## _wp12218

## Source details

**Canonical URL:** [_wp12218](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12218.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12218.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12218.pdf.json)

---

### Introduction and main contributions
- The Great Recession raised concerns about unemployment and job creation:
  - ILO estimates: number of unemployed increased by "27 million units" over "2007–10".
  - Global employment rates fell from "61.2" to "60 percent".
- Cross-country heterogeneity in unemployment response linked to responsiveness of employment to GDP.
- Main contributions:
  - New set of employment–output elasticities for an unbalanced panel of "167" countries over "1991–2009".
  - Assessment of structural and policy determinants of these elasticities.
- Headline findings:
  - Point estimates of elasticities typically fall in the "0–1 range", majority between "0.3 and 0.8".
  - Elasticities vary across regions, income groups, and sectors; highest in most economically developed regions and in industry and services.
  - Structural policies increasing labor and product market flexibility and reducing government size have significant positive impacts on elasticities.
  - Macroeconomic policies reducing macroeconomic volatility and inflation have significant positive impacts.
  - Structural policies should be complemented with macroeconomic policies that increase macroeconomic stability.

### Empirical strategy and data
- Estimation approaches:
  - Country-by-country time-series regressions (equation (1)/(1a)); co-integration relationship for the majority of countries; OLS estimates described as “super-consistent.”
  - Panel framework with country-specific GDP slopes and employment persistence (equation (2)/(2a)) including country dummies.
  - Instrumental Variables (IV) estimator addressing endogeneity using two lags of real GDP as instruments; overidentifying restriction tests indicate instruments valid (tests not reported).
- Computation of long-term elasticities:
  - Country-specific GDP slope coefficients and lag of employment used to compute long-term elasticities.
  - Robustness checks include alternative specifications (1a) and (2a).
- Cross-country determinants regression (equation (3)):
  - Dependent variable: estimated long-term elasticities averaged over sample period.
  - Regressors grouped as S (structural), M (macroeconomic), D (demographic), X (time-invariant controls).
  - Weighted Least Squares (WLS) used; weights based on estimated standard deviations of residuals.
- Determinants included (averaged over sample period):
  - Structural and policy variables:
    - Labor market policies (composite indicator 0–10; higher = more flexible).
    - Product market policies (composite indicator 0–10; higher = more flexible).
    - Government size (share of government consumption in GDP).
  - Macroeconomic variables: GDP per capita; openness (log of GDP’s share of total exports and imports); CPI inflation; GDP growth volatility (coefficient of variation); country size (log population); FDI inflows (% of GDP); Services’ share of value added.
  - Demographic variables: urban population share; population density; log labor force; working-age population growth.
  - Time-invariant controls: log distance from equator; oil-exporting country dummy; infectious deadly diseases.
- Data sources:
  - World Bank Development Indicators (WDI) for employment and real GDP and many controls.
  - Labor market policy data from Global Competitiveness Report (GCR) and Economic Freedom of the World (EFW).
  - Product market policy data from EFW.

### Distribution, magnitudes, and subgroup patterns
- Overall distribution:
  - Point estimates typically in the "0–1 range"; majority between "0.3 and 0.8".
- Regional (eq.(2a) estimates, Total):
  - East Asia and Pacific: 0.27
  - Western Europe: 0.64
  - Eastern Europe and Central Asia: 0.23
  - Latin America and the Caribbean: 0.16
  - Middle East and North Africa: 0.08
  - South Asia: 0.99
  - Sub-Saharan Africa: -0.02
  - North America: 0.81
  - Australia and New Zealand: 0.80
- Income groups:
  - Low-income economies ($1005 or less): 0.02
  - Lower-middle-income economies ($1006 to $3975): 0.10
  - Upper-middle-income countries ($3976-$12,275): 0.18
  - High-income economies ($12,276 or more): 0.46
  - Employment elasticities for high-income countries are about "three times larger" than those for low-income countries.
- Sectoral patterns (selected):
  - South Asia: Agriculture 1.01; Industry 1.04; Services 0.92
  - North America: Agriculture 1.01; Industry 0.85; Services 0.77
  - Western Europe: Agriculture 0.55; Industry 0.62; Services 0.65
  - Cross-sector comparison: industry and services more employment-intensive than agriculture in upper-middle-income to high-income countries.
- Time evolution:
  - Interaction of log(GDP) with time trend suggests employment intensity of growth has increased over time.
  - Table 4 coefficients (N=166, dependent variable log(employment)):
    - log (GDP): 0.102*** (0.010) in (I); 0.131*** (0.006) in (II); 0.125*** (0.006) in (III).
    - log (GDP)* Time-trend: 0.0006*** (0.0002) in (I).
    - log (GDP)* Dummy 1990s: -0.006** (0.0022) in (II).
    - log (GDP)* Dummy 2000s: 0.006** (0.0022) in (III).
    - Adjusted R2: 0.15 in all columns.

