Based on “ “Self-Employment in the Developing World” ( - PowerPoint Presentation

1. Based on ““Self-Employment in the Developing World” (Gindling and Newhouse)and“Who Are the World’s Entrepreneurs and What Constraints Do They Face?” (Margolis and Robalino) Who are the Self-Employed in the Developing World?

2. Motivation (1)Wage and salary employment is lacking in the developing world

3. Motivation (2)Self Employment is a key source of jobs in the developing world

4. OutlineConstrained occupational choice model of employment status determinationDataDescriptive statics: Who are the self employed?Determinants of “successful” self employmentDistribution of unsuccessful self employed with potential for success

5. Occupational choice with constraintsCompare the value of self-employment / entrepreneurship and wage / salary employmentDeterminants of the value of a wage/salary job:Likelihood of finding a jobAgeLevel of education and skills (technical, cognitive, and non-cognitive)Richness of their social network.Stability of the jobIncome from the jobNon-pecuniary benefitsPossible access to social protectionHealth and safety risksDeterminants of the value of self employment/entrepreneurship:Availability of entrepreneurial earnings opportunityStartup capitalStability of income streamAvailability of alternative income insurance mechanismsSkillsIndividual preferences, risk aversion, time preferences, etc… determine weights

6. Occupational choice with constraintsInternal ConstraintsLimited skillsTight market for salaried jobs (may be linked to rural/urban status)Limited capital/collateralExternal constraints“Subsistence activity” is economically non-viableUse of inappropriate technologies Limited access to capitalLack of information about better technologies or opportunitiesMobility restrictions.Regulations that make it difficult to create a business or access creditLack of infrastructureProduct market imperfections that reduce competitionProperty rights Contract enforcement Lack of information and ability to process it.Lack of adequate social protection systems

7. Occupational choice with constraintsA framework for thinking about who becomes self employedAllows us to understand why some individuals might be less successful than othersInternal or External reasonsThe frequency “constrained gazelles” provides an indication of the importance of constraintsWe use micro data to address these questions

8. Data – I2D2 v2(International Income Distribution Database)Compilation of harmonized household micro data sets from nearly 100 countriesMaintained by the World Bank’s Development Economics groupCovers all income levelsSome data sets have extra variablesHousehold consumptionEarnings (validity?)Assets

9. Data – I2D2 v2(International Income Distribution Database)No People’s Republic of China

10. Data – I2D2 v2(International Income Distribution Database)No United States of America

11. Who are the self employed? (Overall)In Low Income and Lower Middle Income countries, less than half of workers are wage and salary employees.Over 70% of workers in LICs are own account workers or non-paid employees, mostly in agriculture.As the GNI per capita increases, share of wage and salaried employees or employers increases, percent of workers who are own account, non-paid or in agriculture falls.

12. Who are the self employed? (Income level)Separating Non-agricultural workers into wage and salary, employer, own account and non-paidSeparating Agricultural workers into wage and salary, employer, own account and non-paidAs per capita GDP rises (to about 600-700 2005 US PPP dollars) workers transition out of non-paid employment and own account in agriculture and into non-agricultural own account. As countries move from low to lower middle income (from about $600 to $1200), there is a shift into wage and salaried work within both agriculture and non-agriculture. As middle income countries grow there is a structural transformation into non-agricultural wage and salary employment, and to a lesser extent non-agricultural employer, and out of all types of agricultural employment and non-agricultural own account.

13. Who are the self employed? (Education)Most educated: Non-agricultural employers and non-agricultural wage and salaried employeesLeast educated: Agricultural workers.Intermediate: Non-agricultural own account workers and Non-agricultural non-paid employees.Similar patterns for countries in all regions and income groups.In particular, as per capita GNI increases employers do not become more educated relative to the own account workers or wage and salaried employees Non-agriculture   Wage and Salaried WorkerNon-paid EmployeesEmployerOwn AccountAgricultureNot EmployedAll Countries9.47.110.46.94.26.7East Asia and Pacific10.38.39.87.55.78.5Europe and Central Asia13.010.512.810.510.010.2Latin America and Caribbean9.88.510.47.74.87.7Middle East and North Africa9.36.810.27.25.78.4South Asia7.06.410.36.23.45.3Sub-Saharan Africa9.65.78.36.24.26.3Low Income6.76.07.85.33.94.9Lower Middle Income8.56.910.16.84.16.2Upper Middle Income10.98.911.08.26.58.8

14. Who are the self employed? (Gender)For countries in all regions and income groups,Women are more likely to be non-employed or agricultural non-paid employees,Men are more likely to be in any other employment category.Men are more likely than women to be employers or own account workers in all regions.In general, women are less likely to be in high quality employment categories than are men.The biggest differences between men and women are in MENA and SA.

15. Who are the self employed? (Age)Proportion of both men and women who are employers increases with age from 15 until about 45 years old, and then remains relatively constant (until around 65, when the proportion of workers in all employment categories falls).Proportion of both men and women who are own account workers increases sharply with age until the late 30s, levels off, and then begins to fall from 40 on.For men, the proportion working as non-paid employees is high for teenagers, then falls sharply from after men reach 20 years old.For women, the proportion of working as non-paid employees is high for teenagers and remains high until they are about 40 years old, after which it begins to fall slowly.

16. Who are the self employed? (Sector)In general, wage and salaried employees are much more likely to be in services than any other industry sector. However, there are some exceptions:In EAP and SA, wage and salaried workers more likely in manufacturing than services.In LMICs, wage and salaried workers more likely in manufacturing than services.

