## wp18150 ## Source details **Canonical URL:** [wp18150](https://www.imf.org/-/media/files/publications/wp/2018/wp18150.pdf) ## Other formats - [Markdown version](/-/media/files/publications/wp/2018/wp18150.pdf.md) - [Structured JSON version](/-/media/files/publications/wp/2018/wp18150.pdf.json) --- ### Aggregate Analysis — demographic context and scope - Population trends and projections: - Population growth in advanced economies (AEs) is slowing; life expectancy changes since 1985: life expectancy at birth increased by about seven years and at age 50 by over five years. - UN projections imply that by the middle of this century total population will be shrinking in almost half of AEs and individuals of working age will be supporting close to double the number of elderly. - Analysis periods and samples: - Aggregate data: past three and a half decades; long-term trends discussed for 1985–2016 based on a sample of 21 AEs. - Individual-level data: 24 European countries during 2000–16. - Figures/tables use samples of 33 AEs and 36 AEs in various displays. ### Aggregate patterns of labor force participation — key statistics and dynamics - Aggregate participation: - The aggregate average labor force participation rate in AEs as a group barely changed over the past 30 years. - Several countries saw aggregate participation gains of more than five percentage points (examples: Germany, Korea, Spain, the Netherlands). - Gender dynamics: - Share of women employed or actively looking for work increased by close to 10 percentage points over the past three decades. - For the median AE, prime-age male participation rate was more than 6 percentage points lower in 2016 than in 1985. - Female gains largely offset male declines among prime-age workers. - Age dynamics: - Youth (ages 15–24) are significantly less likely to be part of the labor force in 2016 than in 1985; increased school enrollment is important. - Share of “idle” youth is small and stable since early 2000s. - Older workers (ages 55 and over) increased participation significantly since the mid 1990s, especially ages 55–64; in the past decade even those older than 65 remained in the labor force longer. - Country-specific: - United States: prime-age male participation decline particularly steep in the past decade; prime-age female participation plateaued in the late 1990s and declined since the Global Financial Crisis. - Labor force and population composition (2015, average AE): - Labor force composition: 37 percent prime-age men; 31 percent prime-age women; 11 percent aged 15–24; 21 percent older than 55. - Population composition: 20 percent prime-age men; 20 percent prime-age women; 12 percent aged 15–24; 31 percent older than 55. ### Drivers and explanatory framework - Candidate drivers examined: - Cyclical versus structural changes. - Labor market policies and institutions (tax–benefit systems, ALMPs, retirement policies). - Policies targeting groups: women, older individuals, migrants. - Technological change: automation of routinizable tasks and shifts in occupation demand. - Population characteristics: education, fertility, marriage, immigration status. - Empirical approaches to routinization: - Cross-country: exploit heterogeneity in initial occupational mix and relative price of investment. - Individual-level: use current and past employment data from 24 European countries (2000–16) to test if workers in routinizable occupations are more likely to be out of the labor force, and whether country policies attenuate that link. ### Main aggregate findings and policy implications - Policies and structural factors: - Tax–benefit systems, ALMPs, policies encouraging group participation, structural change, and gains in educational attainment account for the bulk of the dramatic increase in labor force attachment of prime-age women and older workers over the past three decades. - Technological change: - Automation of tasks where labor is easily substitutable by capital weighed on participation rates of most groups. - Mitigating factors: - Higher spending on ALMPs and education associated with lower likelihood that individuals previously employed in routinizable occupations drop out of the labor force. - Urban residence associated with significantly lower likelihood of dropping out after prior employment in routinizable occupations. - Heterogeneity: - Female participation gains widespread across marital status, parental status, nativity, and education—except relatively low-educated women where gains are less pronounced. - Prime-age male participation declines deepest among those with the lowest educational attainment. - Welfare consequences: - Literature links detachment during peak productive years to lower happiness and life satisfaction for men, poorer health and higher mortality, and depressed employment prospects. - Policy design implications (priorities): - Strengthen tax–benefit incentives encouraging participation. - Expand and target ALMPs and education to improve resilience to automation-induced displacements. - Encourage participation of women, older workers, and migrants. - Promote geographic mobility and urban access to diverse employer pools. - Identify and rank drivers of participation across population groups to prioritize policy levers. - Monitor interactions between automation/structural transformation and national policies. --- ### Men by educational attainment — empirical snapshots and subgroup patterns - Data and sample notes: - Cross-sectional comparisons focus on years 2000 and 2016; panels 4 and 8 use 2004 instead of 2000. - Statistics estimated from a random sample of 10,000 respondents per country per year. - Panels 1 and 5 based on most AEs; panels 2–4 and 6–8 on advanced European economies. - Panels 3 and 7 report statistics for married individuals. Young children = below age 5; older children = age 5–15. - Age groups analyzed (Figures 5 and related text): - All, age 15–24 (2000 and 2016) - Men, age 25–54 (2000 and 2016) - Women, age 25–54 (2000 and 2016) - All, age 55+ (2000 and 2016) - Reasons for inactivity categories (survey responses): - Students; Retired (includes early retirement); Dismissed; Family or childcare; Illness or disability; Temporary contract ended; Other; Have never worked. - Key descriptive findings: - Women more likely to be inactive to look after children; higher fraction of men report illness and disability. - A non-negligible share of inactive individuals may be “involuntarily inactive”; those reporting being dismissed provide a lower-bound indicator. - 2000→2016 trends: student share increased among young and prime-age; early retirement among prime-age fell; share who never worked fell among prime-age women and those over 55; illness/disability became relatively more important. ### Sectoral and occupational concentration of involuntary inactivity - Sector concentration: - Wholesale and retail trade, manufacturing, mining and quarrying, and utilities together account for over half of the involuntary inactive, though they account for less than one third of active workers. - Relationship to routinization: - Excess involuntary inactivity tends to be concentrated in sectors with greater shares of routine jobs vulnerable to automation. - Occupational and income distribution: - Displacement tends to occur disproportionately among lower and middle-skilled occupations; vulnerability pronounced in middle and lower parts of income distribution. - Occupational categories used: - Managers; Professionals; Technicians; Clerical support; Craft and related trades; Plant and machine operators; Agricultural workers; Elementary occupations; Services and sales. - Sector list referenced: - Accommodation and food service; Administrative, health, education, other; Agriculture, forestry, and fishing; Construction; Information, finance, real estate, professional; Manufacturing; Mining and quarrying, utilities; Transport and storage; Wholesale and retail trade. ### Conceptual drivers of participation (policy, institutions, mechanisms) - Individual determinants: gender, educational attainment, previous occupation, household structure. - Policy/institution determinants: labor tax wedge; ALMPs; wage-setting institutions; public spending on childcare and education; proportion of part-time employees; weeks of job-protected maternity leave; statutory retirement age; pension generosity; migration integration policies; cultural attitudes. - Mechanisms highlighted: - Labor tax wedge can reduce net wages and labor demand but net labor supply effect ambiguous. - Childcare and family-friendly flexible arrangements raise opportunity cost of staying home and increase extensive-margin labor supply. - Pension incentives and social transfers important for older workers’ retirement decisions. - Structural growth of services raises female participation; automation may reduce demand for less-skilled labor with impeded adjustment due to skills, relocation costs, or low returns. --- ### Individual-Level Analysis — microdata, methods, and findings ### Data and empirical model - Data: European Labour Force Survey for 24 AEs during 2000–16; random sample of 10,000 individuals per country per year; logit regressions based on a random sample of 10,000 respondents per country per year from 19 countries. - Logit model notation preserved: - Φ(Sj = 1) = βC Ci + βH Hj + βR Rj + πc + υr + τt + ϵj - Sj: dummy for individual j being in or out of the labor force. - C: age, gender for age 55+, born in country, urban/rural, highest education (up to lower secondary, upper secondary, tertiary). - H: family composition (number of children, number of other employed adults, household type baseline = single adult without children; other categories: single adult with children, couple without children, couple with children). - R: routinization score of current or last occupation. - Fixed effects: πc (country), υr (region), τt (year). Standard errors clustered at country-year level. ### Aggregate cross-country associations (summary) - Education: - Increases in the share with secondary and especially tertiary education associated with significantly higher participation, particularly for prime-age women and older workers. - Business cycle and structural change: - Participation rates depend on business cycle, stronger for young and women. - Relative