## wpiea2025184-source-pdf

## Source details

**Canonical URL:** [wpiea2025184-source-pdf](https://www.imf.org/-/media/files/publications/wp/2025/english/wpiea2025184-source-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2025/english/wpiea2025184-source-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2025/english/wpiea2025184-source-pdf.pdf.json)

---

### Introduction
- Japan’s economy is undergoing a profound transformation driven primarily by population aging and advances in artificial intelligence (AI).
- Paper objective: investigate the interplay between population aging and AI in influencing Japan's labor market using regression analysis, occupational-level AI exposure/complementarity measures, and a skill distance index.

### Stylized Facts
- Share of population aged 65 or older in 2023: 29 percent.
- Projected share of population aged 65 or older by 2070: 40 percent.
- Japan has one of the highest robot densities in manufacturing in 2022.
- Japan’s software investments from 2002 to 2022 are negatively correlated with the share of workers aged 34 and younger.
- In 2024, around half of surveyed Japanese firms indicated a lack of qualified full-time employees to meet operational needs.
- New job opening to application ratios remain elevated by historical standards.
- Key policy and institutional details:
  - Employers obliged (as of June 2025) to ensure employment for employees aged 60-65 by raising retirement age to 65, offering continued employment, or abolishing retirement age; for ages 65-70, employers obliged to make efforts to ensure employment.
  - In 2024, 64 percent of companies set the retirement age at 60.
  - Pension benefits cut if total wages plus Employees' Pension Insurance pension for employees aged 65 and over exceeds 510,000 yen (FY25 threshold) per month; this affects 16 percent of pension recipients in 2022.

### Methodology
- Three main approaches:
  - Panel regressions (sector-level with sector and time fixed effects) to assess aging’s impact on labor shortages and productivity.
    - Analysis period for shortages: 2008-2023.
    - Productivity long-run relations use 10-year averages for 2000-2010 and 2010-2021.
    - Data cover 97 sectors.
  - AI exposure and complementarity mapping at occupation level using Japan’s Labor Force Survey, following Felten, Raj and Seamans (2021, 2023) and Cazzaniga and others (2024).
    - Occupational categories: High exposure & high complementarity; High exposure & low complementarity; Low exposure.
  - Skill transferability: construct skill distance between occupation pairs using Occupational Information Network of Japan.
    - Skill distance index ranges from 0 to 1, where 0 = identical skill sets and 1 = orthogonal skill sets.
- Key regression specifications:
  - Labor shortages equation (sector i, time t): y_{i,t} = a_i + b_t + y_{i,t-1} + k x_{i,t-h} + β Z_{i,t} + u_{i,t}, where x_{i,t-h} is h-year lag of share of workers aged 15-34 or 55+, Z includes controls (new job opening rate, value-added growth, labor force growth, wage growth, change in average hours worked).
  - Productivity long-run equation: Δy_{i,t} = α + b_t + θ y_{i,t-1} + β1 X_{i,t} + β2 Z_{i,t} + u_{i,t}, where y is log labor productivity and X is share of workers aged 55+.

