فروش کوتاه مدت ADRs و محدودیت های فروش کوتاه مدت بازار های خارجی

We examine the effect of home market short-sale constraints on securities that also trade in other countries that have more liberal short-sale rules. In particular, we focus on the case of ADRs traded in the US, as in some cases, the home markets of these ADRs prohibit short selling. We find that short sellers more heavily trade ADRs from countries where short selling is prohibited than from markets where short selling is allowed. Furthermore, we find that the greater levels of short selling in ADRs with binding home-market constraints is driven by stocks with greater dispersion of investors’ opinion, low fundamentals-to-price ratios, and recent price increases. Our results support the hypothesis that short sellers target ADRs with home market short-sale constraints because these ADRs are more often subject to temporary misvaluation.

Recent studies that examine international equity markets find that short-sale constraints generally reduce the efficiency of prices and can lead to overvaluation. For example, Bris et al. (2007) and Saffi and Sigurdsson (2010) find that short-sale constraints adversely affect the price efficiency of securities in the home market, while Chang et al. (2007) show that short-sale constraints in Hong Kong lead to stocks becoming overvalued. For the majority of the stocks trading in these constrained markets, investors are limited in their ability to trade on information that might indicate temporary overvaluation. However, for a small subset of these stocks with tradable ADRs, there exists an opportunity for investors to circumvent the effect of short-sale constraints by shorting the ADR. Theory in Miller (1977) suggests short-sale constraints can lead to overvaluation, particularly in the presence of heterogeneous beliefs among investors. In related work, Dechow et al. (2001) find that short sellers target stocks with low fundamentals-to-price ratios. Dechow et al. argue that short sellers’ penchant for these stocks is consistent with the idea that short sellers attempt to actively exploit temporary overpricing. Similarly, Diether et al. (2009a) find that short sellers target stocks that become temporarily overvalued by showing that daily shorting activity is positively related to both contemporaneous and past returns. In this paper, we revisit tenants of Miller’s (1977) theory by examining the short-selling activity of ADRs while conditioning on the underlying securities’ home-market constraints. For exposition, we denote ADRs of foreign stocks that face binding home-market constraints as unfeasible ADRs. Similarly, we denote ADRs of foreign stocks with fewer home-market constraints as feasible ADRs. To determine the feasibility of shorting, we use the classification of short-sale restrictions by Charoenrook and Daouk (2005) and partition ADRs based on the feasibility of shorting in the underlying security’s home country. Our hypothesis is that short-selling will be greater for unfeasible ADRs than for feasible ADRs. Specifically, if ADR prices and the prices of the underlying securities are closely correlated (Kato et al., 1990) and short sellers target stocks that become temporarily overvalued (Dechow et al., 2001; Diether et al., 2009a), then we expect to see more shorting of unfeasible ADRs that exhibit signs of potential overvaluation.1 It is important to note, however, that our study is not concerned with determining whether binding home-market constraints lead to overvaluation in the underlying security, per se. Nor does our study examine whether the shorting of ADRs can reduce overvaluation caused by home-market constraints. 2 To examine the role that overvaluation plays in the short selling of ADRs with binding home-market constraints, we investigate three signals of overvaluation that are discussed in prior work. First, we examine whether short sellers who target stocks with greater heterogeneity among investors’ opinions do so more frequently for unfeasible ADRs.3 Second, we test whether short sellers’ penchant for stocks with low fundamentals-to-price ratios, shown in Dechow et al. (2001), is stronger for unfeasible ADRs. Finally, we determine whether the contrarian behavior of short sellers, documented in Diether et al. (2009a), is more pronounced for unfeasible ADRs. We begin our tests by comparing the shorting activity of feasible ADRs to the shorting activity of unfeasible ADRs. Consistent with our hypothesis, univariate results show that shorting of unfeasible ADRs is more than 40% higher than shorting of feasible ADRs. In multivariate tests these results persist and are robust to controls for a variety of factors that influence the level of short-selling and suggest that ADR shorting activity depends, in part, on the existence of short-sale prohibitions in the ADR’s underlying home market. Next, we examine whether short sellers’ attraction to overvalued securities is stronger for unfeasible ADRs than for feasible ADRs. Uniformly, our results show that short sellers’ proclivity to target securities with (i) high levels of heterogeneity in investor’s opinions, (ii) low fundamentals-to-price ratios, and (iii) recent increases in price is greater for unfeasible ADRs than for feasible ADRs. These results indicate that not only do short sellers circumvent home-market constraints by trading ADRs, but short sellers’ propensity to do so is increasing in the likelihood that the ADR is temporarily overvalued based on the signals discussed above. A natural extension of our analysis is to examine the common negative relation between current short selling and subsequent security returns (Senchack and Starks, 1993, Aitken et al., 1998, Desai et al., 2002, Ackert and Athanassakos, 2005 and Boehmer et al., 2008; Diether et al., 2009a). Theory in Diamond and Verrecchia (1987) posits that when constraints become binding, prices adjust less quickly to negative information. If shorting ADRs provide a mechanism that mitigates the effect of constraints, then the common negative relation between current short sales and subsequent returns should be stronger for unfeasible ADRs than for feasible ADRs. In our final set of tests, we find that short selling of unfeasible ADRs is better at predicting negative returns than short selling of feasible ADRs. In economic terms, we show that the return predictability contained in short sales of unfeasible ADRs is between two to three times greater than the return predictability in short sales of feasible ADRs. Combined with earlier findings, these results suggest that, not only do ADRs provide a mechanism to bypass the effects of binding home-market constraints, but the information contained in shorting is also greater for ADRs with these constraints.

