Marginal Tax Rates and Income: New Time Series Evidence

Transcription

1 COMMENTS ON: Marginal Tax Rates and Income: New Time Series Evidence by Karel Mertens and José Luis Montiel Olea Forthcoming, QJE EBEN LAZARUS (with thanks to Pepe Montiel Olea for figures) Harvard Macro-PF Reading Group April, 8

2 Background Previous paper (Mertens & Ravn, ): Effects of changes in average tax rates In theory, identifies income + AD effect Substitution effect? This paper: Short-run elasticities of income w.r.t. exogenous marginal rate changes Find large effects for entirety of income distribution

3 Previous Evidence Nice discussion of previous literature on elasticity of taxable income (ETI):. PF: Typically quite small (..) Data: Personal tax returns DiD or instrument using statutory changes. Macro: Much larger (.) Data: Macro aggregates Focus on policy reforms Try to reconcile these two Use personal income data as in Saez (), extended to 96 sample Romer and Romer () shocks for identification Not all statutory changes are exogenous Reason to believe that procyclicality in statutory changes induces downward bias in PF-literature estimates Timing also an issue

4 Romer and Romer () Romer and Romer shocks: Four categories of legislated tax changes:. Responding to change in government spending. Responding to some other macro factor likely to affect output. Dealing with budget deficit inherited from past. Trying to achieve some ideological goal (higher growth, increased fairness, smaller gov t) () and () are endogenous (and excluded); () and () are exogenous Debatable is -pg. appendix detailing each tax change and records supporting categorization Mertens and Montiel Olea then use these as instruments in SVAR (and its cousin, local projection)

5 Main Results.. LPIV SVAR /(-AMTR) All Tax Units. LPIV SVAR Real GDP LPIV SVAR Income (All Tax Units) age points LPIV SVAR Unemployment Rate Aggregate Responses to a One Percent Increase in the Marginal Net-of-Tax Rate.

6 Splits. SVAR-IV. LP-IV Top % Top % Top % Top % to % Top % to % Bottom 99% Bottom 9% Top % Top % Top % Top % to % Top % to % Bottom 99% Bottom 9%.... Dynamic Estimates of Tax Elasticities of Income, by Income Group 7 SVAR-IV 7 LP-IV 6 Income excl. CG Income incl. CG Wage Income Nonwage Income excl. CG Nonwage Income incl. CG 6 Income excl. CG Income incl. CG Wage Income Nonwage Income excl. CG Nonwage Income incl. CG Dynamic Estimates of Tax Elasticities of Income, by Income Source

7 Average vs. Marginal A. Cut in Marginal Tax Rate Allowing Impact on Statutory Average Tax Rate /(-AMTR) All Tax Units Income (All Tax Units) - Average Tax Rate - - B. Cut in Marginal Tax Rate Without Impact on Statutory Average Tax Rate /(-AMTR) All Tax Units Income (All Tax Units) - Average Tax Rate - - C. Cut in Average Tax Rate Without Impact on Statutory Marginal Tax Rate /(-AMTR) All Tax Units Income (All Tax Units) -. Average Tax Rate

8 Spillover Distributional Effects (I) /(-AMTR) Top % Tax Units /(-AMTR) Btm 99% Tax Units Income Top % Tax Units Income Btm 99% Tax Units Real GDP.6 Unemployment Rate age points Responses to a % Increase in the Top % Marginal Net-of-Tax Rate 7

9 Spillover Distributional Effects (II) /(-AMTR) Top % Tax Units /(-AMTR) Btm 99% Tax Units Income Top % Tax Units Income Btm 99% Tax Units Real GDP.6 Unemployment Rate age points Responses to a % Increase in the Bottom 99% Marginal Net-of-Tax Rate 8

10 Can We Recover Smaller Results via Misspecification? 9