Testing Benchmark Quality

Coverage --> higher or lower? Positive Active Positions --> More or less? These two seem like odd ways to measure benchmark quality. Can someone please clarify which is an indication of higher quality for both of these? And could yo also explain why we use them as a criteria for benchmark quality? Thanks!

higher coverage and more positive active positions are good attributes for a benchmark.

Cool thanks! What about the logic in using these… WHY is it good to have higher coverage/Positive active position’s? Might be a dumb question, but I’d like to hear it in someone else’s words than the CFAi haha

say you are a large cap manager … so you can select a benchmark with a large cap focus and have 3 index options to consider of which: - LC index 1 has 1000 stocks - LC index 2 has 100 stocks - LC index 3 has 20 stocks I say the manager would be better off selecting the 1000 stock index as his benchmark as it is more likely to cover a wider range of stocks in Large CAp domain from which you can be expected to select your stocks for your portfolio. as regards +ve active positions, if your benchmark is appropriate, normally you are not expected to hold all index stocks but a sample that represents best of opportunity in large cap space. say out of 1000 stocks you hold a sample of 150 stocks (that according to you will outperform the market in near future) … so ideally if you are holding just 150 stocks instead of all 1000, you can expect to hold then in greater weights than they are in the index or a +ve active weight. hope its clear somewhat.

Higher coverage means your portfolio and benchmark have more securities in common. Positive active position means your portfolio has higher weights for certain securities compared to their weights in benchmark. From this, you can say that both are positive for an index to by your benchmark.

I have difficulty in understanding that positive active position indicates that the benchmark is of high quality. Can anyone elaborate please?

happyking02 Wrote: ------------------------------------------------------- > I have difficulty in understanding that positive > active position indicates that the benchmark is of > high quality. Can anyone elaborate please? I guess it speaks for the liquidity of the securities in the benchmark (because you were able to overweight it in your portfolio).

maybe reading the first paragraph on pg 147, vol 6 will help? It basically says, if you(a long only manager) are not holding a high proporotion of securities in your benchmark (creating negative active weight), then that benchmark is not right for you…so i guess by holding +ve active weights, the benchmark is good for you because you have opinion on the security. On a slightly unrelated note - I recall reading somewhere there is a tradeoff between having high Coverage and investing liquidity (or something like that) where if you want high coverage, you may have to buy/sell illiquid securities at a higher cost. I can’t remember the exact concept though, anyone know what I’m talking about?

Ignore my last question…found my answer… Reading 34: International Equity Benchmarks LOS b: Discuss the trade-offs involved in constructing international indices, including (1) breadth versus investability, (2) liquidity and crossing opportunities versus index reconstitution effects, (3) precise float adjustment versus transactions costs from rebalancing, and (4) objectivity and transparency versus judgment.

If you have no opinions on the securities in the benchmark you weigh it in similar proportions ‘as-is’ in the benchmark, indicating you have no opinion and the weights are same. If the manager starts overweighing some select securities and have a +ve alpha weights means he finds the bechmark interesting and can place bets as per his views. So +ve alpha weight represents a good benchmark quality.

But positive active weights in the same securities over a prolonged period of time means you’re benchmark might not be appropriate…right?

Means you are Anchored and overconfident both at the same time :wink:

while on the subject, schweser has indeed vague language in the first quality - systematic bias, i had to read it couple of time to decipher whether beta should be low or high, whether (p-b) should be correlated to (b-m) or otherwise. someone care to explain