### Main findings
- Aging and labor shortages:
  - Sectors with lower share of employees ages 15-34 or higher share of workers 55+ tend to report more firms indicating labor shortages 5 years later.
  - New job opening rate has consistent positive association with labor shortages.
  - Labor shortages display considerable persistence, indicating structural nature and potentially low inter-sector mobility.
- Key regression coefficients and statistics on labor shortages (selected from Table 1):
  - Lagged labor shortages coefficients: 0.417*** (SE 0.0876), 0.298** (0.125), 0.618*** (0.0613), 0.263** (0.126).
  - New job opening as percent of total employment coefficients: 0.455* (0.251), 1.435*** (0.511), 0.380* (0.213), 1.352*** (0.495).
  - Share of workers age 15-34 (5 year lag) coefficients: -0.168* (0.0960), -1.119* (0.649), -0.223*** (0.0827).
  - Share of workers aged 55+ (5 year lag) coefficient: 1.618* (0.809).
  - Observations reported: 1216, 616, 656, 6.
  - R-squared values: 0.82, 0.66, 0.93, 0.30.
  - Dependent variable: Labor shortages reported by firms by sectors (percent balance).
- Aging and productivity:
  - Aging in the labor force is negatively associated with labor productivity growth (Table 2).
  - Investment rate has positive and statistically significant association with labor productivity.
  - No statistically significant link identified between labor productivity growth and share of part-time or female workers in this analysis.
- AI exposure, complementarity, and occupational patterns:
  - In 2023, share of all workers who were seniors (age 65+): 14 percent.
  - Japan has a larger share of workers with low exposure to AI compared with other G7 countries.
  - Share of workers in high exposure & low complementarity is larger compared with the United States or Germany.
  - Occupation-specific vulnerability:
    - Managers and professionals concentrate in high exposure & high complementarity occupations.
    - Clerical support workers, sales workers, telemarketers: high exposure & low complementarity (more vulnerable to displacement).
    - Service workers, security workers, manufacturing processing, packaging/carrying: lower AI exposure (though manufacturing processing is exposed to automation).
  - By gender and age:
    - Women tend to concentrate in high exposure & low complementarity occupations (higher displacement risk).
    - Age cohort 65+ is least exposed to AI; cohort 55-64 has vulnerability similar to middle-aged groups due to concentration in clerical and sales occupations.
- Sectoral detail on seniors and AI potential (Table 3):
  - Share of workers 65+ in top five sectors with most severe labor shortages ranges from 2.5 percent (ICT) to 17 percent (construction).
  - In construction and transportation services, AI may help alleviate aging-induced shortages (share of occupations with high AI exposure & low complementarity relatively low).
  - In medical and healthcare services, AI may have limited room to address shortages because seniors work in occupations with either low AI exposure or high exposure & high complementarity.
- Skill transferability and mobility:
  - Skill distance examples (Table 4): task distance between professionals and cleaning occupations: 0.99; knowledge distance between sales and manufacturing workers: 0.8.
  - Fewer than 20 percent of new hires in high AI exposure occupations that switched jobs came from low AI exposure roles; this share drops to 10 percent for occupations with both high AI exposure and complementarity.
  - About 30 percent of hires in high AI exposure, high complementarity roles moved from jobs with high AI exposure but low complementarity.
  - Occupational mobility matrix (Table 5): total percent of employed persons that switched jobs by occupation shown (examples):
    - Total switching rate for carrying, cleaning, packaging, and related workers: 17.3 percent.
    - Total switching rate for service workers: 16.1 percent.
    - Total switching rate for clerical workers: 12.8 percent.
- Robustness:
  - Findings on AI exposure for Japan are robust to using ILO data for Japan.
  - Using share of seniors (ages 55 and above) instead of share of young workers yields similar qualitative results for labor shortages.

### Policy implications and recommendations
- Improve labor mobility to capitalize on AI benefits:
  - Targeted, occupation-oriented active labor market policies (education, retraining, reskilling) are critical given large skill distances between many low- and high-AI-exposure occupations.
- Address policy frictions discouraging longer working hours among seniors:
  - Reassess income thresholds that trigger pension cuts (e.g., FY25 monthly threshold of 510,000 yen) and tax/deduction phase-outs that incentivize reduced hours among seniors.
- Leverage AI advances selectively:
  - Promote AI adoption in sectors and occupations where AI exposure and complementarity are high and where seniors are prevalent (e.g., some construction, transportation, professional clerical roles), while preparing displaced workers in high exposure & low complementarity occupations (e.g., clerical, sales) for transitions.
- Continue policies that have raised participation rates among women and seniors, while addressing the shorter average hours worked among seniors through reforms that remove disincentives to longer hours.
- Strengthen policies to facilitate transitions from low-AI-exposure occupations to high-demand roles given structural skill mismatches (skill distance indices highlight areas where mobility will be hardest).

*Source — IMF Working Paper (content unit: wpiea2025184-source-pdf) — content provided in the supplied PDF excerpt.*

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

### wpiea2025184-source-pdf - References .............................................................................................................

### Introduction
- Japan’s economy is undergoing a profound transformation driven primarily by population aging and advances in artificial intelligence (AI).
- Paper objective: investigate the interplay between population aging and AI in influencing Japan's labor market using regression analysis, occupational-level AI exposure/complementarity measures, and a skill distance index.

### Stylized Facts
- Share of population aged 65 or older in 2023: 29 percent.
- Projected share of population aged 65 or older by 2070: 40 percent.
- Japan has one of the highest robot densities in manufacturing in 2022; Japan’s software investments from 2002 to 2022 are negatively correlated with the share of workers aged 34 and younger.
- In 2024, around half of surveyed Japanese firms indicated a lack of qualified full-time employees to meet operational needs; new job opening to application ratios remain elevated by historical standards.
- Key policy and institutional details:
  - Employers obliged (as of June 2025) to ensure employment for employees aged 60-65 by raising retirement age to 65, offering continued employment, or abolishing retirement age; for ages 65-70, employers obliged to make efforts to ensure employment.
  - In 2024, 64 percent of companies set the retirement age at 60.
  - Pension benefits cut if total wages plus Employees' Pension Insurance pension for employees aged 65 and over exceeds 510,000 yen (FY25 threshold) per month; this affects 16 percent of pension recipients in 2022.

