UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Neoclassical economics and labour migration theory : a Canadian perspective Olligschlaeger, Andreas Matthias 1986

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1986_A8 O44.pdf [ 7.89MB ]
Metadata
JSON: 831-1.0097140.json
JSON-LD: 831-1.0097140-ld.json
RDF/XML (Pretty): 831-1.0097140-rdf.xml
RDF/JSON: 831-1.0097140-rdf.json
Turtle: 831-1.0097140-turtle.txt
N-Triples: 831-1.0097140-rdf-ntriples.txt
Original Record: 831-1.0097140-source.json
Full Text
831-1.0097140-fulltext.txt
Citation
831-1.0097140.ris

Full Text

NEOCLASSICAL ECONOMICS AND LABOUR MIGRATION THEORY: A CANADIAN PERSPECTIVE by ANDREAS MATTHIAS OLLIGSCHLAEGER B . A . , Concordia University, 1984 A T H E S I S S U B M I T T E D I N P A R T I A L F U L F I L L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F M A S T E R O F A R T S in T H E F A C U L T Y O F G R A D U A T E S T U D I E S Department of Geography, University of Brit ish Columbia We accept this thesis as conforming to the required standard T H E U N I V E R S I T Y O F B R I T I S H C O L U M B I A J U L Y 1986 (c) Andreas Matthias Olligschlaeger, 1986 In p r e s e n t i n g t h i s t h e s i s i n p a r t i a l f u l f i l m e n t o f the r e q u i r e m e n t s f o r an advanced degree a t the U n i v e r s i t y o f B r i t i s h C o l u m b i a , I agree t h a t the L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r r e f e r e n c e and s t u d y . I f u r t h e r agree t h a t p e r m i s s i o n f o r e x t e n s i v e c o p y i n g o f t h i s t h e s i s f o r s c h o l a r l y purposes may be g r a n t e d by the head o f my department o r by h i s o r her r e p r e s e n t a t i v e s . I t i s u n d e r s t o o d t h a t c o p y i n g o r p u b l i c a t i o n o f t h i s t h e s i s f o r f i n a n c i a l g a i n s h a l l not be a l l o w e d w i t h o u t my w r i t t e n p e r m i s s i o n . Department o f The U n i v e r s i t y o f B r i t i s h Columbia 1956 Main Mall V a n c o u v e r , Canada V6T 1Y3 Date ^ x f x ^ DE-6 (3/81) ABSTRACT This thesis examines the theoretical and empirical base of neoclassical migration analysis in economic geography. It is shown that the key assumptions of neoclassical migration analysis stem from the broader marginal equilibrium analysis and theory of resource allocation that defines the neoclassical school. Spefically, the hypothesis that neoclassical economics makes with respect to labour migration is that labour flows from low wage, high unemployment regions to regions with high wages and low unemployment, thus arriving at a state of equilibrium. This hypothesis is tested using Canadian labour migration data for 1976-1981. It is found that the hypothesis is unable to explain labour migration patterns in Canada because: first, the assumptions about human behaviour that the neoclassical model makes are both too simplistic and unrealistic, as are those about the nature of the economy, and second, migration seems to promote cumulative causation rather than move the system towards equilibrium. ii TABLE OF CONTENTS Abstract ii List of Tables v List of Figures ix Acknowledgement x 1. Introduction 1 2. Neoclassical Economics and Economic Geography 3 2.1 The Advent of Neoclassicism 3 2.2 Major Characteristics of Neoclassical Economic Theory 7 2.3 Criticisms of Neoclassical Economic Theory 12 2.4 Geographical Applications of Neoclassical Economics 14 3. Neoclassical Economics and Labour Migration 20 3.1 The Foundations of Neoclassical Labour Migration Theory 20 3.2 Empirical Application of Neoclassical Labour Migration Theory 22 3.2.1 Assumptions of Labour Migration 23 3.2.2 Geographical Models of Labour Migration 24 3.2.3 Empirical Applications 27 3.2.3.1 Micro-Adjustment Models 27 3.2.3.2 Macro-Adjustment Models 28 4. Data and Methodology 32 4.1 Data 32 4.2 Methodology 35 4.2.1 Theoretical Considerations 35 4.2.2 Methodology 38 5. Data Analysis 43 iii 5.1 Gross Migration Flows 43 5.2 Labour Force Adjusted Migration Rates 45 5.3 Migration Rates Adjusted for Labour Force, Distance and Intervening Opportunities 57 5.4 Wages, Unemployment Rates and Adjusted Migration Rates 69 5.5 Net Migration and Changes in Wage and Unemployment Rates 73 5.6 Summary of Findings 76 6. Accounting for Failures of the Neoclassical Migration Model 77 6.1 Maximization 77 6.2 Temporal Considerations 84 6.3 Social Constraints to Labour Migration 86 6.4 Labour Markets 87 6.5 Government Intervention 88 6.6 Spatial Equil ibrium 93 7. S u m m a r y and Conclusion 96 8. Bibliography .• 98 iv LIST OF TABLES 4.1.1 Amalgamation of Occupational Groups 34 5.1.1 Gross Migration, 1976-1981, A l l Occupations 44 5.1.2 Net Exchange Matr ix , 1976-1981, A l l Occupations 44 5.2.1 Migration Rates Adjusted for Size of the Labour Force, 1976-1981, Managerial , Administrative 46 5.2.2 Migration Rates Adjusted for Size of the Labour Force, 1976-1981, Natura l Sciences .' 46 5.2.3 Migration Rates Adjusted for Size of the Labour Force, 1976-1981, Social Sciences47 5.2.4 Migration Rates Adjusted for Size of the Labour Force, 1976-1981, Teachers 47 5.2.5 Migration Rates Adjusted for Size of the Labour Force, 1976-1981, Medicine & Health 48 5.2.6 Migration Rates Adjusted for Size of the Labour Force, 1976-1981, Fine & Commercial Artists 48 5.2.7 Migration Rates Adjusted for Size of the Labour Force, 1976-1981, Clerical Workers 49 5.2.8 Migration Rates Adjusted for Size of the Labour Force, 1976-1981, Sales Occupations 49 5.2.9 Migration Rates Adjusted for Size of the Labour Force, 1976-1981, Service Occupations 50 5.2.10 Migration Rates Adjusted for Size of the Labour Force, 1976-1981, Agriculture.. .50 5.2.11 Migration Rates Adjusted for Size of the Labour Force, 1976-1981, P r i m a r y Occupations 51 5.2.12 Migration Rates Adjusted for Size of the Labour Force, 1976-1981, Processing 5.2.13 Migration Rates Adjusted for Size of the Labour Force, 1976-1981, Manufacturing 52 5.2.14 Migration Rates Adjusted for Size of the Labour Force, 1976-1981, Construction.52 5.2.15 Migration Rates Adjusted for Size of the Labour Force, 1976-1981, Transportation 53 5.2.16 Migration Rates Adjusted for Size of the Labour Force, 1976-1981, Other Occupations 53 5.2.17 Migration Rates Adjusted for Size of the Labour Force, 1976-1981, A l l Occupations 54 5.3.1 Regression of Labour Force Adjusted Migration Rates with Distance and Intervening Opportunities, by Province of Origin 56 5.3.2 Migration Rates Adjusted for Size of the Labour Force, Distance and Intervening Opportunities, 1976-1981, Managerial , Administrative 59 5.3.3 Migration Rates Adjusted for Size of the Labour Force, Distance and Intervening Opportunities, 1976-1981, Natura l Sciences 59 5.3.4 Migration Rates Adjusted for Size of the Labour Force, Distance and Intervening Opportunities, 1976-1981, Social Sciences 60 5.3.5 Migration Rates Adjusted for Size of the Labour Force, Distance and Intervening Opportunities, 1976-1981, Teachers 60 5.3.6 Migration Rates Adjusted for Size of the Labour Force, Distance and Intervening Opportunities, 1976-1981, Medicine & Health 61 5.3.7 Migration Rates Adjusted for Size of the Labour Force, Distance and Intervening Opportunities, 1976-1981, Fine & Commercial Artists 61 5.3.8 Migration Rates Adjusted for Size of the Labour Force, Distance and Intervening Opportunities, 1976-1981, Clerical Workers ; 62 vi 5.3.9 Migration Rates Adjusted for Size of the Labour Force, Distance and Intervening Opportunities, 1976-1981, Sales Occupations 62 5.3.10 Migration Rates Adjusted for Size of the Labour Force, Distance and Intervening Opportunities, 1976-1981, Service Occupations 63 5.3.11 Migration Rates Adjusted for Size of the Labour Force, Distance and Intervening Opportunities, 1976-1981, Agriculture '. 63 5.3.12 Migration Rates Adjusted for Size of the Labour Force, Distance and Intervening Opportunities, 1976-1981, Pr imary Occupations 64 5.3.13 Migration Rates Adjusted for Size of the Labour Force, Distance and Intervening Opportunities, 1976-1981, Processing 64 5.3.14 Migration Rates Adjusted for Size of the Labour Force, Distance and Intervening Opportunities, 1976-1981, Manufacturing 65 5.3.15 Migration Rates Adjusted for Size of the Labour Force, Distance and Intervening Opportunities, 1976-1981, Construction Trades 65 5.3.16 Migration Rates Adjusted for Size of the Labour Force, Distance and Intervening Opportunities, 1976-1981, Transportation 66 5.3.17 Migration Rates Adjusted for Size of the Labour Force, Distance and Intervening Opportunities, 1876-1981, Other Occupations 66 5.4.1 Significance of Interregional Differences in Wage and Unemployment Rates on Residual Migration, by Province of Origin 70 5.4.2 Significance of Interregional Differences in Wage and Unemployment Rates on Reisdual Migration, by Occupation 72 5.5.1 Significance of Net Migration on Changes in Wage and Unemployment Rates, 1976-1981, by Province 74 vii 5.5.2 Significance of Net Migration on Changes in Wage and Unemployment Rates, 1976-1981, by Occupation 75 6.5.1 C M M P Generated Migration as Compared to the Total Internal Migration in Canada, 1966-1976 90 6.5.2 C M M P Generated Migration as Compared to the Total Internal Migration, by Province, 1976-77 and 1977-78 90 6.5.3 The Effects of Selected Changes in Fiscal Structure on Out-Migration Rates F r o m the Atlantic Provinces Including Intra-Atlantic Moves, Lower Income Class, 1968-1977..92 viii LIST OF FIGURES 2.2.1 Feasible Resource Allocation as an Intersection of Resource Constraints 10 2.4.1 Bid-Rent Functions and Locational Equil ibrium 15 2.4.2 Concentric Zone Model of L a n d Use 15 ix ACKNOWLEDGEMENT The two past two years at the University of British Columbia have been most enjoyable. In particular I would like to thank m y supervisor, D r . Trevor Barnes, for his support and patience during all phases of m y studies. Thanks also go to m y second- reader, D r . K e n Denike, for providing help with the methodology. I am particularly indebted to D r . James W . Young, m y undergraduate supervisor at Concordia University, for awakening m y interest in economic geography and - apart from giving me the occasional well-deserved kicks in the rear end - for being such an outstanding teacher. Above all, however, I would like to thank m y parents, Rudolf and Hildegard Olligschlaeger, who have sacrificed so much in providing for their childrens' education, for their affection, love and understanding. It is to them that I dedicate this thesis. Final ly, my graduation from this university marks the end of a six year stay in Canada. During those six years Canadians have at all times made me feel welcome and at home. So, like thanks a lot, eh? x 1. INTRODUCTION. E v e r since the departure from the descriptive type of economic geography (for some examples see Dicken, 1955 and Alexandersson & Norstroem, 1963) in the late 1950s and early 1960s, neoclassical economics has played a major role in the field. This is recognized by K i n g (1979, p.34), who characterizes most of current economic geography as: "positivistic, decidedly neoclassical in its economic foundations, and only weakly spatial". Whereas neoclassical economics has been heavily criticized for quite some time by economists (Robinson, 1962; Dobb, 1973; Rowthorn, 1974), it is only within the last 15 years that economic geographers have begun to question the neoclassical theoretical foundations of their own discipline (Massey, 1973; Sayer, 1976; Clark, 1982, 1983; Clark & Gertler, 1983). To this date two alternative theories have been proposed: the Marxis t school of economic geography (Harvey, 1982, 1985) and more recently the neo-Ricardian one (Barnes, 1984, 1985; Barnes & Sheppard, 1984). Labour migration has been the focus of much attention in economic geography and regional science. This is largely due to the fact that it has traditionally played an important role in regional development and in the spatial economy as a whole. While there are many approaches to labour migration modelling, much of contemporary migration analysis is based upon neoclassical economic theory. Specifically, the notions of utility and resource allocation which are crucial to the neoclassical scheme, are often key assumptions in the analysis of labour migration. These assumptions, however, have come under heavy criticism in recent years (Greenwood, 1975; Clark & Ballard, 1979, 1980; Clark, 1982, 1983; C u r r y , 1985). Thus many empirical studies have shown that such assumptions do not hold in the "real" world. However, in spite of the emergence of the Marxis t and neo-Ricardian schools of economic geography, to date no rigorous alternative theory to neoclassical labour migration analysis has been proposed. This thesis will focus on the neoclassical theory of labour migration and test it using Canadian interprovincial labour migration data for the period 1976-1981. The second chapter 1 of the thesis will provide an overview of the development of the neoclassical school of thought and show how its principles have been applied in economic geography. The third chapter will establish the link between neoclassical economics and labour migration, and also outline two empirically testable hypotheses that stem from the neoclassical analysis. The fourth chapter will discuss the data and methodology employed in testing the two hypotheses. Chapter five analyses the data, and concludes that the neoclassical hypotheses concerning labour migration are invalid at least for Canada during this period. Finally, chapter six will critically examine the neoclassical contentions about labour migration and suggest why the theory does not hold. 2 2. N E O C L A S S I C A L E C O N O M I C S A N D E C O N O M I C G E O G R A P H Y . To understand neoclassical migration analysis, it is first necessary to be familiar with the roots of neoclassicism itself. This introductory chapter consists of four sections. First , a brief history of neoclassical economics is presented. It will concentrate on those features of neoclassicism that particularly have been employed in economic geography. The second section will outline the major characteristics of neoclassical economic theory. The third section will critically examine these characteristics. Final ly , the fourth section will present some geographical applications of neoclassical economics. 