World Bank, June 2012

Transcription

1 Ana Abras Alejandro Hoyos Ambar Narayan Sil Sailesh Tiwari i Poverty Reduction and Equity (PRMPR) World Bank, June 2012

2 Inequality of opportunities in labor markets The problem: How to measure inequality of opportunities in the labor markets? What is an opportunity in the labor market? Minimum level of pay defined by the minimum wage, good working conditions such as reasonable workweek hours, formality and stability represented by contracts, tenure, fringe benefits? How do we measure the effect of circumstances on labor market outcomes?

3 Two measures Inequality adjusted coverage indicators in the labor market. HOI approach. The opportunity is to be employed. We use some restrictions such as working for more than 20 hours per week, having tenure of at least 6 months, having contracts or a professional job. a professional job. Inequality of labor income/consumption due to circumstances. Measure of inequality used: Theil-L index

4 Roadmap Adapt the Human Opportunity Index (HOI) methodology to study inequality of opportunity in the labor market for working age adults using data from the Life in Transition Surveys (LiTS) conducted in Decompose the observed inequalities into components that are attributable to circumstances beyond an individual s control (e.g., gender, parents education, minority status, etc) and characteristics such as education and age. Additional exercises to examine the extent to which measured inequality of opportunity 1. Corresponds with alternative measures of inequality, such as the Theil index; R t ith i di id l ti f lif ti f ti d f i d 2. Resonates with individual perceptions of life satisfaction and fairness; and 3. Compares with inequality of opportunity in labor markets among countries in LAC using the same methodology.

5 Roadmap Our primary finding is that inequality of opportunity in the labor market is a pervasive phenomenon in the Europe and Central Asia region. Gender and parental education are the main contributors to inequality in accessing jobs. The results are robust to different definitions of jobs as opportunities. We also find that our measure of inequality resonates well with expressed life satisfaction and perceptions of fairness in expected ways: higher inequality of opportunity is associated with lower life satisfaction and lower perceptions of fairness

6 Inequality adjusted coverage indicators in labor markets The same approach that we use for the human opportunity index. A measure of access to the opportunity adjusted by inequality. HOI C 1 D Ci, Ca Ci: circumstances Ca: Characteristics D Ci, Ca is an index of labor market inequality. It depends on the circumstances and characteristics, it can be decomposed using a Shapley decomposition in order to identify the contribution of circumstances to the total inequality.

7 Inequality adjusted coverage indicators in labor markets The HOI framework HOI C 1 D Ci, Ca Inequality Index (D) It is the share of the total t number of opportunities that t needs to be reallocated to ensure equality of opportunity, across characteristics and circumstances. n 1 D ( Ci, Ca ) 2CC k 1 C C k

8 Shapley Decomposition What is the relative contribution of each circumstance to the inequality index? The inequality index is a function of the set of circumstances and characteristics considered- sensitive only to inequality between groups (where the groups are defined by the circumstances and characteristics). D=D(z), D( where z is a realization of the vector of circumstances and characteristics. What is the marginal contribution of X variable to the inequality index What is the marginal contribution of X variable to the inequality index among the key set of circumstances and characteristics that are likely to be important drivers of such inequality and considered in the analysis?

9 Inequality adjusted coverage indicators in labor markets Shapley decomposition Once we get the contributions of each of the circumstances, we add them up and obtained the contribution of circumstances. Inequality Index Shapley Decomposition Circumstance 1 Circumstance 2 Circumstance 3 Circumstance 4 Gender, minority status, parents education, and parents political affiliation Characteristic 1 Age and educational attainment Characteristic 2

10 Inequality of opportunities in labor markets indicators: LiTS 2006 What do we consider as an opportunity in labor markets? With regard to working: With opportunities in the labor market: Be occupied With a workweek of more than 20 more than 20h with tenure of six months contracts or professional occupation Without opportunities in the labor market: Be occupied with a workweek of 20 or less hours Be without work and not looking for a job due to discouragement Be looking for a job (unemployed)

11 Labor markets inequality indicators LiTS 2006 Jobs with different characteristics and economic distress 100% 80% 60% 40% 20% 0% Slovenia Kyrgyzstan Lithuania Slovenia Kyrgyzstan Lithuania Slovenia Kyrgyzstan Lithuania Slovenia Kyrgyzstan Lithuania HOI Job with more than 20 hours Sample all Labor Force HOI Job with contract Sample of salaried workers, unemployed, and discouraged HOI Contract or self-emp pprofessional Sample of salaried workers, self-employed, unemployed, and discouraged Job with more than 20 hours and tenure Sample all Labor Force Coverage

