. Offering a Well-Balanced, Flexible Mix of Options To satisfy diverse needs and objectives, and to confront various types of risks, coun- tries should offer a well-balanced, flexible mix of income protection systems and complement them with systems outside income protection. There are several rea- sons to offer a mix of income support systems rather than a single system. Different systems have different objectives. The primary goal of some systems is to compensate for the loss of earnings; other systems and policies may emphasize human resource development. In some stages, the labor reallocation goal may deserve special atten- tion (transition countries are a case in point). Moreover, diverse groups of workers have different needs. Important segments of the workforce in developing countries are engaged in the informal sector and are thus ineligible for social insurance sys- tems and severance pay. Government should provide supplemental income support systems in which anyone can participate (programs such as public works and train- ing). Supplemental systems may be dictated by the regional unevenness of the size of private transfers. Regions that receive lower levels of remittances from abroad, for example, should have wider opportunities for public works. Another reason for offering a mix of systems is that they differ in their ability to confront shocks. The selection of systems should take into account the prevalence and severity of shocks typically confronted. For example, the effectiveness of unem- ployment insurance is reduced during recessions, so supplemental systems (such as public works, training, or conditional cash transfers) may be needed. Similarly, in dealing with industry-level risks, special redundancy systems may be put in place to promote enterprise restructuring. The occurrence of natural disasters points to the need for flexible, quickly deployable systems (such as public works). Countries should also complement income support with systems in other areas, including labor market and financial policies. Well-functioning labor, insurance, and financial markets can substantially increase opportunities for self-protection (by reducing the risk of unemployment) and self-insurance (by contributing to short unemployment spells). Income support can often be effectively complemented by active labor market pro- grams (training, employment subsidies, job search assistance, promotion of self- employment, youth programs). Depending on a country’s fiscal position, objectives, and conditions, these efforts may be used to promote employment opportunities of the unemployed. Two aspects of interaction with income support systems are worth 82 The World Bank Research Observer, vol. 21, no. 1 (Spring 2006) mentioning. First, active programs should be coordinated with income support sys- tems, in order to minimize perverse incentives for enrollment in active programs and weaken incentives for reemployment. Second, active labor market programs may be used as a screening device for participants of income support systems. Another important complementary program is conditional cash transfers, such as Mexico’s PROGRESA and Brazil’s PETI. To increase educational attainment and improve health, these transfers help poor families pay for basic health and schooling expenses, as long as their children regularly visit health clinics and attend school. These programs discourage counterproductive coping mechanisms, such as reduc- ing healthcare expenditures and taking children out of school. Evaluations show that these programs can raise school enrollment and attendance rates and improve child health and nutrition (Rawlings and Rubio 2003). They undoubtedly reduce poverty in the longer run. Diverse needs and circumstances also call for creative, innovative programs. One important example is social funds. Social funds are agencies that finance small-scale projects, predominantly in poor communities. To ensure that truly valuable projects are selected, the communities themselves identify and partly finance projects. Social funds