## labormarket ## Source details **Canonical URL:** [labormarket](https://www.imf.org/-/media/files/publications/reo/whd/2019/october/english/labormarket.pdf) ## Other formats - [Markdown version](/-/media/files/publications/reo/whd/2019/october/english/labormarket.pdf.md) - [Structured JSON version](/-/media/files/publications/reo/whd/2019/october/english/labormarket.pdf.json) --- ### Introduction and key findings - Labor markets in Latin America are characterized by high levels of informality, low female participation rates and rigid employment protection legislation. - Informality continues to be a feature of roughly half the level of employment in the region. - Key summarized findings: - Limited cyclical fluctuations in unemployment are largely explained by larger countercyclical variations in informality, and to a lesser extent by variations in participation rates. - Informality is associated with high and uncertain redundancy costs, and with high minimum wages relative to workers productivity. - High informality lowers the speed of adjustment of the economy to shocks and hampers labor productivity growth. - The chapter examines policy options to reduce informality. ### Stylized trends: unemployment, informality, education, and participation - Unemployment - Unemployment trended downward since the turn of the century, reached a trough with the commodity price bust and started to creep up thereafter. - Regional variance: Brazil’s 2015–16 recession pushed the regional average up; Argentina and Uruguay saw increases since the bust; Mexico reached historical lows. - Informality - Informality accounts for over 50 percent of total employment in Latin America. - Within LAC informality ranges from around 30 to 70 percent among the largest economies. - Informality generally declines with income and higher education levels, with notable cross-country deviations (e.g., Mexico increased informality despite higher GDP per capita). - Education and formalization - Informality decreased partly due to higher skills in the labor force. - Counterfactual patterns: - Bolivia, Chile and Colombia: formality gains largely from increases in the share of skilled workers. - Peru and Ecuador: larger formalization gains occurred within education/skill levels. - Labor force participation - Participation increased since the 1990s, with reductions recently linked to downturns. - Gender gap persists despite steady increases in female participation over three decades. - Age and education patterns (median country): - 25–64 participation rose from about 71 percent to about 79 percent (mid-1990s to 2017). - 15–24 participation dropped by 6 percentage points over the same period. - 25–64 participation by education: low attainment ~66 percent in 1995 to 72 percent in 2017; high attainment ~87 percent. ### Labor market institutions and rigidity indicators - Indicators used: perceptions-based indicators (PBIs) and laws-and-regulations (L&Rs). - Notable rigidities: - Redundancy costs (weeks of salary) are higher than in AEs or other EMDEs. - Dismissal of even one worker often requires third-party approval in many countries. - Permanent contracts are mandatory for permanent tasks in many countries. - Central America stands out with a very high ratio of minimum wage to value added per worker. ### Decomposing unemployment dynamics and the role of informality - Decomposition identity used: u − u* ≈ −(part − part*) + (f − f*) + (pop − pop*) + (wpop − wpop*) - where u = unemployment rate; part = labor force participation rate; pop = working-age population; f = log ratio of formal to total employment; * indicates initial-period values. - Findings from decomposition: - Participation acted as an active margin of adjustment, mitigating unemployment fluctuations. - Labor formalization limited unemployment movements: during the 2000–11 boom formal employment increased substantially, often formalizing jobs rather than lowering unemployment (Mexico an exception). - Country cases: - Colombia: informality fell sharply during early 2000s boom, resumed as shock absorber during GFC; post-2011 continued fall in informality reduced the unemployment decline. - Chile (2007–11): increased formal demand met with increased participation and reduced informality, leaving unemployment stable. - Recent commodity bust period: Brazil and Mexico show lack of adjustment along informality margin (Brazil informality fell 2012–17 while unemployment rose; Mexico informality increased since early 2000s despite low/falling unemployment). Peru and Colombia continued formalization; Colombia aided by payroll tax reduction. ### Okun’s Law and cross-country evidence - Okun’s coefficients estimated with Common Correlated Effects (CCE) for 127 countries, 1990–2017. - Findings: - Unemployment responds less to output fluctuations in developing countries than in AEs. - LAC coefficients are somewhat larger than for EMs overall (except Central America) but smaller than in AEs. - High informality is the main factor behind smaller Okun coefficients: informality accounts for 36 percent of cross-country variation in Okun’s coefficients. - Interpretation: entry/exit into informal sector acts as an adjustment margin; lower unemployment responsiveness can mitigate social impacts of recessions where safety nets are weak. ### Econometric results: informality, participation and the cycle - Informality countercyclical response (averages from CCE estimator, panel of 17 countries, Observations = 356): - For every additional percentage point of GDP growth: - Unemployment rate falls on average by around 0.17 points. - Informality rate falls by 0.3 percentage points. - Specific coefficients (selected specifications): - ∆ GDPt = -0.209*** (0.0543) - ∆ GDPt = -0.266*** (0.0743) - ∆ GDPt = -0.261*** (0.0757) - ∆ GDPt-1 = -0.0607* (0.0325) and -0.0553 (0.0419) - ∆ GDPt-2 = 0.00315 (0.0547) - Heterogeneity: - Informality varies more with the cycle in Central America and Mexico than in South America. - Chile shows comparatively little informality variation with the cycle. - Informality and labor market characteristics (selected regression results, Observations vary across columns: 108, 104, 102, 105, 101): - Job security coefficient examples: 0.148*** (0.0347), 0.154*** (0.0356), 0.0123 (0.0331) - Government effectiveness: -0.547*** (0.0648), -0.557*** (0.0606) - Minimum wage / Labor productivity: 0.0704* (0.0417), 0.0961** (0.0398), 0.0390 (0.0281) - Interaction: Minimum wage / Labor productivity * Government effectiveness: 0.392** (0.154) - R-squared examples: 0.146, 0.027, 0.179, 0.620, 0.574 - Labor force participation (aggregate and gender; CCE estimates, Countries = 122, Observations ≈ 3,100+): - Total: ∆ GDPt = 0.0287** (0.0134); with lag: 0.0222* (0.0125) - Male: ∆ GDPt = 0.0355*** (0.0117); lagged: 0.0246** (0.0113) - Female: ∆ GDPt = 0.0114 (0.0149); lagged: -0.00360 (0.0142) - Interpretation: participation is procyclical but less elastic than unemployment/informality; responsiveness mainly driven by male participation; female participation less responsive in LAC than in AEs/other EMs, with heterogeneity (Chile and Peru more responsive; Caribbean less). ### Female participation in downturns - Evidence of countercyclical female participation during deep recessions. - Fixed-effects panel regressions for downturns (∆ GDP < 0): - Full sample (Countries = 103, Observations = 423): ∆ Female Participation ∆ GDPt = -0.0121 (0.00896) - LAC only (Countries = 16, Observations = 63): ∆ Female Participation ∆ GDPt = -0.355** (0.139) - Interpretation: in LAC downturns, female participation increases (insurance mechanism where women enter labor force to offset lost household income). ### Macro-flexibility and the speed of adjustment (economy-wide ECM) - CCE ECM results for 127/129 countries, 1990–2017 (selected averages and std. dev. from Table 5): - Lagged log employment: -0.205*** (0.021), -0.208*** (0.025), -0.216*** (0.027) - Lagged log GDP: 0.078*** (0.015), 0.085*** (0.016), 0.089*** (0.020) - GDP growth: 0.122*** (0.022), 0.112*** (0.020), 0.121*** (0.023) - Implied long-run elasticity: 0.3818*** (0.0842), 0.4064*** (0.0922), 0.4138*** (0.1051) - Observations examples: 33,203; 20,130; 13,081; Number of countries: 129 - Key numeric interpretation: - Average estimated speed of adjustment coefficient ≈ -0.21 (implied half-life of 3 years to close half the employment gap). - Heterogeneity: about 50 percent of countries have speed of adjustment higher than average; 75th percentile ≈ -0.41 (example countries: Nicaragua and El Salvador), half-life 1.3 years. - Regional medians: median LAC = -0.2; median AE = -0.28; median EM = -0.26. Within LAC, South America has lower median speed than Central America. - Implication: employment reversion to long-run levels is slow on average with important cross-country differences. ### Microeconomic speed-of-adjustment results (CCEM approach) - Two-step estimation following Caballero, Cowan, Engel, and Micco (2013); sample expansions and sector coverage noted. - Manufacturing (UNIDO) average results: - Manufacturing employment gap coefficient: 0.501*** (0.0427) — implies closing 50 percent of employment gap in each period. - Including services and construction: - Employment gap coefficient: 0.459*** (0.0410) — implies closing 45 percent of gap. - Selected interactions and coefficients from Table 6 (examples): - Employment gap * Informality: -0.0596*** (0.0216) - Employment gap * Job security: -0.0374*** (0.00561), -0.0223*** (0.00654) - Employment gap * (Minimum wage/Labor productivity): -0.131*** (0.0175), -0.0542*** (0.0185) - Employment gap * (Minimum wage/Labor productivity) * High government effectiveness: -0.335*** (0.0532) - Employment gap * High government effectiveness: 0.0469*** (0.0157) - Observations examples: 27988, 30895, 27647; Number of groups: 1604, 1693, 1586 (varies by column) - Implications: - LAC countries show slower adjustment: speed ~2.5 percentage points lower than non-LAC average in one specification. - Higher informality and stricter job security measures reduce the microeconomic speed of adjustment. - High relative minimum wages reduce flexibility, especially when enforcement (government effectiveness) is strong. ### Mechanisms: segmentation, enforcement, and minimum wages - Informal-firm types (Ulyssea, 2018): survival firms, opportunistic informal establishments, productive informal establishments. - Evidence suggests a large share of informal firms may be “survival” or opportunistic, limiting their role as a buffer for formal-sector firms. - Costs of switching from informal to formal jobs are large in Mexico and Brazil (Arias et al., 2018), supporting segmentation. - Enforcement matters: - Stronger de facto enforcement makes regulations more binding and can increase informality when regulations are stringent. - High government effectiveness has a negative direct correlation with informality but can amplify the adverse effect of rigid regulations via interaction terms. - Minimum wage effects: - High minimum wage relative to labor productivity increases informality and lowers adjustment speed. - Reducing minimum wages can harm equity and vulnerable groups; calibration and targeting are important. ### Annex quantified impacts (speed, half-lives, and growth correlations) - Half-life formula: 12*(log(0.5)/log(1-speed of adjustment)), measured in months. - Reported values (selected rows): - Average estimated speed: 0.50; implied half life (months): 12.00. - LAC estimated speed: 0.48; implied half life (months): 12.91. - Low (20th percentile) speed: 0.51; half life 11.79 months. - High (80th percentile) speed: 0.47; half life 13.26 months; implied speed differential (low-high): 0.17pp. - Low, High government effectiveness: speed 0.54; half life 10.56 months. - High, High government effectiveness: speed 0.43; half life 14.61 months; implied speed differential: 0.47pp. - Difference in speed between low and high employment protection is 10 percentage points when government effectiveness is high — ~4 additional months to close half the gap. - Difference in speed between low and high relative minimum wages is 16 percentage points when enforcement is high — ~6 additional months to close half the gap. - Growth associations: - Moving from the 80th to the 20th percentile of the speed distribution associated with an increase in medium-term labor productivity growth of 0.5 percentage points per annum. - Reducing minimum wage from 80th to 20th percentile when government effectiveness is high can increase growth by approximately 0.75 percentage points per annum. - Speed of adjustment coefficient is positively correlated with GDP per worker growth (2000–17) and is statistically significant at the 95 percent confidence level. ### Welfare, trade-offs, and policy implications - Main policy messages: - Reducing informality by 10 percentage points would bring LAC’s speed of adjustment roughly in line with EMs in the sample. - Reducing the job security composite index by 1 (e.g., eliminating third-party dismissal notification) would roughly increase LAC’s speed of adjustment to the AE average. - Policies to improve microeconomic flexibility that are most effective when enforcement capacity is high include easing job security regulations and better calibrating minimum wages relative to productivity. - Build unemployment insurance and other benefits concurrently with EPL changes to protect workers. - Structural reforms to increase labor productivity (TFP) are central: raising TFP has larger welfare and growth payoffs than flexibility reforms alone. - Combine payroll tax reductions with reductions in entry barriers