## _wp1554

## Source details

**Canonical URL:** [_wp1554](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp1554.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp1554.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp1554.pdf.json)

---

### A. Inequality and Redistribution
- Market income Gini and disposable income Gini for the working-age population are highly correlated: a simple bilateral regression suggests that, for the working-age population, each point increase in the market income Gini translates into a 0.4 point increase in the disposable income Gini.
- Goodness-of-fit for that regression is about 0.76, and the linear correlation coefficient about 0.87.
- Fiscal redistribution substantially reduces inequality for the elderly: the large gap between market income and disposable income inequality for the elderly is a sign of fiscal burden.
- OECD (2011, 2014) findings cited emphasize that reliance on tax-and-transfer equalization is not efficient or sustainable given population ageing, over-stretched public finances, rising dependency ratio, and low fertility.
- Policy implication noted: Japan might need to take into account the fiscal costs of redistribution given high and rising public debt, increasing share of the elderly population, low fertility rate, and rising dependency ratio; emphasis on structural reforms, including labor market reforms, seems appropriate.

### B. Poverty
- Relative poverty rates defined as households with income below half of median income.
- In 2009, the relative poverty indicator was around 16 percent in Japan.
- Japan’s relative poverty rate in 2009 is 4 percentage points higher than the G7 average and is the second highest among G7 countries following the US.
- Japan’s relative poverty rate is the sixth highest among OECD countries and is above the OECD average by 5 percentage points.
- Child poverty (working households with children in relative poverty) has risen rapidly since 2006 and surpassed the full-sample poverty rate in 2012.
- Single-parent households (majority single-mother households) have relative poverty rates that remained above 50 percent for 1985-2012.
- Poverty by age groups (relative poverty rates, 1985–2009 series shown):
  - Rapid increases in poverty rates among age groups 0-17 and 18-25.
  - Less marked increases for ages 26-65.
  - Declines over time for ages 66-75.
- Long-run concern: high child poverty may produce intergenerational “lock-in” effects, reducing aggregate human capital and potential growth; early education and training are emphasized (Heckman 2000).

### C. Other Dimensions of Inequality
- Two prominent dimensions linked to widening wage inequality among working-age population:
  - Gender gap: Female Labor Participation (FLP) is lower than Male Participation Rate (MPR) by about 20 percentage points in Japan.
  - Labor market duality: share of non-regular workers rose from less than 20 percent in the 1980s to above 35 percent by 2011.
- More than half of employed women are non-regular workers, implying less job security, lower wages, and reduced career opportunities.
- Low FLP and underemployment of women reduce inclusiveness and potential growth.
- Aoyagi and Ganelli (2013) argue excessive duality reduces productivity via a “training channel” and an “effort channel.”

### IV. Data and Empirical Strategy
- Data: prefectural level longitudinal data from the National Population Census (every 5 years); income components include wages, interest, rent, social security and other payments to households.
- Incomes deflated by prefectural GDP deflators from the Cabinet Office.
- Inclusive growth measure: Anand et al. (2013) weighted average of growth in average income and change in an equity index bounded between zero and one (one = perfectly equitable income distribution); measure equals average income growth if distribution unchanged and deviates with equity changes.
- Dependent variable: log-difference (approximate percentage change) of 5-year panel between years 1979-2004; log differences standardized to an average annual rate via geometric mean (e.g., first observation represents average growth 1979-1984).
- Main policy explanatory variables (converted to 5-year averages):
  - Inflation rate: 5-year average, %
  - Ratio of new job openings for part-time to full-time labor: 5-year average, %
  - Female labor force participation ratio: 5-year average, %
  - Growth of labor input in man-hours: 5-year average, %
  - Controls: GDP per capita (initial value at the 5 year periods), Elderly index (5-year average, index; size of population 65+ divided by working-age population)
- Note: female labor participation rates used for entire workforce (15 years and older) due to data limitation at prefectural level.

### V. Results (econometric)
- Model: panel with prefectural fixed effects (Hausman test rejected random effects; Honda test rejected no individual effects). Heteroskedasticity robust standard errors reported.
- Sample: results presented for all households and for working-age households. For each: (1) Inclusive Growth; (2) Average Income Growth; (3) Equity Index Growth.
- Key coefficient estimates (standard errors in parentheses):

Table 1. Results for All Households (Observations 235)
- Inflation (%) — Inclusive Growth: 1.586*** (0.197); Average Income Growth: 1.598*** (0.174); Equity Index Growth: −0.032 (0.069)
- Inflation, squared — Inclusive Growth: −0.353*** (0.041); Average Income Growth: −0.331*** (0.037); Equity Index Growth: −0.020 (0.015)
- Part- to Full-time job openings (%) — Inclusive Growth: −0.028 (0.020); Average Income Growth: −0.033* (0.018); Equity Index Growth: 0.005 (0.007)
- Female labor force participation (%) — Inclusive Growth: 0.165* (0.084); Average Income Growth: 0.125* (0.074); Equity Index Growth: 0.040 (0.029)
- Labor input growth (%) — Inclusive Growth: 0.524*** (0.165); Average Income Growth: 0.489*** (0.146); Equity Index Growth: 0.029 (0.058)
- Initial GDP per capita — Inclusive Growth: −1.564** (0.762); Average Income Growth: −0.985 (0.672); Equity Index Growth: −0.600** (0.267)
- Elderly index — Inclusive Growth: 0.057 (0.037); Average Income Growth: 0.052 (0.033); Equity Index Growth: 0.004 (0.013)
- R2: Inclusive Growth 0.660; Average Income Growth 0.714; Equity Index Growth 0.112

