## sipea2024026

## Source details

**Canonical URL:** [sipea2024026](https://www.imf.org/-/media/files/publications/selected-issues-papers/2024/english/sipea2024026.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/selected-issues-papers/2024/english/sipea2024026.pdf.md)
- [Structured JSON version](/-/media/files/publications/selected-issues-papers/2024/english/sipea2024026.pdf.json)

---

### Data and empirical scope
- Data sources:
  - Japan Household Panel Survey and the Keio Household Panel Survey (JHPS/KHPS), representative annual household panel since 2004 with a total sample size of about 5,000 households.
  - Analysis focuses on 2010-19 (corresponding to survey waves 2011-20).
- Income definitions:
  - Market income = wage income + self-employment income + capital income.
  - Gross income = market income + cash transfers.
  - Wage income split into regular and nonregular employment.
  - Self-employment income includes net business income and home-work income (including agriculture, forestry and fishery income).
  - Capital income includes rent income, interest, and dividends.
  - Cash transfers include public pension, corporate and personal pensions, unemployment benefit, childcare allowances, and other transfers.
- Limitation noted:
  - JHPS/KHPS does not collect tax payment data; disposable income and tax impacts are not analyzed.

### Decomposition methodology
- Overall Gini index (G) is decomposed following Lerman and Yitzhaki (1985).
- Three channels by which income source k affects overall inequality:
  - Share of income source in total income (S_k).
  - Within-source inequality measured by G_k (source Gini).
  - Correlation of income source with total income measured by R_k.
- Concentration coefficient (C_k):
  - C_k = G_k × R_k (quasi-Gini), ranges between -1 and 1.
  - Interpretation:
    - C_k negative: income source is progressive (concentrated among poorer households).
    - C_k positive: income source is regressive (concentrated among richer households).
    - If C_k < G, income source k reduces inequality; if C_k > G, it increases inequality.
- Change decomposition:
  - Change in overall G can be decomposed into:
    - Share effect: changes in S_k over time.
    - Concentration effect: changes in C_k (or G_k and R_k) over time.
  - Overall Gini is G_0 = Σ_k S_k C_k; changes decompose into share effects, concentration effects, and a residual.

### Key empirical findings (aggregate and by source)
- Concordance with macro indicators:
  - Gini coefficients for gross and market income in JHPS are consistent with macro-based indicators; micro data deemed reliable representation for population-level trends.
- Trends 2010-19:
  - Overall gross income inequality has marginally declined by 1.3 percent over the pre-COVID decade.
  - Market income inequality increased by 1.7 percent over the same period.
- Income composition (2019 and averages):
  - Wage income constitutes on average 70 percent of household income (S_k).
  - Transfers account for 20 percent in 2019.
  - Within wage income, regular wage income constitutes two thirds of wage income; its share in total income has increased over time.
  - Share of nonregular wage income remained broadly unchanged but varies substantially across income groups (regular wage share highest among richer households; transfers largest among lower-income households).
  - For elderly: transfers and capital income are relatively larger shares; regular wage income as share of gross income is less than 8 percent.
- Inequality by source:
  - Capital income has the highest within-source Gini (G_k), followed by self-employment income.
  - Wage inequality among nonregular workers is slightly higher than within regular workers (difference small).
  - Concentration patterns (C_k):
    - Regular wage income and capital income are concentrated among richer households (positive C_k).
    - Transfers and nonregular employment are relatively more concentrated among lower-income groups (C_k < G), thus have redistributive effects.
- Marginal impacts (2019):
  - A 1 percent increase in income from transfers, all else equal, reduces overall inequality by close to 14 percent.
  - A 1 percent increase in regular wage income increases overall inequality by 12 percent.
  - A 1 percent increase in capital income increases overall inequality by 4 percent.
  - The unequalizing effects of regular employment income and the equalizing effects of transfers have increased considerably over the years.

### Sources of change in gross income inequality, 2010-19
- Four key drivers identified:
  - (i) Changes in age distribution (ageing of the population) — unequalizing.
  - (ii) Changes in labor market structure (increasing dualism) — unequalizing.
  - (iii) Increase in labor force participation/employment of females and elderly — equalizing.
  - (iv) Changes in distributions of social transfers — equalizing.
- Net aggregate effect:
  - Equalizing impact of transfers and higher female/elderly participation marginally offset unequalizing impact of ageing and increased labor market dualism.

