Adjusting Price History for a Rights Issue

In a rights issue, a company offers existing shareholders the right to buy new shares at a discount to the current market price, typically to raise capital without a full public offering. Because the new shares are cheaper than the market price, the stock's fair value drops mechanically on the ex-rights date to reflect the diluted, blended value of old and new shares together — exactly like a dividend or a stock split, this price drop is not a real loss for anyone who takes up their rights, and any backtest or return series built on unadjusted prices would show a fake crash on that date.

The adjustment factor is built from the ratio of shares held to new shares offered ($N$) and the subscription price ($S$) relative to the cum-rights market price ($P$). The theoretical ex-rights price is $\text{TERP} = \frac{N \times P + S}{N + 1}$, and every historical price before the ex-rights date gets multiplied by the adjustment ratio $\text{TERP} / P$ to bring old prices down to the same basis as new, post-rights prices. Concretely: if a stock trades at $100 cum-rights and the company offers one new share for every four held at a subscription price of $80, then $\text{TERP} = (4 \times 100 + 80)/5 = 96$, so every price before the ex-date gets scaled by $96/100 = 0.96$ — a 4% adjustment, even though nothing fundamental happened to the company that day.

Skipping this adjustment is a common source of a phantom "crash" appearing in unadjusted price feeds around rights-issue dates, and it silently corrupts any return calculation, volatility estimate, or backtest signal that spans the ex-rights date.