The covered interest parity (CIP) hypothesis asserts that the interest differential between two assets, identical in every respect except currency or denomination, should be zero once allowance is made for the cost of cover in the forward exchange market. Profitable deviations from CIP represent riskless arbitrage opportunities and may indicate market inefficiency. The relationships which follow from CIP are often used as identities in international macroeconomic theory, and CIP is often assumed to hold when testing other implications of the efficient markets hypothesis, such as uncovered interest parity. A true deviation from CIP represents a profitable arbitrage opportunity at a particular point in time to a market trader. In order to be able to measure such deviations, therefore, it is important to have data on the appropriate exchange rates and interest rates recorded at the same instant in time and at which a trader could have dealt.
In fact, most empirical tests of CIP have used data from published sources, which do not represent prices facing market traders. In Discussion Paper No. 236, Research Fellow Mark Taylor tests the CIP hypothesis using data on market rates throughout each dealing day, recorded by the chief dealer at the Bank of England. These rates are actual market data on which agents could conceivably have dealt and are recorded contemporaneously. Taylor uses this information to construct a data base from which he calculates whether profitable arbitrage opportunities existed.
His analysis concentrates on the weeks surrounding historical events known to have introduced `news' and turbulence into the markets, including the 1967 devaluation of sterling, the 1972 flotation of sterling, the inception of the European Monetary System, and the 1979 and 1987 UK general elections. Taylor's calculations suggest that during the 1967 devaluation, arbitrage opportunities opened up as the market became more turbulent. Prior to this any profitable opportunities were rather slight: at one-month maturity, £1,000,000 arbitraged into dollars would have produced just £473 profit. Just before devaluation, more significant profit opportunities arose and after the devaluation arbitrage opportunities increased: the riskless return on a £1,000,000 arbitrage varies between £3,800 and £8,000. The size of the arbitrage opportunities then tended to die away in the days following the devaluation.
Taylor found similar results for the 1972 flotation of sterling, but by 1979 there appears to have been an increase in market efficiency with fewer profitable arbitrage opportunities. These results indicate a violation of the efficient markets hypothesis, with the degree of violation apparently a positive function both of the amount of turbulence and of the maturity considered.
Credit limits may explain this `maturity effect', Taylor argues. Foreign exchange and money market dealers are generally not free to deal any amount with any party they choose. This will tend to create a preference for covered arbitrage at the shorter maturities since credit limits are then filled for shorter periods, leaving dealers on average freer to take advantage of other profit opportunities as they arise

Covered Interest Arbitrage and Market Turbulence: An Empirical Analysis
Mark P Taylor

Discussion Paper No. 236, May 1988 (IM/ATE)