## _wp16245 - 2013. It incorporates data from several sources (United Nations University’s World Income

## Source details

**Canonical URL:** [_wp16245 - 2013. It incorporates data from several sources (United Nations University’s World Income](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp16245.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp16245.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp16245.pdf.json)

---

### Data and inequality measures
- Data sources:
  - United Nations University’s World Income Inequality Database; OECD Income Distribution Database; World Bank; Eurostat; the Luxembourg Income Study (standardized; see Solt 2016 for methodology).
  - Additional robustness data: World Top Incomes Databases (WTID) and the share of wage income in GDP from the OECD.
- Gini coefficient bounds and sample ranges:
  - Theoretically bounded between 0 and 100.
  - Sample ranges: 18 to 54 for net measures; 30 to 57 for gross measures.
- Sample period and coverage:
  - Sample period for main analysis: 1990 to 2013.
  - Country coverage for panel estimates: unbalanced panel of 32 economies.

### Identification of monetary policy shocks
- Rationale and approach:
  - Construct “exogenous” monetary policy shocks using forecast errors so changes in monetary policy rates are not driven by contemporary economic conditions affecting inequality.
- Forecast-error construction:
  - Forecast error = actual policy rate − rate expected by analysts (Consensus Economics).
  - When policy rate data unavailable, short-term (typically 3 months) rates used as proxy.
- Purging endogenous responses:
  - Forecast errors of policy rates are regressed on forecast errors of inflation and output growth; the residual captures exogenous monetary policy shocks.
  - Addresses policy foresight and endogeneity from monetary policy responding to contemporaneous news.
- Validation:
  - Comparison with Romer and Romer (2004) for the United States: correlation about 0.8.
  - Local-projection responses of output, unemployment, and inflation to these shocks are statistically significant and comparable to literature benchmarks.

### Empirical methodology (local projections)
- Estimation details:
  - Jorda (2005) local projection method on annual data to estimate impulse response functions.
  - For horizons k = 0, ..., 4 estimate changes in (log of) net or market income inequality on exogenous monetary policy shocks.
  - Controls: country fixed effects, time fixed effects, lagged monetary policy shocks, lagged changes in inequality.
  - Country fixed effects control for unobserved cross-country heterogeneity and measurement differences.
- Horizons and inference:
  - Horizons analyzed: up to five years after the shock (k = 0 to 4).
  - Standard errors: clustered robust standard errors.
  - Sample period: 1990 to 2013.
  - Results robust to different lag specifications.

### Baseline results and key quantitative findings
- Main outcome:
  - Monetary policy tightening leads to a long-lasting increase in income inequality (net Gini for disposable income).
- Magnitudes (verbatim):
  - "An unanticipated policy rate increase of 100 basis increases the Gini index by about 1¼ in the very short term—1 year after the shock—and by about 2¼ percent in the medium term—5 years after the shock."
  - "The effect eventually levels off, after seven years, at about 2½ percent."
- Economic significance:
  - The medium-term effect is approximately equivalent to 1 standard deviation of the change in the Gini coefficient (2.4 percent) in the sample—or to about 8 percentage points.
  - The effect is larger than the sample average cumulative increase of the Gini coefficient over five consecutive years (about 2 percent).
- Comparisons:
  - Effects—particularly for unemployment—are comparable and not statistically different from Romer and Romer (2004) for the United States.
  - Estimated effect is slightly larger—but not statistically different—than Coibion and others (2012) for the United States (about 3-5 percentage points—that is, about 1.5 percent). [text truncated in source]

### Robustness checks
- Gross Gini:
  - Re-estimating equation (3) with gross Gini yields effects slightly smaller but not statistically different from net inequality (Panel A, Figure 5).
- Subsamples:
  - Advanced economies only: results quantitatively similar and not statistically significantly different from full sample (Panel B, Figure 5).
  - 1990-2007 (pre-2008): unanticipated 100 basis shock increases Gini by about 1½ percent after 1 year and by about 2¾ percent after 5 years (Panel C, Figure 5); larger than full-sample baseline but not statistically significantly different.
- Controls for endogeneity:
  - Adding (i) recession dummies and (ii) change in the budget balance confirms robustness (Panel D, Figure 5).

