## _wp0683 - 10.   External Shocks and Labor Adjustment: New Specification

## Source details

**Canonical URL:** [_wp0683 - 10.   External Shocks and Labor Adjustment: New Specification](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp0683.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp0683.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp0683.pdf.json)

---

### I. Introduction
- Labor market flexibility is a key determinant of differences in economic performance across countries; it helps explain the greater capacity of the United States to absorb adverse macroeconomic shocks compared to the euro area and remaining differences in living standards between these regions.
- Canada’s susceptibility to external shocks and highly regionalized economic structure place a premium on economic flexibility, including labor markets; exchange rate flexibility and the ability to shift workers within the country are important buffers.
- The paper examines Canada’s labor mobility across provinces, compares Canada with other countries, and focuses on whether labor mobility differs within Canada and depends on prevailing regional labor market conditions; Canadian Employment Insurance (EI) payments are more generous and more easily accessible where unemployment is high.
- Methodological heritage: extends Blanchard and Katz (1992) and Decressin and Fatas (1995) using a VAR approach on region-specific changes in employment, unemployment (employment ratio), and participation, and explicitly examines roles of migration and real wages for Canada.

### II. Data and Methodology
- Data and coverage:
  - Annual data on employment, unemployment (employment ratio), and the participation ratio for Canadian provinces and regions in other countries.
  - Other countries: United States (all 50 states), France and Germany combined, and Spain (Eurostat regions).
  - Additional Canadian provincial series: real wages and net migration.
  - Data start: mid-1970s for the United States and Canada, mid-1980s for the Europeans.
- Canadian regional groupings:
  - West: Alberta, British Columbia, Manitoba, Saskatchewan.
  - Center: Ontario, Québec.
  - Atlantic: Newfoundland, New Brunswick, Nova Scotia, Prince Edward Island.
- Cycle removal and specification:
  - Aggregate cycle removed by regressing the log of the current change in employment on a constant and the log of the change in national employment; robustness checks include using raw data that include the national cycle.
  - Core VAR specification variables: two lags of annual change in log employment (∆e), log employment ratio (emp), log participation rate (p).
  - Employment entered in changes; employment ratio and participation ratio entered in first differences in the new specification because panel tests cast doubt on stationarity of employment and participation ratios.
  - Traditional specification implicitly assumes employment and participation ratios are stationary and long-run adjustment occurs through population (migration).
- Testing and identification:
  - Table 1 reports two tests of stationarity for Canadian panel data: log employment nonstationary; participation likely nonstationary.
  - Disturbances to employment equation (ε∆e), ordered first, interpreted as labor demand shocks.
  - Supply-side shocks enter through employment ratio and participation ratio (εemp and εp) and feed through with a lag.
  - Unexplained adjustment assumed to reflect population changes (migration) unless migration is explicitly included.

### III. International Comparisons — key empirical findings
- Typical responses (new specification):
  - A typical acyclical annual disturbance to employment in a U.S. state is around 1 percent of overall jobs; this rises over the next few years to a long-term value of 1¾ percent.
  - About one-quarter of the initial shock in U.S. states is absorbed by a rise in the employment ratio (fall in unemployment); this effect rises slightly in year two then becomes minimal over the longer term.
  - Changes in participation explain the lion’s share of initial change in employment in U.S. states and stay relatively stable over time.
  - Changes in population (migration) play a minor role initially but rise over time and play a major role in long-term employment expansion in the United States and Canada.
- Cross-country contrasts:
  - Canada’s labor market adjusts broadly similar to the United States.
  - In EU countries (France, Germany, Spain), population changes play a much less important role; some initial employment gains are lost over time as population fails to expand.
  - Spain notable for the size of cyclical shocks.
  - Conclusion: greater migration distinguishes more flexible North American labor markets from less flexible European counterparts; results robust whether or not national cycle is removed.
- Specification robustness: traditional specification yields broadly similar qualitative conclusions though with different roles over time for unemployment and participation ratios; migration remains more important in North America.

