Research Briefs in Economic Policy No. 103

Bankruptcy and the Cost of Organized Labor: Evidence from Union Elections

By Murillo Campello, Janet Gao, Jiaping Qiu, & Yue Zhang
March 7, 2018

Despite their declining prominence, labor unions still shape workers’ participation in corporate activity. Over eight million private-sector workers in the United States today are represented by unions, and of the largest 100 industrial firms, 33 have a unionized work force. Unions are known to use collective bargaining power to protect workers’ interests such as wages, health care, and job security, but less is known about the role they play in bankruptcy. At the time when workers’ investment in firm-specific human capital is most threatened, the U.S. Bankruptcy Code safeguards only wages and benefits for work already performed. To protect their members’ long-term interests, unions must become active parties in bankruptcy states.

Unions are able to protect their members’ interests in several ways in bankruptcy, and our analysis shows that worker unionization bears significant wealth consequences for other stakeholders of the firm. As recognized creditors, for example, unionized workers may be eligible to hold seats on unsecured creditors’ committees under Chapter 11 of the U.S. Bankruptcy Code. Those committees are commonly favored by the courts and have broad powers to (1) formulate reorganization plans, (2) request the replacement of managers, (3) block asset sales, and (4) move to convert the case to Chapter 7. Nonunionized workers with separate, small claims are not eligible to hold seats on creditors’ committees.

Beyond receiving debtor-like recognition under Chapter 11, unions resort to other tactics to empower workers in bankruptcy. They organize strikes, boycotts, and public denouncements with the goal of forcing managers to acquiesce to their demands, so as to avoid disruptions that invite creditor control. When convenient, unions use their leverage in court so that bankruptcy proceedings allow for disruption of absolute priority rules, whereby unsecured creditors’ claims lose seniority. Unions can also make bankruptcies last longer, using the courts to force parties into repeated, costly negotiations over workers’ demands. In securing continued employment for their members, unions often favor inefficient reorganizations in lieu of liquidation. This is a key concern because firms that emerge from reorganization often reenter bankruptcy, as unions resist asset sales and worker layoffs.

We study the impact of worker unionization on corporate creditors by looking at the price reactions of publicly traded bonds to union elections. Bond prices represent a unique value metric with which to gauge the impact of unionization onto financial stakeholders of the firm. Unlike other creditors 2 (e.g., banks and syndicated lenders), investors of diffusely held bonds find it difficult to renegotiate with borrowers. Bond investors, instead, dispose of their securities in the market in response to innovations to the expected value of their claims. Given the structure of bond payoffs (capped at the issue face values in nonbankruptcy states), bond prices are sensitive to expected losses in bankruptcy states. In particular, as their claims are senior, yet unsecured, bondholders’ expected wealth declines sharply in the face of high bankruptcy costs. Deviations from an orderly bankruptcy process will increase expected bankruptcy costs and lead to declines in the secondary market price of corporate bonds.

Union elections are conducted through secret ballot voting. Once a union wins over 50 percent of the workers’ votes, it attains legal recognition. Union rights are protected by the National Labor Relations Act, and a successful election significantly increases the bargaining power of workers. Naturally, both the occurrence and the results of union elections are influenced by a number of factors. As such, the average union-win firm might differ from its average union-loss counterpart on several dimensions (both observable and unobservable).

To identify these effects, we resort to a design that exploits local variation in the vote share of elections that can lead to discrete shifts in union legal status. In short, our tests contrast bond price reactions to closely won union elections with bond price reactions to closely lost union elections. Workers in closewin elections gain legal representation status while those in close-loss elections do not; yet firm characteristics and workers’ support for unions are ex-ante similar across the two groups. Given the fairly secretive nature (and fast speed) of the union voting process in the United States, it is unlikely for individuals or firms to precisely anticipate or manipulate the outcome of close union elections. Under these conditions, relative differences in bond price reactions to close union election results can be plausibly attributed to the effect of unionization.

We conduct our analysis on a sample of 721 bond issuers witnessing worker unionization attempts between 1977 and 2010 using records from the National Labor Relations Board. In short, our tests show that worker unionization negatively affects the wealth of senior, unsecured creditors.

From a pricing perspective, the decline in bond values that we report could arise from increases in default risk or from bankruptcy costs. We next look for evidence of those effects in our data. We find no evidence that close union winners perform worse, become more likely to enter distress, or are more likely to file for bankruptcy than close union losers for several years after the vote.