### Different measures of inequality
- Top income shares (WTID):
  - A 100 basis points monetary policy shock has significant and persistent effects on top income shares.
  - Largest effect for top 1 percent: about 0.8 percentage point four years after the shock (Figure 6).
- Wage share (OECD):
  - Unexpected 100 basis points increase in policy rates leads to decline in share of wage income in GDP:
    - About ½ percentage point in the short term.
    - About 1½ percentage points in the medium term (Figure 7).

### Type of monetary policy shocks
- Positive versus negative shocks:
  - Positive (contractionary) shocks lead to statistically significant increase in inequality in the medium term—about 4½ percent (Figure 8).
  - Medium-term effect of negative (expansionary) shocks not statistically significantly different from zero (Figure 8).
- Exogenous versus growth-driven shocks (Equation (5) decomposition):
  - Changes in policy rates driven by economic activity (predicted FE) are associated with a decrease in inequality (Figure 9, Panel A).
  - Unanticipated changes in policy rates (forecast errors) increase inequality in short and medium term; effect smaller than for unanticipated exogenous changes (Figure 9, Panel B).
- Responses to policy-rate changes (STt_i rather than forecast errors):
  - A 100 basis point increase in the policy rate raises inequality by about 0.14 percent after one year; effect less persistent than baseline (Figure 9, Panel C).
  - Two potential biases: (i) policy-rate changes may be driven by economic activity; (ii) monetary policy foresight.

### Role of macroeconomic conditions
- Business cycle nonlinearities (equation (6)):
  - On average, monetary policy shocks have larger effects on inequality during expansions than in recessions (Figure 10, Panel A and B).
  - Heterogeneity by shock sign:
    - Positive shocks have larger effects in expansions (Figure 10, Panel C and D).
    - Negative shocks have larger effects during recessions (Figure 10, Panel E and F).
- Labor earnings:
  - Higher share of labor income amplifies inequality effect of monetary policy shocks.
  - Re-estimating with z = normalized labor income share shows larger effects on inequality in countries with higher labor shares over first five years (Figure 11).
- Redistribution policies:
  - Using z = normalized redistribution (market minus net inequality) shows redistribution matters (Figure 12).
  - Monetary policy shocks increase inequality and are statistically significant in countries with limited redistribution (below sample average); effect not statistically significantly different from zero in countries with relatively high redistribution.

### Wealth inequality and asset-price channels
- Asset-type heterogeneity:
  - Bond prices: no significant effects on inequality.
  - Equity price inflation: increases wealth inequality.
  - House price inflation: reduces inequality; magnitude (absolute value) larger than same-size increase in equity prices.
- Asset-price responses to a 100 basis point tightening (Figure 13):
  - Short term: tightening increases housing prices and reduces equity prices.
  - Medium term: tightening reduces housing and equity prices by a similar amount.
- Tentative inference:
  - Expansionary monetary policy increases wealth inequality in the short term but decreases it in the medium term (based on asset-price responses and prior evidence on asset-price effects on inequality).

### Conclusions and policy-relevant findings
- Main quantitative summary (verbatim framing):
  - Using unexpected changes in policy rates orthogonal to innovations in output growth and inflation, an unexpected decrease of 100 basis points in the policy rate:
    - Reduces inequality by about 1¼ percent in the short term.
    - Reduces inequality by about 2¼ percent in the medium term.
- Economic significance:
  - Medium-term effect ≈ standard deviation of change in Gini (2.4 percent) in the sample.
- Heterogeneity and policy implications:
  - Larger inequality effects for positive (tightening) shocks, especially during expansions.
  - Larger effects in countries with higher labor share of income and smaller redistribution policies.
  - Unanticipated increases in policy rates increase inequality, whereas policy-rate changes driven by increased growth are associated with lower inequality.
- Sample note:
  - Analysis covers a panel of 32 advanced and emerging market economies.

