Do Businesses Benefit From Labor Hoarding? — Small Business Economics (2025)
Summary
Small Business Economics (2025) 66:1371-1393, Do businesses benefit from labor hoarding?, is by Nadia Greenhalgh-Stanley, Shawn M. Rohlin and Amanda Weinstein, received 10 February 2025, accepted 28 August 2025 and published online 7 November 2025. The abstract says the paper examines whether firms that retained more employees than predicted during the Great Recession, termed labor hoarding, experienced benefits over 7 years, using establishment-level data from the Infogroup Historical Business Database. It finds that labor hoarding firms were significantly more likely to survive but grew more slowly in both employment and sales than non-hoarding counterparts. The plain English summary says policies supporting workforce retention, such as the Paycheck Protection Program, may improve business survival.
Summary drafted by a model from the document's text below and checked by script against that text before publication. It is a navigation aid, not a reading of what the document proves. Where AI is used
- Issuer
- Source documents
- Document type
- Do Businesses Benefit From Labor Hoarding
- Case
- Do Businesses Benefit From Labor Hoarding
Cite this document
“Do Businesses Benefit From Labor Hoarding? — Small Business Economics (2025).”
Original at doi.org. Archived at Pandemic Darlings, https://pandemicdarlings.com/source-documents/do-businesses-benefit-from-labor-hoarding/
, SHA-256 41e16069e172df00d3630d47deabc63980bd5fb8f291a9c6e38d9224478c2e65.Full text
Small Bus Econ (2025) 66:1371–1393
https://doi.org/10.1007/s11187-025-01116-0
RESEARCH ARTICLE
Do businesses benefit from labor hoarding?
Nadia Greenhalgh‑Stanley · Shawn M. Rohlin ·
Amanda Weinstein
Received: 10 February 2025 / Accepted: 28 August 2025 / Published online: 7 November 2025
© The Author(s) 2025
Abstract This paper examines whether firms that Plain English Summary We examine whether firms
retained more employees than predicted during the realize any benefits of “labor hoarding” during a reces-
Great Recession, often termed “labor hoarding,” sion by looking at their performance measure in the
experienced benefits over 7 years. Using establish- period after the recession. Businesses may face higher
ment-level data from the Infogroup Historical Busi- labor costs during a period of low demand when they
ness Database, we estimate labor hoarding and assess choose to hoard labor during a recession, which could
its impact on firm survival, employment, and sales have long-term negative effects on their business dur-
in the recovery period. We find that labor hoarding ing the recovery. On the other hand, businesses may
firms were significantly more likely to survive, with reduce long-term turnover, building loyalty and skills
the effect strongest for women-owned and women-led in their workforce, acting as a kind of “workforce pres-
businesses. However, these firms grew more slowly in ervation.” We find that labor hoarding amongst firms
both employment and sales than non-hoarding coun- led to increased survival rates of firms after recessions;
terparts, who needed to rehire post-recession. Our however, those firms also experienced slower growth
results suggest that retaining workers acts less as an rates in employment and sales after the recession. We
added cost and more as a strategy of “workforce pres- also examine these effects across gender and find that
ervation,” providing stability at the expense of rapid women-led labor hoarding firms survived at higher
post-recession growth. These findings highlight that rates than men-led labor hoarding firms. Our results
policies supporting workforce retention, such as the suggest that “workforce preservation” may be a more
Paycheck Protection Program, may improve busi- accurate label than “labor hoarding” for firms trying to
ness survival and economic resilience, especially for improve their chances of surviving a recession.
women-led firms.
Keywords Labor hoarding · Recession · Business
survival · Workforce preservation
N. Greenhalgh‑Stanley · S. M. Rohlin (*)
Department of Economics, Kent State University, Kent,
OH 44242, USA JEL Classification E24 · E32 · J65 · L21
e-mail: srohlin@kent.edu
N. Greenhalgh‑Stanley
e-mail: ngreenha@kent.edu 1 Introduction
A. Weinstein
The Paycheck Protection Program (PPP) established
Center On Rural Innovation, Hudson, OH, USA
e-mail: Amanda.Weinstein@ruralinnovation.us by the CARES Act in 2020 was implemented by the
Vol.: (0123456789)
1372 N. Greenhalgh‑Stanley et al.
Small Business Administration to provide small busi- predicts subsequent performance. Without some sort
nesses with loans to help them cover payroll dur- of natural experiment that randomly causes some
ing the COVID-19 pandemic-induced recession. firms to hoard labor, researchers are forced to try to
The goal of the PPP was to save jobs by subsidizing minimize these concerns through econometric and
labor hoarding. Estimates vary greatly on the number methodological choices. Our approach is to match
of jobs saved by the program, ranging from 2 mil- firms that hoard labor with firms that do not, using
lion (Autor et al., 2022) to as many as 18.6 million propensity score matching. We require a match to be
jobs (Faulkender et al., 2020). There are benefits to within the same county and operating in the same
workers in avoiding unemployment and avoiding the industry, with the hope that it minimizes geographic
long-run scarring effects of unemployment (see for variability, differences in local economies, and
example, Gonalons-Pons & Gangl, 2022; Oreopou- growth changes in industries. We hope this minimizes
los et al., 2012; Kahn, 2010; Sullivan & von Wachter, differences in anticipating future demand, expecta-
2009), but less is known about whether the firm may tions in recovery, and similar rehiring or training
benefit from labor hoarding or if it is merely an added costs. Additionally, we match on employment and
cost, specifically coming out of a recession. sales from 2004 through 2007, meaning these busi-
Similar to the concept of “efficiency wages,” nesses were growing in both employment and sales at
whereby firms pay higher than equilibrium wages the same rate before the recession and were similar
to attract and keep a skilled workforce and increase in size. We further attempt to control for additional
productivity or efficiency, labor hoarding may act firm-specific attributes, like productivity and market-
as a type of “efficiency labor” or “workforce pres- ing expenditures. It is important to note that we can
ervation.” By keeping employees on the payroll never fully address omitted variable bias because we
especially during a recession, firms may build firm- have no information on managerial quality, customer
specific human capital and loyalty in employees relationships, or informal credit access.
that similarly boosts productivity or efficiency. It is Firms that retain employees during downturns may
too early to test whether there may be longer term benefit from reduced turnover costs, stronger employee
benefits or costs to businesses of this type of labor loyalty, and preserved firm-specific human capital. By
hoarding that the PPP subsidized. Instead, to test maintaining a stable workforce, firms avoid the costs
this idea, we examine the effect of labor hoarding and risks associated with rehiring and retraining once
in the Great Recession on firm performance after demand recovers. This continuity can also preserve
the Great Recession using the Infogroup Historical organizational knowledge and customer relationships,
Database. We estimate labor hoarding by comparing which may be critical to survival. In this sense, labor
2010 actual employment compared to predicted 2010 hoarding acts less as an added cost and more as an
employment based on existing firms’ employment insurance policy, which helps firms weather downturns
and sales trajectories before the recession from 2004 and positions them for smoother recovery.
and 2007. Previous research finds that women entrepre-
A challenge to studying labor hoarding is that it is neurs were more likely to hoard labor during the
endogenous to firm survival, employment, and sales Great Recession (Matsa & Miller, 2014) and that
growth because the decision to retain excess workers the share of women-owned and managed estab-
is jointly determined. Firms could hoard labor for sev- lishments in counties increased economic stability
eral possibly endogenous reasons including when they in the Great Recession (Deller, Conroy, and Wat-
anticipate stronger future demand; expect a stronger son, 2017). The status of women, beyond mere
recovery; possess valuable firm-specific human cap- ownership, may impact business outcomes such as
ital; or face higher rehiring or training costs. These employment, innovation, and profit. Therefore, we
same expectations and assets also raise the odds examine both women-owned and women-led firms,
of surviving and growing. Additionally, access to where the latter more directly measures women’s
finance further links labor hoarding and future busi- influence over managerial and leadership decisions.
ness outcomes. Firms with better capitalization or We find that women-led firms were more likely to
better access to banking can absorb short-run losses hoard labor during the Great Recession. We further
to preserve workforce capacity, and liquidity itself examine whether the advantages of labor hoarding
Vol:. (1234567890)
Do businesses benefit from labor hoarding? 1373
differed for women-owned/led firms compared to economy (Glaeser et al., 1992; Rosenthal & Strange,
men-owned/led firms. If women-owned/led firms 2003, 2009; Acs & Armington, 2004; Stephens et al.,
experienced larger benefits from labor hoarding, 2013, for example). However, concerns have been
this may explain why women-owned/led firms were mounting over the impact of the declining role of
more likely to hoard labor. We find that labor hoard- startups (defined as firms less than 1 year old) in the
ing firms in general have significantly better sur- US economy (see Fig. 1). The negative demand shock
vival rates after the Great Recession. Accounting of the Great Recession affected sales (and employ-
for gender differences, we find that women-owned/ ment) forcing many firms to close (Fairlie, 2013;
led labor hoarding firms survive at higher rates than Shane, 2011). The Great Recession also significantly
male-led labor hoarding firms; however, we find that reduced the establishment startup share (shown in
women-led labor hoarding firms experience slower Fig. 1).
growth after the Great Recession. It seems that both Following a long trend of the declining share of
men-led and women-led firms who hoarded labor startups, the number of startups dropped dramatically,
during the Great Recession were more likely to sur- more than 25% between 2006 and 2009 as establish-
vive and grow in the wake of the Great Recession. ment entry rates dropped and establishment exit rates
In what follows, we review the literature on labor increased (U.S. Census Business Dynamics Survey).
hoarding and on firm survival; then, we describe Because new and small firms tend to be more flex-
our data and methodology in more detail, followed ible and open to change (and associated with inno-
by our results and conclusion. vation and growth in the economy), the declining
role of startups in the US economy may affect the
economic dynamics in how the USA recovers from
2 Literature review recessions. Clementi and Palazzo (2016) find that the
drop in the number of establishments contributed to
There is a well-documented relationship between the slow recovery from the Great Recession. Many
entrepreneurial activity and the health of the businesses were simply unable to survive the severity
Fig. 1 The decline in the share of startups, 1978–2020. Note: Establishments less than 1 year old are considered startups. The figure
looks similar when we use firms instead of establishments
Vol.: (0123456789)
1374 N. Greenhalgh‑Stanley et al.
of the Great Recession or were slow to recover from than cost retrenchment are better able to cope dur-
the recession. Thus, policies that increase firm sur- ing a recession, thereby increasing survival. Firms
vival during a recession may also help spur a quicker focused on skill accumulation may be more likely to
recovery. pay above equilibrium wages, “efficiency wages,” to
In addition to the business cycle, firm survival rates reduce turnover and retain skilled workers to increase
are affected by other industry and market dynam- productivity or efficiency (see for example, Campbell,
ics (see for example, Agarwal & Gort, 1996, 2002). 1993). Wage cuts (even during a recession) are infre-
Firm survival is also affected by firm characteristics. quent as wages tend to be rigid or “sticky” (Barattieri
For example, growing firms are more likely to survive et al., 2014; Bewley, 2002). Interviews with business
(Phillips & Kirchoff, 1989). More diverse firms in executives suggest that they avoided wage cuts during
terms of education, age, and gender also have higher a recession to help maintain morale (and the produc-
survival rates (Backmand and Kohlhase, 2020). Yet, tivity of their employees)—Bewley (2002). Business
the literature is mixed on whether firm survival rates leaders may choose to pay above equilibrium wages,
(as well as other firm performance measures) differ “efficiency wages,” even during a recession to main-
between women-owned and men-owned firms, with tain productivity, especially in the long term. When
some finding women-owned businesses have lower businesses choose to keep above equilibrium num-
survival rates (see for example, Robb, 2002) while bers of workers, they may be choosing to do this to
others find no difference (see for example, Robb & similarly maintain morale or productivity long term.
