The presence of target zone nonlinearities is generally refuted in empirical research. We argue that this may be due to estimation being performed vis-à-vis official limits when monetary authorities in fact are targeting a narrower band. Estimation results for the Belgian and French franc confirm that nonlinearities are present when narrower zones are accounted for.Keywords: Brownian motion, conditional probability density function, exchange rates, nonlinearities, target zones.JEL classification: F31, F33. * University of Amsterdam, Department of Economics, Roetersstraat 11, 1018 WB Amsterdam, the Netherlands. Tel.: +31-20-5254194, fax: +31-20-5254254, e-mail: Dirk.Veestraeten@uva.nl. I am indebted to Roel Beetsma, Giuseppe Cavaliere, Paul De Grauwe, Casper de Vries, Hans Dewachter, Henk Jager, Eiko Kenjoh, Franc Klaassen, Piet Sercu, Marno Verbeek and Koen Vermeylen for stimulating discussions. Of course, the usual disclaimer applies.
1
Page 1 of 22Editorial Office, Dept of Economics, Warwick University, Coventry CV4 7AL, UK 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 F o r P e e r R e v i e w
Submitted Manuscript
IntroductionEmpirical research on nonlinearities and implicit bands in exchange rate target zones has largely evolved independently. The literature on nonlinearities in the relationship between fundamentals and exchange rates found only sparse evidence of departures from linearity (see, for instance, Flood, Rose and Mathieson, 1991;de Jong, 1994;Lindberg and Söderlind, 1994;Garratt, Psaradakis and Sola, 2001). On the other hand, the empirical literature recorded strong evidence of narrower or implicit zones within official limits in the Exchange Rate Mechanism (see, amongst others, Bartolini and Prati, 1999;Anthony and MacDonald, 1999;Chung and Tauchen, 2001). However, connecting these two distinct lines of the literature could be a worthwhile exercise since the presence of narrower zones implies that expectation formation is to be evaluated with respect to these actual limits rather than vis-à-vis the wider, officially announced zone. This is of the utmost importance for the search for target zone nonlinearities since they are likely to be overlooked when inspection is performed with respect to limits that are too wide. Indeed, the Krugman (1991) model shows that nonlinearities primarily are to be found near the limits. Using official limits in estimation when actually a narrower zone is in place will therefore tend to underestimate nonlinearities. This paper estimates the Krugman (1991) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 F o r P e e r R e v i e w 2 The target zone model of Krugman (1991) The model of Kru...