
Lenny D. answered 05/21/19
Financial Professional with many years of Wall Street Experience
I can give you a few reason why it wouldn't hold and why tests might be fruitless. The fist is that interest rates have two components. an inflation risk premium and and a real interest rate. Inflation is the rate at which the value of a currency is debased. If one country has an inflation rate of 20% the other has inflation of 5% we would expect, in the long-run , that the currency of the high inflation country will depreciate by about 15% per year. If inflation rates in two countries are equal and equally stable then a positive differential might indicate that REAL interest rates are different I rise in a country's real interest and cause the exchange rate to appreciate. So, in simplistic terms an increase in the interest rate differential between a home country and the rest of the world will tend to cause the currency to appreciate if it is because real interest rates have risen as it will attract capital.This may be the result of tighter monetary policies. If the differential rises because increasing domestic inflationary risks it will tend to weaken the currency..
Something that is not discussed much is that there can be a risk premium built into a currency. For example, Turkish real interest rates seem high, However, I may be reluctant to invest in Turkey because political risk and risks that I may not be able They may tax my earnings or freeze my assets . Some currencies are considered "safe havens" as these risks to foreign investors is rather small. The Swiss Franc and US dollar are prime examples. Risk premiums are unobservable and are most likely highly correlated with real interest rates (that is a rising risk premium pushes up real rates to prevent foreign capital outflow. Risk premiums are not constant through time
When testing for uncovered parity these three components . Need to bee included or tests will be biased.
Another real concern on the testing side is that exchange rates move moment to moment . If the move on average 1% per day that is 16 percent per year. So if interest rate differential is 2% and we use that to predict the exchange rate in a year. We would expect that 68% of the the exchange rate will wind up appreciating 2% PLUS OR MINUS 16%. that is a wide margin of error.
I am always happy to discuss this stuff. Not only did I teach it but I traded in these markets on Wall Street for decades.