and hiring/firing costs and TFP-boosting reforms for larger impacts on informality. - Tailor minimum wage policy by groups, regions, industries and cycle-phase to mitigate adverse employment effects. - Caveats and trade-offs: - Stricter enforcement of EPL can reduce informality on the extensive margin but may increase unemployment and reduce welfare if many workers cannot be absorbed by formal firms. - Reductions in minimum wages have distributive costs and require careful calibration. - Some policies that support small firms can unintentionally increase informality; penalties for tax evasion by formal firms can increase informality. ### Model simulations and micro–macro comparisons - DSGE model calibrated to Colombia and simulated: - Payroll tax reduction from 29.5 to 16 percent predicts a fall in informality of ~3 percentage points; empirical impact cited ~2 percentage points. - Steady-state scenario: 50-percent reduction in firm entry costs and hiring costs in formal sector combined with payroll tax cut of 4 percentage points: - Reduces informality by 7 percentage points. - Increases GDP per capita by about 3 percent. - Increases private consumption by about 3 percent. - Comparison: increasing TFP by 50 percent without changing rigidities: - Reduces informality by 7 percentage points. - Increases consumption and GDP by almost 30 percent. - Interpretation: microeconomic flexibility matters but boosting TFP is the key driver of development; flexibility complements TFP-enhancing policies. ### Box 1 — Mexico (informality and productivity) - Around 90 percent of firms operate informally in Mexico; these firms are mostly small and relatively less productive and employ over half the labor force. - Formal firms are on average about twice as productive as informal firms in value added per worker; productivity premiums persist after controlling for sector, location and firm size. - Policy simulation insights: - Reducing payroll taxes, social security contributions and firing costs can generate significant formal employment gains but only moderate aggregate productivity gains. - Reducing firm formalization costs yields much larger productivity gains by enlarging the productive formal-firm sector. - Regulatory distortions explain only a small share of total misallocation in Mexico; structural constraints to larger productive firms also need addressing. ### Methods, data sources and robustness - Econometric approaches: heterogeneous panel regressions with Common Correlated Effects (Pesaran, 2006); CCEM micro approach (Caballero et al., 2013); ECM for speed-of-adjustment. - Data sources and measures: - Informality: ILO share of informal employment in total non-agricultural employment; IDB SIMS database (time series using pension contributions for formality). - Manufacturing: UNIDO INDSTAT Rev. 3 (3-digit sectors); construction/services from OECD STAN or Timmer et al. (2000–2010) 10-sector database. - Labor protection and minimum wage variables: World Bank Doing Business indicators. - Government effectiveness: Kaufmann, Kraay, and Mastruzzi (2010) governance indicators; “High government effectiveness” dummy = 1 if 1996 estimate above global median. - Robustness: - Results robust to inclusion of interaction between GDP per capita and employment gap, sectoral fixed effects, and sector-time fixed effects. - Identification exploits sectoral labor intensity interactions with country-level regulation proxies. *Source: labormarket (PDF chapter), Regional Economic Outlook: Western Hemisphere, International Monetary Fund | October 2019* ### Introduction ### labormarket - Introduction ### Introduction - Labor markets in Latin America are characterized by high levels of informality, low female participation rates and rigid employment protection legislation. - Informality has declined in LAC labor markets but continues to be a feature of roughly half the level of employment in the region (Kugler, 2019 and World Bank, 2019). - Drivers of informality include trends in income per capita and education of the labor force (La Porta and Shleifer, 2014), strict labor regulations and taxation; informality also serves as a buffer for cyclical fluctuations in activity (Lambert, Pescatori and Toscani, 2019, Perry et al., 2007 and references therein). - The chapter studies the role of informality in affecting demand and supply forces in the labor market, particularly for fluctuations in employment and unemployment, and measures relative contributions of trend and cyclical drivers. ### Key findings (summary) - Limited cyclical fluctuations in unemployment are largely explained by larger countercyclical variations in informality, and to a lesser extent by variations in participation rates. - Informality is associated with high and uncertain redundancy costs, and with high minimum wages relative to workers productivity. - High informality lowers the speed of adjustment of the economy to shocks and hampers labor productivity growth. - The chapter reflects on options to reduce informality in labor markets. ### Some Stylized Trends — Unemployment - Unemployment has been on a downward trend since the turn of the century, reached a trough with the commodity price bust and started to creep up since then. - Regional variance is large; Brazil’s sharp recession of 2015–16 pushed the regional average up over the last three years. - Argentina and Uruguay saw increases in unemployment since the commodity price bust; most other countries experienced almost no changes in unemployment—with Mexico reaching historical lows. ### Some Stylized Trends — Informality - Informality in Latin America accounts for over 50 percent of total employment. - Informality levels in LAC are in line with emerging and developing economies, but significantly higher than in advanced economies (AEs). - Within Latin America informality ranges from around 30 to 70 percent among the largest economies. - Labor informality has generally decreased as countries’ income levels rose, though Mexico is a notable exception where informality has increased despite higher GDP per capita. - Even at the same income level there are important cross-country differences (Peru, Mexico, Argentina perform worse than peers at similar development). ### Definition and data notes on informality - Cross-sectional labor informality data comes from the ILO (focus on share of informal employment in total non-agricultural employment). - Time-series labor informality data for Latin American countries comes from the IDB’s SIMS database, which uses pension contributions as the criterion for formality. - In the ILO dataset, informal employment status is determined by criteria including the informal sector nature of the enterprise (ILO, 2018); informal firms are unregistered and/or small-scale private unincorporated enterprises. ### Some