increase in service-sector employment typically followed by entry of prime-age women into labor force. - Urbanization brings gains across groups. - Technology and routinization: - Decline in relative price of investment associated with lower participation in countries with initial occupation mix tilted to routine tasks. - Labor market programs and institutions: - Higher labor tax wedges and more generous unemployment benefits associated with lower labor force attachment. - Public spending on ALMP tends to raise share of young and prime-age women working or seeking employment. - Migrant integration policies encouraging integration associated with stronger prime-age participation, especially for women. - Family reconciliation policies (childcare, maternity leave, flexible work) associated with higher female participation. - Raising statutory retirement age associated with delayed exit; greater pension generosity associated with earlier retirement. - Higher coordination of wage setting associated with greater participation for most groups; union density effects less robust. ### Decomposition 1995–2011 (text summary) - Supportive policies and educational gains drove large increases in participation of prime-age women and older workers. - Structural transformation contributed positively for these groups. - Technological advances weighed on participation for all groups except the young. - For young and some prime-age male workers, a large share of participation decline is attributed to a common component across AEs (time effects) possibly reflecting global forces, policy changes, structural shifts, changing returns to education, rising life expectancy, or scars from the global financial crisis. - For older workers, positive common component may reflect delayed retirement due to suppressed returns on retirement savings (falling global interest rates), losses in financial wealth, and potentially higher indebtedness. ### Micro (logit) estimation — selected exponentiated coefficients and magnitudes (Table 2) - Education and demographics (odds ratios, standard errors in parentheses): - Upper Secondary Education: - Men, ages 25–54: 1.719*** (0.032) - Women, ages 25–54: 1.709*** (0.033) - All, ages 55+: 1.209*** (0.036) - Tertiary Education: - Men, ages 25–54: 2.759*** (0.082) - Women, ages 25–54: 2.961*** (0.077) - All, ages 55+: 1.594*** (0.059) - Age odds ratios: - Men 1.158*** (0.011); Women 1.320*** (0.014); All 55+ 1.396*** (0.113) - Born in Country: - Men 1.489*** (0.035); Women 1.333*** (0.024); All 55+ 1.091** (0.046) - Urban: - Men 1.008 (0.019); Women 1.024* (0.013); All 55+ 1.019 (0.027) - Family composition effects (odds ratios): - Number of Children: - Men 1.049*** (0.009); Women 0.816*** (0.006); All 55+ 0.960* (0.020) - Couple with Children: - Men 1.726*** (0.052); Women 0.757*** (0.028); All 55+ 1.446*** (0.128) - Other Employed Adult(s) in Household: - Men 1.497*** (0.035); Women 1.152*** (0.038); All 55+ 1.703*** (0.091) - Routinization (RTI) effects (odds ratios): - RTI Score of Occupation: - Men 0.825*** (0.010); Women 0.900*** (0.010); All 55+ 0.716*** (0.013) - Interpretation example preserved: about 87 percent of prime-age male managers active vs about 84 percent of prime-age male technicians — RTI score difference can explain about one third of this 3 percentage-point difference. - Business cycle: - Lagged Output Gap odds ratios: - Men 1.037*** (0.006); Women 1.023*** (0.004); All 55+ 1.031*** (0.007) - Sample sizes: - Number of Observations: Men 491,820; Women 474,240; All 55+ 86,441 - Caveat: - Associations not fully causal; family composition and labor supply decisions may be jointly determined. ### Can policies offset routinization effects? — interaction magnitudes (Table 3) - Main finding: - Policies can attenuate negative association between routinization and participation; ALMP spending notably reduces negative association by about one-third at the 75th vs 25th percentile. - Selected δy/δx (effect of unit change in RTI) reported at 25th and 75th percentiles (standard errors in parentheses): - ALMP spending: - Men, ages 25–54: 25th -0.022 (0.002); 75th -0.011 (0.001) - Women, ages 25–54: 25th -0.015 (0.002); 75th -0.008 (0.001) - ALMP spending on training: - Men: 25th -0.012 (0.003); 75th -0.016 (0.002) - Women: 25th -0.014 (0.004); 75th -0.008 (0.003) - Education spending: - Men: 25th -0.016 (0.001); 75th -0.018 (0.001) - Women: 25th -0.017 (0.002); 75th -0.010 (0.002) - Wage-setting coordination: - Men: 25th -0.021 (0.001); 75th -0.013 (0.001) - Women: 25th -0.015 (0.002); 75th -0.010 (0.002) - Location interactions: - Rural: Men -0.019 (0.001); Women -0.017 (0.002) - Urban: Men -0.016 (0.001); Women -0.012 (0.001) - Additional points: - Disaggregated ALMP data suggest training spending drives mitigation for prime-age women. - Policies provide less offset for older workers; negative routinization effects larger for older cohorts and harder to mitigate. ### Policy-relevant implications from micro evidence - Strengthening secondary and tertiary education strongly associated with higher participation, notably for prime-age women and older workers. - ALMPs, particularly training, can attenuate adverse participation effects from routinization, especially for prime-age women. - Family reconciliation policies (childcare, maternity leave, flexible work) support higher female participation. - Retirement incentives matter: raising statutory retirement age delays exit; higher pension generosity encourages earlier retirement. - Labor taxation and unemployment benefit generosity associated with lower participation and may discourage attachment. - Wage-setting coordination associated with higher participation; union density effects are less robust. - Policy effectiveness varies by age: older workers more vulnerable to routinization and less responsive to policy offsets. --- ### Robustness — checks, alternative specifications, and robust determinants ### Robustness checks applied - Alternatives tested include: - Logistic transformation of dependent variable; SUR to account for cross-equation correlation; standard-error corrections via Beck and Katz (1995), HAC, Newey–West; collapsing dataset to five-year averages; excluding 2008–2009; including additional AEs; replacing lagged output gap with lagged unemployment rate; sample selection by dropping one country at a time. - Standard-error conventions preserved: - Driscoll–Kraay SEs in many columns; bootstrapped SEs for SUR; HAC SEs assuming panel-dependent correlation; column (11) reports 10th/90th percentiles. - General outcome: - Results broadly robust across alternative measures, specifications, and error-structure corrections. - Replacing output gap with unemployment rate yields qualitatively comparable results, though some sensitivity for prime-age men and older workers. - Dropping 2008–2009 does not materially change significance or magnitude of coefficients. ### Robust determinants by worker group — selected baseline coefficients (preserved) - Youth (Ages 15–24) — Table 4, column (1) (selected coefficients): - Lag of Output Gap: 0.360*** - Routinization*Relative Price of Investment: 0.303 - Lag of Trade Openness: 0.059*** - Lag of Urbanization: 0.668*** - Education (% Tertiary): -0.275*** - Public Spending on ALMP: 0.041*** - Restrictiveness of Migrant Integration Policies: 0.491*** - Coordination of Wage-setting: 1.104*** - Observations: 571; Countries: 23; R-squared: 0.515 - Prime-age Men (Ages 25–54) — Table 5, column (1): - Lag of Output Gap: 0.072*** - Routinization*Relative Price of Investment: 0.302*** - Lag of Urbanization: 0.101*** - Education (% Secondary): 0.019*** - Unemployment Replacement Ratio: -0.041*** - Restrictiveness of Migrant Integration Policies: -0.047** - Coordination of Wage-setting: 0.131** - Observations: 571; Countries: 23; R-squared: 0.606 - Prime-age Women (Ages 25–54) — Table 6, column (1): - Lag of Output Gap: 0.170* - Routinization*Relative Price of Investment: 1.793*** - Relative Service Employment: 0.015*** - Lag of Urbanization: 0.355*** - Education (% Secondary): 0.211*** - Education (% Tertiary): 0.332*** - Tax wedge: -0.129*** - Public Spending on ALMP: 0.039*** - Restrictiveness of Migrant Integration Policies: -0.462*** - Public Spending on Early Childhood Education and Care: 3.708*** - Share of Part-Time Employment: 0.946*** - Job-Protected Maternity Leave: 0.025*** - Observations: 489; Countries: 23; R-squared: 0.887 - Older Workers (Ages 55+) — Table 7, column (1): - Lag of Output Gap: -0.006 - Routinization*Relative Price of Investment: 0.505* - Lag of Trade Openness: -0.059*** - Education (% Tertiary): 0.389*** - Tax wedge: -0.263*** - Public Spending on ALMP: -0.025** - Union Density: -0.115*** - Statutory Retirement Age: 0.661*** - Public Spending on Old-Age Pension: -0.750*** - Observations: 489; Countries: 23; R-squared: 0.887 - Aggregate Participation Rate — Table 8, column (1): - Lag of Output Gap: 0.183*** - Routinization*Relative Price of Investment: 0.536*** - Lag of Trade Openness: 0.012* - Relative Service Employment: 0.010** - Lag of Urbanization: 0.249*** - Education (% Secondary): 0.063*** - Education (% Tertiary): 0.135*** - Tax wedge: -0.240*** - Unemployment Replacement Ratio: -0.078*** - Public Spending on ALMP: 0.031*** - Restrictiveness of Migrant Integration Policies: -0.207*** - Observations: 570; Countries: 23; R-squared: 0.578 - Significance notation preserved: ***, **, * indicate significance at 1, 5, 10 percent respectively. ### Individual-level robustness (Table 9) — exponentiated logit coefficients (selected) - Men, ages 25–54 (column (1)) — selected: - Age: 1.261***; Age Squared: 0.997***; Upper Secondary: 1.737***; Tertiary: 2.217***; Born in Country: 1.761***; Urban: 0.896***; RTI Score of Occupation: 0.467***; Lagged Output Gap: 1.042***; Predicted Income Decile: 0.952***. - Number of Observations: 474,434 - Women, ages 25–54 (column (2)) — selected: - Age: 1.347***; Upper Secondary: 1.855***; Tertiary: 2.763***; Born in Country: 1.520***; Urban: 0.864***; RTI Score of Occupation: 0.490***; Lagged Output Gap: 1.030***; Predicted Income Decile: 0.950***. - Number of Observations: 443,687 - All, ages 55+ (column (3)) — selected: - Age: 1.356***; Upper Secondary: 1.102**; Tertiary: 1.240***; Born in Country: 1.167**; Urban: 0.866***; RTI