### Structural policies — effects and magnitudes
- Labor market policies:
  - Indicator alone explains more than 25 percent of cross-country variation.
  - Significant and positive across all specifications.
  - Magnitude: increase of indicator equal to one standard deviation of the average change (0.4) increases employment elasticities by about 30 percentage points.
  - Table 6 selected coefficients: 0.747 (6.33)*** in (I); 0.625 (5.49)*** in (III); 0.704 (6.47)*** in (V).
- Product market policies:
  - Significant and positive across specifications.
  - Magnitude: one standard deviation of the average change (0.6) increases elasticities by at least 30 percentage points.
  - Table 6 selected coefficients: 1.249 (5.81)*** in (II); 0.652 (2.90)*** in (IV); 0.885 (4.63)*** in (VI).
- Government size:
  - Explains about 22 percent of cross-country variation; larger government associated with lower elasticities.
  - Magnitude: an increase of 1 percentage point in government consumption in GDP would decrease employment elasticities by about 10 percentage points.
  - Table 6 selected coefficients: -0.166 (-6.16)*** in (III); -0.073 (-1.98)** in (IV); -0.125 (-3.76)*** in (V).
- Component analyses:
  - Labor market components with positive significant effects: Hiring and firing regulations (0.342 (3.81)***), Conscription (0.284 (6.82)***); centralized collective wage bargaining not significant.
  - Product market components with large impacts: price controls 0.707 (4.79)***; bureaucracy costs 0.618 (2.34)***; extra payments/bribes 0.649 (5.85)***; licensing restrictions and cost of tax compliance also significant in some specs.

### Macroeconomic, demographic, and non-linear interactions
- Macroeconomic controls and robustness:
  - Labor market policy effects robust to inclusion of macroeconomic controls; product market effects somewhat sensitive.
  - Growth volatility and inflation negatively associated with employment elasticities.
  - Selected Table 7 coefficients:
    - Growth volatility: -0.712 (-3.82)*** in (IV).
    - Inflation: -0.004 (-2.37)** in (V); -0.006 (-3.48)*** in (VII).
    - Real GDP: 0.677 (3.30)*** in (VI).
    - Services: 0.118 (4.17)*** in (VII).
    - Openness and country size often negative and significant in some specifications.
  - Adjusted R2 range with macro controls: 0.41 to 0.62.
- Demographic determinants (Table 8, selected):
  - Urban population share: positive (0.035 (5.91)***).
  - Population density: negative (-0.603 (-3.33)***).
  - Labor force (log): negative (-0.361 (-3.52)***).
  - Working-age population growth: negative (-0.847 (-4.94)***; -0.914 (-2.80)***).
  - Labor and product market policy results remain robust when demographic controls are included.
- Non-linear interactions:
  - Interaction with level of development:
    - Effect of structural policies on elasticities decreases with higher real GDP per capita (Table 11-A).
    - Labor market policies: 1.979 (2.08)**; interaction Labor market policies* GDP: -0.161 (-1.35)**.
    - Product market policies: 0.337 (1.65)*; Product market policies* GDP: -0.559 (-2.19)**.
  - Interaction with macroeconomic volatility:
    - Lower growth volatility and inflation amplify the effects of structural policies on elasticities (Tables 11-B and 11-C).
    - Examples from Table 11-B: Product market policies: 0.628 (2.71)***; Product market policies* Growth Volatility: -1.967 (-1.53)***.
    - Table 11-C: Labor market policies: 0.869 (8.95)***; Labor market policies* Inflation: -0.004 (-2.51)***.
  - Policy implication: complement structural reforms with macroeconomic policies that increase macroeconomic stability to maximize employment responsiveness.

### Quantitative correlations, model fit, and robustness metrics
- Pairwise and rank correlations:
  - Spearman’s Rank Correlation Coefficients between long-run elasticity estimates: Equation (1) vs (1a): 0.72; (1) vs (2): 1.00; (1) vs (2a): 0.78; (1a) vs (2a): 0.95; (2) vs (2a): 0.88.
- Key correlation statistics (Table 5):
  - Labor market policies: 0.52***.
  - Product market policies: 0.56***.
  - Government size: -0.47***.
- Selected model adjusted R2 ranges:
  - Structural variables only: 0.22 to 0.41.
  - Structural plus macro controls: 0.41 to 0.62.
  - Structural plus demographic controls: 0.46 to 0.58.
  - Non-linear specifications: adjusted R2 up to 0.68 in some columns.

