In November 2012, the EU Regulation on Short Selling (SSR) was introduced in the European Union. The SSR mandates that short sellers disclose short positions to authorities, beginning at 0.2% and continuing every 0.1% above this threshold. In turn, the authorities only publicly disclose short positions starting from 0.5%. However, in January 2017, the Netherlands Authority for the Financial Markets (AFM) accidentally disclosed confidential positions between 0.2% and 0.5%. In our study, we use this accidental disclosure of the full history of short positions as a natural experiment. This enables us to study both the performance of shorted stocks in the period from November 2012 to January 2017, and the immediate effects of a sudden and unexpected increase in transparency, since the entire register of both public (i.e., ≥ 0.5%) and confidential short positions (i.e., ≥ 0.2% and < 0.5%) became publicly available.

We perform three sets of analyses. First, we evaluate the performance of a daily updated portfolio based on the entire register of public and confidential short positions. On the one hand, short positions could be unrelated to future stock returns if these positions arise from hedging demands by institutional investors, or if they are motivated by position holders that are less convinced of their directional views. On the other hand, short positions could be associated with negative stock returns, given that short sellers are generally well-informed investors. We find that a value-weighted portfolio containing small (confidential) short positions underperforms the market index with a statistically significant 4 basis points (bp) per day. We also find that a value-weighted portfolio containing large, public short positions exhibits underperformance as well. These findings suggest there are limitations to arbitraging the predictive ability of short positions.

Second, we perform an event study to investigate the effect of the accidental publication of confidential short positions on stock returns in the post-disclosure period. Consistent with herding concerns voiced by industry bodies (see ESMA report), post-disclosure abnormal returns could be negative if other investors interpret short positions as a negative signal from informed investors, and consequently sell their holdings or establish new short positions. Alternatively, post-disclosure abnormal returns could be positive where the disclosure event results in a (perceived) increase in short-selling risk, or if short sellers cover their positions, for they mostly benefit from them when they are kept secret. Consistent with these explanations, we find evidence for positive abnormal returns after the disclosure of small short positions. The larger the confidential short positions are relative to the already known public position, the higher the abnormal returns.

Finally, we also examine the evolution of borrowing costs surrounding the disclosure event. We document a statistically significant decrease in costs, which points towards a decrease in the demand for shares by short sellers and/or an increase in the supply available for lending. This finding is consistent with short covering and indicates that herding does not seem to play a role after an increase in transparency.

With the SSR in effect for around five years, ESMA is evaluating, amongst other things, the reporting and disclosure requirements related to the transparency of net short positions. On the one hand, ESMA anticipates benefits in the form of reduced information asymmetry from publishing aggregated net short positions on a regular basis (including those that are currently confidential). On the other hand, short-sale disclosure could also provide a coordination mechanism for manipulative short sellers. Interest bodies fear that regular publications might create herding or stimulate abusive practices like squeezing short sellers. Although the accidental publication of short positions was a one-off event, our results suggest that herding is unlikely to be a major concern. Instead, if anything, increased transparency is likely to lead to a reduction in shorting activity.

Rients Galema and Dirk Gerritsen are both Assistant-Professors at the Utrecht University School of Economics.