### Key descriptive statistics (Table 1)
- Inequality:
  - Net Gini: Mean 33.6; SD 7.2; Min 18.0; Max 54.1
  - Gross Gini: Mean 45.5; SD 5.2; Min 30.0; Max 57.3
- Top income shares:
  - Top 1 percent: Mean 11.1; SD 4.3; Min 5.2; Max 23.5
  - Top 5 percent: Mean 24.6; SD 5.7; Min 14.6; Max 38.8
  - Top 10 percent: Mean 35.9; SD 6.6; Min 23.6; Max 50.6
- Share of wage income in GDP: Mean 51.5; SD 4.0; Min 39.2; Max 58.8
- Monetary policy shocks: Mean 0.0; SD 1.13; Min -3.9; Max 4

### Regression evidence on monetary policy shocks and Net Gini (Table 2, selected coefficients)
- Dependent variable: income inequality (Net Gini), sample period 1990-2013. Estimates based on equation (3).
- Monetary policy shock (t) coefficients (coefficient (t-statistic)):
  - K=0: 0.531 (1.56)
  - K=1: 1.176** (2.13)
  - K=2: 1.886*** (2.88)
  - K=3: 2.018*** (2.59)
  - K=4: 2.154** (2.40)
- Gini growth (t-1) selected coefficients (coefficient (t-statistic)):
  - K=0: 0.176** (2.16)
  - K=1: 0.266** (2.53)
  - K=2: 0.080 (0.58)
  - K=3: -0.074 (-0.49)
  - K=4: -0.175 (-1.12)
- Sample sizes and fit:
  - N: K=0 -> 467; K=1 -> 439; K=2 -> 409; K=3 -> 379; K=4 -> 349
  - R^2: K=0 -> 0.16; K=1 -> 0.18; K=2 -> 0.19; K=3 -> 0.24; K=4 -> 0.28
- Note: T-statistics based on robust clustered standard errors in parenthesis. ***,**,* denote significance at 1 percent, 5 percent and 10 percent, respectively.

### Sample composition (selected countries)
- Advanced Economies (selected): Australia; Canada; Czech Republic; France; Germany; Hong Kong SAR; Italy; Japan; Korea; Netherlands; New Zealand; Norway; Singapore; Slovak Republic; Spain; Sweden; Switzerland; Taiwan Province of China; United Kingdom; United States.
- Emerging Market Countries (selected): Argentina; Brazil; Chile; Hungary; India; Indonesia; Malaysia; Mexico; Philippines; Poland; Thailand; Turkey.

*Source: _wp16245 - 2013. It incorporates data from several sources (United Nations University’s World Income*

### 2013. It incorporates data from several sources (United Nations University’s World Income

### _wp16245 - 2013. It incorporates data from several sources (United Nations University’s World Income

### Data and inequality measures
- Data sources incorporated: United Nations University’s World Income Inequality Database, the OECD Income Distribution Database, World Bank, Eurostat, the Luxembourg Income Study; standardized (see Solt 2016 for methodology).
- Additional robustness data: top income shares from the World Top Incomes Databases (WTID) and the share of wage income in GDP from the OECD.
- Gini coefficient bounds and sample ranges:
  - Theoretically bounded between 0 and 100.
  - Sample ranges: 18 to 54 for net measures; 30 to 57 for gross measures.
- Sample period for main analysis: 1990 to 2013.
- Country coverage for panel estimates: unbalanced panel of 32 economies.

### Identification of monetary policy shocks
- Rationale:
  - Changes in monetary policy rates are not driven by inequality, but economic conditions can influence both inequality and monetary policy actions.
  - To estimate causal effects, the paper constructs “exogenous” monetary policy shocks by using forecast errors.
- Forecast-error construction:
  - Forecast error of policy rates is computed as the difference between actual policy rates and the rate expected by analysts (forecasts from Consensus Economics).
  - When policy rate data are unavailable, short-term (typically 3 months) rates are used as a proxy.
- Purging endogenous responses:
  - Forecast errors of policy rates are regressed on forecast errors of inflation and output growth; the residual captures exogenous monetary policy shocks.
  - This approach addresses:
    - Policy foresight by aligning econometrician and agents’ information sets.
    - Endogeneity from monetary policy responding to contemporaneous news in growth and inflation.
- Validation exercises:
  - Comparison with Romer and Romer (2004) for the United States: the constructed shocks exhibit a very similar pattern and are highly correlated (correlation about 0.8).
  - Local-projection responses of output, unemployment, and inflation to these shocks are statistically significant and comparable to literature benchmarks, particularly Romer and Romer (2004).