2.1 T h e A d v e n t of N e o c l a s s i c i s m . Neoclassical theory arose during the 1870's (Walsh & G r a m , 1980). Although the term "neoclassicism" itself was not coined until the early twentieth century. Three authors and their works are commonly associated with its origin: Wil l iam Stanley Jevons {"Theory of Political Economy", 1871), Car l Menger ("Grundsaetze der Volkswirtschaftslehre", 1871) and Leon Walras ("Elements of Pure Economics", 1874) (Dobb, 1973). These three works were essentially a reaction to the classical school of economic thought, as postulated by Smith, Ricardo and Malthus (Walsh & G r a m , 1980). F o r classical economists the primary concern was the reproduction of the economy and the maximization of surplus, where surplus is defined as total output minus those commodities that are used up in the production of that output. In turn the surplus is divided among three social classes: workers, landlords and capitalists. The problem of classical economics was then to locate the mechanisms which determine what proportion of the surplus each class receives. Another feature of the school was its emphasis on the conditions of production. Demand does not play a role because there are constant returns to scale, i.e., changes in output do not affect per unit costs. Neoclassical economics, on the other hand, emphasizes exchange rather than production. Only two sets of actors exist in the neoclassical world: producers and consumers. Prices are set through the demand and supply. While consumers are motivated to buy 3 (demand) goods in order to acquire utility, producers are motivated to sell (supply) goods in order to acquire profits. Through the action of supply and demand - the market mechanism -resources are then allocated to the best possible uses. A s will be illustrated below, Jevons, Walras and Menger each made a unique contribution to this new line of thought, which has been more generally termed the "marginal revolution". Wil l iam Stanley Jevons' (1835-1882) major work, the "Theory of Political Economy" published in 1871, was essentially devoted to the determination of the mechanics of self-interest and utility. The major novelty of Jevons' theory lay in the singling out of the "final degree of utility" and the subsequent equating of this with the exchange value of commodity. In other words, the value (price) of a good depended on how much pleasure ("utility") it would give to an individual. Given that the "utility" of an individual and the value of a commodity are quantifiable factors, Jevons argued that any theory dealing with those variables should necessarily be mathematical (Dobb, 1973). These ideas were directly opposed to the beliefs of the classical economists who postulated that value was uniquely determined by the costs of production, and those of M a r x , who implied that it was determined by the value of labour (labour theory of value). This was most likely one of the major reasons why economists at the time were so reluctant to accept Jevons' principles (Howey, 1973). In fact, Jevons was better known to his contemporaries as an applied economist, rather than as a theorist (Deane, 1978). In constructing a theory that deals with the exchange of commodities in light of consumers' wants and needs, Jevons used a partial equilibrium approach. The term "partial equilibrium", as opposed to "general equilibrium", implies that onty a few selected relationships are examined in detail, while other relationships are assumed to be constant (Hoover, 1978). F o r example, if we want to identify the effect of substituting between various resources in a production process, then the supply price of those resources is fixed at a constant level. General equilibrium theories, on the other hand, simultaneously take into consideration all interdependent variables such as cost schedules, demand functions and prices. 4 C a r l Menger (1840-1921) was the founder of the Austr ian school of neoclassical economic thought. Unlike, Jevons, he was widely recognized for his writings during his lifetime. Menger's theory, put forward in his "Grundsaetze der Volkswirtschaftslehre", published in the same year as Jevon's "Theory of Political Economy", was very similar to that of Jevons, except in that it was not mathematical . The objective of Menger's "Grundsaetze" was to sketch a theory of economic development (Streissler, 1973). The main thrust of his argument was that economic development is brought about by increases in human welfare, which is defined as the "constant widening of the range of goods and the improvement of their quality, i.e. changes in the productive output" (Streissler, 1973, pp. 164-165). In the process of constructing his theory of development, Menger also constructed a theory of value which was very close to that of Jevons. The value of a "first order" (consumer) good was derived from its power of satisfying human wants, whereas the value of a "higher order" (producer) good was determined by its contribution to the production of goods that cater directly to human wants (Dobb, 1973). In both cases the method used to derive the value was the so-called "loss principle", where the value of a good is defined as the loss of satisfaction a person experiences if he or she has to do without it. In effect this is identical to the marginal utility of a good: if a person is deprived of a good then the decrease in the overall utility of that person is equal to the marginal utility times the quantity of the good that is lost. Unlike Jevons, Menger's concern was the detailed study of an economy which is always out of equilibrium. Menger believed that because of information uncertainty, disequilibrium is the norm within an economy. This belief also accounts for Menger's rejection of the standard simplifying assumption of a single equilibrium price in a market. Furthermore, it has also been judged to be the reason why he did not use mathematics. Leon Walras (1834-1910) was the third, and perhaps most influential, of the three founders of neoclassical economic theory. Walras began work on his theory in 1859, but did 5 not complete it until 1872 (Jaffe, 1973). His major work, "Elements of Pure Economics", finally was published in 1874. Walras's theory was expressed in mathematics . One of his main achievements, in fact, was the "synthesis of various aspects of the new approach into a mathematical system of mutual dependence" (Dobb, 1973). The establishment of a "system of mutual dependence" refers to a situation of general equilibrium, where all the central variables are considered simultaneously. This is in direct contrast to Jevons and Menger, with the former adopting a partial equilibrium approach and the latter assuming that the system is constantly out of equilibrium. In his "Elements of Pure Economics" Walras assumed that given first, certain predetermined quantities of productive services and, second, free competition, there will be three so-called "natural effects" (Jaffe, 1973): (1) certain definite quantities of various products; (2) a definite price of each product at each moment in time; and, (3) a definite price of each productive service at each moment in time. The goal of his theory was thus to determine how these quantities and prices are arrived at. His answer to this was to solve a system of linear equations where the three "natural effects" appear as unknowns (Jaffe, 1973). Under these assumptions Walras then proceeded to outline the two key concepts of his work: the theory of exchange and the theory of marginal utility. Both theories were brought together at the market place. Exchange will continue to occur until both sides can no longer gain any more utility from further trade, at which point the equilibrium is reached. In more general terms, the value of commodities is predetermined by their scarcity. The scarcer a good, the greater is its marginal utility and hence the higher its price. Equil ibrium prices are then equal to the ratios of "raretes" (scarcities), where raretes are defined as the "intensities of the last wants satisfied for the holders of the commodities" (Dobb, 1973, p.204). A s mentioned above, Walras is considered by many authors to be the most influential of the three founders of neoclassicism. In fact, contemporary neoclassicism is often referred to as "post-Walrasian" (Walsh & G r a m , 1980). The reason for this is that Walras was far more 6 successful than Jevons or Menger in establishing a clearly defined analytical link between marginal utility and the market price of a commodity (Jaffe, 1973). 2.2 M A J O R CHARACTERISTICS OF NEOCLASSICAL ECONOMIC T H E O R Y Perhaps the central feature of neoclassicism, is its reliance on utility. Utility's importance follows from its role within the neoclassical theory of value. F o r utility is the psychological entity that provides a good with value. It represents the fundamental psychological satisfaction that is the motive behind any purchase. Two features stem from the nation of utility: first, following Jevons et al. , there is a systematic relationship between price (exchange value) and utility (use value). Second, because utility is what everyone wants, it follows that producers and consumers employ their resources in the most efficient way, that is, to realize as much utility as possible given the available means. This is termed "resource allocation". Given these preliminary remarks, the intention of this section is first to summarize the general characteristics of neoclassicism, and second, to discuss the concepts of utility and resource allocation in detail. Not only do the latter concepts play a pivotal role in neoclassical economics, but, by extension, in neoclassical economic geography as well. There are six major characteristics of neoclassical economic theory, all of which have been applied in economic geography and regional science (King, 1979). First , the principal focus of neoclassical economics is on equilibrium analysis, both general and partial. Within this context, the main goal is to determine the conditions which must exist if equilibrium is to be stable. Thus the tasks of neoclassical economics are twofold: first, to show that equilibrium exists, and second, to prove that if at any point in time the system is out of equilibrium, "forces exist that drive the system back to equilibrium" (King, 1979, p.37). Second, it is assumed that only two groups of individuals exist: producers and consumers. E a c h of these groups act in a unified manner, in that both producers and consumers display the same behaviour and make identical decisions under the same circumstances. The aim of both consumers and producers is to seek a state where they are as 7 well off as possible. This leads to the third major characteristic: consumers seek to maximize utility, while producers seek to maximize profit . The fourth characteristic identifies the optimal state of an economy once it has reached equilibrium. Such a state is defined by the so-called "Pareto-optimality" criterion (King, 1979). A n equilibrium is considered to be Pareto optimal if and only if there is