Score of Occupation: 0.488***; Lagged Output Gap: 1.037***; Predicted Income Decile: 0.952***. - Number of Observations: 63,982 - Robustness notes: - Including predicted income decile alters some household composition effects; income itself associated with lower odds of being in the labor force (Predicted Income Decile exponentiated coefficient 0.952***). - Vulnerability to routinization and education effects remain similar to baseline. - Results robust to inclusion of interacted country-year fixed effects in addition to region fixed effects. ### Robust substantive findings and preserved policy implications - Technological advances (automation / routinization) weighed on labor supply for most groups, with baseline coefficients: - Prime-age women 1.793***; prime-age men 0.302***; older workers 0.505*; aggregate 0.536***. - Education heterogeneity: - Education (% Tertiary): Youth -0.275***; Prime-age women 0.332***; Older workers 0.389***; Aggregate 0.135***. - Policies and institutions: - Public spending on ALMP: youth 0.041***; prime-age women 0.039***; aggregate 0.031***; older workers -0.025**. - Public spending on early childhood education and care: prime-age female participation 3.708***. - Statutory retirement age: 0.661***; public spending on old-age pension: -0.750***. - Restrictive migrant integration policies: youth 0.491***; prime-age women -0.462***; aggregate -0.207***. - Individual-level: - RTI Score of Occupation associated with lower odds of being in the labor force (exponentiated coefficients: men 0.467***; women 0.490***; ages 55+ 0.488***). - Policy recommendations preserved: - Invest in education, training, and activation policies to increase resilience to technological progress and globalization. - Reduce disincentives for joining/remain in labor force and help combine family and work life. - Rethink immigration policies to boost labor supply and encourage older workers to postpone retirement, acknowledging political challenges. --- ### Appendix A — sample - Aggregate analysis countries (23): Australia; Austria; Belgium; Canada; Denmark; Finland; France; Germany; Greece; Italy; Ireland; Japan; Korea; Luxembourg; Netherlands; New Zealand; Norway; Portugal; Spain; Sweden; Switzerland; United Kingdom; United States. - Individual-level analysis countries (24): Austria; Belgium; Cyprus; Czech Republic; Denmark; Germany; Estonia; Finland; France; Greece; Iceland; Ireland; Italy; Latvia; Lithuania; Netherlands; Norway; Portugal; Slovakia; Slovenia; Spain; Sweden; Switzerland; United Kingdom. ### Appendix B — variable construction and data sources (selected) - Participation rates: share of relevant population group. Sources: OECD Employment database; Eurostat; National authorities; Barro-Lee. - Cyclical position: output gap (robustness: unemployment rate). Source: IMF, World Economic Outlook database. - Exposure to technological progress: interaction between cross-country average relative price of investment and country exposure to routinization from initial occupational mix (Autor and Dorn (2013) routinization scores). - Trade openness: sum of exports and imports in percent of GDP. Source: IMF, WEO. - Structural transformation: ratio of employment in services relative to industrial employment. Sources: World Bank WDI; EU analyses. - Education: share with highest level reported as primary, secondary, tertiary. Source: Barro-Lee. - Labor tax wedge: ratio between average tax paid by a single-earner family and total labor cost for employer. Sources: OECD Tax database; Bassanini and Duval (2006); IMF (2016b). - Unemployment benefit generosity: gross replacement rate measures, interpolated where needed. Source: OECD Benefits and Wages: Statistics. - ALMP spending: ALMP spending per unemployed person in percent of GDP per capita. Source: OECD Employment database. - Restrictiveness of migration policy: index cumulating major post-entry/integration policy changes since 1980; higher = more restrictive. Source: DEMIG POLICY database. - Wage-setting coordination index: runs from 1 to 5, higher = more centralized bargaining. Source: Amsterdam Institute for Advanced Labour Studies database. - Family reconciliation proxies: public spending on childcare and education as percent of GDP; share part-time employment; job-protected maternity leave (weeks). Sources: OECD Social protection database; OECD Employment and Family databases. - Retirement incentives: statutory retirement age; old-age and incapacity spending as percent of GDP purged of cyclical/demographic factors; alternative measures include implicit tax on continued work and aggregate replacement ratio. Sources: Social Security Programs throughout the World; OECD; Duval (2003); IMF (2016b); Luxembourg Income Study. *Source: IMF working paper (sections 4.1, 4.2, 5.3, Appendices A–B) extracted from wp18150.* ### 4.1 Aggregate Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ### 4.1 Aggregate Analysis ### Demographic context and scope - Population growth in advanced economies (AEs) is slowing, life expectancy is rising, and the number of elderly is increasing steeply; United Nations projections imply that by the middle of this century total population will be shrinking in almost half of AEs and individuals of what is currently considered working age will be supporting close to double the number of elderly that they do now. - Analysis periods and samples: - Aggregate data: past three and a half decades; long-term trends discussed for 1985–2016 based on a sample of 21 AEs for consistency; patterns qualitatively identical when all AEs are included. - Individual-level data: 24 European countries during 2000–16. - Figures and tables draw on samples of 33 AEs and 36 AEs in various displays. ### Key aggregate patterns of labor force participation - Aggregate participation: - The aggregate average labor force participation rate in AEs as a group barely changed over the past 30 years. - Distribution narrowed: several countries saw aggregate participation gains of more than five percentage points (examples cited include Germany, Korea, Spain, and the Netherlands). - Gender dynamics: - The share of women employed or actively looking for work increased by close to 10 percentage points over the past three decades. - Prime-age male participation rates declined almost everywhere; for the median AE, the participation rate among men was more than 6 percentage points lower in 2016 than in 1985. - Divergent trends narrowed gender gaps as female gains offset male declines among prime-age workers. - Age dynamics: - Youth (ages 15–24) are significantly less likely to be part of the labor force in 2016 than in 1985; increased school enrollment and investment in human capital are important contributors. - The share of “idle” youth (neither employed, unemployed, nor enrolled in school) is small and has been stable since the early 2000s. - Older workers (ages 55 and over) increased participation significantly since the mid 1990s, especially ages 55–64; in the past decade even those older than 65 remained in the labor force longer. - Life expectancy changes since 1985: life expectancy at birth increased by about seven years and at age 50 by over five years. - Country-specific notes: - The United States is a notable exception: prime-age male participation decline was particularly steep in the past decade; prime-age female participation plateaued in the late 1990s and declined since the Global Financial Crisis. - Labor force and population composition (2015, average AE): - Labor force composition: 37 percent prime-age men, 31 percent prime-age women, 11 percent aged 15–24, 21 percent older than 55. - Population composition: 20 percent prime-age men, 20 percent prime-age women, 12 percent aged 15–24, 31 percent older than 55. ### Drivers and explanatory framework - Candidate drivers examined: - Cyclical versus structural changes. - Labor market policies and institutions (tax–benefit systems, active labor market programs, retirement policies). - Policies targeting specific groups: women, older individuals, migrants. - Technological change: automation of routinizable tasks and shifts in demand across occupations. - Shifts in population characteristics: education, fertility, marriage, immigration status. - Empirical approaches to routinization: - Cross-country approach: exploit heterogeneity in initial employment mix across occupations to test whether declines in the relative price of investment led to larger participation declines in economies with larger shares of routinizable occupations. - Individual-level approach: use current and past employment data from 24 European countries (2000–16) to estimate whether individuals in more routinizable occupations have higher likelihoods of being out of the labor force, and whether country policies attenuate that link. ### Main findings from aggregate analysis - Policies, institutions, and structural factors: - Policies and institutions (tax–benefit system, active labor market programs, policies that encourage group participation), together with structural changes and gains in educational attainment, account for the bulk of the dramatic increase in labor force attachment of prime-age women and older workers over the past three decades. - Technological change: - Technological advances—automation of tasks where labor is easily substitutable by capital—weighed on the participation rates of most groups of workers. - Mitigating factors: - Higher spending on active labor market programs and education is associated with a lower likelihood that an individual previously employed in a routinizable sector or occupation drops out of the labor force. - Urban residence is associated with a significantly lower likelihood of dropping out after prior employment in routinizable occupations, highlighting the role of access to diverse employer pools. - Heterogeneity across subgroups: - Female participation gains are widespread across marital status, parental status (young children below age 5 and older children below age 15), nativity (natives and immigrants), and educational attainment—with the notable exception of relatively low-educated women where gains are less pronounced. - Prime-age male participation declines were deepest among those with the lowest educational attainment. - Welfare and socioeconomic consequences: - Detachment from the labor force during peak productive years is associated (in cited literature) with lower happiness and life satisfaction for men, poorer health and higher mortality, and depressed employment prospects. ### Analytical implications for policy design - To enable those willing to work and counteract aging pressures, policies should consider: - Strengthening tax–benefit incentives that encourage labor force participation. - Expanding and targeting active labor market programs (ALMPs) and education to improve resilience to automation-induced displacements. - Policies to encourage participation of specific groups—women, older workers, and migrants—given their sizable contribution to aggregate participation trends. - Promoting geographic mobility and urban access to diverse employer pools to facilitate reemployment after displacement. - Assessment priorities: - Identify and rank drivers of participation across population groups to prioritize policy levers. - Monitor how automation and structural transformation interact with national policies to influence participation outcomes. *Source: IMF working paper (section 4.1 Aggregate Analysis, extracted from wp18150).