Hi all, I am wanting to backtest a momentum strategy on the consituents of an index. However i am concerned about the survivorship bias in this sort of backtest. Where can i find data that has the historical index consituents of a particular index say S&P 500 or Nasdaq 100? What would be a better way? any suggestions?
It is an interesting topic you mentioned. If you define your universe as the index constituents only, effectively you are applying a filter. If you handle stocks exiting the benchmark in your backtest in a proper manner, there is no reason you should have any survivourship bias. One possible way of handling this correctly is to close out a position compulsorily at the prevailing price, if a stock exits the index. Replicating this in real life trading is also not very hard. Usually changes in index constituents are announced in advance.
Our in-house Blueshift platform follows the method mentioned above. The historical S&P 500 constituents data is not publicly available. However, on this platform, you will not need it to evaluate your backtest. The NYSE minute dataset automatically tracks the S&P 500 stocks and the platform will auto-close any positions in stocks which have exited the benchmark. The auto-close happens a day after the stock exits the benchmark.
We do not maintain such automatic dataset for NASDAQ 100 yet.
Prodipta
Thank you