Gerald Holtham

Comparative exchange rate simulations

The 1983 Fall Meeting of the Project LINK was held at the time of the tenth anniversary of the commencement of the generalized floating rate regime. It might be debatable whether one should celebrate the birth of this regime or mourn over the death of the Bretton Woods system, but no one would disagree with the view that we shall have to live together with this ten-year old neighbor for some years to come so that we should learn more about the temper of this temperamental system. At the 1982 LINK meeting, one special session was devoted to an investigation of alternative methods of modeling capital flows and exchange rates [see Klein and Krelle (1983)]. During the meeting a plan evolved to organize a comparative study among multi-country or world models that have incorporated exchange rate determination mechanisms. Originally, the Federal Reserve Board Multi-Country Model, the World Economic Model of the Economic Planning Agency (Japan) and the LINK system itself had been planned to participate, but in an early stage of preparation the OECD INTERLINK also joined the group with its latest attempt to endogenize exchange rates among major currencies. In what follows the first four papers report on the methods of endogenizing exchange rates in the framework of multilaterally linked models. They also report on the simulation results designed to facilitate comparisons among alternative approaches. The final paper discusses major similarities and differences of the simulation experiments. Each model. of course has its own history of exchange-rate modeling. An early attempt in FED-MCM can be found in Bemer et al. (1977), which was based on the structural balance of payments approach. Due to difficulties in obtaining stable equation estimates for international capital transactions, among other reasons, a shift has been made to a single-equation method

Exchange rates in the OECD INTERLINK model: Specification and simulation properties

International financial linkages with endogenised exchange rates were introduced into the INTERLINK model to permit simulation analysis of the transmission of financial impulses from one country to another. Apart from simulation analysis, the chief use of the model is in checking the international consistency of the OECD Secretariat’s twice-yearly forecasts. These proceed on the technical assumption of unchanged nominal exchange rates. The model’s role in this context is to help check the consistency of national forecasts with each other and with the exchange-rate assumption. The international financial linkages (a subset of the model’s equations) were conceived as a complete and consistent world system. International identities governing financial flows are binding and 16 exchange rates are determined simultaneously. The INTERLINK system itself consists of 23 country models with sub-models of eight non-OECD zones grouped around a trade model which has a five-commodity breakdown.’ The country models are generally income-expenditure systems with some parameters estimated econometrically and others imposed to reproduce national model properties. For six of the larger OECD countries, the models are relatively more