From Kaplan online:
Both survivorship bias (failure to include funds that ceased to exist over the sample period) and backfill bias (adding past results to historical index returns for funds that are added to the index) tend to bias alternative investment index returns upward.
Am I to understand that survivorship bias is when an index does not include funds that stopped operating during a specific period of time?
Further, backfill bias is when an index creator adds past results to the historical return of the same index? i.e. he just changes the index’s return by adding more observations?
Yes, and yes.
In survivorship bias, poorly performing funds drop out, so the index shows only surviving (i.e., well-performing) funds, so there’s an upward bias.
In backfill bias, it’s more likely that the creator will backfill funds with good performance than those with poor performance, so there’s an upward bias.
Excellent explanations as always.
From the return perspective: Data from only surviving firms are used to compute the return. This can create an overestimation of the returns since the data does not account for the failed companies。
However, from the risk perspective: Only surviving firms are included, so the risks are underestimated, which is quite contradictory to the above conclusions. Understand that risks and returns are generally positively related… Was wondering if you could kindly explain the rationale behind this.