Stylized Trends — Education and Formalization - Informality has decreased as a result of an increase in the level of skills of the labor force. - Informality is lower for workers with higher—say, tertiary—education levels. - A counterfactual exercise keeping initial shares of employment by skill level constant shows two patterns of formalization gains: - (i) Bolivia, Chile and Colombia: formality increased largely due to improvements in education (increases in the share of skilled workers). - (ii) Peru and Ecuador: larger formalization gains occurred within education/skill levels. ### Some Stylized Trends — Labor Force Participation - Labor force participation rate has increased in LAC since the 1990s, with a noticeable reduction in recent years linked to the downturn in several economies. - Since the early 2000s LAC participation tends to be somewhat higher than typical emerging markets or advanced economies. - Large gap between female and male participation persists despite steady increases in female participation over three decades; closing this gap could boost potential growth (Novta and Wong, 2017). - Participation by age groups (median country): - 25–64 age segment increased from about 71 percent to about 79 percent (mid-1990s to 2017). - 15–24 age group participation dropped by 6 percentage points over the same period. - Participation by education level (25–64): - Low educational attainment increased approximately 6 percentage points in the last 20 years (from 66 percent in 1995 to 72 percent in 2017). - Adults with high educational attainment exhibit approximately 87 percent participation. ### Labor Market Institutions — Rigidity and Indicators - Perceptions-based indicators (PBIs) and laws-and-regulations (L&Rs) indicators are used to overview institutions. - World Economic Forum executive survey: flexibility of wage determination similar across Latin America vs. other groups; hiring and firing practices in South America perceived as substantially more rigid. - Summary indicators of employment protection laws do not show that South America has stronger EPL than elsewhere, but Latin American markets display notable rigidities in key dimensions: - Redundancy costs (weeks of salary) are higher than in AEs or other EMDEs. - Dismissal of even one worker often requires third-party approval. - Permanent contracts are mandatory for permanent tasks in many countries. - Ratio of minimum wage to value added: Central America stands out as having a very high ratio; cross-country comparison gives little evidence that minimum wage is more binding in South America than elsewhere. ### Decomposing Unemployment Dynamics - Unemployment in the region has been relatively stable over the commodity price super-cycle because demand fluctuations were met by fluctuations in participation and formalization. - The decomposition used (David, Lambert, and Toscani, 2019) expresses changes in unemployment relative to a reference period as: u − u* ≈ −(part − part*) + (f − f*) + (pop − pop*) + (wpop − wpop*) - where u = unemployment rate; part = labor force participation rate; wap = working-age population; lfF is log formal employment; f is log ratio of formal to total employment; * indicates initial period values. - Changes in labor supply captured by changes in participation and working-age population. - Participation acted as an active margin of adjustment, mitigating unemployment fluctuations; working-age population growth largely stable. - Labor formalization played a crucial role in limiting unemployment movements: - During 2000–11 boom, formal employment increased substantially; this led to formalization of jobs rather than lowering unemployment in most countries (exception: Mexico). - Colombia: informality fell sharply during early 2000s boom, resumed role as shock absorber during the GFC; post-2011 continued fall in informality implying unemployment did not fall as much as it would have otherwise. - Chile (2007–11): strong increase in formal labor demand met with roughly equal shares of increased participation and reduction in informality, leaving unemployment stable. - Recent period (commodity price bust) shows uneven patterns: - Brazil and Mexico show lack of adjustment along informality margin: - Brazil: informality continued to fall over 2012–17 even as unemployment increased strongly. - Mexico: informality increased since early 2000s despite low or falling unemployment. - Peru and Colombia continued strong formalization trends; in Colombia aided by an important reduction in payroll taxes. ### Labor Market Dynamics over the Cycle — Empirical approach - Use heterogeneous panel regressions to assess how unemployment, informality and participation move over the business cycle in LAC. - Estimation allows slope coefficients to vary across countries and accounts for cross-sectional dependency through inclusion of common factors. - Empirical specification (equation 2) uses changes in unemployment rate, informality rate, or participation rate as Z∆; yi∆ is change in log of output (real GDP); αi are country fixed effects; fm,t are common factors with country-specific loadings λi; εit is white noise. - Common Correlated Effects (CCE) estimator (Pesaran, 2006) is used to proxy unobserved common factors by cross-sectional averages of dependent and independent variables. ### Okun’s Law — Unemployment responsiveness to output - Okun’s law relates changes in output to short-term changes in unemployment. - Estimated Okun’s coefficients using the CCE estimator with annual data for 127 countries over 1990–2017. - Findings: - Unemployment responds less to output fluctuations in developing countries than in advanced economies. - LAC coefficients are somewhat larger than for EMs overall (except Central American economies) but smaller than in advanced economies. *Source: labormarket - Introduction (PDF chapter), Regional Economic Outlook: Western Hemisphere, International Monetary Fund | October 2019* ### 1. By Country Group ### By Country Group ### Informality and Okun’s Law - High informality is the main factor behind smaller Okun Law coefficients in EMs and particularly in LAC countries. - The response of unemployment to cyclical variations in output is weaker when informality is higher; informality accounts for 36 percent of the cross-country variation in Okun’s coefficients. - Interpretation notes: - Entry and exit into the informal sector act as a margin of adjustment to cyclical fluctuations. - Other institutional characteristics do not seem to affect the unemployment–output response, except possibly a World Economic Forum indicator on perceptions regarding the flexibility of wages. - Greater sensitivity of unemployment to the cycle is not necessarily desirable: in the absence of unemployment insurance or adequate social safety nets, greater