### Key descriptive statistics (selected, preserved exactly)
- Employment growth: N. obs. 2962; Average 2.2; Standard deviation 1.5; Min -1.6; Max 8.7.
- Real GDP growth: N. obs. 2609; Average 4.1; Standard deviation 3.1; Min -4.8; Max 16.9.
- Labor market policies (S): N. obs. 122; Average 5.6; Standard deviation 1.5; Min 2.7; Max 8.5.
- Hiring and firing regulations: 122; Average 4.7; Standard deviation 1.3; Min 2.3; Max 8.0.
- Conscription: 136; Average 5.7; Standard deviation 4.0; Min 0.0; Max 10.0.
- Product market policies: 135; Average 5.7; Standard deviation 1.0; Min 3.6; Max 8.2.
- Government size (% of GDP): 167; Average 11.7; Standard deviation 6.8; Min 3.1; Max 54.1.
- Openness (log): 167; Average 4.2; Standard deviation 0.5; Min 2.7; Max 5.8.
- Country size (log population): 167; Average 8.8; Standard deviation 1.8; Min 4.6; Max 14.0.
- FDI (% of GDP): 194; Average 5.2; Standard deviation 15.2; Min -6.3; Max 170.3.
- Growth volatility (coefficient of variation): 204; Average 0.9; Standard deviation 0.7; Min -0.1; Max 3.7.
- Inflation (%): 160; Average 40.8; Standard deviation 116.1; Min 0.1; Max 971.8.
- Services (% of total value added): 188; Average 52.8; Standard deviation 14.1; Min 14.5; Max 91.0.
- Urban population (share): 167; Average 50.8; Standard deviation 23.5; Min 7.2; Max 100.0.
- Working-age population growth (%): 191; Average 1.8; Standard deviation 1.5; Min -1.0; Max 10.7.
- Distance from equator (hundreds of km): 203; Average 9.5; Standard deviation 15.3; Min 0.0; Max 55.0.

### Conclusions, policy implications, and caveats
- Main conclusions:
  - Point estimates of employment-GDP elasticities typically in 0–1 range; majority between 0.3 and 0.8.
  - Elasticities vary by region, income, and sector; highest in most developed regions and in industry and services.
  - Structural policies that increase labor and product market flexibility and reduce government size significantly raise employment elasticities.
  - Macroeconomic policies that reduce macroeconomic volatility and inflation significantly raise employment elasticities.
  - Complementarity of structural and macroeconomic policies is important: structural reforms yield larger effects when macroeconomic volatility and inflation are low.
- Policy recommendations:
  - Implement structural reforms to increase labor and product market flexibility and consider reductions in government size where appropriate.
  - Pursue macroeconomic policies that reduce growth and price volatility to amplify the effects of structural reforms on employment responsiveness.
  - Design structural policies to internalize social costs and improve quality of employment, noting the analysis does not address job quality.
- Caveats:
  - Results pertain to medium-term elasticities and do not necessarily imply the same effects during crises; short-term effects may differ or be opposite.
  - No inference about quality of jobs created can be drawn from these results.

### Extensions and future research directions
- Potential extensions:
  - Analyze other structural policies such as tax policy, social benefits, and measures of labor market mismatches.
  - Differentiate between labor market inflows and outflows (job creation vs. job destruction) and assess their responsiveness to business cycle phases and the role of structural policies.

*Source: _wp12218 - References .............................................................................................................*

### References .............................................................................................................

### _wp12218 - References .............................................................................................................

### Introduction
- The Great Recession increased concerns about unemployment and job creation:
  - ILO estimates (ILO’s Global Employment Trends): the number of unemployed increased by "27 million units" over the period "2007–10".
  - Global employment rates fell from "61.2" to "60 percent".
- Cross-country heterogeneity in unemployment response is linked to the responsiveness of employment to GDP:
  - Higher employment intensity of growth is associated with larger increases in unemployment in the crisis (Figure 2).
- Importance of understanding determinants of employment elasticities to accelerate job creation.
- Existing literature: many single-country studies and some cross-country or region-specific studies (examples cited in text).
- Limited prior work on determinants of employment–output elasticities; selected findings from previous studies include:
  - Lower real labor costs, greater labor market flexibility, and less exchange rate volatility increase employment elasticities (Döpke 2001).
  - Job intensity of growth highest in services; mixed statistical significance for labor market reforms (Mourre 2004).
  - Employment elasticities positively related to share of services and negatively related to inflation and taxes on labor; no robust relation with employment protection or globalization measures (Kapsos 2005).

- Main contributions and headline findings of this paper:
  - New set of employment–output elasticities for an unbalanced panel of "167" countries over the period "1991–2009".
  - Assessment of structural and policy determinants of these elasticities.
  - Summary findings:
    - Point estimates of elasticities typically fall in the "0–1 range", with the majority ranging between "0.3 and 0.8".
    - Elasticities vary considerably across regions, income groups, and production sectors; highest for most economically developed regions and in industry and services.
    - Structural policies increasing labor and product market flexibility and reducing government size have a significant and positive impact on employment elasticities.
    - Macroeconomic policies reducing macroeconomic volatility have a significant and positive impact on employment elasticities.
    - To maximize responsiveness of employment to economic activity, structural policies should be complemented with macroeconomic policies that increase macroeconomic stability.