no other point where at least one consumer would be better off and no other consumers worse off (Myrdal , 1953). Thus in Pareto-optimality consumers' utilities are maximized thereby allowing the "greatest happiness for the greatest number". Fifth, it is generally assumed that there are constant returns to scale in production. Production functions are consequently all linear-homogeneous. Because there are constant returns to scale, increases or decreases in the demand for a product will not lead to any increases or decreases in its price. Prices are therefore supply inelastic. Final ly , it is assumed that both the producers and the consumers have perfect knowledge of all prices. Le t us now first turn to the concept of "utilitj'". F r o m the previous discussion we have seen that utility is a key to the theories of all three founders of neoclassicism. For Jevons it was the amount of pleasure a good provided; for Menger it represented the amount of displeasure an individual experienced if he or she was deprived of that good (loss principle); and for Walras the utility of a good was linked to its scarcity: the scarcer a good, the more satisfaction it provided to an individual. In general, however, all three were all referring to the same thing: the use-value of a good. The value of a good is directly derived from the utility it provides. Just as value can be expressed in terms of prices (a physical quantity), the basis of a good's value, utility, can then be expressed in satisfaction (a psychological quantity). A l l individuals act to maximize their utility, i.e., they strive to obtain the max imum satisfaction. The neoclassical argument goes that in pursuing their own utility maximization (see third characteristic above), individuals are also collectively maximizing society's utility. When all individuals' utilities have been 8 maximized the point of equilibrium is reached, i.e. it is not possible to change without decreasing the collective utility. Thus the greatest happiness for the greatest number (Pareto-optimality) is achieved. Therefore, neoclassicists argue, because all individuals have the same goal, namely to maximize their utilities, there is a natural tendency towards a state of equi-librium. Consequently, it is in most cases not necessary to interfere with the economic process in order to maximize human welfare. Therefore the argument is that political economists should adopt a "laissez faire" type of attitude. Connected to the notion of utility is the concept of "marginal utility". The term was not coined until the early 1900's, when it was introduced as an alternative term for Jevons' "final utility" (Howey, 1973). The marginal utility of a commodity is the change in utility an individual receives from consuming an additional unit of a service or a good. Thus the value of a good is directly derived from its marginal utility. Under the assumption of utility maximization neoclassicists explain human behaviour. If it is possible for a person to gain additional utility by obtaining a good or service, then, provided he or she has the funds to do so, it will be purchased. O n the other hand, if an action will lead to disutility, i.e., a reduction in utility, a person will refrain from undertaking it. Let us now turn to the issue of resource allocation. Resource allocation refers to the distribution of two or more given quantities of goods or services between two or more uses. Because consumers are utility maximizers and, by extension, producers are profit maximizers, it is argued that these goods or services will be efficiently distributed such that Pareto optimality is achieved. Let us illustrate this principle by using a typical "post-Walrasian" example (see Figure 2.2.1), where we have two input factors, land (T) and labour (L), and two commodities, wheat (W) and rice (R). The two input factors are constrained in that only a fixed amount is available. The amount of wheat and rice produced is shown on the y and z axes, respectively. A t point L y ^ all of the labour supply is used up for the production of wheat, so that the maximum amount of wheat possible is produced given the fixed supply of labour. Similarly, 9 1 Figure 2.2.1: Feasible Resource Allocation as an Intersection of Resource Constraints. I Wheat-Rice Source: Walsh & G r a m (1980) 10 point L j ^ shows the max imum amount of rice that can be produced if all of the labour supply is used for rice production. Thus any point on the line connecting L w and L - ^ (referred to as the "production possibility frontier") represents a feasible allocation of labour between wheat and rice production, considering the amount of land available. In the same manner the line between Ty^ and T j ^ represents the production possibility frontier of land. Superimposed on the diagram are a number of points (YQ to Yg) . A l l points, with the exception of Y g , Y y and Y g , are feasible locations. Point Y g exceeds both the land and the labour constraint, whereas at Y g the land constraint is exceeded and at Y y there is not enough labour available. The ideal location would be at Y g . Here both input factors are fully utilized and efficiency is at a maximum, i.e., all available resources are put to use. A l l other points are at levels of production where at least one of the resources is underutilized. It should be noted that although a simple case, the above example can be extended to any number of commodities and input factors, or resources. The difference between Walras ian and post-Walrasian neoclassical economics is that for the former point Y g would be considered the only feasible combination of resource allocation. This is because it is assumed that all resources must be fully utilized. Therefore any other point would constitute a "waste" of available resources. In post-Walrasian economics all feasible locations would be considered. For example, consumers in the model m a y all be wheat eaters, in which case point Y-^ would be the ideal allocation. Walras's assumption of fully utilized resources was the subject of considerable debate during the 1930's (Walsh & G r a m , 1980). The most frequent argument was that it is not always possible to fully employ all resources, since, in a two factor - two commodity case, for example, the two production possibility frontiers m a y not intersect at a point where both outputs are non-negative. Both sides, however, agree that at point Y Q production would never occur. A t least one of the resources must always be fully utilized. This happens because individuals wish to maximize their utility, i.e., get the most use (satisfaction) out of what is available in light of their wants and desires. 11 The marginal revolution had significant implications for orthodox economic theory. It provided theorists with a convenient set of analytical tools that could easily and effectively be applied to a wide range of uses. In particular, marginal analysis provided a technique to identify the most efficient allocation of a given set of competing resources. The optimum allocation was at that point where marginal values were equalized. This allocation principle could also be applied to a number of different problems, be it the allocation of a fixed amount of income among a range of consumer goods or a set of production factors. In summary, with techniques enabling the theorists to identify optimal allocations of resources neoclassicists were able to derive a seemingly logically sound explanation of how commodity and factor prices are determined in a market system, while at the same time maximizing consumers' satisfactions. To economists at that time this new method represented substantial analytical power. Thus it is no surprise that many students of economics were attracted by neoclassicism, and that neoclassical economics soon became the leading school of thought. 2.3 CRITICISMS O F NE O C L A S S I C A L EC O N O M I C T H E O R Y . There are numerous criticisms of the above characteristics that have been voiced by economists, regional scientists and geographers alike (King, 1979; Hollis & Nel l , 1975; Robinson, 1962; M y r d a l , 1953; Barnes, 1984, 1985; Blaug, 1973; Holland, 1976; Clark, 1982, 1983; Clark & Gertler, 1983). In particular, critics have tended to focus on utility (and utility maximizing "economic man"), resource allocation (and the inevitable pareto optimal equilibrium) and problems of dynamics in the formulation and testing of neoclassical hypotheses. This section will engage in a preliminary discussion of some of these criticisms. Further criticisms, where appropriate, will be put forward in later chapters. The concept of utility is one of the most heavily criticized aspects of neoclassical theory (Myrdal , 1953; Robinson, 1962, 1973; Dobb, 1973). One of the strongest arguments against utility has been that it is a circular concept (Robinson, 1962, p.48): 12 "Utility is a metaphysical concept of impregnable circularity; utility is the quality in commodities that makes individuals want to buy them, and the fact that individuals want to buy commodities shows that they have utility." Util ity and the maximization assumption have been expressed in a concept that has come to dominate both the disciplines of economics and economic geography: "economic man". Economic man, sometimes also referred to as "rational economic man", is a person who essentially possesses no flaws. In summary, economic man has the following characteristics (Hollis & Nell , 1975): (1) he has perfect knowledge of all factors involved in whatever situation; (2) the principle applies to all organizations, including producers, consumers and landlords; (3) economic actors are maximizers, which implies that producers maximize profits, consumers maximize their utility, and landlords maximize the amount of rent they can obtain under given circumstances; (4) rational economic man will always make a "rational" decision, i.e., he never makes a mistake. There are a number of problems with economic man. First , in using him it is assumed that everybody behaves in the same way. Thus any model based on economic man ignore diversity or cultural differences. Second, it has been argued that economic man is too simplistic an assumption (Pred, 1967, 1969; Hollis & Nell , 1975). For example, Al lan Pred (1967), in outlining his behavioural matrix argues that economic m a n in possessing the ideal combination of perfect information and perfect computational skills is a special case that rarely, if ever, occurs. Final ly , it has been argued that economic man can never be tested (Kaldor, 1972). The argument here is that rational economic man is insulated against empirical falsification by means of a ceteris paribus clause. The problem is that if in testing economic man we find that he is disconfirmed, one never knows whether such disconfirmation is a result of h im actually being falsified or whether things are not in fact equal. A final criticism of neoclassical theory is that its nature largely prohibits the use of dynamic models. The economic system is always considered to be at or near a state of equilibrium. If at any point in time the system falls out of equilibrium, then the perfectly 13 competitive system will ensure that market forces will bring the system back to equilibrium. This occurs because disequilibrium implies that when some people experience disutility, utility-maximizing individuals will strive to make themselves better off. In so doing they eventually bring the system back in to equilibrium. However, all other factors are assumed to be exogenous to the system and stable over time. Thus the dynamics of exogenous factors are ignored. The emphasis on equilibrium analysis and the consequent neglect of dynamics, so the critics contend, is one of the major setbacks of neoclassicism (King, 1979). 2 . 