The basic structural features of the Soviet planned economy emerged in the period following the termination of the New Economic Policy (NEP) and the launching of the First Five Year Plan. With respect to the structure of prices, three principles were formulated as a guide to the formation of industrial wholesale prices. The first is the principle of average cost pricing; the price of a product is based on the average cost of its production in all the enterprises in that branch of industry, plus a normal profit markup of 4 to 5 percent over cost. The second is the principle of permanent prices; once the price is assigned to a product, it endures without limit of time, although from time to time prices are revised. The third is the principle of uniformity; the price of a given product is the same for all sellers and all purchasers. For decades these three principles have been presented in the general literature as the basis of price formation. In fact, certain departures were introduced in the very first years in which they were being formulated, but in the course of time the departures became increasingly massive, and the three principles have become less and less useful as a guide to the actual basis of price formation.

A variety of considerations have contributed to the evolution of price policy. There is, however, one common thread that unifies much of the history of that evolution.

Introduction

At work, women in the USSR earn on the average less than two-thirds of a man's wage; at home they spend more than twice as much time on household chores as do their spouses. In these two respects Soviet urban society resembles urban societies in most developed market economies. However, Soviet women are much more active in the labor force. When coupled with unusually high participation in the labor force, undiminished, perhaps even greater, work at home imposes on women in the Soviet Union an exceptionally heavy double burden.

For women to bear such a heavy work load and then to receive low pay – compared to men – would appear to be in direct conflict with the egalitarian principles of socialist ideology. The obvious embarrassment created by this conflict may be one reason for the shyness of Soviet sources to provide data on male/female wage differentials. On the other hand, the broader questions of the status of women in the family and in society, the double burden of married women, restricted access to various occupations, high participation in the labor force – together with the possible demographic effects of these phenomena – are dealt with in dozens of books and scores of articles by Soviet scholars. Many of these studies emphasize how greater equality for women has been achieved with respect to education, occupational choice, income independence, and social and political activity.

The Soviet Union's system of resource allocation combines the method of direct determination of outputs by the center with the method of central determination of parametric prices and rules of behavior. Direct allocation is used for subaggregates of products, and centrally determined prices – which, because of their considerable stickiness, cannot constitute a “dual” – are used parametrically to determine the product mix of these subaggregates. Why this reliance by the center on both physical planning and centrally determined parametric pricing?

The answer which first suggests itself is that parametric prices are used to deal with a level of disaggregation that cannot be handled by physical planning. Thus, the two methods complement one another. However, given that their stickiness prevents centrally determined prices from acting as a “dual,” why does the center not allow the parametric prices that determine the demand and supply of individual items within each subaggregate to be determined on the marketplace? Not only would this procedure have certain optimizing features currently absent, but it could also be used – if this were desired – to increase the degree of planners' sovereignty over the product mix of consumers' goods.

The explanation that is offered in this chapter for the continuation since the beginning of the 1930s of the foregoing features of the Soviet planning system is that there exists a set of constraints, embodied in the center's objective function, which has not been considered in either the Soviet or Western literature.

Introduction

Modelling the demand for consumer durables is not one of the easiest topics in applied economics. Much of the most creative work in the field was done in Cambridge in the 1950s in the group around Stone. Thus the classic paper by Farrell (1954) was the first systematic application of discrete choice theory to the problem and made a notable contribution also in analysing the interaction of the markets for new and used cars. Under the direct influence of Stone, Cramer (1957), in another classic, first put forward a neoclassical model integrating the demand for durable and non-durable goods with the life cycle theory of Ramsey (1928), Fisher (1930), Tintner (1938) and Modigliani and Brumberg (1955). The essence of the model lies in the assumptions that the budget constraint is linear and known with confidence and that, in efficiency-corrected units, new and used durables are perfect substitutes. Stone and Rowe (1957) simultaneously with Chow (1957) first applied the stock adjustment model to the demand for durables. The latter remains the most popular tool of analysis for aggregate time-series data though more recently Smith (1974; 1975) and Westin (1975) have put forward the ‘discretionary replacement’ model as a simple alternative. Though the neoclassical model of investment has been widely applied since Haavelmo (1960) and Jorgenson (1963), application to consumer durables have been less frequent. Diewert (1974) is one and contains a useful discussion of the theory.