### Empirical strategy and data
- Estimation approaches for long-term employment–GDP elasticities:
  - Time-series regressions estimated country-by-country (equation (1)/(1a)), with note that equation (1) is a co-integration relationship for the vast majority of countries; OLS estimates are “super-consistent.”
  - Panel framework using country-specific GDP slopes and employment persistence (equation (2)/(2a)), with country-specific dummies (ܦ
௜).
  - Instrumental Variables (IV) estimator used to address endogeneity of employment with respect to GDP, using two lags of real GDP as instruments. Overidentifying restriction tests (e.g., Wooldrige score test) indicate instruments are valid (tests not reported).
- Computation of long-term elasticities:
  - Country-specific coefficients for GDP slopes and lag of employment are used to compute long-term elasticities (expression shown as ሺߩሺ1െߚ⁄ሻሻ in text).
  - Robustness: alternative specifications (1a) and (2a) estimated as robustness checks; reference to Kapsos (2005) for similar approach.
- Cross-country determinants regression (equation (3)):
  - Dependent variable: estimated long-term elasticities.
  - Regressors grouped as S (structural variables), M (macroeconomic variables), D (demographic variables), X (time-invariant controls).
  - All regressors averaged over the sample period.
  - Weighted Least Squares (WLS) used to account for differing precision of estimated dependent variables; WLS assumes errors distributed as ߳
௜ܰ~(ߪ,0ଶݏ
௜⁄) where ݏ
௜ are estimated standard deviations of residuals ߱ሺ௜ሻ.
- Determinants included (as specified in text):
  - Structural and policy variables:
    - Labor market policies (composite indicator based on Hiring and firing regulation (GCR); Centralized collective wage bargaining (GCR); Conscription (EFW)), constructed following EFW methodology, rating countries between "0 and 10" (higher = more flexible).
    - Product market policies (composite indicator from EFW based on seven indicators: Price controls; Administrative requirements; Bureaucracy costs; Cost of starting a business; Extra payments/bribes; Licensing restrictions; Cost of tax compliance), rating "0 and 10" (higher = more flexible).
    - Government size (theoretical channels: crowding out private investment, smaller private sector absorption of labor, higher taxes).
  - Macroeconomic variables (M): a) GDP per capita; b) openness (log of GDP’s share of total exports and imports); c) CPI-based inflation rate; d) GDP growth volatility (coefficient of variation of real GDP growth); e) country size (log of total population); f) FDI inflows as a percentage of GDP; g) share of Services’ value added in total GDP.
  - Demographic variables (D): a) share of urban population; b) population density; c) log of total labor force; d) working-age population growth.
  - Additional time-invariant controls (X): a) latitude distance from the equator (log of absolute distance); b) oil-exporting country dummy; infectious deadly diseases also included to control for productivity/employment effects.
- Data sources:
  - Most variables, including employment and real GDP, taken from the World Bank Development Indicators (WDI) database (Table 1).
  - Labor market policy data from the Global Competitiveness Report (GCR) and the Fraser Institute’s Economic Freedom of the World (EFW).
  - Product market policy data from the Fraser Institute’s Economic Freedom of the World (EFW).

### Results
- Estimation methods and robustness:
  - Long-term elasticities estimated using both the time-series and panel approaches (equation (1–1a) and (2–2a)); point estimates vary slightly across methods but country rank is robust (Table 2).
- Distribution and magnitudes:
  - Point estimates typically fall in the "0–1 range", with the majority of elasticities ranging between "0.3 and 0.8" (Figure 3).
- Regional, income, and sectoral variation:
  - Considerable variation across regions, income groups, and production sectors.
  - Panel estimates with region- or income-level-specific dummies for GDP slopes and employment lag show:
    - Highest estimates for South Asia, followed by the most economically developed regions, including Australia and New Zealand, North America, and Western Europe.
    - Modest elasticities in Middle East and North Africa: "0.1 percent".
    - Eastern Europe: "0.23 percent".
    - Sub-Saharan Africa: negative estimates found, "–0.02 percent".
    - Employment elasticities for high-income countries are about "three times larger" than those for low-income countries.
- Policy-relevant findings:
  - Structural policies that increase labor and product market flexibility and reduce government size have a significant and positive impact on employment elasticities.
  - Macroeconomic stability (lower growth volatility and lower inflation) has a significant and positive impact on employment elasticities.
  - Complementarity: structural policies should be complemented with macroeconomic policies aimed at increasing macroeconomic stability to maximize employment responsiveness to economic activity.