### Empirical methodology (local projections)
- Estimation approach:
  - Use Jorda (2005) local projection method on annual data to estimate impulse response functions directly.
  - For each horizon k (k = 0, ..., 4), estimate changes in (log of) net or market income inequality on exogenous monetary policy shocks, controlling for country fixed effects, time fixed effects, lagged monetary policy shocks, and lagged changes in inequality.
  - Country fixed effects control for unobserved cross-country heterogeneity and for differences in measurement (income vs. consumption data).
- Horizons analyzed: up to five years after the shock (k = 0 to 4).
- Standard errors: clustered robust standard errors.
- Sample period: 1990 to 2013.
- Results are robust to different lag specifications.

### Baseline results and key quantitative findings
- Main outcome: Monetary policy tightening leads to a long-lasting increase in income inequality (measured by net Gini for disposable income).
- Magnitudes (preserve exact phrasing and values from source):
  - "An unanticipated policy rate increase of 100 basis increases the Gini index by about 1¼ in the very short term—1 year after the shock—and by about 2¼ percent in the medium term—5 years after the shock."
  - "The effect eventually levels off, after seven years, at about 2½ percent."
- Economic significance:
  - The medium-term effect is approximately equivalent to 1 standard deviation of the change in the Gini coefficient (2.4 percent) in the sample—or to about 8 percentage points.
  - The effect is larger than the sample average cumulative increase of the Gini coefficient over five consecutive years (about 2 percent).
- Comparisons:
  - The estimated effects—particularly for unemployment—are comparable and not statistically different from Romer and Romer (2004) for the United States.
  - The estimated effect is slightly larger—but not statistically different—than Coibion and others (2012) for the United States (about 3-5 percentage points—that is, about 1.1- [text truncated in source]).

### Robustness and interpretation notes
- Robustness checks include:
  - Using top income shares (WTID) and wage-income shares (OECD).
  - Cross-checking shock series against established shock series (Romer and Romer 2004).
  - Verifying macroeconomic responses (output, unemployment, inflation) to shocks via local projections.
- Methodological advantages:
  - Forecast-error approach mitigates policy foresight and endogeneity concerns by aligning information sets and purging growth/inflation news.
- Limitations noted in source:
  - Higher-frequency assessment of impacts on monthly or quarterly variables would require higher-frequency shocks; annual shocks are still informative for validating shocks and transmission channels.
  - Given limited time series, certain country-specific high-frequency validations cannot be performed.

*Source: _wp16245 - 2013. It incorporates data from several sources (United Nations University’s World Income*

### 1.5 percent).

### _wp16245 - 1.5 percent).

### Robustness checks
- Re-estimating equation (3) with the gross Gini (market measure) yields effects on gross inequality that are slightly smaller but not statistically different from those for net inequality (Panel A, Figure 5).
- Re-estimating equation (3) for advanced economies only produces results quantitatively similar to and not statistically significantly different from the full sample (Panel B, Figure 5).
- Re-estimating equation (3) for the 1990-2007 sample (pre-2008) shows larger effects: an unanticipated policy rate shock of 100 basis increases the Gini index by about 1½ percent after 1 year and by about 2¾ percent after 5 years (Panel C, Figure 5). These estimates are larger than the full-sample baseline but not statistically significantly different from it.
- Controlling for possible endogeneity by adding: (i) recession dummies (periods of negative growth) and (ii) change in the budget balance, confirms robustness of results (Panel D, Figure 5).

### Different measures of inequality
- Concerns about SWIID imputations and measurement error motivate using alternative measures: (i) top income share series (10, 5, and 1 percent) from WTID; (ii) share of wage income in GDP from the OECD.
- Top income shares: a 100 basis points monetary policy shock has a significant and persistent effect on all top income shares; the effect is largest for the top 1 percent—about 0.8 percentage point four years after the shock (Figure 6).
- Wage share: an unexpected 100 basis points increase in policy rates leads to a statistically significant and persistent decline in the share of wage income in GDP—about ½ percentage point in the short term and about 1½ percentage points in the medium term (Figure 7).

