March 23, 2009 Presentation for Inclusive Growth Course

1. Growth, Employment and Poverty Reduction Paul Cichello PRMPR The World Bank March 23, 2009 Presentation for Inclusive Growth Course

2. Outline • Background: Understanding the workings of the labor market in developing countries and the resulting impact on growth and poverty outcomes. • Tools for conducting analysis in your country • Special emphasis on JOGGs decomposition software • Also discussing briefly other tools for labor market analysis

3. The quantity and quality of job creation is important for inclusive growth

4. Because … • The creation of more better jobs is an important transmission channel between growth, income inequality and poverty  an essential element of the inclusive growth agenda; • Mismatches and frictions in the labor market are • potential barriers to growth; • critical factors in determining the impact on growth and income distribution of macroeconomic shocks; • important determinant of impact of structural reforms —in trade, education, etc — on growth.

5. Employment and inclusive growth Note: We look at the LM comprehensively, including: • wage employment, • self-employment, and • un-paid work in household enterprises (both agricultural and non-agricultural)

6. Employment and inclusive growth When discussing the LM structure, we mean: the quantity and quality of the available workers, the quantity and quality of the available jobs, and the matching process linking workers and jobs • Do workers get matched where they are best suited? • Are certain types of individuals excluded? • The LM structure is influenced by LM regulations and institutions (and their impact on the above items) • It is also strongly influenced by other factors (exs. education system, financial markets, land markets, legal systems, communication systems, cultural norms, etc.)

7. Employment and inclusive growth The structure of the LM affects inclusive growth through its impact on • growth potentials; • the distribution of the benefits of growth; • the extent to which different groups in the population participate in the growth process. • i.e. the structure of the labor market impacts both efficiency and equity.

8. Employment and Efficiency Growth accounting is commonly usedmethodology: Transmission channels: The LM structure affects growth through its impact on accumulation of E and K.

9. Employment and Efficiency Barriers to growth: The prospects for sustainable growth in per-capita GDP may be severely impaired by: • Labor supply shortages due to adverse demographic trends (ECA) or inefficiently-low participation rates (MENA), • Skill shortages, • Inefficient input allocation and price setting may limit firms’ capacity to choose optimal combination of inputs.

10. Employment and Equity The capacity to sell labor is the main potential source of income for the poor  The impact of growth on employment is an important determinant of the extent to which growth is shared  The employment intensity of growth was conceived as an indicator of how much growth is ‘shared’ Focus on estimates of employment elasticity of growth.

11. Employment and Equity • In LICs, access to a job is not sufficient to guarantee adequate living standards as 18% of the those employed are estimated to be working poor • Segmentation may threaten the equitable distribution of the benefits of growth. • will talk about next

12. Introducing segmented labor markets

13. Essential Features of Segmented LM • The LM consists of various segments that offer qualitatively distinct types of employment for individuals with identical productivity endowments—i.e., ‘preferred segment’ and ‘least preferred segments’ • There is limited access to ‘preferred segments’: • Not all who want a job there (and can reasonably aspire to get one) can get it. • Three main issues to be addressed: • The wage/employment setting mechanisms within each sector, • The way the segments are connected • The role of frictions and barriers to mobility

14. Modeling a Segmented Labor Market URBAN SECTOR High-paySector Public Sector fixed wages and employment Non-employment Barriers? Barriers? Barriers? Private High-wage Sector: non-competitive wage setting Low-pay Sector: self employment/ informal Barriers? Barriers? Migration Decision RURAL SECTOR Rural non-agriculture Agriculture sector

15. Basic Implications of Segmentation • Cross-segment differences in marginal products may persist over the long-run. • Wages may differ from marginal product at least in some sectors (oligopsonistic/ bargaining/efficiency wage models/wage floors). • Labor supply may be abundant in some sectors but restricted in others. • Barriers to mobility prevent optimal reallocation of labor across segments (efficiency losses). • Poverty would be reduced and welfare enhanced if the causes of segmentation were addressed and those working in the least-preferred segments could gain access to the preferred jobs

