Does short interest predict stock returns?
By Mehmet Saruhan, Founder·Opulence Alpha Research · Published · Data through
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.
Keywords: short interest, does short interest predict returns, short selling, short sale volume, days to cover, short squeeze·JEL classification: G11, G12, G14, C12, C58·Concepts: Multiple testing, Point-in-time data, Backtest overfitting
- Universe
- The same for every study: 1,767 US common stocks in the 11 GICS sectors, 422 of them since delisted; S&P 500 members since 1996 plus large and mid-sized companies outside the index
- Period
- Every Wednesday from 4 Jan 1995 to 19 Aug 2026: 1,634 Wednesdays across 7,959 trading sessions
- Sample
- n = 1,763 stocks of the 1,767-stock universe; 1,889,758 stock-weeks, a median of 1,148 stocks per Wednesday
- Outcome
- Return over the next 1, 5, 21 and 63 trading sessions against the same day's median stock
- Inference
- t on non-overlapping dates; proven only when |t| ≥ 3 and the same sign in at least 60% of years
1Introduction
Heavy short interest is read two ways at once: as the verdict of informed short sellers, pointing to a fall, and as fuel for a squeeze, pointing to a jump. The literature leans toward the first. Dechow, Hutton, Meulbroek and Sloan found that short sellers target stocks that look expensive against their fundamentals, and Boehmer, Jones and Zhang found that heavily shorted stocks go on to trail lightly shorted ones. This study tests both readings on one universe, keeping apart the daily flow of short sales, measured since 2009, and the twice-monthly stock of short interest behind days to cover and squeeze scores, measured since 2021.
2Data and method
2.1Sample design
Population and frame. The population is US common stocks listed on the NYSE and Nasdaq; funds, ETFs, trusts, preferred shares, warrants and units are excluded. The sampling frame is a fixed universe of 1,767 companies, drawn once when the platform was built and not re-sampled since, in two strata. Stratum 1 is a census of the S&P 500: every company in the index at any time since 1996 whose price history could be recovered, 1,027 companies of which 413 have since delisted; it holds 76% of the index's members in 1996 and at least 96% in every year from 2010. Stratum 2 is 740 large and mid-sized companies outside the index, selected in proportion to the market's sector weights from the stocks that passed a minimum share price of $15 and a minimum average daily trading value of $25 million; 9 of them have since delisted.
Sample. The unit of observation is a stock-week: one stock on one Wednesday. A stock enters a Wednesday's cross-section when it has a valid close that day, a value of the signal and a measured outcome; bars flagged as bad data and returns that cross a change of issuer are left out. This study's sample is n = 1,763 stocks of the 1,767: 1,889,758 stock-weeks on 1,634 Wednesdays from 4 Jan 1995 to 19 Aug 2026, a median of 1,148 stocks per Wednesday (range 820 to 1,371). 331,954 stock-weeks (17.6%) come from the 422 companies that have since delisted. Every stock and every Wednesday carries equal weight.
Representativeness. Table 1 gives the sample by GICS sector beside the S&P Composite 1500: 7.9% of companies would have to change sector for the two to match exactly. By latest market value, 55% of the listed companies with a recorded market value are large (at least $10bn), 38% mid ($2–10bn) and 6% small. Because Stratum 2 was chosen from companies listed at construction, its history carries survivorship bias. Section 4 repeats the lead result among the stocks that were S&P 500 members on each date, the part of the sample largely free of that bias.
