The Effect of Tax Expenditures on Automatic Stabilizers: Methods and Evidence

• Published date: 01 September 2017
• Author: Hautahi Kingi,Kyle Rozema
• Date: 01 September 2017
• DOI: http://doi.org/10.1111/jels.12155

The Effect of Tax Expenditures on

Automatic Stabilizers: Methods

and Evidence

Hautahi Kingi and Kyle Rozema*

We study theeffect of tax expenditureson the stabilizing powerof the tax system. We propose

a micro-simulation strategy that exploits links that we identify between automatic stabilizers,

tax expenditures, and effectivemarginal tax rates. UsingU.S. tax return micro data from 2000

to 2010, we estimate that, on average, the mortgage interest deduction and the charitable

contributions deduction decreased the ability of the tax system to absorb fluctuations in

aggregateconsumption by an average of 7.4percent and 3.9 percent, respectively.

I. Introduction

The federal government “spends” through the tax code by exempting certain economic activi-

ties from taxation. Research has examined how these “tax expenditures” affect individual

behavior (e.g., Hilber & Turner 2014) and how the benefits of such tax expenditures are

unequally distributed across income groups (e.g., Burman et al. 2008; Cole et al. 2011; Hemel

& Rozema 2017). We broaden these insights by examining tax expenditures within the wider

contextofthefederaltaxsystem.Inparticular,we investigate how tax expenditures affect the

ability of the tax system to stabilize household disposable income and consumption.

The income tax system reduces fluctuations in disposable income by partially

absorbing shocks to market income. Tax expenditures may distort this ability of the tax

system to function as an automatic stabilizer (Listokin 2012). In this article, we study

these potential distortions empirically. We propose a method to estimate the effect of a

tax expenditure on the ability of the tax system to act as an automatic stabilizer. In

developing this method, we exploit underlying theoretical links that we identify between

measures of the automatic stabilizing power of a tax system, the size of tax expenditures,

and effective marginal tax rates.

*Address correspondence to Kyle Rozema, 375 E. Chicago Ave., Chicago, IL 60611; e-mail: kyle.rozema@law.

northwestern.edu. Kingi is Research Associate, IMPAQ International; Rozema is postddoctoral fellow,

Northwestern University Pritzker School of Law.

We thank John Abowd, Aaron Bodoh-Creed, Josh Chafetz, Dawn Chutkow, Charlotte Crane, Michael Frakes,

Brian Galle, Don Kenkel, Jim Lindgren, Yair Listokin, Karel Mertens, Adam Rosenweig, David Schwartz, Nicolas

Ziebarth, two anonymous referees, and participants of the 2015 Conference of Empirical Legal Studies, North-

western University Legal Scholarship Workshop, and the 2015 Midwestern Law and Economics Conference for

helpful comments. All remaining errors are our own.

548

Journal of Empirical Legal Studies

Volume 14, Issue 3, 548–568, September 2017

We measure the automatic stabilization of disposable income using the normal-

ized tax change (NTC) (Auerbach & Feenberg 2000), which captures how much aggre-

gate tax revenue changes in response to a change in aggregate market income (Slitor

1948). To measure how a tax expenditure affects the automatic stabilization of dispos-

able income, we estimate the change in the NTC induced by the removal of the provi-

sion. This change measures the destabilizing effect of a tax expenditure on disposable

income, which we show can be interpreted as (1) the extra proportion of a fluctuation

in market income that would be absorbed by the tax system in the absence of the tax

expenditure, (2) the sensitivity of the tax expenditure to income changes, or (3) the

sensitivity of the effective marginal tax rate to the removal of the tax expenditure.

Automatic stabilizers deal with business cycle fluctuations, which are inherently transi-

tory. Because consumption of rational agents depends on permanent—not transitory—

income, the effect of tax expenditures on disposable income stabilization must be translated

into demand stabilization by adjusting for each household’s marginal propensity to con-

sume. To estimate demand stabilization, we adopt the standard assumption that liquidity-

constrained households have marginal propensitytoconsumeequaltoone,whileallother

households have zero marginal propensity to consume (Auerbach & Feenberg 2000; Dolls

et al. 2012). The adjusted NTC (ANTC) measures how much aggregate tax revenue that

wouldhaveotherwisebeenspentchanges i n response to a change in aggregate market income.

We define our empirical measure of the effect of a tax expenditure on automatic

stabilization as the change in the ANTC induced by the removal of the provision. This

measure, which we call the ANTC shifter, estimates the extra amount of consumption,

as a proportion of a fluctuation in market income, that the tax system would have

absorbed in the absence of the tax provision.

Using U.S. tax return micro data from the Statistics of Income Division of the

Internal Revenue Service from 2000 to 2010, we find evidence that two of the largest tax

expenditures—the mortgage interest deduction and the charitable contributions deduc-

tion—alter the federal tax system’s role as an automatic stabilizer. The mortgage interest

deduction and the charitable contributions deduction decreased the ability of the tax

system to absorb fluctuations in aggregate consumption by an average of 7.4 percent

and 3.9 percent, respectively.

A back-of-the-envelope calculation suggests that in a 3 percent recession, the repeal

of the mortgage interest deduction and the charitable contributions deduction would sta-

bilize annual consumption by an average of $2.4 billion and $1.2 billion (in 2010 dollars).

To contextualize the magnitude of these estimates in terms of the size of discretionary sta-

bilization policies, consider the 2001 tax rebate, which was estimated to have increased

aggregate consumption by $30 billion (in 2010 dollars) (Johnson et al. 2006). Our back-

of-the-envelope calculation suggests that the removal of the mortgage interest deduction

and charitable contributions deduction would have stabilized consumption by about 8 per-

cent and 4 percent of the change in consumption induced by the 2001 tax rebate.

The attractiveness of a tax expenditure, both politically and as a means to increase

social welfare, is rarely analyzed in light of its relationship with the general stabilizing

effect of the federal tax system (Saez 2004a; Listokin 2012). It is our view that the assess-

ment of the desirability of a tax expenditure ought to take this relationship into

549

The Effect of Tax Expenditures on Automatic Stabilizers