## wpiea2019144 - 2005. It is likely that part of the variation in

## Source details

**Canonical URL:** [wpiea2019144 - 2005. It is likely that part of the variation in](https://www.imf.org/-/media/files/publications/wp/2019/wpiea2019144.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2019/wpiea2019144.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2019/wpiea2019144.pdf.json)

---

### Unemployment definitions and measurement
- Two definitions:
  - “Relaxed” definition: individual is unemployed if not employed regardless of search status if willing to accept a job offer.
  - “Strict” definition: individual is unemployed only if not employed and actively searching for a job.
- Several ECCU countries adopted the “relaxed” definition following the International Labor Organization’s recommendation.
- Effects of definition choice:
  - Under the “relaxed” definition the unemployment rate is about 5 percentage points higher than under the “strict” definition in ATG, GRD and LCA.
  - Under the “relaxed” definition unemployment reaches almost 30 percent in GRD.

### Structural and cyclical components of unemployment
- Long-run structural component:
  - Several decades of high unemployment imply an important structural component in most ECCU countries.
  - LCA annual series: unemployment stood at 15 percent in 1994 and stood above 14 percent at the peak of the 2000s tourism-fueled boom and construction for the Cricket World Cup in 2007.
  - HP-Filter suggests structural unemployment in LCA has remained around 20 percent for more than two decades.
  - Structural unemployment trend reflects a large and long-lasting increase following the Great Recession.
- Cyclical component and the global crisis:
  - Global crisis was followed by a substantial increase in unemployment in LCA, adding a significant cyclical component to an already high structural rate.
  - LCA unemployment peaked in 2015, several years after the start of the global crisis, then declined as output approached potential but remained above pre-crisis levels for many years.
  - Unemployment was much higher than before the crisis also in ATG and VCT.
  - GRD: unemployment slightly decreased since 2008, likely a reversal from the major increase that followed two natural disasters and the global financial crisis.
  - KNA experienced a reduction in unemployment following the global crisis, partly due to a major public employment program, by far the largest among ECCU countries as a share of total employment.
  - DMA: no immediate pre-crisis statistics; unemployment in 2009 (latest available) was higher than its 2001 level.
- Quantitative assessment using Okun’s Law (Kandil and others (2014) elasticity):
  - Estimated elasticity of unemployment to the output gap: -0.32.
  - Output gap in LCA increased by 16 percent between 2008 and 2014.
  - Inference: the slowdown in economic activity should have contributed 5 percentage points to LCA’s unemployment rate between 2008 and 2014, about two thirds of its actual increase between those years.

### Youth and gender disparities
- Youth unemployment:
  - Youth unemployment is considerably higher than the overall unemployment rate across ECCU countries.
  - Ratio of youth-to-overall unemployment rates: in all ECCU countries except ATG, this ratio is below its average in most other regions.
  - Actual youth unemployment rates: almost half of the youth population unemployed in GRD, LCA, and VCT.
  - Implication: major social problem potentially linked to high crime rates in those countries.
- Gender gap:
  - Female-to-male unemployment ratios in ECCU are considerably above averages in more gender-egalitarian developing regions; closer to averages in regions where gender discrimination is common.
  - High unemployment ratios translate into large absolute gaps.
  - Example: since 1994, St. Lucia’s female unemployment has been on average 6.6 percentage points higher than male unemployment.
  - The gender gap narrowed significantly following the global crisis due to a faster run-up of male unemployment but widened soon afterwards.
  - Grenada has had a similar sized gender gap in recent years.

### Key statistics and indicators (as reported)
- “Relaxed” vs “strict” difference: about 5 percentage points higher in ATG, GRD and LCA.
- GRD unemployment (relaxed definition): almost 30 percent.
- LCA unemployment: 15 percent in 1994; above 14 percent at peak of 2000s boom.
- Structural unemployment in LCA (HP-Filter): around 20 percent for more than two decades.
- Okun’s Law elasticity used: -0.32.
- LCA output gap increase (2008–2014): 16 percent.
- Implied contribution of output slowdown to LCA unemployment (2008–2014): 5 percentage points (about two thirds of the actual increase).
- Youth unemployment: almost half of youth population unemployed in GRD, LCA, and VCT.
- St. Lucia female-minus-male unemployment gap since 1994: 6.6 percentage points on average.