12 Inequality adjusted coverage indicators in labor markets: LiTS 2006 Job with more than 20 h Job with contract Dissimilarity Index (%) Slovenia Kyrgyzstan Lithuania Dissimilarity Index (%) Slovenia Kyrgyzstan Lithuania Circumstances Age Education Circumstances Age Education Job with contract or Professional occupation Job with more than 20 and tenure (%) (%) Dissimilarity Index S lovenia Kyrgyzstan Lit thuania Dissimilarity Index Slovenia Kyrgyzstan Lithuania Circumstances Age Education Circumstances Age Education

13 Contributions to the observed inequality Age Education Circumstance Note: Job with more than 20h workweek, sample in the Labor Force

14 INEQUALITY IN CONSUMPTION

15 Inequality of consumption due to circumstances Theil-L Index (Consumption) Between Component (Due to circumstances) Within Component (Effort, luck) Decomposition using a Decomposition using a parametric approach

16 Decomposing inequality in consumption between and within groups 45% 40% 35% 30% 25% 20% 15% theil_b theil_w 10% 5% 0% Note: Full sample of adults

17 Theil-L vs. D-Index 40% Theil -Index 30% 20% 10% 0% 5% 10% 15% 20% D-Index Correlation coefficient: 34% Note: Job with more than 20h workweek, sample in the Labor Force. Theil-L calculated with full sample of adults.

18 Exploring the correlation between measured inequality and beliefs Life satisfaction All things considered, I am satisfied with my life now. Individuals reporting to agree or to strongly agree are deemed to be satisfied as opposed to those answering strongly disagree, disagree, or neither disagree nor agree. Fairness In your opinion, which of the factors in this list is the most important to succeed in life in this country now? Individuals answering factors such as effort, hard work, intelligence and skill are deemed to believe in fairness, while political connections, corruption ties or other reasons indicate skepticism about fairness.

19 Exploring the correlation between measured inequality and beliefs Logit of perception dummy personal characteristics: gender, having a 20h job, age and age2, education of the person, education of parents, self reported minority, religion affiliation, household consumption Robustness check Clustering at region X country level Country class dummies Inclusion of other beliefs d f f d l l h d A word of caution: Regressions aim at finding partial correlation, the predictive power is low. There are observations and the R squared of the full regressions are around 0.22 and 0.05 for measures of satisfaction and fairness, respectively.

20 Partial Correlations of Life Satisfaction and Perceptions of Fairness with Various Measures of Inequality dindex dindex gini theil_b theil_w theil_b Index Satisfaction if i Fairness Satisfaction if i Fairness Satisfaction if i Fairness Professional /contract *** *** *** ** *** (0.002) (0.002) (0.005) (0.007) (0.008) (0.007) Job 20h *** *** *** * *** (0.006) (0.005) (0.012) (0.017) (0.015) (0.022) *** *** *** (0.005) (0.004) (0.014) (0.012) (0.011) (0.014) *** *** *** *** * (0.014) (0.014) (0.043) (0.051) (0.035) (0.050) *** *** *** (0.003) (0.003) (0.008) (0.009) (0.008) (0.010) Including within ihi *** *** *** *** ** component (0.015) (0.015) (0.041) (0.051) (0.032) (0.050) theil_w Including between component *** ** *** *** * (0.003) (0.003) (0.007) (0.009) (0.008) (0.009) Clustering (country and no no yes yes yes yes region level) Class dummies no no no no yes yes Personal characteritics yes yes yes yes yes yes Other Beliefs yes yes yes yes yes yes Note: personal characteristics include gender, having a 20h+ job, age and age2, education of the person, education of parents, self reported minority, parents participation in the communist party, religion affiliation. 350 cluster units. Classes are CIS Middle, CIS Low, EU, South East, Other.

21 Partial Correlations of Life Satisfaction and Perceptions of Fairness with Various Measures of Inequality The Theil-b index is statistically and negatively correlated with fairness and satisfaction - robust across specifications The Theil-b does a better job than the Gini or other measures. Th ifi i ihb h ihi db f The specifications with both within and between components of inequality have the expected signs.

22 Final Remarks Overall, our findings show substantial inequality of opportunity in the labor markets in a number of countries the ECA region. Gender and parental education are the main drivers of this inequality. We also find that our measure of inequality resonates well with expressed life satisfaction i and perceptions of fairness in expected ways: higher inequality of opportunity is associated with lower life satisfaction and perceptions of fairness. We also find some suggestive evidence that while higher inequality of opportunity is associated with unfairness, inequality that is associated with differences in things such as effort and luck are generally perceived as being fair.

