Ernest H. Oksanen

A Multi-Country Analysis of the Effects of Data Revisions on an Econometric Model

Abstract This study may be regarded as an empirical contribution to the subject of measurement error in economic statistics and its influence on econometric estimation. A small annual econometric model is specified and estimated using United Nations national accounts data for each of 21 countries for the period 1955–64. The model is fitted first to preliminary data and then to revised data. Changes in coefficient estimates and related statistics resulting from the data revisions are examined. The data revisions themselves are also examined from the point of view of their magnitude, direction, and inter-series correlation.

Measuring the rate of protection with index numbers

Abstract Two approaches to measuring the rate of protection in binary comparisons, involving use of Fishers Ideal Index, are proposed. One is relevant to measuring ‘discrimination’ in tariffs and the other is relevant to measuring a ‘composition effect’.

A Principal Components Analysis of the National Accounting Data of Eighteen OECD Countries

It would appear that the unexpected negative impact of diversification on price-cost margins found in this study is not attributable to the measure of diversification, but rather to the fact that this analysis is based on more broadly defined industries (Enterprise Industry) than the earlier study. This suggests that more widespread diversification, that is probably on average into less closely related activities, may lead to inefficiencies in production, distribution, and management12 that outweigh the positive effect of diversification on margins arising from barriers to entry associated with diversification13 and perhaps from efficiencies of vertical integration and management. While improved and more consistent data reporting by the Census Bureau are needed to resolve questions as to the directional impact of diversification, the regression results presented in this study lend additional support to the proposition that diversification may be treated as a structural variable that affects industry margins. The results also suggest that the barriers to entry variable provides a workable, readily available index of the general conditions of entry in an industry.

INTERNATIONAL COST DIFFERENCES A COMPARISON OF CANADIAN AND UNITED STATES MANUFACTURING INDUSTRIES

T HIS study is based upon a comparison of industrial cost and employment characteristics for manufacturing industries in Canada and the United States. The two economies are highly interdependent, and studies such as this should be relevant to considerations of international economic policy. It begins with extensive pairwise comparisons of means, which indicate considerable similarity in characteristics. These similarities suggest the hypothesis that cost and employment characteristics in Canada do not differ systematically from those in the United States. Cons.equently, multivariate analysis is applied to determine whether or not the CanadianAmerican differences should be regarded as random or systematic. The application of discriminant analysis to a data set for the two countries leads to the conclusion that industries can be classified according to national origin with an extremely high degree of accuracy on the basis of the characteristics in question.