### Global financial crisis and ECCU labor demand — key findings
- The global downturn exacerbated unemployment in ECCU countries; after sizable expansion during the 2000s and up to 2008, real output either stagnated or contracted in all ECCU countries.
- By 2015, only DMA, GRE, and KNA had real output recovered to pre-crisis levels; ATG and LCA remained considerably below pre-crisis levels.
- Employment effects by sector since 2008:
  - Formal employment remained sluggish or contracted in all ECCU countries since 2008.
  - Particularly marked contractions in Hotels and Restaurants and Construction in ATG and LCA.
  - Public administration employment increased in ATG and VCT as governments increased public employment (including temporary employment) in response to the crisis.
- Empirical support: Kandil and others (2014) found a significant relation between output fluctuations and employment in several Caribbean countries, including LCA.

### Sectoral shocks and natural disasters
- Banana/sugar decline and trade preferences:
  - Banana industry accounted for approximately 5 percent of GDP in 1979-1994 in DMA, LCA, and VCT and sugar in KNA; these contracted to less than 2 percent of GDP in 2005.
  - This constituted an annual decline of 0.3 percentage points of GDP on average over 1996-2005.
  - Affected economies grew by an annual average of 2.7 percent over 1996-2005.
  - Conclusion: timing and GDP growth imply banana/sugar decline unlikely to explain very high unemployment rates in the 2000s.
- Natural disasters:
  - Major immediate employment impacts, especially on agriculture.
  - Grenada: Hurricane Ivan (2004) led to overall formal employment contraction of 12 percent.
    - Agricultural output contracted by 40 percent between 2003 and 2005.
    - Formal agriculture employment contracted by 20 percent.
    - Agriculture accounted for 5.3 percent of total formal employment prior to the storms (in 2003); the formal employment contraction in agriculture directly responsible for a 2.3 percent reduction in total formal employment.
  - 2017 Category 5 hurricane hit Dominica with major negative employment impact, especially in agriculture and tourism.

### Labor supply: remittances, household effects, and skills mismatch
- Remittances and participation:
  - Hypothesis: high remittances could raise reservation wages and reduce labor supply.
  - Cross-country glance: remittances as percent of GDP notably high only in DMA among ECCU; remittances much lower than in Jamaica (17 percent).
  - Probit analysis for GRD 2013–2015:
    - Dependent variable: labor force participation dummy.
    - Controls: age, gender, education, wealth, remittances dummy, household employment in tourism/public sector.
    - Results: remittances not a significant factor explaining unemployment in GRD.
    - For youth, higher remittances associated with higher likelihood of participating in the labor force—possible use of remittances as business capital.
- Effect of having a well-paid household member:
  - Rapid growth in tourism and public sectors produced better paid urban employment.
  - Probit regressions indicate household head employed in tourism and public sectors associated with lower labor force participation.
  - Wage differentials: in GRD average salary in 2014 in tourism and public sectors was 66 and 93 percent higher than in formal agriculture; in LCA in 2013 tourism and public sector salaries were 50 percent and 63 percent higher than formal agriculture.
- Skills mismatch:
  - Employer surveys report skill deficiencies.
  - St. Lucia 2012 survey:
    - Job seekers with below secondary education: 60 percent.
    - Job openings requiring secondary or higher education: 75 percent.
    - Job seekers with tertiary education: 7 percent.
    - Job offers requiring tertiary education: 40 percent.
    - Absolute numbers: openings requiring tertiary education (357) vs job seekers with tertiary education (1,547).
    - Filling all tertiary-requiring openings would reduce unemployment by 0.37 percentage points in St. Lucia.
  - Caveats: employer sampling biased against low-skilled employers; evidence does not support broad educational-category mismatch as main driver of very high unemployment.

