Journal Issue
Share
Article

Do Self-Employment Programs Work?

Author(s):
International Monetary Fund. External Relations Dept.
Published Date:
January 1995
Share
  • ShareShare
Show Summary Details

PROGRAMS that help the unemployed start their own businesses have proven to be a potentially cost-effective tool in job creation in a number of industrial countries. They are not a panacea, however.

Job creation is an important concern not only in transition and developing countries with high unemployment rates but also in many industrial countries, especially in Western Europe, where unemployment has risen sharply in recent years. Since the late 1970s, a number of governments have introduced programs that help unemployed workers set up their own businesses.

Self-employment programs are not a cure for unemployment—a recent survey in nine countries belonging to the Organization for Economic Cooperation and Development (OECD) and two transition countries (see box) indicates that such programs attract only a small percentage of the eligible unemployed and generate fewer new jobs than had been believed—and evaluating their impact is difficult because some of the unemployed would start new businesses even if assistance were not available. But, when the costs are roughly equal, providing self-employment assistance is preferable to paying unemployment benefits: studies show that a higher percentage of program participants than nonparticipants succeed in finding long-term employment, even if their businesses fail.

In designing programs, a delicate balance must be struck: benefits must be sufficient to enable participants to set up viable businesses, while costs must be low relative to other forms of unemployment assistance. Mechanisms must be developed for channeling resources to the applicants most likely to benefit and for raising the survival rates of new businesses. These conditions are difficult to meet in the transition countries. In Hungary and Poland, self-employment programs cost more than unemployment benefits. There is a shortage of private sector expertise for screening and advising applicants, and businesses may have trouble getting financing because of underdeveloped capital markets.

How programs work

Self-employment assistance programs are not entitlement programs—to limit the number of participants and keep costs down, applicants are required to meet certain criteria. Most programs rely on the private sector for help with screening and with delivering services.

Screening. Most programs control costs by using self-screening mechanisms and minimizing the role of program administrators and third parties in the selection of participants. The purpose of screening is to weed out less-motivated individuals and business proposals with little chance of succeeding. In Canada, Ireland, the United Kingdom, and the United States, program participants must work full time in their businesses to qualify for assistance. Screening is more rigorous in some countries than others, but, in general, applicants go through a series of steps. At each step, some applicants withdraw as the opportunity cost of their time rises. In theory, those with the weakest proposals withdraw first.

Even light screening procedures, such as information sessions, applications, and interviews, can be effective. Applicants in the United Kingdom must attend an information seminar on the requirements for program participation; nearly half of the applicants drop out after the seminar. Some programs require that applicants obtain pre-entry business advice and training and prepare a business plan. More rigorous screening might include testing an applicant’s business knowledge and requiring successful completion of a business course before admission to the program. The Washington State program (United States) requires that applicants complete training sessions, develop an acceptable business plan, set up a business bank account, satisfy licensing requirements, and obtain adequate financing before they can collect a lump-sum payment of their benefits. In Poland, local labor offices screen applications before passing them to local labor councils, private-public bodies created by the Ministry of Labor and Social Protection. Pre-entry business training is required for applicants who have made it past the first screening of the local labor councils.

Who is eligible for self-employment assistance?

In all of the countries in the survey on which this article is based, only workers who are eligible for unemployment benefits can apply for self-employment assistance. Means-tested benefits may be used in place of insured benefits in all but three countries. In five of the OECD countries, individuals must be unemployed for a minimum period before they are eligible for self-employment assistance.

Countries participating in a survey of self-employment programs
ProgramYear

created
Number of

participants

1992
AustraliaNew Enterprise Incentive Scheme (NEIS)19853,349
CanadaSelf Employment Assistance Component (SEAC)19924,129
DenmarkEnterprise Allowance Scheme (EAS)1985
FranceAide aux Chômeurs Créateurs ou Repreneurs d’Entreprise (ACCRE)197949,337
GermanyÜberbrückungsgeld198631,587
HungaryÚjrakendési kölcsön (Újk)19891
HungaryMunkanélküliek vállalkozóvá válásának elösegitése (Mwe)19915,466
IrelandFÁS Enterprise Programme19871,642
NetherlandsBijstandsbesluit zelfstandigen (Bz Scheme)19851,300
PolandPozyczki no Indywidualna Dzialalnosc Gospodaroza (PNIDZ)19904,907
United KingdomThe Business Start-up Scheme (BSUS)198340,500
United StatesWashington Self Employment and Enterprise Development Demonstration1989755
United StatesMassachusetts Unemployment Insurance Self-Employment Demonstration1990600
Sources: National authorities in nine Organization for Economic Cooperation and Development (OECD) countries responded to a questionnaire about the features, cost, and performance indicators of self-employment programs. Information on independent program evaluations was solicited where available. Annexes containing these data are available in Promotion of Self-Employment Experience in OECD and Transitional Economies, by the authors, World Bank Education and Social Policy Discussion Paper No. 30, June 1994. Similar data were collected by the authors during field studies in Hungary and Poland.