4 G E O G R A P H I C A L A P P L I C A T I O N S O F N E O C L A S S I C A L E C O N O M I C S . Neoclassical economics have been applied in numerous areas of economic geography, notably in theories of location analysis and the analysis of interregional flows, such as trade and migration, for example. In the process both partial and general equilibrium approaches to resource allocation have been adopted. The following discussion will outline some of the major theories within the partial and general equilibrium approaches. Let us begin with partial equilibrium applications and some examples from the oldest branch of regional science and economic geography: location analysis (as we shall see later, however, location analysis is not restricted to the partial equilibrium approach). A very well known location model using neoclassical resource allocation techniques is that developed by Alonso (1964), which analyses the location of a firm within a city using the concept of a bid-rent function. The purpose of the bid-rent function is to show how a firm's willingness to pay rent varies with increasing distance from the C B D while keeping profits constant regardless of location. Since it is assumed that the firm is a profit maximizer, the objective is to find the location with the highest possible profit level and the lowest possible bid-rent. This is the case where the bid-rent function (the slope of which is given by the size of the site, revenue and operating costs) is tangential to the rent gradient (Alonso, 1970). Figure 2.4.1 illustrates this, where R is the actual rent charged with increasing distance from the city (assuming that the slope of the curve is negative), and X-^ to X g are bid-rent functions 14 Figure 2.4.1 Bid-Rent Functions and Locational Equilibrium. K d i s t a n c e Source: Richardson (1978) Figure 2.4.2 Concentric Zone Model of Land Use. distance Source: Richardson (1978) 15 representing the amount of rent the firm is willing to pay with increasing distance from the C B D given a particular level of profits. The firm will then locate at the point of locational equilibrium, L , where the curve is tangential to the rent gradient. The amount of rent will be P at K distance from the C B D . This example can be extended to any number and type of firms (Alonso, 1964), resulting in a land use pattern centered around the C B D . Different firms producing different products (i.e. competing land uses) will have varying bid-rent functions. Thus , if competing land uses are ranked according to the steepness of their bid-rent functions, each will monopolize a concentric area around the C B D . In other words, each type of economic activity is "allocated" a particular zone of land. A n example is shown in figure 2.4.2, where four competing land uses (a,b,c,d) are allocated to four zones of activity, resulting in an "exclusive zoning" pattern (Richardson, 1978). The above model has been extended to a large variety of land uses, notably residential and commercial. However, there are two major weaknesses associated with its assumptions: first, only one rent is determined exogenously, so that the resulting land use patterns do not necessarily reflect reality. The second problem lies with the explicit assumption that the slope of the rent gradient is negative. The advantage of this is, of course, that it enables a neat solution to be derived. However, empirical studies have shown that the bid-rent functions for some industries have a different shape (Richardson, 1978). For example, some manufacturing industries, although initially attracted to the suburbs because of lower land rents, may actually be willing to bid higher rents in peripheral locations than at locations closer to the C B D due to other advantages, such as agglomeration economies and terminal locations. This could in turn imply a positively sloped bid-rent function. Another partial equilibrium approach is that of Weber's (1909) industrial location theory. The assumptions in his model are typical of neoclassical economic theory: producers are profit maximizers; they have perfect knowledge of all input factors and transportation costs (they are all rational); and, in the simple transport oriented model, all variables other 16 than distance are held constant. The argument is that under these assumptions the optimal location is uniquely determined at the site where total transport costs are minimized. In the case of m input factors (raw materials) and c markets, the optimum location will be determined by minimizing the sum of all transport costs (Richardson, 1978): m 1 c • m„ 1 t . r . q . + T t . r . q . ( 2 .4 .1 ) mm TC = l l l . . i i i i = l ]=1 where: t is the transport rate per unit distance; r^  is the distance between the source of the raw material and the production site; and qj is the amount of material moved from the raw material source to the production site. The equation also implies that if there is more than one market, then the location of the optimal site will be sensitive to variations in demand among the different markets because it influences economies of scale (Richardson, 1978). The pull will be towards that market which shows the highest demand for a product because the per unit cost declines with increasing demand. For many years the major problem to be solved in location analysis was to find the minimum transport cost location under the assumption of profit maximization. Subsequent modifications to Weber's theory by Loesch shifted the attention to variations in demand under the assumption of uniform costs. However, this still implies both revenue and profit maximization (Richardson, 1978). Alternative studies have focused on sales maximizing, as opposed to profit maximizing, which implies a different choice of an ideal location if costs change over space. As previously mentioned, most location models adopt the partial equilibrium approach, defined as an economic system is closed off to the rest of the world with all exogenous variables held constant. A general equilibrium theory of location analysis, in contrast, simultaneously considers all variables. Such general equilibrium models are extremely difficult 17 to develop, and although there have been numerous attempts at doing so (Henderson, 1958; Lefeber, 1958; Kuenne, 1963), all have thus far been unsuccessful (Richardson, 1978). One of the major obstacles is that even if an initial equilibrium exists, it is almost impossible to determine the implications of a disturbance of that equilibrium as each disturbance would bring about a chain reaction of locational adjustments throughout the space economy. O n the other hand, for a variety of reasons, such as immobilities and spatial frictions, it is necessary to distinguish between those disturbances that create locational adjustments and those that do not. Furthermore, the prediction of the impact of a disturbance is complicated by the fact that it takes time for certain factors, such as information, for example, to be transmitted over space. One of the earliest attempts at establishing a general equilibrium theory of location was that of Loesch (1954). The basic assumptions of the model are that producers maximize profits and consumers maximize their utility by buying the cheapest products. The competitive struggle between producers results in the elimination of excess profits and thus also in locational equilibrium. Specifically, the conditions for the existence of a general locational equilibrium are (Isard, 1956, 1960): (1) all space is taken up by market areas served by producers; (2) because prices are equal to average costs, all producers earn normal profits; (3) market areas are of a minimum size; (4) all producers are located at the optimum site; and, (5) the market area boundaries are stable because all consumers are indifferent as to their choice between equidistant suppliers. The result of these conditions and assumptions is that regardless of their location all firms producing a specific good have identical costs, market areas, freight rates and f.o.b. market prices. A p a r t from the problems associated with its neoclassical assumptions, the Loeschian model is unrealistic in the sense that it neglects locational interdependencies and intraindustry agglomeration economies. A country could thus never have an industrial heartland in which, for example, most of the nation's automobile production takes place, as it is the case in Canada . Although the model can deal with interindustry agglomerations, the 18 resulting concentrations of population would negate the conditions for a locational equilibrium (Richardson, 1978). Other attempts at formulating a general equilibrium theory of location have been made by Greenhut (1956), Henderson (1958), Lefeber (1958) and Isard & Ostroff (1960, to name a few examples. Although many of these models have been more successful in that they eliminate some of the problems associated with Loesch's model, none of them have been able to account for all aspects of the space economy. Thus a working general equilibrium theory of location has yet to be established. Although the above discussion has not covered all aspects of neoclassicism and its impact on economic geography, it has nevertheless become clear that contemporary economic geography is indeed strongly rooted in neoclassical economics in that it very much reflects neoclassical thought. Just as economists were attracted by the relative parsimony and analytical power of neoclassicism, so were geographers drawn towards the approach due to the apparent ease and efficiency with which it could be applied to a spatial context. The resulting theories and models proved to be extremely popular, illustrated by the fact that, for the most part, they are still in use today, albeit in modified form. Consequently one must ask the question whether neoclassical economics are really applicable to the "real" world, or whether they are just a vision of an idealized "hypothetical" world. 19 3. N E O C L A S S I C A L E C O N O M I C S A N D L A B O U R M I G R A T I O N . In the previous chapter the central features of neoclassical economics were outlined and criticized. It was demonstrated how traditional economic geography is rooted in neoclassicism. In the first part of this chapter it will be shown how neoclassical migration theory is derived from the broader neoclassical assumptions. This is followed by a discussion of several hypotheses about the causes and effects of labour migration that stem from the neoclassical framework. The second part, of the chapter will show how these hypotheses have dominated the theoretical and modeling literature of labour migration. In particular the two dominant schools of thought within the neoclassical approach will be discussed: the "micro-adjustment approach" and the "macro-adjustment" approach. It will be argued that although the original models of these two schools of thought (Sjaastad (1962) developed the micro-adjustment model and Lowrey (1966) the macro-adjustment one), have been modified since their initial presentation, both the current literature on labour migration, and perhaps more significantly public policy, has, and continues to be, dominated by a neoclassical vision of the economy. 3.1 T H E FOUNDATIONS OF NEOCLASSICAL L A B O U R MIGRATION T H E O R Y . Not surprisingly, the key elements of neoclassical labour migration theory are the same as those discussed in the previous chapter: utility, resource allocation determined by the laws of demand and supply and equilibrium. In addition, like all neoclassical theory, it is also highly abstract making a few simple assumptions about human behaviour and the nature of the economy. Perhaps the key assumption in the neoclassical model which is transferred to labour migration is that of utility maximization. F r o m the previous chapter we know that the neoclassical school of thought postulates that all individuals wish to maximize their utility (degree of satisfaction). It was also suggested that utility cannot be measured directly because it consists of a variety of factors in terms of which an individuals determine their level of satisfaction. Consequently some surrogate for utility must be found in order to model labour 20 migration. Wage and unemployment rates offer themselves intuitively because they are the only variables that vary with changes in the supply of labour. Clearly, however, there are other variables that individuals might want to maximize, for example, climate or the quality of life. But climate is unresponsive to changes in economic variables, and therefore difficult to measure, and in the case of quality of life it is indirectly affected by changes in wage and unemployment rates and thereby already partly accounted for. Thus although wage and unemployment rates are not the only explanatory variables considered in neoclassical models, they nevertheless represent the key determinants of labour migration. This is so because they are an indicator of changes in utility at a given place. Perhaps the best way to exemplify how the broader principles of neoclassical economics are employed in neoclassical labour migration theory is to use the example of a hypothetical nation composed of two regions. In both regions we have perfectly competitive markets: there is no government intervention; there are no barriers to movement or entry to any market, thereby allowing each individual at any point in time to move freely from one region to the other; and, finally, information is homogeneous, universal and complete, thus ensuring that all individuals possess perfect knowledge about