محرومیت؟ نگرانی بازار سهام از افشای گزارش حسابرسی

We investigate the stock price reaction to UK going-concern audit report disclosures in the calendar year subsequent to publication. Over this period our firm population underperforms by between 24% and 31% depending on the benchmark adopted. This market underreaction to such an unambiguous bad news release is not a post-earnings announcement drift phenomenon; it is also robust to other potentially confounding explanations. However, whatever the reasons for such stock mispricing, we find costly arbitrage prevents rational investors forcing prices back into line with fundamental value. Our results have implications for the market's ability to impound bad news appropriately and the incompleteness of arbitrage in such small “loser” firm situations.

An increasing body of research suggests that the stock market takes time to assimilate bad news, in contrast to a more timely incorporation of good (positive) news. For example, Womack (1996) shows that analysts’ new buy recommendations provide only small and short-lived price increases whereas new sell recommendations appear to take up to 6 months for their implications to be fully incorporated in market prices, with average 6-month abnormal returns of −9.1%. Dichev and Piotroski (2001) demonstrate that markets respond in similar ways to Moody's bond rating changes, finding no significant abnormal returns following upgrades but negative abnormal returns of between −10% and −14% in the first year following downgrades with a further −3% to −7% in each of the second and third years following the original announcement. Likewise, Dichev (1998) shows that stocks in the highest decile of bankruptcy risk underperform those with low bankruptcy risk (lowest 70% probability of bankruptcy) by around 14% in the year following portfolio formation, and that this pattern continues on a rather more attenuated basis for a further 3 years. Other areas with apparent incomplete stock price reaction to new information events include the post-earnings announcement drift anomaly (e.g. Ball and Brown, 1968; Bernard and Thomas, 1989 and Bernard and Thomas, 1990), dividend cuts and omissions versus increases (Michaely et al., 1995), IPOs/SEOs (e.g. Ritter, 1991; Loughran and Ritter, 1995 and Loughran and Ritter, 1997; Spiess and Affleck-Graves, 1995) and stock splits (e.g. Ikenberry and Ramnath, 2002). Both potential measurement problem issues and behavioral explanations for these apparent empirical irregularities are discussed in the literature. This paper contributes to the “underreaction” literature by providing a sharply defined setting for testing the speed of market reaction to a very serious and unambiguous bad news event, the receipt of a first time going-concern modified (GCM) audit report by a firm. In addition, since most of the market anomalies research has been conducted in US capital markets, our UK context allows us to extend this literature by exploring related issues in an alternative trading environment. Two related studies are those of Willenborg and McKeown (2001) and Weber and Willenborg (2003) who explore the case of going-concern audit opinions in micro-cap IPOs. In particular, Willenborg and McKeown (2001) find IPOs with a GCM audit report on the private-company financial statements contained in their prospectus significantly underperform those with non-going-concern audit reports by around 25% in the first year post-IPO and by three times that on a 3-year horizon. The authors attribute their results to the private information conveyed by the going-concern opinion helping uninformed investors better estimate IPO value in the secondary market. In a companion paper, Weber and Willenborg (2003) break down this analysis by type of audit firm and find that 1-year aftermarket performance is more negative for firms receiving going-concern opinions than those with clean opinions for Big 6 and national firms but not for local firms. We explore the medium-term price reaction to going-concern audit report disclosures by London Stock Exchange (LSE) firms over the one calendar year period subsequent to their publication in the firm's annual report. In particular, we investigate whether such price relevant information is impounded fully by or around the information release date in line with the efficient markets hypothesis, or whether the implications of this bad news disclosure appear to be difficult to assimilate, taking time to be absorbed by the stock market. Our results show that over the 12-month period commencing with the start of the month following the information disclosure event, our sample firms underperform by between 24% and 31%, depending on the benchmark adopted. Equivalent raw returns of −17% in a rising market suggest these results are unlikely to be due to a “bad model” problem. We acknowledge the well-recognized difficulties of measuring abnormal returns for small, less-well followed firms (e.g. Kothari, 2001) which constitute our sample of firms in financial distress. Nonetheless, given the magnitude of our abnormal returns and the extended time period covered in our analysis, such issues as high transaction costs, lack of analyst following and related factors cannot, on their own, fully explain the magnitude of our empirical results. In addition, we demonstrate our results are distinct from any post-earnings announcement drift effects, nor are they driven by penny stock characteristics, financial distress status or momentum. Behavioral explanations for market mispricing depend on arbitrage opportunities being limited. Otherwise, the activities of rational investors will “correct” market prices if investor information processing biases lead to these moving away from fundamental value. We report very high trading costs in our data and demonstrate that, despite the apparent large abnormal returns, simulation of a zero (net) investment arbitrage strategy shows profit opportunities to be illusory. Nonetheless, our stocks are reasonably actively traded. As such, we cannot rule out an irrational investor explanation for our results with investors apparently underreacting to or, in effect, denying the bad news conveyed by a going-concern audit opinion and trading at prices inconsistent with underlying value. Further, since professional investors do not appear to liquidate their holdings in our sample firms on a timely basis, we speculate that such biases are not restricted to naïve investors only, at least in the going-concern context. As the existence of such apparent mispricing is inconsistent with conventional notions of market efficiency, our results are clearly anomalous. Our findings have implications for the market's ability to impound public domain bad news appropriately and the limits to arbitrage in such situations (e.g. Shleifer and Vishny, 1997). Such definitive medium-term underperformance complements the findings of earlier studies demonstrating the adverse short-term price impact on firm valuation of the going-concern modification announcement event (e.g. Fleak and Wilson, 1994; Jones, 1996). Nevertheless, the relative economic importance of our findings needs to be kept in perspective. The remainder of this paper is organized as follows: Section 2 describes our data and Section 3 presents our methodology. Section 4 provides our main empirical results, Section 5 tests for the potential alternative post-earnings announcement drift explanation for these and Section 6 describes additional robustness tests of our findings. Section 7 explores the consequences of costly arbitrage. Section 8 discusses our results and their implications. Section 9 concludes.