17. Who are the self employed? (Consumption)Ranking from richest to poorest:Non-agricultural employersWage and salary workersOwn-Account and Unpaid workersNot EmployedAgricultural workers

18. “Successful” self employmentDefinition 1: EmployerMore robustSmall cell sizes for “success” Region and Income Level NON-AGRICULTURE AGRICULTURE(number of countries in sample)SuccessfulUnsuccessful SuccessfulUnsuccessfulAll Countries (89)2.114.4 0.815.4Low and Middle Income Countries (66)1.815.7 0.918.2Region (Low and Middle Income Countries)East Asia and Pacific (6)1.817.2 1.518.5Europe and Central Asia (13)2.65.0 0.34.7Latin America and the Caribbean (17)3.818.5 1.27.3Middle East and North Africa (4)4.08.7 5.810.2South Asia (4)0.715.6 0.517.8Sub-Saharan Africa (21)1.419.0 1.037.1 Per Capita GNILow Income (19)1.017.9 0.633.7Lower Middle Income (27)1.315.6 1.117.6Upper Middle Income (22)3.614.3 0.85.4High Income (24)3.57.5 0.21.8

19. “Successful” self employmentDefinition 2: Out of poverty ($2 / day consumption)Less cleanLarger cell sizes for “success” Region and Income Level NON-AGRICULTURE AGRICULTURE(number of countries in sample)SuccessfulUnsuccessful SuccessfulUnsuccessfulAll countries (45)7.79.34.314.1RegionEast Asia and Pacific (6)10.38.76.113.9Europe and Central Asia (7)4.60.32.00.8Latin America and the Caribbean (10)19.02.94.33.6Middle East and North Africa (3)10.02.411.94.9South Asia (2)5.110.83.415.1Sub-Saharan Africa (17)5.218.34.931.1Per Capita GNILow Income (13)5.715.04.925.3Lower Middle Income (20)6.89.94.414.6Upper Middle Income (12)13.21.7 3.21.9

20. Probit estimationModels include gender, education dummies and gender/age interactionsMean pseudo R-square for these Probits is 0.0834 for definition 1, and 0.1231 for definition 2 Models are run country by countryRobustness checks:Control for majority social groupWith and without Sector and Urban/Rural“Successful” self employment (Methodology)

21. “Successful” self employment (Results)The probability of being an employer is higher in urban areas than rural areas.In general, among industry sectors the probability of being successful is lowest in manufacturing.Males are more likely to be successfulThe probability of being successful increases with educationFor both men women:Probability of being an employer increases with age from 15 to 49 (and then remains about the same, or falls, for the 50-65 year old group).Probability of being non-poor increases with age, except in low income countries, where the probability of being successful is highest for those between 25 and 49 years old.

22. “Successful” self employment (Marginal effects)Definition 1: EmployerDefinition 2: $2/day consumption

23. Unsuccessful self employed with potential for success (Definition)Using estimated probability of success for each country self employed individual (Grimm, Knorringa and Lay, 2012):Calculate mean probability of success among the successful self employedDetermine a threshold such that the mean probability of success for all individuals with a predicted probability of success above the threshold among the “unsuccessful” is the same as the mean predicted probability for the “successful”Requires “common support” for the predicted probability distributionsThresholds are country specificDefine “potentially successful” as those individuals with a predicted probability of success above the threshold

24. Unsuccessful self employed with potential for success (Results)On average, in Low and Middle Income Countries 36-37% of the non-agricultural own account workers have a high potential to become successful.Share of potential employers increases with GNI per capita, but no such tendency with second definitionBy region:High share of “constrained gazelles” in Europe and Central Asia by both definitionsProportion of self-employed with potential to pull their households out of poverty is much lower in South Asia (only India and Bangladesh, however) than any other region. Sub-Saharan Africa has few self-employed with potential to be employers who are notDefinition 1 (Employer)Definition 2 ($2/day)All Low & Middle Income Countries (50, 38)36%37%Region (Low and Middle Income)East Asia and Pacific (6, 6)34%43%Europe and Central Asia (6, 2)55%63%Latin America and the Caribbean (15, 10)40%47%Middle East and North Africa (4, 3)41%50%South Asia (3, 2)36%29%Sub-Saharan Africa (16, 15)27%52%Per Capita GNILow Income (15, 12)34%42%Lower Middle Income (21, 17)34%35%Upper Middle Income (14, 9)42%47%High Income (23, 0)72%

25. ConclusionsSuccessful self-employed (and those with a high potential to be successful) are more educated, older, more likely to be a household head and less likely to be in agriculture.At all levels of development, most self employed workers are unsuccessful. Approximately 36% of unsuccessful self-employed have characteristics similar to successful self-employed As per capita income increases:Proportion of workers who are unsuccessful self employed falls Most of the unsuccessful self employed are absorbed into wage and salary workA small minority become successful entrepreneurs.Household consumption is correlated with employment statusEmployer households are wealthiestWage and salary worker households better off than own account or unpaid worker householdsAgricultural worker households are the poorest

26. ConclusionsIn the context of the constrained occupational choice model:The characteristics associated with (required by?) wage employment change with level of developmentEmployment opportunities in the wage and salary sector become increasingly prevalent as countries grow, but self employment is quantitatively important in the earlier stages of developmentAs constraints to entering wage employment fall, people who were previously constrained to go to self-employment (and are not likely to be successful) opt out of self employmentConstraints to success seem present in all regions and income levelsVary across regions, and across countries within regionFurther work is needed to identify which constraints are most relevant in each contextSuggests an agenda for policy intervention that needs to be tailored to country specificities and populations within countries