* ### 1. Men by educational attainment ### 1. Men by educational attainment ### Empirical snapshots and subgroup patterns (Figures 5 and related panels) - Reported cross-sectional comparisons focus on years 2000 and 2016; in panels 4 and 8 dark bars show data for 2004 instead of 2000. - Sample construction and coverage: - Reported statistics are estimated from a random sample of 10,000 respondents per country per year. - Panels 1 and 5 are based on data from most AEs, while panels 2 to 4 and 6 to 8 are based on data from advanced European economies. - Panels 3 and 7 report statistics for married individuals. Young children are those below the age of 5; older children are those age 5-15. - Age groups analyzed in Figures 5 and related text: - All, age 15-24 (2000 and 2016) - Men, age 25-54 (2000 and 2016) - Women, age 25-54 (2000 and 2016) - All, age 55+ (2000 and 2016) - Labor market inactivity reasons categorized in survey responses: - Students - Retired (includes early retirement) - Dismissed - Family or childcare - Illness or disability - Temporary contract ended - Other - Have never worked - Key descriptive findings: - Comparing prime-age men and prime-age women reveals important gender differences in reasons for inactivity: women are more likely to drop out to look after children; a higher fraction of men report illness and disability. - A non-negligible share of inactive individuals may be “involuntarily inactive”; those reporting being dismissed from their last job provide a lower-bound indicator for this group. - Over time (2000 to 2016), the share of students increased among both the young and the prime-age; the share in (early) retirement among the prime-age fell; the share who never worked before fell among prime-age women and those over 55; illness and disability became relatively more important as reasons for non-participation. ### Sectoral and occupational concentration of involuntary inactivity (Figure 6) - Sectoral concentration: - Wholesale and retail trade, manufacturing, mining and quarrying, and utilities together account for over half of the involuntary inactive, even though less than one third of active workers (including employed and unemployed) are attached to these sectors. - Relationship to routinization: - Excess involuntary inactivity—measured as the difference between the inactive individuals attached to a sector as a share of all nonparticipants and the active workers attached to the same sector as a share of the labor force—tends to be concentrated in sectors with a greater share of routine jobs vulnerable to automation. - Occupational and income distribution patterns: - Displacement of workers tends to occur disproportionately among lower and middle-skilled occupations. - Vulnerability to routinization is especially pronounced in the middle and lower parts of the income distribution. - Occupational categories referenced in the analysis: - Managers; Professionals; Technicians; Clerical support; Craft and related trades; Plant and machine operators; Agricultural workers; Elementary occupations; Services and sales. - Sector list referenced in panel 1: - Accommodation and food service; Administrative, health, education, other; Agriculture, forestry, and fishing; Construction; Information, finance, real estate, professional; Manufacturing; Mining and quarrying, utilities; Transport and storage; Wholesale and retail trade. ### Drivers of labor force participation: conceptual framing (Section 3) - Individual determinants emphasized: - Gender, educational attainment, previous occupation, household structure — these determine potential market earnings relative to household work. - Policy, institutional, and non-economic determinants: - Tax-benefit system (labor tax wedge) - Active labor market programs (including for migrants) - Wage-setting institutions (union density, degree of coordination in wage setting) - Policies reconciling work and household (public spending on childcare and education as a share of GDP; proportion of employees working part-time; number of weeks of job-protected maternity leave) - Retirement incentives (statutory retirement age; generosity of pension schemes) - Migration integration policies (working rights, access to language and activation programs) - Cultural attitudes toward roles in society - Mechanisms illustrated: - An increase in the labor tax wedge could reduce incentives to work by reducing net wages and suppressing labor demand, but the net effect on labor supply is ambiguous because individuals may work more to maintain income. - Childcare provision and family-friendly flexible work arrangements raise the opportunity cost of staying home and can increase labor supply on the extensive margin. - Pension-system financial incentives and social transfers are important for older workers’ retirement decisions. - Long-lasting shifts in demand for skills (e.g., growth of services) can raise female participation; technological progress enabling automation may reduce demand for less-skilled labor and make certain jobs obsolete, with adjustment impeded by lack of skills, relocation costs, or low perceived returns. ### Empirical strategy and estimation (Section 4 and 4.1) - Two complementary empirical approaches: - Cross-country panel regressions to disentangle the influence of labor policies, technology, and other factors on participation of different population segments. - Individual-level data analysis from 24 European economies to examine individual characteristics (including past occupation routinization) and workforce attachment. - Aggregate analysis coverage: - Historical relationship examined since 1980 across 23 economies classified as AEs for the entire period. - Reduced-form specification estimated (notation preserved from source): - LFP_g,i,t = β_{X,g} X_g,i,t + β_{D,g} D_i,t + β_{GAP,g} GAP_{i,t−1} + β_{Z,g} Z_i,t + π_{g,i} + τ_{g,t} + ε_{g,i,t} - LFP denotes the participation rate of worker group g in country i in year t. - X: set of policies and institutions (group-specific where applicable). - D: factors shifting demand for worker group g (exposure to technological progress, size of service sector, urbanization). - GAP: cyclical position proxied by the output gap (included with one-year lag). - Z: other determinants of labor supply (e.g., educational attainment). - π_i and τ_t are country- and time-fixed effects. - Measurement and novel indices: - Exposure to technological progress measured as interaction between a country’s exposure to automation (time-invariant country-level score computed from initial occupational mix using Autor and Dorn (2013) routinization scores) and the relative price of investment (average relative price of investment across AEs). The index captures that declines in the relative price of capital, driven by global technological progress, induce substitution of labor for capital and have larger labor market consequences in countries with higher shares of routine-occupational employment. - Migration policy restrictiveness index constructed using DEMIG POLICY database; major changes in post-entry rights and migrant integration policies are cumulated starting in 1980, with a higher value denoting more restrictive policies. - Estimation details and inference: - Time series of participation rates are trend stationary per panel unit root tests. - Some endogenous variables (output gap, trade openness) included with a one-year lag. - Standard errors corrected using Driscoll and Kraay (1998) to account for cross-country dependence, autocorrelation, and heteroskedasticity. - Lagrange Multiplier tests indicate serial correlation; modified Wald test indicates group-wise heteroskedasticity; Pesaran and Frees tests reject cross-sectional independence. - Decomposition of contributions: - Contributions of each regressor to changes in participation between t and t′ are computed as C_{S,g,i,t,t′} = β̂_{S,g} (S_{g,i,t′} − S_{g,i,t}) with S = [X, D, GAP, Z]. - Caveat on inference: - Analysis seeks to identify patterns and correlations rather than establish causality. Changes in policies may reflect underlying societal and cultural shifts (e.g., evolving gender norms) that also influence labor supply. *Source: Authors' calculations and associated figures and notes from the original content unit.