Covered Interest Parity and Its Persistent Violations

Covered interest parity held tightly before 2008, and the persistent basis since reflects the cost of bank balance sheet regulation.

Covered interest parity says the forward exchange rate between two currencies must reflect the interest rate differential between them, because otherwise a trader could borrow in one currency, convert, lend in the other, and lock in a profit through the forward market with no net currency risk. The relationship held tightly before 2008, within fractions of a basis point, and it has shown a persistent, measurable basis since, a gap explained by the cost of bank balance sheet and regulatory constraint on the capital that would otherwise arbitrage it away rather than a free lunch nobody had noticed.

Covered interest parity connects the spot exchange rate, the forward exchange rate at some maturity, and the interest rates available in each currency over that maturity. If dollar interest rates run higher than euro interest rates, the euro should trade at a forward premium against the dollar large enough that a hedged dollar deposit and a hedged euro deposit pay the same return once currency risk is fully covered through the forward contract. That is the covered part: the currency exposure is eliminated by construction, so the remaining question is whether the interest rate differential and the forward premium line up.

The logic behind the condition is a straightforward arbitrage argument. Borrow in the lower yielding currency, convert the proceeds at the spot rate, deposit in the higher yielding currency, and simultaneously sell the proceeds forward back into the borrowing currency at the maturity date. If the forward rate does not fully offset the interest rate gap, the trade generates a profit with no net currency exposure, and enough capital chasing that profit should erase it.

Before the 2008 financial crisis, covered interest parity held to within a few basis points across major currency pairs for most of the time, and deviations that did appear were closed quickly by bank trading desks running the arbitrage as part of ordinary treasury and funding operations. Balance sheet capacity to run this arbitrage was effectively unlimited relative to the size of the mispricing, because a bank could add the trade to its books without materially changing its funding cost or its regulatory capital charge.

This made the condition a useful teaching example precisely because it worked as advertised. It was one of the few genuinely close-to-riskless relationships observable in a live market, and it gave students and new hires a clean way to check whether they understood forward pricing.