Reading 10

Hello Everyone,

Please help me to understand time series misspecification 5 : forecasting the past.

Thanks

Any body please help.

Suppose in the real world, consmer debt in the future is a function of current consumer spending

D(t+1)=bo+b1*S(t)

But you work on a flawed specification

D(t+1) = bo+b1(St+1)

then in this example, you are using on the right side of the equation a relationship that normally characterize D(t+2) but you use D(t+1) as an dependent. In doing so, if you use the coefficients to make forecasts, you’re forecasting a variable already described in the real world by S(t). So you’re late to the game…you’re forecasting the past!

Hi

Got your point.

Being a forecaster how one be able to learn about this model’s misspecification because many times trends of past repeat in the future.

Thanks