Academic journal article Economic Inquiry

Empirical Features of the Second-Generation Target Zone Models: Mean-Reverting Fundamentals and Endogenous Devaluation Risk

Academic journal article Economic Inquiry

Empirical Features of the Second-Generation Target Zone Models: Mean-Reverting Fundamentals and Endogenous Devaluation Risk

Article excerpt

I. INTRODUCTION

In 1978 the countries of the European Community founded the European Monetary System (EMS). Its cornerstone is the Exchange Rate Mechanism (ERM), which aims at limiting bilateral exchange fluctuations among participating countries and that became effective March 13, 1979. The ERM allowed for bilateral exchange rate fluctuations to margins of [+ or -]2.25% around predetermined central parities. Italy initially obtained a margin of [+ or -]6% for the lira, which was reduced to the narrow band in January 1990. Greece choose not to participate in the ERM, while Spain, Portugal, and the United Kingdom only joined between 1989 and 1992. After hefty speculations the Italian lira and the British pound were forced out of the ERM in September 1992. After continuing speculative attacks on most of the remaining participating currencies and subsequent devaluations, the margins of the ERM were broadened to [+ or -]15% in August 1993, except for the Dutch guilder/\deutsche Mark exchange rate. Austria joined the ERM on January 9, 1995. On November 25, 1996, the lira reentered the ERM, as participation in the exchange rate system was one of the convergence criteria for participation in the European Economic and Monetary Union.

The ERM is very often considered to be the best known example of a target zone system. Under a purely floating exchange rate regime, the exchange rate will adjust to its fundamental determinants (at least if there are no bubbles), whereas under a fixed exchange rate regime the fundamental must adjust given the exchange rate. Krugman [1991] has shown that a target zone is a nonlinear compromise between these polar cases. His so-called first-generation target zone model was based on two key assumptions: (1) the fluctuation band for the bilateral exchange rate of two currencies is perfectly credible (that is, the bilateral parities are irrevocably fixed), and (2) the fluctuation band is enforced through intervention at the band limits only. As many of the empirical implications of the model were rejected, the Krugman model has been modified and improved upon to accommodate the likelihood of central parity changes as in Bertola and Svensson [1993] and intramarginal interventions as in Lindberg and Soderlind [1994] and Tristani [1994]. These models are therefore often referred to as the second-generation target zone models.

So far, the empirical target zone literature has mainly focused on three issues, as summarized by Ball and Roma [1994]. First, the statistical distribution of exchange rates inside the band has been analyzed in Flood, Rose, and Mathieson [1991] and Lindberg and Soderlind [1994]. In general, these studies document a hump-shaped distribution within the band instead of the U-shaped distribution predicted by the original Krugman model. Second, the behavior of the fundamentals in relation to the exchange rate, the so-called exchange rate function, has been estimated by various authors. Along the lines of the first-generation target zone model Flood, Rose, and Mathieson [1991] have constructed a number of such plots for the different ERM currencies from which no clear (let alone deterministic) relation could be unfolded. Allowing for a time-varying devaluation risk, Rose and Svensson [1995] have reported a weak corroboration of the characteristic S-shape for the French franc/deutsche Mark exchange rate. Third, under the assumption that the expected rate of depreciation within the band is equal to the interest rate differential between the two currencies, a relationship should exist between the position of the bilateral exchange rate within the band and the interest rate differential. On the basis of this notion, Svensson [1991b] has developed a theory of the term structure of interest rate differentials in a first-generation target zone model. Svensson has also examined data on the term structure of Swedish interest differentials and found them to be negatively correlated with the exchange rate, which is consistent with his theory. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.