[R-SIG-Finance] Dynamic asset allocation among hedge fund's strategies

[Alex Bird]

Hi there,

  I am trying to solve a dynamic asset allocation problem among some
hedge fund’s strategies (managed futures, gamma scalping, etc.) For
these purposes I decided to use (after a lot of experiments with
alternative models) MCMC techniques to fit a model similar to a
multivariate stochastic volatility with some minor changes (like
seasonality, regimes, periods of strategy inactivity, etc.) and then
to use n-days ahead forecasted return distributions produced with the
model to decide how much to allocate to a specific strategy via CARA
approximated by Taylor (to account for 1st 4 moments).  Rebalancing
occurs several times per months.

  Now I would like to compare the model with some alternatives I
missed during the experimentation phase. Markowitz mean-variance and
its modifications are of zero interest because of their low
performance compared to the model mentioned above. Could anyone please
give me maybe some advice on where the pitfalls could be or suggest a
good alternative to play around with.

Many thanks in advance!
Alex