Watson, 2010, 2012). It may be more of a strategy of “efficiency labor” or
Lending especially to small businesses fell dramati- “workforce preservation” rather than “labor hoard-
cally during the Great Recession which increased down- ing” as it is typically referred to in the literature. Lay-
sizing particularly for firms that depended on external offs are also associated with reduced morale for the
financing (Duygan-Bump et al., 2015). Men-owned employees that remain at the firm (Iverson & Zatrick,
businesses have outperformed women-owned busi- 2011).
nesses, in part, because men-owned firms have higher Matsa and Miller (2014) find that women busi-
startup capital (Fairlie & Robb, 2009) and women- ness owners, all else equal, were less likely than
owned firms rely more on family and networks domi- their male counterparts to reduce employment dur-
nated by kin (Renzulli et al., 2000). During the Great ing the Great Recession (in their words, more likely
Recession, however, when capital was constrained, to “hoard labor”). With layoffs less likely in women-
women may have had an advantage, being used to oper- owned businesses during the Great Recession, coun-
ating with less capital and relying more on family. ties with a higher share of women-owned businesses
Differences in business leadership styles may also maintained higher stability during the recession (Del-
affect how a business responds to a recession, which ler, Conroy, and Watson, 2017). Yet, the question
may affect business performance during a crisis such of whether this provided more benefits (as a type of
as during a recession. For example, Mano-Negrin and “efficiency labor” or “workforce preservation”) or
Sheaffer (2004) find that women’s more participatory costs (from “labor hoarding”) to the businesses them-
style of decision-making is advantageous for women selves has yet to be answered in the literature. We
business leaders during times of crises. Women estimate how this labor hoarding strategy affected
tend to be more cautious and more oriented toward business outcomes post-recession and how these out-
the long run (Silverman, 2003). Women also tend comes differed between men- and women-owned/led
to be more averse to inequality, preferring fairness firms in the wake of the Great Recession. Specifically,
and exhibiting more altruistic behavior, taking into we test whether the labor hoarding strategy improved
account the interest of others (Adams & Funk, 2012; survival rates and business performance measures
Andreoni & Vesterlund, 2001; Croson & Gneezy, during the recovery from the Great Recession.
2009; Eagly & Carli, 2007). This may also mean that We also contribute to the literature on non-tradi-
women may prefer to avoid necessary layoffs in the tional/marginalized person’s business ownership and
face of a recession to the detriment of their business. potential resources that should be offered to help
On the other hand, Bartoloni et al. (2021) find these businesses survive. Chatterjee et al. (2022) find
that firms focusing more on skill accumulation rather that amongst rural women in poverty, having both
Vol:. (1234567890)
Do businesses benefit from labor hoarding? 1375
work experience and supportive environments leads include every establishment’s level of employment
to realistic expectations and more successful busi- and sales in 2007, 2006, and 2004, which accounts
nesses, while those without support and experience for differences in size both in levels and trends and is
tend to have unrealistic expectations about the busi- represented by 𝛽1 and 𝛽2.2 We also include firm-spe-
ness environment. For transgender business owners, cific measures of productivity (sales per employee)
Darden et al. (2022) find that they face more eco- and a measure of marketing expenditure (size of yel-
nomic disparities and that they value safe working low page advertisement). We also utilize two-digit
environments. Additionally, they find that their fear of SIC fixed effects to account for their industry, IndFE ,
failure drives them in different ways than traditional and US county fixed effects to account for their loca-
business owners, and they suggest that there should tion, CountyFE.
be additional assistance for transgender businesses.
Employmenti,2010 = 𝛽1 Employment2004,2006,2007
Yu et al. (2024) using an experimental design find
that after the Great Recession, which they refer to + 𝛽2 Sales2004,2006,2007
(1)
as the Great Financial Crisis, there was an increase + 𝛽3 X i,2007 + +IndFEi
in funding for women businesses, while using obser- + CountyFEl + 𝜀i
vational data, they find the opposite trend when the
COVID-19 pandemic hit the market. We then use the residual (the difference in the pre-
dicted employment change and their actual employ-
ment change during the recession) as a measure of
3 Data and methodology labor hoarding. We define a binary variable (hoardi )
equal to one if the business had higher employment
Our source of business information comes from the than predicted in 2010 (at the end of the recession)
Infogroup’s Historical Business Database and con- and equal to 0 if they had lower employment than
tains longitudinal data on all establishments in the predicted. As robustness, we defined a firm as labor
USA, with the unit of observation being the establish- hoarding if the firm’s residual is in the top quartile.
ment. The data consist of yearly information on every We first examine how labor hoarding affects a firm’s
business from 1997 to 2017 and are extensive, with economic outcomes (survival rate, employment and
approximately 30 million establishments each year. sales), Y, after the recession (2010 to 2017). Equa-
These data contain a wealth of information, includ- tion 2 shows that we regress these outcome variables
ing exact location, address, age of each establishment, on the previous controls of employment and sales,
number of employees, sales volume, detailed industry represented as X for simplicity, as well as the location
code (up to six digits), and corporate linkages. and industry fixed effects:
To measure labor hoarding, we first predict the
employment level immediately following the Great ΔY2017−2010 = 𝛽1 Hoardi + 𝛽1 X2004,2006,2007
Recession (2010) for every establishment (i) given +𝛽 3 X i,2007 + IndFEi + CountyFEl (2)
their industry, location, and growth in employ- + 𝜀i
ment and sales before the recession to estimate the
expected or predicted amount of employment in 2010 We then test whether the relationship between
given their size, trend, location, and industry (Eq. 1). labor hoarding and post-recession outcomes differs
We chose 2010 because 2009 had the largest loss of for women-owned/led firms. For these regressions,
employment across all major industries, which can be Eq. 3 shows we include an indicator for whether the
seen in Table 11 in the Appendix.1 Specifically, we firm is owned/led by a woman and an interacted term
between labor hoarding and female, in addition to our
previous controls:
1
In results not shown, we changed the reference to 2011 for
manufacturing, construction, and information because 2010
2
had some minor employment losses but this did not impact the We also examine what happens to new firms in 2007. For
results. Just for consistency, we choose to keep the reference those regressions, we obviously condition on 2007 levels since
year as 2010 for all industries. they did not exist in prior years.
Vol.: (0123456789)
1376 N. Greenhalgh‑Stanley et al.
ΔY2017−2010 == 𝛽0 + 𝛽1 Hoardi + 𝛽2 Femalei are larger than women-owned/led firms, which is
consistent with the literature on female entrepreneur-
+ 𝛽3 HoardxFemalei + 𝛽4 Xit +IndFEi (3)
ship. Lastly, men-owned/led firms are twice as likely
+ CountyFEl + 𝜀i
to be in manufacturing than their counterparts. Col-
Similar to Matsa and Miller (2014), we find that umns 5 through 8 show descriptive statistics of firms
women-led businesses are more likely to hoard.3 selected after using nearest-neighbor matching with
Whereas Matsa and Miller (2014) focus on women- three matches using a 0.1 caliper. Generally, we find
led firms, we focus our analysis on establishments the comparison between the labor hoarding firms and
that are either owned by women or led by women. those who did not much closer than without match-
The ability of women in firms to affect significant ing. However, we do find that labor hoarding firms
change, innovative, or otherwise depends on the are still slightly larger than the counterfactual firms
empowerment of women within the firm. Previous regardless of gender. However, the industrial mix is
research has found that the proportion of women in similar between the two groups.
top management positions has a positive effect on Given the observable differences between firms
business outcomes, including closing the gender wage that hold onto their employees during a recession
gap within companies, especially as women manag- and firms that do not, we supplement our analysis by
ers achieve higher status within the firm (Cohen & using propensity score matching to select only those
Huffman, 2007; Smith et al., 2006). Thus, the status non-labor hoarding comparison firms who are simi-
of women, more than the mere presence of a woman lar to the labor hoarding firms. Specifically, we use
(through ownership, for example) within a firm, may nearest-neighbor matching to select three non-labor
affect business outcomes from employment to innova- hoarding firms that are within the same US county,
tion and profit. Thus, we examine both women-owned in the same two-digit SIC industry, and matching on
and women-led firms, where women-led firms may be employment and sales in 2004, 2006, and 2007 with
a more direct measure of women’s influence over the a 10% caliber.4 Panel B of Table 1 displays the differ-
managerial and leadership decisions of the firm. ences between labor hoarding and non-labor hoarding
Table 1 presents descriptive statistics on employ- firms after removing firms without common support.
ment, sales, and number of businesses for all estab- We drop 120,077 or 3.4% of the firms for the exist-
lishments in our data set and across genders, whether ing business analysis and 2648 or 2.48% of firms for
they hoarded labor or not. The first four columns are the new business analysis, which makes sense given
statistics without using propensity score matching, how different these groups of firms were prior to the
while the last four columns are for businesses that matching.
were matched to find better matches between hoarders
and non-hoarders. We present employment and sales
in the pre-periods (2004, 2006, 2007) and immedi- 4 Results
ately after the Great Recession in 2010. Overall, a few
themes emerge. First, labor hoarding firms, those with The results presented in this paper include location
higher employment than predicted given their county, fixed effects (US counties) and two-digit standard
and two-digit industry, as well as their employment industrial code (SIC) fixed effects. We present esti-
and sales in 2004, 2006, and 2007, are quite large mates of firms who have been in business for 4 years
compared to non-labor hoarding firms (roughly five or more in 2007 as “existing” firms first and then
to six times larger). This persists throughout the pre- present estimates for businesses that were started in
period. Second, we find that male-owned/led firms 2007, which we define as “new” firms. For “exist-
ing” firm regressions, we control for the level of
3
Matsa and Miller (2014) examine labor hoarding by exam-
4
ining the degree of downsizing for firms (setting the vari- Our results are quite robust to alternative methods including
able to zero for firms with no downsizing or with employment the number of matches, the size of the caliber, matching within
growth). We will examine similar measures of downsizing in the same state (which is what Matsa and Miller (2014) do), and
our robustness section. one-digit SIC industry.