Table 2. Results for Working-age Households (Observations 240)
- Inflation (%) — Inclusive Growth: 1.514*** (0.197); Average Income Growth: 1.512*** (0.174); Equity Index Growth: −0.008 (0.075)
- Inflation, squared — Inclusive Growth: −0.279*** (0.041); Average Income Growth: −0.264*** (0.037); Equity Index Growth: −0.015 (0.016)
- Part- to Full-time job openings (%) — Inclusive Growth: −0.039* (0.020); Average Income Growth: −0.040** (0.018); Equity Index Growth: 0.001 (0.008)
- Female labor force participation (%) — Inclusive Growth: 0.225*** (0.084); Average Income Growth: 0.155** (0.074); Equity Index Growth: 0.070** (0.032)
- Labor input growth (%) — Inclusive Growth: 0.384** (0.166); Average Income Growth: 0.358** (0.146); Equity Index Growth: 0.023 (0.063)
- Initial GDP per capita — Inclusive Growth: −0.487 (0.764); Average Income Growth: 0.036 (0.673); Equity Index Growth: −0.539* (0.291)
- Elderly index — Inclusive Growth: 0.068* (0.037); Average Income Growth: 0.057* (0.033); Equity Index Growth: 0.011 (0.014)
- R2: Inclusive Growth 0.627; Average Income Growth 0.677; Equity Index Growth 0.065

- Interpretation of inflation coefficients:
  - Inflation modeled quadratically: coefficients imply an initially increasing but then falling effect of inflation on inclusive growth as inflation rises.
  - Estimated maximum effect on inclusive growth reached at around 2 percent inflation rate.
  - Inflation has significant positive effects on inclusive growth and average income growth; negative but small and insignificant effect on equity index growth.
  - Robustness: pseudo-Granger causality tests and IV estimates using lagged inflation (Appendix D) broadly confirm causality from inflation to growth.
- Other key findings:
  - Labor market duality (part- to full-time ratio) has a negative and significant impact on inclusive growth via negative impact on average income growth; effect stronger in working-age sample.
  - Female labor participation positively affects inclusive growth primarily via average income growth; for working-age sample, positive effects on both average income and equity index growth.
  - Labor input growth increases inclusive growth mainly by increasing average income.
  - Initial GDP per capita is a negative and significant determinant of inclusive growth for all-household sample.
  - Elderly index: insignificant for all households, marginally positive and significant for working-age sample.
  - Overall: policy variables affect inclusive growth mostly through average income growth; potential impact of a complete Abenomics package on income equality is relatively small.

### VI. Policy Implications (scenario analysis summary)
- Base level of inclusive growth (national, 2000-04): −1.46 percent for the entire household sample.
- Main policy-relevant findings:
  - Moving towards a 2 percent inflation target (BoJ goal) is estimated to promote growth and raise inclusive growth, with diminishing returns and a turning point around 2 percent inflation.
  - Reducing labor market duality (lower ratio of part-time to full-time new job openings) is estimated to increase inclusive growth by raising average income growth.
  - Increasing female labor force participation is estimated to raise inclusive growth by increasing average income growth and, in the working-age sample, improving equity index growth.
  - Increasing labor input (man-hours) is estimated to boost inclusive growth mainly through average income growth.
- Overall conclusion: a “complete” Abenomics package that combines exiting deflation (toward 2 percent inflation), labor market reforms reducing duality, policies raising female labor participation, and measures increasing labor input is estimated to raise inclusive growth mainly by increasing average income growth. The direct effect on income equality (equity index) is limited in the estimated results.

### Scenario results — Inclusive Growth (Table 3)
- Inflation (0.0 to 2.0%)
  - All Households: Incl. Growth 1.76; Avg. Growth 1.87; Equity Change -0.14
  - Working-Age Households: Incl. Growth 1.91; Avg. Growth 1.97; Equity Change -0.07
- Part- to Full-time job postings ratio (21 to 5%)
  - All Households: Incl. Growth 0.45; Avg. Growth 0.53; Equity Change -0.08
  - Working-Age Households: Incl. Growth 0.63; Avg. Growth 0.64; Equity Change -0.01
- Female Labor Participation (47 to 52%)
  - All Households: Incl. Growth 0.83; Avg. Growth 0.63; Equity Change 0.20
  - Working-Age Households: Incl. Growth 1.13; Avg. Growth 0.78; Equity Change 0.35
- Labor Input (-0.81 to 0.00)
  - All Households: Incl. Growth 0.42; Avg. Growth 0.40; Equity Change 0.02
  - Working-Age Households: Incl. Growth 0.31; Avg. Growth 0.29; Equity Change 0.02