### Methodology
- Three main approaches:
  - Panel regressions (sector-level with sector and time fixed effects) to assess aging’s impact on labor shortages and productivity (analysis period for shortages: 2008-2023; productivity long-run relations use 10-year averages for 2000-2010 and 2010-2021; data cover 97 sectors).
  - AI exposure and complementarity mapping at occupation level using Japan’s Labor Force Survey, following Felten, Raj and Seamans (2021, 2023) and Cazzaniga and others (2024):
    - Occupational categories: High exposure & high complementarity; High exposure & low complementarity; Low exposure.
  - Skill transferability: construct skill distance between occupation pairs using Occupational Information Network of Japan; skill distance index ranges from 0 to 1, where 0 = identical skill sets and 1 = orthogonal skill sets.
- Key regression specifications:
  - Labor shortages equation (sector i, time t): y_{i,t} = a_i + b_t + y_{i,t-1} + k x_{i,t-h} + β Z_{i,t} + u_{i,t}, where x_{i,t-h} is h-year lag of share of workers aged 15-34 or 55+, Z includes controls (new job opening rate, value-added growth, labor force growth, wage growth, change in average hours worked).
  - Productivity long-run equation: Δy_{i,t} = α + b_t + θ y_{i,t-1} + β1 X_{i,t} + β2 Z_{i,t} + u_{i,t}, where y is log labor productivity and X is share of workers aged 55+.

### Main findings
- Aging and labor shortages:
  - Sectors with lower share of employees ages 15-34 or higher share of workers 55+ tend to report more firms indicating labor shortages 5 years later.
  - New job opening rate has consistent positive association with labor shortages.
  - Labor shortages display considerable persistence, indicating structural nature and potentially low inter-sector mobility.
- Key regression coefficients and statistics on labor shortages (selected from Table 1):
  - Lagged labor shortages coefficients: 0.417*** (SE 0.0876), 0.298** (0.125), 0.618*** (0.0613), 0.263** (0.126).
  - New job opening as percent of total employment coefficients: 0.455* (0.251), 1.435*** (0.511), 0.380* (0.213), 1.352*** (0.495).
  - Share of workers age 15-34 (5 year lag) coefficients: -0.168* (0.0960), -1.119* (0.649), -0.223*** (0.0827).
  - Share of workers aged 55+ (5 year lag) coefficient: 1.618* (0.809).
  - Observations reported: 1216, 616, 656, 6 (as shown); R-squared values: 0.82, 0.66, 0.93, 0.30.
  - Dependent variable: Labor shortages reported by firms by sectors (percent balance).
- Aging and productivity:
  - Aging in the labor force is negatively associated with labor productivity growth (Table 2).
  - Investment rate has positive and statistically significant association with labor productivity.
  - No statistically significant link identified between labor productivity growth and share of part-time or female workers in this analysis.
- AI exposure, complementarity, and occupational patterns:
  - In 2023, share of all workers who were seniors (age 65+): 14 percent.
  - Japan has a larger share of workers with low exposure to AI compared with other G7 countries; share of workers in high exposure & low complementarity is larger compared with the United States or Germany.
  - Occupation-specific vulnerability:
    - Managers and professionals concentrate in high exposure & high complementarity occupations.
    - Clerical support workers, sales workers, telemarketers: high exposure & low complementarity (more vulnerable to displacement).
    - Service workers, security workers, manufacturing processing, packaging/carrying: lower AI exposure (though manufacturing processing is exposed to automation).
  - By gender and age:
    - Women tend to concentrate in high exposure & low complementarity occupations (higher displacement risk).
    - Age cohort 65+ is least exposed to AI; cohort 55-64 has vulnerability similar to middle-aged groups due to concentration in clerical and sales occupations.
- Sectoral detail on seniors and AI potential (Table 3):
  - Share of workers 65+ in top five sectors with most severe labor shortages ranges from 2.5 percent (ICT) to 17 percent (construction).
  - In construction and transportation services, AI may help alleviate aging-induced shortages (share of occupations with high AI exposure & low complementarity relatively low).
  - In medical and healthcare services, AI may have limited room to address shortages because seniors work in occupations with either low AI exposure or high exposure & high complementarity.
- Skill transferability and mobility:
  - Skill distance examples (Table 4): task distance between professionals and cleaning occupations: 0.99; knowledge distance between sales and manufacturing workers: 0.8.
  - Fewer than 20 percent of new hires in high AI exposure occupations that switched jobs came from low AI exposure roles; this share drops to 10 percent for occupations with both high AI exposure and complementarity.
  - About 30 percent of hires in high AI exposure, high complementarity roles moved from jobs with high AI exposure but low complementarity.
  - Occupational mobility matrix (Table 5): total percent of employed persons that switched jobs by occupation shown (e.g., Total switching rate for carrying, cleaning, packaging, and related workers: 17.3 percent; for service workers: 16.1 percent; for clerical workers: 12.8 percent).
- Robustness:
  - Findings on AI exposure for Japan are robust to using ILO data for Japan.
  - Using share of seniors (ages 55 and above) instead of share of young workers yields similar qualitative results for labor shortages.