Vol:. (1234567890)
Do businesses benefit from labor hoarding? 1377
Table 1 Summary statistics by labor hoarding and gender for existing (4 years or older) firms in 2007
Without Propensity Score Matching With Propensity Score Matching
Labor Hoarding Firm Not Labor Hoarding Labor Hoarding Firm Not Labor Hoarding
Firm Firm
Male Female Male Female Male Female Male Female
Pre-period (1) (2) (3) (4) (5) (6) (7) (8)
# of Establishments 656,077 243,900 1,946,760 788,496 331,330 153,976 1,699,909 788,496
in 2007
Average # of Employ- 38.57 26.72 6.68 5.37 15 14.05 4.18 4.23
ees in 2007
(136.45) (97.16) (48.8) (29.6) (16.16) (12.9) (6.23) (6.76)
Average Sales (thou- 10261.42 4349.65 1608 881.1 2376.1 1620.7 838 619.2
sands) in 2007
(44272) (22464.9) (58970) (11487) (3793.8) (2118.3) (1731) (1089.38)
Average # of Employ- 37.62 25.84 6.77 5.53 15.27 13.96 4.19 4.24
ees in 2006
(133.57) (88.10) (47.3) (31.8) (16.19) (13.02) (6.23) (6.77)
Average Sales (thou- 9931.4 4158 1665.7 910.67 2357.9 1593.8 835.3 614.87
sands) in 2006
(43310) (21965) (59343) (11796) (3795.0) (2110.2) (1728.9) (1081.8)
Average # of Employ- 35.93 24.5 6.77 5.62 14.96 13.67 4.18 4.25
ees in 2004
(128.48) (86.75) (70.42) (30.66) (16.25) (13.07) (6.21) (6.78)
Average Sales (thou- 8183.5 3752.45 1353.6 812.14 1955.1 1524.02 696.58 545.19
sands) in 2004
(37960) (21926) (21298) (8401.6) (3004.7) (2058.8) (1377.1) (1023.0)
% Manufacturing in 11.42% 2.99% 6.60% 2.94% 5.04% 1.44% 4.58% 2.13%
2007
% Retail in 2007 30.46% 32.92% 26.50% 29.16% 36.66% 35.69% 27.31% 29.28%
% Services in 2007 23.05% 47.52% 32.50% 44.70% 26.10% 47.29% 34.46% 45.87%
Post-period
# of Establishments 656,077 243,900 1,946,760 788,496 331,330 153,976 1,699,909 788,496
in 2010
Average # of Employ- 45.89 32.52 4.88 4.37 20.06 18.79 3.74 3.85
ees in 2010
(165) (117.7) (22.44) (10.05) (47.28) (52.41) (4.03) (4.33)
Average Sales (thou- 12063 5622.5 1218.4 713.3 3325.1 2289.4 764.02 567.32
sands) in 2010
(109528) (168437) (45862) (2965) (9085.5) (10147.7) (1497.0) (781.26)
employment and sales variables in 2007, 2006, and the 2004, 2006, and 2007 employment and sales with
2004 to account for observable differences in pre- both location (US county) and industry (two-digit
trends. Effectively, we are comparing firms that are SIC) fixed effects and then differencing the predicted
in the same county, in the same two-digit SIC indus- 2010 employment with their actual employment. The
try, and we account for differences in employment idea is that the predicted level of 2010 employment is
and sales before the Great Recession. Of course, for what was expected given their employment and sales
“new” firm regressions, we can only include employ- trajectory, location, and industry. Finally, because we
ment and sales in 2007. A reminder is that our “labor are working with a data set that has all establishments
hoarding” variable is determined by first regressing in the USA, the number of observations in the regres-
employment in 2010 (after the Great Recession) on sions is of such magnitude that all of our coefficients
Vol.: (0123456789)
1378 N. Greenhalgh‑Stanley et al.
Table 2 The effect of labor hoarding during the 2009 Recession on the likelihood of survival in 2017
OLS diff-in-diff PSM diff-in-diff
Existing firms in 2007 LH indicator LH indicator Top quartile LH LH indicator LH indicator Top quartile LH
(1) (2) (3) (4) (5) (6)
Labor hoarding (LH) 0.0916*** 0.0900*** 0.100*** 0.0416*** 0.0159*** 0.0629***
(0.000583) (0.000677) (0.000623) (0.000817) (0.00159) (0.000824)
Women-owned and LH 0.00525*** 0.00292** 0.0367*** 0.00902***
(0.00126) (0.00116) (0.000958) (0.00138)
Women-owned −0.0151*** −0.0149*** −0.0142*** −0.0140***
(0.000658) (0.000689) (0.000708) (0.000740)
Constant 0.663*** 0.667*** 0.657*** 0.641*** 0.645*** 0.645***
(0.000281) (0.000339) (0.000357) (0.000336) (0.000397) (0.000397)
Observations 3,635,229 3,635,229 3,635,229 2,898,454 2,898,454 2,898,454
County fixed effects Yes Yes Yes Yes Yes Yes
2-digit SIC fixed effects Yes Yes Yes Yes Yes Yes
Emp. and sales controls in Yes Yes Yes Yes Yes Yes
04, 06, 07
Labor hoarding (LH) is a measure of employment in 2010 minus the expected employment in 2010 conditioned on the location
(county), 2-digit SIC industry, and employment and sales in 2004, 2006, and 2007. Columns 1 and 2 use an indicator variable of
labor hoarding while columns 3 and 4 use an indicator variable of whether they are in the top tercile or quartile of labor hoarding.
are statistically significant, so much of our discus- want to understand whether there are differences for
sion is focused on whether they are economically women-owned/led versus men-owned/led businesses
meaningful.5 when they choose to labor hoard, especially coming
out of a recession. Matsa and Miller (2014) find there
4.1 Survival are differences in downsizing rates between genders,
and the previous literature has also found that differ-
We begin by examining whether keeping more ences between men and women, such as risk aver-
employees than expected, “labor hoarding,” affects sion, may also lead to differences (see Flippin (2022)
the likelihood a business is still operating in 2017 and Holden and Tilahun (2022) as examples of the
(7 years after the Great Recession). Starting with research on women being more risk averse).
“existing” businesses in 2007, which have been in Column 2 of Table 2 adds in the interaction term
business for at least 4 years, the constant term in for being a labor hoarding firm with being women-
Column 1 of Table 2 can be interpreted as the sur- owned/led, while column 3 uses a more restric-
vival rate for non-hoarding firms, which was 66.3%. tive measure of labor hoarding by indicating only
Meanwhile, the indicator variable (dummy variable) those firms in the top quartile of labor hoarders. We
for labor hoarding firms suggests those firms were find very similar effects across both measures. We
9.16% more likely to survive. This higher survival find that men-owned labor hoarding firms are 9% to
rate likely reflects the benefits of workforce pres- 10% more likely to survive than non-hoarding firms.
ervation by reducing costly turnover, maintaining Women-owned labor hoarding firms experience an
firm-specific human capital, and ensuring firms were additional survival advantage of about 0.3% to 0.5%
better positioned to recover when demand returned. compared to their male counterparts. Thus, women-
While this is interesting in its own right, we also owned labor hoarding firms experience a slightly
larger benefit (9.5% to 10.3% more likely to survive
compared to 9% to 10%) from labor hoarding than
5
men-owned firms. Women-owned firms that did not
The decrease in sample size in the employment and sales
results is due to missing data on employment and sales for hoard labor were about 1.5% less likely to survive
these firms. than non-hoarding men-owned firms.
Vol:. (1234567890)
Do businesses benefit from labor hoarding? 1379
These findings are consistent with prior evidence 4.1.1 New firms
that women entrepreneurs often face tighter finan-
cial constraints and have less access to credit markets New firms have a higher propensity to close as they
(Coleman & Robb, n.d.). At the same time, research are more financially vulnerable and less established.
suggests women leaders are more likely to adopt The U.S. Bureau of Labor Statistics estimates that
long-term, participatory decision-making styles and 45% of all businesses fail within their first 5 years,
place greater emphasis on fairness and employee with approximately 20% failing after just 2 years
well-being (Croson & Gneezy, 2009; Mano-Negrin (BLS Table 7). Since new firms are particularly at
& Sheaffer, 2004). Together, these factors make risk, we want to study survival rates for new firms
women-owned businesses less likely to resort to lay- (in 2007) that choose to keep more workers than
offs, treating workforce preservation not only as a expected and present the same analysis but for new
financial strategy to avoid costly turnover but also as firms in Table 3, with the exception that we can only
a means of maintaining firm-specific human capital, add employment and sales controls for 2007. Col-
loyalty, and reputation during challenging economic umn 1 shows the baseline survival rate of non-labor
periods. hoarding new businesses in 2007, which have lower
Columns 4 to 6 present the same analysis using survival rates (57.8%) compared to those of existing
nearest-neighbor propensity score matching to businesses (66.3%) and that new firms benefit simi-
exclude non-labor hoarding firms that are not com- larly to labor hoarding, with an increase in survival
parable to labor hoarding firms. The results remain 7 years later of 9.3%. Columns 2 and 3 present the
largely consistent; there is a reduction in magnitude effects of labor hoarding across genders. Men-owned/
with men-owned labor hoarding firms 2% to 6% more led firms that labor hoard are 8.7% to 9.8% more
likely to survive (compared to 9% to 10%), while the likely to survive, while women-owned/led firms sur-
magnitude of the effect is slightly larger with women- vived an additional 2% compared to men-owned/led
owned hoarding firms experiencing an additional labor hoarding firms. New women-owned/led busi-
survival advantage of 0.9% to 4% (compared to 0.3% nesses that did not hoard labor are roughly 3% less
to 0.5%). Notably, despite the loss of over 2 million likely to survive, which is twice as large as existing
observations, the findings remain robust. women-owned/led non-labor hoarding firms.
Table 3 The effect of labor hoarding during the 2009 Recession on the likelihood of survival in 2017
OLS diff-in-diff PSM diff-in-diff
New firms in 2007 LH indicator LH indicator Top quartile LH LH indicator LH indicator Top quartile LH
(1) (2) (3) (4) (5) (6)
Labor hoarding (LH) 0.0930*** 0.0868*** 0.0982*** −0.0174 −0.0119 0.0295**
(0.00217) (0.00258) (0.00233) (0.000717) (0.0155) (0.0138)
Women-owned and LH 0.0190*** 0.0202*** −0.0126 −0.0407*
(0.00459) (0.00415) (0.0258) (0.0224)
Women-owned −0.0271*** −0.0295*** −0.0347*** −0.0286**
(0.00238) (0.00253) (0.0107) (0.0111)
Constant 0.578*** 0.586*** 0.574*** 0.641*** 0.651*** 0.646***
(0.00104) (0.00128) (0.00136) (0.00565) (0.00643) (0.00652)
Observations 291,376 291,376 291,376 13,472 13,472 13,472
County fixed effects Yes Yes Yes Yes Yes Yes
2-digit SIC fixed effects Yes Yes Yes Yes Yes Yes
Emp. and sales controls in 07 Yes Yes Yes Yes Yes Yes
Labor hoarding (LH) is a measure of employment in 2010 minus the expected employment in 2010 conditioned on the location
(county), 2-digit SIC industry, and employment and sales in 2004, 2006, and 2007. Columns 1 and 2 use an indicator variable of
labor hoarding while columns 3 and 4 use an indicator variable of whether they are in the top tercile or quartile of labor hoarding.
