I'am looking for a book or paper that explain (in a theoretical and empirical way) how to estimate the long-run exchange rate for a country. Could someone suggest me one?
I find tons of references here (search box, to the left):
- this paper recalls that if purchasing power parity holds, real exchange rates is constant (and thus prediction of nominal exchange rate depends on prediction of relative prices). They find evidence in favour of long-run PPP. There are tons of other papers testing PPP.
- forecasting long run exchange rate with cointegration methods (error-correction model).
- a simulation study to test if predicting long-run exchange rates is ever appropriate.
- another test of predictability, using cointegration methods.
- Recent examples of how to predict equilibrium exchange rate for a country with a fixed exchange rate, and one for a flexible exchange rate.
The list is truly enormous.
I kept digging and found the following references. Hope it will help someone else.
(1) There is a large body of research that suggests that the statistical properties of exchange rates are best described by a random walk. In this case your best guess for the long run exchange rate is the exchange rate today.
(2) The theoretical foundation for many exchange rate forecasts is that in the long run some variation of PPP should hold. This means that the exchange rate should adjust to equalize prices in different countries. A forecast for the long run exchange rates could thus be based on the differences in prices today and forecasts of inflation rates.
(3) Be aware that your forecast will most likely be close to worthless which will be obvious to everyone if you include confidence intervals that attempt to correctly reflect the uncertainty in the numbers.