*Source: _wp12218 - References .............................................................................................................*

### 24.      Table 3 also presents estimates of employment elasticities for three main economic

### _wp12218 - 24.      Table 3 also presents estimates of employment elasticities for three main economic

### Employment elasticities: patterns by sector, region, income, and time
- Table 3 presents estimates of employment elasticities for three main economic sectors—agriculture, industry, and services—by region and income group, using value added by economic sector from the WDI economic database.
- Cross-sector comparison:
  - More employment-intensive growth in industry and services than in agriculture in upper-middle-income to high-income countries, where industry and services make a significant contribution to growth.
- Cross-region comparison:
  - Higher elasticities in more advanced economies across each economic sector.
- Evolution over time:
  - Including an interaction term between the log of GDP and a time trend in equation (2) suggests the employment intensity of growth has increased over time (Table 4).
  - Similar results are obtained when dummies for the time periods 1991–2000 and 2001–10 interact with the log of GDP.
- Statistical testing:
  - Estimated elasticities are statistically different from each other for almost all regions and all income groups, as confirmed by pairwise Wald tests (results in Appendix B).

### Determinants of employment elasticities — overview of approach
- Estimation strategy:
  - Econometric results from estimating equation (3) under different specifications and sets of controls.
  - Results presented are based on long-term elasticities from country-specific time series regressions; qualitatively similar results obtained using panel regressions.
- Focus:
  - Role of structural and policy variables in affecting employment-intensity of growth.
  - Table 5 presents pairwise correlations—corrected for different precision in the estimated elasticities—between estimated medium-term elasticities and structural and policy variables.
- Preliminary correlation finding:
  - Countries with higher elasticities are characterized by more flexible labor and product markets and by lower government size.

### Structural policies: labor and product market indicators, government size
- Labor market indicator:
  - The indicator by itself explains more than 25 percent of the cross-country variation in employment elasticities.
  - The labor market indicator has a significant and positive impact across all specifications.
  - Magnitude: an increase of the indicator equal to one standard deviation of the average change (0.4) increases employment elasticities by about 30 percentage points.
- Product market indicator:
  - Significant and positive impact on employment elasticities across all specifications and sets of controls.
  - The indicator by itself explains a significant share of cross-country variation.
  - Magnitude: an increase in the indicator equal to one standard deviation of the average change (0.6) increases employment elasticities by at least 30 percentage points.
- Government size:
  - Explains about 22 percent of cross-country variation in employment elasticities.
  - Larger government associated with lower employment elasticities.
  - Magnitude: an increase of 1 percentage point in the share of government consumption in GDP would decrease employment elasticities by about 10 percentage points.

### Macroeconomic policies and other macro controls
- Robustness:
  - Effects of labor market policies, and to a lesser extent product market policies, are robust to inclusion of macroeconomic controls (Table 7).
- Macroeconomic volatility:
  - Policies aimed at reducing macroeconomic (output and price) volatility significantly increase employment elasticities.
- Development and trade:
  - Different levels of economic development (proxied by real GDP per capita) and trade openness help explain cross-country variations.
  - Employment elasticities tend to be higher in more advanced and closed economies.
- Sectoral composition:
  - Employment intensity of growth tends to be higher in countries with a larger service sector.

### Demographic determinants
- Table 8 results indicate employment elasticities vary with demographic characteristics:
  - Higher share of urban population → larger employment elasticities.
  - Higher population density, labor force growth, and working-age population growth → negatively correlated with employment-output elasticities.
- Robustness:
  - Results for labor market policies, and to a lesser extent product market policies, remain robust when demographic controls are included.

### Labor market indicators — component analysis
- Re-estimation of equation (3) with components of the composite labor market indicator shows:
  - Statistically significant and positive components:
    - (i) Hiring and firing regulations;
    - (ii) Conscription (Table 9).
  - No evidence of a statistically significant relationship between employment elasticities and the indicator of centralized collective wage bargaining.
- Consistency:
  - Results align with Bernal-Verdugo et al. (2012a and b) showing these indicators significantly impact unemployment outcomes.
- Robustness:
  - Product market regulation and government size results generally robust to inclusion of different labor market indicators.

### Product market indicators — component analysis
- Estimation using components of product market policies (Table 10) shows:
  - Most product market policy indicators have a positive and statistically significant effect on employment elasticities.
  - Indicators with largest and most significant impact: price controls, bureaucracy costs, and extra payments/bribes.
- Interaction with government size:
  - Labor market policies and government size remain robust across specifications.
  - Exception: government size lacks significance when price controls and extra payments/bribes are included, driven by strong negative correlation—countries with larger governments tend to have higher price controls and extra payments/bribes.