all economic indicators (wages, prices, employment, etc.). Also, assume that initially the two regions are in perfectly competitive equilibrium, that is, there are no regional differences in any of the economic variables. Now let us assume that for some reason, say, a shortage in the supply of labour, wage rates in one of the regions, call it A , increase. A s a result the system is no longer in equilibrium. Because information is assumed to flow freely across space, workers in the other region, call it B , will learn of this wage increase and, because of their desire to maximize their utility, respond by moving to region A . In turn, the migration from region B to region A will cause an increase in the supply of labour in region A , while at there same time decreasing it in region B . The increase in the supply of labour will, ceteris paribus, through the law of demand and supply, reduce wage rates in region A while the reduction in the pool of workers in B will 21 cause wage rates to rise in that region. Workers will migrate from region B to region A , and this process will continue until the system is eventually brought back in to equilibrium. The same argument can be made for unemployment rates: if unemployment is lower in region A than in region B , the unemployed will move to A because of the greater probability of finding a job. In broader terms, labour migration can be viewed as an "adjustment process" where labour is reallocated according to regional variations in the demand and supply of labour: workers respond to regional differences in labour market conditions by migrating between regions until the system is back in equilibrium. Thus, just as resources in the wider economy are efficiently allocated between uses through the price system, so too, ceteris paribus, is labour efficiently "allocated" among regions through the push and pull of market forces, namely changes in wage and unemployment rates. In general then, two hypotheses can be derived from this simple model. The first concerns the causes of labour migration: all other things being equal, labour flow from regions of low wage and high unemployment rates to regions of high wage and low unemployment rates. The second hypotheses is that the effect of labour migration in regions experiencing net outmigration will be for wage rates to rise and unemployment rates to fall, while in those areas experiencing net inmigration wage rates will fall and unemployment will rise. The net result is that the spatial economic system will be driven back towards equilibrium, i.e., wage and unemployment rates will be equal across the country. 3.2 Empirical Applications of Neoclassical Labour Migration Theory The neoclassical labour migration theory presented in the previous section is only a simple version; it derives from Hicks' (1932) early work on wages and labour allocation. Since Hicks' pioneering study a number of modifications to the theory have been made. In part, such changes are a result of the realization that the assumptions of neoclassicism rarely, if ever, apply. Nevertheless, despite those modifications labour migration theory remains 22 fundamentally neoclassical. This part of the chapter will first present the main assumptions of current labour migration literature, and then show some examples of geographical models employing neoclassical labour migration theory. 3.2.1 Assumptions of Labour Migration. M c K a y & Whitelaw's (1977) work identifies the six most common assumptions in current literature on labour migration. The first assumption is that the decision of when and where to migrate is made by individuals who are free to do as they please. Such decisions, however, are influenced by prevailing economic conditions, spatially biased information flows and existing social ties. This first assumption is essentially a modification of the pure neoclassical model and reflects the realization that the real world is more complicated than the simplifying assumptions of neoclassicism. The second assumption is that individual decision making takes place within a "free market" framework in which there is a wide range of alternatives. A person can choose between as many destinations as there are regions within the spatial system. Also, once a person has reached a destination, then he or she is free to compete with the local labour force for any job openings. Third , an individual always has the alternative to remain in the same location. No pressures exist that could force a person to move. The dislocation of workers due to plant closures, for example, could by assumption never happen. Fourth, although it is an accepted fact that mobility rates differ by socioeconomic groups, it is assumed that all individuals are subject to a single migration process, albeit with different intensities. This implies that each person responds in the same way to those variables influencing labour migration. It is recognized, however, that the intensity (probability) with which people respond does vary. This leads us directly to the fifth most common assumption, which is that since a single process is assumed, highly aggregated data is sufficient to test a migration model 23 because everybody behaves (reacts) to the economic environment in the same way. If everybody did not act in the same way the hypothesis on the causes of labour migration would not hold. Finally, and perhaps most importantly, migration is assumed to operate as an efficient equilibrating mechanism, reducing differences between regions: wage and unemployment rates will, through changes in the supply of labour resulting from migration, be equalized among regions. 3.2.2 Geographical Models of Labour Migration. Within the realm of neoclassical labour migration theory two distinct approaches have been adopted: the "micro-adjustment" approach (commonly also referred to as the "human investment" approach) and the more widely used "macro-adjustment" approach. Whereas the former approach reflects the highly individualistic nature of neoclassical theory and is theoretically more "pure" than the other, the latter is more practical in that it can actually be applied empirically. In our discussion of the two approaches let us begin with the micro-adjustment approach and its ties to human capital or human investment theory. The origins of human capital theory go back to A d a m Smith (Carline et al, 1985). The theory is closely linked to the concepts of marginal utility and marginal product. Not only has it been used in migration analysis, but also in a wide range of other areas. H u m a n capital theory is used to determine whether an individual will make a decision to "invest" in an undertaking or not. For example, assuming a wealth maximizing individual, if the marginal increase in future wages (discounted in terms of the remaining lifetime working horizon) is greater than the cost of such training that enables a person to earn those higher wages, then that person will decide to go ahead with the training. Generally speaking, if the marginal increase in utility or the marginal product is greater than the cost of an undertaking, then, ceteris paribus, an individual will always choose such an undertaking. 24 This highly individualistic approach to decision making was first applied in migration analysis by Sjaastad (1962). His argument is that when an individual is considering moving from region A to region B, he or she computes the present value (PVt) of the move with respect to the following variables (Clark & Ballard, 1979): Potential earnings (Yg) in region B at time t; current earnings (Y^) in region A at time t; the individual's lifetime working horizon (T); the cost of moving from A to B (C); and a discount rate (r) which individuals apply to their earnings stream. Migration will only occur if the anticipated returns are greater than the cost of moving. In formal terms, define the present value of migrating as: T YB - YA P V ( t ) a = I Z- - C (3.2.1) t=0 (1 + r ) T 1 where = t+1. A present value can also be calculated for the decision not to migrate, call it PV(t) . Therefore, migration will occur if and only if PV(t) is greater than PV(t) . In the case of multiple regions the individual will migrate to that region which promises the highest anticipated returns. The micro-adjustment model has a number of distinct advantages (Schwartz, 1976) which are linked to the significance it attaches to the attributes of individuals. Demographic characteristics such as age, occupation, and education are easily incorporated. Furthermore, contrary to the neoclassical assumption that migrants flow from low wage to high wage regions, it can account for the fact that there are also reverse flows: some people might be able to maximize their future earnings by migrating to a depressed region. However, there are two major problems with this model (Clark & Ballard, 1979). The most important one is that it is essentially untestable. Individual case histories would be necessary to test the propositions. Even if such data did exist, one would need a sufficiently large number of regionally stratified observations in order to make any general conclusions as to its validity. The second problem is the assumption that the individual will always go to that 25 region which promises the highest future gain. In real life the decision of when and where to migrate might well not be decided by an individual. A person might be forced to move (because of a plant closure, for example; for a discussion of this see Gordus et al . , 1981) even though the move will result in a net loss. O n the other hand, a move could be the result of a company transfer, in which case the decision is made by a higher authority, not the individual (see M c K a y & Whitelaw, 1977). Let us now turn to the macro-adjustment approach. The assumptions it makes are typical of neoclassical theory: workers have perfect knowledge of all factors involved; information is homogeneous, universal and complete; there are no differences between regions other than in terms of wage and unemployment rates; and there is no significant social or economic cost associated with the migration process. Work in this area was pioneered by Lowrey (1966). In his model wage and unemployment characteristics at the origin and destination points are assumed to be the determinants of migration. In its basic version, the model can be specified as: u. w. L. L . , 1 * 1 * i : u. w. D. . where: Mjj is the gross migration from i to j ; U j and Uj are the unemployment rates in i and j ; Wj and Wj are the wage rates in i and j ; and L j , L j and Djj are the population and distance components of the gravity model. The major criticism of this model is that its assumptions conflict with reality: perfect information does not exist and there are numerous social and economic costs associated with the migration process. Furthermore, there has been very little empirical evidence to support the Lowrey model. Clark (1983) maintains that although it has been quite successful for prediction purposes, probably due to its gravity component, it is inadequate in explaining the labour migration process. 26 In sum, both the micro- and the macro-adjustment models are representative of the first neoclassical hypothesis, namely that labour will migrate in accordance with the spatial pattern of demand and supply as reflected in interregional differences in wage and unemployment rates. It is assumed that in the long run migration will tend to decline as the system approaches equilibrium. 3.2.3 Empirical Applications. There are numerous examples of applications of neoclassical theory in labour migration modelling (Lansing and Mueller, 1967; Renshaw, 1970; Fields, 1976, 1978; Milne, 1981; Simmons, 1982; Young, 1984). The first part of this section will discuss applications of the micro-adjustment model, while the second will deal with examples of applications of the macro-adjustment model. 3.2.3.1 Micro-Adjustment Models. Empirical applications of the micro-adjustment approach to migration analysis are few and far between. This is not surprising, considering the data requirements of the model. O f the few examples that do exist, most adopt the micro-adjustment (human investment) approach in their theoretical sections, but revert to an aggregated version when empirically testing the model. One example of this is Vanderkamp's (1968) study of interregional mobility in Canada. He hypothesizes that individuals are more likely to migrate from region A to region B depending on the net advantage they derive from migrating. A p a r t from regional income differentials, Vanderkamp uses unemployment and return migration as variables (all aggregate variables). H e found that the flow of migration is negatively related to unemployment (as is predicted by the neoclassical hypothesis) and that return migration is positively related to the unemployment variables. The positive relationship is explained by the push of the sending region's employment being stronger than the pull of the region to which they return. 