Vol.: (0123456789)
1380 N. Greenhalgh‑Stanley et al.
Including only non-labor hoarding firms that are from having higher employment than anticipated of
similar to labor hoarding firms lowers the estimated about half a percent (0.4% to 0.64% in OLS and 0%
effect of labor hoarding, with only the top quartile of to 0.28% in PSM). The results from Table 4 suggest
new firms seeing an increased likelihood of survival that overall labor hoarding is associated with slower
of 3% and no additional advantage for women-owned employment growth (about 5.5%) during the recovery
new firms that labor hoard. and about 5.7% for men and 5% for women-led busi-
nesses. This is likely due to businesses that did not
4.2 Employment hoard labor being forced to catch up on hiring after
the recession, whereas businesses that kept their labor
Tables 4 and 5 explore how employment changes were already right-sized for the recovery. Columns 4
from 2010 to 2017 in the post period of the reces- to 6 present results using propensity score matching
sion for labor hoarders vs non-labor hoarders and to improve the comparison in the pre-period of non-
comparing women-owned/led versus men-owned/led labor hoarding businesses. We continue to find a sim-
businesses. Results from Table 4 suggest that firms ilar pattern across all the results, with labor hoarding
with more employment in 2010 than anticipated (our businesses owned/led by men with about 7% lower
definition of labor hoarders) had a slower growth rate employment growth post-recession and 4% lower
from 2010 to 2017 than those that had less employ- employment growth for women-owned firms (a 3%
ment than anticipated in 2010. Remember, Table 1 smaller effect for women-owned businesses).
suggests that labor hoarders typically maintained (or
grew) their employment during the recession while
non-labor hoarders lowered their employment. Essen- 4.2.1 New firm employment
tially, if a firm let go of a lot of employees during the
recession, then they had to hire more people back For new firms, results from Table 5 find that labor
after the recession, which makes intuitive sense. The hoarding had a smaller effect on employment in the
breakdown by gender shows that women-owned/led period after the recession, with coefficients of 0.7%
business hoarders grew faster or they benefited more with no difference for women-owned businesses.
Table 4 The effect of labor hoarding during the 2009 Recession on employment growth from 2010 to 2017
OLS diff-in-diff PSM diff-in-diff
Existing firms in 2007 LH indicator LH indicator Top quartile LH LH indicator LH indicator Top quartile LH
(1) (2) (3) (4) (5) (6)
Labor hoarding (LH) −0.0553*** −0.0570*** −0.0876*** −0.0686*** −0.0695*** −0.0704***
(0.000389) (0.000451) (0.000581) (0.000717) (0.000835) (0.000698)
Women-owned and LH 0.00652*** 0.00403*** 0.00279** −0.000256
(0.000843) (0.00108) (0.00137) (0.00117)
Women-owned −0.00287*** −0.00372*** −0.00206*** −0.00140**
(0.000430) (0.000661) (0.000606) (0.000637)
Constant 0.0103*** 0.0103*** −0.00190*** 0.0187*** 0.0193*** 0.0251***
(0.000183) (0.000183) (0.000195) (0.000287) (0.000338) (0.000350)
Observations 2,871,029 2,871,029 2,871,029 1,893,741 1,893,741 1,893,741
County fixed effects Yes Yes Yes Yes Yes Yes
2-digit SIC fixed effects Yes Yes Yes Yes Yes Yes
Emp. and sales controls Yes Yes Yes Yes Yes Yes
in 04, 06, 07
Labor hoarding (LH) is a measure of employment in 2010 minus the expected employment in 2010 conditioned on the location
(county), 2-digit SIC industry, and employment and sales in 2004, 2006, and 2007. Columns 1 and 2 use an indicator variable of
labor hoarding while columns 3 and 4 use an indicator variable of whether they are in the top tercile or quartile of labor hoarding.
Dependent variable is the Δ ln(emp) in from 2010 to 2017.
Vol:. (1234567890)
Do businesses benefit from labor hoarding? 1381
Table 5 The effect of labor hoarding during the 2009 Recession on employment growth from 2010 to 2017
OLS diff-in-diff PSM diff-in-diff
New firms in 2007 LH indicator LH indicator Top quartile LH LH indicator LH indicator Top quartile LH
(1) (2) (3) (4) (5) (6)
Labor hoarding (LH) −0.00138 −0.00681*** −0.0666*** −0.0836*** −0.0861*** −0.0987***
(0.00168) (0.00196) (0.00175) (0.0142) (0.0168) (0.0139)
Women-owned and LH 0.00305 0.00271 0.00778 0.0189
(0.00347) (0.00311) (0.0283) (0.0225)
Women-owned −0.00652*** −0.00633*** −0.00447 −0.00731
(0.00178) (0.00191) (0.00987) (0.0102)
Constant 0.0250*** 0.0270*** 0.0336*** 0.0436*** 0.0448*** 0.0480***
(0.000775) (0.000949) (0.00102) (0.00550) (0.00611) (0.00615)
Observations 209,880 209,880 209,880 8391 8391 8391
County fixed effects Yes Yes Yes Yes Yes Yes
2-digit SIC fixed effects Yes Yes Yes Yes Yes Yes
Emp. and sales controls in 07 Yes Yes Yes Yes Yes Yes
Labor hoarding (LH) is a measure of employment in 2010 minus the expected employment in 2010 conditioned on the location
(county), 2-digit SIC industry, and employment and sales in 2004, 2006, and 2007. Columns 1 and 2 use an indicator variable of
labor hoarding while columns 3 and 4 use an indicator variable of whether they are in the top tercile or quartile of labor hoarding.
Dependent variable is the Δ ln(emp) in from 2010 to 2017.
Columns 4 to 6 show results with propensity score maintain their productivity to maintain employment
matching and find that new firms have more simi- or got used to the lower productivity level required
lar results to existing firms, with firms that kept their during a time of low demand. The result is similar
employees through the recession seeing about 9% for women-and men-owned firms, though slightly
lower employment growth post-recession (2010–2017). lower for women-owned firms in the top quartile.
Results by gender did not find significant differences Table 7 shows that new firms had a similar effect
between men and women-owned/led businesses. with slightly larger magnitudes. Labor hoarding
new firms grew their sales 7% slower than their
non-hoarding counterparts, though it is 14% slower
4.3 Sales in the PSM results. Again, our PSM results suggest
that this is true across genders, though smaller for
Our results for sales show a similar story. Table 6 women-owned businesses in the top quartile that
shows sales at existing businesses grew slower for hoard labor (7% slower growth in sales).
those who have higher employment than anticipated,
and the propensity score matching suggests it is 4.4 High‑skilled vs. low‑skilled impacts
consistent across gender. Our results for the impact
of labor hoarding on sales, similar to the effect on Given that training costs and replacement costs can
employment, show that labor hoarding during the vary by high- versus low-skilled workers and in light
recession is associated with lower sales growth dur- of recent work highlighting that recessions tend to
ing the recovery period, with OLS results suggest- accelerate the replacement of low-skilled jobs with
ing a loss of 7%, while PSM results suggest a loss capital (Jaimovich and Siu,2020, Fedotenkov et al.,
of 6%. This could be indicative of (1) the increased 2024), we want to examine effects between firms
costs associated with labor hoarding affecting busi- that rely on high- and low-skilled workers. To do so,
ness performance during the recovery or (2) a disin- we follow Doms et al. (2010) and calculate the per-
centive effect associated with labor hoarding where cent of a firm’s industry’s workforce that is college
employees who avoided layoffs during the recession educated, calculated at the three-digit NAICS code
when demand was low may not feel as motivated to using data from the Bureau of Labor and Statistics.
Vol.: (0123456789)
1382 N. Greenhalgh‑Stanley et al.
Table 6 The effect of labor hoarding during the 2009 Recession on sales growth from 2010 to 2017
OLS diff-in-diff PSM diff-in-diff
Existing firms in 2007 LH indicator LH indicator Top quartile LH LH indicator LH indicator Top quartile LH
(1) (2) (3) (4) (5) (6)
Labor hoarding (LH) −0.0662*** −0.0734*** −0.0687*** −0.0558*** −0.0581*** −0.0517***
(0.00106) (0.00122) (0.00111) (0.00139) (0.00161) (0.00134)
Women-owned and LH 0.0259*** 0.0248*** 0.00834*** 0.00991***
(0.00229) (0.00208) (0.00264) (0.00224)
Women-owned −0.0296*** −0.0313*** −0.0264*** −0.0276***
(0.00119) (0.00126) (0.00115) (0.00121)
Constant −0.105*** −0.0972*** −0.0921*** −0.118*** −0.110*** −0.115***
(0.000506) (0.000607) (0.000644) (0.000551) (0.000645) (0.000633)
Observations 2,333,308 2,333,308 2,333,308 1,833,357 1,833,357 1,833,357
County fixed effects Yes Yes Yes Yes Yes Yes
2-digit SIC fixed effects Yes Yes Yes Yes Yes Yes
Emp. and sales controls in Yes Yes Yes Yes Yes Yes
04, 06, 07
Labor hoarding (LH) is a measure of employment in 2010 minus the expected employment in 2010 conditioned on the location
(county), 2-digit SIC industry, and employment and sales in 2004, 2006, and 2007. Columns 1 and 2 use an indicator variable of
labor hoarding while columns 3 and 4 use an indicator variable of whether they are in the top tercile or quartile of labor hoarding.
Dependent variable is the Δ ln(sales) in from 2010 to 2017.
Table 7 The effect of labor hoarding during the 2009 Recession on sales growth from 2010-17
OLS diff-in-diff PSM diff-in-diff
New firms in 2007 LH indicator LH indicator Top quartile LH LH indicator LH indicator Top quartile LH
(1) (2) (3) (4) (5) (6)
Labor hoarding (LH) −0.0624*** −0.0656*** −0.0643*** −0.121*** −0.142*** −0.144***
(0.00338) (0.00399) (3.99e−05) (0.0270) (0.0320) (0.0264)
Women-owned and LH 0.00911 0.0139** 0.0685 0.0741*
(0.00713) (0.00638) (0.0540) (0.0429)
Women-owned −0.0330*** −0.0313*** −0.0628*** −0.0692***
(0.00364) (0.00326) (0.0187) (0.0194)
Constant 0.118*** 0.128*** 0.134*** −0.0695*** −0.0527*** −0.0481***
(0.00157) (0.00191) (0.00205) (0.0104) (0.0115) (0.0116)
Observations 201,806 201,806 201,806 8230 8230 8230
County fixed effects Yes Yes Yes Yes Yes Yes
2-digit SIC fixed effects Yes Yes Yes Yes Yes Yes
Emp. and sales controls in 07 Yes Yes Yes Yes Yes Yes
Labor hoarding (LH) is a measure of employment in 2010 minus the expected employment in 2010 conditioned on the location
(county), 2-digit SIC industry, and employment and sales in 2004, 2006, and 2007. Columns 1 and 2 use an indicator variable of
labor hoarding while columns 3 and 4 use an indicator variable of whether they are in the top tercile or quartile of labor hoarding.
