Considerable difficulty is presented for students of classical economics by the ambiguous use of the term “capital.” In particular there is much room for debate concerning the importance of technological capital as distinct from wage-goods capital in classical economic literature. This problem exists in the Wealth of Nations as well as in the work of Adam Smith's successors.

Malthus was quick to point out that serious error may result from Smith's analysis of wages wherein “every increase in the stock or revenue of the society” is regarded as “an increase in the funds for the maintenance of labour.” More recent writers recognize the references to “fixed capital” but imply that mere Up service was paid to such items; the emphasis was predominantly upon wage goods. In particular, it has been argued, Smith emphasized “the superior physical productivity of time-using processes” and the requirement that subsistence goods be available to permit the adoption of such techniques. “These accumulations are the simplest form of capital; machinery, tools, stocks of all sorts and human skill and knowledge are simply more complicated forms.

Perhaps the strongest criticism is by Marx. Marx did not deny that “constant” capital makes an appearance in the Wealth of Nations, but, he argued, it does so only at the micro-economic level of analysis: “Adam Smith, by a fundamentally perverted analysis, arrives at the absurd conclusion, that even though each individual capital is divided into a constant and a variable part, the capital of society resolves itself only into variable capital, i.e., is laid out exclusively in payment of wages.”

Previous studies have noted that Canadians tend to support one political party at the provincial level of government and another party at the federal level, yet no attempt has been made to discover which voters switch their electoral support. This study is concerned with an examination of one electoral district as both a federal and a provincial constituency in two 1964 elections to determine (a) the nature and degree of vote switching from one election at one level of government to a second election at another level and (b) the characteristics of the people who switched their votes.

The city of Saskatoon was selected for the study for two closely interrelated reasons. First, two elections were held remarkably close together in time: a provincial general election on April 22, 1964, followed by a federal by-election on June 22, 1964. Second, Saskatoon had returned CCF members to the provincial legislature since 1944, but had elected as its member of parliament the Progressive-Conservative candidate since 1957. The Conservative candidates had a poor record in provincial elections, generally running a modest third, while in federal elections the party's candidate polled more votes than his combined opposition in 1958, 1962, and 1963. As Table I indicates the voters of Saskatoon had on six occasions in the period from 1956 to 1963 consistently supported the CCF provincially and the Progressive-Conservatives federally. The 1964 elections followed basically the same pattern The main differences were that provincially the Liberals and Progressive-Conservatives improved their position considerably in comparison with past elections (with the Liberals electing one member in the five-member district), while federally the Progressive-Conservative vote dropped substantially with a lower voter turn-out than had been the case in recent federal elections (see Table II).

A Robinsonian model is a two-sector model with fixed technical coefficients. Aside from Mrs Robinson's own work, similar models, without technical change, have been discussed in two major papers, one by I. M. D. Little and the other by R. Findlay. Mrs Robinson has criticized both these contributions on the ground that they do not treat the rate of growth or accumulation as an exogenous variable, but rather regard it as endogenous, emerging from the model after the real-wage rate has been given. In the first section of this paper I explore the conditions for stable equilibrium growth on Mrs Robinson's assumption of a desired rate of accumulation (a function of “animal spirits” and the rate of profit), and then add to the model neutral technical change of an appropriate sort.

In section II, the effects of introducing biased technical change are examined. I show that the model has “technological stability,” by which I mean that there is a tendency for technical change to be neutral. Predictions and tests about the movement over time of the consumption-good price of capital are made in section III. Finally, I end with a short comment on stability in the short run.

In their efforts to analyse the behaviour of households and firms, economists have traditionally assumed that each of these units maximizes an index, such as a utility function, a profit function, or a present value function, which is taken to characterize its own interest. However, in analysing government policies these same economists have generally shied away from the maximization-of-own-interest assumption and have instead assumed that governments maximize a social welfare function or the common good; or they have proceeded by developing ad hoc hypotheses for each different problem. This is regrettable, for the returns to the maximization-of-own-interest assumption in the analysis of household and firm behaviour have been impressive, even when due consideration is given to the simplification of reality that is involved. One cannot help suspecting that they might also be high in the analysis of government behaviour.

In consequence, in the present paper, following the lead of Schumpeter and Downs, I suppose that governments are not primarily interested in maximizing a social welfare function, or the common good, or the public interest—whatever these may be—but that instead they seek to maximize their own interest. Specifically, following Downs and Wilson, I assume that governments maximize the probability of their re-election, a probability which is dependent on the policies that are implemented as well as on the degree to which they are implemented. This relationship rests on the view that governments and their electorates are engaged in a sort of exchange in which policies are traded for votes; to maximize the probability of their re-election, governments endeavour to “produce” those policies which can be exchanged for the largest possible quid pro quo, that is, for the largest number of votes. Thus an analysis of government behaviour following a change in the preferences of consumer-voters, for example, would be to a large extent an analysis of changes in government policies. The potential number and variety of these policies, however, is so large that at first glance it seems impossible to proceed with a systematic analysis.