«Working Paper = = Unionization, Cash, and Leverage Martin C. Schmalz Stephen M. Ross School of Business University of Michigan Ross School of ...»
Unionization, Cash, and Leverage
Martin C. Schmalz
Stephen M. Ross School of Business
University of Michigan
Ross School of Business Working Paper
Working Paper No. 1215
This work cannot be used without the author's permission.
This paper can be downloaded without charge from the
Social Sciences Research Network Electronic Paper Collection:
ÜííéWLLëëêåKÅçãL~Äëíê~ÅíZOORQMOR=rkfsbopfqv=lc=jf`efd^k= Unionization, Cash, and Leverage Martin C. Schmalz∗ Abstract What is the eﬀect of unionization on corporate ﬁnancial policies? The average unionized ﬁrm responds with lower cash and higher leverage to a unionization election than the average ﬁrm escaping unionization. However, using a regression discontinuity design I ﬁnd that the causal eﬀect of unionization is close to zero on average, but heterogeneous across ﬁrms. For the subset of large and ﬁnancially unconstrained ﬁrms, the causal eﬀect is positive on leverage and negative on cash; the opposite is true for small and ﬁnancially constrained ﬁrms. These results help reconcile controversially discussed views on how corporate ﬁnance and labor interact.
JEL Classiﬁcation: J50, G32 Keywords: Unionization, cash, leverage, capital structure, labor adjustment costs ∗ email@example.com. Stephen M. Ross School of Business, University of Michigan. This paper is based on the ﬁrst chapter of my dissertation at Princeton University. I am deeply grateful to my adviser Markus Brunnermeier, as well as to David Sraer, Alexandre Mas, and David Lee for invaluable support, ideas, and advice, and to David Lee and Alexandre Mas for generously providing data. For helpful discussions and comments, I would like to thank Sergey Zhuk and other Princeton classmates, John DiNardo, Henry Farber, Mireia Giné, Bo Honoré, Oleg Itskhoki, Martin Kanz, Hyunseob Kim, David Lee, Thomas Mertens, Roni Michaely, Benjamin Moll, Ulrich Müller, José Scheinkman, Matthew Serﬂing, Hyun Song Shin, Toni Whited, as well as seminar participants at Bocconi, Boston University, Columbia GSB, Harvard Business School, HSE Moscow, IESE, INSEAD, London School of Economics, University of Michigan (Finance; Labor Economics), NBIM, Ohio State University, Princeton, Rochester, SMU, UC Irvine, University of Vienna, USC, and USI Lugano. Laurien Gilbert and Alexis Furuichi provided excellent research assistance. I gratefully acknowledge generous ﬁnancial support through a Fellowship of Woodrow Wilson Scholars and an NTT Fellowship from the Mitsui Life Financial Center.
1 Introduction Two polar views shape the debate on the interaction between labor and corporate ﬁnance.
According to one view, ﬁrms strategically choose their ﬁnancial policies to attain a better bargaining position in future wage negotiations with their employees. In particular, according to this “bargaining view,” ﬁrms respond with higher leverage and lower cash to unionization or the threat thereof [Bronars and Deere, 1991, Matsa, 2010]; see also Baldwin , Dasgupta and Sengupta , Perotti and Spier , Sarig . The intuition is simple: unionized workers have a decreased ability to extract rents when the ﬁrm has low cash reserves and already committed future cash ﬂows to debt holders. Therefore, reducing ﬁnancial ﬂexibility can reduce workers’ incentive to unionize.
The alternative view holds that an increase in fundamental or operating risk arising from more rigid labor inputs leads to an increased need for ﬁnancial ﬂexibility [Mauer and Triantis, 1994, Gamba and Triantis, 2008, Pratt, 2011, Kahl et al., 2013, Chen et al., 2013].1 Unionization is one reason why an increase in labor rigidity can occur. Therefore, according to the “ﬁnancial ﬂexibility” view, unionization should cause an accommodating response with more cash and less leverage.