Dependent variable is the Δ ln(Sales) in from 2010 to 2017.
We define a high-skilled firm as a firm whose work- industries include education services, professional
force is at least 50% college educated and a low- scientific services, information services, and man-
skilled firm as a firm whose workforce is less than agement of companies, while low-skilled three-digit
25% college educated. High-skilled three-digit NAICS industries include apparel manufacturing, repair and
Vol:. (1234567890)
Do businesses benefit from labor hoarding? 1383
maintenance, truck transportation, and food service, more firm-specific knowledge, and may be more criti-
to name a few. Table 8 compares the effects of labor cal to maintaining productivity and innovation. Thus,
hoarding for firms in high-skill versus low-skill indus- preserving high-skilled labor may yield greater ben-
tries. Overall, the results are consistent across both efits for firm performance during recovery compared
groups, suggesting that the survival and employment to retaining lower-skilled workers.
effects of workforce preservation are not confined to
one skill segment of the labor market. However, one
small difference does emerge: for high-skill intensive 5 Alternative specification
firms, the negative effect on sales growth from labor
hoarding is smaller for high-skill intensive firms. Our measure of “labor hoarding,” estimating the
High-skilled workers may be harder to replace, carry expected amount of labor conditional on location,
Table 8 Comparing the effect of labor hoarding between high-skilled reliant firms and low-skilled reliant firms on existing firm sur-
vival, and the growth rate of employment and sales from 2010 to 2017
OLS diff-in-diff
Existing firms in 2007 Survival in 2017 Employment Δ (2017–2010) Sales Δ (2017–2010)
Low-skill High-skill Low-skill High-skill Low-skill High-skill
(1) (2) (3) (4) (5) (6)
Labor hoarding (LH) 0.0849*** 0.0852*** −0.0801*** −0.0950*** −0.0826*** −0.0728***
(0.00144) (0.00149) (0.00119) (0.00154) (0.00218) (0.00256)
Women-owned and LH 0.0405*** −0.0223*** 0.0114*** 0.0121*** 0.0336*** −0.00918**
(0.00274) (0.00237) (0.00225) (0.00240) (0.00409) (0.00395)
Women-owned −0.0219*** 0.0263*** −0.00985*** 0.00232* −0.0227*** 0.0339***
(0.00163) (0.00119) (0.00136) (0.00122) (0.00248) (0.00196)
Constant 0.684*** 0.652*** 0.0296*** 0.0225*** 0.0458*** −0.227***
(0.000728) (0.000697) (0.000604) (0.000726) (0.00110) (0.00116)
Observations 734,780 936,966 505,600 620,769 499,962 568,826
PSM diff-in-diff
Existing firms in 2007 Survival in 2017 Employment Δ (2017–2010) Sales Δ (2017–2010)
Low-skill High-skill Low-skill High-skill Low-skill High-skill
(1) (2) (3) (4) (5) (6)
Labor hoarding (LH) 0.0436*** 0.0398*** −0.0772*** −0.0724*** −0.0866*** −0.0475***
(0.00187) (0.00199) (0.00147) (0.00191) (0.00263) (0.00321)
Women-owned and LH 0.0512*** −0.0116*** 0.00805*** 0.00963*** 0.0359*** −0.00855*
(0.00322) (0.00300) (0.00246) (0.00281) (0.00441) (0.00463)
Women-owned −0.0241*** 0.0284*** −0.00703*** 0.00409*** −0.0238*** 0.0370***
(0.00176) (0.00127) (0.00136) (0.00116) (0.00244) (0.00188)
Constant 0.673*** 0.631*** 0.0213*** 0.0190*** 0.0381*** −0.233***
(0.000818) (0.000803) (0.000628) (0.000740) (0.00112) (0.00124)
Observations 602,915 774,793 409,625 504,144 406,200 470,201
County fixed effects Yes Yes Yes Yes Yes Yes
2-digit SIC fixed Yes Yes Yes Yes Yes Yes
effects
Emp. and sales controls Yes Yes Yes Yes Yes Yes
in 04, 06, 07
Labor hoarding (LH) is a measure of employment in 2010 minus the expected employment in 2010 conditioned on the location
(county), 2-digit SIC industry, and employment and sales in 2004, 2006, and 2007. A firm is considered a high-skilled reliant firm
if their industry’s workforce has at least 50% college educated, while a low-skilled reliant firm’s industry workforce is less than 25%
college educated.
Vol.: (0123456789)
1384 N. Greenhalgh‑Stanley et al.
industry, growth, and level of their employment and main measure of labor hoarding measures a positive
sales, is unique to the literature. To alleviate concerns impact on employment, so similar results would have
that our results are driven by this uniqueness, opposite signs to our main results. Also, note that a
we present estimates using whether or not a firm firm that lowers its employment less than expected
downsizes or not, which is similar to Matsa and would be indicated as a “labor hoarder” in our main
Miller (2014). Now, the variable of interest is results, as more of its employment was maintained
whether the firm had lower employment in 2010 at given its industry, size, etc.
the end of the recession compared to 2007 before Beginning with Table 9, we find the same exact
the recession. Tables 9 and 10 present the results pattern as our previous results. When firms downsize
for existing and new firms, respectively, examining during the recession, it lowers the likelihood that they
survival rates, employment growth, and sales growth. survive in the post-recession period of 2010 to 2017
Technically, an indicator of downsizing identifies (about 5% less likely for the PSM). For the growth
a “negative” impact on employment, whereas our rates of employment and sales, we continue to find
Table 9 The effect of downsizing during the 2009 Recession on existing firm survival, and the growth rate of employment and sales
from 2010 to 2017
OLS diff-in-diff
Existing firms in 2007 Survival in 2017 Employment Δ (2017–2010) Sales Δ (2017–2010)
(1) (2) (3) (4) (5) (6)
Downsized −0.0269*** −0.0261*** 0.0749*** 0.0780*** 0.0650*** 0.0622***
(0.000701) (0.000830) (0.000725) (0.000853) (0.00138) (0.00162)
Women-owned and −0.00257* −0.0107*** 0.0108***
downsized
(0.00154) (0.00161) (0.00307)
Women-owned −0.0146*** −0.000538 −0.0243***
(0.000628) (0.000587) (0.00112)
Constant 0.688*** 0.692*** −0.00509*** −0.00493*** −0.127*** −0.121***
(0.000264) (0.000319) (0.000245) (0.000296) (0.000470) (0.000564)
Observations 3,635,229 3,635,229 2,432,359 2,432,359 2,333,308 2,333,308
PSM diff-in-diff
Existing firms in 2007 Survival in 2017 Employment Δ (2017–2010) Sales Δ (2017–2010)
(1) (2) (3) (4) (5) (6)
Downsized −0.0438*** −0.0447*** 0.0528*** 0.0525*** 0.0521*** 0.0491***
(0.000820) (0.000985) (0.000776) (0.000929) (0.00148) (0.00176)
Women-owned and 0.00340* 0.00108 0.0109***
downsized
(0.00175) (0.00166) (0.00316)
Women-owned −0.0112*** −0.00259*** −0.0270***
(0.000696) (0.000589) (0.00112)
Constant 0.649*** 0.652*** 0.0104*** 0.0112*** −0.125*** −0.117***
(0.000341) (0.000400) (0.000290) (0.000339) (0.000558) (0.000648)
Observations 2,898,454 2,898,454 1,893,741 1,893,741 1,833,357 1,833,357
County fixed effects Yes Yes Yes Yes Yes Yes
2-Digit SIC fixed Yes Yes Yes Yes Yes Yes
effects
Emp. and sales con- Yes Yes Yes Yes Yes Yes
trols in 04, 06, 07
Downsized is an indicator that the firm had lower employment in 2010 compared to 2007.
Vol:. (1234567890)
Do businesses benefit from labor hoarding? 1385
the same pattern regardless of whether we use OLS similar in magnitude. A comparison of results for sales
or propensity score matching. Firms that downsized between the two measures finds our labor hoarding
during the recession grew their employment post- measure having a loss of −7% (OLS) and −6% (PSM)
recession more. Again, this could simply be these compared to a growth of 7% (OLS) and 5% (PSM) for
firms trying to rehire workers to get back to their the downsizing measure. Table 10 shows the analysis
optimal level of employment, and with the added for downsizing on new firms, and the results are
workers, they can produce more and increase their similar. The firms that labor hoard or do not downsize
sales. A comparison of results for our main labor have higher survival rates but slower growth in
hoarding measure with the downsizing measure shows employment and sales in the post-recession. Again, we
an employment loss of −6% (OLS) and −7% (PSM) expect the signs to reverse for the results to be similar.
compared to a growth of 7% (OLS) and 5% (PSM), so
Table 10 The effect of downsizing during the 2009 Recession on new firm survival, and the growth rate of employment and sales
from 2010 to 2017
OLS diff-in-diff
New firms in 2007 Survival in 2017 Employment Δ (2017–2010) Sales Δ (2017–2010)
(1) (2) (3) (4) (5) (6)
Downsized −0.0503*** −0.0549*** 0.101*** 0.108*** 0.0917*** 0.0892***
(0.00192) (0.00230) (0.00255) (0.00306) (0.00437) (0.00519)
Women-owned and 0.0143*** −0.0206*** 0.00816
downsized
(0.00411) (0.00546) (0.00949)
Women-owned −0.0258*** −0.00637** −0.0372***
(0.00220) (0.00271) (0.00469)
Constant 0.611*** 0.619*** 0.0211*** 0.0230*** −0.0948*** −0.0838***
(0.000956) (0.00118) (0.00116) (0.00143) (0.00198) (0.00241)
Observations 348,605 348,605 201,904 201,904 182,435 182,435
PSM diff-in-diff
New firms in 2007 Survival in 2017 Employment Δ (2017–2010) Sales Δ (2017–2010)
(1) (2) (3) (4) (5) (6)
Downsized −0.0812*** −0.0870*** 0.0446*** 0.0631*** 0.0669** 0.0690*
(0.0132) (0.0163) (0.0163) (0.0205) (0.0309) (0.0388)
Women-owned and 0.0199 −0.0497 −0.00140
downsized
(0.0274) (0.0337) (0.0636)
Women-owned −0.0352*** −0.00436 −0.0581***
(0.0106) (0.00976) (0.0184)
Constant 0.672*** 0.683*** 0.0230*** 0.0242*** −0.0915*** −0.0752***
(0.00432) (0.00533) (0.00385) (0.00473) (0.00726) (0.00889)
Observations 13,472 13,472 8391 8391 8230 8230
County fixed effects Yes Yes Yes Yes Yes Yes
2-digit SIC fixed Yes Yes Yes Yes Yes Yes
effects
Emp. and sales controls Yes Yes Yes Yes Yes Yes
in 04, 06, 07
Downsized is an indicator that the firm had lower employment in 2010 compared to 2007.
