I suspect that cross-sectional strategies and time-series strategies run at substantially different volatilities given same weights.
Things like payroll, ISM, you mean? I spent a few years looking at this. I found that the market leads the indicator not the other way round. Personally I wouldn't waste my time on this. GAT
That's a short summary after years of research, but all the more useful. I thought that FX and fixed income could perhaps be traded given the link of fundamentals with monetary policy.
You're right - good spot. The way to deal with this is to work out the p&l of each strategy individually, measure the vol, and then take that into account in the allocation of forecast weights. Personally I don't bother since I only have a couple of cross sectional strategies whose vol isn't especially low, and it doesn't bother me if they're run at a lower vol. GAT
Because of the low frequency you won't get statistical significance unless there is a strong effect: and there isn't a strong effect. GAT
Hmm, so in this example, there are maybe 123 good data points to calculate carry between GBM Dec and March, and I get similar numbers: -5% average carry over those 123 days. Between March and June I have fewer good data points for the same date range, about 65. Averaging carry over that range is 4%. So still very different numbers even smoothed, and not really 180 data points to smooth across. Or do you mean smooth the carry calculation after rolling multiple contracts over that 180 day period? (I've seen that being used in some cases to smooth commodity seasonality.) Thanks!
Great, that makes sense, thank you. I did not know that there was a German bond effect that needed to be smoothed. Do you know why this is? Is it some sort of "seasonality" effect that is unique to German bonds?