Stocks that have lost value should also be sold according to pre‑set rules — don’t try to even out costs by buying more

37. Stocks that have lost value should also be sold according to pre‑set rules — don’t try to even out costs by buying more

Cost: No cost at all. The hard part is admitting the loss and actually selling the stock.

In plain terms: Human nature drives us to sell winners and hold losers. In seven years of data from 10,000 U.S. accounts, stocks that rose were sold about 50% more often than those that fell. The same investors also tended to add to positions in losing stocks. Yet over the following year those unloved losers underperformed the sold winners by an average of 3.4 percentage points.

Benefit: This study examined every trade made by 10,000 accounts at a U.S. discount broker from 1987 to 1993. The proportion of profitable stocks sold (PGR) was 0.148, while the proportion of losing stocks sold (PLR) was 0.098 — a difference of 0.050 with a t‑value exceeding 35. In other words, rising stocks were sold more than 50% more often than falling ones. Investors also bought losing stocks more frequently: the re‑purchase rate for losers was 0.135 versus 0.094 for winners, again with a strong statistical significance. The data also contradicted the belief that “stuck stocks will eventually bounce back.” Over the next year, the excess returns of sold profitable stocks exceeded those of kept losing stocks by 3.4 percentage points. These excess returns represent performance above the market index. As a concrete illustration, selling $1,000 worth of a losing stock instead of a winning one would yield roughly 4.4% more return in the following year; this figure includes tax savings from the early sale under U.S. tax rules (1987‑1993).

Evidence grade: A

Sources: Odean T (1998). Are Investors Reluctant to Realize Their Losses? The Journal of Finance 53(5):1775-1798. <https://doi.org/10.1111/0022-1082.00072>

Notes: This point addresses the decision “to sell or not to sell”; point 15 (“avoid excessive trading”) deals with how often to trade — the two are not contradictory. Following a pre‑defined rule actually reduces the number of trades. Mathematically, buying more shares does lower the average cost per share, but treating this as a way to recover losses is misguided; it merely increases the amount of capital tied up in a single stock (see point 19, “don’t bet everything on one stock”). The tax calculations apply only to the U.S. system and cannot be directly applied in China; all other conclusions remain valid. The data originate from U.S. accounts in the 1980s‑1990s; trading costs and tax regimes differ for A‑share investors. This information is not intended as investment advice.