Bad news disclosures and market bias : do investors underreact?

Abstract

The seminal market efficiency paradigm in finance is being increasingly challenged by evidence apparently inconsistent with its predictions. Such "anomalies" tend to show that the market does not fully incorporate information upon its release in an unbiased way. Recent literature in finance identifies two potential types of anomalous market reaction to news disclosuresl overreaction, and underreaction. The overreaction phenomenon does not find much empirical support but market underreaction, on the other hand, appears quite robust, particularly in the case of bad news which the market appears to take time to process in many situations. My PhD explores these issues. The first part of my thesis tests the hypothesis that if investors rationally incorporate new pessimistic (optimistic) information then after controlling for risk, bad (good) news firms will not under- (over-) react. I test this hypothesis in the going-concern modified audit report disclosure domain. Going-concern opinions offer an appropriate test for the underreaction model as such information releases are associated with acute psychological stress and where a clear distinction between bad and good news can easily be made by considering the parallel case of going-concern withdrawal events. The second part of my thesis extends my work to investigate the market underreaction phenomenon conditional on the underlying bankruptcy regime of the institutional environment. Specifically, I explore the market response to the information content of closely related going-concem modified audit report disclosures (bad news) conditional on the underlying bankruptcy codes in very similar institutional and market environments differing only in the nature of bankruptcy regimes. More specifically, I work with the debtor-friendly U. S. and the creditor- friendly U. K. legal regimes. I hypothesize that investors in a creditor-friendly bankruptcy regime (the U. K. ) will react more adversely to the publication of first-time going-concem modified audit report indicating increased risk of loss than do investors in a debtor- friendly bankruptcy regime (the U. S. ). This is because of a remarkable divergence across the bankruptcy codes of the two different countries with regards to the rights of claimholders in the event of a default on debt contracts. The idea is to test whether there is any difference in investor response to similar bad news signals highlighting financial distress across different institutional environments. In the first part of my thesis, I find that there is asymmetric market response to first- time going-concern modified audit report disclosures (bad news) and withdrawal of the going-concern modified audit report disclosure (good news). Using a large sample of 845 U. S. firms from 1994-2002,1 find that the market underreacts to going-concern modified audit report disclosures (bad news), resulting in a downward drift of around -16% over the one-year period subsequent to the publication of going-concern modified audit opinion, but treats its withdrawal (good news) consistent with theory with no subsequent abnormal returns. To ensure that my empirical results are robust I employ various methodologies and also conduct additional tests to control for alternative explanations to my market underreaction story such as post-earnings announcement drift, momentum etc. My main results on market underreaction to going-concern modified audit report disclosures remain unchanged. I also test if there is an opportunity to eam profits by trading on this underreaction anomaly but find that any profit opportunity is illusory and highly risky. I conduct additional analyses that explore the trading behaviour of different classes of investors in my sample firms. This analysis is important as it could highlight whether institutional investors and retail investors exhibit similar trading biases. Results reveal that institutional investors reduce their holdings in such stocks on a timely basis in contrast to retail investors who appear to increase their holdings in such distressed stocks. The evidence is even clearer when such analysis is conducted by splitting my going-concern sample by subsequent outcomes. I conclude that de. spite clear adverse signals about the firm's continuing financial viability being conveyed to investors by the publication of the going-concem modified audit report, this information is not being fully impounded by the market, in contrast to the good news conveyed by going-concern withdrawals. My findings add to the existing literature calling into question the ability of the market to rationally price stocks in the case of acute public-domain bad news disclosures, as opposed to good news releases. My results suggest that my evidence of stock mispricing and extended post-goi ng-c once rn drift might then be explained by a limits to arbitrage argument with naYve (retail) investors keeping stock prices artificially high by trading inappropriately in these stocks due to behavioural biases identified in the literature and skilled investors (professional arbitrageurs) having limited incentive to trade in these small firms because of high costs. In terms of the second main theme of my thesis, my empirical analysis comparing the market response to going-concern modified opinions in the U. S. and the U. K. shows that, as hypothesized, investors in a creditor-friendly regime (the U. K. ) react more adversely, -3 1 %, than investors in a debtor-friendly regime (the U. S. ), - 18%, in the eight year time-period (1995-2002). One particular reason is that the U. S. bankruptcy regime is biased more towards the rights of' debtors, whereas the U. K. regime is biased more towards the rights of creditors and once a firrn enters bankruptcy proceedings in the U. K., it is unlikely that stockholders' equity has any residual value. These results provide evidence of the important role of legal regimes on the informativeness of accounting inforination. My results suggest that as standard-setters pursue uniform accounting and auditing standards across the world, they need to take into account how such standards interact with local legal regimes and consequently their informativeness to investors and other financial statement users. As such, these results present crucial empirical evidence that adds to the ongoing debate about the relevance of global standards among standard- setters, regulators and academics. Overall, my thesis makes important theoretical and empirical contributions to the behavioural finance, market pricing, and accounting literature in the bad news disclosure domain.