### Type of monetary policy shocks
- Positive versus negative shocks:
  - Positive (contractionary) monetary policy shocks lead to a statistically significant increase in inequality in the medium term—about 4½ percent (Figure 8).
  - Medium-term effect of negative (expansionary) shocks is not statistically significantly different from zero (Figure 8).
- Exogenous versus growth-driven shocks (Equation (5) decomposition):
  - Changes in policy rates driven by changes in economic activity (predicted FE) are associated with a decrease in inequality (Figure 9, Panel A). This likely reflects improved economic conditions rather than a pure causal effect of the policy rate.
  - Unanticipated changes in policy rates (forecast errors) increase inequality in the short and medium term; their effect is smaller than for unanticipated exogenous changes (Figure 9, Panel B).
- Response to changes in policy rates themselves (STt_i rather than forecast errors):
  - A 100 basis point increase in the policy rate raises inequality by about 0.14 percent after one year and the effect is less persistent than the baseline (Figure 9, Panel C).
  - Two potential biases: (i) changes in policy rates may be driven by economic activity (confounding effect); (ii) monetary policy foresight—agents receive news in advance and adjust behavior prior to rate changes.

### Role of Macroeconomic Conditions
- Role of the business cycle (nonlinear specification, equation (6)):
  - On average, monetary policy shocks have larger effects on inequality during expansions than in recessions (Figure 10, Panel A and B).
  - Heterogeneity by shock sign:
    - Positive shocks have larger effects in expansions (Figure 10, Panel C and D).
    - Negative shocks have larger effects during recessions (Figure 10, Panel E and F).
- Role of labor earnings:
  - Hypothesis: higher share of labor income amplifies the inequality effect of monetary policy shocks because bottom-of-distribution labor earnings are more cyclical.
  - Re-estimating with z equal to normalized labor income share shows larger effects of monetary policy shocks on inequality in countries with higher labor shares over the first five years after the shock (Figure 11).
- Role of redistribution policies:
  - Using z as normalized redistribution (market minus net inequality) shows that redistribution matters (Figure 12).
  - Monetary policy shocks increase inequality and are statistically significant in countries with limited redistribution (below sample average), while the effect is not statistically significantly different from zero in countries with relatively high redistribution.

### Wealth inequality
- Asset-price channels differ by asset:
  - Bond prices: no significant effects on inequality.
  - Equity price inflation: increases wealth inequality.
  - House price inflation: reduces inequality; magnitude (absolute value) larger than same-size increase in equity prices.
- Implication: examine housing and equity price responses to infer wealth inequality effects.
- Estimating equation (3) with housing and equity prices as dependent variables (Figure 13):
  - Monetary policy tightening increases housing prices and reduces equity prices in the short term.
  - In the medium term, tightening reduces housing and equity prices by a similar amount.
  - Tentative inference: expansionary monetary policy increases wealth inequality in the short term but decreases it in the medium term (based on asset-price responses and previous findings on asset-price effects on inequality).

### Conclusions and key quantitative findings
- Using unexpected changes in policy rates orthogonal to innovations in output growth and inflation, an unexpected decrease of 100 basis points in the policy rate:
  - Reduces inequality by about 1¼ percent in the short term.
  - Reduces inequality by about 2¼ percent in the medium term.
- The medium-term effect is economically significant: approximately equivalent to a standard deviation of the change in the Gini coefficient (2.4 percent) in the sample.
- Heterogeneity in effects:
  - Larger effects for positive (tightening) shocks, especially during expansions.
  - Larger effects in countries with higher labor share of income and smaller redistribution policies.
  - Unanticipated increases in policy rates increase inequality, whereas changes in policy rates driven by increased growth are associated with lower inequality.
- Sample note: analysis covers a panel of 32 advanced and emerging market economies.