demonstrate the ability to foster cooperation among local actors, provide valu- able products and services to vulnerable segments of population, and engage the poor in the implementation of projects, thereby providing them with income protec- tion (Rawlings, Sheburne-Benz, and Van Domelen 2003). Summary Each system of income support for the unemployed has its strengths and its weaknesses. • Unemployment insurance facilitates a high degree of consumption smoothing, performs well under various types of shocks, and acts as an automatic macroeco- nomic stabilizer, but it creates reemployment disincentives and wage pressures, which increase the equilibrium unemployment rate and make unemployment persistent. Because its successful performance relies on conditions that are typi- cally lacking in developing and transition countries, the case for introducing unemployment insurance is less compelling in these countries than in developed countries. • Unemployment assistance allows more effective targeting, but it may not yield savings over unemployment insurance. It provides a lower level of protection for high-income workers, is more expensive to administer than unemployment insurance, and creates similar employment disincentives. Its applicability is thus limited, perhaps to countries with relatively developed administrative capacity and a small informal sector. Milan Vodopivec 83 • By internalizing the costs of unemployment benefits, unemployment insurance savings accounts ameliorate the moral hazard inherent in traditional unemploy- ment insurance and thus improve reemployment incentives. Given the weak monitoring capacity of developing countries, this is an important advantage. When integrated version with social insurance, which pools risk across partici- pants, the system may improve work incentives while in principle offering income protection similar to that of traditional unemployment insurance. The scheme also has the potential to attract informal sector workers. By allowing individuals to borrow from their accounts, however, some versions of integrated accounts create incentives to withdraw from a formal sector in order to avoid repaying the debt. Because the system has been largely untested, further investigation of its effects and design parameters, including piloting of the system, is needed. • Severance pay is an important option because it is already provided in many developing and transition countries. It offers few advantages, however. To improve its protection and efficiency effects, countries may consider streamlin- ing these systems and reducing their costs if they are too generous. A more radi- cal reform could introduce prefunding, to improve nonperformance and ameliorate labor market rigidities. • Public works program are effective in reaching the poor, have good targeting properties and substantial capacity to redistribute income from the rich to the poor, are able to attract informal sector workers, can provide a flexible and fast response to shocks, and are administratively less demanding than other public income support systems. Although public works also have high nonwage costs, are likely to have a countercyclical pattern of funding, and may stigmatize par- ticipants, they are suitable for developing countries, particularly as a comple- mentary program. Notes Milan Vodopivec is a senior economist in the Social Protection Unit, Human Development Network, South Asia Region at the World Bank. His e-mail address is mvodopivec@worldbank.org. This article draws heavily on Vodopivec (2004). It is part of a larger effort by the World Bank’s Labor Market Group, Human Development Network, to understand better and synthesize the effects of labor market policies and programs. 1. For a recent summary of the efficiency effects of unemployment insurance as well as other key income support programs, see Vodopivec (2004). 2. While the association between more generous benefits and longer durations of unemployment has long been established, recent empirical studies also persuasively show that more generous unem- ployment benefits cause exits from unemployment to slow down (see, for example, van Ours and Vodopived forthcoming). 