RONALD C. DUNCAN During the 1980s, substantial advances were made in the global modelling of commodity markets in several areas, advances which were reflected in many of the papers delivered to the Applied Econometrics Association meeting held at the World Bank in Washington, DC, in October 1988. The several areas where I see advances being made, some of which the International Commodity Markets Division of the World Bank has taken part in, are the following: (a) in the theoretical specification of commodity price behaviour; (b) in the increased emphasis on modelling imperfect markets; (c) in the incorporation of the interrelationships between macro­ economic and commodity market variables; (d) in the specification of supply response, particularly in respect of perennial crops; and (e) in the realization of complementarity between time series analysis and economet­ rically estimated structural models. Improvements in the specification of the commodity price formation process have probably been the most important of the above advances. Until the early 1980s, prices were modelled as a simple linear function of stocks. Gilbert has played an important role in introducing the rational expecta­ tions hypothesis into the specification of commodity prices. Recent work by Gilbert, Trivedi, and Deaton and Laroque offers the possibility of non-linear specification of the relationship between prices and stocks within an expectational framework and of thereby capturing the phenomenon of sharp run-ups in commodity prices. Gilbert has also played an important role in clarifying the interrelation­ ships between macroeconomic variables and primary commodity prices.