- Interpretation of inflation scenario:
  - Initial CPI inflation value: 0.0 percent (value in 2012)
  - BoJ target: 2.0 percent
  - Using all-sample coefficients, raising inflation from 0.0 percent to 2.0 percent yields an inclusive growth boost of 1.76 percentage point.
  - For the working-age household sample, the analogous boost is 1.91 percentage point.
  - Quadratic-form implications:
    - The estimated optimal inflation level (all-sample coefficients) is 1.586/0.706 = 2.246 %; the maximum value of inclusive growth at that inflation is 1.781%.
    - With an initial value of 0.00% inflation, the upper threshold for marginal gains in inclusive growth from higher inflation is about 4.5% (explicitly 4.492%).
    - The policy scenario of increasing inflation from 0.00% to 2.00% produces a marginal effect on inclusive growth of 1.76%.

- Other scenario channel notes:
  - Duality historical/benchmark values: 21 percent in 2000-04; pre-bubble level around 5 percent in 1979-84.
  - Female participation historical: around 48 percent; male around 70 percent in 2004.
  - Labor input historical trend: decline since 1979; negative growth nationally since 1990-1994 period.

### Appendix D. Causality Direction between Inflation and Growth
- Bilateral regressions (lag l = 2) specification:
  - y_t = a_1 y_{t-1} + a_2 y_{t-2} + b_1 m_{t-1} + b_2 m_{t-2} + μ_i + v_t
- Results (Dependent variable: Growth and Inflation; Observations = 144):
  - lag(dys, 1): Growth -0.461*** (0.110); Inflation 0.080** (0.039)
  - lag(dys, 2): Growth -0.078 (0.104); Inflation 0.094** (0.037)
  - lag(Inflation, 1): Growth 2.052*** (0.251); Inflation 0.962*** (0.090)
  - lag(Inflation, 2): Growth 1.431*** (0.093); Inflation 0.816*** (0.033)
  - Goodness of fit: R2 Growth 0.748; Inflation 0.918. Adjusted R2 Growth 0.478; Inflation 0.587.
  - F Statistic (df = 4; 92): Growth 68.230***; Inflation 258.300***
- Inference:
  - Both variables are good leading indicators of the other at lags 1 and 2.
  - Coefficients on the impact of inflation on growth tend to be larger than those of the impact of growth on inflation.

- Instrumental variables estimations (lagged inflation as instrument; Balestra & Varadharajan-Krishnakumar, 1987):
  - Results for All Income (IV), Observations = 235:
    - Inflation (%) — dys: 4.707*** (1.154); dy: 4.592*** (1.071); dw: 0.051 (0.259)
    - Inflation, squared — dys: -0.876*** (0.195); dy: -0.832*** (0.181); dw: -0.034 (0.044)
    - Female labor force participation (%) — dys: 0.270* (0.137); dy: 0.225* (0.127); dw: 0.042 (0.031)
    - Initial GDP per capita — dys: -5.476*** (1.812); dy: -4.737*** (1.683); dw: -0.707* (0.407)
    - Elderly index — dys: 0.450*** (0.153); dy: 0.429*** (0.142); dw: 0.014 (0.034)
    - Goodness of fit: R2 dys 0.486; dy 0.539; dw 0.105. Adjusted R2 dys 0.374; dy 0.415; dw 0.081. F Statistic (df = 7; 181): dys 5.173***; dy 7.330***; dw 3.021***
  - Results for Working-age Income (IV), Observations = 235:
    - Inflation (%) — dys.work: 4.926*** (1.226); dy.work: 4.371*** (1.047); dw.work: 0.534 (0.324)
    - Inflation, squared — dys.work: -0.852*** (0.207); dy.work: -0.743*** (0.177); dw.work: -0.106* (0.055)
    - Female labor force participation (%) — dys.work: 0.338** (0.146); dy.work: 0.250** (0.124); dw.work: 0.088** (0.038)
    - Initial GDP per capita — dys.work: -4.770** (1.926); dy.work: -3.545** (1.645); dw.work: -1.230** (0.509)
    - Elderly index — dys.work: 0.497*** (0.162); dy.work: 0.417*** (0.139); dw.work: 0.079* (0.043)
    - Goodness of fit: R2 dys.work 0.444; dy.work 0.514; dw.work 0.007. Adjusted R2 dys.work 0.342; dy.work 0.396; dw.work 0.005. F Statistic (df = 7; 181): dys.work -0.283; dy.work 5.466***; dw.work -4.416
- Inference:
  - IV results are broadly consistent with those discussed in the main text.

