My question is quite general and is about the coherence between forecast horizon, then forecast dates, and related actions. As example we can keep in mind the asset allocation problem. It seems me that the natural way is to produce forecast about expected return and risk with horizon one step ahead and check/modify the asset allocation at the end of the period (start of new period). Formally we can think in term of $E[r_{t+1}|I_t]$ and $V[r_{t+1}|I_t]$ where $r$ is return vector and $I_t$ stand for information at time t (present). Time horizon =$1$ (one: week, month, year, etc …); $t+1$ is “forecast dates” and it coincide to “asset allocation adjustment dates”.

However exist the possibility to make other things. For example we can produce long run forecast but modify them at predetermined dates. At same predetermined dates we can adjust the asset allocation. In this case time horizon =$\infty$, “forecast date never arrive” while “asset allocation adjustment date” are any: week, month, year, etc. ($t+1$, $t+2$, …,$t+n$, ... etc …).

More in general forecast dates and decision dates can be different (or not?); or in any case something like this sometimes happen. At least theoretically, situations like this seems me not easy to justify.

My question is: this mismatch is coherent? What coherent conditions we need between forecast dates and operational dates if them exist?

Accordingly to the above example, my question is probably related to multi-period asset allocation choices problem that is quite note; however I do not find precise answer for my question.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.