responsiveness could reduce welfare (Ahn et al., 2019). Low elasticity of unemployment to GDP growth, tightly associated with informality, can mitigate adverse social impacts of recessions (Loayza, 2018). _Source data and measures_ - Informality measure: share of informal employment on total non-agricultural employment (ILO 2018 latest available year). ### Informality over the Cycle (Econometric Results) - Informality exhibits clear countercyclical behavior across LAC countries. - Quantitative comparison with unemployment: - For every additional percentage point of GDP growth: - The unemployment rate falls on average by around 0.17 points. - The informality rate falls by 0.3 percentage points. - CCE estimator results for responsiveness of informality to output fluctuations (Table 1; averages and standard deviations shown in source): - Specification (1): ∆ GDPt coefficient = -0.209*** (0.0543) - Specification (2): ∆ GDPt coefficient = -0.266*** (0.0743) - Specification (3): ∆ GDPt coefficient = -0.261*** (0.0757) - ∆ GDPt-1 coefficients: -0.0607* (0.0325) and -0.0553 (0.0419) in specifications including one lag - ∆ GDPt-2 coefficient: 0.00315 (0.0547) - Constants reported as 0.178 (0.247), 0.283 (0.388), 0.215 (0.587) - Observations = 356; Countries = 17 - Heterogeneity: - Informality rates vary more with the cycle in Central American countries and Mexico compared to South America. - Chile is noteworthy: informality does not vary much with the cycle relative to other countries in the region. ### Informality and Labor Market Characteristics - Cross-country associations: - Informality (ratio of informal employment over total non-agricultural employment) is positively associated with: - Redundancy costs (weeks of salary) - Minimum wage ratio (minimum wage to value added per worker) - De jure vs de facto enforcement: stronger enforcement makes regulations more binding in practice and can increase informality when regulations are stringent. - CCE regression results linking informality and de facto labor market regulations (Table 2): - Job security coefficient: 0.148*** (0.0347) and 0.154*** (0.0356) in different specifications; 0.0123 (0.0331) in another - Job security * Government effectiveness: 0.102** (0.0481) - Government effectiveness: -0.547*** (0.0648) and -0.557*** (0.0606) - Minimum wage / Labor productivity: 0.0704* (0.0417), 0.0961** (0.0398), 0.0390 (0.0281) - Minimum wage / Labor productivity * Government effectiveness: 0.392** (0.154) - Constants: 0.307*** (0.0468), 0.112 (0.205), -0.185 (0.206), 0.703*** (0.0511), 0.502*** (0.143) - Observations: 108, 104, 102, 105, 101 (across columns) - R-squared: 0.146, 0.027, 0.179, 0.620, 0.574 - Interpretation: - Redundancy costs and dismissal approval requirements are significantly correlated with informality (David, Lambert, and Toscani, 2019). - Government effectiveness has a negative direct correlation with informality but can amplify the effect of regulations on informality via interaction terms. ### Labor Force Participation over the Cycle - Aggregate and gender-specific participation rates move procyclically. - CCE estimator results for changes in labor force participation on changes in GDP (Table 3; coefficient averages as outlier-robust means): - Total (∆ GDPt): 0.0287** (0.0134) - Total with one lag: 0.0222* (0.0125) - Male (∆ GDPt): 0.0355*** (0.0117) and 0.0246** (0.0113) with one lag - Female (∆ GDPt): 0.0114 (0.0149) and -0.00360 (0.0142) with one lag - ∆ GDPt-1 coefficients generally small and not significant (e.g., 0.0123, 0.00111) - Constants vary across specifications (e.g., 0.0905 (0.118), 0.0351 (0.146), -0.197* (0.110), -0.286* (0.154), 0.144 (0.102), 0.185 (0.138)) - Observations: ~3,100+ across columns; Countries = 122 - Findings: - Sensitivity of participation rates to output changes is smaller than for unemployment or informality. - Responsiveness mainly driven by changes in male participation; female participation tends to be less elastic. - Significant heterogeneity across countries and regions: female participation less responsive to GDP fluctuations in LAC compared to AEs or other EMs; within LAC, Chile and Peru show more responsiveness for female participation, Caribbean countries show less. ### Female Labor Force Participation in Downturns - Evidence that female participation can be countercyclical during deep recessions (Novta and Wong, 2017; historical cases in Argentina, Mexico, Peru). - Scatter evidence: during periods with negative GDP growth, female participation tends to increase (moves countercyclically). - Fixed-effects panel regressions for downturn periods (∆ GDP < 0) — Table 4: - Full sample: ∆ Female Participation ∆ GDPt = -0.0121 (0.00896) - LAC only: ∆ Female Participation ∆ GDPt = -0.355** (0.139) - Constants: 0.0366 (0.362) for full sample; -0.323 (1.013) for LAC only - Country fixed-effects: Yes; Time fixed-effects: Yes - Observations: 423 (full), 63 (LAC only) - Countries: 103 (full), 16 (LAC only) - Interpretation: - For LAC economies, female participation increases during downturns (coefficient negative and statistically significant when focusing on LAC), consistent with an insurance mechanism where women enter the labor force to offset lost household income. ### Macro-Flexibility, Speed of Adjustment, and Employment Dynamics - Methodology: - Error-correction model (ECM) estimated using Pesaran (2006) heterogeneous panel (CEC approach) to capture speed of adjustment of employment to deviations from equilibrium employment implied by GDP. - Specification parameters: βi captures contemporaneous response of employment growth to GDP shocks; θi is long-run elasticity of employment to GDP; αi is speed of adjustment parameter. - CCE ECM estimation results for 127/129 countries over 1990–2017 (Table 5; averages and standard deviations): - No lags / 1 lag / 2 lags specifications: - Lagged log employment: -0.205*** (0.021), -0.208*** (0.025), -0.216*** (0.027) - Lagged log GDP: 0.078*** (0.015), 0.085*** (0.016), 0.089*** (0.020) - GDP growth: 0.122*** (0.022), 0.112*** (0.020), 0.121*** (0.023) - Implied long-run elasticity: 0.3818*** (0.0842), 0.4064*** (0.0922), 0.4138*** (0.1051) - Observations: 33,203; 20,130; 13,081 across specifications - Number of countries: 129; Number of lags of the cross-sectional averages noted - Key numeric summary and interpretation: - Average estimated speed of adjustment coefficient is -0.21 (implied from lagged log employment average), which implies a half-life of 3 years to close half the employment gap. - Heterogeneity: - Approximately 50 percent of countries have a speed of adjustment higher than the average. - 75th percentile speed of adjustment ≈ -0.41 (example countries: Nicaragua and El Salvador), implying a half-life of 1.3 years. - Regional medians: - Median country in LAC: speed of adjustment coefficient = -0.2 - Median advanced economy (AE): -0.28 - Median emerging market (EM): -0.26 - Within LAC, South America has a lower median speed of adjustment than Central America. - Implication: - Employment reversion to long-run levels is relatively slow on average, with meaningful cross-country and regional differences in adjustment speed that affect recovery dynamics and welfare costs of shocks. *Source: IMF staff calculations and analyses as presented in the PDF chapter "labormarket - 1. By Country Group."