### Policy implications and recommendations
- Improve labor mobility to capitalize on AI benefits:
  - Targeted, occupation-oriented active labor market policies (education, retraining, reskilling) are critical given large skill distances between many low- and high-AI-exposure occupations.
- Address policy frictions discouraging longer working hours among seniors:
  - Reassess income thresholds that trigger pension cuts (e.g., FY25 monthly threshold of 510,000 yen) and tax/deduction phase-outs that incentivize reduced hours among seniors.
- Leverage AI advances selectively:
  - Promote AI adoption in sectors and occupations where AI exposure and complementarity are high and where seniors are prevalent (e.g., some construction, transportation, professional clerical roles), while preparing displaced workers in high exposure & low complementarity occupations (e.g., clerical, sales) for transitions.
- Continue policies that have raised participation rates among women and seniors, while addressing the shorter average hours worked among seniors through reforms that remove disincentives to longer hours.
- Strengthen policies to facilitate transitions from low-AI-exposure occupations to high-demand roles given structural skill mismatches (skill distance indices highlight areas where mobility will be hardest).

*Italic: Source — IMF Working Paper (content unit: wpiea2025184-source-pdf) — content provided in the supplied PDF excerpt.*

### References

### References

### Demographics, Aging, and Japan-specific Studies
- Asao, K., Smirnov, D., & Xu, T. (2024). Japan’s Fertility: More Children Please. IMF Selected Issues Paper No. 2024/025. Washington, D.C.: International Monetary Fund.
- International Monetary Fund (IMF). 2018. Japan: Selected Issues. IMF Country Report No. 18/334. Washington, DC.
- International Monetary Fund (IMF). April 2025. World Economic Outlook, Chapter 2. “The Rise of the Silver Economy: Global Implications of Population Aging”. Washington, DC.
- Kikuchi, S., I. Fujiwara, and T. Shirota. 2024. “Automation and Disappearing Routine Occupations in Japan.” Journal of the Japanese and International Economies 74(C).
- Maestas, N., K. J. Mullen, and D. Powell. 2023. “The Effect of Population Aging on Economic Growth, the Labor Force and Productivity.” American Economic Journal: Macroeconomics 15 (2): 306–332.
- Westelius, N., and Y. Liu. 2016. “Aging, Automation, and the Labor Market: Evidence from Japan.” IMF Working Paper No. 16/176, International Monetary Fund, Washington, DC.

### Artificial Intelligence, Automation, and Labor Markets
- Cazzaniga, M., F. Jaumotte, L. Li, G. Melina, A. J. Panton, C. Pizzinelli, E. J. Rockall, and M. M. Tavares. 2024. “Gen-AI: Artificial Intelligence and the Future of Work.” IMF Staff Discussion Note No. 2024/001, International Monetary Fund, Washington, DC.
- Cazzaniga, M., Pizzinelli, C., Rockall, E., and Tavares, M. M. (2024). Exposure to Artificial Intelligence and Occupational Mobility: A Cross-Country Analysis. IMF Working Paper 24/116.
- Felten, E., M. Raj, and R. Seamans. 2021. “Occupational, Industry, and Geographic Exposure to Artificial Intelligence: A Novel Dataset and Its Potential Uses.” Strategic Management Journal 42 (12): 2195–2217.
- Felten, E., M. Raj, and R. Seamans. 2023. “How Will Language Modelers Like ChatGPT Affect Occupations and Industries?” arXiv.org Working Paper.
- Acemoglu, D., and D. Autor. 2011. “Skills, Tasks and Technologies: Implications for Employment and Earnings.” In Handbook of Labor Economics, Vol. 4B, ed. O. Ashenfelter and D. Card, 1043–1171. Amsterdam: Elsevier.
- Gathmann, C., and U. Schönberg. 2010. “How General Is Human Capital? A Task-Based Approach.” Journal of Labor Economics 28 (1): 1–49.

### Convergence, Regional Growth, and Macroeconomic Context
- Barro, R. J. 2015. “Convergence and Modernization.” Economic Journal 125 (585): 911–942.
- Sala-i-Martin, X. 1996. “Regional Cohesion: Evidence and Theories of Regional Growth and Convergence.” European Economic Review 40 (6): 1325–1352.

### Other Working Papers and Notes
- Cazzaniga, M., Pizzinelli, C., Rockall, E., and Tavares, M. M. (2024). Exposure to Artificial Intelligence and Occupational Mobility: A Cross-Country Analysis. IMF Working Paper 24/116.
- Westelius, N., and Y. Liu. 2016. “Aging, Automation, and the Labor Market: Evidence from Japan.” IMF Working Paper No. 16/176, International Monetary Fund, Washington, DC.

*wpiea2025184-source-pdf - References*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2025/english/wpiea2025184-source-pdf.pdf_