Khalifa H. Ghali, Musaed Ben Eid
This paper argues that conventional tests of financial market integration may be misleading because they neglect to consider two important issues: (i) testing the existence of transaction costs and (ii) testing the existence of structural change. While the existence of transaction costs may inhibit the one-to-one correspondence between changes in real interest rates in different countries, it could also be misleading to assume the existence of such costs while they do not. Therefore, this paper’s proposition is to test for the existence of transaction costs to motivate the inclusion of an intercept term in a cointegration equation of real interest rates. It is shown that this is possible using the Johansen (1994) procedure to test for the inclusion of deterministic components in the cointegration space. It is also noted that the empirical literature has neglected the problem of stability of financial markets and, hence, its consequences on financial integration tests. To solve these two issues, this paper proposes a model of financial market integration that allows for the existence of transaction costs and for the possibility of structural change. This alternative model generates much stronger support for interest parity than is found in the existing literature.
© 2001-2024 Fundación Dialnet · Todos los derechos reservados