* ### 1. By Country Group ### 1. By Country Group ### Differences in speed of adjustment across countries - Differences in the coefficients of speed of adjustment of employment may be related to the characteristics of each country’s labor market and the institutions and regulations governing them. - The estimated coefficient for informality implies that if LAC decreased its informality rate by 10 percentage points—roughly the difference between the region’s average and the average for EMs in our sample of 110 countries—the region’s speed of adjustment would be roughly the same as the EM average. - If the region decreased its average job security composite index by 1 (for example, by eliminating third party dismissal notification), the speed of adjustment of employment would roughly increase to the estimated average for AEs. - Higher informality appears to attenuate the impact of shocks to GDP on unemployment, but it also makes the adjustment process more protracted. ### Microeconomic flexibility and empirical approach - The analysis follows CCEM (Caballero, Cowan, Engel, and Micco, 2013) to estimate a microeconomic speed of adjustment using the following specification (as presented): - ∆eit = αi + λi (e*it − eit−1) + gapt + εit - where ∆ is the first difference operator, eit is the natural logarithm of employment, e* is the (log) equilibrium level of employment, (e* − e t−1) is the employment gap in t-1, and λi is the speed of adjustment. - Two-step estimation: first construct a proxy of the employment gap following CCEM; second estimate the error-correction equation using that gap proxy. - Empirical expansions relative to CCEM: - Period changed to 2000–2017. - Assessment of informality and policies governing worker compensation in addition to labor market regulations. - Comparison of speed of adjustment using UNIDO manufacturing sectors versus a combined dataset including manufacturing, services and construction. ### Key empirical results (microeconomic sample results) - When focusing on manufacturing (UNIDO), on average manufacturing sectors close 50 percent of the employment gap in each period. - Employment gap coefficient (UNIDO, Column (1)): 0.501*** (standard error 0.0427) - When including services and construction, the coefficient drops to approximately 45 percent. - Employment gap coefficient (UNIDO + services and construction, Column (2)): 0.459*** (standard error 0.0410) - Table 6 reported coefficients (selected columns and interactions): - Employment gap: 0.501*** (0.0427) [UNIDO, Col (1)] - Employment gap: 0.459*** (0.0410) [10S+OECD, Col (2)] - Employment gap * LAC: -0.0255* (0.0143) - Employment gap * Informality: -0.0596*** (0.0216) - Employment gap * Job security: -0.0374*** (0.00561) and -0.0223*** (0.00654) (reported across specifications) - Employment gap * (Minimum wage/Labor productivity): -0.131*** (0.0175) and -0.0542*** (0.0185) - Employment gap * (Minimum wage/Labor productivity) * High government effectiveness: -0.335*** (0.0532) - Employment gap * High government effectiveness: 0.0469*** (0.0157) and 0.0973 (0.121) - Constant terms reported (examples): 0.00173*** (0.000611), 0.00460*** (0.000675), 0.00186*** (0.000618), 0.000832 (0.000688) - Observations examples: 27988, 30895, 27647, 20123, 27056, 26942, 27647, 27585 (reported in Table 6) - Number of groups examples: 1604, 1693, 1586, 1141, 1553, 1549, 1586, 1582 (reported in Table 6) - LAC countries exhibit slower speed of adjustment: - Column (3) shows the speed of adjustment for the average country in LAC is approximately 2.5 percentage points lower compared to countries outside the region. - Higher informality levels are associated with slower adjustment of employment (Column (4) and related specifications). ### Interpretation and mechanisms - Informality and segmentation: - Not all informal employment is equal; Ulyssea (2018) classifies informal establishments in Brazil into three types: “survival” firms, opportunistic informal establishments, and productive informal establishments that would formalize if costs were low enough. - Ulyssea finds roughly half of informal firms in Brazil are “survival” establishments and close to 40 percent are opportunistic, indicating a large share of informal employment may not be available for formal firms to tap into during downturns. - Arias et al. (2018) find large costs of switching from an informal job to a formal job in the same industry in Mexico and Brazil—costs comparable to switching jobs across sectors—supporting segmentation between formal and informal employment. - Labor market regulations and enforcement: - LAC has stricter EPLs compared to other countries and informality is associated with stricter EPLs. - Strict EPLs tend to hamper an economy’s reaction to economic shocks; the stiffening effect of EPLs increases with higher government effectiveness (more binding enforcement). - High minimum wages relative to labor productivity also erode microeconomic flexibility. - The negative effects of job security regulations and high relative minimum wages on the speed of adjustment are amplified when countries have a higher ability to enforce labor market regulations. ### Policy-relevant implications - Reducing informality by 10 percentage points is estimated to bring LAC’s speed of adjustment roughly in line with EMs in the sample. - Reducing the job security composite index by 1 (for example, eliminating third party dismissal notification) would roughly increase LAC’s speed of adjustment to the estimated average for AEs. - Easing job security regulations and lowering the minimum wage relative to labor productivity (or improving the calibration of minimum wages) could increase microeconomic flexibility, particularly in environments with strong enforcement capacity. - Policy reforms should account for labor market segmentation: many informal workers may not be easily absorbed by formal firms during downturns, so policies to reduce the share of “survival” and opportunistic informal establishments and lower costs of formalization for productive informal firms could improve adjustment dynamics. *Source: IMF staff calculations based on David, Pienknagura, and Roldos (2019).