While often discussed controversially, there is no contradiction between these forces. In theory, both can be present at the same time in the same ﬁrm, whereas it is an empirical question which one is quantitatively more important in the average ﬁrm, or in particular subsets of ﬁrms. Also, it is conceivable that the “bargaining” motive is more important ex ante (before a unionization attempt), whereas the “ﬁnancial ﬂexibility” motive is more important ex post (in case actual unionization occurred).2 Lastly, it is important to understand if the response of ﬁrms becoming unionized is the This prediction is also consistent also with models emphasizing the eﬀect of asset-market frictions on liquidity risk management [Bolton et al., 2011]. Note that human capital is an asset without market liquidity from the perspective of the ﬁrm. A similar prediction arises from ﬁrms’ internalization of expected costs of bankruptcy born by their employees [Titman, 1984, Jaggia and Thakor, 1994, Berk et al., 2010]: implicit promises of wage insurance are only credible when the ﬁrm has the ﬁnancial reserves to honor them. Agrawal and Matsa , Brown and Matsa , Kim , Chemmanur et al.  provide empirical support for variations of the insurance hypothesis, see also Caggese and Cuñat , Sraer and Thesmar , Shivdasani and Stefanescu , Ellul et al. .
The eﬀect could also be non-monotonic: too much ﬁnancial ﬂexility can hurt a ﬁrm’s bargaining position, whereas too little ﬁnancial ﬂexility can reduce shareholder value because of ﬁnancial distress costs. The empirical question is what the marginal eﬀect of unionization is on the optimal level of ﬁnancial ﬂexibility.
eﬀect of unionization per se (i.e., the eﬀect of changes in labor rigidity as well as how bargaining is conducted as a result of unionization), or if the response is driven by potentially unobservable variables that are correlated with more or less strong support for unionization, such as employee morale. In other words, what is the causal eﬀect of unionization on corporate ﬁnancial policies?
As is evident from an active empirical literature on the eﬀect of unionization (and labor rigidity more generally) on corporate ﬁnancial choices, there is much interest in these questions. Indeed, approaching answers to them is of ﬁrst-order importance for advancing our understanding of which theories are most important in describing corporate ﬁnancial decision making.
To answer these questions, this paper employs a data set that contains all unionization elections in U.S. establishments belonging to Compustat ﬁrms between 1961 and 1999. The data set and research design I use was ﬁrst developed by DiNardo and Lee , and extended by Lee and Mas  to study the causal eﬀect of unionization on productivity and wages as well as on stock returns, respectively. In contrast to earlier papers in the corporate ﬁnance literature that have used variation in the ability to form unions at the state or industry levels, I directly examine ﬁrms experiencing a unionization attempt. The research design also allows me to determine the causal eﬀect of unionization on ﬁrm policies.3 I ﬁnd that the average ﬁrm experiencing a unionization election reduces its cash-to-asset ratio and increases leverage. This observation is consistent with the bargaining view, but also with the notion that ﬁrms in bad ﬁnancial condition impose more unemployment risk on employees, and employees facing higher unemployment risk have higher incentives to (attempt to) unionize in order to enjoy better protection of their interests in bankruptcy. Next, ﬁrms in which more than 50% of workers voted for unionization (which almost surely leads to unionization of the establishment) respond with 2% lower cash-to-asset ratios and +0.8% higher leverage to an election, compared to ﬁrms in which less than 50% voted for unionization (which almost never leads to unionization). This The data go back much further in time than other papers using the election design to investigate the unionization eﬀect on other outcome variables, such as innovation [Bradley et al., 2013] or staﬃng levels and quality [Sojourner et al., 2012].
result is consistent with the predictions of the bargaining view. Importantly, however, this analysis ignores the information content of the precise share of the vote cast in favor of unionization, i.e., it gives equal weight to unionization events in which 100% of workers voted for unionization and events in which only 50.1% voted for unionization.
Because the degree of union support, reﬂected in the vote share for unionization, may be correlated with unobserved characteristics of the ﬁrm or the workforce, estimating the causal eﬀect of unionization requires comparing observations in which only unionization but none of the other characteristics changed signiﬁcantly. The discontinuity in unionization outcomes at the 50% vote share threshold aﬀords such a comparison. I ﬁnd that the ﬁnancial policy response of close “winners” (from the perspective of the union) and close “losers” is almost identical, both in terms of cash and in terms of leverage. Thus, the causal eﬀect of unionization per se is zero; the average eﬀect is entirely driven by the tails of the vote share distribution. From this result alone, one can either conclude that the “bargaining” and “ﬂexibility” eﬀects exactly cancel out for the subset of ﬁrms at the discontinuity, or that unionization per se is not a ﬁrst-order driver of corporate ﬁnancial policy.
