Rule out the rate gap

The short-term interest-rate gap between two countries is a good place to start any currency analysis. Unfortunately, the simplistic approach of assuming the currency with the higher nominal interest rate will be the one to attract capital inflows often leads to the incorrect answer.

Markets often assume, correctly, that high nominal interest rates are unsustainable. Borrowing (selling) this currency even at the high nominal rate, and swapping it into (buying) a lower yielding unit often proves profitable if the high-rate policy is ended. The most famous example of this is George Soros’ breaking of the Bank of England in September 1992.

In the Brazilian case, BRL weakness between 2000 and 2002 led to ever higher short-term interest rates. The market presumed Brazil’s Committee on Monetary Policy (COPOM) would keep raising the target rate (SELIC) to support the BRL. The market presumed the opposites when the BRL started to firm — that is, the COPOM would relax its credit policies. The somewhat odd end result is the BRL led changes in the spread between six month Brazilian and U.S. swap rates by six months.

The more common pattern globally is for expected changes in interest rates to lead changes in the currency rate. If we take the interest-rate analysis out of a static spread and move it into the expectational structure of the two money-market curves, will we see the expected results? The answer is “yes.”

We can measure short-term interest rate expectations using the forward-rate ratio between six and nine months (FRR6,9) for both currencies. The FRR is the rate at which we can borrow for three months starting six months from now, divided by the nine-month rate. This number provides a measurement of the money-market conditions that are expected to prevail when the standard three-month no deliverable forward is unwound. A FRR in excess of 1.00 indicates a positively sloped money-market curve; a FRR less than 1.00 indicates an inverted money-market curve.

If we subtract the BRL FRR6,9 from the USD FRR6,9 we get a measure of relative money-market expectations. This FRR gap was consistently in favor of the U.S. from March 2003 until September 2006. It led the strengthening of the BRL by six months on average over this period. This is hardly a rule, however, and we must caution against using it as a one-factor model for the BRL or any other currency. The BRL weakened sharply after 9/11 and into mid-2002; the thinking was the U.S. was headed into a recession and any monetary lassitude in the U.S. would be more than matched by looser credit in Brazil.

The BRL strengthened sharply between mid 2002 and early 2003 even as the BRL FRR6,9 grew relatively steep. This was a period of asset price uncertainty in the U.S. but a time of strong asset returns in emerging markets. The Federal Reserve may have meant only to stimulate the U.S. consumer, but it stimulated the Brazilian producer as well. (Recall the bullish course of the Bovespa during this period.