Model Validation and Further Research

The allure of a higher frequency system is compelling, but will higher transaction costs eat into results?

It is always nice to see another analyst validate your research. Formula Research took a crack at the timing model , and after adding a few tweaks(substituting corporate bonds for Treasuries for one) found the results to be consistent(returns and risk were higher with corporates).

In another study, an individual on the Motley Fool Mechanical Investing Board reported results from a daily version of the model, below:

1972-2005

Paper’s results: CAGR 11.92% GSD 6.61%

Daily Data Results: CAGR 13.41% GSD 4.84%

He found that that the biggest improvements were on the real estate and commodity portions, a finding that is logical (they are the two most volatile asset classes). I am not sure exactly what data was used (some of the series in my paper do not have daily data back to 1972), but it is nice to see confirming results either way.

In a somewhat related note, here is an interesting Panagora presentation on risk parity.