For example, a manager may stop reporting either because of very poor performance or because the manager has had strong performance and is no longer raising assets. Some of the more significant concerns that should be understood include: II Survivorship bias. Hedge fund managers are dropped from an index if they stop reporting to the index provider. It is clear that periods of nonreporting can coincide with periods of significant loss. Since this lost information is not included in the index's construction, index performance is biased upwards and downside volatility is possibly understated. II Backfill bias. Hedge funds are added after they have a few successful years managing money, at which point their entire return history is put on the database. This biases the data toward firms that managed to survive the first few, difficult years. II InvestibiUty. Indexes potentially include funds that are no longer accepting new assets. The index is therefore not investible, so it is not a true benchmark. II Transparency. Some index providers reveal the number of managers in each category but not the actual names of the managers. These indexes are therefore not known in advance, and so are not useful for true benchmarking. This feature makes comparing any one fund to the index less effective. II Incorrect fund categorization. Funds can identify their own categories, and some funds report themselves in categories in which they do not manage capital. 11 Frequency of reporting. For many traditional investment products, index performance can be calculated on a daily basis. Hedge fund indexes, by contrast, are typically reported only monthly. This occurs because many hedge fund managers report results only on a monthly basis. Monthly data tends to understate a fund's true peak-to-trough losses. It is not unusual to encounter short-term periods of significant loss that would be revealed if daily data, as opposed to monthly data, were available. II Leverage measurement. "With conventional indexes (e.g., the S&P), there is no ambiguity about what it means to be fully invested versus the index. The same degree of certainty does not exist with respect to hedge fund indexes. Differences in returns among managers are caused, to some degree, by different levels of leverage inherent in each manager's strategy. Of course, statistical methods can be used to infer the effective leverage of a manager relative to an index. For example, one can calculate the beta of any manager's returns to the index. However, there will likely be a wide confidence interval around any statistical estimate since the hedge fund index providers have only monthly data. II Constituent weightings. Some indexes equal-weight the funds in their indexes, while others use weights based on assets under management. An equally weighted index is particularly suspect because this construction process gives equal weight to returns from both small and large funds. In fact, large funds and small funds, even if they operate in the same investment space, are often not comparable. To the extent that large returns may be easier to achieve on smaller rather than larger amounts of capital under management, this approach overstates the performance of the investment sector the index purports to measure.