Changing Regional Income Disparities in Canada, 1951-1989

Abstract

Regional income disparities are one of the most important topics in the disciplines of economic geography, regional science and economics. The well-known study by Williamson (1965) made clear that they follow an inverted U-shape curve temporally, implying that disparities diverge in an early phase and then converge in a later phase. In the 1980's, however, advanced capitalist countries experienced a divergence or slowing-down of convergence in regional income disparities. These new trends have recently stimulated theoretical and empirical investigations on this theme. This paper, taking up the Canadian experience in the postwar period, seeks to elucidate minute temporal changes in the income disparities and to explain it.To begin with, theoretical standpoints in the existing literature are outlined. As is well known, there are three dominant perspectives on this theme: the equilibrium, disequilibrium and transition points of view (Lipshitz, 1992). However, it is important to recognize not only the differences among the three, but also similarities: each perspective has referred to the direction of convergence of regional income disparities in advanced countries. Note that even Myrdal, who has been regarded as a typical proponent of the disequilibrium perspective, paid attention to the possibilities of temporal reduction in the disparities (Myrdal, 1957, pp. 37-38). Therefore, taking the experiences of the 1980's into consideration, the findings below must have the significance of revisiting the existing explanatory frameworks.In the second section, the long-term trend of regional income inequalities in Canada during the period 1951-1989 is clarified. To avoid a one-sided approach associated with a particular measure, the changing inter-provincial differentials of personal income per capita are specified using Gini's coefficient and the coefficient of variation (Figure 2), and Theil's measure (Figure 3). The curves in the figures show a generally similar trend: the disparities gradually decreased from the early 1950's to the late 1960's and dropped sharply in the early 1970's. During the decade since the mid-1970's, it was rather steady in spite of a slight expansion at the beginning of the 1980's. Worthy of note in the second half of the 1980's is an obvious widening of the diparities. Such trend resembles those found in other advanced countries such as the United Kingdom and Japan.As shown in Figure 3, Theil's measure enables us to decompose the observed total inequalities into three additive components: the disparities within the eastern provinces (Newfoundland, Prince Edward Island, Nova Scotia, New Brunswick and Quebec), the one within western provinces (Ontario, Manitoba, Saskatchewan, Alberta and British Columbia) and the one between these two regions. The converging total disparities is, to a certain degree, attributed to the third component of the disparities, which is the largest contributor among the three and showed a remarkable reduction from the 1950's to the 1970's. The expanding disparities within western provinces during the 1980's are responsible for the diverging total disparities.In the third section, the role of fiscal transfers on the changing regional income inequalities is examined. Myrdal and Williamson suggested that central government's policies, including fiscal transfers, can reduce regional disparities. This suggestion is verified here. Province-specific regression analysis is undertaken to explain the trend of regional disparities (Tt), shown at Figure 3, by the transition of the rate of government transfer income in provincial personal income (Gpt). The results obtained based on the equation (6) (see Table 2) imply that, while increase of the government transfer in personal income has raised a particular province's income level to near the national level for the cases of the Atlantic Provinces and Quebec