### Effects of social transfers
- Redistribution magnitude:
  - Fiscal redistribution through social transfers reduced the Gini coefficient on market income by 33 percent in 2010 and by 37 percent in 2019 (this includes pension benefits).
  - Social transfers helped reduce the Gini coefficient on total gross income by 6 percent between 2010-19.
- Targeting and progressivity:
  - Transfers are more effective for the elderly than for the working-age population.
  - For the elderly, concentration coefficient is positive in most countries (including Japan) because pensions are linked to past income.
  - For the working-age population, concentration coefficient is usually negative in OECD countries (progressive), but in Japan it is positive, indicating relatively weak targeting toward low-income working-age households.
  - Pensions constitute close to 40 percent of the total transfers in the dataset.
- Efficiency caveat:
  - Reliance on tax-and-transfer redistributive mechanisms may be inefficient and financially unsustainable; authors note importance of tackling market income inequality through other means.

### Effects of demographics
- Ageing:
  - Share of population aged above 65 increased by 10 percentage points over the last decade.
  - Ageing increased overall income inequality because elderly incomes are lower on average and many have no labor income; wage distribution is more unequal for older workers.
- Offsetting factors:
  - Redistributive transfers and increased labor force participation/employment of elderly have partially offset ageing’s unequalizing impact.

### Female and elderly labor supply effects
- Labor force additions and participation changes 2010-19:
  - Japan added 2.9 million women and 3.5 million elderly (age 65 and above) to the labor force between 2010-19.
  - Female and elderly labor force participation rates each went up by 6 ppt.
  - Unemployment rate declined from 5 to 2.3 percent.
- Distributional impact:
  - Increased participation and employment of females and elderly acted as an income-equalizing force by increasing employment income among previously nonparticipating individuals and increasing incidence of dual and multi-income earner households.
  - This helped reduce the Gini index for market income among females by close to 5 percent between 2010-19, and among the elderly by 3 percent.

### The increasing duality in the labor market and structural barriers to full-time employment

#### Major findings on inequality and labor-market duality
- Overall wage inequality constitutes close to 80 percent of the gross income inequality and 90 percent of the market income inequality.
- More than half of the overall income inequality is associated with inequality in regular wages; the influence of regular wages on inequality has increased by 10 percentage points over time.
- The average share of regular wage income in total gross income has grown and increasingly benefited higher-income households; households with regular employment have become richer over time, widening the income divide with the rest of the Japanese population.
- The growing share of nonregular workers increased from 33 percent in 2010 to 38 percent in 2019 and, together with declining hours of work for nonregular workers, widened the income gap between regular and nonregular workers.
- On an hourly basis, part-time workers—who account for 70 percent of nonregular workers—were paid only 57 percent as much as full-time workers in 2019.
- Wage inequality within regular workers and within nonregular workers, respectively, has remained fairly unchanged since 2010; the widening gap is driven mainly by changes in shares and hours rather than within-group wage dispersion.
- Most nonregular workers do not receive bonus payments and retirement benefits: 70 percent of part-time workers do not receive bonus payments and 90 percent do not receive the lump-sum retirement benefit paid by firms.
- Opportunities for career progression and training for nonregular workers are typically limited, further enlarging the wage gap between regular and irregular workers over time.

#### Mechanisms and distortions
- Declining work hours of nonregular workers are in part due to distortionary effects of tax and social security benefits that force second earners in the family to reduce working hours.
- Limited mobility opportunities for nonregular workers exacerbate income inequality between households with regular employment and others.
- Under Japan’s seniority-based wage system, firms often cannot afford to keep regular elderly workers, prompting mandatory retirement and transformation of regular workers to nonregular status thereafter; this discourages some employees from continuing to work and lowers the productivity of those who stay.

#### Distributional patterns by gender and age
- 70 percent of females employed during 2010-19 worked in nonregular jobs.
- 85 percent of the elderly employed during 2010-19 worked in nonregular jobs.
- Even women who are employed as regular workers and leave the labor force to care for children tend to be relegated to nonregular status when they return, exacerbating gender wage gaps.
- For the elderly, the shift from regular to nonregular employment has contributed to widening inequality among the elderly.

#### Data, alternative explanations, and methodology notes
- Results may vary with dataset choice; e.g., Yoshino and others (2018) using the quarterly Family Income and Expenditure Survey find that a rise in the price of financial assets (capital gains) that benefited richer households is a key driver of increased income inequality in Japan.
- The Japan Household Panel Survey and the Keio Household Panel Survey (JHPS/KHPS) collect 11 income categories for respondents, spouses, and aggregates of other family members; annual employment incomes of respondents and spouses include bonuses and monthly salaries and are segregated into regular and non-regular employment.
- The KHPS asks the amount of severance pay but does not ask duration of employment; severance pay is therefore not considered in the model, following Kawade (2018).
- The chapter adopts a factor decomposition of the Gini index following Shorrocks (1982), Lerman and Yitzhaki (1985), and Stark, Taylor, and Yitzhaki (1986); key definitions include S_k (share of component k), G_k (Gini of component k), R_k (Gini correlation), and C_k (concentration index, where C_k = G_k ∙ R_k). The overall Gini is G_0 = Σ_k S_k C_k, and changes decompose into share effects, concentration effects, and a residual.

