A company announces a 1:1 bonus issue. On the ex-date, the share count doubles and the quoted price roughly halves. Nothing happened to the business and nothing happened to your wealth.
Feed that raw price series into a momentum rule and the strategy sees a 50% single-day loss. It sells. The rule did exactly what you told it to, against data that lied to it.
The main types
| Action | Effect on the quoted price | Effect on your holding |
|---|---|---|
| Stock split | Falls in proportion to the split ratio | Share count rises proportionally |
| Bonus issue | Falls as new shares are issued free | Share count rises, value unchanged |
| Dividend | Falls by roughly the dividend on the ex-date | Cash is received separately |
| Rights issue | Adjusts for the discounted subscription price | You must subscribe or let the entitlement lapse |
| Merger or demerger | The listed entity or entities change | Holdings are exchanged for new securities |
| Symbol or name change | No price effect | Historical series must be relinked across the old and new ticker |
| Delisting or suspension | Trading stops | The position becomes unexitable |
Why adjustment is necessary
Any rule that reads a return, a moving average, a high or a ratio is reading a price series. If that series contains an unadjusted split or bonus, every derived value spanning the ex-date is wrong, and the error is large enough to trigger or suppress trades.
Corporate actions applied to prices in saral.money backtests are splits and bonus issues, which covers the discontinuities large enough to break a technical rule.
An honest limit
Dividends are not accrued into returns here. Prices are adjusted for splits and bonuses, but dividend income is not added to the equity curve.
The consequence is specific and worth stating plainly. A backtest of a high-dividend or income strategy understates what you would have earned, because the yield is missing from the result while the ex-date price fall is not. Read those backtests as price return, not total return, and treat any comparison against a total-return benchmark as unfair to your strategy.
Caveats
Back-adjustment rewrites history by design. The adjusted price for a date in 2015 is not the price anyone saw in 2015, which is fine for computing returns and misleading if you are reasoning about round numbers or price levels.
Demergers are the hardest case, because the pre-event company and the post-event companies are not the same asset, and any single adjusted series through that event is an approximation. When a strategy’s result depends heavily on one name that went through a demerger, check that leg by hand.