برابری بهره واقعی و هزینه های معامله برای گروه از 10 کشور

Equality of ex ante real interest rates is investigated allowing for variations within a transaction costs band. Transaction costs are estimated in foreign exchange markets and in Eurocurrency markets directly from the bid–ask spreads. Two one-sided t tests show that observed transaction costs are too small to account for differences among real interest rates. Moreover, there is a clear evidence that transaction costs tend to decrease overtime.

Many tests are used in the literature to investigate equality of ex ante real interest rates across countries. If these rates are equal, then this implies that the real interest rate is determined in the world credit market rather than in national credit markets. This means that governments are less able to conduct independent economic policies to affect the real interest rates since they are not determined by the conditions in the national markets. The empirical literature that employs regression methods to test whether two real interest rates are equal, which was led by Cumby and Mishkin (1986), Mark (1985), and Mishkin (1984), suffers in general from three measurement problems. First, it ignores new developments in time series analysis regarding the invalidity of tests when one or more of the variables are nonstationary. Second, these tests focus on the short run equality of real interest rates and ignore any dynamics and the long run validity of the real interest parity (RIP). Third, as Goodwin and Grennes (1994) argue, conventional testing of equality of real rates that uses regression methods is misleading because it ignores transaction costs. They argue further that national real interest rates may fluctuate independently within a transaction costs band, even if financial markets are well integrated. The empirical work that uses cointegration methods (for example, Chinn & Frankel, 1995, Modjtahedi, 1988 and Throop, 1994) does not explicitly acknowledge this transaction costs band. This paper goes beyond the previous literature by investigating equality among real interest rates allowing for variations within a transaction costs band. Transaction costs are estimated in foreign exchange markets and Eurocurrency markets from the bid–ask spreads. Two one-sided t tests (TOST) methodology is used to investigate whether transaction costs are large enough to account for differences among national real interest rates. The paper is organized as follows. Section 2 describes the methods used to test for equality among ex ante real interest rates including the TOSTs. Section 3 provides a brief summary of measuring transaction costs in the literature and describes measurement of transaction costs in this paper. The empirical results are presented in Section 4, and some concluding remarks are presented in Section 5.

Transaction costs are too small to account for differences among real interest rates. TOSTs reject equality among real interest rates within a transaction costs band for most pairs of currencies. The estimated average transaction costs is around 0.23% while a 0.86% average is needed to account for differences among these real interest rates. However, it is possible for the after-tax real interest rates to be equal within a transaction costs band. This question is left for future research. Moreover, there is a clear evidence of the tendency of transaction costs to decrease overtime. Therefore, deviations from RIP may be better explained by other factors such as the deviations from the purchasing power parity and/or the exchange rate premia rather than by transaction costs.