### Non-linear effects: interaction with development and volatility
- Interaction with level of development:
  - Including interaction terms between real GDP per capita and structural variables shows the effect of structural policies on employment elasticities is a decreasing function of the level of economic development (Table 11.A).
  - Implication: policies increasing labor and product market flexibility and reducing government size have larger effects in less developed countries.
  - Inclusion of interaction terms significantly increases the regression’s fitting power.
- Interaction with macroeconomic volatility:
  - Growth volatility and inflation shape the response of employment elasticities to structural variables (Table 11.B and Table 11.C).
  - The lower the economic and price volatility, the larger the effects of structural policies on employment elasticities.
  - Policy implication: to maximize structural reforms’ effects on employment responsiveness, complement structural reforms with macroeconomic policies aimed at increasing macroeconomic stability.

### Conclusions — key results and policy implications
- Two contributions:
  - Provides a new set of employment-GDP elasticities for a large set of advanced and developing economies.
  - Assesses determinants of cross-country variations in employment elasticities, focusing on structural variables.
- Key results summarized:
  - Point estimates of elasticities typically fall in the 0–1 range, with the majority ranging between 0.3 and 0.8.
  - Elasticities vary considerably across regions, income groups, and production sectors; highest estimates typically in the most economically developed regions, and in industry and services sectors.
  - Structural policies that increase labor and product market flexibility and reduce government size have a significant and positive impact on employment elasticities.
  - Macroeconomic policies that reduce macroeconomic volatility have a significant and positive impact on employment elasticities.
  - To maximize the positive impact on employment responsiveness, structural policies should be complemented by macroeconomic policies that increase macroeconomic stability.
- Caveats and limitations:
  - Results refer to medium-term elasticities; they do not necessarily imply the same effects during periods of crisis and may suggest the opposite in the short term.
  - The analysis does not address the quality of jobs created; no inference can be drawn about job quality from these results.
  - Policy makers should design structural policies to internalize social costs and improve quality of employment.

*Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12218.pdf*

### 47.      Building on our study, a number of interesting extensions can be pursued. First,

### Building on our study, a number of interesting extensions can be pursued. First,

### Extensions and future research directions
- Several other types of structural policies, beyond those analyzed in the paper, could in principle affect employment-GDP elasticities: tax policy, social benefits and measures of labor market mismatches are natural candidates.
- Differentiate between labor market inflows and outflows (job creation vs. job destruction) and assess their respective responsiveness to upturns and downturns in economic activity, and the role that structural policies may have in influencing their responsiveness.

### Key descriptive statistics (selected, preserved exactly as in source)
- Employment growth: N. obs. 2962; Average 2.2; Standard deviation 1.5; Min -1.6; Max 8.7. (Source: WDI)
- Real GDP growth: N. obs. 2609; Average 4.1; Standard deviation 3.1; Min -4.8; Max 16.9. (Source: WDI)
- Labor market policies (S): N. obs. 122; Average 5.6; Standard deviation 1.5; Min 2.7; Max 8.5. (Source: A)
- Hiring and firing regulations: 122; Average 4.7; Standard deviation 1.3; Min 2.3; Max 8.0. (Source: GCR)
- Centralized collective bargaining: 122; Average 6.5; Standard deviation 1.3; Min 3.0; Max 8.6. (Source: GCR)
- Conscription: 136; Average 5.7; Standard deviation 4.0; Min 0.0; Max 10.0. (Source: EFW)
- Product market policies: 135; Average 5.7; Standard deviation 1.0; Min 3.6; Max 8.2. (Source: EFW)
- Price controls: 123; Average 4.8; Standard deviation 1.9; Min 0.0; Max 9.5. (Source: EFW)
- Administrative requirements: 122; Average 3.7; Standard deviation 1.0; Min 1.8; Max 7.2. (Source: EFW)
- Bureaucracy costs: 122; Average 5.3; Standard deviation 0.8; Min 0.5; Max 7.0. (Source: EFW)
- Starting a business: 133; Average 7.5; Standard deviation 1.5; Min 0.2; Max 9.6. (Source: EFW)
- Extra payment/Bribes: 122; Average 5.7; Standard deviation 1.8; Min 2.3; Max 9.6. (Source: EFW)
- Licensing restrictions: 133; Average 6.6; Standard deviation 2.1; Min 0.0; Max 10.0. (Source: EFW)
- Cost of tax compliance: 133; Average 6.5; Standard deviation 2.4; Min 0.0; Max 9.9. (Source: EFW)
- Government size (% of GDP): 167; Average 11.7; Standard deviation 6.8; Min 3.1; Max 54.1. (Source: WDI)
- Oil dummy: 231; Average 0.1; Standard deviation 0.3; Min 0.0; Max 1.0. (Source: Authors)
- Distance from equator (hundreds of km): 203; Average 9.5; Standard deviation 15.3; Min 0.0; Max 55.0. (Source: CIA)

