GOVERNMENT SPENDING, ENTRY, AND THE CONSUMPTION CROWDING‐IN PUZZLE

• DOI: http://doi.org/10.1111/iere.12241
• Published date: 01 August 2017
• Author: Roland Winkler,Vivien Lewis
• Date: 01 August 2017

INTERNATIONAL ECONOMIC REVIEW

Vol. 58, No. 3, August 2017

GOVERNMENT SPENDING, ENTRY, AND THE CONSUMPTION

CROWDING-IN PUZZLE∗

BYVIVIEN LEWIS AND ROLAND WINKLER1

KU Leuven, Belgium, and Deutsche Bundesbank, Germany; TU Dortmund University,

Germany

This article documents empirically that net firm entry robustly rises after a U.S. government spending expan-

sion. We use this new finding to test the empirical validity of various model features that have been proposed to

generate consumption crowding-in after positive expenditure shocks. Endogenous-entry models typically fail to

generate the observed joint increase in consumption and entry. Model features that dampen the wealth effect,

such as rule-of-thumb households or complementarity between labor and consumption in preferences, tend to

reduce entry. We show that utility- or productivity-enhancing public spending can reconcile the model with our

documented fact and performs well empirically.

1. INTRODUCTION

We document that a government spending expansion in the United States stimulates firm

entry and private consumption. We use the joint dynamics observed in the data to test the

empirical validity of various model features that address the so-called consumption crowding-in

puzzle. That puzzle describes the inability of standard business cycle models to account for the

crowding-in of private consumption after a positive government spending shock, which has been

observed in numerous empirical studies. What we find is that a combination of countercyclical

markups and a reduced wealth effect, which has been the literature’s answer to the puzzle,

cannot generate a rise in both consumption and entry. Instead, we propose useful government

spending as a way to reconcile our empirical finding with dynamic stochastic general equilibrium

models.

The effect of government spending expansions on private economic activity is much debated

among economists and policy makers alike, especially since the onset of the Great Recession

in 2007. This article sheds light on a particular form of economic activity, the entry into the

market of new firms (and products), which represents investment along the extensive margin.

Understanding the determinants of entry is important from a policy maker’s perspective, as

firm turnover is associated with a substantial amount of job creation and destruction. Davis

and Haltiwanger (1990) attribute 25% of U.S. annual job destruction to firm exit and 20%

of annual job creation to entry, whereas Spletzer (1998) reports over one-third for these two

measures. For a government wishing to fight unemployment through a fiscal stimulus package,

∗Manuscript received January 2015; revised February 2016.

1We thank our editor, Harold Cole, two anonymous referees, Christian Bredemeier, Andrea Colciago, Pedro

Gomes, Punnoose Jacob, Falko J¨

ussen, Mathias Klein, Christopher Krause, Ludger Linnemann, Stefania Villa, Jo van

Biesebroeck, and Nils Wittmann as well as seminar participants at Antwerp University, the Bank of England, Deutsche

Bundesbank, Hamburg University, Heidelberg University, the Kiel Institute for the World Economy, the 2013 CEF

Conference, the 2013 Annual Conference of the German Economic Association, the 2014 EEA Annual Congress,

and the 2015 ASSA Annual Meeting for helpful comments. The views expressed herein do not reflect those of the

Bundesbank or the Eurosystem. All remaining errors are ours.

Please address correspondence to: Vivien Lewis, KU Leuven, Belgium, Department of Economics, Naamsestraat

69, 3000 Leuven, Belgium. E-mail: vivien.lewis@kuleuven.be.

943

C

(2017) by the Economics Department of the University of Pennsylvania and the Osaka University Institute of Social

and Economic Research Association

944 LEWIS AND WINKLER

TABLE 1

EFFECTS OF U.S.GOVERNMENT SPENDING EXPANSIONS:EMPIRICAL RESULTS

Consumption Investment

Study Response Response

Alesina et al. (2002) n.a. negative

Blanchard and Perotti (2002) positive negative

Ben Zeev and Pappa (2017) positive positive

Burnside et al. (2004) insignificant positive

Caldara and Kamps (2008) positive method-dependent

Edelberg et al. (1999) negative positive

Fat´

as and Mihov (2001) positive positive

Fisher and Peters (2010) positive n.a.

Gal´

ı et al. (2007) positive insignificant

Monacelli and Perotti (2008) positive negative

Mountford and Uhlig (2009) positive negative

Ramey (2011) negative negative

Ravn et al. (2012) positive n.a.

it is therefore useful to understand, in a first step, the effects of government spending on firm

creation.2

Although empirical studies using vector autoregressions (VARs) overwhelmingly suggest

that private consumption responds positively to an expansionary government spending shock,

several find that capital investment falls; see Table 1.

Evidence on the effect of fiscal expansions on extensive-margin investment, that is, firm and

product entry, is so far missing in the literature. In the first part of the article, we estimate

a VAR model on U.S. data and identify government spending shocks by applying different

identification methods. As our benchmark, we follow Blanchard and Perotti (2002) and apply

a recursive identification scheme with government spending ordered first. This implies the

identifying assumption that government spending reacts only to its own shocks within the

quarter. Our empirical results indicate that government spending expansions lead to higher

private consumption and a rise in net business formation. We show that this finding is robust

to controlling for monetary and tax policies, to adding capital investment, to controlling for

anticipated changes in government spending, and to an alternative identification method using

sign restrictions. Additional evidence using a panel data estimation on annual U.S. state-level

data also shows a significant increase in consumption and firm entry in response to military

purchases.

Neither the canonical real business cycle model nor the standard New Keynesian model is

able to generate a crowding-in of private consumption. The reason is the strength of the wealth

effect, which is the prediction that the expectation of higher lump-sum taxes induces people to

consume less and to work more. A substantial research effort has tried to extend the models in

various ways so as to overcome this inconsistency between the model and the data.

After providing the evidence, we thus turn to the second part of the article where we inves-

tigate which model features are needed to generate a crowding-in of consumption and entry,

as observed in the data. To this end, we specify a benchmark business cycle model with en-

dogenous entry. The entry mechanism builds on Bilbiie et al. (2012), where entry costs are

sunk costs incurred once at the beginning of a firm’s existence, and the number of producers is

a state variable. We allow for countercyclical markups through changes in competition.3Hall

(2009) demonstrates that countercyclical markups have the potential to generate a consumption

2Colciago and Rossi (2012) show that job creation due to firm entry amplifies the response of labor market variables

to technology shocks in a model with endogenous entry and unemployment.

3There is compelling evidence for a negative relationship between entry and markups. Domowitz et al. (1988)

report a positive effect of concentration on markups. Bresnahan and Reiss (1991) and, more recently, Campbell and

Hopenhayn (2005) find that markups depend negatively on the number of competitors in an industry.