### Appendix E. Contributions of regressors to variation in equity
- Definitions:
  - Standardized Regression Coefficients: expected change of dependent variable in standard deviations per one standard deviation change in explanatory variable.
  - Semi-Partial Correlation Coefficients: unique contribution of each explanatory variable to variation of dependent variable.
  - Extra Sum of Squares: amount by which sum of squared residuals is reduced by removing one variable from the full model.
- Contributions (table entries preserved):
  - Inflation
    - Standardized Regression Coefficient: -0.089
    - Semi-Partial Correlation Coefficient: 0.005***
    - Extra Sum of Squares: 0.707***
  - Inflation, squared
    - Standardized Regression Coefficient: -0.056
    - Semi-Partial Correlation Coefficient: 0.006***
    - Extra Sum of Squares: 0.321***
  - Job Openings
    - Standardized Regression Coefficient: 0.132***
    - Semi-Partial Correlation Coefficient: 0.001
    - Extra Sum of Squares: 0.089
  - FLP (Female labor force participation)
    - Standardized Regression Coefficient: 0.373***
    - Semi-Partial Correlation Coefficient: 0.007***
    - Extra Sum of Squares: 0.254
  - Labor Growth
    - Standardized Regression Coefficient: 0.069
    - Semi-Partial Correlation Coefficient: 0.0005
    - Extra Sum of Squares: 0.032
  - GDP PC
    - Standardized Regression Coefficient: -0.58***
    - Semi-Partial Correlation Coefficient: 0.004
    - Extra Sum of Squares: 0.757***
  - Elderly Index
    - Standardized Regression Coefficient: 0.079
    - Semi-Partial Correlation Coefficient: 0.0001
    - Extra Sum of Squares: 0.017
  - Total (semi-partial sum): 0.0236
- Inference:
  - Top contributors (denoted by ***) in semi-partial correlation and extra sum of squares include Inflation, Inflation squared, FLP, and GDP PC depending on the measure.

*Source: IMF staff analysis (text extracted from the provided PDF chapter/section).*

### 0.7 to close to 0.3 in most recent years

### 0.7 to close to 0.3 in most recent years

### A. Inequality and Redistribution
- Market income Gini and disposable income Gini for the working-age population are highly correlated: a simple bilateral regression suggests that, for the working-age population, each point increase in the market income Gini translates into a 0.4 point increase in the disposable income Gini.
- Goodness-of-fit for that regression is about 0.76, and the linear correlation coefficient about 0.87.
- Fiscal redistribution substantially reduces inequality for the elderly: the large gap between market income and disposable income inequality for the elderly is a sign of fiscal burden.
- OECD (2011, 2014) findings cited in the text emphasize that reliance on tax-and-transfer equalization is not efficient or sustainable given population ageing, over-stretched public finances, rising dependency ratio, and low fertility.
- Policy implication noted: Japan might need to take into account the fiscal costs of redistribution given high and rising public debt, increasing share of the elderly population, low fertility rate, and rising dependency ratio; emphasis on structural reforms, including labor market reforms, seems appropriate.

### B. Poverty
- Relative poverty rates are defined as households with income below half of median income.
- In 2009, the relative poverty indicator was around 16 percent in Japan.
- Japan’s relative poverty rate in 2009 is 4 percentage points higher than the G7 average and is the second highest among G7 countries following the US.
- Japan’s relative poverty rate is the sixth highest among OECD countries and is above the OECD average by 5 percentage points.
- Child poverty (working households with children in relative poverty) has risen rapidly since 2006 and surpassed the full-sample poverty rate in 2012.
- Single-parent households (majority single-mother households) have relative poverty rates that remained above 50 percent for 1985-2012.
- Poverty by age groups (relative poverty rates, 1985–2009 series shown):
  - Rapid increases in poverty rates among age groups 0-17 and 18-25.
  - Less marked increases for ages 26-65.
  - Declines over time for ages 66-75.
- Long-run concern: high child poverty may produce intergenerational “lock-in” effects, reducing aggregate human capital and potential growth; early education and training are emphasized (Heckman 2000).

### C. Other Dimensions of Inequality
- Two prominent dimensions linked to widening wage inequality among working-age population:
  - Gender gap: Female Labor Participation (FLP) is lower than Male Participation Rate (MPR) by about 20 percentage points in Japan.
  - Labor market duality: share of non-regular workers rose from less than 20 percent in the 1980s to above 35 percent by 2011.
- More than half of employed women are non-regular workers, implying less job security, lower wages, and reduced career opportunities.
- Low FLP and underemployment of women reduce inclusiveness and potential growth (cited IMF emphasis).
- Aoyagi and Ganelli (2013) argue excessive duality reduces productivity via a “training channel” and an “effort channel.”