23 Robustness check: Selection correction bias Kyrgyzstan No Correction Correction for observables Correction for non-observables O f S Note: Opportunity of having a job with more than 20 h per week. Selection on observables partition the data with a dummy on age (youth or under 36 years old), gender, person s education, and parents education. Lithuania Slovenia

24 Satisfaction Fairness Full (1) () (2) () (3) () (1) () (2) () (3) () Regressions Results Good job? *** *** *** *** *** *** (0.036) (0.055) (0.051) (0.035) (0.041) (0.036) Age *** *** *** *** *** *** (0.005) (0.007) (0.007) (0.005) (0.006) (0.006) Squared age *** *** *** *** *** *** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Gender *** *** *** (0.032) (0.048) (0.047) (0.031) (0.035) (0.034) Secondary education *** *** *** (0.047) (0.075) (0.070) (0.047) (0.099) (0.093) Vocational training * * (0.047) 047) (0.082) 082) (0.073) 073) (0.046) 046) (0.068) 068) (0.061) 061) College educated *** ** * *** ** *** (0.057) (0.087) (0.083) (0.055) (0.081) (0.073) coll_post * (0.177) (0.215) (0.211) (0.167) (0.200) (0.211) second_father *** *** (0.048) (0.070) (0.068) (0.046) (0.070) (0.070) voc_father ** * (0.045) (0.066) (0.057) (0.044) (0.074) (0.057) father_coll_degree ** * (0.059) (0.081) (0.080) (0.057) (0.092) (0.080) father_coll_post (0.222) (0.250) (0.234) (0.229) (0.279) (0.294) parent_communist_party t t *** ** *** *** * * (0.044) (0.073) (0.058) (0.041) (0.050) (0.043) consider_minority (0.051) (0.069) (0.066) (0.050) (0.083) (0.080) lexp *** *** *** *** * * (0.024) (0.053) (0.036) (0.023) (0.048) (0.042) D Index *** ** * *** (0.006) (0.025) (0.033) (0.005) (0.019) (0.027) country_exp *** *** * ** (0.017) (0.088) (0.078) (0.016) (0.043) (0.074) Satisfied with the economy? *** *** *** *** *** *** (0.039) (0.101) (0.094) (0.039) (0.076) (0.053) reduce_eco ** * *** *** *** (0.041) (0.077) (0.075) (0.040) (0.057) (0.056) better_now *** *** *** *** *** *** (0.038) (0.066) (0.062) (0.038) (0.056) (0.055) better_market *** *** *** *** *** *** (0.035) (0.043) (0.039) (0.035) (0.053) (0.055) better_demo *** *** *** *** *** *** (0.035) (0.055) (0.056) (0.034) (0.063) (0.068) buddhist *** *** * (0.116) (0.193) (0.195) (0.135) (0.117) (0.217) jewish ** ** ** ** ** (0.416) (0.286) (0.306) (0.350) (0.380) (0.377) christian * *** (0.053) (0.104) (0.089) (0.051) (0.102) (0.094) muslim *** * *** ** (0.064) (0.222) (0.185) (0.064) (0.191) (0.159) other_religion * *** *** (0.141) (0.160) (0.165) (0.142) (0.104) (0.091) CI Midle (0.386) (0.222) EU (0.384) (0.295) Other (0.335) (0.260) South East ** (0.312) (0.255) _cons *** *** *** *** *** *** (0.212) (0.514) (0.550) (0.206) (0.440) (0.484) r2_p N Clustering (country and region level) No Yes Yes No Yes Yes Country group dummies No No Yes No No Yes Personal characteritics Yes Yes Yes Yes Yes Yes Other Beliefs Yes Yes Yes Yes Yes Yes

25 Inequality of opportunities in labor markets indicators: LiTS 2006 Job with more than 20 h Job with contract Gender 0% 20% 40% 60% 80% 100% Relative Contribution to Inequality due to Circumstances Parents at communist party Father's education Job with contract or Professional occupation Self-reported minority 0% 20% 40% 60% 80% 100% Relative Contribution to Inequality due to Circumstances Gender 0% 20% 40% 60% 80% 100% Relative Contribution to Inequality due to Circumstances Parents at communist party Father's education Job with more than 20 and tenure Self-reported minority 0% 20% 40% 60% 80% 100% Relative Contribution to Inequality due to Circumstances Gender Father's education Gender Father's education Parents at communist party Self-reported minority Parents at communist party Self-reported minority