38 Pages Posted: 7 Oct 2003
Date Written: July 8, 2003
In this article we examine the backfill bias or instant history bias for hedge funds using additional information from the Tass database. This is information about the exact date a hedge fund starts reporting to Tass. Using this information we are able to reveal the length of the instant histories. We find these to be just over 3 years on average. This number is far greater than previously documented. More than half of the recorded returns in the database are backfilled. The magnitude of the overall backfill bias is about 4 percent per annum on average. Again this number exceeds all previous estimates of the backfill bias we are aware of. We elaborate further across different time periods styles. Next, we eliminate backfilled returns and use survivorship free data to create a universe in which we could invest in real time. We introduce an investor who invests an equal amount in each fund that is in the universe. Conditional on this investment strategy our results indicate that the backfill bias is underestimated, and has a substantial downward effect on the returns across most hedge fund styles and is consistent over time for the whole sample. We have no reason to believe that our conclusions are limited to the Tass database.
Keywords: Backfill bias, Hedge funds, Performance Persistence, Self-selection bias
JEL Classification: G10, G11, G23, G29
Suggested Citation: Suggested Citation
By Bing Liang