### IV. Data and Empirical Strategy
- Data: prefectural level longitudinal data from the National Population Census (every 5 years), income components include wages, interest, rent, social security and other payments to households.
- Incomes deflated by prefectural GDP deflators from the Cabinet Office.
- Inclusive growth measure: Anand et al. (2013) weighted average of growth in average income and change in an equity index bounded between zero and one (one = perfectly equitable income distribution); measure equals average income growth if distribution unchanged and deviates with equity changes (see Appendix A).
- Dependent variable: log-difference (approximate percentage change) of 5-year panel between years 1979-2004; log differences standardized to an average annual rate via geometric mean (e.g., first observation represents average growth 1979-1984).
- Main policy explanatory variables (converted to 5-year averages):
  - Inflation rate: 5-year average, %
  - Ratio of new job openings for part-time to full-time labor: 5-year average, %
  - Female labor force participation ratio: 5-year average, %
  - Growth of labor input in man-hours: 5-year average, %
  - Controls: GDP per capita (initial value at the 5 year periods), Elderly index (5-year average, index; size of population 65+ divided by working-age population)
- Note: female labor participation rates used for entire workforce (15 years and older) due to data limitation at prefectural level.

### V. Results (econometric)
- Model: panel with prefectural fixed effects (Hausman test rejected random effects; Honda test rejected no individual effects). Heteroskedasticity robust standard errors reported.
- Sample: results presented for all households and for working-age households. For each: (1) Inclusive Growth; (2) Average Income Growth; (3) Equity Index Growth.
- Key regression coefficient estimates (Tables summarized with significance levels; standard errors in parentheses):

Table 1. Results for All Households (Observations 235)
- Inflation (%) — Inclusive Growth: 1.586*** (0.197); Average Income Growth: 1.598*** (0.174); Equity Index Growth: −0.032 (0.069)
- Inflation, squared — Inclusive Growth: −0.353*** (0.041); Average Income Growth: −0.331*** (0.037); Equity Index Growth: −0.020 (0.015)
- Part- to Full-time job openings (%) — Inclusive Growth: −0.028 (0.020); Average Income Growth: −0.033* (0.018); Equity Index Growth: 0.005 (0.007)
- Female labor force participation (%) — Inclusive Growth: 0.165* (0.084); Average Income Growth: 0.125* (0.074); Equity Index Growth: 0.040 (0.029)
- Labor input growth (%) — Inclusive Growth: 0.524*** (0.165); Average Income Growth: 0.489*** (0.146); Equity Index Growth: 0.029 (0.058)
- Initial GDP per capita — Inclusive Growth: −1.564** (0.762); Average Income Growth: −0.985 (0.672); Equity Index Growth: −0.600** (0.267)
- Elderly index — Inclusive Growth: 0.057 (0.037); Average Income Growth: 0.052 (0.033); Equity Index Growth: 0.004 (0.013)
- R2: Inclusive Growth 0.660; Average Income Growth 0.714; Equity Index Growth 0.112

Table 2. Results for Working-age Households (Observations 240)
- Inflation (%) — Inclusive Growth: 1.514*** (0.197); Average Income Growth: 1.512*** (0.174); Equity Index Growth: −0.008 (0.075)
- Inflation, squared — Inclusive Growth: −0.279*** (0.041); Average Income Growth: −0.264*** (0.037); Equity Index Growth: −0.015 (0.016)
- Part- to Full-time job openings (%) — Inclusive Growth: −0.039* (0.020); Average Income Growth: −0.040** (0.018); Equity Index Growth: 0.001 (0.008)
- Female labor force participation (%) — Inclusive Growth: 0.225*** (0.084); Average Income Growth: 0.155** (0.074); Equity Index Growth: 0.070** (0.032)
- Labor input growth (%) — Inclusive Growth: 0.384** (0.166); Average Income Growth: 0.358** (0.146); Equity Index Growth: 0.023 (0.063)
- Initial GDP per capita — Inclusive Growth: −0.487 (0.764); Average Income Growth: 0.036 (0.673); Equity Index Growth: −0.539* (0.291)
- Elderly index — Inclusive Growth: 0.068* (0.037); Average Income Growth: 0.057* (0.033); Equity Index Growth: 0.011 (0.014)
- R2: Inclusive Growth 0.627; Average Income Growth 0.677; Equity Index Growth 0.065

- Inflation modeled quadratically: coefficients imply an initially increasing but then falling effect of inflation on inclusive growth as inflation rises.
- Estimated maximum effect on inclusive growth reached at around 2 percent inflation rate.
- Inflation has significant positive effects on inclusive growth and average income growth; negative but small and insignificant effect on equity index growth.
- Robustness: pseudo-Granger causality tests and IV estimates using lagged inflation (Appendix D) broadly confirm causality from inflation to growth.
- Labor market duality (part- to full-time ratio) has a negative and significant impact on inclusive growth via negative impact on average income growth; effect stronger in working-age sample.
- Female labor participation positively affects inclusive growth primarily via average income growth; for working-age sample, positive effects on both average income and equity index growth.
- Labor input growth increases inclusive growth mainly by increasing average income.
- Initial GDP per capita is a negative and significant determinant of inclusive growth for all-household sample (consistent with Solow/Kuznets perspectives).
- Elderly index: insignificant for all households, marginally positive and significant for working-age sample (possible explanation via asset returns included in income measure).
- Overall: policy variables affect inclusive growth mostly through average income growth; potential impact of a complete Abenomics package on income equality is relatively small.
- Robustness checks using alternative weights on the equity index and different assumptions on income distribution confirm robustness (details not reported here).