27 Other examples are those of Brennan (1965), Diehl (1966), Bodenhofer (1967), Rabianski (1970) and Kottis (1972). Typically, these studies relate the benefits of migration to factors such as variations in basic wage rates, opportunities for job training and education, and economic rewards in alternative regions for attained skills (Shaw, 1975). 3.2.3.2 Macro-Adjustment Models. F a r more popular than the micro-adjustment approach is the macro-adjustment approach. The majority of past and current literature on labour migration has adopted it in one form or another. A s mentioned earlier it is essentially an aggregated version of the micro-adjustment approach and reflects the neoclassical notion that labour flows from areas of low wages and high unemployment to high wage and low unemployment regions. Gra nt & Vanderkamp (1976) examine the causes and effects of labour migration in Canada during different time periods. In a model almost identical to that of Lowrey's , the significance of wage rates, unemployment rates and job information were examined in determining migration flows. The significance of all the variables was tested. To allow for the fact that information about other provinces declines with increasing distance, a gravity type variables was also included. Although it was found that the standard linear equation did not satisfactorily explain the variation in mobility rates (both income and unemployment were insignificant as explanatory variables), a general observation was that income was a strong pull factor, but only a weak push factor. A s expected, distance played a major inhibiting role. Unemployment had a strong push and a negative pull effect. There was, however, considerable variation in the significance of variables between time periods, suggesting that the migration pattern m a y be subject to cyclical fluctuations. Another neoclassical study of labour migration is that by Roseman (1977) of changing migration patterns in the United States. Aga in the approach was identical to that of Lowrey, with one major modification: the decision to move and the decision where to move were considered as two separate models (equations). The results were essentially commensurate to 28 those of Grant & Vanderkamp's (1976), with the additional observation that there was also a considerable variation in mobility rates between age groups. Greenwood (1981) proposes a typical neoclassical model of metropolitan growth and migration. Arguing along traditional lines, Greenwood maintains that outmigration of labour will put upward pressure on wage levels whereas inmigration will apply downward pressure. The magnitude of change in wages, however, depends on the relative magnitude of the change in the demand and suppty of labour. The same argument holds for unemployment rates. Also, the impact of migration on both variables is contingent upon a variety of other factors, such as the occupational composition of the inmigrating labour force visa-vis the occupational structure of the existing labour force. This represents a departure from the normal assumption where the labour force is taken to be homogeneous, that is any labour can do any job. Young (1985) analysed changes in the significance of selected origin and destination characteristics on migration flows to Western Canada and on return migration from Western Canada between 1971-1976 and 1976-1981 using a model along the lines of Grant & Vanderkamp (1976) and Courchene (1970). These characteristics included a gravity term, annual average employment in 1971 dollars (income), percentage growth in employment over the five year census period, and a cultural variable (percentage of the population of French mother tongue). It was found that the significance of origin and destination variables declined significantly between the two time periods under analysis. While for 1971-1976 almost all of the variables were statistically significant at the 1% level, for 1976-1981 only two of the variables were significant at 5%. Although the results of the analysis have to be interpreted with caution due to the five year census interval, they do suggest that there could be cyclical fluctuations in the sensitivity of migrants to interregional differences in economic variables. One of the most extensive reviews of migration literature is that conducted by Shaw (1975). His work marks the beginning of a growing criticism of neoclassical migration analysis. His criticism, however, is limited to methodological rather than theoretical problems. Thus he argues that aggregate migration measures do not permit the identification of non-29 economic motives for migrants. Furthermore, return migration and multiple moves are not accounted for in the neoclassical approach. In addition, even if crude economic measures (such as wages and unemployment) do seem significant in explaining migration patterns, the results are only rough indicators of the real forces at work in migration. A p a r t from the methodological criticisms of Shaw, theoretical failings of the neoclassical approach have recently been identified. Most critiques have either focused on the assumptions of human capital theory (Greenwood, 1981; Peek & Standing, 1982) or on the effects of migration (Clark & Ballard, 1979, 1980; Clark, 1982, 1983; Clark & Gertler, 1983). There is, however, a general consensus that migrants are not as sensitive to economic factors as neoclassical theory suggests. In addition, empirical evidence has shown that the broader framework that accounts for migration is one of cumulative causation (Curry, 1985) rather than a Pareto-optimal state of equilibrium (for a more detailed discussion see chapter 6). Despite such criticisms, a complete alternative labour migration theory has yet to be proposed. These methodological and theoretical criticisms of neoclassical labour migration modelling (for further examples see Kriesberg & Vining, 1978; Stone, 1978, 1979; Plane, 1981; Stock, 1981; Foot & Milne, 1982; Plane et al . , 1984), are being increasingly recognized by neoclassicists themselves. The result is that a growing number of studies have adopted increasingly sophisticated methods, thereby recognizing, for example, that information is not homogeneous and universally available, people are not always free to move or to stay, and that the labour market is not perfectly competitive. However, despite evidence to the contrary, most contemporary migration analysts hold on to the belief that interregional migration will eventually lead to a Pareto-optimal equilibrium. One of the recent examples of this persistence is a study by Cebula (1979). H e proposes two models of labour migration, one along the lines of the micro-adjustment approach and the other an aggregate macro-adjustment model. Surprisingly, only very few modifications are made to the standard Lowrey and Sjaastad models. The point is, however, that assuming perfectly competitive markets (!) and maximizing behaviour he argues that while a perfect 30 spatial equilibrium may never be reached, this is only because the marginal interregional differences in economic benefits (marginal utility) do not outweigh the costs of moving. This imperfect spatial equilibrium is thus only a reflection of rational decision making by utility-maximizing economic man: people will not make themselves better off by moving because the costs are greater than the benefits. Therefore, while wage equalization will never occur, labour migration will nevertheless eventually result in a Pareto-optimal equilbirium (nobodjr can make him or herself better off by moving). In summary, neoclassical economic theory as embodied in the micro-adjustment and the macro-adjustment approach to modelling labour migration is, despite a growing body of literature critical of its assumptions and hypotheses, still very much alive. Although neoclassicists do recognize and allow for the inadequacies of the standard assumptions, they continue to argue that labour migration will lead to a spatial equilibrium. The next two chapters will examine the validity of the neoclassical h^potheses outlined in section 3.1 in a Canadian context. 31 4 . D A T A A N D M E T H O D O L O G Y . The objective of this chapter is to discuss the data and methodology used in testing the two neoclassical hypotheses presented in chapter 3. Specifically, these are that (1) differences in unemployment and wage rates are directly related to migration rates, and (2) that net migration figures influence changes in unemployment and wage rates within a region. This chapter is divided into two parts. The first discusses problems of data, while the second addresses methodological issues. 4 . 1 D A T A . D a t a availability and reliability have posed a problem to researchers in many areas of migration study. While the micro-adjustment model probably represents the purest transformation of neoclassical theory, it requires unobtainable data. A s a result researchers have turned to aggregate data models, in particular the macro-adjustment approach. Despite the data availability for aggregate modelling there are disadvantages to using gross flows. For example, return migration (individuals who return to their place of origin after having moved to another place in search of a job) are almost impossible to estimate. However, studies have shown that this return migration can account for up to 40% of total migration flows (Vanderkamp, 1968). In the United States a new source of migration data has recently become more available (Clark, 1983). It is the Continuuous Work History Sample (CWHS) compiled by the Social Security Administration and the Bureau of Economic Analysis . This data source provides micro-level migration estimates based on individual work histories (Clark & Gertler, 1983). The C W H S is recompiled each year, making it possible to conduct in-depth temporal analyses. Unfortunately the data source does not extend to Canada and thus the first part of this analysis is forced to use gross migration flows, and hence a macro-adjustment model. In particular, the analysis takes data from the 1981 Canadian Census (although estimates of net migration at the provincial level are available annually, complete figures are only available 32 every five years from the census). Data listed in the census volume dealing with population mobility are based on a 20% sample of private households. Gross migration flows are given for each province of origin and destination, broken down by occupation. A s migration is defined in terms of the difference in place of residence (province) between 1976 and 1981, multiple moves during the five year period are not accounted for. Multiple moves would also include return migration as well as "hypermobile" individuals who move more than once every five years (particularly important in the white collar occupational groups; see Stone, 1978). A p a r t from this problem with the census figures, another difficulty is finding wage and unemployment data on which to regress the gross migration figures. A s the migration data refers to a five year period, the decision to migrate could have been based on information obtained in any one of the five years, depending upon when that decision was made. .On the other hand, information about wages and unemployment takes time to diffuse over space (Clark, 1986). Furthermore, there will be a time lag between the decision to migrate and the actual migration. Ideally, then, one would use annual migration flows and regress these on wage and unemployment rates of the previous year (if one assumed that the time lag was one year). Since this was not possible, 1976 was selected as the base year. E v e n if (as it is assumed) the error associated with the use of 1976 data is constant, the results of the following analysis must nontheless be interpreted with caution. Wage and unemployment figures were compiled from original tapes of the 1976 and 1981 censi (the latter being used for the second part of the analysis). A s with the migration data, the tapes are based on a 20% sample. A problem encountered here was that the definitions of the occupational groups had changed between 1976 and 1981. In 1876 there were fewer groups than in 1981. Thus the 1981 data (including the migration figures) had to be amalgamated in order to facilitate a direct comparison between the two years (see table 4.1.1). A further problem is the availability of data. Whereas in 1981 data were available for all provinces and territories, the 1976 tapes did not include Prince Edward Island, the Yukon 33 Table 4.1.1 Amalgamations of Occupational Groups. Group # Description (Census Occupational Group #) 1 Managerial , Administrative and Related Occupations (11) 2 Natural Sciences, Engineering and Mathematics (21) 3 Social Sciences and Related Fields (23) 4 Teaching and Related Occupations (27) 5 Occupations in Medicine and Health (31) 6 Fine and Commercial Artists (33) 7 Clerical and Related Fields (41) 8 Sales Occupations (51) 9 Service Occupations (61) 10 Farming , Horticultural and A n i m a l Husbandry Occupations (71) 11 Pr imary Occupations: - Fishing, Hunt ing and Trapping (73) - Forestry and Logging (75) - Mining and Quarry ing (77) 12 Processing Occupations (81) 13 Manufacturing Occupations (85) 14 Construction Trades (87) 15 Transportation Equipment Operators (91) 16 Other Occupations: - Machining and Related Occupations (85) - Materials Handl ing Occupations (93) - Other Crafts and Equipment Operating Occupations (95) - Occupations Not Elsewhere Classified (99) 34 and the Northwest Territories. These regions were thus omitted from the analysis. In the case of the two territoties this omission is probably just as well, as the migration figures there are extremely high in some occupations. Inclusion of the Northwest Territories and the Yukon therefore would have distorted the results of the analysis. For the second part of the analysis (the test of the second neoclassical hypothesis) net mobility rates were used (calculated from the gross migration matrix). The 1976 and 1981 data were used to calculate changes in unemployment and wage rates. 