Graveyard / Equity events / US stocks
Shorting IPO lock-up expiries
Do IPO stocks fall when the 180-day lock-up ends?
The idea
After an IPO, insiders are usually barred from selling for 180 days. When the lock-up ends, they can sell, so the price may drop. Short a few days before expiry and cover just after.
How we tested it
| Data | IPOs with a first-month median close of at least $5 and first-three-month median dollar volume of at least $2M, with SPAC-like names excluded. Day 0 is the first session on or after IPO plus 180 calendar days. Short from the close of day −3 to the close of day +2, long an equal dollar amount of IWM, 5 slots of 20% each. |
|---|---|
| Period | July 2005 to June 2026, split at July 2022 (538 events). |
| Costs | 10 bp per side on the stock, 1 bp per side on IWM, 20% a year borrow. Stress case: 25 bp per side and 60% a year borrow. |
| Holdout | In-sample and out-of-sample Sharpe. |
| Control | No random-entry control was run. |
Result
| Events | 538 |
|---|---|
| Stock minus IWM over days −2 to +2 | +0.42% (median +0.32%, t 1.43) |
| Sharpe, no cost (full / in / out) | −0.26 / −0.17 / −0.68 |
| Sharpe, base cost (full / in / out) | −0.67 / −0.60 / −0.98 |
| Sharpe, stress cost (full / in / out) | −1.40 / −1.37 / −1.52 |
| Max drawdown, no cost / base / stress | −46% / −68% / −90% |
| Profit factor, no cost / base / stress | 0.83 / 0.65 / 0.41 |
Why it failed
The sign was wrong before any cost: the stocks went up around expiry, not down. Lock-up dates are widely published, so the drop looks priced in. Delisted IPOs are missing from the data, but over five sessions that bias is worth a few tenths of a percent at most, not enough to turn +0.42% into the 1.5% or more needed to pay for costs and borrow.
What would make it worth retesting
Not without a point-in-time IPO list that includes delisted names and the exact lock-up terms.
Past results of a backtest say nothing certain about the future.