### Structural rigidities, unit labor costs (ULCs), and unions
- High ULCs and unemployment:
  - ECCU wages higher than peers: World Bank (2005) for 2002 found wages for unskilled and semi-skilled labor about 25 percent higher in ECCU than other upper-middle income countries; wages for professionals about 45 percent higher.
  - Post-global crisis: average wages in ECCU continued to increase despite substantial decline in productivity, leading to significant increases in ULCs.
  - Tourism sector wages: ratio of average wage in Hotels and Restaurants to GDP per capita considerably higher in ECCU than in Bahamas, Dominican Republic, and Mexico.
  - Minimum wage indicators: minimum wage-to-GDP per capita ratio elevated relative to “competitive middle-income” countries (top 50 Ease of Doing Business). St. Lucia does not have an official minimum wage.
  - Example wage levels (2002, selected): Unskilled hourly: ECCU 2.3, Upper-Middle Income 1.8, Microstates 1.9; Professional annual: ECCU 22.3, Upper-Middle Income 15.5, Microstates 19.0.
- ULCs correlated with unemployment:
  - Approximation: ULCs = average wage in social security statistics / GDP per capita.
  - Back-of-the-envelope slope estimates:
    - 10 percent higher ULCs associated with higher structural unemployment of 3.4 percent in the early 2000s.
    - 10 percent higher ULCs associated with higher structural unemployment of 5.1 percent in the mid-2010s.
  - Example: VCT with ULCs more than 30 percent above ATG and KNA is expected to have about 11 to 16 percentage points higher unemployment (based on slope estimates).
  - Caution: small sample size and methodological limitations; results suggestive.
- Unions and CBAs:
  - Trade union density not significantly different on average in ECCU relative to other regions.
  - Collective Bargaining Agreements affect more than 80 percent of labor force on average in ECCU.
  - Unions cover large, strategic sectors; many unions historically aligned with major political parties.
  - Labor regulatory framework not particularly rigid on hiring/firing: dismissal without third-party approval is allowed; no prohibitions to fixed-term contracts for permanent tasks; probation periods commonly allowed (up to 6 years in DMA and VCT).
- Public sector wage demonstration effects:
  - Social security data and dynamic panel regressions (ATG, GRD, LCA, VCT) show private sector wages strongly related to lagged public-sector wage.
  - VAR and Granger causality tests on Grenada find bi-directional causality between public and private wages.
  - Implication: public sector wage increases can rapidly transmit to private sector wages, raising economy-wide ULCs and affecting private sector profitability and job creation.
  - System GMM lagged public-sector wage coefficient reported: 0.820***.

### Conclusions and policy-oriented findings
- Unemployment in most ECCU countries reflects both temporary (natural disasters, global financial crisis) and structural components; persistence of very high unemployment in DMA, GRD, LCA, and VCT implies enduring structural factors.
- High ULCs are likely an important factor behind high structural unemployment; moderation of wage growth may be prudent to reduce unemployment.
- Contributing factors to high ULCs and wage inflation:
  - Strong union activity with CBAs covering over 80 percent of the labor force.
  - Public sector wage “demonstration effects” transmitting public wage increases into private sector wages.
- Labor supply factors with limited support:
  - Remittances: not statistically significant in reducing labor force participation in GRD; for youth remittances correlate with higher participation.
  - Skills mismatch at broad educational categories: filling all tertiary-required openings in St. Lucia would reduce unemployment by only 0.37 percentage points.
- Sectoral shocks:
  - Banana/sugar decline contributed to sectoral employment shifts but unlikely to explain very high unemployment in the 2000s given overall growth 1996-2005.
  - Natural disasters have immediate and large negative employment impacts (e.g., Grenada: 12 percent overall formal employment contraction; agriculture output -40 percent between 2003 and 2005; formal agriculture employment -20 percent).
- Policy-relevant empirical relationships:
  - Positive correlation between approximated ULC and unemployment: early 2000s slope implies 10 percent ULC ↑ → 3.4 percent unemployment ↑; mid-2010s slope implies 10 percent ULC ↑ → 5.1 percent unemployment ↑.
  - Strong lagged relation: Lagged public sector wage → current private sector real wage (System GMM lagged public coefficient 0.820***).

### Policy recommendations (numbered items from source)
- 1. Implement “flexicure” labor market arrangements that combine flexible labor rules with unemployment safety nets that include training (see annex 1).
  - If reducing collective bargaining coverage is not possible, promote wage flexibility by establishing social partnership among governments, firms and workers to incorporate macroeconomic considerations into wage-setting discussions.
- 2. Control public sector wage growth by linking public sector wages to productivity and involving unions more closely during the budget process to inform them about fiscal budget constraints.
- 3. Improve the effectiveness of countercyclical policies to minimize cyclical unemployment; without a flexible exchange rate, countries need to build fiscal space to undertake countercyclical fiscal policies, though most ECCU countries are in a weak fiscal situation with high public debt.
- 4. Build resilience to natural disasters by adopting a fiscal framework that finances prevention, relief and reconstruction efforts; developing the capacity to design and implement such a framework deserves priority attention.
- 5. Foster productivity and human capital growth to lower ULCs and thus promote structural employment by:
  - increasing economies of scale through measures to increase international and intraregional integration;
  - improving economy-wide competitiveness;
  - strengthening the educational system and training programs, including those that provide international certification.
- 6. Narrow the unemployment gender gap through deeper analysis and decisive policy responses to reduce an apparent bias against female employment.