While US equity markets permit short selling, many foreign markets do not. However, US traded ADRs can provide a mechanism that has the potential to partly circumvent country specific short-sale constraints by allowing investors to short the ADR in a market with less restrictions. In this paper, we test whether short-selling levels are greater for ADRs with home-market constraints than for ADRs without home-market constraints. This test is motivated by theory in Miller (1977) that suggests that constraints can lead to overvaluation. Our basic hypothesis, that the shorting of ADRs will be driven by ADRs with binding home-market constraints, contains several components. First, if constrained short sellers target stocks with greater heterogeneity or dispersion among investors’ opinions (Miller, 1977), then we expect that higher levels of short selling in ADRs with home-market constraints will be driven by ADRs with increasing dispersion. Second, following Dechow et al. (2001) who find that short selling occurs in stocks that are temporarily overpriced, we expect that higher shorting levels in ADRs with home-market constraints will be driven by ADRs with low fundamentals-to-price ratios. Third, based on the Diether et al. (2009a) finding that short selling relates directly to both contemporaneous and past returns, we expect that contrarian short selling should be driven by ADRs with binding home-market constraints. Results from our univariate analysis show that shorting activity is more than 40% higher for ADRs with binding home-market constraints than for ADRs without constraints. This result holds when we control for other factors that influence short selling in our multivariate tests and indicates that ADRs provide an opportunity for short sellers to migrate from a constrained foreign market to the US market that has more liberal short-selling restrictions. We also find that the higher shorting levels in ADRs with home-market constraints is driven by ADRs with (i) greater dispersion of opinion, (ii) low fundamentals-to-price ratios, and (iii) increasing prices. Each of these findings indicate that not only are ADRs with binding home-market constraints shorted more heavily, but that these ADRs also exhibit signs of overvaluation which is targeted by short sellers. In our final test, we condition the return predictability of ADR short sellers on the tightness of home-market constraints. The motivation for this test is based on both theoretical and empirical work showing that prices of short-sale constrained stocks respond more slowly to negative information than stocks that are unconstrained. If ADRs provide an avenue for otherwise constrained short sellers to trade, then the information (or return predictability) contained in short sales should be greater for ADRs with binding home-market constraints. We find that the common negative relation between current short selling and subsequent returns is stronger for ADRs with home-market constraints than for ADRs without constraints. In economic terms, the return predictability contained in short selling is two to three times greater for unfeasible ADRs than for feasible ADRs. Combined with our earlier findings, these results indicate that not only do short sellers attempt to profit from the overvaluation effect caused by foreign-market short-sale constraints by targeting these security’s ADRs, but the information contained in short sales of these ADRs is also greater.