Firms rely extensively on debt financing. While debt has many benefits, value can be destroyed when lenders worry they will not be repaid. For example, a firm at risk of liquidating may lose customers who derive utility from future interactions with a stable business. Likewise, a firm may be perceived by consumers as having low quality if it is reorganizing in Chapter 11 bankruptcy. Fearing this consumer response, distressed firms may avoid an otherwise beneficial Chapter 11 reorganization to prevent such a loss of customers. While these indirect costs of bankruptcy have been studied for decades, they are notoriously difficult to quantify. In our recent paper, we use two incentivized experiments in which participants make willingness-to-pay decisions regarding bankrupt and non-bankrupt firms to estimate the causal effect of corporate bankruptcy on consumer demand for a bankrupt firm’s products. We find that learning about a Chapter 11 bankruptcy filing causally reduces a consumer’s willingness to pay for the bankrupt firm’s products by 18-35%, depending on the industry.
We consider three reasons why consumers might care about a corporate bankruptcy. First, consumers might worry that a bankruptcy could lead to liquidation, preventing valuable future interactions with a firm such as the use of warranties, return policies, and reward programs (referred to as “future-interactions hypothesis” below). Similarly, consumers with a preference for brand familiarity and loyalty dislike liquidations and the associated switching costs. Second, consumers might fear that a firm’s bankruptcy will cause the firm to reduce the quality of its products (referred to as “current-quality hypothesis” below) during bankruptcy. In this hypothesis, consumers worry a bankrupt firm will try to conserve cash by firing employees, reducing inventory, failing to maintain its assets, or increasing prices. Third, consumers might be concerned that a bankruptcy is a negative signal of a firm’s inherent quality. We show that the first two concerns both significantly contribute to the effect of bankruptcy on consumer demand. In contrast, the third concern about the signal of a bankruptcy filing appears to have little effect.
Measuring the effect of bankruptcy on consumer demand is difficult because of an omitted-variable problem: Unobservable adverse economic shocks can cause both a firm’s bankruptcy filing and a reduction in consumer demand for the firm’s products. To isolate the decline in demand caused by a firm entering bankruptcy, we need an estimate of what demand would have been had the same firm not entered bankruptcy. We form such an estimate by asking experiment participants to report their willingness to pay for various firms’ products, randomly changing the bankruptcy status of each firm while holding all other firm characteristics fixed.
We ran our first experiment during Hertz’s 2020 bankruptcy. We asked participants to make a series of choices between gift cards at Hertz and Enterprise. Participants were entered into a lottery to win one of their chosen gift cards, incentivizing them to report their true preferences. We find that 26% of participants were aware that Hertz was bankrupt. Randomly informing the other participants of Hertz’s bankruptcy, we show that learning about Hertz’s bankruptcy causally reduces willingness to pay for Hertz by 35 cents on the dollar. Learning that similar firms have survived bankruptcy in the past eliminates two thirds of this effect. However, learning that Hertz secured a one-billion-dollar loan to help them through bankruptcy had no effect on participants’ willingness to pay.
In our second experiment, we ask participants to make a series of choices involving two firms, Firm A and Firm B. In some questions these are hypothetical firms, whereas in other questions they correspond to real firms. We give a reference price for some product at Firm A and ask participants to report their willingness to pay for the same product at Firm B. We follow a recent methodology developed by Kessler, Low, and Sullivan (2019) to incentivize participants to honestly report their preferences in hypothetical choices.[i] Because the questions are hypothetical, we can vary the bankruptcy status of “Firm B” for both real and fictional firms. We can also vary other information about Firm B’s bankruptcy, allowing us to understand the mechanisms by which bankruptcy affects consumer decision making. By randomly disclosing the bankruptcy status of a firm to consumers, we estimate the causal effect of the bankruptcy on consumer preferences.
We consider three industries: airlines, car manufacturers, and retail stores. We focus on these consumer-facing industries because of the high number of large historical bankruptcies. We incentivize participants to honestly report their preferences in a series of willingness-to-pay decisions involving real and fictional firms. We find that knowledge of a firm’s bankruptcy causally reduces a consumer’s willingness to pay for that firm’s products by 22%, 19%, and 18% for airlines, car manufacturers, and retailers, respectively.
In order to determine which of the three hypotheses is leading to the reduction in willingness to pay, we add a few treatment groups in which we provide additional information about the hypothetical firms. To test the current-quality hypothesis, we inform participants that some third-party agency has verified that each bankrupt firm’s quality has not changed since entering bankruptcy. To test the future-interactions hypothesis, we add that financial experts have declared that the firm will almost surely survive bankruptcy and continue operating. We find that, for airlines and retailers, the current-quality hypothesis accounts for two-thirds of the decrease in willingness to pay and future interactions account for the remaining third. For car manufacturers, this relationship is reversed. This is intuitive because cars are a durable good, for which future interactions such as warranty usage and replacement-part purchases are important.
We additionally have a treatment group in which participants are told that financial experts have estimated each bankrupt firm has a 50% chance of surviving bankruptcy. Comparing the willingness to pay between the control group, the baseline treatment group, the 50% survival group, and the 100% survival group, we can back out what beliefs participants must have about the survival prospects of bankrupt firms. While consumers are close to correct regarding survival rates of large U.S. airlines, they dramatically underestimate the survival rates of car manufacturers and retail stores. Therefore, educating consumers about the survival rates of bankrupt firms in the car manufacturing and retail industries would substantially reduce the effect of bankruptcy on consumer demand.
While we see large decreases in willingness-to-pay when consumers learn a firm is bankrupt, this result only matters to firms if consumers are aware when firms go bankrupt. We find that a substantial fraction of consumers is aware when a large firm files for bankruptcy. We present experiment participants with a list of firms and ask them to select which, if any, have ever filed for bankruptcy. A large fraction of consumers correctly identifies historical bankruptcies. For example, 48% of participants are aware that J.C. Penney filed for bankruptcy. Similarly, when asked to rate firms based on how close they ever came to bankruptcy, participants give high ratings to firms that have filed for bankruptcy. In contrast, among firms that never filed for bankruptcy, this perceived distress measure has no correlation with empirical distress measures like credit ratings. Thus, while participants are aware of bankruptcy filings, they are not aware of pre-bankruptcy financial distress.
The negative consumer response to corporate bankruptcies that we document harms both bankrupt firms, which lose market share, and consumers, who may have previously derived surplus from a bankrupt firm’s products. To quantify these losses, we use our experiment to estimate a structural model. Using the model as well as historical data on market shares and prices, we can explore counterfactual scenarios in which various historical bankruptcies never occurred. For example, what would have happened if American Airlines had never filed for bankruptcy in 2011? We show that American Airline’s bankruptcy reduced its market share by roughly 9% and consumer welfare by roughly 3%. We find similar results for Delta’s 2005 bankruptcy and United’s 2002 bankruptcy. Our model also shows that bankrupt firms typically set prices slightly lower than they would in the absence of bankruptcy, while competitors opportunistically increase prices.
Samuel Antill is an Assistant Professor of Business Administration in the Finance Unit at the Harvard Business School.
Megan Hunter is an Assistant Professor of Marketing at the Carroll School of Management of Boston College.
This post is adapted from their paper, “Consumer Choice and Corporate Bankruptcy,” available on SSRN.
The views expressed in this post are those of the author and do not represent the views of the Global Financial Markets Center or Duke Law.
[i] Kessler, J. B., C. Low, and C. D. Sullivan. 2019. Incentivized resume rating: Eliciting employer preferences without deception. American Economic Review 109:3713–44