Anticipation bias is often treated as an abstract behavioral trait. Investors observe data, form expectations and make decisions. But in reality, these biases shape real trades. Some investors extrapolate from recent performance. Others react in the opposite direction, and these differences affect which stocks they buy, which stocks they sell, and how they move money into stocks. This paper presents a new perspective. Prediction bias is not just about stated beliefs. It appears directly in the individual stock selection.
Extrapolators and contrarians: Forecast bias and stock trading of individual investors
- Steffen Andersen, Stephen G. Dimmock, Kasper Meisner Nielsen, Kim Peijnenburg
- Journal of Financial Economics, 2026
- A version of this paper can be found here here
- Want to read our summaries of academic finance papers? Check out ourAcademic Research OverviewCATEGORY
Key academic insights
Forecast bias affects real stock markets
The paper relates a laboratory measure of forecast bias to actual administrative trade data. This allows the authors to move beyond surveys and show that biased expectation formation affects stock selection in the real world. Investors don’t just report different beliefs. They act on them.
Extrapolators buy recent winners
Investors with stronger extrapolation bias tend to buy stocks with higher past returns. A one-standard deviation increase in forecast bias is associated with buying stocks whose past one-year return is about three percentage points higher. This supports the idea that extrapolators follow recent performance.
Contrarians buy relatively weaker performers
Investors with a negative forecast bias behave differently. They are more likely to buy stocks with lower past returns compared to extrapolators. This shows that the same return history can lead different investors to make opposite choices depending on how they process the information.
Sales depend more on capital gains than on past returns
For sell decisions, the key performance measure is not the stock’s past one-year return. It is the investor’s own capital gain. Extrapolators are less likely to sell stocks with high capital gains, suggesting that the benefits of personal experience are particularly salient when deciding what to sell.
Forecast bias also affects flows in stocks
The paper finds that investors with higher forecast bias increase stock allocations following strong market returns. However, forecast bias does not strongly predict when investors trade. It matters more to what they do since they are already trading.
The behavior is consistent for buying and selling
Investors who buy recent winners also tend to sell recent losers. This pattern appears not only in the experimental sample but also in the full universe of Danish investors. This consistency supports the idea that a common behavioral mechanism drives buying and selling decisions.
Practical applications for investment advisors
Diagnose callback behavior
Customers may not just be “aggressive” or “conservative”. Some may systematically extrapolate recent performance into the future. This can lead them to chase winners and hold or sell positions based on biased expectations.
Separate beliefs from preferences
A client’s portfolio choices may reflect biased forecasts rather than true risk tolerance. Advisors must distinguish between what clients want and how they interpret recent performance.
Use disciplined rebalancing rules
Because forecast bias can affect both buying and selling, systematic rebalancing can help reduce emotionally driven or performance-driven security selection.
Pay attention to individualized capital gains
For sales, investors react strongly to their gains and losses. Advisors must recognize that tax amounts, purchase prices and unrealized gains can shape behavior as much as market-wide performance.
How to explain this to customers
“Two investors can look at the same stock chart and come to very different conclusions. One can see recent gains and expect more gains. Another can see the same growth and expect a reversal. This paper shows that those forecast differences matter. They affect which stocks people buy, which stocks they sell, and how they move money in the market. A disciplined process helps prevent portfolio performance from being decidedly biased.”
The most important chart from the paper
Figure 4: his figure shows the results of the regressions of the relationship between the extrapolation purchases ratio and the extrapolation sales ratio. A rollover purchase is defined as a stock purchase following a positive excess return over the previous 12 months and a rollover sale as a stock sale following a negative capital gain. The extrapolation buy (sell) ratio is the number of extrapolation buys (sells) divided by the total purchases (sales) calculated during the period 2013 to 2021. The sample is limited to investors with at least two buys, two sells and ten total trades. Panel A includes all subjects in our experiment conditional on meeting the constraint. Panel B includes all Danish investors subject to meeting the restrictions. The number of observations in panels A and B are 417 and 316,770, respectively. The solid line is the linear regression line and the dots represent 25 bins of observations.
Results are hypothetical results and are NOT an indication of future results and do NOT represent returns actually achieved by any investor. Indices are not managed and do not reflect management or trading fees, and one cannot invest directly in an index.
ABSTRACT
We test whether forecast bias affects individual investors’ stock trading by combining measures of bias from laboratory experiments with administrative trading data. Forecast bias is positively related to past excess returns of the stocks purchased: Compared to the opposite, extrapolators buy stocks with higher past returns. Forecast bias is negatively related to capital gains on shares sold. Forecast bias also explains investor heterogeneity in the relationship between market returns and net flows. Taken together, forecast bias provides a unifying mechanism through which different measures of performance—past stock returns, capital gains, and past market returns—shape relevant buy, sell, and net flow decisions.
Extrapolators and contrarians: Forecast bias and stock trading of individual investors originally published in Alpha Architect. Please read the Alpha Architect FINDINGS at your convenience.
