Study · Short interest

Does short interest predict stock returns?

Measured Sep 24, 2026 · outcomes through Aug 21, 2026

The short answer

A little, through short-sale volume, not days to cover. Since 2009, the fifth of US stocks with the heaviest share of volume sold short beat the median stock over the next month 49 times in 100, against 51 for the lightest, the same way in 15 of 18 years: a tilt across hundreds of stocks, a nudge for any one. Days to cover (since 2021) showed no measurable edge.

Heavy short selling, then a slightly weaker month

Every Wednesday since 2009, US stocks are split into fifths by the share of their trading volume that was sold short over the last 21 sessions, and each fifth is followed against the same day's median stock. Over the next month the lightest-shorted fifth beat it 51 times in 100 and the heaviest 49; over three months, 51 and 49. The dashed line is chance.

Share that beat the median stock over the next month

Fifths by share of volume sold short over 21 sessions, lightest to heaviest

505644chance 5051lowest50low50middle49high49highest
Share that beat the median stock over the next three months

Fifths by share of volume sold short over 21 sessions, lightest to heaviest

505644chance 5051lowest50low50middle49high49highest

For the curious: rank correlation -0.017 at one month (t -3.1) and -0.022 at three months (t -3.0); the 5-session version measured -0.019 at three months (t -3.3).

How small is small?

Over the next three months the lightest-shorted fifth averaged +0.64 percentage points against the average stock and the heaviest -0.42: a gap of 1.06 points, the same way in 14 of 18 years. That is a tilt across hundreds of stocks and a nudge for any one of them, small next to how far a single stock can move in three months.

Out of 100 stocks, over the next month
51 of 100 of the lightest-shorted stocks beat the median stock
49 of 100 of the heaviest-shorted stocks beat the median stock
Average return against the average stock over three months

Percentage points, fifths by share of volume sold short, lightest to heaviest

+0.64lowest+0.18low-0.08middle-0.40high-0.42highest

Days to cover and squeeze scores: no measurable edge

Days to cover divides the shares sold short by average daily volume: roughly how many days of normal trading short sellers would need to buy them back. Our history for it starts in 2021. Over the next month the fifth with the most days to cover beat the median stock 49 times in 100 and the fifth with the fewest 50. Yet on average the fifth with the most days to cover did better against the average stock: +0.19 points, against +0.11. Over 6 years of data, neither gap can be told from chance. The squeeze score, short-side crowding and the change in short interest also showed no measurable edge.

Share that beat the median stock over the next month, by days to cover

Fifths by days to cover, fewest to most, since 2021

505644chance 5050lowest51low50middle50high49highest
Every short-interest measure we track

Of the 6 measures, only the short-sale volume ratios carried a measured tendency; 4 showed no measurable edge at any horizon.

MeasureData sinceWhat it showed
Share of volume sold short, last 21 sessions2009Proventhe next month, the next three months
Share of volume sold short, last 5 sessions2009Proventhe next three months
Days to cover2021No measurable edgeat any horizon
Squeeze score (days to cover, weighted by its rank)2021No measurable edgeat any horizon
Short-side crowding (days to cover, short volume and change, blended)2021No measurable edgeat any horizon
Change in shares sold short since the last report2021No measurable edgeat any horizon

Why short interest arrives late

Short interest, the number of shares sold short and not yet bought back, is reported twice a month and published days after the date it describes; this study dates every reading by the day it was published. Short-sale volume is faster: the share of each day's reported volume that was sold short, published every evening. Much of it is market makers selling short to fill buy orders, so a high reading is not simply a crowd betting against the company.

Context, not a call

Short-interest data describes who is betting against a stock and how crowded the exit is. In this data only the flow of short selling leaned on what came next, and only a little; the stock of short positions, its change and the squeeze scores built on it did not. Read them as context about a stock, not as a forecast of its price.

Volume and liquidity: another small, steady tilt

How this was measured

Every Wednesday since 1995, the stocks in our universe are compared on the signal and followed for 1, 5, 21 and 63 trading sessions. A reading counts when the stock beat the median stock on the same day, so chance is always 50 in 100 and a market-wide rally cannot flatter the result. The strength of a relation is a rank correlation (IC) between the signal and the forward return; its t-statistic is taken on non-overlapping dates, and a relation is called proven only when |t| is at least 3, it held in at least 60% of years and it is large enough to matter. With hundreds of tests, a looser bar would pass about one in twenty by luck. Studies run Sep 24, 2026; outcomes measured through Aug 21, 2026.

Research, not advice. A measured tendency across hundreds of stocks is a nudge for any one of them, never a forecast of its price.

Questions

Does high short interest predict a falling stock?
Not in a way we could measure. Since 2021, the fifth of US stocks with the most days to cover beat the median stock over the next month 49 times in 100, against 50 for the fifth with the fewest: no measurable edge. The measure that did carry a small tilt is the share of daily volume sold short.
Does heavy short selling predict lower returns?
Slightly. Since 2009, the fifth of US stocks with the heaviest share of volume sold short beat the median stock over the next three months 49 times in 100, against 51 for the lightest fifth, the same way in 14 of 18 years. A tilt across hundreds of stocks, a nudge for any one.
Do heavily shorted stocks get squeezed higher?
Not as a tendency. Our squeeze score, days to cover weighted by its own rank, showed no measurable edge from the next session to the next three months on US stocks since 2021. Squeezes happen to single stocks; across the most-shorted stocks they did not add up to a measurable edge.
Does rising short interest mean a stock will fall?
Not measurably. The change in shares sold short between two reports showed no measurable edge at any horizon from the next session to three months, on data since 2021. Short interest is reported twice a month and published days later, so by the time a rise is public it is already dated.
What is days to cover, and does it predict returns?
It is the shares sold short divided by average daily volume: roughly how many days of normal trading short sellers would need to buy them back. On US stocks since 2021 it showed no measurable edge. Over three months the fifth with the most days to cover beat the median stock 49 times in 100 and the fifth with the fewest 51. Yet on average the fifth with the most days to cover did better against the average stock: +0.75 points, against +0.40. Over 6 years of data, neither gap can be told from chance.

References

  • Boehmer, E., Jones, C. M. and Zhang, X. (2008). Which Shorts Are Informed? Journal of Finance 63(2).
  • Dechow, P. M., Hutton, A. P., Meulbroek, L. and Sloan, R. G. (2001). Short-Sellers, Fundamental Analysis, and Stock Returns. Journal of Financial Economics 61(1).