A
regression of popularity (Government considered excellent/good, source
Datafolha, very similar to Ibope or Census, but longer data) in four
determinants: (i) lag of popularity, (ii) GDP growth, (iii) inflation, (iv)
real exchange rate.
Is unemployment
irrelevant? Yes, in contrast with some specifications I saw around, which forgot
to put lagged variables and had nonstationary residuals, I found unemployment
to be irrelevant. I’m happy with this, as the unemployment series is total
crap.
Are the
regressors endogenous? Shit yeah.
In the
first picture a forecast. Assuming growth at 2%, inflation at 6% and BRL at
2.40, popularity trends up to 45. Shit yeah.
In the
second picture, which shows the residual of the regression, a curiosity. The
effect of the “passe livre” street movement was huge, but vanished.
Oh, boy, I think popularity is a substitute for trust which affects investment decisions which affect GDP !! Do not bullshit me with lagged values
ReplyDeleteI totally agree it's endogenous. Choice was made only to improve fitness.
ReplyDelete