### VI. Policy Implications (scenario analysis summary)
- Scenarios present marginal effects of changes in each policy variable on inclusive growth, holding other things constant, using base level of inclusive growth at national level during 2000-04:
  - Base level of inclusive growth (national, 2000-04): −1.46 percent for the entire household sample.
- Main policy-relevant findings:
  - Moving towards a 2 percent inflation target (BoJ goal) is estimated to promote growth and raise inclusive growth, with diminishing returns and a turning point around 2 percent inflation.
  - Reducing labor market duality (lower ratio of part-time to full-time new job openings) is estimated to increase inclusive growth by raising average income growth.
  - Increasing female labor force participation is estimated to raise inclusive growth by increasing average income growth and, in the working-age sample, improving equity index growth.
  - Increasing labor input (man-hours) is estimated to boost inclusive growth mainly through average income growth.
- Overall conclusion: a “complete” Abenomics package that combines exiting deflation (toward 2 percent inflation), labor market reforms reducing duality, policies raising female labor participation, and measures increasing labor input is estimated to raise inclusive growth mainly by increasing average income growth. The direct effect on income equality (equity index) is limited in the estimated results.

*Source: IMF staff analysis (text extracted from the provided PDF chapter/section)._

### 1.15 percent for the working-age household sample, as references to evaluate the magnitude

### _wp1554 - 1.15 percent for the working-age household sample, as references to evaluate the magnitude

### Scenario results — Inclusive Growth (Table 3)
- Scenario dimensions reported for "All Households" and "Working-Age Households" with three component outcomes: Incl. Growth, Avg. Growth, Equity Change.
- Inflation (0.0 to 2.0%)
  - All Households: Incl. Growth 1.76; Avg. Growth 1.87; Equity Change -0.14
  - Working-Age Households: Incl. Growth 1.91; Avg. Growth 1.97; Equity Change -0.07
- Part- to Full-time job postings ratio (21 to 5%)
  - All Households: Incl. Growth 0.45; Avg. Growth 0.53; Equity Change -0.08
  - Working-Age Households: Incl. Growth 0.63; Avg. Growth 0.64; Equity Change -0.01
- Female Labor Participation (47 to 52%)
  - All Households: Incl. Growth 0.83; Avg. Growth 0.63; Equity Change 0.20
  - Working-Age Households: Incl. Growth 1.13; Avg. Growth 0.78; Equity Change 0.35
- Labor Input (-0.81 to 0.00)
  - All Households: Incl. Growth 0.42; Avg. Growth 0.40; Equity Change 0.02
  - Working-Age Households: Incl. Growth 0.31; Avg. Growth 0.29; Equity Change 0.02

### Interpretation of inflation effects (quadratic specification)
- Baseline and target used in scenarios:
  - Initial CPI inflation value: 0.0 percent (value in 2012)
  - BoJ target: 2.0 percent
- Key quantitative interpretations:
  - Using all-sample coefficients, raising inflation from 0.0 percent to 2.0 percent yields an inclusive growth boost of 1.76 percentage point.
  - For the working-age household sample, the analogous boost is 1.91 percentage point, sufficient to raise a negative growth outcome above zero.
  - Gains from higher inflation arise mostly from an increase in average income growth; the equity index deteriorates (inequality rises) as inflation increases.
- Quadratic-form implications:
  - The estimated optimal inflation level (all-sample coefficients) is 1.586/0.706 = 2.246 %; the maximum value of inclusive growth at that inflation is 1.781%.
  - Marginal effects of inflation depend on the initial inflation level.
  - With an initial value of 0.00% inflation, the upper threshold for marginal gains in inclusive growth from higher inflation is about 4.5% (explicitly 4.492% from symmetry around the optimal 2.246%).
  - The policy scenario of increasing inflation from 0.00% to 2.00% produces a marginal effect on inclusive growth of 1.76%.

### Other scenario interpretable channels
- Labor market duality (flow measure: ratio of new part-time to full-time job postings)
  - Historical/benchmark values: 21 percent in 2000-04 (5-year average); pre-bubble level around 5 percent in 1979-84.
  - Scenario: ratio returns to pre-bubble level (21% to 5%) → inclusive growth up by 0.45 percentage points (all-sample).
  - Effect composition: bulk of boost from average income growth; small deterioration in equality.
  - Duality measure correlation: correlated with net capital investment (correlation coefficient around 0.6).
- Female labor participation
  - Historical: female participation around 48 percent; male around 70 percent in 2004.
  - Scenario: female participation increases by 5 percentage points (47 to 52% used in table).
    - All-sample: inclusive growth +0.83 percentage points.
    - Working-age sample: inclusive growth +1.13 percentage points; Avg. Growth +0.78; Equity +0.35.
  - Interpretation: ambitious but feasible; increases both average income growth and income equality (especially in working-age sample).
- Labor input (man-hours)
  - Historical trend: decline since 1979; negative growth nationally since 1990-1994 period.
  - Scenario: shift labor input growth from -0.81 to 0.00 (i.e., stop decline).
    - All-sample: inclusive growth +0.42 percent.
    - Relative magnitude: about one third of the latest level; less effective than achieving the inflation target or addressing dual labor market duality.