16. Segmentation and Efficiency • Segmented structure of the labor market is a source of inefficiency as TFP losses may arise from the resulting factor misallocation. • These losses may be substantial: • Reallocation of labor away from agriculture explains 20% of cross-country variations in TFP growth [Temple & Woessmanm, 2006]. LM conditions—i.e., regulation, skill composition, etc.— are important barriers to inter-sectoral labor mobility High-productivity Sector • Link between • labor market conditions and • growth reinforced by potential negative impact on labor mobility Labor Market Conditions Low-productivity Sector Labor

17. Segmentation and Equity • Segmentation may threaten the equitable distribution of the benefits of growth. • The sectoral pattern of growth may be as important as the overall level in determining the extent to which it is poverty-reducing. • - If the ‘preferred’ sector(s) is also the most • dynamic, growth may reinforce pre-existing • differentials in MVP and earnings. • Barriers to cross-segment labor reallocation • may prevent workers in the ‘least preferred’ • sector(s) from benefiting from increased • employment opportunities in the ‘preferred’ • sector.

18. Policy Implications of Segmentation • Education may not guarantee access to the preferred segments if employment in these segments is rationed.  Increased education is an effective policy option only if (i) it increases sectoral mobility of labor and/or (ii) it increases returns in the least-preferred, non-rationed segments • It may be more effective to target the sectors in which the poor are not, so that growth in these sector can pull the poor out of the less productive sectors • Identifying and addressing barriers to mobility and other sources of segmentation (ie, matching efficiency, inefficient regulation) is a priority objective of development policies.

19. The need to revisit the role of employment in inclusive growth strategies

20. Required data • Cross-sectional data • Need data from two different periods • with employment status and category • with individual earnings (at least for wage workers) • with poverty data • Ex. Typical LSMS • Macro level data • GDP Value Added, by sector

21. Tools • Toolkit available to guide you every step of the way • Country Studies (Making Work Pay series) provide examples of analysis with cross-sectional data • Step-by-step guide explaining the methods for analyzing the data • JoGGs Excel program for decomposing GDP per capita • will see next • ADePT Labor: Linking Employment, Earnings & Job Type to Poverty • Skills Shortages/Mismatches • Direct examination of Labor Market Segmentation

22. Website with tools http://www.worldbank.org/employment

23. Decomposing the Change in GDP per capita March 23, 2009 Presentation for Inclusive Growth Course Using Rwanda as an Example Paul Cichello

24. Basic Trends, 2000-2006 Source: MINECOFIN (GDP), eicv1 & eicv2 (Employment and Population) * 94 2001 RWF = $1 US (PPP)

25. Decomposing change in GDP per capita Y/N = (Y/E)(E/A)(A/N) GDP (value added) per capita can be decomposed into 3 parts: • Output per worker • Employment to (Working Aged) Population Ratio • Working Aged Population to Total Population Decomposition of the change in GDP uses Shapley approach to divide each component As with all such decompositions, this does not provide a direct causal interpretation.

26. Increased share of the population was working aged (15-64) Decomposing change in GDP per capita

27. Lower participation is associated with a sizable decline in GDP per capita of 6,023 RWF 7.4% of GDP per cap in 2000; 36% of the ∆ in GDP per capita This was driven by 15 to 24 year-olds (which had sizeable gains in enrollment rates) Decomposing change in GDP per capita

28. Decomposing change in GDP per capita due to employment changes Note: Values represent the contribution to the change in GDP per capita Percentage values are in percent of total change in GDP per capita

29. Notice the increase in output per worker. This drives growth in GDP per capita. 25% of GDP per cap in 2000; 122% of the ∆ in GDP per capita Decomposing change in GDP per capita

30. Decomposing change in GDP per capita due to changes in output per worker Note: Values represent the contribution to the change in GDP per capita Percentage values are in percent of total change in GDP per capita

