The authors examine whether a decrease in the cost of short selling affects earnings management, fraud detection, and price efficiency. The authors find that earnings management declined for “pilot” firms that the SEC temporarily exempted from short-sale price tests (i.e., the uptick rule). Likewise, during the program the probability that earnings fraud was detected increased for pilot firms. Finally, pilot firms’ stock prices reflected future earnings information better (i.e., more efficiently) than those of non-pilot firms.

Practitioners may wish to monitor measures of short-selling costs as a flag for the likelihood and intensity of earnings management and for informed trading around earnings announcements.

There is little research on how the prospect of short selling (as opposed to actual short selling) affects financial reporting and stock price efficiency. This study examines a quasi-experimental situation (the SEC pilot program) that captures conditions very favorable for determining the effects of easing short-sale restrictions. Performance-matched discretionary accruals of pilot firms are clearly lower than those of non-pilot firms, and the marginal effect on fraud detection corresponds to a 38% increase in the unconditional probability of detection, making that aspect of the study potentially valuable for professionals. Although the research findings are interesting, they have limited explanatory power (low R²).

The authors also find that post–earnings announcement drift (PEAD) after negative news diminished for pilot firms, but only for the most negative news. The likelihood of a pilot firm just beating analysts’ earnings estimates also declined during the program. To the extent that other factors affect the ease of short selling, these findings may be useful for situations in which those factors are themselves changing in magnitude. That is, as certain stocks become easier or tougher to short, the effects ought to vary accordingly.

The authors use a difference-in-differences (DiD) methodology to compare pilot and non-pilot firms over time (before, during, and after the SEC pilot program). For their analysis of discretionary accruals, the authors match pilot and non-pilot firms on ROA to better control for performance that may otherwise be driving results.

The initial 986 pilot stocks are from a list of members of the 2004 Russell 3000 Index that the SEC selected in order to examine the effect of relaxing short-selling restrictions. From this set, the authors exclude financial services and utilities, resulting in a sample of approximately 741 pilot and 1,504 non-pilot firms, which further shrinks to 388 pilot and 709 non-pilot firms when the authors require all firms to have full data over the study period.

The authors’ analytical statistics follow established research. The Kothari–Leone–Wasley performance-matched discretionary accruals measure, for example, decreases for pilot firms during the program and reverts to pre-pilot levels afterward. The Lundholm–Myers returns–earnings model, along with its statistic that proxies for price efficiency, shows that pilot firms’ stock prices better reflect future earnings, consistent with greater price efficiency. The Boehmer–Wu methodology for the PEAD measure is the basis for the comparative analysis, which shows a difference in PEAD between pilot and non-pilot firms but only for the firms with the most negative earnings surprises.

Overall, this research is interesting and confirms what practitioners might find intuitive: Earnings management diminishes and price efficiency improves in the face of increased market forces—here, in the form of an easing of short-selling restrictions. It would be interesting to know what effects, if any, the subsequent changes (see https://www.sec.gov/investor/pubs/regsho.htm) to Regulation SHO have had and whether they are comparable in size to those during the original pilot program. Pragmatically, do the findings translate into actionable investment strategies or tactics? Unfortunately, this study does not break down the effects by firm size or industry—where they might have been concentrated and more tradable (or not) in a particular grouping of stocks.