Survivorship Bias

Survivorship bias is one of the most pervasive and underappreciated statistical distortions in financial markets, affecting how investors perceive the historical performance of mutual funds, stock market strategies, and investment strategies based on back-tested data.

In the Indian mutual fund context, survivorship bias is embedded in the performance databases used by most investors and comparison platforms. When a mutual fund scheme is merged into another (often due to regulatory rationalisation under SEBI's mutual fund categorisation rules of 2017–18, which eliminated many duplicate schemes) or wound up due to poor performance, its track record disappears from the live fund universe. Databases that show the 10-year average returns of equity mutual funds implicitly only count the funds that survived the full 10-year period. Funds that were merged or closed — disproportionately the underperformers — are excluded, causing the average historical return figure to appear higher than what the average rupee actually earned.

SEBI's mutual fund regulations have resulted in significant consolidation, with well over 100 schemes being merged into other schemes between 2017 and 2020 alone. An honest assessment of actively managed fund performance must include the returns of discontinued schemes to avoid the survivorship bias distortion.

In equity market history, survivorship bias makes the market appear to have always been populated by quality companies. The companies that went bankrupt, were delisted for non-compliance, or saw their stocks become worthless (think IL&FS, DHFL, Yes Bank's pre-recapitalisation period, Jet Airways) are removed from indices and performance calculations. Long-run stock market return databases computed from index data implicitly include the benefit of index reconstitution — losing stocks are removed and replaced by winners — a form of institutional momentum strategy that retail buy-and-hold portfolios do not automatically replicate.

For back-tested investment strategies, survivorship bias is especially pernicious. A strategy back-tested on current index constituents (say, all Nifty 500 stocks as of today) will produce artificially elevated returns because it includes only companies that were successful enough to remain listed and prominent — it has no historical exposure to the companies that failed during the same period. Properly constructed back-tests use point-in-time databases that include all companies that existed at each historical date, including those that subsequently failed. The gap between survivorship-biased and point-in-time back-tests can range from 3–7% annually, fundamentally changing the viability assessment of a strategy.