*Source: IMF working paper content provided in the supplied PDF excerpt.*

### References

### _wp16245 - References

### Key bibliographic sources cited
- Acemoglu, Daron and Simon Johnson (2012) ‘Who Captured the Fed?’, New York Times, 29.
- Auerbach, Alan, and Yuriy Gorodnichenko. 2013a. “Fiscal Multipliers in Recession and Expansion.” In Fiscal Policy After the Financial Crisis, eds. Alberto Alesina and Francesco Giavazzi, NBER Books.
- Auerbach, Alan, and Yuriy Gorodnichenko. 2013b. “Measuring the Output Responses to Fiscal Policy.” American Economic Journal: Economic Policy 4 (2): 1–27.
- Bernanke, B. S. (2015), “Monetary Policy and Inequality” (Brookings blog).
- Bernanke, B. S., and K. N. Kuttner (2005), “What explains the stock market’s reaction to Federal Reserve policy?”, The Journal of Finance, Vol. 60/3, pp. 1221-1257.
- Coibion, O., Y. Gorodnichenko, L. Kueng and J. Silvia (2012) ‘Innocent bystanders? Monetary policy and inequality in the US’, NBER Working Paper 18170.
- Jaumotte, F. and C. Osorio Buitron (2015), “Inequality and labor market institutions”, IMF Staff Discussion Note, No. 15/14.
- Leeper, Eric M., Alexander W. Richter, and Todd B. Walker. 2012. “Quantitative Effects of Fiscal Foresight.” American Economic Journal: Economic Policy 4 (2): 115–44.
- Yellen, J. (2014), “Perspectives on inequality and opportunity from the survey of consumer finances”, speech at the Conference on Economic Opportunity and Inequality, Federal Reserve Bank of Boston, October 17.
- Zdzienicka, A., S. Chen, F. Diaz Kalan, S. Laseen, and K. Svirydzenka (2015) “Effects of Monetary and Macroprudential Policies on Financial Conditions: Evidence form the United States”, IMF Working Paper, 15/288.

(Note: full list of references appears in the source document; above are selected citations that appear in the supplied content.)

### Empirical findings and visual evidence (figures)
- Figure 1: Evolution of Inequality (1990-2013)
  - Displays “Inequality in Net Income (1990-2013)” and “Inequality in Market Income (1990-2013)”.
  - Series plotted for: All sample, AE, EM; solid line indicates average mean estimates; shaded area indicates difference between minimum and maximum values.
  - Source: Standardized World Income Inequality Database v5.0 (Solt, 2014).

- Figure 2: Exogenous Monetary Policy Shocks for the United States (percentage points)
  - Two series: “Ours” (shocks from equations (1) and (2)) and “R&R” (Romer and Romer, 2004).
  - Reported correlation = 0.79.

- Figure 3: The effect of monetary policy on output, unemployment and inflation
  - Responses to an unanticipated increase in monetary policy rates of 100 basis points; 90 percent confidence bands shown.
  - Panels: Output (%), Unemployment (ppt), CPI level (%). t=0 is year of shock. Estimates based on equation (3).

- Figures 4–13: Responses of inequality, top income shares, wage share, housing and equity prices
  - Figure 4: Effect on income inequality (Net Gini), responses to a 100 basis point unanticipated increase; t=0 is year of shock; 90 percent confidence bands; estimates based on equation (3).
  - Figure 5: Robustness checks (Gross Inequality; Advanced economies; Pre-2008; controlling for recessions and fiscal stance).
  - Figure 6: Effect on top income shares — Panels for Top 10 percent, Top 5 percent, Top 1 percent; responses to 100 basis point shocks; 90 percent confidence bands.
  - Figure 7: Effect on share of wage income in GDP (percentage points); responses to 100 basis point shocks; 90 percent confidence bands.
  - Figure 8: Effect on inequality, positive vs. negative shocks; Panels A (Negative shocks) and B (Positive shocks); solid yellow lines denote unconditional (baseline) response from Figure 3.2; estimates based on equation (4).
  - Figure 9: Effect on inequality, exogenous vs. growth-driven shocks; Panels for growth-driven shocks, innovations in policy rates, changes in policy rates; estimates based on equation (5).
  - Figure 10: Role of business cycle — multiple panels showing responses during recessions vs. expansions, and for positive/negative shocks; estimates based on equation (6).
  - Figure 11: Role of labor earnings — Panels A (Very low labor share) and B (Very high labor share); comparisons to baseline response.
  - Figure 12: Role of redistribution — Panels A (Low redistribution) and B (Low initial equality); comparisons to baseline response.
  - Figure 13: Effects on equity and house price index — Panel A Housing prices; Panel B Equity prices; responses to 100 basis point shocks; 90 percent confidence bands.

