VIII МЕЖДУНАРОДНАЯ КОНФЕРЕНЦИЯ ПО ПРОБЛЕМАМ ЭКОНОМИЧЕСКОГО РАЗВИТИЯ В СОВРЕМЕННОМ МИРЕ 22 АПРЕЛЯ 2011 ГОДА, Г.ЕКАТЕРИНБУРГ КЛАРА САБИРЬЯНОВА ПИТЕР, УНИВЕРСИТЕТ СЕВЕРНОЙ КАРОЛИНЫ НА ОСНОВЕ СТАТЬИ TAX PROGRESSIVITY AND INCOME INEQUALITY (JOINTLY WITH DENVIL DUNCAN, INDIANA UNIVERSITY) ПРОГРЕССИВНОСТЬ НАЛОГОВОЙ СИСТЕМЫ И НЕРАВЕНСТВО ДОХОДОВ

Глобальное снижение максимальной ставки подоходного налога, Source: Sabirianova Peter, Buttrick, and Duncan (2010) No. of country-year observations=3613; no. of countries = 189

Глобальное снижение структурной прогрессивности подоходного налога Source: Sabirianova Peter, Buttrick, and Duncan (2010); Notes: Rates are GDP-weighted; y denotes the level of gross income equivalent to a countrys GDP per capita No. of country-year observations = 3082 ; no. of countries =175

Мотивация Evidence points to flattening of personal income tax (PIT) schedules or declining structural PIT progressivity throughout the world Structural progressivity is defined as a slope of average tax rates along the income distribution (as opposed to effective progressivity that captures the tax-induced changes in income distribution); Musgrave and Thin (1948) Equity cost of flattening PIT schedules implies the trade-off between equity and efficiency Theory: Ramsey (1927), Mirrlees (1971) Empirical: Paulus and Peichl (2008), Fuest, Peichl, and Schaefer (2008) But, does a flatter PIT schedule necessarily imply increased income inequality?

Мотивация Gross Income Personal Income Tax Function Observed Net Income Tax Liability Reported Gross Income

Мотивация Gross Income Personal Income Tax Function Evasion effect Productivity effect Redistribution Gross Income Actual Net Income Observed Net Income Tax Liability Reported Gross Income Hidden Income

Цели исследования What is the effect of declining structural progressivity on income inequality throughout the world? Is there a differential effect of PIT flattening on observed versus true income inequality? What factors contribute to the differential effect?

Теоретические гипотезы Неравенство в наблюдаемом/декларируемом доходе (Kuznets ratio)

Теоретические гипотезы Неравенство в ненаблюдаемом/истинном доходе (Kuznets ratio)

Теоретические гипотезы Testable hypotheses H1: Negative relationship between tax progressivity and observed income inequality. H2: Negative effect is reinforced by pro-poor redistributive environments. H3: Structural progressivity has a smaller effect on actual than on observed net income inequality. H4: A positive effect on actual inequality is more likely to be found in countries with weaker legal institutions.

Данные по прогрессивности Unbalanced panel Countries: 140+ Years: 25 (1981 – 2005) Progressivity measures Top PIT rate Time-varying estimates of structural progressivity Linear slope of marginal and average tax rates along income scale Two different income ranges around countrys GDP per capita Source: Peter, Buttrick, and Duncan (2009)

Данные по неравенству в доходах Inequality measure GINI - World Institute for Development Research (WIDER v2b) Gross income, net income and consumption based measures. Dummies to control for differences across GINI measures Income adjustment (e.g., use of equivalence scale, per capita adjustment, no adjustment) Coverage (national vs. other)

Данные по неравенству в доходах

Эмпирический анализ неравенства в наблюдаемом/декларириуемом доходе Baseline model Gross and net income inequality only, I it P it = progressivity measures, β is expected negative Z it = quadratic log of GDP per capita, inflation, services in GDP (%), and industry in GDP (%) W it = dummies to control for differences across GINI measures Endogeneity of progressivity is addressed by IV-2sls IVs are population-distance weighted averages of neighboring PIT measures

Эмпирический анализ неравенства в наблюдаемом/декларириуемом доходе Base Specification for Inequality in Observed Income 1.Inequality in observed (reported) income is negatively associated with top personal income tax (PIT) rate:

Эмпирический анализ неравенства в наблюдаемом/декларириуемом доходе Base Specification for Inequality in Observed Income 2. Personal income tax progressivity reduces observed income inequality:

Роль демократических институтов Results with the index of civil liberties and the democracy score are similar. Structural Progressivity, Gini in Observed Income, and the Role of Democratic Institutions 3. The negative effect of structural progressivity on observed income inequality is strongest in countries with developed democratic institutions.

Эмпирический анализ неравенства в потреблении Consistent underreporting of income relative to expenditures in developing countries (Deaton 1997 and 2005, Gorodnichenko et al 2009 Observed income vs. consumption P it = progressivity measures D g = 1 if gross income GINI D n = 1 if net income GINI Consumption GINI is the omitted category Parameters of interest are: β?, λ 1

Эмпирический анализ неравенства в потреблении Differential effect of structural PIT progressivity on true vs. observed income inequality. Significantly smaller effect of structural PIT progressivity on inequality in consumption than its effect on inequality in observed net disposable income. The differential effect is largest in environments conducive to high levels of evasion.

Эмпирический анализ неравенства в потреблении Base Specification for Inequality in Consumption 4. The effect of structural progressivity on consumption inequality could be positive in countries with weak law and order.

Заключение Differential effect of structural progressivity on true vs. observed income inequality The differential effect is largest in environments conducive to high levels of evasion. Policy implications: Equity cost of efficiency are likely smaller than we think. Countries with very large informal sectors may not be restricted by the equity-efficiency trade-off. Flatter PIT schedule may be optimal in countries with large shadow economies. Redistribution policy should target gross income rather than the progressivity of the tax schedule.

Спасибо.

The Caveat Actual income inequality is not observed Use consumption data Consistent underreporting of income relative to expenditures in developing countries (Deaton 1997 and 2005, Gorodnichenko et al 2009) Consumption vs. actual net income inequality Consumption distributed differently than income Savings, private transfers, and other smoothing mechanisms Meyer and Sullivan (2009) and Blundell, Pistaferri, and Preston (2008) Our estimates will be valid under the assumption of no response in the distribution of saving and private transfers to changes in tax progressivity.

Marginal Rate Progression: Example MR = statutory marginal rates adjusted for standard deductions, basic allowance, local taxes, and surcharges MRP1 = marginal rate progression for the level of income up to 4 times GDP per capita MRP2 = marginal rate progression for the level of income up to 2 times GDP per capita

Data

Эмпирический анализ неравенства в потреблении