### Annex 1 — Recommended labor market reforms (summary)
- Mix a more flexible labor market environment with safety nets for the unemployed, including training and certification.
- Strengthen social security for the unemployed; introduce active labor-market policies that train the unemployed and give them new qualifications.
- Design unemployment benefits to balance support and incentives, taking into account limited fiscal space; implement as part of expenditure switching towards better targeted social programs.
- Consider lessons from Barbados’ unemployment benefit scheme.
- Increase wage-setting flexibility given high coverage of collective bargaining agreements; possible measures include:
  - prioritize firm-level bargaining (example: Spain moved from sector-region to firm-level agreements, eased opt-outs, reduced ultra-activity period);
  - reduce barriers to part-time work (example: Germany).
- Alternatively or additionally, strengthen social dialogue using Social Partnerships as the platform.

### Road map for fostering/strengthening social dialogue and social partnerships
- Establish a well-articulated legal framework to commit to a social partnership.
- Define objectives through consensus with all partners.
- Ensure social partnership structure is flexible and inclusive.
- Conduct social dialogue at all appropriate stages of the decision-making process.
- Ensure engagement is consultative and transparent.
- Involve social partners in monitoring and evaluation of policy outcomes.

*Source: wpiea2019144 (IMF staff analysis and national statistical sources included in the provided content).*

### 2005. It is likely that part of the variation in

### wpiea2019144 - 2005. It is likely that part of the variation in

### Unemployment definitions and measurement
- Two definitions discussed:
  - “Relaxed” definition: considers an individual to be unemployed if she is not employed regardless of her search status if she is willing to accept a job offer.
  - “Strict” definition: considers an individual to be unemployed only if she is not employed and is actively searching for a job.
- Several ECCU countries have adopted the “relaxed” definition following the International Labor Organization’s recommendation.
- Effects of definition choice:
  - Under the “relaxed” definition the unemployment rate is about 5 percentage points higher than under the “strict” definition in ATG, GRD and LCA.
  - Under the “relaxed” definition unemployment reaches almost 30 percent in GRD.

### Structural and cyclical components of unemployment
- Long-run structural component:
  - Several decades of high unemployment imply an important structural component in most ECCU countries.
  - LCA annual series: unemployment stood at 15 percent in 1994 (pre-fallout of the banana industry) and stood above 14 percent at the peak of the 2000s tourism-fueled boom and construction for the Cricket World Cup in 2007.
  - Filtering annual unemployment (HP-Filter) suggests structural unemployment in LCA has remained around 20 percent for more than two decades.
  - Structural unemployment trend reflects a large and long-lasting increase following the Great Recession.
- Cyclical component and the global crisis:
  - The global crisis was followed by a substantial increase in unemployment in LCA, adding a significant cyclical component to an already high structural rate.
  - LCA unemployment peaked in 2015, several years after the start of the global crisis, then declined as output approached potential but remained above pre-crisis levels for many years.
  - Unemployment was much higher than before the crisis also in ATG and VCT.
  - GRD: unemployment slightly decreased since 2008, but this reduction was most likely a reversal from the major increase that followed two natural disasters and the global financial crisis.
  - KNA experienced a reduction in unemployment following the global crisis, partly due to a major public employment program, by far the largest among ECCU countries as a share of total employment.
  - DMA: no immediate pre-crisis statistics; unemployment in 2009 (latest available) was higher than its 2001 level.
- Quantitative assessment using Okun’s Law (Kandil and others (2014) elasticity):
  - Estimated elasticity of unemployment to the output gap: -0.32.
  - Output gap in LCA increased by 16 percent between 2008 and 2014.
  - Inference: the slowdown in economic activity should have contributed 5 percentage points to LCA’s unemployment rate between 2008 and 2014, about two thirds of its actual increase between those years.