This paper explores the stock price behavior of firms subsequent to publishing a GCM audit report for the first time. We find highly significant continuing adverse price reaction over the full 1-year period commencing at the start of the month immediately following the GCM publication date of between −24% and −31%, depending on expected return benchmark used, and demonstrate this is not a post-earnings announcement drift effect. We test for alternative explanations to market underreaction but find we cannot explain our results in terms of poor firm matching or bad model problems, nor can we find any evidence that this post-GCM announcement drift is driven by, or significantly associated with, financial distress, penny stock bias or prior stock returns. Tests of systematic mispricing stories depend, inter alia, on the crucial assumptions of investor decision errors and lack of incentives for rational investors to trade against those exhibiting information processing biases, thereby “correcting” market prices. We find that profitable opportunities to arbitrage GCM stocks are severely inhibited by the high costs of trading. This is quite apart from the issue that many such stocks are unborrowable. On the other hand, our analysis does not allow us to reject the behavioral proposition that investors are, in fact, biased in their ability to process the bad news conveyed by a going-concern audit opinion appropriately leading to market underreaction. The relatively high levels of trading activity in the stocks of our small losing firms and the fact that both institutional investors and firm insiders do not sell their holdings post-GCM are consistent with the idea of “denial” of the implications of a GCM audit report in stock valuation judgments. Our findings add to the existing literature questioning the market's ability to impound bad news appropriately. They also have implications for the problems of arbitrage in such situations. The (public policy) argument is that trades are apparently taking place in the light of “full information” but clearly not at a fair (fundamental) value, whether or not such systematic mispricing can be exploited by market traders. Rubinstein (2001) divides market rationality into three categories, maximally rational markets, where all investors are rational, rational markets, where asset prices are set as if all investors are rational, and minimally rational markets. In this case, although markets are not rational, no abnormal profit opportunities exist for investors who are rational. Although our findings clearly violate the traditional market efficiency paradigm, they are still consistent with minimally rational markets on the basis that arbitrage opportunities for investors are limited in practice. Notwithstanding this, it should be pointed out that our analysis does not allow us to reject the hypothesis that, absent any limitations on arbitrage, if all investors, including existing stockholders, are biased in their ability to deal with the bad news conveyed by the GCM event appropriately, the stock market would still misprice GCM stocks. Nonetheless, whatever the explanations for our anomalous results, our evidence is clearly consistent with GCM audit reports having major longer-term negative price impact, supportive of those studies suggesting the going-concern opinion event itself has market value. However, although the empirical results of our paper demonstrate market mispricing of GCM firms, it must be stressed their number and size are small with total market capitalization of our entire firm population of only £2.1 bn (approximately $3.4 bn) and total assets £5.2 bn (approximately $8.4 bn). As such, the relative economic importance of our findings needs to be kept in perspective as well as their relevance to asset-pricing theory.