an Alternative Approach Theory and Empirical ... - Semantic Scholar

The NATREX: an Alternative Approach Theory and Empirical Verifications

Giancarlo Gandolfo Faculty of Economics, and CIDEI (director) University “La Sapienza”, Rome

Alberto Felettigh University “La Sapienza”, Rome

No. 52 – November 1998

The NATREX: an Alternative Approach Theory and Empirical Verifications

Giancarlo Gandolfo Faculty of Economics, and CIDEI (director) University “La Sapienza”, Rome

Alberto Felettigh University “La Sapienza”, Rome

ABSTRACT The research on equilibrium real exchange rate determination is still in progress. In the recent past the NATREX (NATural Real EXchange rate) approach by Jerome Stein and his co-authors has formulated a theoretical framework for the long-run natural real exchange rate. The model has satisfactorily undergone empirical verifications for several industrialised countries. In this paper we develop and estimate a NATREX model for the multilateral real exchange rate of the Italian lira. The theoretical model qualifies the original framework in order to take into account the peculiarities of the Italian experience in the period under examination (1976-1995). The estimation of the empirical model, a system of simultaneous nonlinear dynamic equations, is done via full information maximum likelihood, which constitutes an alternative approach to the NATREX “tradition”. Our estimates confirm the validity of the NATREX theory for the Italian economy. In particular, our in-sample simulations for the long-run equilibrium real exchange rate confirm the analysis of the real misalignment of the Italian lira made by the Bank of Italy.

JEL classification number: F31, F41, F43, F47 Key words: equilibrium real exchange rate, real misalignment, NATREX

CONTENTS

1.

INTRODUCTION .................................................................................................1

2.

THE NATREX THEORY......................................................................................1

3.

THE ECONOMETRICS OF NATREX ................................................................2

4.

A NATREX MODEL FOR THE ITALIAN LIRA: THEORY...............................3

5.

6.

4.1.

THE INVESTMENT EQUATION..................................................................4

4.2.

THE CONSUMPTION EQUATION ..............................................................5

4.3.

THE TRADE BALANCE EQUATION ..........................................................6

4.4.

THE REAL INTEREST RATE EQUATION..................................................7

A NATREX MODEL FOR THE ITALIAN LIRA: ESTIMATION ......................8 5.1.

THE DATA ....................................................................................................8

5.2.

THE ECONOMETRIC MODEL.....................................................................9

5.3.

ESTIMATION RESULTS ............................................................................ 13

THE LONG TERM NATREX: DYNAMIC SIMULATIONS ............................. 16 6.1. THE MISALIGNMENT OF THE ITALIAN LIRA AND THE 1992 CURRENCY CRISIS .............................................................................................. 18

7.

CONCLUSION ................................................................................................... 19

ECONOMETRIC APPENDIX................................................................................... 20 REFERENCES .......................................................................................................... 21

1. INTRODUCTION There is a widespread opinion in the literature that economic theory is not yet able to explain, and most of all to predict, the dynamics of the nominal exchange rate (see Meese, 1990, pp.117,132 and Dornbusch, 1995, pp.197,201)1. More encouraging results have been achieved in the analysis of the real exchange rate (RER hereafter), especially when the time horizon has been the long run. Among the theories that have been developed in the recent past, the NATREX (NATural Real EXchange rate) approach to the equilibrium RER determination, originally formulated by Stein (1990), has been able to satisfactorily explain the medium-to-long run dynamics of the RER in several industrial countries: USA (Stein 1995a, 1995b), Australia (Lim and Stein, 1995), Italy, Germany and France (Stein and Paladino, 1998), Belgium (Verrue and Colpaert, 1998). In this paper we develop and estimate a NATREX model for the multilateral RER of the Italian lira. The theoretical framework of the NATREX approach is briefly summarised in the next section. In the third section the econometric techniques of the NATREX “tradition” are examined, and an alternative approach is suggested. Sections 4-6 are concerned with the theory, the econometrics and the numerical simulations of our NATREX model for the Italian lira respectively. The concluding section dwells on the claim that our model constitutes an alternative approach within the NATREX “tradition”.

2. THE NATREX THEORY The NATREX theory explains the dynamics of the medium-to-long run equilibrium RER. It is not a single model but rather a class of models each tailored to the particular features of the economy under study. In this section we briefly recall the theoretical framework of the NATREX approach: for a complete treatment, the reader should consult Allen (1995) and Stein (1995a). The NATREX is the intercyclical equilibrium real exchange rate that ensures the balance of payments’ equilibrium in the absence of cyclical factors, speculative capital movements and movements in international reserves. In other words, the NATREX is the equilibrium real exchange rate that would prevail if the above-mentioned factors could be removed and the GNP were at capacity. Since it is an equilibrium concept, the NATREX should guarantee both the internal and the external equilibrium, the focus being on the long run. The long-run internal equilibrium is achieved when the economy is at capacity output, that is when the GNP is at its potential level. The long-run external equilibrium is achieved when the longterm accounts of the balance of payments are in equilibrium. Short term (speculative) capital movements and movements in official reserves are bound to be short term transactions, since they are unsustainable in the long run. In the long-run equilibrium they must average out at zero; hence, the excess of national (private plus public) investment over national saving must be entirely financed through international long term borrowing.

1

See, however, Gandolfo et al. (1990) and Gandolfo and De Arcangelis (1997) for a medium-size model that outperforms the random walk and predicts almost all the turning points.

1

Under these conditions long term capital inflows and excess national investment over saving coincide, so that also the real market long-run equilibrium condition and the long-term external equilibrium condition coincide: [2.1]

S-I=CA

where CA is the balance of payments’ current account, the private and public sectors having been aggregated into a single one. Relation [2.1] is intended in real terms: the model assumes neutrality of money and that the monetary policy keeps inflation at a level compatible with internal equilibrium (at least in the long run). Therefore, the focus being on the real part of the economy, there is no need to model the money market. Perfect international capital mobility is assumed: the real interest rate is driven by the portfolio equilibrium condition or real interest parity condition (supposed to hold instantaneously), possibly with a risk premium. The system is assumed to be self-equilibrating (hence the adjective natural in the acronym NATREX). Take for example an initial position of full equilibrium (SI=CA=0) and suppose an exogenous shock leads to a situation where S-I0) leads to a decrease in the marginal product of capital and hence, ceteris paribus, in the rate of investment itself. The NATREX theory does not include GDP among the determinants of investment. Our model, on the contrary, assumes a positive relation between investment and GDP, for the following reasons: 1. the Keynesian tradition explicitly models a positive relation between investment and GDP, although with a small propensity (which is confirmed by our estimations); 2. the investment under exam is the social (private plus public) investment, whose public component is not necessarily driven by the aim of profit maximisation typical of the private agent. Given the peculiarities of the Italian fiscal policies in the period under exam, it is plausible to assume that public investments are a positive function of GDP; 3. from the econometric point of view, GDP plays the role of scale variable. Our investment equation is based upon the following partial-adjustment mechanism: [4.1]

It – It-1 = α1 (I*t-1 – It-1)

0< α1