I am looking for some guidance regarding model predictions of revenue and expenditure numbers.
I have done a predictive model using ETS to predict revenue for a client and received a question regarding the net present value of the results. The question is: "Should the predicted values for future years be discounted to be nominal or should they be adjusted for inflation?"
My thinking is that the historical revenue (10 years) have grown with inflation as well as growth in volume and the ETS model use the trend (no seasonality) to forecast revenue for future years. The forecast should therefore include an assumption for inflation as well as volume growth. I therefore assume that the future values include an adjustment for inflation and should thus be discounted to get the nominal values at time p0.
I will appreciate any feedback about the accuracy of this assumption or any additional advice.
With the disclaimer that I'm not a statistician (so take my comments with a grain of salt). I think that since the ETS model doesn't include exogenous variables (i.e. inflation) there is no guarantee (or messure) that inflation is being taked into account or just lost in the remaining errors (I think this is very likely since inflation could be non parametric) and also removing effects of inflation from the forecasted values is going to depend on estimated inflation values, thus adding another source of error.
Assuming that the goal for this is to perform project evaluation using economic engineering techniques, I think that a better aproach would be to remove the effect of inflation from the historical data (which has known values) before modeling so you would get forecasted values in terms of "todays dolars" making the calculation of present value much easier.
The goal is to produce entity wide forecasted financial statements as far out as 30 years and use infrastructure investment to adjust for future growth and then measure sustainability based on a set of predefined ratios and norms. Its pretty ambitious but a nice challenge. My approach is to forecast revenue and expenses and then calculate the balances and cash flows. Adjusting the historical numbers for inflation is an interesting approach.