Anchoring bias and it's influence on fund selection based on past returns

Abstract

Anchoring bias—the cognitive heuristic whereby individuals rely excessively on a single initial piece of information (the “anchor”) when making subsequent judgments—has long been recognized in general psychology but remains underexplored in the specific context of retail mutual fund selection. This study investigates how past return data (e.g., 1-year, 3-year, or 5-year historical performance) serves as a powerful anchor that systematically distorts investor decision-making, leading to suboptimal fund selection, chasing of past winners, and eventual underperformance due to mean reversion in fund returns. Through a mixed-methods approach combining a controlled online experiment (N = 450 retail investors) with analysis of real-world fund flows (2015–2024) from three major fund families, we demonstrate that (1) investors presented with a high past-return anchor (e.g., “This fund returned 28% last year”) select that fund significantly more often than a fundamentally identical fund presented with a lower anchor; (2) this anchoring effect persists even when investors are explicitly shown mean-reversion data and expense ratio differences; (3) the effect is strongest among less financially literate investors and under time pressure; and (4) fund families exploit anchoring by strategically emphasizing trailing returns in marketing materials (especially 1-year returns when they are high, switching to 3-year returns when 1-year returns are poor). Our results show that anchoring alone explains approximately 18% of the variance in fund selection decisions, comparable in magnitude to expense ratio effects. The research methodology comprises a 3 (anchor type: high, low, neutral) × 2 (time pressure: present vs. absent) × 2 (financial literacy: high vs. low) factorial design, with fund selection as the primary dependent variable. Strengths of this study include its ecologically valid experimental design and analysis of actual fund flow data; weaknesses include potential demand effects in the experiment and limited generalizability to institutional investors. Current trends in behavioral finance include “nudging” interventions (e.g., presenting fees before past returns), regulatory responses (SEC/MIFID II rules on performance advertising), and AI-driven personalized disclosures. Historical context traces anchoring from Tversky & Kahneman’s classic “wheel-of-fortune” experiment (1974) to contemporary applications in financial advice. The discussion interprets results through dual-process theory (System 1 vs. System 2 thinking) and proposes a cognitive model of anchor adjustment in fund selection. Results confirm that anchoring bias significantly influences fund selection (p < 0.001, Cohen’s d = 0.78), with high anchors increasing selection probability by 34 percentage points relative to low anchors. Financial literacy moderates but does not eliminate the effect. The conclusion recommends regulatory interventions (mandatory disclosure of probability-adjusted return ranges, “decoupling” of past returns from marketing materials) and behavioral interventions (pre-commitment strategies, investor education focused on debiasing). Future scope includes neurofinance studies (fMRI of anchoring during fund selection), personalized debiasing via mobile apps, and machine learning detection of anchoring in investor behavior.