3. Several transition countries ran deficits financed out of state budgets in the 1990s (Vodopivec, Wörgötter, and Raju 2005). 84 The World Bank Research Observer, vol. 21, no. 1 (Spring 2006) 4. For an excellent review and evaluation of unemployment insurance savings accounts in Latin- American countries, see Ferrer and Riddell (forthcoming). 5. Coady, Grosch, and Hoddinott (2004) compare four targeting methods: individual assessment (based on means testing or “proxy means testing,� that is, construction of a targeting index from income and wealth attributes); categorical targeting (such as child allowances); self-selection methods (such as public works and food subsidies); and community assessment. 6. Coady, Grosch, and Hoddinott (2004) find that community involvement in some countries does help improve targeting. In contrast, Galasso and Ravallion (2005) find evidence that the more unequal distribution of assets in a village, the better positioned the nonpoor are to capture the benefits of the program. 7. For a careful analysis of different unemployment insurance savings account designs and their effects on protection and efficiency, see Parsons (forthcoming). 8. Smetters (2000) assesses the risk of having high rather than low to medium administrative costs of private pension accounts in the United States. A similar assessment is valid for unemployment insur- ance savings accounts and for other countries. To keep the costs of private accounts low, Smetters pro- poses that investment funds be approved and regulated by the government and subject to standard auditing controls to reduce fraud. He also proposes limits on investment charges and on the free move- ment of money between funds. In such a case, most of the administrative costs would come from col- lecting contributions from individual workers, at few extra costs compared with the public system. 9. Countries are grouped by per capita income, a proxy for key institutional and labor market conditions. 10. To affect the unemployment insurance system, the following administrative procedures would have to be put in place: making, assessing, and approving claims for compensation; making payments to approved claimants; monitoring the continuing eligibility of beneficiaries; providing employment and training services to beneficiaries; and auditing, monitoring, and evaluating the system. 11. Other factors are susceptibility to political risk (for example, the danger of using unemployment insurance funds for other purposes) and the nature of collective bargaining (under fragmented bar- gaining, the introduction of unemployment insurance is more likely to generate wage pressures). See Vodopivec (2004). References Acemoglu, D., and R. Shimer. 1999. “Efficient Unemployment Insurance.� Journal of Political Economy 107 (5): 893–928. ———. 2000. “Productivity Gains from Unemployment Insurance.� European Economic Review 44 (7): 1195–1224. Acevedo, G., P. Eskenazi, and C. Pages. Forthcoming. “The Chilean Unemployment Insurance: A New Model of Income Support for the Unemployed?� In R. Holzmann and M. Vodopivec, eds, Improving Severance Pay: An International Perspective. World Bank, Washington, D.C. Amadeo, E. J., I. S. Gill, and M. C. Neri. 2002. “Assessing the Impact of Regulations on Informal Workers in Brazil.� In I.S. Gill, C.E. Montenegro, and D. Domeland, eds, Crafting Labor Policy: Techniques and Lessons from Latin America. World Bank, Washington, D.C. Atkinson, A. B. 1995a. Incomes and the Welfare State: Essays on Britain and Europe. Cambridge: Cambridge University Press. ———. 1995b. “On Targeting Social Security: Theory and Western Experience with Family Benefits.� In D. van de Walle and K. Nead, eds, Public Spending and the Poor: Theory and Evidence. Baltimore, Md.: John Hopkins University Press. Atkinson, A. B., and J. Micklewright. 1991. “Unemployment Compensation and Labor Market Transi- tions: A Critical Review.