Behavioral Effects of Withholding Taxes on Labor Supply*

• Published date: 01 October 2019
• DOI: http://doi.org/10.1111/sjoe.12345
• Date: 01 October 2019
• Author: Melanie Steinhoff,Johannes Becker,Jonas Fooken

Scand. J. of Economics 121(4), 1417–1440, 2019

DOI: 10.1111/sjoe.12345

Behavioral Effects of Withholding Taxes on

Labor Supply*

Johannes Becker

University of M¨unster, DE-48143 M¨unster, Germany

johannes.becker@wiwi.uni-muenster.de

Jonas Fooken

University of Queensland, St Lucia QLD, Australia

jonas.fooken@gmail.com

Melanie Steinhoff

KPMG AGWirtschaftspr ¨ufungsgesellschaft,DE-20459 Hamb urg,Ger many

msteinhoff@kpmg.com

Abstract

Inmost advanced economies, income tax collection uses third-party withholding for wage income.

Because withholding taxes do not necessarily reflect true effective taxes, these may give false

signals of net-of-tax pay.We test labor supplyresponses to such misconceptions using laborator y

experiments.Withholding taxes (and the resulting tax refunds) should be behaviorally neutral, but

our results show that tax adjustments lead to effort adjustments, which suggests that withholding

blurs tax incentives. While there is no statisticallysignificant response in the overall sample, the

participants in the subgroup who self-assess to be motivated by monetary incentives (about half

of the total sample) reduce their effort in response to withholding taxes and increase their effort

after receiving tax refunds.

Keywords: Experiment; tax perceptions; withholding taxes

JEL classification:G32; H25; M41

I. Introduction

Are you expecting a tax refund this season? If so, you’re

probably dreaming about all the possibilities of spending what

feels like newly “found” money. However, it’s important to

*We thank Benedikt Herrmann, Lars Siemers, Kay Blaufus, Denvil Duncan, David Rowell,

conference participants at ACE Brisbane, ESA Sydney, ZEW Mannheim and IMEBESS

Rome, and seminar participants in Hamburg, Mannheim, Middlesex, and Siegen for useful

comments.

C

The editors of The Scandinavian Journal of Economics 2019.

1418 Behavioral effects of withholding taxes on labor supply

remember that this is your money, that you earned – not a bonus.

A tax refund is issued when you overpaid the government.1

Most wage earners receive an income tax refund at the end of the tax

year. As the above quote suggests, this money is often treated as “newly

found”, as a surprising one-time increase in income, and it triggers an

increase in consumption spending (Shapiro and Slemrod, 1995). Because the

refund can be anticipated, this finding is hard to reconcile with fully rational

behavior (in the absence of liquidity constraints). If individuals change their

behavior when the money is returned, it might well be that behavior also

changes when the money is taken away, and hence when withholding taxes

reduce income throughout the year. In particular, withholding taxes and the

resulting tax adjustments at the end of the year can affect labor supply by

changing the perceived immediate return from working (i.e., the perceived

net wage).

In this paper, we investigate labor supply responses to withholding rates

and tax adjustments using a laboratory experiment, which allows us to

make an experimental policy change that cannot easily be done in reality.

Participants perform a real effort task and are paid a piece rate for each

completed task. Their income is taxed. Our treatments use different levels

of withholding rates. In the absence of withholding, wage earners receive

“monthly” gross wage payments; at the end of the “year”, they pay their

tax liability as one large installment. In the case of withholding, taxes

are directly subtracted from the gross wage each “month”; at the end

of the “year”, a tax adjustment takes place, which implies an additional

payment or a tax refund, depending on the withholding rate and on effort.

Thus, withholding shifts tax payments forward in time and increases their

frequency. We use our experiment to test if withholding rates and tax

adjustments influence labor supply.

The experiment is designed such that withholding rates and tax

adjustments should be irrelevant for behavior. However, our results show

that tax adjustments (refunds or additional payments) change effort

provision in the subsequent period. After observing their true tax liability

and end-of-year income adjustments, participants strongly adjust their labor

supply. We interpret this as an update of beliefs on the actual tax burden.

Furthermore, on closer examination, we find that these findings are driven

by participants who describe themselves as money-motivated (about half

of our sample). Within this group, higher withholding rates significantly

reduce effort, while refunds increase it. Because these two effects are in

1Taken from the article “Here’s How You Should Spend Your Tax Refund” by Jon Lal,

published in US News on 7 April 2016 (available online at http://money.usnews.com/money/

blogs/my-money/articles/2016-04-07/heres- how-you-should-spend-your-tax- refund).

C

The editors of The Scandinavian Journal of Economics 2019.