Regarding the first year, i don't use a model, I use economists median type forecast and check if they are in the range of the futures for the HICPxT. For others index I take the economists forecast and that's all. Because I beleive they are better at this game than I am.For the following years I proceed in a 2 step method.First extract the seasonnal component from historical Data ( via TRAMO SEAT or X-12 ARIMA ) then I use a simple model (like the dummy one suggested) to estimate the "mean" seasonnal component of the extracted series.( Because what I want is a simple long term behavior of seasonality and because all the sophisticated econometric stuff has already been done in step one)(This method is the one suggested in the Belgrade paper)Of course every month I have to reassess those factors and the P&L is subject to a jump if my new seasonnal factors differ from form the previous set of factors. This is model risk and it's clearly hard to evaluate ...

Hi there, I am trying to interpolate implied inflation volatilities. May I know which is the common practice here? Is it similar to the CPI index interpolation convention? For example, EUR is interpolated piecewise while USD is interpolated linearly. Thanks in advance, Rgds,xdlyn