� Journal of Economic Literature 29 (4): 1679–1727. Milan Vodopivec 85 Attanasio, O., and J. V. Rios-Rull. 2000. “Consumption Smoothing in Island Economies: Can Public Insurance Reduce Welfare?� European Economic Review 44 (7): 1225–58. Beegle, K., E. Frankenberg, and D. Thomas. 1999. “Measuring Change in Indonesia.� Labor and Popu- lation Program Working Paper Series 99–07. RAND, Santa Monica, Calif. Besley, T., and R. Burgess. 2004. “Can Labor Regulation Hinder Economic Performance? Evidence from India.� Quarterly Journal of Economics 119 (1): 91–134. Betcherman, G., and R. Islam. 2001. East Asian Labor Market and the Economic Crisis: Impacts, Responses, and Lessons. World Bank, Washington, D.C. Betcherman, G., K. Olivas, and A. Dar. 2004. “Impacts of Active Labor Market Programs: New Evi- dence from Evaluations with Particular Attention to Developing and Transition Countries.� Social Protection Discussion Paper Series 0402. World Bank, Washington, D.C. Blanchard, O. 2000. “The Economics of Unemployment: Shocks, Institutions, and Interactions.� Lionel Robbins Lectures, London School of Economics. [http://econ-www.mit.edu/faculty/ download_pdf.php?id=800]. Blanchard, O., and J. Wolfers. 2000. “The Role of Shocks and Institutions in the Rise in. European Unemployment: the Aggregate Evidence.� Economic Journal 110 (462): 1–33. Boone, P., P. Fredriksson, B. Holmlund, and J. van Ours. 2001. “Optimal Unemployment Insurance with Monitoring and Sanctions.� Discussion Paper 3082. Centre for Economic Policy Research, London. Botero, J., S. Djankov, R. La Porta, F. Lopez-de-Silanes, and A. Shleifer. 2002. The Regulation of Labor. NBER Working Paper 9756. Cambridge, Mass.: National Bureau of Economic Research. Chimerine, T., T. S. Black, and L. Coffey. 1999. “Unemployment Insurance as an Economic Stabilizer: Evidence of Effectiveness over Three Decades.� Unemployment Insurance Occasional Paper 99–8. U.S. Department of Labor, Washington, D.C. Coady, D., M. Grosch, and J. Hoddinott. 2004. The Targeting of Transfers in Developing Countries: Review of Experiences and Lessons. World Bank, Washington, D.C. Davis, S., and J. Haltiwanger. 1999. “Gross Job Flows.� In Orley Ashenfelter and David Card, eds, Hand- book of Labor Economics. Amsterdam: North Holland. De Ferranti, D., G. E. Perry, I. S. Gill, and L. Serven. 2000. Securing Our Future in a Global Economy. World Bank Latin and Caribbean Studies. World Bank, Washington, D.C. Dungan, P., and S. Murphy. 1995. “The UI System as an Automatic Stabilizer in Canada.� Human Resources Development Canada, Hull, Quebec. Ehrlic, I., and G. Becker. 1972. “Market Insurance, Self-Insurance, and Self-Protection.� Journal of Political Economy 80 (4): 623–48. Esguerra, J., M. Ogawa, and M. Vodopivec. 2001. “Options of Public Income Support for the Unem- ployed in the Philippines.� Philippine Review of Economics 38 (2): 37–66. Fallon, P. R., and R. E. B. Lucas. 1991. “The Impact of Changes in Job Security Regulations in India and Zimbabwe.� World Bank Economic Review 5 (1): 395–413. Feldstein, M., and D. Altman. 1998. Unemployment Insurance Savings Accounts. NBER Working Paper 6860. Cambridge, Mass.: National Bureau of Economic Research. Ferrer, A., and C. W. Riddell. Forthcoming. “Unemployment Insurance Savings Accounts in Latin America: Overview and Assessment.� In R. Holzmann and M. Vodopivec, eds, Improving Severance Pay: An International Perspective. World Bank, Washington, D.C. Forster, M. 2000. “Trends and Driving Factors in Income Distribution and Poverty in the OECD Area.� OECD Labour Market and Social Policy Occasional Paper No. 42, Paris. Galasso, E., and M. Ravallion. 2005. “Decentralized Targeting of an Anti-Poverty Program.� Journal of Public Economics. Journal of Public Economics 89 (4): 705–27. 86 The World Bank Research Observer, vol. 21, no. 1 (Spring 2006) Gill, I. S., and N. Ilahi. 2000. “Economic Insecurity, Individual Behavior and Social Policy.� Paper pre- pared for the Regional Study “Managing Economic Insecurity in Latin America and the Caribbean.