Companies in this study's sample, of them those since delisted, the size of the listed companies with a recorded market value, the sample's share of stock-weeks and the sector's share of the S&P Composite 1500.
| Sector | Sample | Delisted | Large | Mid | Small | Stock-weeks | S&P 1500 |
|---|---|---|---|---|---|---|---|
| Information Technology | 282 | 67 | 122 | 76 | 18 | 13.9% | 12.7% |
| Financials | 256 | 60 | 114 | 74 | 8 | 14.6% | 17.2% |
| Industrials | 255 | 53 | 126 | 67 | 11 | 15.9% | 17.5% |
| Health Care | 237 | 47 | 92 | 81 | 16 | 12.4% | 10.9% |
| Consumer Discretionary | 209 | 39 | 69 | 87 | 12 | 12.3% | 12.9% |
| Energy | 105 | 30 | 39 | 31 | 5 | 5.9% | 4.7% |
| Consumer Staples | 101 | 38 | 40 | 17 | 5 | 5.9% | 4.9% |
| Materials | 94 | 34 | 35 | 24 | 2 | 5.6% | 5.1% |
| Real Estate | 84 | 10 | 40 | 32 | 2 | 5.7% | 6.9% |
| Communication Services | 80 | 31 | 28 | 13 | 5 | 3.9% | 3.3% |
| Utilities | 60 | 13 | 35 | 12 | 0 | 4% | 4% |
| All sectors | 1,763 | 422 | 740 | 514 | 84 | 100% | 100% |
Sample: n = 1,763 of the 1,767 companies. Size by latest market value for the 1,338 listed companies with one recorded: large ≥ $10bn, mid $2–10bn, small < $2bn. S&P 1500 shares count the constituents of the S&P 500, MidCap 400 and SmallCap 600 (1,506 companies, lists read 25 Sep 2026). Sectors are each company's current GICS sector.
2.2Signals
| Signal | Definition | Since | Wednesdays |
|---|---|---|---|
| Short-sale share of volume, last 21 sessions | The share of each day's reported volume that was a short sale, over the last 21 sessions | 5 Aug 2009 | 880 |
| Short-sale share of volume, last 5 sessions | The same share over the last five sessions | 5 Aug 2009 | 880 |
| Days to cover | Shares held short divided by average daily volume, dated by the day the report was published | 30 Jun 2021 | 266 |
| Squeeze score | Days to cover, weighted by its own rank across stocks | 30 Jun 2021 | 266 |
| Short-side crowding | Days to cover, short-sale volume and the change in short interest, blended | 30 Jun 2021 | 266 |
| Change in short interest | The change in shares held short since the previous report | 14 Jul 2021 | 264 |
2.3Measurement
Outcome. The return from the close on the Wednesday to the close 1, 5, 21 and 63 trading sessions later, on closes adjusted for splits and dividends, compared with the same day's median stock; half of all stocks beat the median by construction, so chance is 50 in 100 on every date. No delisting return is added.
Each Wednesday the stocks are ranked on the signal. The study reports how often each fifth of that ranking beat the median stock, and the rank correlation (IC) between the signal and the return that followed.
The t-statistic uses non-overlapping dates only. A result is called proven when |t| is at least 3, it held in at least 60% of years and it is large enough to matter; with 1,470 tests across the studies, a looser bar would pass dozens by luck.
Full data and methods3Results
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.
Fifths by share of volume sold short over 21 sessions, lightest to heaviest
Fifths by share of volume sold short over 21 sessions, lightest to heaviest
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.
Percentage points, fifths by share of volume sold short, lightest to heaviest
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.
Fifths by days to cover, fewest to most, since 2021
Of the 6 measures, only the short-sale volume ratios carried a measured tendency; 4 showed no measurable edge at any horizon.