### Policy implications and conclusions
- Primary policy message:
  - Fully implementing structural reforms (the "third arrow") is necessary for Abenomics to foster inclusive growth that combines higher average income growth with improved or stable income equality.
- Specific implications drawn from scenarios and regressions:
  - Achieving the BoJ 2 percent inflation target alone (first arrow) stimulates average income growth but tends to worsen income equality.
  - Combining monetary policy (inflation target) with structural reforms (higher female labor participation, reduced duality, increased labor input) raises average income growth while keeping inequality broadly unchanged (all-household sample) or improving it (working-age household sample).
  - If structural reforms are not implemented and monetary policy becomes overburdened, runaway inflation above the BOJ’s 2 percent target would reduce overall inclusive growth (not only equality).
- Concluding assessment:
  - If all arrows of Abenomics are fully launched, the policy framework can be effective in promoting both growth and income equality, thereby fostering inclusive growth.

### Data sources and key variable definitions (selected)
- Income Deciles: Every 5 years, 10 thousands yen, 1975 price; National survey of family income and expenditure, Ministry of Internal Affairs and Communications (MIC)
- Inflation rate: 5-year average, %; CPI reports, MIC
- Ratio of new job openings for part-time to full-time labor at job security office: 5-year average, %; Public Employment Security Office
- Female labor force participation ratio: 5-year average, %; e-stat database, MIC
- Growth of labor input in man-hours (workers * annual workhours / 1000): 5-year average, %; JIP Database, Research Institute of Economy, Trade and Industry
- Real GDP per capita of Prefectures: Initial value at the 5 year periods, 1 thousand yen, 1975 price; Prefectural Statistics, Cabinet Office

*Source: _wp1554 - 1.15 percent for the working-age household sample, as references to evaluate the magnitude (IMF PDF content provided).*

### Appendix D. Causality Direction between Inflation and Growth

### Appendix D. Causality Direction between Inflation and Growth

### Bilateral regressions between growth and inflation (endogeneity test)
- Regression specification:
  - y_t = a_1 y_{t-1} + a_2 y_{t-2} + b_1 m_{t-1} + b_2 m_{t-2} + μ_i + v_t
  - where y is the dependent variable and m is the (supposedly endogenous) explanatory variable.
  - Notes: lag limited by data availability (l = 2); regressions run separately; specification contains fixed effect for each prefecture; not a conventional Granger causality test.

- Results (Dependent variable: Growth (column 1) and Inflation (column 2); Observations = 144):
  - lag(dys, 1)
    - Growth: -0.461***
      - (0.110)
    - Inflation: 0.080**
      - (0.039)
  - lag(dys, 2)
    - Growth: -0.078
      - (0.104)
    - Inflation: 0.094**
      - (0.037)
  - lag(Inflation, 1)
    - Growth: 2.052***
      - (0.251)
    - Inflation: 0.962***
      - (0.090)
  - lag(Inflation, 2)
    - Growth: 1.431***
      - (0.093)
    - Inflation: 0.816***
      - (0.033)
  - Goodness of fit:
    - R2: Growth 0.748; Inflation 0.918
    - Adjusted R2: Growth 0.478; Inflation 0.587
    - F Statistic (df = 4; 92): Growth 68.230***; Inflation 258.300***

- Inference:
  - Both variables are good leading indicators of the other at lags 1 and 2.
  - Coefficients on the impact of inflation on growth tend to be larger than those of the impact of growth on inflation.
  - Note on significance notation: * p<0.1; ** p<0.05; *** p<0.01

### Instrumental variables estimations (lagged inflation as instrument; Balestra & Varadharajan-Krishnakumar, 1987)
- Setup: Same model as benchmark case but with lagged (l = 1) inflation (5 year averages) as an instrument variable.