### Key descriptive statistics (Table 1)
- Inequality
  - Net Gini: Mean 33.6; SD 7.2; Min 18.0; Max 54.1
  - Gross Gini: Mean 45.5; SD 5.2; Min 30.0; Max 57.3
- Top income shares
  - Top 1 percent: Mean 11.1; SD 4.3; Min 5.2; Max 23.5
  - Top 5 percent: Mean 24.6; SD 5.7; Min 14.6; Max 38.8
  - Top 10 percent: Mean 35.9; SD 6.6; Min 23.6; Max 50.6
- Share of wage income in GDP: Mean 51.5; SD 4.0; Min 39.2; Max 58.8
- Monetary policy shocks: Mean 0.0; SD 1.13; Min -3.9; Max 4

### Regression evidence on monetary policy shocks and Net Gini (Table 2)
- Dependent variable: income inequality (Net Gini), sample period 1990-2013.
- Specifications reported for K=0, K=1, K=2, K=3, K=4 (lags included).
- Selected coefficients (coefficient (t-statistic)):
  - Gini growth (t-1):
    - K=0: 0.176** (2.16)
    - K=1: 0.266** (2.53)
    - K=2: 0.080 (0.58)
    - K=3: -0.074 (-0.49)
    - K=4: -0.175 (-1.12)
  - Gini growth (t-1) [second row reported]:
    - K=0: 0.071 (1.17)
    - K=1: -0.139 (-1.58)
    - K=2: -0.274** (-2.41)
    - K=3: -0.346*** (-2.81)
    - K=4: -0.305** (-2.27)
  - Monetary policy shock (t):
    - K=0: 0.531 (1.56)
    - K=1: 1.176** (2.13)
    - K=2: 1.886*** (2.88)
    - K=3: 2.018*** (2.59)
    - K=4: 2.154** (2.40)
  - Monetary policy shock (t-1):
    - K=0: 0.441 (1.62)
    - K=1: 0.389 (1.08)
    - K=2: 0.074 (0.16)
    - K=3: 0.160 (0.30)
    - K=4: 0.221 (0.34)
  - Monetary policy shock (t-2):
    - K=0: -0.034 (-0.16)
    - K=1: -0.114 (-0.36)
    - K=2: 0.278 (0.74)
    - K=3: 0.364 (0.84)
    - K=4: 0.235 (0.49)
- Sample sizes and fit:
  - N: K=0 -> 467; K=1 -> 439; K=2 -> 409; K=3 -> 379; K=4 -> 349
  - R^2: K=0 -> 0.16; K=1 -> 0.18; K=2 -> 0.19; K=3 -> 0.24; K=4 -> 0.28
- Note: T-statistics based on robust clustered standard errors in parenthesis. ***,**,* denote significance at 1 percent, 5 percent and 10 percent, respectively. Estimates based on equation (3).

### Sample composition (Annex, Table A1)
- Advanced Economies included (selected): Australia; Canada; Czech Republic; France; Germany; Hong Kong SAR; Italy; Japan; Korea; Netherlands; New Zealand; Norway; Singapore; Slovak Republic; Spain; Sweden; Switzerland; Taiwan Province of China; United Kingdom; United States.
- Emerging Market Countries included (selected): Argentina; Brazil; Chile; Hungary; India; Indonesia; Malaysia; Mexico; Philippines; Poland; Thailand; Turkey.

*Source: _wp16245 - References (extracted figures, tables, and reference list from the supplied PDF).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2016/_wp16245.pdf_