* ### 4.2 Individual-Level Analysis ### 4.2 Individual-Level Analysis ### Microdata, sample, and empirical approach - Data: European Labour Force Survey for 24 AEs during 2000–16; random sample of 10,000 individuals per country per year (analysis based on a random sample of 10,000 respondents per country per year from 19 countries for the logit regressions reported). - Empirical model: Logit model - Φ(Sj = 1) = βC Ci + βH Hj + βR Rj + πc + υr + τt + ϵj - Sj: dummy for individual j being in or out of the labor force. - C: individual characteristics (age, gender for age 55+, born in country, urban/rural, highest education: up to lower secondary, upper secondary, tertiary). - H: family composition (number of children, number of other employed adults, household type baseline = single adult without children; other categories: single adult with children, couple without children, couple with children). - R: routinization score of current or last occupation. - Fixed effects: πc (country), υr (region), τt (year). - Standard errors clustered at country-year level. ### Aggregate cross-country findings (summary of key associations) - Education - An increase in the share of workers with secondary and especially tertiary education is associated with significantly higher participation, particularly for prime-age women and older workers. - Business cycle and structural change - Participation rates depend on the state of the business cycle, with stronger associations for the young and women. - A relative increase in service sectors employment is typically followed by the entry of prime-age women into the labor force. - Urbanization brings gains in participation across groups. - Technology and routinization - A decline in the relative price of investment is associated with lower participation rates in countries where the initial occupation mix is tilted toward routine-task occupations. - Technological progress and differential exposure to routine occupations are linked to declines in labor share and employment losses in local labor markets. - Labor market programs and institutions - Tax-benefit system: Higher labor tax wedges and more generous unemployment benefits are associated with lower labor force attachment for most groups. - Unemployment benefit generosity (gross benefit replacement rate) is negatively related to participation, consistent with potential discouragement effects and pathways to early retirement. - Active labor market policies and migrant integration: - Higher public spending on active labor market programs (ALMP) tends to raise the share of young and prime-age women working or seeking employment. - Policies encouraging migrant integration are associated with stronger prime-age participation, especially for women. - Family reconciliation policies: - Better access to childcare, longer maternity leave, and greater flexibility in work arrangements are associated with higher female labor force participation. - Retirement incentives: - Raising statutory retirement age is associated with delayed exit from the labor market. - Greater pension scheme generosity is associated with earlier retirement. - Wage-setting institutions: - Higher coordination of wage setting is associated with greater participation for most groups; union density effects are less robust. - Decomposition 1995–2011 (text summary) - Supportive policies and educational gains drove large increases in participation of prime-age women and older workers. - Structural transformation contributed positively for these groups. - Technological advances weighed on participation for all groups except the young. - For young and some prime-age male workers, a large share of participation decline is attributed to a common component across AEs (time effects) possibly reflecting global forces (technological progress, globalization), concurrent policy changes, structural shifts, changing returns to education, rising life expectancy, or scars from the global financial crisis. - For older workers, positive common component may reflect delayed retirement due to suppressed returns on retirement savings (falling global interest rates), losses in financial wealth, and potentially higher indebtedness. ### Individual-level (micro) estimation results — key findings - Education and demographics (from logit exponentiated coefficients, Table 2) - Upper Secondary Education (odds ratios): - Men, ages 25–54: 1.719*** (standard error 0.032) - Women, ages 25–54: 1.709*** (0.033) - All, ages 55+: 1.209*** (0.036) - Tertiary Education (odds ratios): - Men, ages 25–54: 2.759*** (0.082) - Women, ages 25–54: 2.961*** (0.077) - All, ages 55+: 1.594*** (0.059) - Age: - Age odds ratios: Men 1.158*** (0.011); Women 1.320*** (0.014); All 55+ 1.396*** (0.113) - Age Squared: - Men 0.998*** (0.000); Women 0.997*** (0.000); All 55+ 0.998*** (0.000) - Born in Country: - Men 1.489*** (0.035); Women 1.333*** (0.024); All 55+ 1.091** (0.046) - Urban: - Men 1.008 (0.019); Women 1.024* (0.013); All 55+ 1.019 (0.027) - Family composition effects (odds ratios) - Number of Children in Household: - Men 1.049*** (0.009); Women 0.816*** (0.006); All 55+ 0.960* (0.020) - One Adult with Children: - Men 1.042 (0.059); Women 0.846*** (0.026); All 55+ 1.785*** (0.394) - Couple without Children: - Men 1.356*** (0.035); Women 0.906*** (0.034); All 55+ 0.842*** (0.025) - Couple with Children: - Men 1.726*** (0.052); Women 0.757*** (0.028); All 55+ 1.446*** (0.128) - Other Household Structure: - Men 0.937** (0.027); Women 0.868*** (0.030); All 55+ 0.812*** (0.038) - Other Employed Adult(s) in Household: - Men 1.497*** (0.035); Women 1.152*** (0.038); All 55+ 1.703*** (0.091) - Routinization (RTI) of occupation (odds ratios) - RTI Score of Occupation: - Men 0.825*** (0.010); Women 0.900*** (0.010); All 55+ 0.716*** (0.013) - Interpretation: A unit change in routinization score roughly corresponds to the difference in routinization score between technicians and managers; differences in RTI can explain substantial parts of observed participation differentials (example: about 87 percent of prime-age male managers active vs about 84 percent of prime-age male technicians — the RTI score difference can explain about one third of this 3 percentage-point difference). - Business cycle (Lagged Output Gap) - Lagged Output Gap odds ratios: - Men 1.037*** (0.006); Women 1.023*** (0.004); All 55+ 1.031*** (0.007) - Sample sizes for micro regressions (Table 2) - Number of Observations: Men 491,820; Women 474,240; All 55+ 86,441 - Notes on interpretation: - Findings are associations rather than fully causal estimates; family composition and labor supply decisions may be jointly determined. - Country-by-country estimates confirm negative and significant effects of routinization in most countries. ### Can policies offset routinization effects? (interaction analysis and magnitudes) - Approach: Augment logit with interaction between individual routinization score and country-level policy measures; report δy/δx (effect of a unit change in routinization score) estimated at 25th and 75th percentiles of policy distributions. - Main result: Policies can attenuate the negative association between routinization and participation, notably ALMP spending. The negative association is about one-third as large in countries at the 75th percentile of ALMP spending versus the 25th percentile. - Table 3: Effects of policies on relationship between participation and routinization (δy/δx and standard errors) - ALMP spending: - Men, ages 25–54: 25th -0.022 (Std. error 0.002); 75th -0.011 (0.001) - Women, ages 25–54: 25th -0.015 (0.002); 75th -0.008 (0.001) - ALMP spending on training: - Men: 25th -0.012 (0.003); 75th -0.016 (0.002) - Women: 25th -0.014 (0.004); 75th -0.008 (0.003) - Education spending: - Men: 25th -0.016 (0.001); 75th -0.018 (0.001) - Women: 25th -0.017 (0.002); 75th -0.010 (0.002) - Product market deregulation: - Men: 25th -0.019 (0.002); 75th -0.012 (0.002) - Women: 25th -0.011 (0.003); 75th -0.014 (0.003) - Wage-setting coordination: - Men: 25th -0.021 (0.001); 75th -0.013 (0.001) - Women: 25th -0.015 (0.002); 75th -0.010 (0.002) - Employment protection: - Men: 25th -0.019 (0.001); 75th -0.013 (0.001) - Women: 25th -0.013 (0.002); 75th -0.009 (0.002) - Location interactions: - Rural: Men -0.019 (0.001); Women -0.017 (0.002) - Urban: Men -0.016 (0.001); Women -0.012 (0.001) - Additional points - Disaggregated ALMP data suggest training spending drives mitigation of negative routinization effects for prime-age women. - Policies provide less offset for older workers; negative routinization effects are larger for older cohorts and harder to mitigate. ### Policy-relevant implications (drawn from the empirical findings) - Strengthening education (secondary and tertiary) is strongly associated with higher labor force participation, notably for prime-age women and older workers. - Active labor market programs, particularly training, can attenuate adverse participation effects from routinization, especially for prime-age women. - Family reconciliation policies (childcare, maternity leave, flexible work) support higher female participation. - Retirement incentives matter: raising statutory retirement age delays exit; higher pension generosity encourages early retirement. - Labor taxation and unemployment benefit generosity are associated with lower participation and may discourage labor force attachment. - Wage-setting coordination is associated with higher participation; union density effects are less robust across specifications. - Policy effectiveness varies by age: older workers are more vulnerable to routinization and less responsive to policy offsets. *Source: IMF Working Paper — section 4.2 "Individual-Level Analysis" (authors’ calculations reported in Tables 1–3).