### Youth and gender disparities
- Youth unemployment:
  - Youth unemployment is considerably higher than the overall unemployment rate across ECCU countries, as globally.
  - Cross-regional comparison: the ratio of youth-to-overall unemployment rates shows that in all ECCU countries except ATG, this ratio is below its average in most other regions.
  - Actual youth unemployment rates are alarmingly high: almost half of the youth population unemployed in GRD, LCA, and VCT.
  - Implication: this constitutes a major social problem that could partly explain the high crime rates in those countries.
- Gender gap in unemployment:
  - Female-to-male ratios of unemployment in ECCU countries are considerably above the average level in the more gender egalitarian developing regions of East Asia and the Pacific and Eastern Europe and Central Asia, while closer to averages in other regions where gender discrimination is common.
  - Given high unemployment ratios in most ECCU countries, this translates into a large absolute gap.
  - Example: since 1994, St. Lucia’s female unemployment has been on average 6.6 percentage points higher than male unemployment.
  - The gender gap narrowed significantly following the global crisis due to a faster run-up of male unemployment but widened soon afterwards.
  - Grenada has had a similar sized gender gap in recent years.

### Key statistics and indicators (as reported)
- “Relaxed” vs “strict” definition difference: about 5 percentage points higher in ATG, GRD and LCA.
- GRD unemployment (relaxed definition): almost 30 percent.
- LCA unemployment: 15 percent in 1994; above 14 percent at peak of 2000s boom.
- Structural unemployment in LCA (HP-Filter): around 20 percent for more than two decades.
- Okun’s Law elasticity used: -0.32.
- LCA output gap increase (2008–2014): 16 percent.
- Implied contribution of output slowdown to LCA unemployment (2008–2014): 5 percentage points (about two thirds of the actual increase).
- Youth unemployment: almost half of youth population unemployed in GRD, LCA, and VCT.
- St. Lucia female-minus-male unemployment gap since 1994: 6.6 percentage points on average.

*Source: wpiea2019144 - 2005. It is likely that part of the variation in (IMF staff analysis and national statistical sources included in the provided content).*

### 24.  EAP = East Asia and

### III. LABOR DEMAND: GLOBAL CYCLES, STRUCTURAL SHOCKS, AND DISASTERS (ECCU)

### A. Global financial crisis and ECCU unemployment — key findings
- The global economic downturn exacerbated unemployment in ECCU countries; after sizable expansion during the 2000s and up to 2008, real output either stagnated or contracted in all ECCU countries.
- By 2015, only DMA, GRE, and KNA had real output recovered to pre-crisis levels; ATG and LCA remained considerably below pre-crisis levels.
- Employment effects by sector since 2008:
  - Formal employment remained sluggish or contracted in all ECCU countries since 2008.
  - Particularly marked contractions in Hotels and Restaurants and Construction in ATG and LCA.
  - Public administration employment increased in ATG and VCT as governments increased public employment (including temporary employment) in response to the crisis.
- Empirical support: Kandil and others (2014) found a significant relation between output fluctuations and employment in several Caribbean countries, including LCA.

### B. Sectoral demand: banana/sugar decline and public sector employment — key findings
- Banana and sugar exports phase-out (EU preferences) triggered decline in exports; additional effects from hurricanes, pests and disease further dampened industries.
- Numerical relevance:
  - Banana industry accounted for approximately 5 percent of GDP in 1979-1994 in DMA, LCA, and VCT and sugar in KNA; these contracted to less than 2 percent of GDP in 2005.
  - This constituted an annual decline of 0.3 percentage points of GDP on average over 1996-2005.
  - Affected economies grew by an annual average of 2.7 percent over 1996-2005.
- Conclusion: Timing and GDP growth imply the banana/sugar decline is unlikely to explain very high unemployment rates in the 2000s.
- Evidence: No increase in unemployment is observed between mid-1990s decline of these industries and 2008 (pre-global crisis); DMA saw unemployment fall from 23 percent in 1997 to 12 percent in 2001 (banana decline started after 1997).

