- Tamim Bayoumi, Guy Meredith, and Bijan Aghevli
- Published Date:
- June 1998
Koichi Hamada focused his discussion on the degree to which the research paper took account of intertemporal factors. He noted that modern analysis of the current account looked on it as an intertemporal phenomenon. While he acknowledged that the paper formally took such intertemporal issues into account by its inclusion of the real interest rate, he continued to have the feeling that the analytic framework was somewhat static. To his taste, the analysis seemed to be somewhat old-fashioned—more like the undergraduate textbook of Dornbusch and Fischer than the more advanced analysis in Blanchard and Fischer. The dynamics, if any, in the model were unstructured. The analysis must be clear on whether the path is a long-run intertemporal consumption choice or a short-run adjustment path with sticky prices, as in the Dornbusch overshooting model. To illustrate his concerns, Hamada mentioned that the simulations using alternative fiscal policies should depend on the assumed degree of Ricardian equivalence. For example, in the extreme case of complete Ricardian equivalence, tax policies would not matter for the results. He would also have preferred to have seen more time-series econometric work, and more discussion of the role of Balassa-Samuelson effects (i.e., the behavior of relative productivities across sectors and countries), both for past and future projections. Finally, he wondered what was the significance of long-range projections and what the implications of the analysis were for exchange rate policies.
Kazuo Ueda noted that the research paper for the session was broad-ranging, involving discussions of stylized facts, the manner in which the IMF’s World Economic Outlook forecasts are constructed, and at least two approaches to the equilibrium exchange rate—one using a reduced form and the other a large econometric model. His own discussion would focus upon the results from the model. He had several concerns about the analysis, mainly related to the dynamic properties of MULTIMOD, the econometric model used in the simulations. The first concern was that the results were clearly dependent upon the policy reaction functions incorporated in the model, and, in particular, given that the model incorporated sticky prices, they depended upon the monetary policy responses that were being assumed. There was also the issue of whether private sector behavior was more forward-looking than was assumed in the model. This was a particular concern, in his view, in the case of private consumption, which was often modeled in a more static manner than it appeared to actually exhibit, Ueda also noted that the analysis assumed a continuation into the future of the trend appreciation of the yen, due to Balassa-Samuelson effects. Given that the yen was significantly above the level implied by purchasing power parity, however, it was equally reasonable to expect that this trend appreciation would be reversed at some point in the future. Ueda also wondered if the issue of the equilibrium value of the yen could not be approached in a simpler fashion. In particular, given that relative export prices showed no trend over the long term, he wondered if the theory of purchasing power parity could not be applied to this variable. Finally, on the issue of the failure of uncovered interest parity to hold over the past, he noted that recent survey evidence implied that it may have been due to a risk premium.
Several participants contributed to the general discussion. Joseph Gagnon pointed to recent work linking trends in real exchange rates to movements in net foreign asset positions, and noted that this would imply a continued appreciation of the yen, regardless of Balassa-Samuelson effects. Masahiro Kawai was surprised that the paper made no mention of the work of John Williamson, who had championed the macroeconomic balance approach (which formed the basis for the analysis) over an extended period. Ralph Bryant agreed that the approach had much in common with that taken by John Williamson, but noted that the current paper included several important new insights. In particular, the framework emphasized that government policies were endogenous to the results, and hence that any given assessment of the exchange rate had to be predicated on a particular set of assumptions about future policies. Given this, he felt that more discussion could have been devoted to the monetary and fiscal reaction functions used in the analysis.
In addition, this also made it important to understand the assumptions being made about private sector responses to such policies more fully, such as the level of Ricardian equivalence and the existence of a risk premium in the uncovered interest parity relationship. He and Steven Symansky both felt strongly, however, that MULTIMOD did include most of the dynamic mechanisms discussed in the literature, and hence felt that the analysis was sufficiently intertemporal. Finally, Masaru Yoshitomi emphasized the increasing role of services in overall trade, which could make future projections of the trade balance particularly difficult.