Recency Bias
ES: Sesgo de Recencia PT: Viés de Recência
The tendency to give disproportionate weight to what has just happened when predicting the future. Traders overestimate the persistence of the current trend: bullish at the tops and bearish at the bottoms. Everyone invests in growth in 2021 and in value in 2022.
What Recency Bias Is
Recency bias is the psychological tendency to give excessive weight to recent events when predicting the future, underweighting historical patterns and long-term trends. Put bluntly: the trader extrapolates whatever has just happened indefinitely. If the market rises, they expect it to keep rising; if it falls, to keep falling; if technology leads, they assume it always will; if emerging markets lag, they abandon them. Every one of those extrapolations ignores the historical frequencies.
The bias operates across every horizon. Intraday, a 3% fall leads you to expect more of the same and sell, frequently at the session low. Weekly, a 10% rally invites you to buy just before the top. Monthly, a sector up 20% attracts buyers who arrive late to the rotation. And annually, the technology euphoria of 2021 drew late buyers who lost half their capital in 2022.
Academic research documented it through Kahneman and Tversky’s availability heuristic: what is easily recalled — and the recent is recalled better — seems more probable. Benartzi and Thaler showed in 1995 that it also affects long-term investors and significantly reduces their returns.
The historical examples are eloquent. Japanese investors in 1989 extrapolated the Nikkei’s rise to a peak of 39,000 points followed by three decades of bear market. Americans in 1999 extrapolated the technology boom before a 78% fall in the Nasdaq. In 2006 it was taken as given that property prices only go up. And in 2021 the slogan was that bitcoin would reach $100,000, shortly before it fell 77%. The pattern is universal: extrapolating the recent trend leads you to buy at the tops and capitulate at the bottoms.
Evolutionary and Neurological Roots
The bias has deep evolutionary roots. Early humans found pattern recognition useful: if there had been predators nearby lately, it paid to expect more; if food was abundant, continuity was a reasonable assumption. It worked in stable environments. Financial markets are nothing of the sort: they revert to the mean over the long run and undergo periodic structural changes. The same cognitive mechanism that kept our ancestors alive costs the modern investor money.
At the neurological level, the amygdala and hippocampus interact during memory encoding, so recent emotionally charged events are burned in strongly while unemotional historical data fades. The result is that the brain’s probability estimate disproportionately weights recent emotional experiences.
The availability heuristic explains a specific effect: after a prolonged bull market, investors underestimate the frequency and severity of crises, because they remember them poorly; and during a crisis they overestimate the difficulty of recovering, because they forget how many times it has happened.
Media amplification makes the problem worse. Continuous financial coverage emphasises the immediate — what the market is doing today, the week’s most active stocks — and historical context almost never appears. Social media is even more recency-biased, because its algorithms reward the conversation of the moment.
Survey data confirms the pattern: after six bullish months, more than 70% of retail investors expect continuity; after six bearish months, more than 60% expect the same. The historical reality is that mean reversion is the norm and heavily extended trends are the exception. Positioning reports also show that professional funds are usually positioned on the opposite side to retail.
How It Shows Up in Trading
The bias translates into ten concrete behaviours.
The first is entering trends late: professional momentum strategies use rules, not emotional extrapolation, and the retail trader arrives once the move has already happened. The second is missing mean reversion, which happens frequently but looks improbable under the bias.
The third is sector rotation errors: abandoning a lagging sector just before it rotates in your favour, and entering the leading one just before it rotates against. The fourth is mispricing volatility: after calm periods like 2017 people sell volatility aggressively, and after the spikes they overpay for protection.
The fifth is compulsively switching strategy, abandoning the one that has lagged for six months just as it starts working. The sixth is misjudging risk, underestimating it after calm and overestimating it after crisis. The seventh is anchoring to recent prices, which makes a stock look expensive because it has risen, regardless of its fundamentals.
The eighth is earnings expectations: three quarters of beats generate the expectation of a fourth, and the disappointment comes as a surprise. The ninth is economic forecasting, extrapolating recent growth and inflation, which is why recessions surprise people again and again. And the tenth is sizing errors: after a winning run people increase size out of confidence, and the next large loss takes the recent gains and then some.