### C. Natural disasters — key findings
- Natural disasters had major, immediate employment impacts, especially on agriculture.
- Grenada: Hurricane Ivan (2004) led to the most severe contraction in formal employment in the last two decades: overall formal employment contraction of 12 percent.
  - Agricultural output contracted by 40 percent between 2003 and 2005.
  - Formal agriculture employment contracted by 20 percent.
  - Agriculture accounted for 5.3 percent of total formal employment prior to the storms (in 2003); the formal employment contraction in agriculture was directly responsible for a 2.3 percent reduction in total formal employment.
- All largest natural disasters this century had particularly strong impact on agriculture employment.
- More recent: 2017 Category 5 hurricane hit Dominica with major negative employment impact, especially in agriculture and tourism.
- Employment index around disasters: Agriculture, Tourism, and Total employment show sharp declines indexed to 100 one year before the disaster for Antigua and Barbuda (Storm, 2008); Grenada (Storm, 2004); St. Lucia (Storm, 2010); St. Vincent and the Grenadines (Storm, 2010) and Flood (2013).

---

### IV. LABOR SUPPLY: RESERVATION WAGES AND SKILLS MISMATCH

### A. Remittances and labor force participation — key findings
- Hypothesis: High remittances could raise reservation wages and reduce labor supply.
- Cross-country glance: Remittances as percent of GDP are notably high only in DMA among ECCU countries; remittances as share of GDP are much lower than in Jamaica (17 percent).
- Probit analysis using GRD’s 2013, 2014 and 2015 labor force surveys:
  - Dependent variable: dummy for labor force participation.
  - Controls: age, gender, education, wealth, remittances dummy, household employment in tourism/public sector.
  - Results:
    - Probit results do not support that remittances are a significant factor explaining unemployment in GRD.
    - For youth population, higher remittances are associated with higher likelihood of participating in the labor force.
  - Possible interpretation: Remittances may be used as business capital, increasing entrepreneurship and labor force participation among youth.
- Graphic numeric: Workers' remittances, 2016 (Percent of GDP) shown for KNA, ATG, LCA, GRD, VCT, DOM (no numeric values provided beyond chart).

### B. Effect of having a well-paid household member — key findings
- Rapid growth in tourism and public sectors produced better paid urban employment.
- Probit regressions (Table 1) indicate having a household head employed in tourism and public sectors is associated with lower labor force participation.
- Evidence on wage differentials:
  - In GRD, average salary in 2014 in tourism and public sectors was 66 and 93 percent higher than in formal agriculture.
  - In LCA in 2013, tourism and public sector salaries were 50 percent and 63 percent higher than formal agriculture.

### C. Skills mismatch — key findings and caveats
- Employer surveys report skill deficiencies (e.g., Grenada employers’ survey, De Koning and Jong, 2014).
- St. Lucia 2012 national labor market needs survey:
  - Job seekers with below secondary education: 60 percent.
  - Job openings requiring secondary or higher education: 75 percent.
  - Job seekers with tertiary education: 7 percent.
  - Job offers requiring tertiary education: 40 percent.
- Caveats:
  - Employer sampling biased against low-skilled employers (excludes agricultural farm holdings/firms).
  - Absolute numbers: job openings requiring tertiary education (357) much lower than job seekers with tertiary education (1,547).
  - Filling all tertiary-requiring openings would reduce unemployment by 0.37 percentage points in St. Lucia — unlikely to explain very high structural unemployment.
- Conclusion: No strong support for broad educational-category mismatch (primary, secondary, tertiary) as main driver of high unemployment.

---

### V. STRUCTURAL RIGIDITIES AND THE ROLE OF GOVERNMENT

### A. High unit labor costs (ULCs) and unemployment — evidence
- ECCU countries show evidence of high unit labor costs compared with countries of similar income per capita.
- World Bank (2005) findings for 2002:
  - Wages for unskilled and semi-skilled labor about 25 percent higher in ECCU than in other upper-middle income countries.
  - Wages for professionals about 45 percent higher.
- Post-global crisis: Average wages in ECCU continued to increase despite substantial decline in productivity, leading to significant increases in ULCs.
- Tourism sector wages (Chen, 2018, 2016 data): ratio of average wage in Hotels and Restaurants to GDP per capita considerably higher in ECCU than in Bahamas, Dominican Republic, and Mexico.
- Minimum wage indicators:
  - Minimum wage-to-GDP per capita ratio elevated relative to “competitive middle-income” countries (top 50 Ease of Doing Business).
  - St. Lucia does not have an official minimum wage (union representatives argue imposing minimum wage could lower current wages).
- Numeric evidence (selected):
  - Private wages, productivity, unit labor cost index (3 yr. moving average, 2001=100) plotted for ECCU 2001-2013.
  - ECCU and comparator selected wage rates in 2002 (hourly and annual US$ levels) presented (e.g., Unskilled hourly: ECCU 2.3, Upper-Middle Income 1.8, Microstates 1.9; Professional annual: ECCU 22.3, Upper-Middle Income 15.5, Microstates 19.0).