To try and tell apart these interpretations, I analyze the average and causal eﬀects of unionization in subsamples of the data. If there is truly no causal eﬀect of unionization, that should be true in subsamples of the data as well. By contrast, if both mechanisms are equally important for the average marginal ﬁrm, perhaps one is more important than the other in subsets of ﬁrms. To investigate this possibility, I split the sample by the degree to which the ﬁrms can freely adjust their ﬁnancial policy at the time of the unionization election, as measured by ranking below (unconstrained) or above (constrained) the median of the Whited and Wu  index of ﬁnancial constraints. A ﬁrm presently unconstrained can become constrained in the future, and therefore has a motive for risk management; according to the ﬂexibility view, it should increase cash and decrease leverage in response to an increase of labor rigidity. By contrast, a ﬁrm presently constrained faces a tradeoﬀ between two exclusive uses of the existing levels of cash and debt capacity: using (and thus reducing) it now for investment, or keeping it for the future [Rampini et al., 2014]. Thus, a presently constrained ﬁrm will, if anything, reduce its ﬁnancial ﬂexibility in response to an increase of labor rigidity, but deﬁnitely not increase it (because it can’t). By contrast to these predictions, the simplest version of the bargaining view prescribes a tightening of ﬁnancial ﬂexibility for both types of ﬁrms.
Using the regression discontinuity design, I ﬁnd that unionization causes a 6.2% increase of cash-to-assets ratios among ﬁnancially unconstrained ﬁrms. This eﬀect, corresponding to a 0.43percentage-point increase from 6.86% to 7.29%, or $23 million for the average ﬁrm, is not only highly statistically signiﬁcant but also economically sizable: the increase corresponds to about 13% of the average ﬁrm’s wage bill. I also ﬁnd a highly statistically signiﬁcant 2.4% reduction of on-balancesheet ﬁnancial leverage, which corresponds to a 1-percentage-point decrease of leverage from 42.8% to 41.8%, or a $63 million reduction in debt ﬁnancing. By contrast, ﬁnancially constrained ﬁrms reduce cash and increase leverage, although these estimates are less precise. I view these results as most easily compatible with the view that ﬁnancial ﬂexibility is a relatively more important consideration for ﬁrms that can actively and freely adjust their ﬁnancial policies. (Indeed, ﬁnancial ﬂexibility is considered the most important driver of corporate ﬁnancial decisions in general [Graham and Harvey, 2001].) This interpretation is also corroborated by the ﬁnding that the average eﬀect of unionization reported above (a decrease of cash and increase in leverage) is entirely driven by the subsample of ﬁnancially constrained ﬁrms, which presumably cannot freely adjust their ﬁnancial policies.
An alternative interpretation is that bargaining considerations are relatively more important in ﬁrms that are already ﬁnancially constrained, whereas they are less important compared to ﬁnancial ﬂexibility in less ﬁnancially constrained ﬁrms. Given the ﬁnding by Benmelech et al.  of a link between ﬁnancial distress and wage concessions, this interpretation seems plausible; however, a diﬀerent explanation would need to address why bargaining considerations are unimportant in ﬁnancially unconstrained ﬁrms.4 Under either interpretation, the RD results indicate that unionGiven that the measure of ﬁnancial constraints I use is highly negatively correlated with ﬁrm size, the interpretation could also be that the bargaining motive is more important for small ﬁrms and ﬁnancial ﬂexibility is a more important concern for larger ﬁrms, e.g. because an increased cost in quasi-ﬁxed [Oi, 1962] labor costs has the potential to spill over to other plans. Also, there is an ongoing debate on how well measures of ﬁnancial constraints ization per se is indeed an important determinant of corporate ﬁnancial policies, but the eﬀect is heterogeneous across ﬁrms.