[论文解读] Survivorship Bias in Emerging Market Small-Cap Indices: Evidence from India's NIFTY Smallcap 250

Harjot Singh Ranse|arXiv (Cornell University)|Mar 19, 2026

Financial Markets and Investment Strategies被引用 0

一句话总结

这篇论文使用 1,437 只股票（2016–2025）在印度的 NIFTY Smallcap 250 指数上量化存活偏差，显示仅存活股回测对年化收益率高估 4.94 个百分点、夏普比率高估 0.097。它记录了高换手率以及下市/退网/降级成分，并提出基于价格-成交量的历史重构用于无存活偏差的回测。

ABSTRACT

This study quantifies survivorship bias in India's NIFTY Smallcap 250 index using a dataset of 1,437 stocks over nine years (2016-2025). By reconstructing historical index composition through market capitalization ranking and comparing equal-weight portfolios of current constituents versus all historical members, I show that survivor-only backtesting overstates annual returns by 4.94 percentage points (23.3%) and Sharpe ratios by 0.097 (9.1%). The analysis reveals an 82.5% turnover rate, including delisted (16.1%), graduated (33.1%), and demoted stocks (33.2%), with all categories contributing to bias. Using bhavcopy data that includes delisted securities, the reconstruction achieves 100% accuracy for current constituents and an estimated 85-90% accuracy historically. These findings highlight that survivorship bias is materially larger in emerging market small-caps and that using only current index members can significantly overstate strategy performance. Briefly, the methodology reconstructs historical index membership using a price-volume-based ranking approach to enable survivor-free backtesting.

研究动机与目标

• Quantify the magnitude of survivorship bias in India’s NIFTY Smallcap 250 index.
• Assess how survivor-only backtesting differs from survivor-free backtesting over nine years.
• Identify the contribution of different stock status categories (delisted, graduated, demoted) to bias.
• Develop a historical reconstruction method to enable accurate backtesting using non-survivor data.

提出的方法

• Reconstruct historical index composition by market-cap ranking to emulate backtests with and without survivorship.
• Compare equal-weight portfolios of current constituents against all historical members.
• Use bhavcopy data including delisted securities to improve historical accuracy (target ~100% current constituents, 85–90% historical).
• Quantify differences in annual returns and Sharpe ratios between survivor-only and survivor-free frameworks.

实验结果

研究问题

• RQ1 What is the magnitude of survivorship bias in India’s NIFTY Smallcap 250 when using survivor-only backtests?
• RQ2 How do equal-weight portfolios formed from current constituents compare to portfolios that include all historical members in performance metrics?
• RQ3 What proportion of historical constituents are delisted, graduated, or demoted, and how do these categories contribute to bias?
• RQ4 Can a price-volume-based historical reconstruction method achieve survivor-free backtesting accuracy?

主要发现

• Survivor-only backtests overstate annual returns by 4.94 percentage points (23.3%).
• Survivor-only backtests overstate Sharpe ratios by 0.097 (9.1%).
• Turnover is 82.5%, with delisted (16.1%), graduated (33.1%), and demoted stocks (33.2%) contributing to bias.
• Historical reconstruction with bhavcopy data achieves near 100% accuracy for current constituents and ~85–90% accuracy historically.