¹ Many economists have adduced from the existence of substantial wage dispersion “proof” of serious error in the marginal productivity theory of wages. For bibliography and rebuttal, see Douty, H. M., “Some Aspects of Wage Statistics and Wage Theory” in Proceedings of Eleventh Annual Meeting of Industrial Relations Research Association, 1958.Google Scholar For union policy aspects, see Goldfinger, N. and Kassalow, E. M., “Trade Union Behaviour in Wage Bargaining” in Taylor, George W. and Pierson, Frank C., eds., New Concepts in Wage Determination (New York, 1957), 78–81.Google Scholar With regard to governmental aspects, in Canada the administration of the Fair Wages and Prevailing Rates legislation continually involves consideration of problems arising from wage diversity as does the American Walsh-Healey Act.

² This can be explained in large part by the fact that no establishment rate data are published. The present study is based on unpublished data made available by the Economics and Research Branch of the Department of Labour, Ottawa.

³ The modal rate was chosen for these establishments with several labouring rates since, upon observation of the data, it was apparent that certain of the rates within any particular establishment were probably beginners rates or rates for workers somewhat more skilled or more experienced than labourers. In almost all cases the majority of workers were clustered at one rate.

⁴ Some manufacturing industries do not report a labouring rate; some establishments within an industry which does normally report labouring rates for one reason or another fail to report a rate for this occupation in a particular year. This sample of “all” manufacturing is thus limited by the extent of the wage rate survey in this particular respect. It consists of 1,253 establishments.

⁵ Establishment wage data are available for earlier years but in a form which makes any extensive analysis quite unfeasible.

⁶ The merit of this measure is described by R. M. MacDonald, who says that common labour is “one of the few occupational groups which meets the test of rough comparability between plants. Common labour involves only physical work, requires a minimum of skill and experience, imposes little responsibility. …” In chapter on Pulp and Paper in L. G. Reynolds and Taft, Cynthia, eds., The Evolution of Wage Structure (New Haven, 1958), 130.Google Scholar Clark Kerr points out that common labour is one of the most important “ports of entry” between the external and internal markets, that is, it is very important from the viewpoint of labour recruitment and turnover. See “The Balkanisation of Labor Markets” in Labor Mobility and Economic Opportunity (New York, 1954), 102.Google Scholar

⁷ The inter-decile range (the “middle” 80 per cent) is used throughout this study as a measure of dispersion. While its limitations were recognized, it was none the less considered the most suitable measure given the nature of the raw data. Standard deviation (or coefficient of variation) is preferable on purely statistical grounds but, after some experimentation, was rejected because a large part of the analysis (for example, the examination of individual industries or specific size categories of plants within specific industries and market areas, etc.) deals with a sample of establishments too small and too scattered to merit the degree of statistical manipulation involved in these techniques. The inter-decile range is an easily understood and straightforward absolute monetary measure. It was decided to use an absolute rather than a percentage measure since variation in the means of various distributions of rates in the two provinces tended to exaggerate (often in an unpredictable fashion) the comparisons of dispersion.

⁸ One consequence of this very considerable spread in rates in both provinces is that the so-called regional differential, based on averages, is revealed to be a whole universe of differentials created by the relationship between two rather wide distributions of rates. For a method of analysing this universe, see Ostry, S. W., Cole, H. D., and Knowles, K. G. J. C., “Wage Differentials in a Large Steel Firm,” Bulletin of the University of Oxford Institute of Statistics, Aug., 1958.Google Scholar

⁹ The fact that the inter-decile range for the ten industries is somewhat greater than for manufacturing as a whole has no significance for the question of inter-industry variation but arises from the fact that the industries selected include a substantial number of very low-wage and very high-wage establishments.