� World Bank, Washington, D.C. Gruber, J. 1997. “Consumption-Smoothing Effects of Unemployment Insurance.� American Economic Review 87 (1): 192–205. Guasch, J. L. 1999. “An Alternative to Traditional Unemployment Insurance Programs: A Liquidity- Based Approach against the Risk of Earnings Losses.� World Bank, Latin America and Caribbean Region, Finance, Private Sector, and Infrastructure, Washington, D.C. Haddad, L., and M. Adato. 2001. “How Efficiently Do Public Works Programs Transfer Benefits to the Poor? Evidence from South Africa.� FCND DP 108. International Food Policy Research Institute, Washington, D.C. Haffner, R., S. Nickell, G. Nicoletti, S. Scarpetta, and G. Zoega. 2001. “European Integration, Liberaliza- tion and Labour Market Performance.� In G. Bertola, T. Boeri, and G. Nicoletti, eds, Welfare and Employment in a United Europe. Cambridge, Mass.: MIT Press. Haltiwanger, J, S. Scarpetta, and M. Vodopivec. 2003. “How Institutions Affect Labor Market Out- comes: Evidence from Transition Countries.� Paper presented at the World Bank Economists Forum. Washington, D.C., April. Hamermesh, D. S. 1992. “Unemployment Insurance for Developing Countries.� World Bank Policy Research Working Paper No. 897. Washington, D.C. Hamermesh, D. S., and D. T. Slesnick. 1995. Unemployment Insurance and Household Welfare: Microeco- nomic Evidence 1980–93. NBER Working Paper 5315. Cambridge, Mass.: National Bureau of Eco- nomic Research. Heckman, J. J., and C. Pages. 2000. The Cost of Job Security Regulation: Evidence from Latin American Labor Markets. NBER Working Paper 7773. Cambridge, Mass.: National Bureau of Economic Research. Holmlund, B. 1998. “Unemployment Insurance in Theory and Practice.� Scandinavian Journal of Eco- nomics 100 (1): 113–41. Holzmann, R., K. Iyer, and M. Vodopivec. Forthcoming. “Severance Pay Programs around the World: Rationale, Status, and Reforms.� In R. Holzmann and M. Vodopivec, eds, Improving Severance Pay: An International Perspective. World Bank, Washington, D.C. Hopenhayn, H., and J. C. Hatchondo. Forthcoming. “The Welfare Consequences of Alternative Designs of Unemployment Insurance Savings Accounts.� In R. Holzmann and M. Vodopivec, eds, Improving Severance Pay: An International Perspective. World Bank, Washington, D.C. ILO (International Labour Organization). 2001. LABORSTA. Geneva: International Labour Office. Koman, R., U. Schuh, and A. Weber. Forthcoming. “The Austrian Severance Payments Reform: Toward a Funded Pension Pillar.� In R. Holzmann and M. Vodopivec, eds, Improving Severance Pay: An International Perspective. World Bank, Washington, D.C. Kugler, A. 1999. “The Impact of Firing Costs on Turnover and Unemployment: Evidence from the Colombian Labour Market Reform.� International Tax and Public Finance Journal 6 (3): 389–410. ———. 2002. “From Severance Pay to Self-Insurance: Effects of Severance Payments Savings Accounts in Colombia.� IZA DP No. 434, Institute for the Study of Labor, Bonn, Germany. Lazear, E. P. 1990. “Job Security Provisions and Employment.� Quarterly Journal of Economics 105 (3): 699–726. Lipsett, B. 1999. “Supporting Workers in Transition: Income Support Programs for the Unemployed in Brazil and Argentina.� Paper prepared for the World Bank, Human Development Network, Washington, D.C. Milan Vodopivec 87 MacIsaac, D., and M. Rama. 2000. “Mandatory Severance Pay in Peru: An Assessment of Its Coverage and Effects Using Panel Data.� World Bank, Development Research Group, Public Service Delivery, Washington, D.C. Mansor, N., T. E. Chye, A. Boehanoeddin, F. Said, and S. M. Said. 2001. “Malaysia: Protecting Workers and Fostering Growth.� In G. Betcherman and R. Islam, eds, East Asian Labor Market and the Eco- nomic Crisis: Impacts, Responses, and Lessons. World Bank, Washington, D.C. Marquez, G. 1999. “Labor Markets and Income Support: What Did We Learn from the Crisis?� Inter- American Development Bank, Office of the Chief Economist, Washington, D.C. Mazza, J. 1999. “Unemployment Insurance: Case Studies and Lessons for Latin America and the Caribbean.