Labor Responses, Regulation, and Business Churn

Date01 February 2021
Published date01 February 2021
DOIhttp://doi.org/10.1111/jmcb.12694
DOI: 10.1111/jmcb.12694
MARTA ALOI
HUW DIXON
ANTHONY SAVAGAR
Labor Responses, Regulation, and Business Churn
We develop a model of sluggish rm entry to explain short-run labor re-
sponses to technology shocks. We show that the labor response to technol-
ogy and its persistence depend on the degree of returns to labor and the rate
of rm entry. Existing empirical results support our theory based on short-
run labor responses across U.S. industries. Wederive closed-form transition
paths that show the result occurs because labor adjusts instantaneously while
rms are sluggish, and closed-form eigenvalues showthat stricter entry reg-
ulation results in slower convergence to steady state. Finally,we show that
our theoretical results hold in a quantitativemodel with capital accumulation
and interest rate dynamics.
JEL codes: D25, E20, L11, O33
Keywords: deregulation, dynamic entry, endogenous entry costs, short-run
labor responses
T -    hours to technology
shocks has been widely debated in macroeconomics. Empirical studies, such as
Chang and Hong (2006), show that labor responses to technology shocks differ across
U.S. manufacturing industries. Using four-digit manufacturing sector data, Chang
and Hong show that while some industries exhibit a temporary reduction in employ-
ment in response to a permanent increase in technology, many more industries ex-
hibit a short-run increase in both employment and hours per worker. However, the
theory underlying these responses is not fully understood. In this paper, we identify
a novel mechanism based on dynamic rm entry to explain short-run labor responses
and subsequent persistence. Cross-industry data supports our theory. In addition, we
show that persistence of labor responses depends on rm sluggishness, which regu-
lation affects through endogenous entry costs.
Thanks to Mathan Satchi, Christian Siegel, Mehdi Sahneh, Vincent Sterk, Petr Sedlacek, Kjetil
Storesletten, and Matteo Cacciatore for useful discussions.
M A is at University of Nottingham (E-mail: marta.aloi@nottingham.ac.uk). H D
is at Cardiff University (E-mail: dixonh@cardiff.ac.uk). A S is at University of Kent
(E-mail: a.savagar@kent.ac.uk).
Received September 18, 2018; and accepted in revised form August 27, 2019.
Journal of Money, Credit and Banking, Vol. 53, No. 1 (February 2021)
© 2020 The Ohio State University
120 :MONEY,CREDIT AND BANKING
Our mechanism focuses on endogenous variation in labor per rm, which occurs
when rm creation is sluggish but labor adjusts instantaneously. Endogenous varia-
tion in labor per rm is important for aggregate labor responses except in the special
case of a constant marginal product of labor (MPL) in the rm’s production function.
For example, if a positive technology shock increases hours, but the stock of rms
is xed, hours per rm increase. With short-run increasing MPL, the rise in hours
per rm increases MPL, increases wages, and increases hours. Subsequent rm entry
decreases hours per rm, decreases MPL, decreases wage, and decreases labor to its
long-run level.1This channel is typically overlooked because either labor per rm is
xed (due to instantaneous free entry) or the MPL is constant.2
Wedevelop a DGE small open economy (SOE) model in continuous time extended
to include dynamic rm entry. Output is produced with labor, and there is an inter-
nationally traded bond with world interest rates equal to the household discount rate.
Hence, the household perfectly smooths utility, so consumption dynamics do not play
a role, which allows a closed-form analysis of rm dynamics. Households can invest
in new rms by paying an endogenous entry cost. Once operational, rms compete
under monopolistic competition and pay a xed overhead cost each period. The re-
striction to one state variable (number of rms) keeps eigenvalues tractable, so we
can study persistence and short-run versus long-run effects analytically.
To model dynamic entry, we assume that the entry costs depend on the ow of
entry due to a congestion effect caused by red tape (Datta and Dixon 2002). As a
result, output per rm and operating prots vary in the short run, while in the long run
rms fully adjust so that there is a free-entry, zero-prot steady state. In steady state,
average rm size is independent of technology. The speed of convergence captured
in the stable eigenvalue depends on the ow of rm creation, which in turn depends
on the level of regulation in an economy. We characterize deregulation as cut in red
tape, which causes less congestion in the entry process decreasing the endogenous
sunk entry cost and speeding-up business churn. Our model is parsimonious in order
to derive general analytic results and provides testable implications consistent with
empirical literature.
Wealso consider a quantitative RBC model with capital and a variable interest rate,
keeping sluggish rm entry and allowing for variationin the slope of the marginal cost
curve. We nd very similar results to our simple SOE model, which shows that the
simplifying assumptions we make for an analytical solution are not necessary for our
results to hold in larger quantitative models.
Related literature. Recent literature in macroeconomics has focused on the impor-
tance of rm entry dynamics for business cycle uctuations Bilbiie, Ghironi, and
Melitz (2012) (BGM) developed a popular model of sluggish entry based on a xed
1. With decreasing MPL, the signs are reversed.
2. In principle, other mechanisms that cause variation in employment at the rm level could cause
similar effects. We focus on the sluggishness of rm entry, but equally slow aggregate labor adjustment
could also affect employment at the rm level—providing the adjustment of rms is not exactly propor-
tional to the adjustment of labor, such that labor per rm does not vary, which is the case in zero-prot,
free-entry models.
MARTAALOI, HUW DIXON, AND ANTHONY SAVAGAR :121
250
200
150
100
50
0
0 5 10 15 20
Procedures to Start a Business
Days to Start a Business
Fig 1. Red Tapeand Business Churn.
entry cost and a time-to-build lag in discrete time. We extend the idea of sluggish
entry adjustment to a continuous-time model of endogenous entry costs. This has the
benet of allowing for a tractable analysis and offers a new angle to study deregu-
lation. The endogenous entry cost creates an intertemporal zero-prot condition that
equates the cost of entry in each instant to the net present value of incumbency. This
causes the number of rms to gradually adjust to its long-run value. Entry costs are en-
dogenous because they depend on the number of entering rms. Lewis (2009) shows
that these so-called entry congestion effects are important for macro-economic prop-
agation in empirical work, and recent theoretical papers also include this mechanism
(Bergin and Lin 2012).
Cantore, Ferroni, and Leon-Ledesma (2017, Figure 1, p. 70) provide empirical ev-
idence that, at the aggregate level, short-run responses (SRRs) of labor to technology
have reversedover the past century in the United States from decreasing to increasing,
and that the deviation now persists for longer.We show that increased persistence can
occur because of slower business churn caused by higher entry regulation. Our anal-
ysis contributes a novel angle to existing studies of entry regulation. Most literature
focuses on the effect of decreasing xed entry costs. This determines the stock of rms
operating in the long run which has implications for static allocations (Barseghyan
and DiCecio 2011). However, we analyze deregulation of endogenous entry costs
that affect the speed at which rms transition to arbitrage prots, and therefore de-
termine the persistence of aggregate variables. Cacciatore and Fiori (2016) explore
deregulation in goods and labor markets. They nd that reforms are benecial in the
long run, but can have short-run recessionary effects. Similarly to our work, they in-
clude sluggish rm adjustment, but they also have sluggish labor adjustment due to
search frictions.

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT