Financial development, exchange rate regimes and the Feldstein–Horioka puzzle: evidence from the MENA region

Abstract

This article investigates whether the Feldstein and Horioka (1980) argument on domestic saving-investment relationship is supported by the data of the countries in the Middle East and North Africa region when financial development levels and exchange rate regimes are taken into account. To this end, we employ both the Auroregressive Distributed Lag bounds cointegration test and panel mean group procedures. The results support the view that a successful international financial integration requires compatible levels of financial intermediation. The evidence also suggests that saving-investment cointegration is not invariant to exchange rate regimes.

Notes

¹ See Coakley et al . (2004), Obstfeld and Rogoff (2000) and Taylor (2002) for recent surveys.

² The results of the ADF and KPSS tests suggested that sav _t and inv _t , are I(1) for Algeria, Egypt, Iran, Syria and Turkey. The variable sav _t is I(1) for Morocco, S. Arabia and I(0) for Israel, Jordan, Tunisia whilst inv _t is I(1) for Israel, Jordan. The financial development variable fin _t is I(0) for Iran and I(1) for the others. The results for the rest of the variables are found not be to definitive whether they are I(0) or I(1). However, the necessary condition for the implementation of the MG and Pesaran et al . (2001) bounds procedures, the degree of integration being less than two for each of the variables in a system, appears to be satisfied. See the working paper version of this study (Özmen, 2004) for these results and the individual country estimates to obtain the MG coefficients.

³ Due to the constancy of the exchange rate regimes for Morocco, S. Arabia, Syria and Tunisia, the augmented FH equations do not contain the ERR variables for these countries.