* ### Annex 2 suggests that the economy-wide impact of informality may be larger, given that the elasticity for industry is lo ### labormarket - Annex 2 suggests that the economy-wide impact of informality may be larger, given that the elasticity for industry is lo ### Speed of adjustment, informality, and key quantified impacts - Half-life is calculated as 12*(log(0.5)/log(1-speed of adjustment)) and is measured in months. - Average estimated speed of adjustment: 0.50; implied half life (in months): 12.00. - LAC estimated speed of adjustment: 0.48; implied half life (in months): 12.91. - Low (20th percentile) speed of adjustment: 0.51; implied half life (in months): 11.79. - High (80th percentile) speed of adjustment: 0.47; implied half life (in months): 13.26; implied speed differential (low-high): 0.17pp for one comparison row. - Low, High government effectiveness: speed 0.54; implied half life 10.56 months. - High, High government effectiveness: speed 0.43; implied half life 14.61 months; implied speed differential: 0.47pp. - Low, Low government effectiveness: speed 0.52; implied half life 11.39 months. - High, Low government effectiveness: speed 0.49; implied half life 12.32 months; implied speed differential: 0.11pp. - Low, High government effectiveness (another row): speed 0.55; implied half life 10.56 months. - High, High government effectiveness (another row): speed 0.39; implied half life 16.85 months; implied speed differential: 0.74pp. - Low, Low government effectiveness (another row): speed 0.51; implied half life 11.72 months. - High, Low government effectiveness (another row): speed 0.49; implied half life 12.48 months; implied speed differential: 0.09pp. - Difference in speed of adjustment between countries with low and high employment protection is 10 percentage points when government effectiveness is high, translating into approximately 4 additional months to close the employment gap in half. - Difference in speed of adjustment parameter between countries with low and high relative minimum wages is 16 percentage points when enforcement is high, implying six additional months to close half the employment gap and restore equilibrium. ### Relationship between speed of adjustment and growth - Moving a country from the 80th percentile of the speed of adjustment distribution to the 20th percentile is associated with an increase in medium-term labor productivity growth of 0.5 percentage points (pp) per annum. - Changing the minimum wage from the 80th to the 20th percentile of the distribution when government effectiveness is high can increase growth by approximately 0.75 pp per annum. - A positive correlation is observed between GDP per worker growth (2000–17) and the implied speed of adjustment coefficient; the speed of adjustment coefficient is statistically significant at the 95 percent confidence level in the reported specification. ### Econometric robustness and identification - Results in Table 6 are robust to inclusion of interaction between GDP per capita and the employment gap, sectoral fixed effects, and sector-time fixed effects. - Identification strategy: labor market regulations are presumed to affect disproportionately labor-intensive sectors; empirical specification estimates sector j’s labor productivity growth gijt with interaction terms including the sector’s labor share αjt and country-level proxies for employment protection legislation or minimum wage over labor productivity. - Table 8 results: higher employment protection legislation (EPL) and higher minimum wages lower labor productivity growth, with larger adverse effects in labor-intensive sectors. The effect of EPL does not consistently affect capital-intensive sectors. ### Labor market regulations, informality, and welfare trade-offs - Stricter enforcement of EPL reduces the speed of adjustment to shocks and lowers labor market flexibility. - Increasing enforcement of EPL is effective in reducing informality on the extensive margin but may increase unemployment and reduce welfare if many workers lack skills to be hired by formal firms (Ulyssea, 2010). Contrasting evidence: Meghir et al. (2015) find tightening enforcement does not increase unemployment and improves wages, output and welfare by reallocating workers to higher productivity jobs. - General micro literature message: lowering entry costs to formality (including for firms) is preferable to increasing costs to informality. - Effects of minimum wages: - High relative minimum wages increase informality and lower the speed of adjustment to shocks. - Reductions in minimum wages could have significant negative impacts on equity and on low-skilled, young and immigrant employment; they must be well-calibrated (Duval and Loungani (2018); Kugler (2019)). - Reductions in payroll taxes contribute to lower informality but are most effective when combined with reductions in entry barriers and hiring/firing costs and reforms that increase total factor productivity (TFP). - Colombia’s experience: a 13 percentage point reduction in payroll taxes lowered informality (case evidence cited). - Policy interactions can produce unintended consequences: some policies supporting small firms can increase informality (Dabla-Norris et al., 2018). Penalties for tax evasion by formal firms can increase informality (Vargas, 2015). ### Informality’s role in labor market dynamics and measurement implications - Informality responds more strongly to GDP cycles than unemployment does. - The formal/informal adjustment margin reduces the importance of the employment/unemployment margin in Okun’s law estimates. - Informality dampens the speed of adjustment of labor markets and negatively affects productivity growth. - In high-informality economies, unemployment rate and job creation alone may be insufficient statistics to assess labor market performance; the informality rate is important to gauge cyclical position. ### Policy recommendations and priorities - Reduce expected cost of firing procedures: make firing procedures more transparent, predictable, and less administratively burdensome to tackle informality and improve labor market functioning (recommendation from Duval and Loungani (2019)). - Build up unemployment insurance and other benefits concurrently to guarantee adequate protection of workers when EPL changes are made. - Embark on structural reforms that increase labor productivity as the best way to address informality stemming from high minimum wages relative to productivity (IMF (2019)). - Combine reductions in payroll taxes with reductions in entry barriers and hiring/firing costs and reforms that increase TFP to achieve larger and more effective reductions in informality. - Tailor minimum wage policy carefully (differentiation across population groups, regions, industries, and business-cycle adjustments) to mitigate adverse employment effects (Duval and Loungani (2019)). ### Box 1 — Informality and Productivity: The Case