#### Policy recommendations
- Reduce labor market dualism and improve labor mobility:
  - Address factors that encourage firms to hire non-regular workers, mainly lower labor costs and greater employment flexibility.
  - Consider introducing a Single Open-Ended Contract for all newly hired workers, complemented by a model that combines labor market flexibility and security (as suggested in Aoyagi and Ganelli (2013)).
  - Implement programs to increase skill training to enhance career opportunities and mobility for nonregular workers, raising their productivity and real incomes and reducing income gaps.
- Further boost labor participation of females and the elderly and remove disincentives to employment in social benefit policies:
  - Advance implementation of work-style reforms, including flexible work arrangements such as telework, to bring more women and elderly into full-time employment (IMF, 2022, Annex IX).
  - Eliminate social security and tax distortions related to dependent spouses to allow voluntary increases in working hours (Xu and Chahande, 2023).
- Better target existing social benefit programs:
  - Given fiscal debt constraints, shift the allocation of social spending to increase the share received by low-income households to help lower income inequality (IMF, 2024).

#### Key statistics and exact figures
- Overall wage inequality: close to 80 percent (gross income) and 90 percent (market income) of income inequality.
- Increase in influence of regular wages on inequality: 10 percentage points over time.
- Share of nonregular workers: 33 percent in 2010; 38 percent in 2019.
- Share of part-time workers among nonregular workers: 70 percent.
- Hourly pay of part-time workers relative to full-time workers in 2019: 57 percent.
- Share of females employed in nonregular jobs during 2010-19: 70 percent.
- Share of elderly employed in nonregular jobs during 2010-19: 85 percent.
- Share of part-time workers who do not receive bonus payments: 70 percent.
- Share of part-time workers who do not receive lump-sum retirement benefit: 90 percent.

*Source: sipea2024026 - 25. The increasing duality in the labor market and structural barriers to full-time employment (IMF).*

### 2. Methodology for Factor Decomposition of the Gini Index _________________________ 14

### 2. Methodology for Factor Decomposition of the Gini Index

### Data and empirical scope
- Data sources:
  - Japan Household Panel Survey and the Keio Household Panel Survey (JHPS/KHPS), representative annual household panel since 2004 with a total sample size of about 5,000 households.
  - Analysis focuses on 2010-19 (corresponding to survey waves 2011-20).
- Income definitions:
  - Market income = wage income + self-employment income + capital income.
  - Gross income = market income + cash transfers.
  - Wage income split into regular and nonregular employment.
  - Self-employment income includes net business income and home-work income (including agriculture, forestry and fishery income).
  - Capital income includes rent income, interest, and dividends.
  - Cash transfers include public pension, corporate and personal pensions, unemployment benefit, childcare allowances, and other transfers.
- Limitation noted:
  - JHPS/KHPS does not collect tax payment data; disposable income and tax impacts are not analyzed.

### Decomposition methodology
- Overall Gini index (G) is decomposed following Lerman and Yitzhaki (1985).
- Three channels by which income source k affects overall inequality:
  - Share of income source in total income (S_k).
  - Within-source inequality measured by G_k (source Gini).
  - Correlation of income source with total income measured by R_k.
- Concentration coefficient (C_k):
  - C_k = G_k × R_k (quasi-Gini), ranges between -1 and 1.
  - Interpretation:
    - C_k negative: income source is progressive (concentrated among poorer households).
    - C_k positive: income source is regressive (concentrated among richer households).
    - If C_k < G, income source k reduces inequality; if C_k > G, it increases inequality.
- Change in overall G can be decomposed into:
  - Share effect: changes in S_k over time.
  - Concentration effect: changes in C_k (or G_k and R_k) over time.

