It is now widely accepted that stocks with high short interest ratios underperform the market. This is a very recent bit of conventional wisdom, based largely on the evidence in Asquith and Meulbroek’s (1995) unpublished working paper for New York Stock Exchange (NYSE) and American Stock Exchange (Amex) stocks, and Desai, Ramesh, Thiagarajan, and Balachandran’s (2002) article for Nasdaq stocks. Both Asquith and Meulbroek and Desai et al report negative and significant abnormal returns for firms with short interest ratios of 2.5% or more, where the short interest ratio is defined as the ratio of short interest to shares outstanding.
Both papers also report large secular increases in short interest ratios, and skewed cross sectional distributions, with most stocks having short interest ratios of less than 0.5%, and very few firms having a ratio exceeding 10%. Prior to these papers, the conventional wisdom was that large short positions presaged positive future returns, caused by the flow demand from short sellers covering their positions.
The rationale for why high short interest should presage negative abnormal returns relies upon the fact that short selling stocks is costly or constrained relative to taking long positions. There are many reasons for this, primarily regulatory constraints. If short selling is costly, the “votes” of optimistic investors carry a greater weight than those of pessimistic investors in stock valuation. Two predictions flow from these observations. First, firms with a high dispersion of opinion will be overvalued, and thus have low subsequent returns, as first discussed by Miller (1977). Second, since short sellers will concentrate their positions in the firms that are most overvalued, firms with high short interest ratios will have low subsequent returns, as modeled by Diamond and Verrecchia (1987).
This paper shows that the new conventional wisdom regarding short interest ratios and return predictability, as well as continual increases in short interest ratios over time, is premature and incomplete. Using a longer time period for both NYSE-Amex and Nasdaq stocks and examining short interest and stock returns in more detail than any previous study, we find that the patterns are more ambiguous than the recent literature suggests. While the result that high short interest stocks underperform is generally true for equally weighted portfolios, it does not hold for value weighted portfolios. Specifically, for the period from 1976 to 2002, equally weighted (EW) portfolios of NYSE-Amex and (beginning in July 1988) Nasdaq firms with high short interest reliably underperform.
These same portfolios over the same calendar time period, however, do not reliably underperform on a value weighted (VW) basis. For example, firms in the 99th percentile of short interest ratios have monthly abnormal returns of –53 basis points (t= -2.44) on an EW basis, but only –5 basis points (t= -0.17) on a VW basis. This difference is not because highly shorted stocks are all micro-caps, however. While highly shorted stocks have a larger median equity value than the median of non-highly shorted stocks, they are disproportionately small stocks, with both micro-cap stocks and large-cap stocks underrepresented. This is true for the entire sample and for the subsamples of NYSE-Amex and Nasdaq firms.
Download
Short Interest and Stock Returns