� Technical Study RE2/SO2, Inter-American Development Bank, Washington, D.C. OECD (Organisation for Economic Co-operation and Development). 1999. Employment Outlook. Paris. ———. 2000. Employment Outlook. Paris. Orszag, J. M., P. R. Orszag, D. J. Snower, and J. E. Stiglitz. 1999. “The Impact of Individual Accounts: Piecemeal vs. Comprehensive Approaches.� Paper presented at the Annual Bank Conference on Development Economics, World Bank, Washington, D.C., April 29. Parsons, D. Forthcoming. “Savings Accounts as an Alternative to Job Separation Pay: Theoretical Con- siderations.� In R. Holzmann and M. Vodopivec, eds, Improving Severance Pay: An International Per- spective. World Bank, Washington, D.C. Ravallion, M. 1991. “Reaching the Rural Poor through Public Employment: Arguments, Evidence, and Lessons from South Asia.� World Bank Research Observer 6 (2): 153–75. ———. 1999. “Appraising Workfare.� World Bank Research Observer 14 (1): 31–48. Rawlings, L. B., and G. M. Rubio. 2003. “Evaluating the Impact of Conditional Transfer Programs: Les- sons from Latin America.� Policy Research Working Paper 3119. World Bank, Washington, D.C. Rawlings, L. B., L. Sheburne-Benz, and J. Van Domelen. 2003. Evaluating Social Funds: A Cross-Country Analysis of Community Investments. World Bank, Washington, D.C. Smetters, K. 2000. “The Design and Cost of Pension Guarantees.� Brookings/SIEPR/TIAA-CREF Con- ference on Public Policies and Private Pensions, Washington, D.C., September. Subbarao, K. 1997. “Public Works as an Anti-Poverty Program: An Overview of Cross-Country Experi- ence.� American Journal of Agricultural Economics 79 (2): 678–83. ———. 2003. “Systemic Shocks and Social Protection: Role and Effectiveness of Public Works Pro- grams.� Social Protection Discussion Paper 0302. World Bank, Washington, D.C. U.S. Social Security Administration. 1999. Social Security Programs throughout the World: 1999. Washington, D.C. ———. 2002a. Social Security Programs throughout the World: Asia and the Pacific 2002. Washington, D.C. [www.ssa.gov/policy/docs/progdesc/ssptw/2002-2003/asia/index.html]. ———. 2002b. Social Security Programs throughout the World: Europe. Washington, D.C. [www.ssa.gov/policy/docs/progdesc/ssptw/2002-2003/europe/index.html]. van Ours, J., and M. Vodopivec. Forthcoming. “How Shortening the Potential Duration of Unemploy- ment Benefits Affects the Duration of Unemployment: Evidence from a Natural Experiment.� Journal of Labor Economics. Vodopivec, M. 2004. Income Support for the Unemployed: Issues and Options. World Bank, Washington, D.C. ———. 2005. “Introducing Unemployment Insurance in Sri Lanka.� World Bank, Human Develop- ment Network, South Asia Region, Washington, D.C. Vodopivec, M., and T. Rejec. Forthcoming. “Unemployment Insurance Savings Accounts: Simulation Results for Estonia.� In R. Holzmann and M. Vodopivec, eds, Improving Severance Pay: An Interna- tional Perspective. World Bank, Washington, D.C. 88 The World Bank Research Observer, vol. 21, no. 1 (Spring 2006) Vodopivec, M., A. Wörgötter, and D. Raju. 2005. “Unemployment Benefit Systems in Central and East- ern Europe: A Review of the 1990s.� Comparative Economic Studies 47 (4): 615–51. Vroman, W. 2002. “Unemployment Insurance and Unemployment Assistance: A Comparison.� Social Protection Discussion Paper 0203. World Bank and Urban Institute, Washington, D.C. Wodon, Q. 2000. Poverty and Policy in Latin America and the Caribbean. World Bank Technical Paper 46. Washington, D.C. World Bank. 2003. World Development Indicators and Global Development Finance central database. Washington, D.C. Yoo, K. S. 2001. “The Feasibility of Introducing an Employment Insurance Scheme in the Philippines.� Paper prepared for the World Bank–International Labour Organization–Japanese Ministry of Labor–Philippine Department of Labor and Employment seminar on “Labor Market Policies: Their Implications for East and South East Asia.� March 1–2, Manila. Yun, J. 2001. “On the Integration of Unemployment Insurance with Pension through Individual Sav- ings Account.� World Bank, Human Development Network, Washington, D.C. Milan Vodopivec 89