Charta is on Google Play: practise on the go. Get the app
Charta

Outline the performance characteristics of mutual funds and hedge funds, and describe how measurement biases distort the assessment of performance.

FRM® Part I · Financial Markets and Products · Fund Management · FRM1.FM.3.g

The idea

Mutual funds on average earn roughly the market return before expenses and slightly less than comparable passive funds after expenses, with little evidence that past winners keep winning. Hedge fund returns are more varied and often show negative skew, fat tails and serial correlation. Reported hedge fund performance is also distorted by biases in the databases. Survivorship bias arises because funds that fail drop out, so average returns of survivors overstate what an investor in the whole universe earned. Backfill or incubation bias arises when a fund is added to a database and its earlier good record is filled in, while poor records are never listed. Smoothed or stale valuations of illiquid assets understate volatility and correlation and overstate the Sharpe ratio. Self-reporting adds selection bias. The exam point is that every bias inflates measured performance, and you should know which bias a scenario describes.

A practice question

From the question bank, with the reason each wrong option is wrong.

What is survivorship bias in hedge fund return databases?

  1. Funds are added to the database with their earlier returns filled in, usually when those returns were good.Why not: Describes backfill bias, not survivorship bias.
  2. Illiquid assets are valued at stale prices, which understates volatility.Why not: Describes smoothing or stale pricing.
  3. Only large funds report returns, so the data understate average performance.Why not: Reporting bias would not systematically understate; survivorship bias overstates.
  4. Funds that closed are dropped from the data, so average returns of the remaining funds overstate those of the whole universe. - correct

Poorly performing funds tend to close and disappear, so a database of survivors overstates the return an investor in the whole group earned.

A harder one

A database contains 100 funds for the year. The 80 surviving funds had an average return of 9 percent. The 20 funds that closed during the year had an average return of 2 percent and are excluded. By how much does the survivors-only average overstate the average of all 100 funds, closest to?

  1. 1.4 percentage points - correct
  2. 7.0 percentage pointsWhy not: Takes the difference between survivors and failures rather than between the survivors and the whole group.
  3. 7.6 percentage pointsWhy not: Reports the true average of all funds, not the overstatement.
  4. 1.8 percentage pointsWhy not: Multiplies the 9 percent by the 20 percent of funds that closed.

True average = (80 x 9 + 20 x 2) / 100 = 7.6 percent. Survivors-only average is 9.0, so survivorship bias = 9.0 - 7.6 = 1.4 percentage points.

The hardest tier

A manager runs ten small incubator funds with the firm's own money. Their average annual return over three years is 6.0 percent. The two best, averaging 15.0 percent, are opened to outside investors, and a database adds them with their three-year incubation records backfilled. By how much does the backfilled record overstate what the incubation programme as a whole earned, closest to?

  1. 15.0 percentage pointsWhy not: Reports the level of the backfilled return rather than its excess over the whole programme.
  2. 21.0 percentage pointsWhy not: Adds the two averages instead of finding their difference.
  3. 6.0 percentage pointsWhy not: Reports the average of all ten funds as the bias.
  4. 9.0 percentage points - correct

The backfilled record shows 15.0 while the whole programme earned 6.0: 15.0 - 6.0 = 9.0 points. This incubation (backfill) bias arises because only the successes get listed with their earlier history.

In the app: 6 practice questions on this outcome, spaced review that brings it back before you forget it.

Start free →

More in Fund Management