### Key empirical findings (aggregate and by source)
- Concordance with macro indicators:
  - Gini coefficients for gross and market income in JHPS are consistent with macro-based indicators; micro data deemed reliable representation for population-level trends.
- Trends 2010-19:
  - Overall gross income inequality has marginally declined by 1.3 percent over the pre-COVID decade.
  - Market income inequality increased by 1.7 percent over the same period.
- Income composition (2019 and averages):
  - Wage income constitutes on average 70 percent of household income (S_k).
  - Transfers account for 20 percent in 2019.
  - Within wage income, regular wage income constitutes two thirds of wage income; its share in total income has increased over time.
  - Share of nonregular wage income remained broadly unchanged but varies substantially across income groups (regular wage share highest among richer households; transfers largest among lower-income households).
  - For elderly: transfers and capital income are relatively larger shares; regular wage income as share of gross income is less than 8 percent.
- Inequality by source:
  - Capital income has the highest within-source Gini (G_k), followed by self-employment income.
  - Wage inequality among nonregular workers is slightly higher than within regular workers (difference small).
  - Concentration patterns (C_k):
    - Regular wage income and capital income are concentrated among richer households (positive C_k).
    - Transfers and nonregular employment are relatively more concentrated among lower-income groups (C_k < G), thus have redistributive effects.
- Marginal impacts (2019):
  - A 1 percent increase in income from transfers, all else equal, reduces overall inequality by close to 14 percent.
  - A 1 percent increase in regular wage income increases overall inequality by 12 percent.
  - A 1 percent increase in capital income increases overall inequality by 4 percent.
  - The unequalizing effects of regular employment income and the equalizing effects of transfers have increased considerably over the years.

### Sources of change in gross income inequality, 2010-19
- Four key drivers identified:
  - (i) Changes in age distribution (ageing of the population) — unequalizing.
  - (ii) Changes in labor market structure (increasing dualism) — unequalizing.
  - (iii) Increase in labor force participation/employment of females and elderly — equalizing.
  - (iv) Changes in distributions of social transfers — equalizing.
- Net aggregate effect:
  - Equalizing impact of transfers and higher female/elderly participation marginally offset unequalizing impact of ageing and increased labor market dualism.

### Effects of social transfers
- Redistribution magnitude:
  - Fiscal redistribution through social transfers reduced the Gini coefficient on market income by 33 percent in 2010 and by 37 percent in 2019 (this includes pension benefits).
  - Social transfers helped reduce the Gini coefficient on total gross income by 6 percent between 2010-19.
- Targeting and progressivity:
  - Transfers are more effective for the elderly than for the working-age population.
  - For the elderly, concentration coefficient is positive in most countries (including Japan) because pensions are linked to past income.
  - For the working-age population, concentration coefficient is usually negative in OECD countries (progressive), but in Japan it is positive, indicating relatively weak targeting toward low-income working-age households.
  - Pensions constitute close to 40 percent of the total transfers in the dataset.
- Efficiency caveat:
  - Reliance on tax-and-transfer redistributive mechanisms may be inefficient and financially unsustainable; authors note importance of tackling market income inequality through other means.

### Effects of demographics
- Ageing:
  - Share of population aged above 65 increased by 10 percentage points over the last decade.
  - Ageing increased overall income inequality because elderly incomes are lower on average and many have no labor income; wage distribution is more unequal for older workers.
- Offsetting factors:
  - Redistributive transfers and increased labor force participation/employment of elderly have partially offset ageing’s unequalizing impact.

### Female and elderly labor supply effects
- Labor force additions and participation changes 2010-19:
  - Japan added 2.9 million women and 3.5 million elderly (age 65 and above) to the labor force between 2010-19.
  - Female and elderly labor force participation rates each went up by 6 ppt.
  - Unemployment rate declined from 5 to 2.3 percent.
- Distributional impact:
  - Increased participation and employment of females and elderly acted as an income-equalizing force by increasing employment income among previously nonparticipating individuals and increasing incidence of dual and multi-income earner households.
  - This helped reduce the Gini index for market income among females by close to 5 percent between 2010-19, and among the elderly by 3 percent.

*sipea2024026 - 2. Methodology for Factor Decomposition of the Gini Index _________________________ 14*

### 25.      The increasing duality in the labor market and structural barriers to full-time

### 25.      The increasing duality in the labor market and structural barriers to full-time employment

### Major findings on inequality and labor-market duality
- Overall wage inequality constitutes close to 80 percent of the gross income inequality and 90 percent of the market income inequality.
- More than half of the overall income inequality is associated with inequality in regular wages; the influence of regular wages on inequality has increased by 10 percentage points over time.
- The average share of regular wage income in total gross income has grown and increasingly benefited higher-income households; households with regular employment have become richer over time, widening the income divide with the rest of the Japanese population.
- The growing share of nonregular workers increased from 33 percent in 2010 to 38 percent in 2019 and, together with declining hours of work for nonregular workers, widened the income gap between regular and nonregular workers.
- On an hourly basis, part-time workers—who account for 70 percent of nonregular workers—were paid only 57 percent as much as full-time workers in 2019.
- Wage inequality within regular workers and within nonregular workers, respectively, has remained fairly unchanged since 2010; the widening gap is driven mainly by changes in shares and hours rather than within-group wage dispersion.
- Most nonregular workers do not receive bonus payments and retirement benefits: 70 percent of part-time workers do not receive bonus payments and 90 percent do not receive the lump-sum retirement benefit paid by firms.
- Opportunities for career progression and training for nonregular workers are typically limited, further enlarging the wage gap between regular and irregular workers over time.