- Macroeconomic (selected):
  - Openness (log): 167; Average 4.2; Standard deviation 0.5; Min 2.7; Max 5.8. (WDI)
  - Country size (log population): 167; Average 8.8; Standard deviation 1.8; Min 4.6; Max 14.0. (WDI)
  - FDI (% of GDP): 194; Average 5.2; Standard deviation 15.2; Min -6.3; Max 170.3. (WDI)
  - Growth volatility (coefficient of variation): 204; Average 0.9; Standard deviation 0.7; Min -0.1; Max 3.7. (WDI)
  - Inflation (%): 160; Average 40.8; Standard deviation 116.1; Min 0.1; Max 971.8. (WDI)
  - Real GDP (log): 167; Average 8.5; Standard deviation 1.3; Min 5.4; Max 11.2. (WDI)
  - Services (% of total value added): 188; Average 52.8; Standard deviation 14.1; Min 14.5; Max 91.0. (WDI)
- Demographic (selected):
  - Urban population (share of total population): 167; Average 50.8; Standard deviation 23.5; Min 7.2; Max 100.0. (WDI)
  - Population density (log): 167; Average 3.9; Standard deviation 1.4; Min 0.4; Max 8.6. (WDI)
  - Labor force (log): 166; Average 14.9; Standard deviation 1.8; Min 10.9; Max 20.3. (WDI)
  - Working-age population growth (%): 191; Average 1.8; Standard deviation 1.5; Min -1.0; Max 10.7. (WDI)

### Correlations and distributions
- Spearman’s Rank Correlation Coefficient for Long-run Elasticities Estimates:
  - Equation (1) vs Equation (1a): 0.72
  - Equation (1) vs Equation (2): 1.00
  - Equation (1) vs Equation (2a): 0.78
  - Equation (1a) vs Equation (2a): 0.95
  - Equation (2) vs Equation (2a): 0.88

- Distribution of Long-Run Employment Elasticities: plotted in the source (range shown from -0.4 to 1.2 on x-axis; frequency on y-axis).

### Long-term employment elasticities (selected values, eq.(2a) estimates)
- By region (Total):
  - A. East Asia and Pacific: 0.27
  - B. Western Europe: 0.64
  - C. Eastern Europe and Central Asia: 0.23
  - D. Latin America and the Caribbean: 0.16
  - E. Middle East and North Africa: 0.08
  - F. South Asia: 0.99
  - G. Sub-Saharan Africa: -0.02
  - H. North America: 0.81
  - I. Australia and New Zealand: 0.80
- By income:
  - J. Low-income economies ($1005 or less): 0.02
  - K. Lower-middle-income economies ($1006 to $3975): 0.10
  - L. Upper-middle-income countries ($3976-$12,275): 0.18
  - M. High-income economies ($12,276 or more): 0.46
- Sectoral elasticities (selected, based on restricted sample where available):
  - South Asia: Agriculture 1.01; Industry 1.04; Services 0.92
  - North America: Agriculture 1.01; Industry 0.85; Services 0.77
  - Western Europe: Agriculture 0.55; Industry 0.62; Services 0.65

### Time evolution
- Table 4 (dependent variable log(employment), N=166):
  - log (GDP): 0.102*** (0.010) in (I); 0.131*** (0.006) in (II); 0.125*** (0.006) in (III).
  - log (GDP)* Time-trend: 0.0006*** (0.0002) in (I).
  - log (GDP)* Dummy 1990s: -0.006** (0.0022) in (II).
  - log (GDP)* Dummy 2000s: 0.006** (0.0022) in (III).
  - Adjusted R2: 0.15 in all columns.

### Structural policy correlations and effects (selected)
- Table 5: Correlation between employment elasticities and structural policy variables:
  - Labor market policies: 0.52***.
  - Product market policies: 0.56***.
  - Government size: -0.47***.
  - Note: *** denotes significance at 1 percent.

- Table 6 (Effects of structural variables on employment elasticities; selected coefficients and significance):
  - Labor market policies: 0.747 (6.33)*** in (I); 0.625 (5.49)*** in (III); 0.704 (6.47)*** in (V).
  - Product market policies: 1.249 (5.81)*** in (II); 0.652 (2.90)*** in (IV); 0.885 (4.63)*** in (VI).
  - Government size: -0.166 (-6.16)*** in (III); -0.073 (-1.98)** in (IV); -0.125 (-3.76)*** in (V).
  - Distance from equator: -0.051 (-2.30)** in (V).
  - Sample sizes: N = 111, 117, 134, 111, 111, 111, 117, 111 across specifications.
  - Adjusted R2 range reported: 0.22 to 0.41 depending on specification.

