ABSTRACT
One key characteristic of the model is its simplicity. For this, there were three good reasons.2
(i) At the analytical level, a simple model was all that was possible at the time-
the early post-war years-in view of the paucity of data (e.g. the absence of reliable national income or GNP figures) for many of the Fund’s member countries and the total absence of econometric models to describe their economies. In these circumstances there were obvious advantages in the choice of a model that could produce some useful results with the input of only two sets of statistics, both generally available, banking and trade data. [Although GNP appears in Eqs. (1) and (2) of the model presented below, its structure is such that the reduced form of its import equation contains the coefficients m and k only in the combination m/k, that is the ratio of imports to money, if the marginal propensity to import can be assumed to equal the average propensity; but to estimate Y, a separate estimate of m is needed (Polak and Boissonneault, 1977, pp. 72-3).] Although the limitations on statistical data and on the practicability of interpreting these data by means of econometric tools have to a considerable extent subsided, it would still be a questionable undertaking to design an empirical, ‘Lucas-proof’ model for many of the Fund’s customer countries, both in the developing world and among the ‘transition economies’. Thus, while the Sri Lanka workbook contains a considerable number of behavioral equations fitted to annual data for that country, these do not add up to a country model and are used only, together with more ad hoc methods, in an iterative process to estimate future values of individual variables.
