Search
× Search
Tuesday, November 19, 2024

Archived Discussions

Recent member discussions

The Algorithmic Traders' Association prides itself on providing a forum for the publication and dissemination of its members' white papers, research, reflections, works in progress, and other contributions. Please Note that archive searches and some of our members' publications are reserved for members only, so please log in or sign up to gain the most from our members' contributions.

Estimation of risk aversion density distribution in a given market

photo

 private private,

 Monday, December 4, 2017

Many algorithms and trading methods try to predict the market response to a given strategy: in other words taking into account the reaction of traders who trades against you. These approaches generally rely on a risk parameter (or distribution density) which can be related to investors risk aversion in the market. Academic papers describing these methods remain generally silent on the way one can estimate market's risk aversion, what about in practice?


Print

2 comments on article "Estimation of risk aversion density distribution in a given market"

photo

 Vasily Nekrasov, Quantitative Developer at IDS GmbH – Analysis and Reporting Services

 Sunday, December 10, 2017



It is very hard topic alone due to the problem of robust(!) estimation of market parameters.

To understand what I am talking about by means of a simple Monte-Carlo experiment consider the last R-script from this post https://letyourmoneygrow.com/2016/09/16/stripping-down-the-robo-advisors-sparrow-brains-inside/

Back to your question: Ziemba did some work and found out that e.g. Buffett is a full Kelly investor, whereas Keynes was 80% Kelly (being more than 100% Kelly asymptotically means taking idiosyncratic risk).

I also use Kelly Criterion to determine the risk preference of investors, however, only those whom I can personally talk to.

In a sense, if you manage to have a representative(!) sampling of investors, your problem is solved.


photo

 private private,

 Tuesday, December 12, 2017



Thank you Vasily, for your insights. You're right, in general, it is very difficult to get a robust estimation of market parameters. Therefore, the optimal enveloppe in the strategic layer is quite wide.

It seems that having a good sample of investors is the most popular way to get an idea about the distribution of risk aversion. This is done also in other areas. In general, one obtains a Gamma or Log-normal like distribution.

Please login or register to post comments.

TRADING FUTURES AND OPTIONS INVOLVES SUBSTANTIAL RISK OF LOSS AND IS NOT SUITABLE FOR ALL INVESTORS
Terms Of UsePrivacy StatementCopyright 2018 Algorithmic Traders Association