### B. ULCs correlated with unemployment — quantitative association
- Approximation: ULCs = average wage in social security statistics / GDP per capita.
- Periods analyzed: early 2000s (structural) and mid-2010s.
- Results:
  - Positive and significant relation between approximated ULCs and unemployment in both periods.
  - Back-of-the-envelope slope estimates:
    - 10 percent higher ULCs associated with higher structural unemployment of 3.4 percent in the early 2000s.
    - 10 percent higher ULCs associated with higher structural unemployment of 5.1 percent in the mid-2010s.
  - Example: VCT with ULCs more than 30 percent above ATG and KNA is expected to have about 11 to 16 percentage points higher unemployment (based on slope estimates).
- Caution: Small sample size and methodological limitations; results are suggestive not definitive.

### C. Strong union activity and collective bargaining coverage
- Trade union density (membership as share of workforce) is not significantly different on average in ECCU relative to other regions.
- Collective Bargaining Agreements (CBAs) affect more than 80 percent of labor force on average in ECCU — multiple times higher than other regions.
- Many unions historically aligned with major political parties, increasing influence.
- Unions cover large, strategic sectors: public sector, ports, finance, manufacturing, tourism, privately-run utilities.
- Public sector is a focal point: government employment is relatively large; public wage negotiations significantly influence the entire labor market.
- Despite strong union activity, labor regulatory framework is not particularly rigid regarding hiring/firing: dismissal without third-party approval is allowed; no prohibitions to fixed-term contracts for permanent tasks; probation periods commonly allowed (up to 6 years in DMA and VCT).

### D. Public sector wage demonstration effects
- Empirical evidence from social security data (ATG, GRD, LCA, VCT) and dynamic panel regression:
  - Private sector wages are strongly related to lagged public-sector wage (lagged public-sector wage coefficient significant in OLS, Random Effects, Fixed Effects, Difference GMM, System GMM).
  - Lagged CPI and lagged productivity are not robustly significant predictors of private sector wages.
- VAR regression and Granger causality tests on Grenada (IMF, 2016) find bi-directional causality between public and private wages — both “signaling” and “competition” effects.
- Implication: Public sector wage increases can rapidly transmit to private sector wages, raising economy-wide ULCs and affecting private sector profitability and job creation.

---

### VI. CONCLUSIONS AND POLICY-ORIENTED FINDINGS

- Unemployment in most ECCU countries reflects both temporary (natural disasters, global financial crisis) and structural components; the persistence of very high unemployment in DMA, GRD, LCA, and VCT implies enduring structural factors.
- High ULCs are likely an important factor behind high structural unemployment; moderation of wage growth may be prudent to reduce unemployment.
- Contributing factors to high ULCs and wage inflation:
  - Strong union activity with CBAs covering over 80 percent of the labor force.
  - Public sector wage “demonstration effects” that transmit public wage increases into private sector wages.
- Labor supply factors with limited support:
  - Remittances: not statistically significant in reducing labor force participation in GRD; for youth remittances correlate with higher participation.
  - Skills mismatch at broad educational categories: evidence does not support a major role for broad tertiary/secondary/primary mismatches in explaining very high unemployment (example: filling all tertiary-required job openings in St. Lucia would reduce unemployment by only 0.37 percentage points).
- Sectoral shocks:
  - Banana/sugar industry decline contributed to sectoral employment shifts but, given overall growth in 1996-2005, is unlikely to explain very high unemployment in the 2000s.
  - Natural disasters have immediate and large negative employment impacts, particularly in agriculture (e.g., Grenada: 12 percent overall formal employment contraction; agriculture output -40 percent between 2003 and 2005; formal agriculture employment -20 percent).
- Policy-relevant empirical relationships identified:
  - Positive correlation between approximated ULC (average wage-to-GDP per capita) and unemployment (early 2000s: slope implies 10 percent ULC ↑ → 3.4 percent unemployment ↑; mid-2010s: 10 percent ULC ↑ → 5.1 percent unemployment ↑).
  - Strong, statistically significant lagged relation: Lagged public sector wage → current private sector real wage (System GMM lagged public coefficient 0.820***).
- Recognizing the mix of cyclical and structural drivers (natural disasters, global shocks, high ULCs, union and public wage dynamics, limited role of remittances and broad skills mismatch) underscores the need for targeted policies addressing wage-setting dynamics, public sector wage management, disaster resilience in agriculture, and labor market flexibility to foster private sector job creation.