31. Changes in Value Added and Employment across sectors

32. Further analysis- details in guide • Examine inter-sectoral changes accounted for by sector • Decompose into changes in output per worker (within sectors) into TFP and capital-labor ratio components • Need information on: • share of national income going towards capital • Capital STOCK

33. Can it be done in practice??? This decomposition sounds complicated! Will it be time-consuming to compute??? LET’S SEE

34. Does it always give the same story? Review Exercises

35. Is this the end of our analysis? No- this is not mean to be “the answer” This was an accounting framework, not a causal framework Want to follow leads with further analysis in the micro-data

36. Looking at Micro DataQuick Note on Labor Market Indicatorsfor low/middle-income countries March 23, 2009 Presentation for Inclusive Growth Course

37. Motivation • Standard indicators inadequate in depicting labor market conditions in low/middle-income countries. • LM characteristics in low-income settings: • low labor productivity, segmentation, subsistence or survival employment, low regulation, low unemployment, self-employment and household labor activities, multiple jobs. • Need for indicators that describe the “quality of employment”. Primary focus: labor earnings.

38. Why aren’t they appropriate? Excluding some highly educated individuals, most of working-age population is employed, but workers are often underemployed and working poor. How much are they earning? • Median earnings • All workers, wage and salaried workers, self-employed, employers and household enterprise workers • Low earnings rate • All workers, wage and salaried workers, self-employed, employers and household enterprise workers

39. Low Earnings Line Low earnings line 1 = individual poverty line Why individual poverty line? To see if the worker earnings enough to individually escape poverty

40. Low Earnings Line Low earnings line 2 = individual poverty line x scaling factor for household dependency on individual earnings Why scale up? Earnings of worker also typically used to support other members of the household Proposed country-specific low earnings line: national individual poverty line x the median ratio of household members to working-age employed household members

41. Using ADePT Labor forLabor Market Analysis March 23, 2009 Presentation for Inclusive Growth Course

42. Data requirements • Nationally-representative household survey. • Key: individual level earnings data; want greatest coverage of workers possible. • Differentiate over worker type and sector of employment • Best if repeated over time with a labor market module and an income/consumption expenditure module.

43. An example

44. First set: Overview of the Labor Force • Table 1.1 gives a broad overview. • It will help you determine if unemployment is a problem to focus on or if you should be focused on quality of employment (low earnings, etc.). • It also helps you understand the level of child labor and changes over time and other interesting information. • Tables 1.3 gives an overview of where employment is located by type of work • Allows use of any designation you like not just employment type Ex. Formal/informal. • It will also break into ag/non-ag. • Table 1.4: Welfare status for each worker type • Including median earnings, low earnings rates, inequality of earnings, etc. • Table 1.5a: Shares of workers by sector • Table 1.5b: Share of workers by education (and broken ag/non-ag)

45. Next: Linking Poverty and Labor Market • Tables help you identify where poverty is concentrated • Type of workers & sectors • This section uses two ways of calculating outcomes a) Calculate it using individual sector of employment; b) Calculate it using the sector of employment of the household head. • The Tables in the Annex are used for follow up once an initial investigation reveals an area of interest - They give a disaggregation of main indicators by different groups for unemployment, low earnings, child labor, etc.

46. Where can you get ADePT- Labor? • ADePT software and resources are available on the web: • General guide for completing labor market analysis; • ADePT users guide with definitions for all labor market variables needed by ADePT and what the output files look like; • Technical users guide for running ADePT. • Link to download ADePT software • This requires STATA to run but can use either STATA or SPSS data files as inputs

47. What other analysis is there? • Example analytical work on: • Skill Mismatch/Shortages analysis • Direct examination of Labor Market Segmentation

48. What else is on the horizon? • Simulation model • Forecasting employment impact of shocks with multi-sector labor market models • Earnings mobility analysis • Taking advantage of panel data (where available) to better understand: • Typical earnings paths (and persistence of low earnings outcomes) • What explains paths to higher earnings? • Who is vulnerable to earnings declines?