What is recency bias and why does it persist?
Recency bias is the tendency to overweight recent observations relative to the full historical record when forming estimates and expectations. In statistical terms, it is a kind of nonstationary prior: the investor implicitly assumes that the most recent period is more informative about the future than longer historical patterns.
Recency bias persists because recent events are more vivid, easier to recall, and feel more causally relevant than distant ones. The availability heuristic -- the tendency to judge probability by how easily examples come to mind -- drives recency bias. A market crash that happened last year is far more available to recall than a crash that happened 40 years ago, so the recent crash receives more weight in probability estimates.
Recency bias is amplified by experience. Investors who lived through the 2008 financial crisis have visceral recency bias toward financial sector risk that investors who only read about it do not share. This creates systematic age-related differences in risk tolerance and portfolio construction: older investors who experienced multiple crashes maintain higher cash balances and more defensive allocations than historical base rates would justify.
Recency bias is not always wrong. Recent data can legitimately be more informative than distant data when structural conditions have changed. A company that recently shifted its business model has financials that are more predictive of future results than the pre-transition financials. The problem is distinguishing genuine regime changes from temporary deviation -- recency bias causes investors to treat temporary deviations as regime changes.
How recency bias affects asset allocation decisions
Return chasing is the most directly observable consequence of recency bias in asset allocation. Investors who observe a strong run in an asset class increase their allocation to it precisely when mean reversion toward long-run expected returns is most likely. Flows into equity funds peak at or near market tops; flows into bond funds peak at or near interest rate peaks. In both cases, investors are chasing the recent return pattern.
Risk tolerance shifts based on recent market conditions are a related effect. After a prolonged bull market, investors report higher risk tolerance, accept more equity exposure, and reduce cash and bond allocations. After a significant drawdown, investors report lower risk tolerance and shift to defensive positioning -- often at the time when expected equity returns are highest and the case for equity is strongest.
Recency bias in correlation estimation affects portfolio construction. Correlations between asset classes change over time, and recency bias causes investors to anchor to correlations that were observed recently. Correlations between equities and bonds that were negative during the 2000s and 2010s shifted toward positive in 2022. Investors who built portfolios expecting negative equity-bond correlations experienced less diversification than their models predicted because they were anchoring to a recent correlation regime that had ended.
Extrapolation of recent earnings growth rates into future earnings estimates is a specific form of recency bias in equity analysis. Companies that have compounded earnings at 25% per year for five years rarely maintain that rate indefinitely, but analysts consistently project recent growth rates forward beyond the point that competitive economics would support.
Correcting for recency bias in investment analysis
Base rate anchoring with long-run data provides the primary correction. Before extrapolating a recent trend, identify the longest available data series for the same variable and use the full-period base rate as the prior. Equity return base rates over 100+ years provide a more stable anchor than the last decade. Mean reversion tendencies in valuations and rates are more visible over long periods than over short ones.
Cycle awareness involves explicitly identifying where current market and economic conditions sit within historical cycles. A measure of the credit cycle, the equity valuation cycle, and the interest rate cycle all provide context that is invisible when looking only at recent data. Recency bias prevents investors from asking "has this happened before?" and instead causes them to treat current conditions as novel.
Structured scenario analysis that includes historical analogues reduces recency bias. Before projecting current conditions forward, identify three or four historical periods with similar characteristics and examine what happened next. The range of outcomes across analogues provides a more calibrated distribution than extrapolation from the most recent period alone.
Equal-period weighting in correlation and volatility estimation deliberately counters the recent-data bias built into exponentially weighted moving average models. An equally weighted 20-year estimate of equity-bond correlation gives the same weight to 2005 as to 2024; an exponentially weighted estimate gives much more weight to recent periods. Both have uses, but over-reliance on exponentially weighted estimates embeds recency bias into the risk model.
Frequently asked questions
What is recency bias in investing?
Recency bias is the tendency to give excessive weight to recent events and observations relative to the full historical record when forming expectations. Investors with recency bias extrapolate recent market returns, volatility, correlations, and business conditions forward beyond what the longer data record would support. After a strong bull market, they expect more bull market; after a drawdown, they expect more drawdown -- both overweighting the recent sample relative to the long-run base rate.
How does recency bias affect asset allocation decisions?
Recency bias drives return chasing: investors increase allocations to recently strong-performing asset classes when mean reversion suggests they should reduce them. It causes risk tolerance to rise after bull markets and fall after drawdowns -- both procyclical to market movements. It distorts correlation estimates by anchoring to recent correlation regimes that may not persist. And it causes earnings growth extrapolation beyond what competitive economics support for specific companies.
How do you correct for recency bias in investment analysis?
Key techniques include: base rate anchoring with long-run data before extrapolating any trend; cycle awareness -- identifying where current conditions sit in historical market, credit, and valuation cycles; structured historical analogue analysis that finds comparable prior periods and examines the range of subsequent outcomes; and equal-period weighting in statistical estimation to avoid over-representing recent data in volatility and correlation estimates. None eliminates recency bias, but each provides a competing anchor grounded in longer-run history.
What is the relationship between recency bias and market cycles?
Recency bias amplifies market cycles. When recent returns are strong, recency bias causes investors to increase risk exposure near peaks; when recent returns are poor, recency bias causes investors to reduce risk near troughs. This procyclical behavior creates self-reinforcing dynamics: return chasing pushes prices above fundamental value in bull markets and selling pressure pushes prices below fundamental value after corrections. Recency bias is one of the behavioral mechanisms that drives mean reversion in asset returns -- the overweighting of recent trends creates the overextension that then corrects.