*Source: IMF staff chapter on ECCU labor markets (excerpt provided).*

### 1. Increase wage-setting flexibility to avoid misalignments between wages and productivity

### 1. Increase wage-setting flexibility to avoid misalignments between wages and productivity

### Key findings and diagnosis
- Wage/productivity disconnects can lead to both cyclical and/or structural unemployment.
- Wage flexibility is particularly critical as an adjustment mechanism in a fixed nominal exchange rate context to adjust labor costs in case of external shocks.
- Strong union activity generates significant wage premiums (Bryson, 2014), contributing to high unit labor costs (ULCs).
- There are currently no unemployment benefits programs in these countries; social security for the unemployed is substantially underdeveloped.
- Skills mismatches may exist in specific technical areas but are unlikely to explain the very high unemployment rates.
- Example evidence: in Barbados, a social partnership contributed to a reduction in unemployment from 20 percent in 1995 to 9 percent in 2000.

### Policy recommendations (numbered items from source)
- 1. Implement “flexicure” labor market arrangements that combine flexible labor rules with unemployment safety nets that include training (see annex 1).
  - If reducing collective bargaining coverage is not possible, promote wage flexibility by establishing social partnership among governments, firms and workers to incorporate macroeconomic considerations into wage-setting discussions.
- 2. Control public sector wage growth by linking public sector wages to productivity and involving unions more closely during the budget process to inform them about fiscal budget constraints.
- 3. Improve the effectiveness of countercyclical policies to minimize cyclical unemployment; without a flexible exchange rate, countries need to build fiscal space to undertake countercyclical fiscal policies, though most ECCU countries are in a weak fiscal situation with high public debt.
- 4. Build resilience to natural disasters by adopting a fiscal framework that finances prevention, relief and reconstruction efforts; developing the capacity to design and implement such a framework deserves priority attention.
- 5. Foster productivity and human capital growth to lower ULCs and thus promote structural employment by:
  - increasing economies of scale through measures to increase international and intraregional integration;
  - improving economy-wide competitiveness;
  - strengthening the educational system and training programs, including those that provide international certification.
- 6. Narrow the unemployment gender gap through deeper analysis and decisive policy responses to reduce an apparent bias against female employment.

### Annex 1 — Recommended labor market reforms (summary)
- Aim for labor market arrangements that mix a more flexible labor market environment with safety nets for the unemployed, including training and certification.
- Strengthen social security for the unemployed; introduce active labor-market policies that train the unemployed and give them new qualifications.
- Design unemployment benefits to strike a balance between providing support and encouraging labor participation, taking into account limited fiscal space; implement as part of expenditure switching towards better targeted social programs.
- Consider lessons from Barbados’ unemployment benefit scheme.
- Increase wage-setting flexibility given high coverage of collective bargaining agreements; possible measures include:
  - prioritize firm-level bargaining (example: Spain moved from sector-region to firm-level agreements, eased opt-outs, reduced ultra-activity period);
  - reduce barriers to part-time work (example: Germany).
- Alternatively or additionally, strengthen social dialogue using Social Partnerships as the platform.

### Road map for fostering/strengthening social dialogue and social partnerships
- There needs to be a well-articulated legal framework to commit to a social partnership.
- Objectives should be clearly defined through consensus with all partners.
- Social partnership structure should be flexible and inclusive.
- Social dialogue should take place at all appropriate stages of the decision-making process.
- Engagement should be consultative and transparent.
- Social partners should be involved in the monitoring and evaluation of the policy outcomes.

*Source: IMF working paper chapter: 1. Increase wage-setting flexibility to avoid misalignments between wages and productivity (wpiea2019144).*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2019/wpiea2019144.pdf_