In the USSR, consumption and demand undergo both constant growth in volume and qualitative structural change. The total growth of consumption and demand may be illustrated by the data on retail sales which by 1978 had increased 2.4-fold over the 1965 level. In addition, consumption and demand patterns constantly undergo considerable qualitative changes due to increases in the consumption of, and demand for, high-calorific food, for non-foods that satisfy developing needs, and for products that make house-keeping easier and save time. Thus the total sale of meat and meat products increased 2.1 times between 1965 and 1977. For the same period the corresponding figure for milk and milk products was 2.1, for eggs 3.7, and for fruit 2.2. At the same time, the consumption of bread and potatoes per capita has decreased and the level of sugar and vegetable oil consumption has been constant, in accordance with physiological standards.

Among non-foods the demand for knitted garments and carpets had the highest rate of growth during this period (i.e. 1965–77) increasing 3.3-fold and 5.7 respectively. Furniture sales increased 2.3-fold, articles for cultural and domestic needs 2.4-fold and so on. The provision of most durable goods has also been improved. Thus the supply of TV sets increased 3.3 times during the period, refrigerators by 6.5 times, washing machines by 3.3 times, vacuum cleaners by 3.1 times and so on.

It is a matter of common observation that the quantities consumed of many goods and services are not directly under the control of those who consume them. The level of provision of public goods cannot be varied to taste by any single consumer: shortages or formal rationing of market goods may place an upper limit on consumption: transactions costs or market imperfections, particularly in asset markets, may prevent the short-run adjustment of stocks to their optimum levels so that consumers may have to consume too much as well as too little. Perhaps most importantly, involuntary unemployment in the labour market can be thought of as an enforced consumption of an undesirably large amount of leisure. All these situations involve quantity constraints on consumer behaviour, and although rationing is only one possibility, we shall use the term to deal with all, including those situations where more is consumed than would be freely bought.

As one might expect, much of the early work on rationing was done during and immediately after the second world war. This work is surveyed in the classic paper by Tobin (1952). For a considerable period subsequently, there appeared to be little interest in the subject and little was published, although see the two papers by Pollak (1969; 1971). In the last few years, however, rationing has once again become a major focus of attention.

In a volume dedicated to Sir Richard Stone, it is appropriate that first consideration should be given to the theory and measurement of commodity demands. Sir Richard's own great monograph, The Measurement of Consumers' Expenditure and Behaviour in the United Kingdom [51] retains its classic status in applied econometrics to this day. The research programme established there and in the 1954 Economic Journal article [48] on the linear expenditure system is still flourishing and the five papers in part one represent several aspects of it.

The first set of topics concern the appropriate choice of functional form for empirical demand equations. In [51], Sir Richard and his coworkers adopted a largely pragmatic approach using a loglinear constant elasticity form. This has great advantages in computation and allows a much more flexible research strategy than is possible with more complex non-linear equations in which all commodities are dealt with simultaneously. However, as has been known for a long time, loglinear demand functions for all commodities are inconsistent with utility theory in that they cannot permit the predicted demands to add up to the predetermined sum of expenditures. This reflects a quite general problem: how do we choose functional forms which are convenient to work with, which allow the easy incorporation of such information as we possess about the nature of individual demands, and which are consistent with the theory? The first two papers in this section are addressed to that question.

Introduction

The impetus for our work on the timing and spacing of children has come from two surveys done by the University of Montreal in 1971 (Henripin and Lapierre-Adamcyk, 1974 and 1975). These surveys are unusual in that they contain questions on work experience before marriage, after marriage but before the birth of the first child, at the time of the interview, and the number of years worked after marriage. The questions enable one to reconstruct the proportion of a woman's time spent working during the child-rearing period. The usual questions are asked concerning socioeconomic background and pregnancy history. Because the time of the mother spent with her children is thought to be an important determinant of child ‘quality’ – begging the question of just what that is – and because female labour force participation is known to be greatly inhibited by the presence of young children (Sweet, 1973), it was clear to us that we had an almost unique opportunity to explore the joint relationship among the timing and the spacing of children and female labour force participation. In addition, the surveys contained an impressiveset of questions related to the couple's preferences for children. These questions included not only the usual inquiry concerning the ideal number of children and the number of children wanted by the couple, but also more abstract questions concerning couples in general, and questions related to preferences about the timing and spacing of children.

Sir Richard Stone retired from his chair in Cambridge in September 1980. To mark the occasion, this volume has been written in his honour. It is not a festschrift after the usual mould, where friends and colleagues contribute a diverse collection of papers. Sir Richard's achievements have been too broad and his disciples too many to permit a single collection along such lines. Instead, I have taken one single field in which Sir Richard has been preeminent, and attempted to bring together a first-rate collection of papers in that field. Many of the authors are close friends or ex-colleagues of Sir Richard's, but several have had little more than professional contact. However, all are indebted to him through his scientific work, and in contributing to this volume are united in their wish to honour him and to acknowledge their indebtedness. In editing the volume it is my hope that the best way of honouring Sir Richard and commemorating his retirement is the preparation of a volume of the best current work in the economics of consumer behaviour. The papers published here are representative of a wide range of contemporary research in the field and only a few important topics are not covered at some point. They provide a good indication not only of the state of the art but also of the extraordinary area over which Sir Richard's own work has been an influence.