### IV. Results across Canadian provinces
- Regional patterns (new specification, Figure 4):
  - Western provinces:
    - Face large employment shocks that grow over time.
    - Population change plays a dominant role in adjustment.
  - Atlantic provinces:
    - Face similarly sized initial shocks as the west, but employment does not expand thereafter due to limited role for population change (though the population response remains larger than in EU countries).
  - Central provinces:
    - Face smaller employment shocks (size and industry diversity imply more adjustment within the province).
    - Employment rises mildly over time, much less than in the west; population change plays a less important role.
- Ontario versus Québec (Figure 5):
  - Similar sized employment disturbances and initial responses.
  - Ontario: employment expands over time and population change plays a dominant role in long-term adjustment.
  - Québec: employment remains relatively stable over time and population change plays a minimal role.
- Inclusion of national cycle (Figure 6):
  - Sizes of employment shocks across regions become more similar.
  - Responses depend more on whether shocks to employment and participation ratios are considered permanent; assuming permanence produces less cyclical responses.
- Overall: labor markets become more flexible moving west; migration central to adjustment in west and Ontario, limited in Atlantic provinces and Québec.

### V. Extending the Canadian model — migration and wages
- Migration inclusion:
  - Migration variable defined as mig_t = log(pop_t) - log(pop_{t-1}) - (log(pop_t - m_t) - log(pop_{t-1} - m_{t-1})); migration data include those under 15.
  - Migration ordered last in the VAR (people become unemployed or leave labor force before relocating).
- Findings with migration (Figure 7 and Figure 8):
  - Overidentifying restriction (that adjustment unexplained by employment ratio and participation reflects migration) is much more closely satisfied for the new specification than the traditional one.
  - Migration is the major source of long-term adjustment in western provinces (represents some two-thirds of long-term labor market adjustment) and in Ontario (over half of long-term adjustment).
  - Further east, migration is significantly a less important adjustment channel.
  - Shock dynamics:
    - Employment-ratio shock: “discouraged worker effect” — people switch between unemployment and dropping out of labor force, little impact on employment or migration.
    - Participation-ratio shock: “discouraged migrant effect” — individuals not in labor force leave a province, boosting participation, followed by a fall in employment but a small increase in the employment ratio.
    - Inward migration: associated with subsequent increases in employment and participation — migrants move to places with jobs and join the workforce.
- Real wages inclusion (Figure 9):
  - Real wages defined as income from employment divided by employment and provincial CPI.
  - Change in log real wages included immediately after change in log employment to help identify whether employment shocks are labor demand (real wages rise) or supply (real wages fall).
  - Results:
    - Real wages rise in response to positive employment shocks — consistent with labor demand shocks.
    - Size of percentage changes in real wages relatively small — between one-quarter and one-third of the change in employment.
    - Comparing full VAR with variant where real wages are made exogenous indicates real wages play a relatively minor role in regional labor market adjustment.
    - Long-term adjustment occurs mainly via population changes (migration) rather than through relative wage changes.
- External shocks specification (equation (9), Figure 10):
  - First three variables: changes in log real nonenergy commodity prices in U.S. dollars (p_nec), log real energy commodity prices in U.S. dollars (p_en), and real effective exchange rate (reer).
  - These identify responses to: nonenergy commodity price shocks, oil price shocks not linked to nonenergy commodity prices, exchange rate changes not linked to commodity price developments, and other labor demand shocks.
  - Findings:
    - External price shocks produce generally similar adjustment patterns to other labor demand shocks; labor market responses largely independent of shock type.
    - Nonenergy commodity prices and exchange rate shocks are relatively transient; oil shocks have a more permanent component.
    - Nonenergy and energy commodity price shocks increase employment in the west with minor effects elsewhere; exchange rate shocks have more significant effects in all regions.
    - Regional pattern remains: larger employment shocks in the west resulting in significant population changes; less mobility elsewhere.

### VI. Conclusions and policy implications
- Key conclusions:
  - Canada has a relatively flexible labor market; like the United States, migration plays a significant role in adjusting to labor demand shocks.
  - Migration is much less important in Spain, France, and Germany; this difference helps explain differing levels of labor market flexibility across countries.
  - Labor markets are more flexible in the west than in the east in Canada.
  - Migration is a major mechanism for long-term adjustment in western provinces and Ontario; much less so in Atlantic provinces and Québec.
  - Real wage differentials are a minor factor relative to migration.
  - Adjustment patterns are similar across different types of macroeconomic disturbances.
- Policy implications and determinants:
  - Cross-country literature points to minimum wages, union representation, unemployment and social benefits, and housing-move constraints as policy levers; Canada’s relative flexibility shifts focus to the EI system.
  - Empirical research indicates economic opportunities, cost of migration, age, and language are primary determinants of interprovincial migration in Canada.
  - EI appears to have a modest negative effect on mobility (recent studies), particularly for those with limited labor-force attachment, though some studies suggest EI may have improved mobility by providing resources to the unemployed.
  - Eliminating differences in EI eligibility and benefits across localities would tend to reduce the regional differences in labor market flexibility identified in this analysis.