* ### 5.3 Robustness ### wp18150 - 5.3 Robustness ### Robustness checks and alternative specifications - The cross-country panel regression results reported in Tables 4–8 are tested for robustness to: - Applying the logistic transformation to the dependent variable (column (2)). - Accounting for cross-equation correlation via SUR (column (3)). - Correcting standard errors for cross-sectional dependence and other issues using Beck and Katz (1995) (column (4)), HAC standard errors (column (5)), and Newey–West standard errors (column (6)). - Collapsing the dataset to five-year averages (column (7)). - Excluding global financial crisis years 2008 and 2009 (column (8)). - Including additional advanced economies (column (9)). - Replacing the lag of the output gap with the lag of the unemployment rate (column (10)). - Sample selection checks by dropping one country at a time and reporting median and 10th/90th percentiles (column (11)). - Standard errors and reporting conventions: - Driscoll–Kraay standard errors reported in parentheses in columns (1), (2), (7)–(11) in many tables. - Bootstrapped standard errors reported in column (3) for SUR. - HAC standard errors assuming a panel-dependent correlation structure reported in column (4) in many tables. - Column (11) reports the 10th and 90th percentile of coefficients in parentheses. - General outcome of robustness checks: - Results are broadly robust across these alternative measures, specifications, and error-structure corrections. - Replacing the output gap with the unemployment rate returns qualitatively comparable results, although prime-age men and older workers’ participation rates can be sensitive to this cyclical measure. - Dropping 2008–2009 does not materially change the significance or magnitude of coefficients. - SUR estimation yields similar results where estimated (aggregate participation SUR not estimated). ### Robust determinants by worker group (selected baseline coefficients and robustness) - Youth (Ages 15–24) — selected baseline coefficients (Table 4, column (1)): - Lag of Output Gap: 0.360*** - Routinization*Relative Price of Investment: 0.303 - Lag of Trade Openness: 0.059*** - Lag of Urbanization: 0.668*** - Education (% Tertiary): -0.275*** - Tax wedge: -0.103 - Public Spending on ALMP: 0.041*** - Restrictiveness of Migrant Integration Policies: 0.491*** - Coordination of Wage-setting: 1.104*** - Observations: 571; Countries: 23; R-squared: 0.515 - Prime-age Men (Ages 25–54) — selected baseline coefficients (Table 5, column (1)): - Lag of Output Gap: 0.072*** - Routinization*Relative Price of Investment: 0.302*** - Lag of Trade Openness: -0.005 - Lag of Urbanization: 0.101*** - Education (% Secondary): 0.019*** - Unemployment Replacement Ratio: -0.041*** - Restrictiveness of Migrant Integration Policies: -0.047** - Coordination of Wage-setting: 0.131** - Observations: 571; Countries: 23; R-squared: 0.606 - Prime-age Women (Ages 25–54) — selected baseline coefficients (Table 6, column (1)): - Lag of Output Gap: 0.170* - Routinization*Relative Price of Investment: 1.793*** - Relative Service Employment: 0.015*** - Lag of Urbanization: 0.355*** - Education (% Secondary): 0.211*** - Education (% Tertiary): 0.332*** - Tax wedge: -0.129*** - Public Spending on ALMP: 0.039*** - Restrictiveness of Migrant Integration Policies: -0.462*** - Public Spending on Early Childhood Education and Care: 3.708*** - Share of Part-Time Employment: 0.946*** - Job-Protected Maternity Leave: 0.025*** - Observations: 489; Countries: 23; R-squared: 0.887 - Older Workers (Ages 55+) — selected baseline coefficients (Table 7, column (1)): - Lag of Output Gap: -0.006 - Routinization*Relative Price of Investment: 0.505* - Lag of Trade Openness: -0.059*** - Education (% Tertiary): 0.389*** - Tax wedge: -0.263*** - Public Spending on ALMP: -0.025** - Union Density: -0.115*** - Statutory Retirement Age: 0.661*** - Public Spending on Old-Age Pension: -0.750*** - Observations: 489; Countries: 23; R-squared: 0.887 - Aggregate Participation Rate — selected baseline coefficients (Table 8, column (1)): - Lag of Output Gap: 0.183*** - Routinization*Relative Price of Investment: 0.536*** - Lag of Trade Openness: 0.012* - Relative Service Employment: 0.010** - Lag of Urbanization: 0.249*** - Education (% Secondary): 0.063*** - Education (% Tertiary): 0.135*** - Tax wedge: -0.240*** - Unemployment Replacement Ratio: -0.078*** - Public Spending on ALMP: 0.031*** - Restrictiveness of Migrant Integration Policies: -0.207*** - Observations: 570; Countries: 23; R-squared: 0.578 Notes on significance notation preserved: - ***, **, and * indicate statistical significance at 1, 5, and 10 percent, respectively (as reported). ### Individual-level robustness (Table 9) — determinants of being in the labor force (exponentiated logit coefficients) - Men, ages 25–54 (column (1)): - Age: 1.261*** - Age Squared: 0.997*** - Upper Secondary Education: 1.737*** - Tertiary Education: 2.217*** - Born in Country: 1.761*** - Urban: 0.896*** - Number of Children in Household: 1.094*** - Couple with Children: 2.141*** - Other Employed Adult(s) in Household: 0.992 - RTI Score of Occupation: 0.467*** - Lagged Output Gap: 1.042*** - Predicted Income Decile: 0.952*** - Number of Observations: 474,434 - Women, ages 25–54 (column (2)): - Age: 1.347*** - Age Squared: 0.997*** - Upper Secondary Education: 1.855*** - Tertiary Education: 2.763*** - Born in Country: 1.520*** - Urban: 0.864*** - Number of Children in Household: 0.869*** - One Adult with Children: 0.846*** - Couple with Children: 1.248*** - Other Employed Adult(s) in Household: 0.601*** - RTI Score of Occupation: 0.490*** - Lagged Output Gap: 1.030*** - Predicted Income Decile: 0.950*** - Number of Observations: 443,687 - All, ages 55+ (column (3)): - Age: 1.356*** - Age Squared: 0.998*** - Upper Secondary Education: 1.102** - Tertiary Education: 1.240*** - Born in Country: 1.167** - Urban: 0.866*** - Number of Children in Household: 1.039 - Couple with Children: 2.429*** - Other Employed Adult(s) in Household: 0.636*** - RTI Score of Occupation: 0.488*** - Lagged Output Gap: 1.037*** - Predicted Income Decile: 0.952*** - Number of Observations: 63,982 - Notes on individual-level robustness: - Including actual or predicted income decile changes some effects: once included, the effect of being part of a couple and having children on women’s participation turns positive, the effect of other employed adults in the household turns negative, and income itself has a negative effect (Predicted Income Decile exponentiated coefficient 0.952***). - Results on vulnerability to routinization (RTI Score of Occupation) and education remain very similar to baseline. - Results robust to inclusion of interacted country-year fixed effects in addition to region fixed effects. ### Key substantive findings and policy implications (preserving language and numeric emphasis) - Heterogeneity across groups: - Technological advances (automation / routinization) weighed on labor supply for most groups, with particularly large baseline coefficients for prime-age women (1.793***) and substantial effects for prime-age men (0.302***), older workers (0.505*), and the aggregate (0.536***). - Educational attainment matters heterogeneously: tertiary education is associated with lower youth participation (Education (% Tertiary): -0.275***), but higher participation among prime-age women (0.332***), older workers (0.389***), and aggregate participation (0.135***). - Role of policies and institutions: - Public spending on ALMP has positive baseline associations with participation for youth (0.041***), prime-age women (0.039***), and aggregate participation (0.031***), but negative for older workers (-0.025**). - Public spending on early childhood education and care has a large positive association with prime-age female participation (3.708***). - Job-protected maternity leave and higher share of part-time employment are positively associated with prime-age female participation (0.025*** and 0.946***, respectively). - Restrictive migrant integration policies show mixed associations: strongly positive for youth (0.491***) but strongly negative for prime-age women (-0.462***) and negative for aggregate participation (-0.207***); prime-age men show a small negative effect (-0.047**); older workers show an insignificant baseline coefficient (0.056). - Statutory retirement age (0.661***) and public spending on old-age pension (-0.750***) are strong correlates of older worker participation. - Individual-level evidence: - Higher vulnerability to routinization (RTI Score) is associated with lower odds of being in the labor force (exponentiated coefficients 0.467*** men, 0.490*** women, 0.488*** ages 55+). - Predicted income decile is associated with lower odds of being in the labor force (0.952***), suggesting upper-income deciles may be able to afford labor force detachment. - Policy recommendations emphasized (preserved text): - Further investment in education, training, and activation policies can encourage labor market activity and increase workforce resilience to technological progress and globalization. - Policies that reduce disincentives for joining or remaining in the labor force and that help workers combine family and work life can broaden participation gains. - Rethinking immigration policies to boost labor supply and policies encouraging older workers to postpone retirement are highlighted as important for many advanced economies, noting possible political challenges of immigration but also potential population-growth and labor-supply benefits. *Source: Authors’ calculations, wp18150 - 5.3 Robustness (IMF working paper).* ### References ### References ### Bibliographic citations - Abraham, Katharine G and Melissa S Kearney (2018). “Explaining the Decline in the US Employment-to-Population Ratio: A Review of the Evidence”. In: National Bureau of Economic Research Working Paper No. 24333. - Acemoglu, Daron and David Autor (2011). “Skills, Tasks and Technologies: Implications for Employment and Earnings”. In:Handbook of Labor Economics. Vol. 4. Elsevier, pp. 1043–1171. - Acemoglu, Daron and Pascual Restrepo (2017). “Robots and Jobs: Evidence from US Labor Markets”. In: National Bureau of Economic Research Working Paper No. 23285. - Aguiar, Mark, Mark Bils, Kerwin Kofi Charles, and Erik Hurst (2017). “Leisure Luxuries and the Labor Supply of Young Men”. In: National Bureau of Economic Research Working Paper No. 23552. - Aiyar, Shekhar, Bergljot Bjørnson Barkbu, Nicoletta Batini, Helge Berger, Enrica Detragiache, Allan Dizioli, Christian Ebeke, Huidan Lin, Linda Kaltani, Sebastian Sosa, Antonio Spilimbergo, and Petia Topalova (2016). “The Refugee Surge in Europe”. In: International Monetary Fund, Staff Discussion Note 16/02. - Arulampalam, Wiji, Alison L Booth, and Mark P Taylor (2000). “Unemployment Persistence”. In: Oxford Economic Papers52.1, pp. 24–50. - Arulampalam, Wiji, Paul Gregg, and Mary Gregory (2001). “Unemployment Scarring”. In:The Economic Journal111.475, pp. 577–584. - Autor, David H and David Dorn (2013). “The Growth of Low-Skill Service Jobs and the Polarization of the US Labor Market”. In:American Economic Review103.5, pp. 1553–97. - Autor, David H, David Dorn, and Gordon H Hanson (2016). “The China Syndrome: Local Labor Market Effects of Import Competition in the United States”. In:American Economic Review 103, pp. 2021–68. - Banerji, Angana, Ms Hannah Huidan Lin, and Mr Sergejs Saksonovs (2015). “Youth Unemployment in Advanced Europe: Okun’s Law and Beyond”. In: International Monetary Fund Working Paper 16/160. - Baron-Cohen, Simon, Rebecca C Knickmeyer, and Matthew K Belmonte (2005). “Sex Differences in the Brain: Implications for Explaining Autism”. In:Science310.5749, pp. 819–823. - Bassanini, Andrea and Romain Duval (2006). “Employment Patterns in OECD Countries”. In: OECD Economic Department Working Paper 486, Organisation for Economic Co-operation and Development, Paris, France. - — (2009). “Unemployment, Institutions, and Reform Complementarities: Re-assessing the Aggregate Evidence for OECD Countries”. In:Oxford Review of Economic Policy25.1, pp. 40–59. - Bertola, Giuseppe, Francine D Blau, and Lawrence M Kahn (2007). “Labor Market Institutions and Demographic Employment Patterns”. In:Journal of Population Economics20.4, pp. 833–867. - Blanchard, Olivier and Justin Wolfers (2000). “The Role of Shocks and Institutions in the Rise of European Unemployment: The Aggregate Evidence”. In:The Economic Journal110.462, pp. 1–33. - Blau, David M and Ryan M Goodstein (2010). “Can Social Security Explain Trends in Labor Force Participation of Older Men in the United States?” In:Journal of Human Resources45.2, pp. 328–363. - Blau, Francine D and Lawrence M Kahn (2013). “Female Labor Supply: Why Is the United States Falling Behind?” In:American Economic Review103.3, pp. 251–56. - Bl ̈ondal, Sveinbj ̈orn and Stefano Scarpetta (1999). “The Retirement Decision in OECD Countries”. In: OECD Economics Department Working Paper 202, Organisation for Economic Co-operation and Development, Paris. - Blundell, Richard, Eric French, and Gemma Tetlow (2016). “Retirement Incentives and Labor Supply”. In: 1, pp. 457–566. - Blundell, Richard and Thomas MaCurdy (1999).Labor Supply: A Review of Alternative Approaches. Vol. 3. Elsevier, pp. 1559–1695. - B ̈orsch-Supan, Axel and Irene Ferrari (2017). “Old-Age Labor Force Participation in Germany: What Explains the Trend Reversal among Older Men? And What the Steady Increase among Women?” In: National Bureau of Economic Research Working Paper No. 24044. - Card, David, Jochen Kluve, and Andrea Weber (2010). “Active Labour Market Policy Evaluations: A Meta-Analysis”. In:The Economic Journal120.548. - Carrasco, Raquel, Juan F Jimeno, and A Carolina Ortega (2008). “The Effect of Immigration on the Labor Market Performance of Native-Born Workers: Some Evidence for Spain”. In:Journal of Population Economics21.3, pp. 627–648. - Case, Anne and Angus Deaton (2017). “Mortality and Morbidity in the 21st Century”. In: Brookings Papers on Economic Activity. - Cattaneo, Cristina, Carlo V Fiorio, and Giovanni Peri (2015). “What Happens to the Careers of European Workers When Immigrants “Take Their Jobs”?” In:Journal of Human Resources 50.3, pp. 655–693. - CEA (2016). “The Long-Term Decline in Prime-Age Male Labor Force Participation”. In: Council of Economic Advisers, Executive Office of the President of the United States. - Christiansen, Lone Engbo, Huidan Lin, Joana Pereira, Petia Topalova, and Rima Turk (2016). Gender Diversity in Senior Positions and Firm Performance. International Monetary Fund. - Cipollone, Angela, Eleonora Patacchini, Giovanna Vallanti, et al. (2013). “Women Labor Market Performance in Europe: Novel Evidence on Trends and Shaping Factors”. In: IZA Discussion Papers 7710, IZA Institute of Labor Economics, Bonn, Germany. - Clements, Benedict J, Kamil Dybczak, Vitor Gaspar, Sanjeev Gupta, and Mauricio Soto (2015). “The Fiscal Consequences of Shrinking Populations”. In: Staff Discussion Note, 15/21, International Monetary Fund. - Cortes, Guido Matias, Nir Jaimovich, and Henry E Siu (2018). “The “End of Men” and Rise of Women in the High-Skilled Labor Market”. In: National Bureau of Economic Research Working Paper No. 24274. - D’Amuri, Francesco and Giovanni Peri (2014). “Immigration, Jobs, and Employment Protection: Evidence from Europe Before and During the Great Recession”. In:Journal of the European Economic Association12.2, pp. 432–464. - Dao, Mai Chi, Ms Mitali Das, Zsoka Koczan, and Weicheng Lian (2017). “Why Is Labor Receiving a Smaller Share of Global Income? Theory and Empirical Evidence”. In: International Monetary Fund Working Paper 17/169. - Dao, Mai, Davide Furceri, Tae-Jeong Kim, Meeyeon Kim, and Jisoo Hwang (2014). “Strategies for Reforming Korea’s Labor Market to Foster Growth”. In: International Monetary Fund Working Paper 14/137. - De Resende, Carlos (2014). “An Assessment of IMF Medium-Term Forecasts of GDP Growth”. In: IEO Background Paper No. BP/14/01 (Washington: Independent Evaluation Office of the IMF). - De Serres, Alain, Fabrice Murtin, and Christine De La Maisonneuve (2012). “Policies to Facilitate the Return to Work”. In:Comparative Economic Studies54.1, pp. 5–42. - Driscoll, John C and Aart C Kraay (1998). “Consistent Covariance Matrix Estimation with Spatially Dependent Panel Data”. In:Review of Economics and Statistics80.4, pp. 549–560. - Duval, Romain (2003). “The Retirement Effects of Old-Age Pension and Early Retirement Schemes in OECD Countries”. In:OECD Economic Studies2, pp. 7–50. - — (2004). “Retirement Behaviour in OECD Countries”. In:OECD Economic Studies2003.2, pp. 7–50. - Eberstadt, Nicholas (2016).Men Without Work: America’s Invisible Crisis. Templeton Foundation Press. - Eliason, Marcus and Donald Storrie (2009). “Does Job Loss Shorten Life?” In:Journal of Human Resources44.2, pp. 277–302. - Elsby, Michael WL, Bart Hobijn, and Ay ̧seg ̈ul S ̧ahin (2015). “On the Importance of the Participation Margin for Labor Market Fluctuations”. In:Journal of Monetary Economics72, pp. 64–82. - Feingold, Alan (1994). “Gender Differences in Personality: A Meta-Analysis.” In:Psychological bulletin116.3, p. 429. - Fern ́andez, Raquel (2013). “Cultural Change as Learning: The Evolution of Female Labor Force Participation over a Century”. In:American Economic Review103.1, pp. 472–500. - Foged, Mette and Giovanni Peri (2016). “Immigrants’ Effect on Native Workers: New analysis on Longitudinal Data”. In:American Economic Journal: Applied Economics8.2, pp. 1–34. - Gal, Peter and Adam Theising (2015). “The macroeconomic Impact of Policies on Labour Market Outcomes in OECD Countries: A Reassessment”. In: OECD Economic Department Working Paper 1271, Organisation for Economic Co-operation and Development, Paris, France. - Galor, Oded and David N Weil (1996). “The Gender Gap, Fertility, and Growth”. In:American Economic Review86.3, pp. 374–387. - Genda, Yuji, Ayako Kondo, and Souichi Ohta (2010). “Long-Term Effects of a Recession at Labor Market Entry in Japan and the United States”. In:Journal of Human Resources45.1, pp. 157–196. - Genre, Veronique, Ram ́on G ́omez-Salvador, and Ana Lamo (2005). “The Determinants of Labour Force Participation in the European Union”. In:Labour Supply and Incentives to Work in Europe. - Genre, V ́eronique, Ram ́on G ́omez Salvador, and Ana Lamo (2010). “European Women: Why Do (n’t) They Work?” In:Applied Economics42.12, pp. 1499–1514. - Gerdtham, Ulf-G and Magnus Johannesson (2003). “A Note on the Effect of Unemployment on Mortality”. In:Journal of Health Economics22.3, pp. 505–518. - Goos, Maarten, Alan Manning, and Anna Salomons (2014). “Explaining job polarization: Routine-biased technological change and offshoring”. In:American Economic Review104.8, pp. 2509–26. - Grigoli, Francesco, Zsoka Koczan, and Petia Topalova (2018). “A Cohort-Based Analysis of Labor Force Participation for Advanced Economies”. In: International Monetary Fund Working Paper 18/120. - Gruber, Jonathan and David Wise (2000). “Social security programs and retirement around the world”. In:Research in Labor Economics. Emerald Group Publishing Limited, pp. 1–40. - Heckman, James J, Robert J LaLonde, and Jeffrey A Smith (1999). “The Economics and Econometrics of Active Labor Market Programs”. In:Handbook of Labor Economics. Vol. 3. Elsevier, pp. 1865–2097. - Holzer, Harry J, Paul Offner, and Elaine Sorensen (2005). “Declining Employment among Young Black Less-Educated Men: The Role of Incarceration and Child Support”. In:Journal of Policy Analysis and Management24.2, pp. 329–350. - Hurd, Michael and Susann Rohwedder (2011). “Trends in Labor Force Participation: How Much is Due to Changes in Pensions?” In:Journal of Population Ageing4.1-2, pp. 81–96. - IMF (2003). “Unemployment and Labor Market Institutions: Why Reforms Pay Off”. In: World Economic Outlook, Chapter 4, International Monetary Fund. - — (2016a). “Spillovers from China’s Transition and from Migration”. In: World Economic Outlook, Chapter 4, April, International Monetary Fund. - — (2016b). “Time for a Supply-Side Boost? Macroeconomic Effects of Labor and Product Market Reforms in Advanced Economies”. In: World Economic Outlook, Chapter 3, April, International Monetary Fund. - IMF (2017a). “Recent Wage Dynamics in Advanced Economics: Drivers and Implications”. In:World Economic Outlook, Chapter 2, October, International Monetary Fund. - — (2017b). “Understanding the Downward Trend in Labor Income Shares”. In: World Economic Outlook, Chapter 3, April, International Monetary Fund. - — (2018). “Shifts in Global Manufacturing Activity”. In: World Economic Outlook, Chapter 3, April, International Monetary Fund. - IMF/WB/WTO (2017). “Making Trade an Engine of 23 Growth for All: The Case for Trade and for Policies to Facilitate Adjustment”. In: Policy Papers for Discussion at the Meeting of G20 Sherpas, March 23–24. Frankfurt, Germany. - Jaumotte, Florence (2003). “Female Labour Force Participation: Past Trends and Main Determinants in OECD Countries”. In: OECD Economics Department Working Paper 376, Organisation for Economic Co-operation and Development, Paris. - Kahn, Lisa B (2010). “The Long-Term Labor Market Consequences of Graduating from College in a Bad Economy”. In:Labour Economics17.2, pp. 303–316. - Knabe, Andreas and Steffen R ̈atzel (2011). “Scarring or Scaring? The Psychological Impact of Past Unemployment and Future Unemployment Risk”. In:Economica78.310, pp. 283–293. - Koczan, Zsoka and Benjamin Hilgenstock (2018a). “Permanently displaced? Increasingly disconnected? Labor force participation in U.S. states and metropolitan areas”. In: International Monetary Fund Working Paper (forthcoming). - — (2018b). “Still attached? Are social safety nets working? Labor force participation in European regions”. In: International Monetary Fund Working Paper (forthcoming). - — (2018c). “Storm Clouds Ahead? Migration and Labor Force Participation Rates”. In: International Monetary Fund