Introduction

A price index refers to a pair of consumption periods, and price-index formulae usually involve demand data from the reference periods alone. When there are many periods, a price index can be determined from any one period to any other, in each case using the data from just those two periods. But then consistency questions arise for the set of price indices so obtained. Especially, they must have the consistency that would follow from their being ratios of ‘price levels’. The well-known tests of Irving Fisher have their origin in such questions. When these tests are regarded as giving identities to be satisfied by a standard formula and are taken in combination, it is impossible to satisfy them. Such impossibility remains even with partial combinations. Eichhorn and Voeller (1976) have given a full account of the inconsistencies between Fisher's tests. Reference is made there for their results and for the history of the matter.

Fisher recognizes the consistency question also in his idea of the ‘rectification of pair comparisons’. In this the price indices are all calculated, as usual, separately and regardless of any consistency they should have together, and then they are all adjusted in some manner so that they can form a consistent set. For instance, by ‘crossing’ a formula with its ‘antithesis’ you got one that satisfied the ‘reversal’ test. Here he takes one of the tests separately as if any one could mean anything on its own, and contrives a formula to satisfy it.

The theory and measurement of economic index numbers presents side by side some of the most difficult and abstruse theory with the most immediately practical issues of everyday measurement. The construction of index numbers is an essential part of all social accounting; without compression and aggregation the mass of quantities and prices thrown up by the economic system would be incomprehensible. Yet from the outset such aggregation has been known to be meaningful only in the context of welfare measurement. But to what extent are welfare-based index numbers practical? In his book on index numbers [56] written for the OEEC (now OECD), Sir Richard Stone addressed the question of whether practical international standards for index number construction could be established in line with his earlier standardized system of national accounts. Sir Richard gives the following reasons why the welfare approach is useful:

Introduction

The objective of this chapter is to discuss the relationship between the technique of principal components, first applied to economic data in the pioneering paper by Stone (1947), and the analysis of consumer demand.

The purpose of principal component analysis in statistics is to formulate a set of variables that are in some way ‘more basic’ than the observed variables. Factor analysis has a similar objective. Starting in 1956, Gorman and his associates (Boyle et al., 1977; Gorman, 1956; 1959; 1976b) applied these statistical techniques to consumer demand in order to assess the consumer's basic wants, while Lancaster (1966, 1971) and Becker (1965) pursued similar goals by extending the economic theory of the consumer rather than using statistical tools. The independence transformation, which originated with Theil (1967; 1975–76; 1977), Brooks (1970), and Laitinen and Theil (1978), is related to both approaches and is in fact ‘between’ them. The transformation requires no extension of the theory of the utility-maximizing consumer, although it can handle such extensions without difficulty (see the example on leisure in section 2). At the same time, the transformation has a simple statistical interpretation, viz., that of a constrained principal component transformation.

To provide adequate motivation, we discuss some examples in section 2. Sections 3 and 4 describe the preference independence transformation and the underlying differential approach to consumption theory. Section 5 gives a brief discussion of similar results in the theory of the firm.

One of the most difficult areas in consumer theory is that which concerns itself with intertemporal choice. Theory which ignores uncertainty lacks credibility while theory taking it into account is complex and hard to implement. On the empirical side, the analysis of the consumption function and of durable goods presents all the classic time-series problems of dynamics, seasonally and serial correlation. Even after 40 years of continual econometric activity, and in spite of their contemporary treatment as standard classroom examples of applied econometric analysis, both durable goods and consumption functions are still subject to lively controversy.

Much the most influential work on what is still called the ‘modern’ theory of the consumption function is that of Modigliani and Brumberg (1955a, b) on the life-cycle model, although the ideas go back to Fisher and Ramsey. The life-cycle model has the inestimable advantage of viewing the theory of the consumption function as a part of consumption theory in general. This work, together with that of Friedman (1957), provided the basic model of the consumption function which has dominated theoretical and empirical discussion ever since. In its applied form at least, the basic regression is one of consumption on its own lagged value and on income although a number of other variables (lagged income, wealth, liquid assets, the distribution of income) make periodic appearances.