### Key table — stationarity test p-values (Table 1)
- Log employment 0.99 0.00 **
- ∆Log employment 0.00 ** 0.88
- Log employment ratio 0.18 0.32
- ∆Log participation rate 0.07 0.00 **
- Migration 0.00 ** 0.57
- Log real wage 0.83 0.00 **
- ∆Log real wage 0.00 ** 0.08
- Notes: One and two asterisks indicate the result is significant at the 1 and 5 percent levels, respectively.

*Source: _wp0683 - 10.   External Shocks and Labor Adjustment: New Specification...............................................27*

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

### _wp0683 - References

### References
- References....................................................................................................................................15

### Table
- 1. Stationary Tests on Canadian Panel Data ............................................................................17

### Figures
- 1.     Labor Market Developments Across Canadian Regions, 1980–2004 .................................18
- 2.     International Responses to Shocks in Employment Demand ..............................................19
- 3.     International Responses to Shocks in Employment Demand Including Economic  
     Cycle in the Data..............................................................................................................20
- 4.     Canadian Regional Labor Market Adjustment ....................................................................21
- 5. Canadian Labor Market Adjustment in Ontario, Québec ....................................................22
- 6.     Canadian Regional Labor Market Adjustment with Economic Cycle Included  
     in the Data ........................................................................................................................23
- 7.     Canadian Regional Labor Market Adjustment with Migration Effects...............................24
- 8.     Regional Adjustment with Migration Effects: New Specification ......................................25
- 9.     Real Wages and Labor Adjustment: New Specification......................................................26

*Source: _wp0683 - References.*

### 10.   External Shocks and Labor Adjustment: New Specification...............................................27

### 10.   External Shocks and Labor Adjustment: New Specification...............................................27

### I. INTRODUCTION
- Labor market flexibility is a key determinant of differences in economic performance across countries; it helps explain the greater capacity of the United States to absorb adverse macroeconomic shocks compared to the euro area and remaining differences in living standards between these regions.
- Canada’s susceptibility to external shocks and highly regionalized economic structure place a premium on economic flexibility, including labor markets; exchange rate flexibility and the ability to shift workers within the country are important buffers.
- The paper examines Canada’s labor mobility across provinces, compares Canada with other countries, and focuses on whether labor mobility differs within Canada and depends on prevailing regional labor market conditions, noting that Canadian Employment Insurance (EI) payments are more generous and more easily accessible where unemployment is high.
- Methodological heritage: follows and extends Blanchard and Katz (1992) and Decressin and Fatas (1995) by using a VAR approach on region-specific changes in employment, unemployment (employment ratio), and participation, and explicitly examining roles of migration and real wages for Canada.

### II. DATA AND METHODOLOGY
- Data:
  - Annual data on employment, unemployment (employment ratio), and the participation ratio for Canadian provinces and regions in other countries.
  - Other countries: United States (all 50 states), France and Germany combined, and Spain (Eurostat regions).
  - Additional Canadian provincial series: real wages and net migration.
  - Data start: mid-1970s for the United States and Canada, mid-1980s for the Europeans.
- Regional groupings for Canada:
  - West: Alberta, British Columbia, Manitoba, Saskatchewan.
  - Center: Ontario, Québec.
  - Atlantic: Newfoundland, New Brunswick, Nova Scotia, Prince Edward Island.
- Cycle removal: aggregate cycle taken out by regressing the log of the current change in employment on a constant and the log of the change in national employment; robustness checks include using raw data that include the national cycle.
- Core VAR specification:
  - Variables: two lags of annual change in log employment (∆e), log employment ratio (emp), log participation rate (p).
  - Employment entered in changes; employment ratio and participation ratio entered in first differences in the new specification (not levels) because panel tests cast doubt on stationarity of employment and participation ratios.
  - Implicit assumption in traditional specification: employment and participation ratios are stationary and long-run adjustment occurs through population (migration).
- Testing and specification choice:
  - Table 1 reports two tests of stationarity for Canadian panel data: log employment nonstationary; employment and participation ratios less clear-cut, participation likely nonstationary.
  - Result: focus on VAR specification including employment and participation ratios in first differences; also report results from the traditional specification for comparison.
- Identification assumptions:
  - Disturbances to employment equation (ε∆e), ordered first, interpreted as labor demand shocks.
  - Supply-side shocks enter through employment ratio and participation ratio (εemp and εp) and feed through with a lag.
  - Unexplained adjustment assumed to reflect population changes (migration) unless migration is explicitly included.