- Results for All Income (IV) — Dependent variables: dys (col 1), dy (col 2), dw (col 3); Observations = 235
  - Inflation (%)
    - dys: 4.707***
      - (1.154)
    - dy: 4.592***
      - (1.071)
    - dw: 0.051
      - (0.259)
  - Inflation, squared
    - dys: -0.876***
      - (0.195)
    - dy: -0.832***
      - (0.181)
    - dw: -0.034
      - (0.044)
  - Part- to Full-time job openings (%)
    - dys: 0.031
      - (0.038)
    - dy: 0.023
      - (0.035)
    - dw: 0.007
      - (0.009)
  - Female labor force participation (%)
    - dys: 0.270*
      - (0.137)
    - dy: 0.225*
      - (0.127)
    - dw: 0.042
      - (0.031)
  - Labor input growth (%)
    - dys: 0.156
      - (0.291)
    - dy: 0.136
      - (0.270)
    - dw: 0.019
      - (0.065)
  - Initial GDP per capita
    - dys: -5.476***
      - (1.812)
    - dy: -4.737***
      - (1.683)
    - dw: -0.707*
      - (0.407)
  - Elderly index
    - dys: 0.450***
      - (0.153)
    - dy: 0.429***
      - (0.142)
    - dw: 0.014
      - (0.034)
  - Goodness of fit:
    - R2: dys 0.486; dy 0.539; dw 0.105
    - Adjusted R2: dys 0.374; dy 0.415; dw 0.081
    - F Statistic (df = 7; 181): dys 5.173***; dy 7.330***; dw 3.021***

- Results for Working-age Income (IV) — Dependent variables: dys.work (col 1), dy.work (col 2), dw.work (col 3); Observations = 235
  - Inflation (%)
    - dys.work: 4.926***
      - (1.226)
    - dy.work: 4.371***
      - (1.047)
    - dw.work: 0.534
      - (0.324)
  - Inflation, squared
    - dys.work: -0.852***
      - (0.207)
    - dy.work: -0.743***
      - (0.177)
    - dw.work: -0.106*
      - (0.055)
  - Part- to Full-time job openings (%)
    - dys.work: 0.025
      - (0.040)
    - dy.work: 0.014
      - (0.034)
    - dw.work: 0.011
      - (0.011)
  - Female labor force participation (%)
    - dys.work: 0.338**
      - (0.146)
    - dy.work: 0.250**
      - (0.124)
    - dw.work: 0.088**
      - (0.038)
  - Labor input growth (%)
    - dys.work: -0.020
      - (0.309)
    - dy.work: 0.018
      - (0.264)
    - dw.work: -0.039
      - (0.082)
  - Initial GDP per capita
    - dys.work: -4.770**
      - (1.926)
    - dy.work: -3.545**
      - (1.645)
    - dw.work: -1.230**
      - (0.509)
  - Elderly index
    - dys.work: 0.497***
      - (0.162)
    - dy.work: 0.417***
      - (0.139)
    - dw.work: 0.079*
      - (0.043)
  - Goodness of fit:
    - R2: dys.work 0.444; dy.work 0.514; dw.work 0.007
    - Adjusted R2: dys.work 0.342; dy.work 0.396; dw.work 0.005
    - F Statistic (df = 7; 181): dys.work -0.283; dy.work 5.466***; dw.work -4.416
      - Note: significance notation as above.

- Inference:
  - IV results are broadly consistent with those discussed in the main text.

### Appendix E — Contributions of regressors to variation in equity
- Definitions:
  - Standardized Regression Coefficients: expected change of dependent variable in standard deviations per one standard deviation change in explanatory variable (example provided: one standard deviation change in FLP is associated with a 0.373 standard deviation change in equity index growth).
  - Semi-Partial Correlation Coefficients: unique contribution of each explanatory variable to variation of dependent variable (do not account for joint explanation).
  - Extra Sum of Squares: amount by which sum of squared residuals is reduced by removing one variable from the full model (example: removing inflation increases SSR by 0.707; removing inflation squared increases SSR by 0.321).

- Contributions of regressors to variation in equity (table entries preserved):
  - Inflation
    - Standardized Regression Coefficient: -0.089
    - Semi-Partial Correlation Coefficient: 0.005***
    - Extra Sum of Squares: 0.707***
  - Inflation, squared
    - Standardized Regression Coefficient: -0.056
    - Semi-Partial Correlation Coefficient: 0.006***
    - Extra Sum of Squares: 0.321***
  - Job Openings
    - Standardized Regression Coefficient: 0.132***
    - Semi-Partial Correlation Coefficient: 0.001
    - Extra Sum of Squares: 0.089
  - FLP (Female labor force participation)
    - Standardized Regression Coefficient: 0.373***
    - Semi-Partial Correlation Coefficient: 0.007***
    - Extra Sum of Squares: 0.254
  - Labor Growth
    - Standardized Regression Coefficient: 0.069
    - Semi-Partial Correlation Coefficient: 0.0005
    - Extra Sum of Squares: 0.032
  - GDP PC
    - Standardized Regression Coefficient: -0.58***
    - Semi-Partial Correlation Coefficient: 0.004
    - Extra Sum of Squares: 0.757***
  - Elderly Index
    - Standardized Regression Coefficient: 0.079
    - Semi-Partial Correlation Coefficient: 0.0001
    - Extra Sum of Squares: 0.017
  - Total (semi-partial sum)
    - 0.0236

- Inference:
  - Top contributors (denoted by ***) in semi-partial correlation and extra sum of squares include Inflation, Inflation squared, FLP, and GDP PC depending on the measure.

*Source: Appendix D and Appendix E, _wp1554 - Appendix D. Causality Direction between Inflation and Growth*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2015/_wp1554.pdf_