Graveyard / Equity events
The UK short-seller register
Does following UK short disclosures work?
The idea
UK holders of a net short position of 0.5% or more must disclose it. Informed shorts should predict declines, and a cover should predict a rise. Short after a new disclosure and go long after a cover.
How we tested it
| Data | The UK Financial Conduct Authority's historic register: 15.6 thousand new-short events and 14.9 thousand covers. Of 292 securities searched, 173 could be priced and 119 (delisted) could not. |
|---|---|
| Period | 2012 to 2026. |
| Costs | 80 to 84 bp per round trip, mostly UK stamp duty of 0.5%. |
| Holdout | Held out; the exact start date is not stated in the note. |
| Control | A random twin on the same stocks. |
Result
| Short after new disclosure, gross / net | +36 bp / −48 bp (twin p 0.97) |
|---|---|
| Long after cover, gross / net | +3 bp / −77 bp (twin p 0.002) |
| Costs | 80 bp or more per round trip |
Why it failed
Names that new shorts targeted rose rather than fell, so the short earned a gross loss against its direction. The cover signal was worse than random. At 80 bp of costs neither direction has room. Delisted names are missing, which is serious for the short leg.
What would make it worth retesting
Survivor-free UK prices including delisted names, plus a rule that works net of 80 bp, or a test that sizes by the track record of the individual holder.
Test id: ukshortreg-equities-event-v1. Past results of a backtest say nothing certain about the future.