- Table 7 (Structural and macroeconomic controls; selected):
  - Labor market policies: coefficients range from 0.372 (2.89)*** to 0.773 (6.51)*** across specifications.
  - Product market policies: 0.949 (3.98)*** in (I); becomes insignificant or smaller in some specifications (e.g., 0.250 (1.08) in II).
  - Government size: -0.082 (-2.20)** in (I); other columns show mixed significance.
  - Openness: -1.335 (-2.64)*** in one specification; -1.761 (-3.17)*** in another.
  - Country size: -0.440 (-4.24)*** in (II); -0.666 (-2.67)*** in (VIII).
  - Growth volatility: -0.712 (-3.82)*** in (IV).
  - Inflation: -0.004 (-2.37)** in (V); -0.006 (-3.48)*** in (VII).
  - Real GDP: 0.677 (3.30)*** in (VI).
  - Services: 0.118 (4.17)*** in (VII).
  - Adjusted R2 range: 0.41 to 0.62.

- Table 8 (Structural and demographic controls; selected):
  - Labor market policies: 0.880 (7.51)*** in (I); 0.912 (7.65)*** in (IV).
  - Product market policies: significant in some specs (0.484 (2.19)** in II; 0.446 (2.15)** in IV).
  - Urban population: 0.035 (5.91)*** in (I).
  - Population density: -0.603 (-3.33)*** in (II).
  - Labor force: -0.361 (-3.52)*** in (III).
  - Working-age population growth: -0.847 (-4.94)*** in (IV); -0.914 (-2.80)*** in (V).
  - Adjusted R2 range: 0.46 to 0.58.

### Specific policy indicators (selected)
- Table 9 (Labor market policy indicators):
  - Hiring and firing regulations: 0.342 (3.81)*** in (I).
  - Conscription: 0.284 (6.82)*** in (II).
  - Product market policies: 0.528 (2.58)** in (I); 0.702 (3.59)*** in (II).
  - Adjusted R2: 0.31 in (I); 0.48 in (II).

- Table 10 (Product market policy indicators):
  - Price controls: 0.707 (4.79)*** in (I).
  - Bureaucracy costs: 0.618 (2.34)*** in (II).
  - Extra payment/Bribes/Favoritism: 0.649 (5.85)*** in (III).
  - Licensing restrictions: 0.240 (2.35)** in (IV).
  - Cost of tax compliance: 0.259 (3.71)*** in (V).
  - Labor market policies remain significant across these specifications (e.g., 0.902 (7.70)*** in (III)).
  - Government size shows negative coefficients in several specs (e.g., -0.099 (-2.97)*** in II).
  - Adjusted R2 range: 0.39 to 0.51.

### Non-linear effects (selected)
- Table 11-A (Non-linear effects on GDP):
  - Labor market policies: 1.979 (2.08)** in (I); interaction Labor market policies* GDP: -0.161 (-1.35)**.
  - Product market policies: 0.337 (1.65)* in (I); Product market policies* GDP: -0.559 (-2.19)** in (II).
  - Government size: -1.257 (-3.66)*** in (III).
  - Adjusted R2: 0.58–0.59 across columns.

- Table 11-B (Non-linear effects on Growth Volatility):
  - Labor market policies: 0.569 (1.96)* in (I); interaction Labor market policies* Growth Volatility: -0.129 (-0.76).
  - Product market policies: 0.628 (2.71)*** in (I); Product market policies* Growth Volatility: -1.967 (-1.53)*** in (II).
  - Government size: -0.194 (-3.71)*** in (III); Government size* Growth Volatility: 0.153 (4.71)***.
  - Adjusted R2: 0.47, 0.68, 0.56 across columns.

- Table 11-C (Non-linear effects on Inflation):
  - Labor market policies: 0.869 (8.95)*** in (I); Labor market policies* Inflation: -0.004 (-2.51)*** in (I).
  - Product market policies: 0.225 (1.20) in (I).
  - Government size: coefficients small and generally insignificant; Government size* Inflation: 0.001 (1.96)* in (III).
  - Adjusted R2: 0.63–0.64.

### Notable empirical patterns and interpretation
- Labor market policies and product market policies show consistently positive and often highly significant associations with higher employment-GDP elasticities in multiple specifications.
- Government size tends to be negatively associated with employment elasticities in many specifications (e.g., -0.166 (-6.16)***), though significance and sign vary with controls and non-linear interactions.
- Macroeconomic variables such as openness, country size, growth volatility, inflation, and real GDP enter significantly in some specifications and alter the magnitudes of structural-policy coefficients.
- Demographic factors (urban population, population density, labor force size, working-age population growth) significantly affect employment elasticities and can strengthen or weaken the impact of structural policies.

### Sample coverage
- Country sample includes a broad set across income groups and regions (Low-Income; Lower-Middle-Income; Upper-Middle-Income; High-Income), covering regions such as East Asia and Pacific, Eastern Europe and Central Asia, Latin America and the Caribbean, Middle East and North Africa, North America, South Asia, Sub-Saharan Africa, Western Europe, Australia and New Zealand.
- Appendix Table A1 lists countries in the sample by region and income classification (as presented in the source).

*International Monetary Fund Working Paper content (as provided in the source excerpt).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2012/_wp12218.pdf_