Knowledge that Transforms

I investigate the impact of implementing SFAS No. 86, which provides an exception to the GAAP requirement of the immediate expensing of research and development (R&D), on information asymmetry. Using bid-ask spread and share turnover as proxies for information asymmetry, I find that after the introduction of SFAS No. 86, information asymmetry decreases for software firms relative to that of other high-tech firms. Within the software industry, I find that information asymmetry is significantly lower for firms that capitalize (capitalizers) than for those who expense (expensers) software development costs. Thus, accounting for software development costs per SFAS No. 86 reduces information asymmetry and, consequently, the cost of capital. As well, investors' uncertainty about the future benefits of software development costs is reduced when firms capitalize these costs.

This study investigates whether paying higher wages motivates employees to provide higher effort and whether firm profit moderates this relation. Consistent with gift exchange (Akerlof 1982) and reciprocity (Rabin 1993) models, my experimental results show that workers provided more effort when they were paid higher wages even though there was no ex post financial reward for doing so. Moreover, firm profit influenced the relation between wages and effort. Workers provided higher effort when firm profit decreased compared to when it increased. This suggests that the degree of reciprocity is affected by firm profit. However, workers' responded asymmetrically to firm profit, in that they behaved as if they expected to share in firm profit increases but not decreases. Although firms were fairly adept at predicting the profit-maximizing wage strategy, they apparently did not anticipate workers' reluctance to share in firm profit decreases.

This study investigates the effect of firms' adoption of SFAS No. 131 segment disclosure rules on the stock market's ability to predict the firms' earnings, as captured by the forward earnings response coefficient (FERC). The FERC is the association between current-year returns and next-year earnings. SFAS No. 131, effective for fiscal years beginning after December 15, 1997, arguably increased both the quantity and quality of segment disclosure. Consistent with the standard's intended qualitative effects, pre-131 multi-segment firms experienced a significant increase in FERC after adopting SFAS No. 131. Consistent with the standard's intended quantitative effects, many pre-131 single-segment firms began disclosing multiple segments, and those that did experienced an increase in FERC. However, pre-131 single-segment firms that remained single segment (i.e., were unaffected by SFAS No. 131) had no change in FERC, indicating that the increase in FERC for 131-affected firms is not due to some other event concurrent to the adoption of SFAS No. 131. These results are robust under numerous procedures that control for characteristics of the sample firms and their earnings, providing strong evidence that SFAS No. 131 resulted in an increase in stock price informativeness for affected firms. Thus, we provide the first empirical price-based evidence that SFAS No. 131 provided more information (about future earnings) to the market, as the standard's proponents have suggested.

We investigate whether investors price losses conditional on the likelihood of the firm's return to profitability, consistent with the abandonment option hypothesis (Hayn 1995). We first develop a loss-reversal model to define subsamples of persistent and transitory losses. We find that on average investors price transitory losses positively over the sample period, as if a transitory loss indicates a low likelihood of exercising the abandonment option. We also observe that early in the sample period investors do not price persistent losses, as predicted by the abandonment option hypothesis. By contrast, later in the sample period, larger persistent losses correspond to higher returns, inconsistent with the prediction of the abandonment option hypothesis. To understand why we observe a change in the valuation of persistent losses over the sample period, we study their components and establish the key role of R&D for their valuation. We find that investors do not price persistent losses without an R&D component, consistent with these losses indicating financial distress and a higher likelihood of exercising the abandonment option. However, when persistent losses contain R&D, investors separately value the R&D component as an asset and the non-R&D component as if it is a transitory loss. Thus, investors do not consider losses to be homogeneous, but consider the causes and nature of the loss to assess its long-term implications for firm value.

Information that firms provide about financial instruments and derivatives should help investors judge risk. However, this paper reports that such information often is not effective for this purpose. Through a series of experiments, we demonstrate that the labels firms use to describe financial instruments and derivatives cause investors to assess economically equivalent instruments as different in terms of risk. We also show that loss-only disclosures that companies use to describe their risks cause investors to assess the same level of risk for firms with differing underlying exposures. Moreover, we establish that loss-only disclosures cause investors to make risk judgments that correspond to infrequently used risk-management strategies. We test two possible remedies for these judgment problems. Our results show that additional information describing the underlying economic exposures of a financial instrument does not eliminate the labeling effects. However, we do find that providing investors with upside and downside (i.e., two-sided) risk disclosures help them distinguish among firms using different risk-management strategies.

I examine the association between audit committee quality and the quality of corporate internal control. While information on the quality of internal control is not generally available, companies changing auditors are required to disclose any internal control problems that were pointed out by their predecessor auditors. The empirical results are based on a comparison of companies disclosing such internal control problems with a control sample of companies changing auditors but not disclosing internal control problems. Audit committee quality is measured in three dimensions: its size, its independence, and its expertise. The internal control problems are observed at two levels of increasing seriousness: reportable conditions and material weaknesses. The sample time period precedes the effective dates of recent policy changes regarding audit committees. The results indicate that independent audit committees and audit committees with financial expertise are significantly less likely to be associated with the incidence of internal control problems. This is true for both levels of internal control problems. The results are consistent with recent policy emphasis on audit committee independence and expertise.