Using firms that continue to trade on a major stock exchange while in bankruptcy, the author finds that the market fails to process Chapter 11 bankruptcy signals in a timely manner. The results indicate a statistically significant and negative post–Chapter 11 drift over the six-month period after the bankruptcy filing date. The findings are robust to various alternative specifications and are inconsistent with the efficient market hypothesis.
The author examines the medium-term equity market performance of firms that are in bankruptcy. He finds a negative and statistically significant post-bankruptcy drift of at least –12.5%, on average, over the six-month period after the bankruptcy filing date. This drift is on top of the average adverse market reaction of –26% on the Chapter 11 announcement date. The finding is robust to various alternative specifications, and the result is inconsistent with the efficient market hypothesis. It adds to the literature that shows the market is unable to process bad public news events in a timely manner.
How Is This Research Useful to Practitioners?
The author finds that bad news is not incorporated quickly into stock prices. As he notes, this finding is inconsistent with the efficient market hypothesis, which suggests that stock prices incorporate all value-relevant information very quickly.
The author also helps identify which types of firms are more susceptible to this issue. Hong and Stein (Journal of Finance 1999) predicted that prices will adjust more slowly when firm-specific information is more difficult to access. The author finds a more pronounced effect for firms that are small, are expensive to trade, have small institutional holdings, and are poorly covered by securities analysts or media outlets. In this sample, more than two-thirds of the firms are covered by zero, one, or two analysts. These results are not surprising, but the empirical validation is important. Information about small firms may spread more slowly if there is a high fixed cost to acquire information and if there are market microstructure issues (e.g., increased trading costs and wider bid–ask spreads). As for analysts, they play an important role not only in the dissemination of value-relevant information but also in anticipating Chapter 11 announcements. Therefore, it should be expected that the fewer the analysts covering a firm, the slower the rate that information is transmitted.
How Did the Author Conduct This Research?
The author uses a sample consisting of 275 nonfinancial, nonutility US industrial firms that continued trading on a major US stock exchange after filing for Chapter 11 bankruptcy. He starts by identifying a list of 1,812 Chapter 11 cases between October 1979 and October 2005. He then excludes firms with no data available from CRSP or Compustat (167 firms), firms whose stocks do not trade on the NASDAQ, AMEX, or NYSE after filing for Chapter 11 (1,311 firms), and firms that are classified as utilities or financial firms (59 firms).
The abnormal performance of the sample firms around their Chapter 11 announcement dates is examined by calculating buy-and-hold abnormal returns (BHARs). The BHARs are winsorized at the 1% and 99% levels and are computed over a maximum period of approximately six months post-bankruptcy to minimize the impact of delisting firms in the results. Following the methodology of recent distressed firm stock price performance studies, the author benchmarks each firm against a single control firm. The single control firm is used to determine the expected return of the subject firm. Because abnormal returns differ for small and large firms and firms with higher bankruptcy risk earn lower-than-average returns, the firms are benchmarked on size and pre-bankruptcy distress risk.
With a median firm size of just $8.1 million and a common difficulty of selling short stocks with prices of less than $5 per share, the observed behavior may be difficult to take advantage of. These limits to arbitrage serve to prevent such anomalies from being arbitraged away on a timely basis.
Because of issues surrounding the measurement of abnormal returns over long horizons, the author performs a series of robustness tests. Specifically, the main test is rerun while considering alternative control samples based on size and earnings surprise; size and momentum; and the industry, size, and book-to-market ratio. The results of these alternative analyses are similar to the main result.
The author’s findings are interesting but may be a function of the sample he selected rather than a conclusion that can be generalized. Prior research has demonstrated a slower response to negative news relative to positive news, but such responses are still within minutes of the information disclosure, which is still consistent with market efficiency. It would be interesting to further explore an explanation from behavioral finance, in which the disposition effect suggests that investors are slow to realize losses.