4.2 Methodology. 4.2.1 Theoretical Considerations. In the previous discussion it was shown that most studies testing the first neoclassical hypothesis either use first, a gravity type model or, second, a model that directly relates migration rates (in the form of gross or net migration or mobility rates) to wages and unemployment. In particular, both models generally employ multiple regression techniques, where the independent variables include differences in unemployment and wage rates. In gravity type models population figures are often weighted by wage and/or unemployment o rates. The r and the partial regression coefficients are then deemed to be an indication of the explanatory power of the hypothesis. This approach has been quite popular not only in migration analysis, but also in other studies of interregional flows, for example, the analysis of trade patterns. For quite some time now it has been realized that there are a number of problems associated with this approach (Olsson, 1965, 1970; Lycan, 1969; Curry, 1972; Johnston, 1973). While some studies recognize these problems and accordingly adjust their methodology, a surprisingly large number of studies do not. Significantly, these include the testing of the neoclassical hypotheses. The first problem with the traditional approach concerns the use of unconstrained gravity models, especially those that incorporate wage and unemployment rates. Regressions 35 of gross migration rates with this type of model characteristically show extremely high correlations, usually greater than .8. While possessing a large degree of predictive power, gravity type models do not explain any of the variation. The resulting correlation coefficient is only a measure of the "goodness-of-fit" of the model. Equal ly high r squares could most probably be achieved by simply using population or labour force figures. Thus it is not possible to either reject or accept a hypothesis on the basis of a gravity model, because it is not known whether the high correlation is a result of wage and unemployment differentials or simply the friction of distance. The second problem is related to the regression of raw gross migration figures on differences in unemployment and wage rates and, in most cases, a distance variable. The obvious implication here is that the number of migrants depends on the size of the population (or the labour force if the concern is with labour migration). Thus, the greater the population size of the region of origin, the greater the number of migrants. In order to facilitate a comparison of migration rates between regions of varying sizes, gross migration rates are usually transformed to mobility rates, i.e., the number of migrants per thousand population. However, this eliminates just one part of the problem. Mobility rates to regions with large populations will still be greater than to those with smaller populations. One of the reasons for this is that the larger a region's population base, the larger the number of employment opportunities and, by extension, the greater its attraction. Nelson (1959) has shown that a large part of total gross migration is a result of a "random" allocation process. In other words, migrants are randomly allocated according to the size of the regions of origin and destination. F o r example, if region j has 10% of the total population of a country, then, ceteris paribus, the number of migrants moving from i to j should be one tenth of all outmigrants from i. Conversely, if region i has a 5% share of the country's population, then, all other things being equal, the number of migrants from i to j should equal 5% of all inmigrants to j . Wage rates, unemployment rates, or distance have nothing to do with this aspect of migration. In order to test the influence of these variables gross migration 36 rates thus have to be adjusted to allow for this factor by removing the random component of gross migration. Once the effects of the sizes of origins and destinations have been removed, the effects of spatial factors have to be considered and also removed before the influence of any socioeconomic variables can be tested. Distance and intervening opportunities have been judged to be the most influential inhibitors to migration (Stouffer, 1960, 1962; L y c a n , 1969). The influence of distance is in itself a compound of two influences (Johnston, 1973). First , the friction of distance varies by origin. In terms of migration this would mean that some regions experience outmigration to more distant destinations, whereas in others, migrants tend to go to a more restricted portion of the system, concentrating on their "nearest neighbour". The second influence is that of the map pattern. Simply because of the location of the region of origin in relation to the destination regions, the average distance migrants have to travel varies. For example, Ontarians, being located in the middle of Canada, do not have to travel as far on average as Brit ish Columbians to reach their desired destination. Therefore the influence of distance (degree of inhibition) will, all other things being equal, not be as high for migrants from Ontario as it will be for migrants from British Columbia. The implication is that aggregate estimates of the influence of distance (based on regressions that are not broken down by origin) will be inaccurate for any kind of detailed analysis. A systemwide regression would result in an overestimation of the influence of distance on migration from regions located in the centre of a country, whereas it would be underestimated for peripheral regions. Therefore migration rates adjusted on the basis of a singular regression (not broken down by origin) would include this error. The other major spatial factor influencing migration is intervening opportunity. Intervening opportunity, like distance, is in itself the result of a combination a several variables, such as employment opportunities, and social and cultural considerations. The assumption is that intervening opportunities will have an impeding effect on migration over space. While intervening opportunities are closely related to distance, they do tend to "de-37 emphasize spatial separation across sparsely populated areas and to stimulate the absorptive effect of densely populated intervening areas" (Lycan, 1969, p.242). In other words, if, for example, a potential migrant has a choice between two alternate destinations, then the higher probability of moving to the closest one could be negated by the number of employment opportunities lying in between (if that number is less for the further destination). A s with distance, the influence of intervening opportunities will vary by province of origin for much the same reasons, i.e., the range of intervening opportunities will differ. What this discussion suggests is that in order to test the neoclassical migration model we must first remove both the spatial and population size-related influences from observed migration patterns. If we did not, we could not conclude much from our analysis. For it would be not clear whether it was the spatial and population factors that were deterring migration, or the wage and unemployment rates that are the focus of the neoclassical model. So what this chapter will do is to first account for observed migration patterns only in terms of spatial and population size-related factors. Once this has been done, we can test the residuals for differences in the effect of wage and unemployment rates. 4.2.2 Methodology. The methodology used in the first part of the statistical analysis closely follows that of Nelson (1959) and L y c a n (1969), albeit with some modifications. Gross occupational migration rates will be adjusted for the size of the labour force in each occupation, distance and intervening opportunities. Interprovincial differences in wage and unemployment rates will then be tested for their significance in exaplining the residual migration. First , census estimates of gross interprovincial migration are adjusted according to the size of the labour force in each occupation in each province. The rationale for this change is that the number of migrants in occupation k moving between two provinces should be directly proportional to the product of the provincial shares of the labour force in occupation k. The resulting adjusted migration rates (which are essentially ratios of expected to actual migration) 38 are defined as equal to 1.00 if the share of migrants moving from a given province of origin to a given province of destination is exactly proportional to the product of each of the provinces' share of the Canadian labour force in occupation k, less than 1.00 if it is less, and greater than 1.00 if it is greater (Lycan, 1969). The formula used to arrive at the labour force adjusted migration rates is: „ k k k • = - ^ - r / £• * —I i , j = 1 9 (4 .2 .1 ) . . i l h l u. i i j J i j J where my is the labour force adjusted migration rates; M ^ is the gross number of migrants between i and j in occupation k; and and is the size of the labour force in provinces i and j in occupation k. Second, the influence of distance and intervening opportunities is estimated by means of multiple regression. Migration rates are then adjusted a second time to account for this influence. Distance is measured in the number of road kilometers separating the major population centres of the provinces. Intervening opportunity is defined as the percentage of the total Canadian labour force in a particular occupation lying between the two provinces. Although intervening opportunities are most likely related to wage and unemployment rates (the higher the difference in wage rates, for example, the greater the intervening opportunities), it is assumed that this relationship is equal across the country. Because Canadian migration is unique in the sense that almost all (with exception of the Marit ime provinces) of the migration occurs in an easterly or westerly direction (the Y u k o n and Northwest Territories are not included in the analysis), it is possible to simply add the percentage shares of the labour forces of the provinces lying between i and j . A multiple regression will be performed on each province of origin under the assumption that: l n ( m i j k ) = f(ln(D i j); I C ^ k ) (4.2.2) 39 where: k 10.. i o . k 1 D jv1 k n = i + l I L. 1 = 1 J A L  n = j + l v * j = l i f j g . t . i , or i f j 1. t . i ; i , j ,n = 1 , . . . , 9 ( 4 . 2 . 