| Measure | Data since | What it showed |
|---|---|---|
| Share of volume sold short, last 21 sessions | 2009 | Proventhe next month, the next three months |
| Share of volume sold short, last 5 sessions | 2009 | Proventhe next three months |
| Days to cover | 2021 | No measurable edgeat any horizon |
| Squeeze score (days to cover, weighted by its rank) | 2021 | No measurable edgeat any horizon |
| Short-side crowding (days to cover, short volume and change, blended) | 2021 | No measurable edgeat any horizon |
| Change in shares sold short since the last report | 2021 | No 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 tilt4Robustness
Table 3 re-measures both over the next month. The short-sale share of volume kept its negative sign in 11 of 11 sectors and in 2 of the decades its history covers; among S&P 500 members on the date it was about as large as across all stocks (-0.018 against -0.017) but, with fewer stocks, short of the bar (t -2.2 against t -3.0). Days to cover showed no measurable edge overall (-0.010, t -0.9) and kept its sign in only 7 of 11 sectors. Among S&P 500 members more days to cover did come with a weaker month (-0.023, t -3.1), but on data only since 2021, in one of several subsets tested and without support from all stocks together, so it is not counted as a result.
Rank IC and its t-statistic on non-overlapping dates.
| Sample | IC | t | Wednesdays |
|---|---|---|---|
| Short-sale share of volume, last 21 sessions, next month | |||
| All stocks, whole period | -0.0168 | -3.0 | 876 |
| By decade: 1995–2004 | |||
| By decade: 2005–2014 | -0.0191 | -3.0 | 280 |
| By decade: 2015– | -0.0157 | -1.8 | 596 |
| S&P 500 members on the date | -0.0176 | -2.2 | 876 |
| Sectors with the overall sign | 11 / 11 | ||
| Days to cover, next month | |||
| All stocks, whole period | -0.0102 | -0.9 | 262 |
| By decade: 1995–2004 | |||
| By decade: 2005–2014 | |||
| By decade: 2015– | -0.0102 | -0.9 | 262 |
| S&P 500 members on the date | -0.0233 | -3.1 | 262 |
| Sectors with the overall sign | 7 / 11 |
5Limitations
- The universe and period are the same as in every study, but the data starts later: short-sale volume in 2009, and days to cover, the squeeze score, crowding and the change in short interest in 2021. Verdicts on the short-interest family therefore rest on fewer years and can change as its history grows.
- Before costs. Averages exclude trading costs, taxes and market impact.
- Same-close timing. Returns start at the close the signal is computed from; a real trade would start later.
- Survivors among smaller companies. The non-index names were chosen from companies listed when the universe was built; no delisting returns are added.
6Conclusion
Heavy short interest did not reliably forecast a squeeze, nor, measured as days to cover, a fall. The flow of short sales carried a small tilt the way the literature expects: stocks with the heaviest short-sale share of volume beat the median stock slightly less often over the next month and three months, the same way in 15 of 18 years over the month. It is a tilt across hundreds of stocks and a nudge for any one; the stock of short positions and the scores built on it are best read as context about who is positioned against a company.
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).
- Benjamini, Y. and Hochberg, Y. (1995). Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. Journal of the Royal Statistical Society, Series B 57(1).
- Grinold, R. C. and Kahn, R. N. (2000). Active Portfolio Management, 2nd ed. McGraw-Hill.
- Harvey, C. R., Liu, Y. and Zhu, H. (2016). … and the Cross-Section of Expected Returns. Review of Financial Studies 29(1).
- Shumway, T. (1997). The Delisting Bias in CRSP Data. Journal of Finance 52(1).
Appendix A. Questions readers ask
- 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.
Data availability and citation
Every figure in this paper is quoted from one frozen snapshot, published as JSON with the sample description, the robustness results and the test counts. The same snapshot feeds the research console, so the two cannot disagree. The paper and the snapshot are licensed under CC BY 4.0: share and adapt them with credit to Opulence Alpha and a link to this page.
Mehmet Saruhan (2026). Does short interest predict stock returns? Opulence Alpha Research, Opulence Alpha Studies, 25 Sep 2026. https://opulencealpha.ai/studies/short-interest
Research, not advice. A measured tendency across hundreds of stocks is a nudge for any one of them, never a forecast of its price.
More evidence
All studiesWhere this is read on the S&P 500 company pages
The company pages show each S&P 500 company's readings from the latest close; a ranking lists every company by one measure.