Creating Benchmarks

The Dalbar study is the only one that I am aware of that attempts to create Benchmarks for Investor Behaviour. There are obvious advantages to having a widely established benchmarks that you can use to measure your performance. The key though is to establish that they make sense.

Here are the characteristics of a good benchmark as according to the AIMR Benchmark & Performance Attribution Subcomitee Report (1998) by 'Insurance Finance & Investment', Oct 31 2006 issue, a publication of World Trade Executive. Comment welcome.

1. Representative of the asset class or mandate
2. Investable
3. Constructed in a disciplined and objective manner
4. formulated from publicly available information
5. acceptable by the manager as a neutral position
6. consistent with underlying investor status (regarding tax, time horizon etc.)

A 'Behavioural Benchmark' would obviously be different from a fund benchmark. There are a few issues that come up.

How do you measure what the client did with the money when they withdrew it? Did it sit in cash or did they maybe put it in another vehicle that actually outperformed? Is it possible to measure the impact of client behaviour without knowledge of their entire portfolio?

You can use systematic behaviour patterns as benchmarks since they are implementable (i.e. like being 'investable'). This allows comparison with methods such as dollar-cost averaging and buy and hold strategies.

It is possible to know the actual return of the investor or IRR. Why is an attempt to work out the average investor behaviour better than just using the actual return they got? Why should we model what they would have got based on 'average behaviour' such as in the Dalbar study?

Traditionally we try stripping out the effect of cash flows by giving the Time-Weighted Rate of Return rather than the IRR. The fund manager has no control of the cash flows and so we attempt to give a fairer view of the value add.

In order to accurately measure the impacts of a clients behaviour, you would need to have a detailed knowledge of their actual cash flows and then measure their decisions against their default or passive allocations. I am a bit concerned that the Dalbar measures don't really do this. By giving a simple flat $10,000/20 years it ignores the impact of cash flows and of what the clients alternative choices were.

Obviously though, benchmarks have to be simple and implementable as well which is what the Dalbar's benchmark is.