One of the most significant post-war development in economics has been the contribution of econometrics to the refinement of techniques of analysis. Econometrics is now an integral part of economics teaching in most of the known universities throughout the world and large econometric models are being used for economic policies in industrialized countries. The CANDIDE model (Canadian Disaggregated Inter-Departmental Econometric Model) is a medium-term policy oriented model and an indication of some degree of maturity in the Art of model building in CANADA.
The purpose of this special issue of L'Actualité Économique is to initiate undergraduate students in economics as well as the interested public to the model, to show how the model can be applied and finally to discuss some of its deficiencies. It is hoped that this special issue will be a useful tool for the teaching of econometrics in Canada. It has three parts. The first part comprising three papers, explains the nature and the characteristics of the model. The second part through five papers shows various applications of the model. Finally, in the third part, seven papers discuss some of the deficiencies of the major blocks of the model1.
1 Only the first two parts are included in this issue. The third part will be published in the next issue.
This introductory text deals in very general terms with CANDIDE and development and application of large models.
A brief sketch of the history of CANDIDE precedes a discussion of the extensive disaggregation and medium-term nature of CANDIDE; the two characteristics that set it apart from more traditional models.
From his experience with CANDIDE the author then comments on model design and model use. Development of multi-purpose models, in his view, consists of many difficult and often arbitrary choices. He warns against excessive disaggregation, and use of models as black boxes. Finally, he calls for co-ordination of impact studies.
This text describes the CANDIDE model from three critical points of view. First, it presents the structure of the model with the aid of a table of "supersectors" and a flow chart. The discussion of the linkages between these supersectors brings out the general equilibrium character of the model.
The center section of the paper deals with the theoretical foundations of the behavioural equations. A statement of the equation specification in each sector of the model is followed by a brief evaluation. The authors confess to an eclectic attitude and stress the practical difficulties of making stylized theories bring order in variegated facts. Exogenous influences on the model are discussed as well.
The same eclectic approach is evident in the third part of the text which treats of estimation and validation. A variety of tests and checks—some novel ones included—are applied in the process of developing the model. The authors conclude with the observation that recent tumultuous developments in the economy raise fundamental questions about the validity and usefulness of models like CANDIDE.
The CANDIDE model has been used by the Economic Council to examine Canada's economic potential, to analyze the effects of economic forces, and to consider the appropriateness of alternative policies in reaching economic objectives. For its Annual Reviews, the model provides an analytical basis for taking into account the interdependence of a number of phenomena, including those related to demographic trends, external economic conditions and domestic policies influencing supply and demand, and thus facilitates estimation of the potential development of the economy over the longer term. Within this context, a realizable set of medium-term objectives can then be established. These have been presented by the Council as performance indicators for the three years immediately ahead and they are subsequently used to monitor and assess economic developments. The model also is used by the Council to examine how various economic influences work their way through the Canadian economy. In its Annual Reviews, for example, the effects of alternative scenarios for energy investment and prices have been considered. In a special study of the construction industry, the model was used to trace the causes and effects of instability in this sector. Some illustrative results from each of these impact studies are provided. The model has also been employed to explore the implications of certain past and future changes in commercial policy, including separating out the impact of the Canada-United States Automobile Agreement, and in examining changes that have been taking place in labour markets. Each of these areas have been the subject of special studies carried out by the Council.
This paper simulates the economy of the period 1963-71 under the assumption that the federal sales tax on construction materials had never been introduced.
Evidently, under such circumstances the revenues of the federal government would have been lower than with the construction materials tax in force. We attempted to compensate for the "shortfall" in revenues by raising other taxes or by reducing federal government expenditures.
The simulations suggest that the absence of the construction materials sales tax would have led at best to an additional 9,300 housing starts in 1971. Construction costs in 1971 would have been 5%-8.5% lower than actual, depending on the compensatory policy. Somewhat less inflation and somewhat more housing starts can be obtained at the price of lower real growth and higher unemployment. The choice remains, ultimately, a political one.
In this paper, a large scale econometric model of the Canadian economy (CANDIDE) is employed to study questions concerning the evolution of Canadian prices over the period 1960-1974. After a brief description of the wage-price sectors of the model, "trade-off" or Phillips curves are derived on the basis of simulations of the entire model, for the periods 1960-1965 and 1965-1970 separately. On the basis of these simulations, it is found that the trade-off curve does not appear to be very sensitive to the choice of the instrument utilized to move the Canadian economy from positions of lower to fuller utilization of its labour resources. The position of the curve is, however, quite sensitive to the rate of increase of U.S. prices; also, the full curve appears to have shifted upward considerably from the first to the second subperiod. The analysis of the period 1970-1974 suggests, among other conclusions, the possibility that an even rate of U.S. inflation over this period of rather violent fluctuations in U.S. prices could have had in itself favourable repercussions on the Canadian trade-off curve. The concluding section presents another summary of the major findings, as well as some important caveats and qualifications.
Monetary policy has been recognized as one of the principal reasons for cyclical instability in residential construction. It has been therefore suggested that residential construction should be isolated from monetary policy. This paper is designed to measure the impact of such suggestion.
We have used the CANDIDE model for our experimentation. Residential construction is isolated through appropriate adjustments of the intercepts of the equations for residential construction. This led of cause of disturbances in various aggregates of the economy. Hence, a series of simulations was run in such a way that the disturbed aggregates would come close to the corresponding aggregates obtained by the control-solution. This meant additional adjustment in money supply and Treasury Bond yield, which are regarded as proper indicators of monetary policy.
It was found that the additional adjustment varied between 9 to 16% in case of money supply and 38 to 45% in case of Treasury Bond yield. In addition, variation in non-residential construction and investment in machinery and equipment was found to be significant. In short, though efficient in stabilizing residential construction, the isolation of residential construction from monetary policy appears to be costly in terms of its impact on monetary policy and capital investment.
CANDIDE-R is a huge simultaneous macro-economic model which raises estimations difficulties. We avoid the problem of identification assuming that the great number of variables in our model makes it impossible that the necessary condition be not satisfied. We assume that our system converges to a solution solving this way the problem of identification. The core of the paper gives justifications of the procedure we adopted to estimate CANDIDE-R. Because of the presence of regional equations and the limited amount of regional data, we are bound to pool cross sections and time series data. We then justified the use of Zellner's approach instead of the error components models within the class of regional models built on national premises.