### III. INTERNATIONAL COMPARISONS
- Responses to regional employment shocks are compared for Canada, United States, Spain, France, and Germany (France and Germany combined due to instability).
- Key empirical observations (new specification):
  - A typical acyclical annual disturbance to employment in a U.S. state is around 1 percent of overall jobs; this rises over the next few years to a long-term value of 1¾ percent.
  - About one-quarter of the initial shock in U.S. states is absorbed by a rise in the employment ratio (fall in unemployment); this effect rises slightly in year two then becomes minimal over the longer term.
  - Changes in participation explain the lion’s share of initial change in employment in U.S. states and stay relatively stable over time.
  - Changes in population (migration) play a minor role initially but rise over time and play a major role in long-term employment expansion in the United States and Canada.
- Cross-country contrast:
  - Canada’s labor market adjusts broadly similar to the United States.
  - In EU countries (France, Germany, Spain), population changes play a much less important role; some initial employment gains are lost over time as population fails to expand.
  - Spain notable for the size of cyclical shocks.
  - Conclusion: greater migration distinguishes more flexible North American labor markets from less flexible European counterparts; results robust whether or not national cycle is removed.
- Specification robustness:
  - Traditional specification yields broadly similar qualitative conclusions though with different roles over time for unemployment and participation ratios; migration remains more important in North America.

### IV. RESULTS ACROSS CANADIAN PROVINCES
- Regional economic structure and potential implications for adjustment:
  - Western provinces: specialized in raw material production; remote location could slow labor adjustment.
  - Central provinces: manufacturing heartland; size (Ontario and Québec ~ one-third of national employment each) and Québec’s cultural differences may affect adjustment.
  - Atlantic provinces: smaller, declining industries; migration discouraged by dwindling pool of young/mobile and by more generous, easily accessed unemployment benefits.
- New-specification findings (top panels, Figure 4):
  - Western provinces:
    - Face large employment shocks that grow over time.
    - Population change plays a dominant role in adjustment.
  - Atlantic provinces:
    - Face similarly sized initial shocks as the west, but employment does not expand thereafter due to limited role for population change (though the population response remains larger than in EU countries).
  - Central provinces:
    - Face smaller employment shocks (size and industry diversity imply more adjustment within the province).
    - Employment rises mildly over time, much less than in the west; population change plays a less important role.
- Ontario vs Québec (Figure 5):
  - Similar sized employment disturbances and initial responses.
  - Ontario: employment expands over time and population change plays a dominant role in long-term adjustment.
  - Québec: employment remains relatively stable over time and population change plays a minimal role.
  - Consistent across both specifications; aligns with micro evidence that French-speakers and Québec residents are less likely to migrate.
- Inclusion of Canadian cycle (Figure 6):
  - Sizes of employment shocks across regions become more similar.
  - Responses depend more on whether shocks to employment and participation ratios are considered permanent; assuming permanence produces less cyclical responses.
- Overall intra-Canada pattern: labor markets become more flexible moving west; migration central to adjustment in west and Ontario, limited in Atlantic provinces and Québec.