Countermeasures
Fighting it requires systematic approaches, and there are ten that work.
The first is the practice of historical context: before deciding, asking how often this has happened over fifty years. The second is contrarian positioning at sentiment extremes, which investor surveys, options ratios and volatility levels let you quantify. The third is fundamental discipline, evaluating assets by their metrics against their historical averages rather than by their recent behaviour.
The fourth is spacing decisions out over time, rebalancing by calendar rather than reacting. The fifth is long historical testing, evaluating any strategy across several market regimes and not just the last bull cycle.
The sixth is filtering other people’s opinions, bearing in mind that analyst recommendations are heavily recency-biased and that buy ratings cluster near the tops. The seventh is having a written plan defining decisions in advance.
The eighth is Buffett’s rule of being fearful when others are greedy, which is nothing other than explicitly anti-recency positioning. The ninth is following sentiment indicators — the volatility index, the put-call ratio, retail sentiment surveys, positioning indicators — and treating their extreme readings as turning signals. And the tenth is cultivating patience: some of the best trades consist of waiting for the reversal.
In options the bias directly affects the price of implied volatility: after calm periods IV is too low and it pays to buy cheap protection; after crises it is too high and it pays to sell premium. Systematic contrarian positioning on IV tends to be profitable.
The central idea is that valuations revert to the mean over the long run. Trends extend further than expected, but they eventually turn. Recency bias leads you to trade exactly against that reversion, buying tops and selling bottoms, and the discipline to trade with it is one of the core competencies of professional management.
Recency Bias vs Legitimate Trend Following
The process distinguishes them, even if the outcome can look similar.
| Aspect | Recency bias | Legitimate analysis |
|---|---|---|
| Emotional extrapolation | Systematic indicators | |
| Late, near the highs | Early, on a confirmed signal | |
| None | Criteria defined in advance | |
| Grows with the gains | Constant, by rule | |
| Dismissed | Acknowledged and monitored | |
| Ignored | Consulted regularly |
Frequently Asked Questions
How do I distinguish recency bias from legitimate trend following?
Recency bias, by contrast, is emotional extrapolation, enters late and near the tops, increases size in line with recent gains, and dismisses the possibility of reversal.
The decisive test is simple: can you explain under what specific conditions you would exit the position? Trend following has exit signals defined in advance; recency bias has none at all.
Why do professional investors often bet against the trend?
The basis is mathematical: valuations revert to the mean over long periods, and extended trends generate overvaluation or undervaluation. Taking the other side captures that reversion.
The retail investor, biased toward the recent, takes the opposite position and ends up providing liquidity — and losses — to the professional.
How do sentiment indicators help?
A volatility index above 40 reflects extreme fear and usually coincides with market bottoms; below 12 it reflects complacency and often precedes corrections. A bullish percentage above 50% in retail surveys indicates euphoria. An elevated put-call ratio signals fear. And positioning reports let you compare how professionals are placed against retail.
The way to use them is as a contrarian indicator: when readings reach extremes, position on the opposite side. Historical studies show systematic profitability in that kind of positioning.
Does recency bias affect long-term investors?
The classic errors are three: buying international equities just as their period of outperformance ends; selling bonds while rates rise, just before their recovery; and abandoning the value style at the peak of growth’s outperformance.
The solution is a strategic allocation with periodic rebalancing, annual or quarterly, which prevents reactive changes. Buffett’s advice for his wife’s inheritance — a simple, largely indexed allocation — responds to exactly this logic: a simple structure eliminates the decisions subject to the bias.
How do I exploit this bias trading options?
After calm periods, with the volatility index below 12, it pays to buy cheap protection: historically those levels do not hold for long. After volatility spikes, above 30, it pays to sell premium, since the index usually returns to its 15-18 average within months.
The corresponding structures are iron condors when implied volatility is elevated and long straddles when it is compressed. It demands patience, because the exact moment of reversion is unpredictable.
The same logic applies to individual stocks: implied volatility around an event frequently overestimates the actual size of the move, and selling that inflated expectation is profitable on average.