Vol.: (0123456789)
1386 N. Greenhalgh‑Stanley et al.
6 Conclusion to explain why women-owned businesses are more
likely to hoard labor or preserve their workforce during
This study examines the effects of labor hoarding a recession. However, these survival benefits come with
during the Great Recession on firm performance trade-offs: labor hoarding firms tended to grow more
(survival, employment, and sales) during the recovery slowly in employment and sales during the recovery,
to discover whether labor hoarding serves as a strategic while non-hoarding firms “caught up” by rehiring and
advantage or an added cost, with a focus on the expanding more aggressively.
differential impact of labor hoarding for women-owned/ Thus, programs like the Paycheck Protection Pro-
led businesses (as women-owned businesses are more gram (PPP) may have helped boost business survival
likely to hoard labor in a downturn). rates post-COVID. Labor hoarding or really work-
Estimating the causal effects is challenging given force preservation helps to reduce the volatility associ-
concerns about endogeneity and omitted variable ated with the boom-and-bust cycle of a recession and
bias, but we mitigate these issues by using propensity recovery. Framing labor hoarding as workforce pres-
score matching within counties and industries and ervation highlights its role not as inefficiency but as a
controlling for pre-recession growth trajectories deliberate strategy for resilience and stability—espe-
(2004–2007). It is important to note that firm-specific cially valuable in the face of systemic shocks.
unobservables, such as managerial ability and access
to informal credit, may still exist.
We find that firms engaging in labor hoarding, better Data availability The paper uses proprietary data that can be
purchased from Infogroup, a private company.
described as “workforce preservation,” were significantly
more likely to survive in the wake of the recession. The Declarations
cumulative effect is sizable: firms that preserved their
workforce were 9–10% more likely to survive over the Conflict of interest The authors have no interests that are di-
rectly or indirectly related to the work submitted for publication.
2010–2017 period, with slightly stronger effects for This research is not sponsored or funded by anyone.
women-owned and women-led businesses. This helps
Appendix
Table 11 US national employment by industry, 2007–2015 (in yearly percent change)
Industry 2008 2009 2010 2011 2012 2013 2014 2015
Total nonfarm − 0.60% − 5.20% − 0.40% 1.70% 1.80% 2.00% 1.90% 2.00%
Construction − 4.20% − 21.50% − 1.80% 1.80% 1.80% 1.70% 5.10% 4.80%
Manufacturing − 2.20% − 11.90% − 1.70% 0.90% 0.90% 0.80% 1.70% 1.70%
Wholesale trade − 0.80% − 8.50% 0.00% 0.90% 0.90% 0.90% 0.90% 0.90%
Retail trade − 1.30% − 5.90% 0.70% 0.70% 0.70% 1.40% 1.30% 1.30%
Transportation and warehousing 1.10% − 4.30% 0.00% 1.10% 1.10% 1.10% 2.20% 2.10%
Information − 0.60% − 6.20% − 3.10% 0.00% 0.00% − 1.80% − 0.40% 0.40%
Financial activities 0.00% − 4.70% − 2.50% 0.00% 0.60% 0.60% 0.60% 0.60%
Professional and business services − 1.10% − 7.50% − 0.60% 3.10% 3.00% 3.50% 3.40% 3.80%
Education and health services 1.60% 0.50% 1.00% 2.00% 2.50% 2.90% 2.80% 2.80%
Leisure and hospitality − 0.80% − 4.60% 0.00% 2.40% 2.30% 1.50% 2.30% 2.20%
Other services − 0.90% − 3.60% − 0.90% 0.00% 0.90% 0.90% 0.90% 0.90%
Government 0.00% − 0.90% 0.00% 0.50% 0.00% 0.00% 0.00% 0.00%
Notes: Data comes from the Bureau of Labor Statistics
Vol:. (1234567890)
Do businesses benefit from labor hoarding? 1387
Table 12 The effect of labor hoarding during the 2009 Recession on the likelihood of survival in 2013
OLS diff-in-diff PSM diff-in-diff
Existing firms in 2007 LH indicator LH indicator Top quartile LH LH indicator LH indicator Top quartile LH
(1) (2) (3) (4) (5) (6)
Labor hoarding (LH) 0.0494*** 0.0472*** 0.0502*** 0.0283*** 0.0258*** 0.0291***
(0.000444) (0.000516) (0.000475) (0.000843) (0.000994) (0.000877)
Women-owned and LH 0.00765*** 0.00620*** 0.00791*** 0.00802***
(0.000963) (0.000880) (0.00165) (0.00144)
Women-owned − 0.00914*** − 0.00937*** − 0.00769*** − 0.00840***
(0.000501) (0.000526) (0.000717) (0.000747)
Constant 0.843*** 0.846*** 0.841*** 0.826*** 0.829*** 0.828***
(0.000214) (0.000258) (0.000272) (0.000388) (0.000443) (0.000449)
Observations 3,635,229 3,635,229 3,635,229 1,647,425 1,647,425 1,647,425
County fixed effects Yes Yes Yes Yes Yes Yes
2-digit SIC fixed effects Yes Yes Yes Yes Yes Yes
Emp. and sales controls Yes Yes Yes Yes Yes Yes
in 04, 06, 07
Labor hoarding (LH) is a measure of employment in 2010 minus the expected employment in 2010 conditioned on the location
(county), 2-digit SIC industry, and employment and sales in 2004, 2006, and 2007. Columns 1 and 2 use an indicator variable of
labor hoarding while columns 3 and 4 use an indicator variable of whether they are in the top tercile or quartile of labor hoarding.
Table 13 The effect of labor hoarding during the 2009 Recession on the likelihood of survival in 2013
OLS diff-in-diff PSM diff-in-diff
New firms in 2007 LH indicator LH indicator Top quartile LH LH indicator LH indicator Top quartile LH
(1) (2) (3) (4) (5) (6)
Labor hoarding (LH) 0.0683*** 0.0607*** 0.0622*** 0.0222*** 0.0214** 0.0312***
(0.00180) (0.00214) (0.00194) (0.00695) (0.00846) (0.00725)
Women-owned and LH 0.0242*** 0.0229*** 0.00326 0.0207*
(0.00381) (0.00345) (0.0136) (0.0114)
Women-owned − 0.0192*** − 0.0213*** − 0.0158*** − 0.0108*
(0.00198) (0.00210) (0.00555) (0.00580)
Constant 0.778*** 0.785*** 0.778*** 0.823*** 0.827*** 0.825***
(0.000864) (0.00107) (0.00114) (0.00312) (0.00351) (0.00354)
Observations 291,376 291,376 291,376 29,163 29,163 29,163
County fixed effects Yes Yes Yes Yes Yes Yes
2-digit SIC fixed effects Yes Yes Yes Yes Yes Yes
Emp. and sales controls in 07 Yes Yes Yes Yes Yes Yes
Labor hoarding (LH) is a measure of employment in 2010 minus the expected employment in 2010 conditioned on the location
(county), 2-digit SIC industry, and employment and sales in 2004, 2006, and 2007. Columns 1 and 2 use an indicator variable of
labor hoarding while columns 3 and 4 use an indicator variable of whether they are in the top tercile or quartile of labor hoarding.
Vol.: (0123456789)
1388 N. Greenhalgh‑Stanley et al.
Table 14 The effect of labor hoarding during the 2009 Recession on employment growth from 2010 to 2013
OLS diff-in-diff PSM diff-in-diff
Existing firms in 2007 LH indicator LH indicator Top quartile LH LH indicator LH indicator Top quartile LH
(1) (2) (3) (4) (5) (6)
Labor hoarding (LH) − 0.0546*** − 0.00275*** − 0.0553*** − 0.0355*** − 0.0364*** − 0.0367***
(0.000381) (0.000419) (0.000402) (0.000655) (0.000769) (0.000648)
Women-owned and LH 0.00593*** 0.00548*** 0.00269** 0.00305***
(0.000826) (0.000747) (0.00125) (0.00104)
Women-owned − 0.00275*** − 0.00292*** − 0.000975** − 0.00114**
(0.000419) (0.000441) (0.000493) (0.000515)
Constant 0.00966*** 0.0104*** 0.0149*** 0.00957*** 0.00985*** 0.0112***
(0.000179) (0.000215) (0.000228) (0.000272) (0.000309) (0.000312)
Observations 3,015,458 3,015,458 3,015,458 1,331,108 1,331,108 1,331,108
County fixed effects Yes Yes Yes Yes Yes Yes
2-digit SIC fixed effects Yes Yes Yes Yes Yes Yes
Emp. and sales controls Yes Yes Yes Yes Yes Yes
in 04, 06, 07
Labor Hoarding (LH) is a measure of employment in 2010 minus the expected employment in 2010 conditioned on the location
(county), 2-digit SIC industry, and employment and sales in 2004, 2006, and 2007. Columns 1 and 2 use an indicator variable of
labor hoarding, while columns 3 and 4 use an indicator variable of whether they are in the top tercile or quartile of labor hoarding.
Dependent variable is the Δ ln(emp) from 2010 to 2013.
Table 15 The effect of labor hoarding during the 2009 Recession on employment growth from 2010 to 2013
OLS diff-in-diff PSM diff-in-diff
New firms in 2007 LH indicator LH indicator Top quartile LH LH indicator LH indicator Top quartile LH
(1) (2) (3) (4) (5) (6)
Labor hoarding (LH) − 0.00138 0.00588*** − 0.0108*** − 0.0478*** − 0.0519*** − 0.0566***
(0.00168) (0.00200) (0.00192) (0.00584) (0.00712) (0.00566)
Women-owned and LH − 0.0245*** − 0.0195*** 0.0116 0.00798
(0.00348) (0.00337) (0.0111) (0.00871)
Women-owned − 0.00693*** − 0.00725*** − 0.00710* − 0.00739*
(0.00219) (0.00234) (0.00401) (0.00419)
Constant 0.0255*** 0.0279*** 0.0348*** 0.00851*** 0.0105*** 0.0117***
(0.000968) (0.00118) (0.00127) (0.00225) (0.00252) (0.00254)
Observations 264,472 264,472 264,472 23,504 23,504 23,504
County fixed effects Yes Yes Yes Yes Yes Yes
2-digit SIC fixed effects Yes Yes Yes Yes Yes Yes
Emp. and sales controls in 07 Yes Yes Yes Yes Yes Yes
Labor hoarding (LH) is a measure of employment in 2010 minus the expected employment in 2010 conditioned on the location
(county), 2-digit SIC industry, and employment and sales in 2004, 2006, and 2007. Columns 1 and 2 use an indicator variable of
labor hoarding, while columns 3 and 4 use an indicator variable of whether they are in the top tercile or quartile of labor hoarding.
