Zero leverage puzzle: Do labour laws matter?

• Published date: 01 September 2023
• Author: Hamid Boustanifar,Arnt Verriest
• Date: 01 September 2023
• DOI: http://doi.org/10.1111/eufm.12385

DOI: 10.1111/eufm.12385

ORIGINAL ARTICLE

Zero leverage puzzle: Do labour laws matter?

Hamid Boustanifar

1

|Arnt Verriest

2

1

EDHEC Business School, Nice, France

2

Faculty of Business and Economics, KU

Leuven, Leuven, Belgium

Correspondence

Hamid Boustanifar, EDHEC Business

School, Nice, France.

Email: hamid.boustanifar@edhec.edu

Abstract

Exploiting the staggered passage of labour protection

laws in the United States, we find that higher labour

adjustment costs increased the likelihood of observing

zero leverage firms by 22%. This effect is significantly

larger in states with stronger unionization, in indus-

tries with higher volatility and concentration, and in

firms with higher labour intensity. Both within‐firm

changes in debt policies and higher propensity of

newer firms to be debt‐free are important in explaining

these patterns. Overall, our work contributes to the

literature on the relation between financial and labour

markets by highlighting the role of labour laws in

explaining the zero‐leverage puzzle.

KEYWORDS

employee protection, labour protection, net debt, zero leverage

JEL CLASSIFICATION

G32, K31

1|INTRODUCTION

A growing literature studies the interaction between financial and labour markets (Benmelech

et al., 2021; Bos et al., 2018; Boustanifar et al., 2017; Boustanifar, 2014; Chodorow‐Reich, 2013;

Ellul & Pagano, 2019; Geng et al., 2022; Michaels et al., 2018). Our study contributes to this

literature by showing that the rigidity of labour laws has strong power in explaining why a

significant portion of firms adopt very conservative financial policies.

Prior literature has documented that a substantial fraction of firms around the world are

run with zero or almost zero leverage (more than 20% in the United States, for example), a

Eur Financ Manag. 2023;29:1119–1159. wileyonlinelibrary.com/journal/eufm © 2022 John Wiley & Sons Ltd.

|

1119

EUROPEAN

FINANCIAL MANAGEMENT

The authors would like to thank the Editor, John Doukas, and one anonymous referee for very helpful comments

during the review process.

phenomenon called zero‐leverage puzzle. This puzzling fact has been shown by Strebulaev and

Yang (2013) for the United States, Dang (2013) for the United Kingdom, Huang et al. (2017) for

China, and Bessler et al. (2013) and El Ghoul et al. (2018) for a large sample of countries around

the world. This phenomenon is typically referred to as a puzzle since, as argued in Strebulaev

and Yang (2013), almost no model of capital structure predicts such a high proportion of firms

with zero or almost zero leverage.

2

The extant literature trying to explain this phenomenon has

highlighted the role of CEO ownership and family firms (Strebulaev & Yang, 2013),

macroeconomic conditions and financial constraints (Dang, 2013), credit constraints (Devos

et al., 2012), and the real option value of having debt (Lotfaliei, 2018). Differences in the degree

of zero leverage phenomenon across countries have been shown to be related to culture and

trust (El Ghoul et al., 2018), creditor protection, legal origin and dividend taxation (Bessler

et al., 2013).

Our hypothesis is that higher labour protection leads to a higher tendency of firms to adopt

zero‐leverage policies. Specifically, in an economy where firms face a high cost of firing

employees, labour expenses are unavoidable in the short run as firms will not be able to

discharge employees at a low cost in case of a temporary economic downturn

(Boustanifar, 2014; Hamermesh & Pfann, 1996; Oi, 1962). In such an economy, firms will

have higher operating leverage since labour cost is not only a variable cost but rather has a fixed

cost component. Higher operating leverage makes firms riskier and as a result firms should

respond with adopting conservative financial policies (Aiyagari, 1994). Put differently, an

increase in operating leverage induces the requirement for more financial flexibility, to offset

the increased risk and to reduce the probability and costs of business termination or

bankruptcy. Therefore, we predict that increases in employee protection should result in a

higher likelihood of observing zero or almost zero leverage among firms.

Empirically, investigating the impact of labour protection laws on adopting zero leverage

policies is challenging due to endogeneity and omitted variable issues. For example, any cross‐

sectional correlation between labour protection and the zero‐leverage phenomenon would be

subject to criticism that regions with high labour protection are fundamentally different from

those with low labour protection and the differences in tendency for adopting zero leverage

may be the result of those (unobservable) characteristics and not labour protection laws

themselves. Using changes in labour protection laws within a country would also be subject to

the fact that the estimated effect cannot isolate the effect of any other macro shocks that affect

all firms in the economy (access to finance, growth prospects, etc.) around the change in

labour laws.