Working Paper (forthcoming). - Krause, Eleanor and Isabel Sawhill (2017). “What We Know and Don’t Know about Declining Labor Force Participation: A Review”. In: Center on Children and Families at Brookings, The Brookings Institution, Washington, DC. - Krueger, Alan B (2017). “Where Have All the Workers Gone? An Inquiry into the Decline of the US Labor Force Participation Rate”. In:Brookings Papers on Economic Activity, pp. 7–8. - Lucas, Richard E, Andrew E Clark, Yannis Georgellis, and Ed Diener (2004). “Unemployment Alters the Set Point for Life Satisfaction”. In:Psychological Science15.1, pp. 8–13. - Murtin, Fabrice, Alain De Serres, and Alexander Hijzen (2014). “Unemployment and the Coverage Extension of Collective Wage Agreements”. In:European Economic Review71, pp. 52–66. - Ngai, L Rachel and Barbara Petrongolo (2017). “Gender Gaps and the Rise of the Service Economy”. In:American Economic Journal: Macroeconomics9.4, pp. 1–44. - Nickell, Stephen (1981). “Biases in Dynamic Models with Fixed Effects”. In:Econometrica, pp. 1417–1426. - Olivetti, Claudia and Barbara Petrongolo (2017). “The Economic Consequences of Family Policies: Lessons from a Century of Legislation in High-Income Countries”. In:Journal of Economic Perspectives31.1, pp. 205–30. - Oreopoulos, Philip, Till Von Wachter, and Andrew Heisz (2012). “The Short- and Long-Term Career Effects of Graduating in a Recession”. In:American Economic Journal: Applied Economics4.1, pp. 1–29. - Pager, Devah, Bruce Western, and Naomi Sugie (2009). “Sequencing Disadvantage: Barriers to Employment Facing Young Black and White Men with Criminal Records”. In:The ANNALS of the American Academy of Political and Social Science623.1, pp. 195–213. - Rendall, Michelle (2010). “Brain versus Brawn: The Realization of Women’s Comparative Advantage”. In: IEW Working Papers 491, University of Zurich. - Schmitt, John and Kris Warner (2010). “The Changing Face of US Labor, 1983–2008”. In:Journal of Labor and Society13.2, pp. 263–279. - Soskice, David (1990). “Wage Determination: The Changing Role of Institutions in Advanced Industrialized Countries”. In:Oxford Review of Economic Policy6.4, pp. 36–61. - Sullivan, Daniel and Till Von Wachter (2009). “Job Displacement and Mortality: An Analysis Using Administrative Data”. In:The Quarterly Journal of Economics124.3, pp. 1265–1306. - Th ́evenon, Olivier (2013). “Drivers of Female Labour Force Participation in the OECD”. In: OECD Economics Department Working Paper 145, Organisation for Economic Co-operation and Development, Paris. - UN (2017). “The Impact of the Technological Revolution on Labour Markets and Income Distribution”. In: Frontier Issues. - Winkelmann, Liliana and Rainer Winkelmann (1995). “Happiness and Unemployment: A Panel Data Analysis for Germany”. In:Applied Economics Quarterly41.4, pp. 293–307. *Appendix and methodological citations continue in the source document.* --- ### Appendix A. Sample ### Countries included in the analysis - Aggregate analysis: Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Greece, Italy, Ireland, Japan, Korea, Luxembourg, Netherlands, New Zealand, Norway, Portugal, Spain, Sweden, Switzerland, United Kingdom, United States. - Individual-level analysis: Austria, Belgium, Cyprus, Czech Republic, Denmark, Germany, Estonia, Finland, France, Greece, Iceland, Ireland, Italy, Latvia, Lithuania, Netherlands, Norway, Portugal, Slovakia, Slovenia, Spain, Sweden, Switzerland, United Kingdom. --- ### Appendix B. Variable Construction and Sources ### Variable definitions and data sources - Aggregate and group-specific labor force participation rates are expressed as a share of the relevant population group. Sources: OECD, Employment database; Eurostat; National authorities, Barro-Lee Educational Attainment dataset. - The cyclical position is captured with the output gap, while robustness tests use the unemployment rate. Source: IMF, World Economic Outlook database. - Exposure to technological progress is measured as the interaction between the cross-country average relative price of investment and the country’s exposure to routinization through its initial occupational mix. The latter consists of scores that rely on occupation-level measures, which order occupations by their share of routine tasks, and then use the employment shares of these occupations to construct country-level measures of routinizability. Sources: IMF, World Economic Outlook database, Autor and Dorn (2013), Eurostat, and Population censuses. - Trade openness is measured as the sum of exports and imports in percent of GDP. Source: IMF, World Economic Outlook database. - Potential shifts in the demand for different types of labor due to structural transformation are measured as the ratio of employment in the services sector relative to employment in the industrial sector. Sources: World Bank, World Development Indicators; European Union, Level analysis of Capital, Labour, Energy, Materials, and Service inputs. The share of urban population is also used. Sources: World Bank, World Development Indicators. - Educational attainment is measured as the share of the population within a specific age-gender group with highest level of education reported as primary, secondary, or tertiary. Source: Barro-Lee Educational Attainment dataset. - The labor tax wedge is defined as the ratio between the average tax paid by a single-earner family (one parent at 100 percent of average earnings with two children) and the corresponding total labor cost for the employer. The labor tax wedge is available from the OECD for 2000 to 2016, and was extended back to 1980 using Bassanini and Duval (2006) and IMF (2016b). The latter series is available only in uneven years; the value of the labor tax wedge in even years is obtained by linear interpolation. Sources: OECD, Tax database; Bassanini and Duval (2006); IMF (2016b). - The generosity of the unemployment benefits system is measured as the gross replacement rate, gross unemployment benefit levels as a percentage of previous gross earnings. The summary measure with the best coverage is the average of the gross unemployment benefit replacement rates for two earnings levels, three family situations, and three durations of unemployment. Such measures are available in uneven years, and are interpolated to obtain their values for even years. The reported values are for the average worker from 2001 to 2011, and average production worker from 1961 to 2005. The two series are spliced. Source: OECD, Benefits and Wages: Statistics. - Public expenditure on active labor market programs is calculated as active labor market programs spending per unemployed person in percent of GDP per capita, following Gal and Theising (2015). Source: OECD, Employment database. - Restrictiveness of migration policy is an index with information about all changes to the existing legal framework relevant for migration (see also De Resende, 2014). Focus is on major changes in policies guiding the post-entry rights or other aspects of migrants’ integration. These changes are cumulated starting 1980 to construct an index for each country, with a higher value denoting more restrictive policies. Source: International Migration Institute, DEMIG POLICY database. - Union density is measured as net union membership as a proportion of wage earners in employment. Source: OECD, Employment database. - Coordination of wage setting is an index of the centralization of bargaining. The index runs from 1 to 5 with values defined as (1) Fragmented wage bargaining, confined largely to individual firms or plants, (2) mixed industry and firm-level bargaining, weak government coordination through minimum wage setting or wage indexation, (3) negotiation guidelines based on centralized bargaining, (4) wage norms based on centralized bargaining by peak association with or without government involvement, and (5) maximum or minimum wage rates/increases based on centralized bargaining. Source: Amsterdam Institute for Advanced Labour Studies, Database on Institutional Characteristics of Trade Unions, Wage Setting, State Intervention, and Social Pacts. - Policies that help reconcile work inside and outside the household are proxied by public spending on childcare and education as a percent of GDP; the proportion of employees with a part-time contract to total employees; and job-protected maternity leave, defined as the total number of weeks of job-protected maternity, parental, and extended leave available to mothers, regardless of income support. Sources: OECD, Social protection database; OECD, Employment database; OECD, Family database. - Retirement incentives are proxied by the statutory retirement age and by the generosity of pension schemes. The measure with the best country and time coverage is old-age and incapacity spending as a percent of GDP. This measure is first purged from fluctuations due to cyclical and demographic factors (namely, share of the population in different age groups and health status, proxied by life expectancy) that may mechanically generate a negative correlation with the labor force attachment of older workers. As a robustness check, the analysis considers the implicit tax on continued work, calculated as the change in the present value of the stream of future pension payments net of contributions to the system from working five more years for typical workers at different ages. An alternative measure also considered is the aggregate replacement ratio, calculated as the ratio of the mean disposable income of ages 65–74 to the mean disposable income of ages 50–59. This variable is computed for select years based on the availability of household survey data and is interpolated for the missing years. Sources: Social Security Programs throughout the World; OECD, Social protection database; Duval (2003); IMF (2016b); Luxembourg Income Study database. *33 The cross-country average relative price of investment across all AEs is used to minimize endogeneity concerns and capture changes that are due to global technological progress (rather than, for example, country-specific capital taxation policies).* *Notes: Specific methodological details, interpolation approaches, and data coverage issues are described in the source document.* --- _Source: https://www.imf.org/-/media/files/publications/wp/2018/wp18150.pdf_