Dependent variable is the Δ ln(emp) from 2010 to 2013.
Vol:. (1234567890)
Do businesses benefit from labor hoarding? 1389
Table 16 The effect of labor hoarding during the 2009 Recession on sales growth from 2010 to 2013
OLS diff-in-diff PSM diff-in-diff
Existing firms in 2007 LH indicator LH indicator Top quartile LH LH indicator LH indicator Top quartile LH
(1) (2) (3) (4) (5) (6)
Labor hoarding (LH) − 0.0424*** − 0.0513*** − 0.0553*** − 0.0305*** − 0.0303*** − 0.0272***
(0.000781) (0.000902) (0.000402) (0.00133) (0.00155) (0.00131)
Women-owned and LH 0.0320*** 0.00548*** − 0.000125 0.00149
(0.00169) (0.000747) (0.00254) (0.00211)
Women-owned − 0.0313*** − 0.00292*** − 0.0226*** − 0.0229***
(0.000851) (0.000441) (0.000995) (0.00104)
Constant 0.0901*** 0.0990*** 0.0149*** 0.0842*** 0.00985*** 0.0112***
(0.000362) (0.000435) (0.000228) (0.000547) (0.000309) (0.000312)
Observations 2,924,809 2,924,809 2,924,809 1,307,411 1,307,411 1,307,411
County fixed effects Yes Yes Yes Yes Yes Yes
2-digit SIC fixed effects Yes Yes Yes Yes Yes Yes
Emp. and sales controls in Yes Yes Yes Yes Yes Yes
04, 06, 07
Labor hoarding (LH) is a measure of employment in 2010 minus the expected employment in 2010 conditioned on the location
(county), 2-digit SIC industry, and employment and sales in 2004, 2006, and 2007. Columns 1 and 2 use an indicator variable of
labor hoarding, while columns 3 and 4 use an indicator variable of whether they are in the top tercile or quartile of labor hoarding.
Dependent variable is the Δ ln(sales) from 2010 to 2013.
Table 17 The effect of labor hoarding during the 2009 Recession on sales growth from 2010 to 2013
OLS diff-in-diff PSM diff-in-diff
New firms in 2007 LH indicator LH indicator Top quartile LH LH indicator LH indicator Top quartile LH
(1) (2) (3) (4) (5) (6)
Labor hoarding (LH) − 0.0169*** − 0.0207*** − 0.0309*** − 0.0671*** − 0.0770*** − 0.0741***
(0.00282) (0.00334) (0.00317) (0.0140) (0.0171) (0.0135)
Women-owned and LH 0.0104* 0.0125** 0.0300 0.0162
(0.00602) (0.00572) (0.0266) (0.0208)
Women-owned − 0.0355*** − 0.0374*** − 0.0467*** − 0.0467***
(0.00359) (0.00384) (0.00957) (0.0100)
Constant 0.116*** 0.127*** 0.133*** 0.117*** 0.130*** 0.131***
(0.00156) (0.00190) (0.00204) (0.00536) (0.00600) (0.00605)
Observations 243,776 243,776 243,776 23,179 23,179 23,179
County fixed effects Yes Yes Yes Yes Yes Yes
2-digit SIC fixed effects Yes Yes Yes Yes Yes Yes
Emp. and sales controls in 07 Yes Yes Yes Yes Yes Yes
Labor hoarding (LH) is a measure of employment in 2010 minus the expected employment in 2010 conditioned on the location
(county), 2-digit SIC industry, and employment and sales in 2004, 2006, and 2007. Columns 1 and 2 use an indicator variable of
labor hoarding, while columns 3 and 4 use an indicator variable of whether they are in the top tercile or quartile of labor hoarding.
Dependent variable is the Δ ln(Sales) from 2010 to 2013.
Vol.: (0123456789)
1390 N. Greenhalgh‑Stanley et al.
Table 18 The effect of downsizing during the 2009 Recession on existing firm survival, and the growth rate of employment and
sales from 2010 to 2013
OLS diff-in-diff
Existing firms in 2007 Survival in 2013 Employment Δ (2013–2010) Sales Δ (2013–2010)
(1) (2) (3) (4) (5) (6)
Downsized − 0.00440*** − 0.00468*** 0.0534*** 0.0566*** 0.0452*** 0.0444***
(0.000533) (0.000631) (0.000493) (0.000581) (0.000999) (0.00118)
Women-owned and down- 0.00123 − 0.0112*** 0.00351
sized
(0.00117) (0.00109) (0.00221)
Women-owned − 0.00819*** 0.000437 − 0.0244***
(0.000478) (0.000397) (0.000808)
Constant 0.855*** 0.858*** − 0.00835*** − 0.00847*** 0.0760*** 0.0829***
(0.000200) (0.000232) (0.000166) (0.000201) (0.000338) (0.000407)
Observations 3,640,604 3,640,604 3,015,458 3,015,458 2,924,809 2,924,809
PSM diff-in-diff
Existing firms in 2007 Survival in 2013 Employment Δ (2013–2010) Sales Δ (2013–2010)
(1) (2) (3) (4) (5) (6)
Downsized − 0.0120*** − 0.0136*** 0.0319*** 0.0325*** 0.0286*** 0.0303***
(0.000847) (0.00102) (0.000689) (0.000832) (0.00139) (0.00167)
Women-owned and down- 0.00549*** − 0.00194 -0.00431
sized
(0.00179) (0.00144) (0.00291)
Women-owned − 0.00688*** − 0.000762 -0.0225***
(0.000707) (0.000485) (0.000979)
Constant 0.829*** 0.831*** 0.00779*** 0.00803*** 0.0827*** 0.0894***
(0.000392) (0.000446) (0.000274) (0.000310) (0.000551) (0.000623)
Observations 1,647,425 1,647,425 1,331,108 1,331,108 1,307,411 1,307,411
County fixed effects Yes Yes Yes Yes Yes Yes
2-digit SIC fixed effects Yes Yes Yes Yes Yes Yes
Emp. and sales controls in Yes Yes Yes Yes Yes Yes
04, 06, 07
Downsized is an indicator that the firm had lower employment in 2010 compared to 2007.
Table 11.
Table 12.
Table 13.
Table 14.
Table 15.
Table 16.
Table 17.
Table 18.
Table 19.
Vol:. (1234567890)
Do businesses benefit from labor hoarding? 1391
Table 19 The effect of downsizing during the 2009 Recession on new firm survival, and the growth rate of employment and sales
from 2010 to 2013
OLS diff-in-diff
New firms in 2007 Survival in 2013 Employment Δ (2013–2010) Sales Δ (2013–2010)
(1) (2) (3) (4) (5) (6)
Downsized − 0.0161*** − 0.0198*** 0.0666*** 0.0711*** 0.0590*** 0.0580***
(0.00159) (0.00191) (0.00187) (0.00225) (0.00313) (0.00373)
Women-owned and downsized 0.0115*** − 0.0145*** 0.00288
(0.00342) (0.00400) (0.00677)
Women-owned − 0.0152*** − 0.0121*** − 0.0322***
(0.00183) (0.00201) (0.00338)
Constant 0.796*** 0.801*** 0.0104*** 0.0142*** 0.0980*** 0.108***
(0.000795) (0.000981) (0.000865) (0.00107) (0.00144) (0.00176)
Observations 348,605 348,605 264,472 264,472 243,776 243,776
PSM diff-in-diff
New firms in 2007 Survival in 2013 Employment Δ (2013–2010) Sales Δ (2013–2010)
(1) (2) (3) (4) (5) (6)
Downsized − 0.00854 − 0.00502 0.0315*** 0.0310*** 0.0114 0.0194
(0.00667) (0.00818) (0.00566) (0.00698) (0.0135) (0.0166)
Women-owned and downsized − 0.00948 0.00167 − 0.0210
(0.0140) (0.0119) (0.0283)
Women-owned − 0.0121** − 0.00714* − 0.0429***
(0.00555) (0.00399) (0.00953)
Constant 0.844*** 0.848*** 0.00149 0.00357* 0.104*** 0.116***
(0.00227) (0.00280) (0.00162) (0.00199) (0.00386) (0.00474)
Observations 29,163 29,163 23,504 23,504 23,179 23,179
County fixed effects Yes Yes Yes Yes Yes Yes
2-digit SIC fixed effects Yes Yes Yes Yes Yes Yes
Emp. and sales controls in 04, Yes Yes Yes Yes Yes Yes
06, 07
Downsized is an indicator that the firm had lower employment in 2010 compared to 2007.
Open Access This article is licensed under a Creative Com- References
mons Attribution 4.0 International License, which permits
use, sharing, adaptation, distribution and reproduction in any Acs, Z., & Armington, C. (2004). Employment growth and
medium or format, as long as you give appropriate credit to the entrepreneurial activity in cities. Regional Studies, 38(8),
original author(s) and the source, provide a link to the Crea- 911–927. https://doi.org/10.1080/0034340042000280938
tive Commons licence, and indicate if changes were made. The Adams, R., & Funk, P. (2012). Beyond the glass ceiling: Does
images or other third party material in this article are included gender matter? Management Science, 58(2), 219–235.
in the article’s Creative Commons licence, unless indicated https://doi.org/10.1287/mnsc.1110.1452
otherwise in a credit line to the material. If material is not Agarwal, R., & Gort, M. (1996). The evolution of markets and
included in the article’s Creative Commons licence and your entry, exit and survival of firms. The Review Of Econom-
intended use is not permitted by statutory regulation or exceeds ics And Statistics, 78(3), 489–498. https://doi.org/10.
the permitted use, you will need to obtain permission directly 2307/2109796
from the copyright holder. To view a copy of this licence, visit Agarwal, R., & Gort, M. (2002). Firm and product life cycles
http://creativecommons.org/licenses/by/4.0/. and firm survival. The American Economic Review,
Vol.: (0123456789)
1392 N. Greenhalgh‑Stanley et al.
92(2), 184–190. https://doi.org/10.5089/9781498305433. Applied Economics, 49(56), 5686–5697. https://doi.org/
001 10.1080/00036846.2017.1332745
Andreoni, J., & Vesterlund, L. (2001). Which is the fair sex? Doms, M., Lewis, E., & Robb, A. (2010). Local labor force
Gender differences in altruism. The Quarterly Journal education, new business characteristics, and firm per-
of Economics, 116(1), 293–312. https://doi.org/10.1162/ formance. Journal of Urban Economics, 67(1), 61–77.