Hungary’s Újk program proved too costly and was replaced by the less generous Mvve program. …. Data not available.

Sources: National authorities in nine Organization for Economic Cooperation and Development (OECD) countries responded to a questionnaire about the features, cost, and performance indicators of self-employment programs. Information on independent program evaluations was solicited where available. Annexes containing these data are available in Promotion of Self-Employment Experience in OECD and Transitional Economies, by the authors, World Bank Education and Social Policy Discussion Paper No. 30, June 1994. Similar data were collected by the authors during field studies in Hungary and Poland.

Hungary’s Újk program proved too costly and was replaced by the less generous Mvve program. …. Data not available.

Private sector delivery. Increasingly, governments contract with the private sector to provide the specialized business expertise needed to screen applicants and deliver services for self-employment programs. In 1991, the United Kingdom’s Department of Employment transferred the delivery of its training and business assistance schemes to Training and Enterprise Councils, new limited companies. Canada’s Self-Employment Assistance Component relies on a public-private organization, and private sector organizations have delivered services under Australia’s New Enterprise Incentive Scheme since 1990. Private sector agents are also involved in monitoring and evaluating service delivery.

Specialized business services are scarce in transition economies. To overcome this shortage, Poland’s Ministry of Labor and Social Protection has financed the development of private sector organizations to deliver pre-and post-entry business advisory services. The Mvve program in Hungary has attempted to stimulate demand for business services by paying half of their cost, but few participants have taken advantage of this opportunity.

Benefits. Most programs provide financial assistance in the form of either a periodic allowance or a lump-sum payment. Six of the nine OECD countries in the survey provide periodic allowances. The duration ranges from 26 weeks in Germany and the United States (Massachusetts program), to 180 weeks in Denmark; most countries offer allowances for up to one year. To avoid distorting incentives for participation, periodic allowances are closely linked to the benefits available to participants from unemployment insurance and social assistance programs.

Some countries offer lump-sum payments up front, to provide participants with capital to invest in their new businesses. However, upfront payments attract a greater number of applicants, putting pressure on screening mechanisms. (The Washington State program has twice as many participants as the Massachusetts program.) To reduce the incentive for abuse of lump-sum payments, France’s ACCRE program requires that participants operate their enterprise for at least 341 days or repay the lump-sum grant. The Washington State program demonstrates that lump-sum payments (made only after participants have passed five program milestones) increase beneficiary costs by about 20 percent but lead to the establishment of enterprises that are more capital-intensive—20 percent of the Washington State participants are involved in manufacturing, compared with 14 percent of the Massachusetts participants. Findings in France also indicate that lump-sum payments are used to support capital investments.

The Netherlands is the only OECD country in the survey that provides program participants with access to loans, in addition to paying benefits matching unemployment compensation. Applicants are not eligible for financial assistance unless they have been turned down by a commercial lender first, however.

Access to credit is an important feature of self-employment assistance programs in Poland and Hungary. Poland’s PNIDZ program offers, in addition to other benefits, loans that are three times as high, on average, as the maximum allowable unemployment benefits. Payback features are attractive—recipients who are current on their loan payments during the first 24 months can request forgiveness of up to 50 percent of the loan. Under Hungary’s Ujk program, the Ministry of Labor paid the interest on bank loans of up to $5,000; the maximum loan amount was subsequently raised to $6,300. The scheme was extremely favorable to participants, because of Hungary’s high inflation and interest rates, but too costly, and the loan program was dropped. Participants in Hungary’s Mvve program receive a six-month extension of unemployment benefits, plus a 50 percent rebate of the cost of business training, counseling, and credit security insurance.

Post-entry business services are compulsory in Canada, France, and the United States. They include training, monitoring of new businesses, and helping entrepreneurs to get over rough spots, and are provided by a mix of public and private agencies.

Other financial benefits may be offered to program participants. France, for example, exempts its participants from social security contributions for six months. Germany provides participants with an additional one third of their weekly self-employment allowance to cover social security costs. Canada offers supplementary allowances to help cover the cost of dependent care, travel, and living away from home.