### V. EXTENDING THE CANADIAN MODEL
- Add migration explicitly:
  - Migration variable defined as: mig_t = log(pop_t) - log(pop_{t-1}) - (log(pop_t - m_t) - log(pop_{t-1} - m_{t-1})) [as in equation (5)]; measures change in growth of population caused by migration (migration data include those under 15).
  - Migration ordered last in the VAR (people become unemployed or leave labor force before relocating).
- Findings from VAR including migration (Figure 7):
  - Overidentifying restriction (that adjustment unexplained by employment ratio and participation reflects migration) is much more closely satisfied for the new specification than the traditional one.
  - Migration is the major source of long-term adjustment in western provinces (represents some two-thirds of long-term labor market adjustment) and in Ontario (over half of long-term adjustment).
  - Further east, migration is significantly a less important adjustment channel.
- Shock dynamics (Figure 8):
  - Employment-ratio shock: “discouraged worker effect” — people switch between unemployment and dropping out of labor force, little impact on employment or migration.
  - Participation-ratio shock: “discouraged migrant effect” — individuals not in labor force leave a province, boosting participation, followed by a fall in employment but a small increase in the employment ratio.
  - Inward migration: associated with subsequent increases in employment and participation — migrants move to places with jobs and join the workforce.
- Add real wages to VAR:
  - Real wages defined as income from employment divided by employment and provincial CPI.
  - Change in log real wages included immediately after change in log employment to help identify whether employment shocks are labor demand (real wages rise) or supply (real wages fall).
  - Results (Figure 9, full VAR):
    - Real wages rise in response to positive employment shocks — consistent with labor demand shocks.
    - Size of percentage changes in real wages relatively small — between one-quarter and one-third of the change in employment.
    - Comparing full VAR with variant where real wages are made exogenous indicates real wages play a relatively minor role in regional labor market adjustment.
    - Long-term adjustment occurs mainly via population changes (migration) rather than through relative wage changes.
- External shocks specification (equation (9), Figure 10):
  - First three variables: changes in log real nonenergy commodity prices in U.S. dollars (p_nec), log real energy commodity prices in U.S. dollars (p_en), and real effective exchange rate (reer).
  - These identify responses to: nonenergy commodity price shocks, oil price shocks not linked to nonenergy commodity prices, exchange rate changes not linked to commodity price developments, and other labor demand shocks.
  - Findings:
    - External price shocks produce generally similar adjustment patterns to other labor demand shocks; labor market responses largely independent of shock type.
    - Nonenergy commodity prices and exchange rate shocks are relatively transient; oil shocks have a more permanent component.
    - Nonenergy and energy commodity price shocks increase employment in the west with minor effects elsewhere; exchange rate shocks have more significant effects in all regions.
    - Regional pattern remains: larger employment shocks in the west resulting in significant population changes; less mobility elsewhere.

### VI. CONCLUSIONS
- International comparison:
  - Canada has a relatively flexible labor market; like the United States, migration plays a significant role in adjusting to labor demand shocks.
  - Migration is much less important in Spain, France, and Germany; this difference helps explain differing levels of labor market flexibility across countries.
- Intra-Canada findings:
  - Labor markets are more flexible in the west than in the east.
  - Migration is a major mechanism for long-term adjustment in western provinces and Ontario; much less so in Atlantic provinces and Québec.
  - Real wage differentials are a minor factor relative to migration.
  - Adjustment patterns are similar across different types of macroeconomic disturbances.
- Policy implications and determinants:
  - Cross-country literature points to minimum wages, union representation, unemployment and social benefits, and housing-move constraints as policy levers; Canada’s relative flexibility shifts focus to the EI system.
  - Empirical research indicates economic opportunities, cost of migration, age, and language are primary determinants of interprovincial migration in Canada.
  - EI appears to have a modest negative effect on mobility (recent studies), particularly for those with limited labor-force attachment, though some studies suggest EI may have improved mobility by providing resources to the unemployed.
  - Eliminating differences in EI eligibility and benefits across localities would tend to reduce the regional differences in labor market flexibility identified in this analysis.

*Source: _wp0683 - 10.   External Shocks and Labor Adjustment: New Specification...............................................27*