of Mexico (key findings) - Around 90 percent of firms operate informally in Mexico; they comprise mostly small and relatively less productive firms that employ over half of the labor force. - Formal firms are on average about twice as productive as informal firms in terms of value added per worker; productivity premiums from formality remain after accounting for sector, location and firm size. - Staff model results: - Reducing payroll taxes, social security contributions and firing costs associated with salaried contracts can lead to significant gains in formal employment but only moderate gains in aggregate productivity. - Reducing firm formalization costs can lead to much larger productivity gains because entry costs affect the extensive margin and increase the volume of productive formal firms. - Regulatory distortions explain only a small share of total resource misallocation in Mexico; authorities should also address structural shortcomings preventing development of larger productive firms irrespective of formality. *Source: IMF staff calculations and analysis as presented in the chapter "LABOR MARKET DYNAMICS AND INFORMALITY OVER THE BUSINESS CYCLE IN LAC," International Monetary Fund | October 2019.* ### 1. Extensive and Intensive Margins of Informality ### labormarket - 1. Extensive and Intensive Margins of Informality ### Value Added per Worker Premium in Formal Firms - Regression coefficient on formal sector dummy reported; all coefficients are significant at the 95 percent significance level. - (Figure elements in source indicate formal firms have a positive value-added per worker premium relative to informal firms across firm-size categories; exact plotted values are in the source charts.) ### Shares: Extensive and Intensive Margins - The source distinguishes two margins of informality: - Extensive margin: share of informal firms (by firm size categories shown in figures: <3; 3 to 5; 6 to 50; 51 to 100; 101 to 500). - Intensive margin: share of non-salaried contracts at formal firms (share of non-salaried contracts among formal firms shown in figures). - Figure annotations show policy-control scenarios: No controls; Sector controls; Sector and location controls; Firm size, sector, and location controls (used for robustness checks in regressions). ### Gains from Reducing Informality Distortions (Model Simulations) - Model: small open-economy dynamic stochastic general equilibrium with two sectors (formal and informal); calibrated to Colombian steady state and simulated-method-of-moments used to set persistence and volatility of shocks (TFP, consumer preferences, firing probabilities, and commodity prices). - Policy/structural reform simulation results: - Reduction in payroll tax rate from 29.5 to 16 percent (Colombian 2012 reform) → model predicts a fall in informality of around 3 percentage points; empirical impact cited is about 2 percentage points (Fernandez and Villar, 2017). - Steady-state simulation of a 50-percent reduction in firm entry costs and hiring costs in the formal sector, combined with a further reduction in payroll taxes by 4 percentage points: - Reduces labor informality by 7 percentage points. - Increases GDP per capita by about 3 percent. - Increases private consumption (used as a welfare proxy) by about 3 percent. - Comparison: increasing TFP by 50 percent without changing rigidities: - Also reduces informality by 7 percentage points. - Increases consumption and GDP by almost 30 percent. - Interpretation: - Labor market flexibility matters for informality and macro outcomes, but boosting TFP is the key driver of development; flexibility should complement TFP-enhancing policies. ### Cyclical Properties and Labor Market Rigidities - High labor market rigidities imply a more muted reaction of unemployment to a temporary negative commodity price shock because informality acts as an additional margin of adjustment. - (Source figure “Reaction of Unemployment Rate to Temporary Negative Commodity Price Shock (Percent)” shows unemployment responses over 12 quarters for Baseline and Low frictions scenarios; vertical axis scaled from –0.5 to 2.0 percent.) ### Empirical Estimates: Participation and GDP - Annex Table 1.1 regressions for Active Population as a share of Working Age Population: - ∆ GDP_t: 0.0610** (column 1), 0.0698*** (column 2) - ∆ GDP_t-1: -0.0324 - Constant: 0.142 (col 1), -0.0423 (col 2) - Observations: 439 - Countries: 18 - Note: Standard errors in parentheses. Significance notation: *** p<0.01, ** p<0.05, * p<0.1. ### Methodology and Estimation (Annex 3: Caballero, Cowan, Engel, and Micco (2013) Methodology) - The sectoral representative firm faces isoelastic demand and Cobb-Douglas production in labor and hours per worker; equilibrium conditions yield an employment gap equation and dynamics. - Estimation strategy: - Two-step procedure to estimate parameter φ (phi). First construct instrumented variable zijt = (∆vijt − ∆wijt0)^(1−αj)/γj (approximations explained in source), instrument ∆vijt − ∆v.jt0 with lagged ∆wijt − ∆w.jt0. After φ estimation, construct employment gap and estimate adjustment parameter using Murphy-Topel standard errors to account for first-stage estimation error. - Sector-specific median labor share used to approximate αjγj; manufacturing uses UNIDO, services and construction use OECD STAN. - Log labor productivity (v) = log(output per worker); wijt0 proxied with average labor productivity across countries. - Growth implication: - Using an AK growth model with microeconomic structure, CCEM derive a formula approximating the long-term growth difference between two countries differing only in speed-of-adjustment coefficients (λ), with parameters including depreciation rate δ and a constant ε depending on labor share, productivity volatility, and demand elasticity. ### Data Sources and Variable Construction - Manufacturing sectors: 3-digit UNIDO INDSTAT Rev. 3 (output, employment, wages, capital for 28 sectors). - Construction and services: - European countries: OECD STAN dataset (2005–17). - Non-European countries: Timmer, de Vries and de Vries (2015) 10-sector database (2000–2010). - Employment protection and minimum wage variables from World Bank Doing Business indicators: - Labor protection index = sum of three normalized variables (time to notify dismissals, approval of dismissals by a third party, severance payments). - Minimum wage variable = ratio between national minimum wage and GDP per worker (labor productivity). - Government effectiveness from Kaufmann, Kraay, and Mastruzzi (2010) governance indicators; “High government effectiveness” = dummy equal to one if 1996 estimate above global median. *This summary is based solely on the supplied content.* --- _Source: https://www.imf.org/-/media/files/publications/reo/whd/2019/october/english/labormarket.pdf_