We then set out to investigate the effects of unionization on bankruptcy costs. This is a difficult task and our analysis is limited by the fact that we have to focus on explicit (observable) bankruptcy costs. The examination necessitates data from actual bankruptcy events and we first expand our dataset to include information from the UCLA-LoPucki bankruptcy database. In this investigation, we compare the duration, costs, and outcomes of court proceedings across bankrupt firms with unionized workers and those without. We find that unionized firms experience more prolonged court proceedings and are also more likely to go through inefficient reorganizations, as evidenced by a higher likelihood of emerging from bankruptcy and refiling for bankruptcy shortly thereafter. Unionized firms are also more likely to reorganize under debtor-in-possession financing, which is often highly detrimental to preexisting bondholders. In addition, firms with labor unions incur significantly higher expenses and fees paid in bankruptcy court. The results we report are consistent with the notion that unionization is associated with higher in-court bankruptcy costs. Admittedly, these tests may allow for a noncausal interpretation.

We thus set out to more granularly identify the welfare costs of labor unions in bankruptcy court by exploiting statutory variation in the number of seats assigned to unions on unsecured creditors’ committees (UCCs). Section 1102(a) of the Bankruptcy Code charges the U.S. Trustee with the duty of organizing a committee composed of the largest unsecured creditors of the bankrupt firm (including both unionized workers and bondholders). Following this guideline, the Trustee shall assign union representatives to seats on UCCs if they represent labor claims whose amount ranks among the largest liabilities of the firm. It is difficult to ascertain and calculate the claims of various corporate creditors, and as a result the number of UCC seats eventually assigned to unions-seats that may come at the expense of other unsecured creditors-has a considerable degree of variation. We use this source of variation to gauge the marginal effect of unions’ empowerment in bankruptcy court on bondholders’ wealth in bankruptcy. We collect information on the composition of UCCs of firms filing for bankruptcy between 1988 and 2010 and combine it with Moody’s data on in-court loss given default (LGD) rates. Our tests show that bondholders’ losses monotonically increase with the assignment of seats to unions on unsecured creditors’ committees. Notably, the LGD rates of secured creditors (e.g., banks) on the same firms are found to be insensitive to the number of UCC seats assigned to unions.

We also exploit firm and union heterogeneity in our framework to help characterize how unionization affects bond 3 values through expected bankruptcy costs. First, we compare subsamples of financially distressed and financially healthy firms, expecting bond price reactions to news of unionization to be particularly pronounced for firms in distress. Across several measures of financial distress, the results show that unionization has a much greater impact on the bonds of distressed firms. We also look at the funding status of firms’ pension plans. Unionized workers’ pensions are entitled to the same (high) priority assigned to their wages in bankruptcy. As such, underfunded plans will aggravate bondholders’ expected bankruptcy costs. We partition our sample using firms’ pension funding status and find the effect of unionization to be significantly stronger for firms with underfunded plans. Finally, we examine the argument that the value impact of unions is related to their bargaining powers. The adoption of right-towork (RTW) laws by some state legislatures allows nonunion workers to enjoy the benefits of collective bargaining without paying union dues. These laws constrain unions’ financial resources, diminishing their powers. Partitioning our sample according to whether a union election is held in a state with RTW laws, we find that the negative impact of unionization on bond values is much weaker in states with RTW laws in place (where unions are weaker).

We must caution readers about limits in the generalization of our inferences. First, our methodology focuses on contrasts between closely won and closely lost elections-a narrow band. Second, our estimates refer to firms with access to bond markets that witness union elections after 1976. Our results, therefore, do not directly speak to union elections won by large margins, to votes conducted in small firms, or to firms that do not observe votes for unionization in their plants after the 1970s. With these limitations in mind, our findings are important in assessing the impact of labor force unionization on the bondholders of large, public firms over the past four decades.

This research brief is based on Murillo Campello, Janet Gao, Jiaping Qiu, and Yue Zhang, “Bankruptcy and the Cost of Organized Labor: Evidence from Union Elections,” Review of Financial Studies 31, no. 3 (March 2018): 980-1013.

Download the Full Research Brief in Economic Policy

Murillo Campello, Cornell University; Janet Gao, Indiana University; Jiaping Qiu, McMaster University; and Yue Zhang, Université Catholique de Louvain.