¹⁰ See: Glasser, Carrie, Wage Differentials: The Case of the Unskilled, 161–4Google Scholar; Reynolds, and Taft, , Evolution of Wage Structure, 81–20.Google Scholar

¹¹ In Ontario there are many more metropolitan and other major industrial areas than in Quebec; a substantially larger proportion of Quebec industry is found in relatively small “outlying” communities in th e province. See “Classification of Labour Market Areas, Jan. 1, 1959” Labour Gazette, Ottawa, Jan., 1959.Google Scholar

¹² This allows us to compare two market areas of approximately similar degrees of concentration, but it cannot deal with the factor of the relative location of these communities within the two regions.

¹³ We tried to find a reasonable number of industries with at least five establishments, but these proved to be too few industries for analysis so that we were forced to use, as a criterion, a minimum of four establishments in each city.

¹⁴ See Beausoleil, Gilles, Wages in Quebec and Ontario (Montreal, 1954), 48.Google Scholar Also, for similar findings related to the North-South differential in the United States, see Reynolds, and Taft, , Evolution of Wage Structure, 342.Google Scholar

¹⁵ Ideally, one should hold constant other factors suspected of affecting wage levels such as size of plant and unionization. The present sample for any one industry is, however, too scanty to permit this type of analysis.

¹⁶ A lack of association between community size and wage level has, however, been noted by others. Thus MacDonald, in his study of the pulp and paper industry in the United States, says: “It is important to note the relevant consideration in appraising the influence of location on plant wage level is not community size per se but rather the presence of other high wage plants.” Reynolds, and Taft, , eds., Evolution of Wage Structure, 140.Google Scholar Also Jesse Markham, stressing the importance of industry mix as a factor giving rise to the North-South differential in the United States, emphasizes the fact that “the relative size of urban centres” in the two areas “is not a fundamental cause” of lower wages. “Some Comments on the North-South Differential,” Southern Economic Journal, Jan., 1950, 283.Google Scholar

¹⁷ Each establishment is coded, by size, into one of nine codes ranging from one (1–7 employees) to nine (1,000 and over). Actually the present sample ranges from size two (8–14) to size nine.

¹⁸ These distributions were calculated for all manufacturing at the provincial level, for all manufacturing at the city level, and for the group of thirteen industries in Montreal and Toronto. Since these exercises yielded no different results from those described above, it was felt unnecessary to reproduce them here.

¹⁹ A11 the high-wage plants in the larger size categories (with one exception) are pulp and paper mills.

²⁰ This conclusion is similar to that found in the classic study “Hourly Earnings of Employees in Large and Small Enterprises,” carried out by Jacob Perlman for the T.N.E.C. (Washington, 1940). Perlman found no evidence “on the independent character of the relationship between size of establishments and average hourly earnings” (p. 46). Pierson, F. C., in an analysis of Los Angeles County, Community Wage Patterns (Berkeley, 1953), 120–3Google Scholar, found a non-significant correlation between plant size and average annual earnings. It is not clear whether Pierson is referring to plants or firms, but the footnote to Table 15, p. 122, seems to indicate he is using establishment data. See also similar findings in Knowles, K. G. J. C. and Hill, T. P., “The Variability of Engineering Earnings,” Bulletin of the Oxford Institute of Statistics, 05, 1956, 107.Google Scholar For contrary view, see Cleland, S., The Influence of Plant Size on Industrial Relations (Princeton, 1955)Google Scholar and Douty, H. M., “Union and Non-Union Wages” in Woytinsky, W. S., et al., Employment and Wages in the United States (New York, 1953).Google Scholar This latter study employs averages which may be very misleading since even the weakest of associations will result in high correlation if the raw data are sufficiently condensed.

²¹ Ross, A. M., “The External Wage Structure” in Taylor, and Pierson, , New Concepts in Wage Determination, 189.Google Scholar

²² The Quebec figure is probably somewhat inflated since some of the establishments reporting a union may in fact only be under a Parity Committee. It was too great a clerical task to check each of the nearly 500 establishments in the Quebec sample.

²³ Ross, , “External Wage Structure,” 197.Google Scholar