Results

Four measures of the performance of self-employment assistance programs were examined: the participation rate, the number of new business start-ups that would have happened without the program (the so-called deadweight effect), the survival rate of new businesses, and the cost per participant. The two US programs were designed specifically for the purpose of measuring such impacts—participants were chosen at random from a group of eligible unemployed workers; those not chosen formed a control group.

Participation rate. The participation rate is low in both the OECD and the transition countries, below 3 percent of the unemployed, except for the Washington State program, which has a participation rate of approximately 4.5 percent. The size of program budgets and the rigor of the screening mechanisms were possible factors in the participation rate, but, even in France, which has the largest target group and the lightest screening, the participation rate is below 2 percent. The ratio of participants in eastern Germany, although double that in western Germany, is below 2 percent. In Poland, after rising unemployment expenditures forced the Government to cut back its program, the participation rate dropped from 2.9 percent to less than 1 percent. Interest-free loans and light screening in Hungary’s Ujk program attracted nearly half the unemployed in 1989, but fewer than 2 percent of the eligible unemployed participated in the Mvve program.

Education, age, and sex seem to be the main factors determining participation in self-employment programs—most participants are better-educated males in their 30s. The programs reach a surprisingly small number of women. In France, for example, only 20 percent of program participants are women, even though women account for more than half of unemployed workers. In Germany, a requirement that businesses meet a minimum income target is seen as a deterrent to women, who are likely to enter low-income, service activities. Self-employment programs are unlikely to affect rising unemployment rates for recent graduates in many countries—not only are the young less inclined to participate when self-screening is used but the evidence also suggests that the enterprises of participants over 35 years old are likely to have higher success rates.

Employment creation. Job creation has been less than expected. The US programs indicate that one of every four new businesses would have been started even without them. A less rigorous estimate, based on asking participants what they would have done had the self-employment program not been available, produced even higher deadweight estimates in Australia, Denmark, France, and the United Kingdom. Also, the businesses set up by program participants generated few additional jobs—after a year, only about 10 percent of the self-employed in Australia’s program had hired full-time workers; 12 percent had hired part-time workers. The numbers were similar in the United Kingdom. In France, four years after completing the program, only 29 percent of the participants had hired full-time and part-time workers. The multiplier effect in these cases is about one half job for each self-employed person.

Table 1How many businesses set up by program participants survive?
Intake yearMonths since

leaving program
Survival rate

(percent)
Australia19901254
Denmark 119891240
France19865451
Netherlands19853652
United Kingdom19917271
United States (Massachusetts)1990-921377
United States (Washington)19901563
Source: Authors’ survey.

This estimate is based on an extrapolation of register and survey data of participants who had been in the program for two years.

The number of weeks participants spent in the program varies; the mean was 48.4 weeks. The survey was conducted approximately 18 months after participants started the program, which can last 26-66 weeks.

Source: Authors’ survey.

This estimate is based on an extrapolation of register and survey data of participants who had been in the program for two years.

The number of weeks participants spent in the program varies; the mean was 48.4 weeks. The survey was conducted approximately 18 months after participants started the program, which can last 26-66 weeks.

Survival rate. The number of jobs created depends, in part, on how many businesses survive. In the OECD countries, the survey indicates that approximately one of every two businesses fails during the first year (Table 1). But is this rate higher or lower than that of new businesses set up by entrepreneurs who do not participate in self-employment programs?

Table 2Cost of self-employment programs varies widely
YearBenefit term

(weeks)
Annual cost per

participant1

(1990 dollars)
Australia19915210,700
Canada1989407,800
Denmark199218013,500
France1992..24,400
Germany (western)1992132,000
Germany (eastern)1992131,800
Ireland1992406,600
Netherlands19924019,0003
United Kingdom199126-663,300
United States (Massachusetts)1990-92246,700
United States (Washington)1989-90265,400
Source: Authors’ survey.

Excluding administrative costs, except in Australia and Canada.

France has no maximum term for receiving unemployment benefits.

Includes small enterprise loans.

Source: Authors’ survey.

Excluding administrative costs, except in Australia and Canada.

France has no maximum term for receiving unemployment benefits.

Includes small enterprise loans.

In France, the survival rate is lower for ACCRE participants than for the self-employed who have not received ACCRE assistance. However, ACCRE-assisted enterprises tend to have a smaller capital base, and the cultural, professional, and social backgrounds of the participants are perceived to be less oriented to enterprise creation. Similar conclusions are reached for the programs in Denmark and the Netherlands.

However, when businesses do not survive, there is evidence that program participants find it easier to get wage or salary employment than nonparticipants. The US demonstration programs showed statistically significant differences in employment rates for participants and nonparticipants. Of participants whose businesses did not survive, 27 percent found full-time jobs; 8 percent found part-time jobs; and 2 percent sold their businesses. Follow-up surveys in four countries find that approximately eight out of ten participants are either employed or self-employed.