To overcome the empirical issues mentioned above, we exploit the introduction of the good

faith exception, one particular law in the Wrongful Discharge Laws (WDLs), which constitutes

a drastic change in labour protection across US states. The good faith exception prevents

employers from firing people out of bad faith, malevolence or retaliation. This exception has

been interpreted quite broadly and its adoption is often considered as the strongest increase in

employee protection, compared to other protection laws such as implied contract exception and

the public policy exception (Autor et al., 2007).

3

An important feature of these laws was that

they happened in a staggered way across the states. The fact that different states adopted good

faith exemption laws in different years provides us with an excellent laboratory to test the effect

2

Baker et al. (2020) model rational leverage choices in the presence of mispricing risk. Within this (behavioural)

framework, some firms rationally choose very low or very high leverage.

3

We provide detailed descriptions of different employment protection laws in the United States in Section 2.

1120

|

EUROPEAN

FINANCIAL MANAGEMENT

BOUSTANIFAR AND VERRIEST

of labour protection laws on the likelihood of adopting zero leverage policies by firms in a

Difference‐in‐Differences framework. In other words, we can estimate the change in probability

of firms adopting zero leverage policies after the law within the same state, controlling for

similar change in states that did not pass the law. The treatment (control) group includes firms

that are located in states that have adopted the good faith exception (are located in non‐

adopting states). Of course, an important assumption behind our analysis is that the regulatory

changes across states were exogenous to the zero leverage phenomenon, in presence of our

control variables. We provide multiple tests and analyses, and virtually all of them support this

assumption. Most of our analyses are at the firm level and robust to inclusion of industry, year,

state and firm level fixed effects.

Our main findings indicate that the good faith exception is positively associated with the

incidence of having zero leverage and negative net debt (hereafter negative debt or ND). These

effects are not only statistically significant but also economically meaningful. Specifically, our

results suggest that the introduction of the good faith regulation increases the probability of

observing zero leverage firms from 11.5% to 14.7%, or by 22%. In addition, and importantly, we

find that the documented positive link between higher employee protection and lower leverage

in prior literature is predominantly driven by the zero leverage and negative debt firms. More

precisely, after controlling for these firms, employment protection has no (statistically or

economically) significant effect on firm leverage.

If as we predict, labour laws affect the decision to be zero levered through the labour

adjustment channel, we expect predictable cross‐sectional variation in the proportion of zero

leverage firms across the population. First, our results should be stronger in states with low

unionization rates since firms in highly unionized states are most likely to face higher firing

costs even in absence of the Good faith laws. Second, we expect a larger effect in industries with

higher volatility and in those with more intense competition since firms in these industries are

the ones that are more likely to face negative shocks and need more flexibility. Finally, since

our argument is based on labour adjustment costs, we expect that our effects should be focused

mostly on firms with relatively high labour intensity. We show that the effect of labour

protection laws on zero leverage varies across states, industries, and firms in a theoretically

predictable way that are consistent with our main arguments. These results increase our

confidence that the association we document is indeed the result of labour protection laws and

is unlikely to be due to other confounding factors. In other words, any alternative explanation

should also be consistent with these heterogeneous effect of labour laws on different states,

industries and firms.

Finally, we conduct a battery of additional analyses to investigate the robustness of our

results as well as to better understand the mechanism behind the effect we document.

Specifically, we show that the effect on labour laws on zero leverage policies is not driven by

one influential state. Moreover, the statistical significance and economic magnitude of the

effect barely change when we use alternative regression models such as probit or linear

probability models, or when we run regressions at the state‐level. We also show that the results

are robust when we use the full set of industry‐year fixed effects to control for any unobserved

time‐varying industry‐specific factor that could be correlated with both our measure of labour

law and zero leverage policies. Finally, we find that within‐firm changes in policies in favour of

zero leverage and negative debt (i.e., when we add firm fixed effects) are responsible for

between 30% and 50% of the overall effect we document.

Our paper contributes to several strands of literature. First, our findings help to further

understand the zero‐leverage puzzle and its determinants. Previous papers have highlighted the

BOUSTANIFAR AND VERRIEST EUROPEAN

FINANCIAL MANAGEMENT

|

1121