Friday, July 6, 2012

Temasek Holdings Forecast Performance

Temasek Holdings in the interest of setting standards in transparency have been reporting their annual performance to the Singaporean public. I am not sure if they had taken out adverts in international leading publication as previously, which might appear later.

This year's performance report has one unusual feature: Monte Carlo simulations of returns suggesting that by 31 March 2013 returns can be expected to fall between -18% and 26%. At least they try to avoid "how many sigmas" but I still think these graphs and explanation would be over the head for most readers.

I am skeptical you can trust a Monte Carlo simulation of the portfolio under various scenarios. It is just too theoretical, i.e, the errors are wider than the out numbers and range. This take me back to the ideas of Nassim Taleb eloquently argued Black Swan. Gosh, we are not dealing with hard disc MTBF here. It is a stage full of manic depressive human actors with a wide range of asset sizes. You can be sure Monte Carlo have zero chance of  modeling herd behavior even if here is no lack of literature pointing to the opposite. Lots of vested interests needing protection out there. I can think of some who had tried to sell me their mathematical products before. 

I hope they do not assign responsibility and accountability away from themselves to these in-house models. It can easily be the substitute object of a committee's product where everyone can take credit upon success and nobody is punished for catastrophic failure. Remember how they had blamed it on the VAR models when trading losses mounted. 

1 comment:

  1. Cynical investor has made a fantastic suggestion:
    Madam Ho Ching should pay F&N and FCOT to teach Temasek financial engineering.