Cost of programs. The cost of self-employment programs is a function of the services offered and varies widely (Table 2). Excluding administrative costs, expenditures per participant range from $1,800 in eastern Germany to $19,000 in the Netherlands. The costs of the Washington State program are about $1,100 higher per participant than the unemployment benefits paid to the control group. The higher cost per participant must be weighed against the higher overall employment rates for the self-employed and higher productivity rates where the self-employed work in more capital-intensive enterprises. In Massachusetts, self-employment assistance actually costs less than unemployment benefits paid to the control group.

In the OECD countries, where allowances paid by self-employment programs are closely linked to the benefits of income-support programs, the choice between paying unemployment benefits or providing self-employment assistance is not a difficult one, even with deadweight losses and low survival rates. The situation in the transition economies is different. After accounting for the repayment of PNIDZ loans in Poland, program costs are still 1.5 times greater than unemployment benefits. Hungary’s Ujk program was costly because the Ministry of Labor assumed interest payment obligations while failing to screen participants effectively. Participation exploded to nearly half the unemployed in 1989, before falling to 37 percent in 1990. Interest payments a year after the program was closed absorbed one third of the Government’s Employment Fund resources.

The evidence suggests that rigorous screening results in a bigger deadweight effect; a balance needs to be struck between controlling costs and reducing the deadweight effect. Screening is particularly important when program benefits are higher than unemployment benefits and attract many participants. The objective should be to increase the social and private benefits of self-employment assistance by identifying the unemployed who would need program support to start a new business or those whose new businesses would have a better chance of survival and create a greater number of new jobs with program support. Expenditures on training and advisory services need to be based on their impact on enterprise survival rates.

Funding of self-employment programs needs to be separated from unemployment benefits, because the latter will crowd out the former in times of high unemployment. In Poland, 14 percent of the Employment Fund budget was allotted to the self-employment assistance program in 1990, when unemployment was low. When unemployment rose sharply in 1991, resources were shifted to pay for unemployment benefits, and the PNIDZ program’s share of the budget fell to 1 percent in 1993.

The right environment

Programs are more cost-effective in the OECD countries than in the transition economies, in part because of the enabling environment: a legal infrastructure that creates a transparent investment environment, competitive markets for raw materials and supplies, a network of technical assistance services, and efficient capital markets. The transition economies need to take several steps to reduce program costs and improve performance. First, they need to create an enabling environment with sound macroeconomic policies. Second, they need to address credit market weaknesses; until the banking sector in these countries is more highly developed, they might consider lump-sum payments. Third, they need to develop a base of supporting business services by opening the market to organizations offering private sector expertise. Fourth, they need to separate financing of self-employment programs from unemployment programs.

Not a panacea

Although self-employment assistance programs are potentially beneficial, they cannot solve the problem of rising unemployment. The question in developing and transition countries is whether such programs will pay for themselves: they seem to have a modest impact on business start-ups and job creation, and there tend to be no unemployment or social assistance programs against which to offset their costs. Strategies focused on improving the enabling environment for small businesses may produce higher economic returns than self-employment assistance. In the industrial countries, governments need to include self-employment assistance in a menu of labor market policies designed to meet the diverse needs of the unemployed for productive employment.

This article is based on the authors’ study, Self-Employment for the Unemployed: Experience in OECD and Transition Economies, World Bank Discussion Paper No. 263, Washington, DC, 1994.

THE INTERNATIONAL MONETARY FUND PRESENTS

Effective Government Accounting

by A. Premchand

This book, the first of its type, seeks to bridge fiscal economics and government accounting. It examines historical developments, ingredients of existing systems, features of accounting standards, and other issues that governments must address to perform their tasks economically and efficiently. Readers gain insights into institutional linkages and the role of electronic technology. Arguing that government accounting systems are too important to be left to accountants alone, the book presents an interdisciplinary approach that goes beyond adapting commercial accounting formats to governments.

Available in English. 190 pp. (paper). 1995. ISBN 1-55775-485-3.

US$32.50

Stock # EGA-EA.

International Monetary Fund

Publication Services

Box FD-395

700 19th Street, N.W.

Washington, D.C 20431 U.S.A.

Telephone (202) 623-7430

Telefax: (202)623-7201

Cable Address: INTERFUND

INTERNET: PUBLICATIONS@IMF.ORG

American Express, MasterCard, and VISA credit cards

Other Resources Citing This Publication