003355301556419 https://doi.org/10.1016/j.jue.2009.10.002
Armington, C., & Acs, Z. J. (2002). The determinants of Duygan-Bump, B., Levkov, A., & Montoriol-Garriga, J.
regional variation in new firm formation. Regional Studies, (2015). Financing constraints and unemployment: Evi-
36(1), 33–45. https://doi.org/10.1080/00343400120099843 dence from the Great Recession. Journal of Monetary
Autor, D., Cho, D., Crane, L. D., Goldar, M., Lutz, B., Economics, 75, 89–105. https://doi.org/10.1016/j.jmone
Montes, J., Peterman, W. B., Ratner, D., Villar, D., & co.2014.12.011
Yildirmaz, A. (2022). The $800 billion paycheck pro- Eagly, A. H., Johannesen-Scmidt, M. C., & van Egen, M. L.
tection program: Where did the money go and why did (2003). Transformational, transactional, and laissez-faire
it go there? Journal of Economic Perspectives, 36(2), leadership styles: A meta-analysis comparing women and
55–80. https://doi.org/10.1257/jep.36.2.55 men. Psychological Bulletin, 129(4), 569–591. https://doi.
Backman, M., & Kohlhase, J. E. (2020). Labor force diversity org/10.1037/0033-2909.129.4.569
and firm survival. Journal Of Regional Science, 60(5), Eagly, A. H., & Carli, L. L. (2007). Through the labyrinth: The
903–928. https://doi.org/10.1080/00343400120099843 truth about how women become leaders. Harvard Busi-
Bartoloni, E., Arrighetti, A., & Landini, F. (2021). Recession ness School Press.
and firm survival: Is selection based on cleansing or skill Fairlie, R. W., & Robb, A. M. (2009). Gender differences
accumulation? Small Business Economics, 57, 1893– in business performance: Evidence from the charac-
1914. https://doi.org/10.1007/s11187-020-00378-0 teristics of business owners survey. Small Business
Barattieri, A., Basu, S., & Gottschalk, P. (2014). Some evi- Economics, 33(4), 375–395. https://doi.org/10.1007/
dence on the importance of sticky wages. American Eco- s11187-009-9207-5
nomic Journal: Macroeconomics, 6(1), 70–101. https:// Fairlie, R. W. (2013). Entrepreneurship, economic conditions,
doi.org/10.1257/mac.6.1.70 and the great recession. Journal of Economics & Manage-
Bewley, T. F. (2002). Fairness, reciprocity, and wage rigid- ment Strategy, 22(2), 207–231. https://doi.org/10.1111/
ity. Cowles Foundation Discussion Papers. 1648. https:// jems.12017
elischolar.library.yale.edu/cowles-discussion-paper-series/ Faulkender, M., Jackman, R., & Miran, S. (2020). The Job-
1648. preservation effects of paycheck protection program loans.
Campbell, C. M. (1993). Do firms pay efficiency wages? Evi- The Department of the Treasury Office of Economic Pol-
dence with data at the firm level. Journal of Labor Eco- icy Working Paper 2020–01. https://home.treasury.gov/
nomics. https://doi.org/10.1086/298303 system/files/226/Job-Preservation-Effects-Paycheck-Prote
Carr, D. (1996). Two paths to self-employment? Women’s and ction-Program-Loans.pdf. Accessed Feb 2022.
men’s self-employment in the United States, 1980. Work Fedotenkov, I., Kvedaras, V., & Sánchez-Martínez, M. (2024).
and Occupations, 23, 26–53. https://doi.org/10.1177/ Employment protection and labour productivity growth
0730888496023001003 in the EU: Skill-specific effects during and after the great
Chatterjee, I., Shephard, D., & Wincent, J. (2022). Wom- recession. Empirica, 51(1), 209–262. https://doi.org/10.
en’s entrepreneurship and well-being at the base of the 1007/s10663-023-09585-w
pyramid. Journal of Business Venturing, 37(4), Article Flippin, A. (2022) Gender differences in risk attitudes. IZA
106222. https://doi.org/10.1016/j.jbusvent.2022.106222 World of Labor: 100. https://doi.org/10.15185/izawol.100.
Clementi, G. L., & Palazzo, B. (2016). Entry, exit, firm dynam- v2.
ics, and aggregate fluctuations. American Economic Jour- Glaeser, Edward, L., Saiz, A. (2003). “The rise of the skilled
nal: Macroeconomics, 8(3), 1–41. https://doi.org/10.1257/ city.” NBER Working Paper No. 10191, https://doi.org/10.
mac.20150017 3386/w10191.
Coleman, S., & Robb, A. (n.d.). The next wave: Financing Glaeser, E. L., Kallal, H. D., Scheinkman, J. A., & Shleifer, A.
women’s growth-oriented firms. Stanford University Press. (1992). Growth in cities. Journal of Political Economy,
Croson, R., & Gneezy, U. (2009). Gender differences in prefer- 100, 1126–1152. https://doi.org/10.1086/261856
ences. Journal Of Economic Literature, 47(2), 448–474. Glaeser, E., Rosenthal, S., & Strange, W. (2010). Urban eco-
https://doi.org/10.1257/jel.47.2.448 nomics and entrepreneurship. Journal Of Urban Econom-
Cohen, P. N., & Huffman, M. L. (2007). Working for the ics, 67, 1–14. https://doi.org/10.1016/j.jue.2009.10.005
woman? Female managers and the gender wage gap. Gonalons-Pons, P., & Gangl, M. (2022). Regulated earnings
American Sociological Review, 72(5), 681–704. https:// security: The relationship between employment protection
doi.org/10.1177/000312240707200502 and unemployment scarring over the Great Recession.
Darden, B., Kellermanns, F., & Best, E. (2022). Entrepreneur- Socio-Economic Review, 20(1), 169–198. https://doi.org/
ial fear of failure: Transgender versus cis-gender entrepre- 10.1093/ser/mwaa049
neurs. Journal of Business Venturing, 18, Article e00332. Holden, S., & Tilahun, M. (2022). Are risk preferences explain-
https://doi.org/10.1016/j.jbvi.2022.e00332 ing gender differences in investment behavior? Journal of
Deller, S., Convoy, T., & Watson, P. (2017). “Women business Behavioral and Experimental Economics, 101, Article
owners: A source of stability during the great recession? 101949. https://doi.org/10.1016/j.socec.2022.101949
Vol:. (1234567890)
Do businesses benefit from labor hoarding? 1393
Iverson, R. D., & Zatrick, C. (2011). The effects of downsizing Of Economics And Statistics, 94(3), 764–788. https://doi.
on labor productivity: The value of showing consideration org/10.1162/REST_a_00193
for employees’ morale and welfare in high-performance Rosenthal, S. S., & Strange, W. C. (2003). Geography, indus-
work systems. Human Resource Management, 50(1), trial organization, and agglomeration. Review Of Econom-
29–44. https://doi.org/10.1002/hrm.20407 ics And Statistics, 85, 377–393. https://doi.org/10.1162/
Jaimovich, N., & Siu, H. E. (2020). Job polarization and job- 003465303765299882
less recoveries. The Review Of Economics And Statistics, Rosenthal, Stuart, & Strange, Will. (2009). Small establishments/
102(1), 129–147. https://doi.org/10.1162/rest_a_00875 big effects: Agglomeration, industrial organization, and
Kahn, L. B. (2010). The long-term labor market consequences entrepreneurship. In Z. J. Acs & D. B. Audrestsch (Eds.),
of graduating from college in a bad economy. Labour Handbook of entrepreneurship research: An interdiscipli-
Economics, 17(2), 303–316. https://doi.org/10.1016/j. nary survey and introduction (pp. 277–302). Springer.
labeco.2009.09.002 Shane, S. (2011). The great recession’s effect on entrepreneur-
Mano-Negrin, R., & Sheaffer, Z. (2004). Are women ‘cooler’ ship. Federal Reserve Bank of Cleveland Economic Com-
than men during crises? Exploring gender differences in mentary. https://doi.org/10.26509/frbc-ec-201104
perceiving organisational crisis preparedness proneness. Silverman, I. W. (2003). Gender differences in delay of grati-
Women In Management Review, 19(2), 109–122. https:// fication: A meta-analysis. Sex Rote, 49(9–10), 451–463.
doi.org/10.1108/09649420410525315 https://doi.org/10.1023/A:1025872421115
Matsa, D., & Miller, A. (2014). Workforce reductions at Smith, N., Smith, V., & Verner, M. (2006). Do women in top
women-owned businesses in the United States. ILR management affect firm performance? A panel study of
Review, 67(2), 422–452. https://doi.org/10.1177/00197 2,500 Danish firms. International Journal of Productivity
9391406700206 and Performance Management, 55(7), 569–593. https://
Oreopoulos, P., von Wachter, T., & Heisz, A. (2012). The doi.org/10.1108/17410400610702160
short- and long-term career effects of graduating in a Stephens, H., Patridge, M., & Faggian, A. (2013). Innovation,
recession. American Economic Journal: Applied Eco- entrepreneurship and economic growth in lagging regions.
nomics, 4(1), 1–29. https://doi.org/10.1257/app.4.1.1 Journal Of Regional Science, 53(5), 778–812. https://doi.
Phillips, B. D., & Kirchoff, B. A. (1989). Formation, growth org/10.1111/jors.12019
and survival; Small firm dynamics in the U.S. economy. Sullivan, D., & von Wachter, T. (2009). Job displacement and
Small Business Economics, 1(1), 65–74. mortality: An analysis using administrative data. The
Renzulli, L. A., Aldrich, H., & Moody, J. (2000). Family matters: Quarterly Journal of Economics, 124(3), 1265–1306.
Gender, networks, and entrepreneurial outcomes. Social https://doi.org/10.1162/qjec.2009.124.3.1265
Forces, 79(2), 523–546. https://doi.org/10.1093/sf/79.2.523 Thebaud, S., & Sharkey, A. (2016). Unequal hard times: The
Robb, A. M. (2002). Entrepreneurial performance by women and influence of the great recession on gender bias in entrepre-
minorities: The case of new firms. Journal of Developmen- neurial financing. Sociological Science, 3, 1–31. https://
tal Entrepreneurship, 7(4), 383. doi.org/10.15195/v3.a1
Robb, A. M., & Watson, J. (2010). Comparing the performance Yu, W., Fei, J., Peng, G., & Bort, J. (2024). When a crisis hits:
of female-and male-controlled SMEs: Evidence from the An examination of the impact of the global financial cri-
United States and Australia. Frontiers of Entrepreneurship sis and the COVID-19 pandemic on financing for women
Research, 30(8), Article Article 1. entrepreneurs. Journal of Business Venturing. https://doi.
Robb, A. M., & Watson, J. (2012). Gender differences in firm org/10.1016/j.jbusvent.2024.106379
performance: Evidence from new ventures in the United
States. Journal of Business Venturing, 27, 544–558.
https://doi.org/10.1016/j.jbusvent.2011.10.002 Publisher’s Note Springer Nature remains neutral with regard
Rosenthal, S. S., & Strange, W. (2012). Female entrepreneur- to jurisdictional claims in published maps and institutional
ship, agglomeration, and a new spatial mismatch. Review affiliations.
Vol.: (0123456789)
File and source
- File
- do-businesses-benefit-from-labor-hoarding.pdf
- Size
- 711,572 bytes
- SHA-256
- 41e16069e172df00d3630d47deabc63980bd5fb8f291a9c6e38d9224478c2e65
- Original
- doi.org