### REFERENCES

### _wp0683 - REFERENCES

### References cited
- Audas, Rick, and James McDonald, “Employment Insurance and Geographic Mobility: Evidence from the SLID,” SRDC Working Paper Series 03–03 (Ottawa: University of Brunswick Social Research and Demonstration Corporation).
- Bayoumi, Tamim, Douglas Laxton, and Paolo Pesenti, 2004, “Benefits and Spillovers of Greater Competition in Europe: A Macroeconomic Assessment,” CEPR Discussion Paper No. 4481 (London: Centre for Economic and Policy Research).
- Blanchard, Olivier, and Lawrence Katz, 1992, “Regional Evolutions,” Brookings Papers on Economic Activity, No. 1, pp. 1–75.
- Day, Kathleen, and Stanley Weiner, 1994, “Internal Migration and Public Policy,” in Issues in the Taxation of Individuals, ed. by A. Masgrove (Toronto: University of Toronto Press).
- Day, Kathleen, and Stanley Weiner, 2001, “Policy-Induced Migration in Canada: An Empirical Study,” Carleton University Working Paper 2001–08, Ottawa.
- Decressin, Jorg, and Antonio Fatas, 1995, “Regional Labor Market Dynamics in Europe,” European Economic Review, Vol. 39 December, pp.1627–55.
- European Commission, 2005, “Integration Guidelines for Growth and Jobs (2005–08)”; Communication to the Spring European Council.
- Finnie, R., 2000, “Who Moves? A Panel Logit Model Analysis of Inter-provincial Mobility in Canada,” Statistics Canada Analytic Studies Branch Research Paper 142 (Ottawa: Statistics Canada).
- Gersbach, 2003, “Structural Reforms and the Macroeconomy: The Role of General Equilibrium Effects,” CEPR Discussion Paper 4043 (London: Centre for Economic and Policy Research).
- International Monetary Fund, 2005, Canada: 2005 Article IV Consultation—Staff Report; Staff Statement; and the Public Information Notice on the Executive Board Discussion for Canada, IMF Country Report 05/118, Washington.
- Lin, Zhengxi, 1995, Interprovincial Labor Mobility in Canada: The Role of Unemployment Insurance and Social Assistance, HRDC Unemployment Insurance Evaluation Series 21 (Ottawa: Human Resources Development Canada).
- Lin, Zhengxi, 1998, Foreign-Born versus Native-Born Canadians: A Comparison of their Inter-Provincial Labour Mobility, Statistics Canada Analytic Studies Branch Research Paper 125 (Ottawa: Statistics Canada).
- Luzio, Rodolfo, and Vladimir Klyuev, 2006, “Regional Dimensions of the Canadian Economy,” unpublished paper available from the authors and forthcoming in International Monetary Fund Country Report, Canada: Selected Issues.
- OECD, 2005 Economic Policy Reforms: Going for Growth (OECD: Paris, France).

### Key table (stationarity tests) — reported values and notes
- Log employment 0.99 0.00 **
- ∆Log employment 0.00 ** 0.88
- Log employment ratio 0.18 0.32
- ∆Log participation rate 0.07 0.00 **
- Migration 0.00 ** 0.57
- Log real wage 0.83 0.00 **
- ∆Log real wage 0.00 ** 0.08
- Notes: One and two asterisks indicate the result is significant at the 1 and 5 percent levels, respectively.
- Table 1. Stationary Tests on Canadian Panel Data (p-values)
  - Hadri Z-test Null: No Unit Root
  - Null: Unit Root
  - Inverse Fisher Chi-square Test
  - Implication: Non-Stationary / Stationary / Ambiguous / Non-Stationary / Stationary / Non-Stationary / Stationary

### Figures and captions (visuals and focal variables)
- Figure 1. Labor Market Developments Across Canadian Regions, 1980–2004
  - Sources: Haver Analytics; and Fund staff calculations.
  - Series shown: Employment, 1980 = 100; Unemployment rate, in percent; Net migration inflows (percent of working-age population); Labor force participation rate, in percent.
  - Region groupings labeled: Central, Western, Atlantic.
- Figure 2: International Responses to Shocks in Employment Demand
  - New specification and Traditional specification panels for: United States, Canada, Spain, France & Germany.
  - Variables plotted: Unemployment, Participation, Employment (horizontal scale: periods 1 through 10; vertical scale from -0.5% to 2.0% in panels).
  - Source: Fund staff calculations.
- Figure 3: International Responses to Shocks in Employment Demand Including Economic Cycle in the Data
  - New specification and Traditional specification panels for: United States, Canada, Spain, France & Germany.
  - Variables plotted: Unemployment, Participation, Employment (horizontal scale: periods 1 through 10; vertical scale from -0.5% to 5.0% in panels).
  - Source: Fund staff calculations.
- Figure 4: Canadian Regional Labor Market Adjustment
  - New specification and Traditional specification panels for: Western Provinces, Central Provinces, Atlantic Provinces.
  - Variables plotted: Unemployment, Participation, Employment (horizontal scale: periods 1 through 10; vertical scale 0.0% to 2.5%).
  - Source: Fund staff calculations.
- Figure 5: Canadian Labor Market Adjustment in Ontario and Quebec
  - New specification and Traditional specification panels for: Ontario, Quebec.
  - Variables plotted: Unemployment, Participation, Employment (horizontal scale: periods 1 through 10; vertical scale -0.5% to 2.5%).
  - Source: Fund staff calculations.
- Figure 6: Canadian Regional Labor Market Adjustment with Economic Cycle Included in the Data
  - New specification and Traditional specification panels for: Western Provinces, Central Provinces, Atlantic Provinces.
  - Variables plotted: Unemployment, Participation, Employment (horizontal scale: periods 1 through 10; vertical scale 0.0% to 3.0% in some panels).
  - Source: Fund staff calculations.
- Figure 7: Canadian Regional Labor Market Adjustment with Migration Effects
  - New specification and Traditional specification panels for: Western Provinces, Ontario, Quebec, Atlantic Provinces.
  - Variables plotted: Unemployment, Participation, Net migration, Employment (horizontal scale: periods 1 through 10; vertical scale -0.5% to 2.5%).
  - Source: Fund staff calculations.
- Figure 8: Regional Adjustment with Migration Effects: New Specification
  - Shock types and responses shown: Employment shock, Unemployment shock, Participation shock, Net migration shock; responses plotted for Employment response, Unemployment response, Participation response, Net migration response (horizontal scale: periods 1 through 10; vertical scale -1.0% to 2.0%).
  - Source: Fund staff calculations.
- Figure 9: Real Wages and Labor Adjustment: New Specification
  - Panels for Wages endogenous and Wages exogenous across Western Provinces, Central Provinces, Atlantic Provinces.
  - Variables plotted: Unemployment, Participation, Net migration, Employment, Wages (horizontal scale: periods 1 through 10; vertical scale -0.5% to 2.5%).
  - Source: Fund staff calculations.
- Figure 10: External Shocks and Labor Adjustment: New Specification
  - Shock types: Non-energy commodity price shock, Energy commodity price shock, Exchange rate shock unconnected to commodity prices, Other demand shocks.
  - Panels for Western Provinces, Central Provinces, Atlantic Provinces.
  - Variables plotted: Unemployment, Participation, Employment, Commodity prices, Energy prices, Real exchange rate (horizontal scale: periods 1 through 10; various vertical scales shown from -1.0% to 2.5%).
  - Source: Fund staff calculations.

### Data appendix — panels and data sources (by region/country)
- United States
  - Panel: Fifty states plus the District of Columbia.
  - Data sources:
    - labor force Bureau of Labor Statistics
    - employment Bureau of Labor Statistics
    - unemployment Bureau of Labor Statistics
    - unemployment rate Bureau of Labor Statistics
    - population aged 16-64 = labor force / labor force participation rate /100
    - labor force participation rate Bureau of Labor Statistics
- Canada
  - National panel includes 10 provinces: Alberta, British Columbia, Manitoba, New Brunswick, Newfoundland & Labrador, Nova Scotia, Ontario, Prince Edward Island, Quebec, and Saskatchewan.
  - Western panel includes: Alberta, British Columbia, and Manitoba.
  - Central panel includes: Ontario and Quebec.
  - Atlantic panel includes: New Brunswick, Newfoundland & Labrador, Nova Scotia, and Prince Edward Island.
  - Data sources:
    - labor force Statistics Canada
    - employment Statistics Canada
    - unemployment Statistics Canada
    - unemployment rate Statistics Canada
    - population aged 15-64 Statistics Canada
    - domestic migration Statistics Canada
    - labor force participation rate Statistics Canada
    - total population Statistics Canada
    - personal income Statistics Canada
    - average hourly wages Statistics Canada
- Europe
  - French panel includes 8 regions: Bassin Parisien, Centre-Est, Est, Isle de France, Mediterraneo, Nord - Pas-de-Calais, Ouest, and Sud-Ouest.
  - German panel includes 17 regions: Bayern, Baden-Wittemberg, Berlin, Brandenburg, Bremen, Hamburg, Hessen, Mecklenburg-Vorpommern, Niedersachsen, Nordrhein-Westfalen, Rheinland-Pfalz, Saarland, Sachsen, Sachsen-Anhalt, Schleswig-Holstein, Sonneberg, and Thuingen.
  - Spanish panel includes 7 regions: Canarias, Centro, Este, Comunidad de Madrid, Noreste, Noroeste, and Sur.
  - Data sources:
    - labor force = employment + unemployment
    - employment Eurostat
    - unemployment Eurostat
    - unemployment rate = unemployment / labor force * 100
    - population aged 15-64 Eurostat
    - labor force participation rate = labor force / population aged 15-64 * 100

*Source: _wp0683 - REFERENCES (PDF), Data Appendix, and accompanying figures and tables as provided in the content unit.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2006/_wp0683.pdf_