### Mechanisms and distortions
- Declining work hours of nonregular workers are in part due to distortionary effects of tax and social security benefits that force second earners in the family to reduce working hours.
- Limited mobility opportunities for nonregular workers exacerbate income inequality between households with regular employment and others.
- Under Japan’s seniority-based wage system, firms often cannot afford to keep regular elderly workers, prompting mandatory retirement and transformation of regular workers to nonregular status thereafter; this discourages some employees from continuing to work and lowers the productivity of those who stay.

### Distributional patterns by gender and age
- 70 percent of females employed during 2010-19 worked in nonregular jobs.
- 85 percent of the elderly employed during 2010-19 worked in nonregular jobs.
- Even women who are employed as regular workers and leave the labor force to care for children tend to be relegated to nonregular status when they return, exacerbating gender wage gaps.
- For the elderly, the shift from regular to nonregular employment has contributed to widening inequality among the elderly.

### Data, alternative explanations, and methodology notes
- Results may vary with dataset choice; e.g., Yoshino and others (2018) using the quarterly Family Income and Expenditure Survey find that a rise in the price of financial assets (capital gains) that benefited richer households is a key driver of increased income inequality in Japan.
- The Japan Household Panel Survey and the Keio Household Panel Survey (JHPS/KHPS) collect 11 income categories for respondents, spouses, and aggregates of other family members; annual employment incomes of respondents and spouses include bonuses and monthly salaries and are segregated into regular and non-regular employment.
- The KHPS asks the amount of severance pay but does not ask duration of employment; severance pay is therefore not considered in the model, following Kawade (2018).
- The chapter adopts a factor decomposition of the Gini index following Shorrocks (1982), Lerman and Yitzhaki (1985), and Stark, Taylor, and Yitzhaki (1986); key definitions include S_k (share of component k), G_k (Gini of component k), R_k (Gini correlation), and C_k (concentration index, where C_k = G_k ∙ R_k). The overall Gini is G_0 = Σ_k S_k C_k, and changes decompose into share effects, concentration effects, and a residual.

### Policy recommendations
- Reduce labor market dualism and improve labor mobility:
  - Address factors that encourage firms to hire non-regular workers, mainly lower labor costs and greater employment flexibility.
  - Consider introducing a Single Open-Ended Contract for all newly hired workers, complemented by a model that combines labor market flexibility and security (as suggested in Aoyagi and Ganelli (2013)).
  - Implement programs to increase skill training to enhance career opportunities and mobility for nonregular workers, raising their productivity and real incomes and reducing income gaps.
- Further boost labor participation of females and the elderly and remove disincentives to employment in social benefit policies:
  - Advance implementation of work-style reforms, including flexible work arrangements such as telework, to bring more women and elderly into full-time employment (IMF, 2022, Annex IX).
  - Eliminate social security and tax distortions related to dependent spouses to allow voluntary increases in working hours (Xu and Chahande, 2023).
- Better target existing social benefit programs:
  - Given fiscal debt constraints, shift the allocation of social spending to increase the share received by low-income households to help lower income inequality (IMF, 2024).

### Key statistics and exact figures
- Overall wage inequality: close to 80 percent (gross income) and 90 percent (market income) of income inequality.
- Increase in influence of regular wages on inequality: 10 percentage points over time.
- Share of nonregular workers: 33 percent in 2010; 38 percent in 2019.
- Share of part-time workers among nonregular workers: 70 percent.
- Hourly pay of part-time workers relative to full-time workers in 2019: 57 percent.
- Share of females employed in nonregular jobs during 2010-19: 70 percent.
- Share of elderly employed in nonregular jobs during 2010-19: 85 percent.
- Share of part-time workers who do not receive bonus payments: 70 percent.
- Share of part-time workers who do not receive lump-sum retirement benefit: 90 percent.

*Source: sipea2024026 - 25. The increasing duality in the labor market and structural barriers to full-time employment (IMF).*

---


_Source: https://www.imf.org/-/media/files/publications/selected-issues-papers/2024/english/sipea2024026.pdf_