my job is to duplicate a model in eviews to carry out a sensitivity analysis. the model has three equations: y, for output; π, for inflation; and r, central bank taylor rule. all 3 use a set of fixed parameters denoted as :series: series, as well as, other formulas all of which are calculated in the model in eviews.
however, i have great difficulties implementing the mean squared forecasting errors (MSFEs), which consists among others of a SUM and also a variables that denotes geometrically declining weights (w or omega). basically, the model introduces a learning factor for economic agents. the series of declining weights serves as a memory for forecasting and is quite essential to the model.
this is the formula:
currently, my model calculates this formula with:
' FORECAST ERROR OPTIMISTS REGARDING Y
END_FORECER_Y_OPT = (END_Y - END_EXP_Y_OPT(-1))^2
and
' ACCUMULATED WEIGHTED FORECAST ERROR OPTIMISTS
END_U_OPT_Y = PAR_PHI * END_U_OPT_Y(-1) - END_FORECER_Y_OPT(-1)
where parameter PHI is set at 0.5, and in no way geometrically declining.
i have thought inserting variables manually for PHI. so i calculated 491 observations in Excel basically running from 0.99 to 0.01 and copy pasted to simulate something at least, but that didn't work out as Eviews gave an error.
END_Y is my main formula for y, or output; and the END_EXP_Y_OPT is set at 0
there is also a formula for pessimists (I have to insert this formula twice), there END_EXP_Y_PES is set at END_Y (-1)
any help or insights are tremendously appreciated!