3 ) 10y is the percentage of the Canadian labour force in occupation k lying between i and j ; L j and Ljk is the size of the labour force in occupation k in i and j , respectively; and Djj is the distance between i and j . Labour force adjusted migration rates and distance were logged in order to normalize the distribution. B y conducting a separate regression for each province of origin, differential influences of distance and intervening opportunities due to map location can be accounted for. Furthermore, intervening opportunities can also be broken down by occupation, thus reducing the problem of aggregation. For both variables it is expected that the partial regression coefficients will be negative. In formal terms, to adjust migration rates in a manner such that distance and intervening opportunities are considered, one simply subtracts the migration due to distance and intervening opportunities from the adjusted total migration, to arrive at the residual migration (rm-). Thus (following the example of L y c a n (1973)): 40 r m ^ = lnCm^) - ((a1 + b1i(ln(Dij)) + b 2 iaOy k)) (4.2.4) where r m ^ is the residual migration; a* is the constant for province i; and b-^ and b2* are the regression parameters for province i. If the residual is less than zero, migration was less than expected based on labour force, distance and intervening opportunities. Conversely, if the residual is greater than zero, then migration was greater than expected. Spatial and labour force size related factors have now been removed as far as possible given the available data. However, some spatial biases still remain due to factors other than distance and intervening opportunities. For example, migration from the Maritime provinces will tend to be between Maritime provinces (in some occupations) due to the social welfare system in those provinces. Fishermen, for example, will tend to stay in the Maritimes, because it is easier for them to obtain unemployment benefits there than in other, non-Maritime provinces. This and other factors will be discussed in more detail in the next chapter. The final task (and a major objective of this part of the thesis) is to test the significance of differences in wage and unemployment rates in explaining the residual migration. This will again involve the use of multiple regression. The formula used is that: k u k - u k w k - w k r m i i = f { - J n r * 1 0 0 ; -J~ir * 1 0 0 > ( 4 . 2 . 5 ) 3 u K wK 1 1 where U j k and U j k are unemployment rates in provinces, j and i in occupation k, respectively; and Wjk and Wj k are wage rates in occupation k in provinces j and i, respectively. According to the neoclassical hypothesis stated in chapter 3, the expected sign of the partial regression coefficients would be positive for wage rates and negative for unemployment rates. This section of the analysis will consist of two parts. In the first a separate regresion will be run on each province of origin to test interregional differences, while in the second it will be broken down by occupational group. 41 The second neoclassical hypothesis concerning the effects of labour migration suggests that positive net migration will lower wage rates, while at the same time increasing unemployment rates. Thus: W j k = f ( N E T M O B j k ) (4.2.6) and U j k = f ( N E T M O B j k ) (4.2.7) where W j k and U j k are percent changes in wage and unemployment rates in province i between 1976 and 1981 in occupation k; and N E T M O B j k is the net loss (gain) per thousand of the 1976 labour force in occupation k in province i between 1976 and 1981. In keeping with the neoclassical hypothesis the expected sign of the correlation coefficient should be positive for unemployment rates and N E T M O B and negative for wage rates and N E T M O B . 42 5. DATA ANALYSIS. This chapter will discuss the results of the data analysis as outlined in the previous chapter. The first section will provide a description of overall labour migration patterns in Canada between 1976 and 1981, while the second section will concentrate on the labour force adjusted migration rates. The third section will examine the residual migration rates not accounted for by the size of the labour force, distance and intervening opportunities. In the fourth section the results of the test of the first hypothesis will be presented, while in the fifth those of the second hypothesis will be discussed. Final ly , this chapter will conclude with a summary of the findings. Where appropriate, comparisons will be made to a similar study conducted by L y c a n (1969) on Canadian migration patterns between 1955 and 1961. 5.1 Gross Migration Flows. Table 5.1.1 shows gross migration figures within Canada for all occupations between 1976 and 1981. Prince Edward Island, the Yukon and Northwest Territories are included, as sufficient data were available for this stage of the analysis. Also included are flows from outside Canada. B y far the most popular destination for migrants from outside Canada was Ontario, followed by British Columbia, Alberta and Quebec. Considering the population size of Quebec, the inflow of foreigners is somewhat lower than expected. However, this relative lack of attractivity could in part well be attributed to the fact that the official language of Quebec is French, creating a cultural barrier for overseas or American immigrants. Furthermore, the Quebec Department of Immigration (a separate entity from the federal Department of Immigration with veto rights) tends to favour immigrants from French speaking countries such as France, Hait i and some African nations. Immigrants from these countries constitute a minority in the total immigration to Canada. In order to interpret migration patterns within Canada, the gross migration matrix was converted into a net exchange matrix for all occupations (see table 5.1.2; the table does 43 Table 5.1.1 Gross Migration, 1976-1981, A l l Occupations. From/to NFLD P.E.I. N.S. N.B. QUEBEC ONT. MAN. SASK. ALTA. B.C. YUKON N. • W.T. FOREIGN 1225 610 4195 3595 44715 136570 14260 6365 45310 50880 265 380 NFLD 27565 320 3905 1350 750 8810 895 490 6100 1940 90 365 PEI 165 5770 1485 795 105 1535 135 145 1935 455 25 35 NS 1820 1135 43790 4915 1660 12425 1260 940 10920 5205 105 350 NB 725 965 5435 33855 3950 8630 905 625 8455 2580 65 175 Quebec 760 500 3475 524 5 592165 73250 2715 1485 22025 13935 235 415 Ontar io 4960 1705 11865 7230 24015 753955 13615 8135 92 45 49810 1015 1820 Man i t . 385 50 1015 485 1105 8170 49510 8005 22930 14880 225 560 Sask. 70 115 4 60 290 525 3855 4275 69135 23840 9510 180 450 Alta. 425 375 2280 1050 1895 15105 4650 11195 168775 41785 680 1155 B.C. 345 310 2175 915 2890 16750 3900 5755 43310 267275 1645 800 Yukon 15 20 65 30 45 345 70 180 1220 2225 820 95 N.W.T. 105 20 115 40 285 725 285 410 2960 1245 340 1925 Note: Exchange between same province is internal migration Source: Statistics Canada (1981) Table 5.1.2 Net Exchange Matrix, 1976-1981, A l l Occupations. From/to NFLD P.E.I. N.S. N.B. QUEBEC ONT. MAN. SASK. ALTA. B.C. YUKON N.W.T. NFLD ****** 155 2085 625 -10 3850 510 420 5675 1595 75 260 PEI -155 ****** 350 -170 -395 -170 85 30 1560 145 5 15 NS -2085 -350 ****** -520 -1815 560 245 480 8640 3030 40 235 NB -625 170 520 ****** -1295 1400 420 335 7405 1665 35 135 Quebec 10 395 1815 1295 ****** 49235 1610 960 20130 11045 190 130 Ontario -3850 170 -560 -1400 -49235 * * * * * * 5445 4280 77140 33060 670 1095 Manit. -510 -85 -245 -420 -1610 -5445 ****** 3730 18280 10980 155 275 Sask. -420 -30 -480 -335 r960 -4280 -3730 ****** 12645 3755 0 40 Alta. -5675 -1560 -8640 -7405 -20130 -77140 -18280 -12645 ****** -1525 -540 -1805 B.C. -1595 -145 -3030 -1665 -110 4 5 -33060 -10980 -3755 1525 ****** -580 -445 Yukon -75 -5 -40' -35 -190 -670 -155 0 5*0 580 ****** -245 N.W.T. -260 -15 -235 -135 -130 -1095 -275 -40 1805 445 245 ****** NET IN -15240 -1300 -8460 -10165 -86815 -66815 -25105 -6205 155345 64745 295 -310 Source: Statistics Canada (1981) 44 not include inmigrants from outside Canada). Here the migration pattern is more readily apparent than in the previous table. In terms of total net inmigration from all provinces, Alberta was able to attract 155,345 more migrants than it lost to other provinces. The only other province that experienced positive net migration was B . C . , with 64,775. This tremendous gain for Alberta is no doubt related to the boom of the oil industry and the resulting economic spinoffs. The mass exodus from Quebec is again not surprising, considering that the Partie Quebecois came to power in 1976. The subsequent introduction of Bi l l 101, the controversial language law requiring all business transactions to be conducted in French and prohibiting advertising in English, is most likely to be one of the major determinants of the net loss. With the exception of Newfoundland, Quebec was the only province to lose migrants to very other province. Perhaps a bit surprising is the relatively large loss for Ontario (66,815). For the Maritimes the loss was to be expected. Newfoundland, Prince E d w a r d Island and New Brunswick have traditionally been areas of low employment opportunities. O f the opportunities that do exist, m a n y are seasonal jobs, particularly in fisheries and related occupations. Overal l , there was a pronounced east to west movement between 1976 and 1981 (see table 5.1.2). This is indicated by the fact that, again with the exception of Quebec, almost all of the values above the diagonal in the matrix are positive, whereas below they are negative. F r o m Ontario westwards, provinces gained from eastern provinces and lost to western provinces. Clearly the pole of attraction was Alberta, obtaining a net gain even from neighbouring Brit ish Columbia which, like Alberta, gained from all other provinces. 5.2 Labour Force Adjusted Migration Rates. Labour force adjusted migration rates according to formula (4.2.1) display a more detailed pattern. Tables 5.2.1 to 5.2.16 show migration rates broken down by occupation. Tables 5.2.17 shows migration rates for all occupations. F o r reasons discussed above Prince E d w a r d Island, the Yukon and the Northwest Territories are not included in the tables. The 45 Table 5.2.1 Migration Rates Adjusted for Size of the Labour Force, 1976-1981, Managerial, Administrative. FROM/TO NFLD N s. N B . QUEBEC ONT . MAN. SASK. ALTA . B C . NEWFOUNDLAND 0 0 16 36 6 24 O 31 1 06 1 01 O 94 2 86 0 49 NOVA SCOTIA 12 70 0 0 22 97 0 39 1 82 1 04 2 26 5 45 1 76 NEW BRUNSWICK 4, 52 20 75 0 O 0 90 1 09 0 77 1 40 4 32 1 08 QUEBEC 0 28 0 86 1 21 0 0 1 50 0 27 0 18 1 56 0 75 ONTARIO 0 79 1 59 1 03 0 38 0 0 1 00 0 96 3 75 1 58 MANITOBA 0 51 0 90 0 28 0 15 1 12 0 0 12 13 10 24 4 55 SASKATCHEWAN 0 16 0 97 0 43 O 07 0 58 5 76 0 0 14 1 1 3 94 ALBERTA 0 43 1 05 0 84 0 10 0 83 2 01 8 41 0 0 7 62 BRITISH COLUMBIA 0 25 1 08 0 30 0 15 0 72 1 22 3 22 8 22 0 0 Table 5.2.2 Migration Rates Adjusted for Size of the Labour Force, 1976-1981, Natural Sciences. FROM/TO NFLD N s. N B . QUEBEC ONT.. MAN. SASK. ALTA . B C. NEWFOUNDLAND 0 o 15 66 3 48 0 84 1 34 1 63 0 95 8 89 1 08 NOVA SCOTIA 12 05 0 0 10 44 0 49 2 06 0 94 0 82 1 1 96 3 90 NEW BRUNSWICK 4 17 8 63 0 0 0 54 0 88 0 63 2 69 4 32 0 72 QUEBEC 0 63 1 12 0 63 0 0 1 40 0 45 0 58 2 23 0 94 ONTARIO 0 79 1 65 0 49 0 38 0 0 0 96 0 96 4 78 1 49 MANITOBA 1 36 0 94 0 8 1 0 26 1 01 0 0 1 1 25 9 78 3 63 SASKATCHEWAN 0 48 1 10 0 32 0 12 0 46 4 95 0 0 16 29 3 02 ALBERTA 0 73 1 43 O 29 0 20 0 61 0 9 1 5 19 0 0 5 10 BRITISH COLUMBIA 0 0 0 90 0 28 0 22 0 47 0 69 1 70 7 01 0 0 46 Table 5.2.3 Migration Rates Adjusted for Size of the Labour Force, 1976-1981, Social Sciences. FROM/TO NFLD N. , S . N, ,B. QUEBEC ONT . MAN . SASK . ALTA . B , C . NEWFOUNDLAND 0. .0 10 . 38 2 . 9 1 0. . 12 1 . 75 3 . 7 1 1 .80 2 . 65 0 . 30 NOVA SCOTIA 19 . . 28 0. .0 8 . . 73 0. . 73 2 . 10 1 . 18 1 . 43 9 . 13 2 . 73 NEW BRUNSWICK 8 . . 73 9 . . 75 0. .0 0. .64 0 . 98 0 .51 0 . 93 3 . 98 1 . 16 QUEBEC 0. .46 0. . 73 0. .96 0. 0 1 . .07 0. .55 0 .4 1 1 . . 15 0. .83 ONTARIO 0. .99 2 . .02 0. .69 0. .44 0. .0 1. . 4 1 1, . 25 4 . 85 2 . 19 MANITOBA O. 74 1 . 57 0. 51 0. 12 1 . . 55 0. .0 7 . 63 10. . 19 5 . .07 SASKATCHEWAN 0. 0 1 . .91 0. 62 0. 04 0. 51 4 . 05 0. O 9 . 10 2 . .83 ALBERTA 0. 88 1 . 64 0. 46 0. 13 0. 82 2 . 93 5 . . 26 0. .0 6 . . 38 BRITISH COLUMBIA 0. 30 1 . 45' 0. 1 1 0. 1.3 0. .77 1 . 69 3 . 32 4 .84 0. .0 Table 5.2.4 Migration Rates Adjusted for Size of the Labour Force, 1976-1981, Teachers FROM/TO NFLD N . S . N . B . QUEBEC ONT . MAN. SASK . ALTA . B . C . NEWFOUNDLAND O. .0 6 . 75 1 . 87 0 . 34 0 .84 1 . 67 0 .47 2 . 37 1 . 59 NOVA SCOTIA 10. . 29 0 .0 8 .63 0 . 50 1 . 54 1 . 44 2 .02 6 . 72 2 .56 NEW BRUNSWICK 0. 70 9 .00 0, ,0 0 .83 0 .60 0 . 67 1 .07 2 .96 0. . 57 QUEBEC 0. 47 0 .98 1 . 02 0. .0 1 . 2 1 0 . 48 0 . 3 1 1 .07 0. .91 ONTARIO 1 . 00 1 . 7 1 0. 61 0. 62 0. .0 1 . . 22 1 . . 35 2 .97 2 . 13 MANITOBA 0. 84 1 . .31 0. 29 0. 33 0. .94 0. O 5 . . 35 4 . 37 3 . 88 SASKATCHEWAN 0. 23 0. 37 0. 0 0. 08 0. 63 3 . 54 0. 0 9. 53 3 . 91 ALBERTA 1 . 13 2 . 03 0. 77 0. 18 0. 86 1 . 47 5. 92 0. 0 6 . 65 BRITISH COLUMBIA 0. 59 1 . 94 0. 06 0. 28 0. 89 1 . 09 2 . 49 5 . 34 0. 0 47 Table 5.2.5 Migration Rates Adjusted for Size of the Labour Force, 1976-1981, Medicine & Health. FROM/TO NFLD N. . S . N. .B. QUEBEC ONT . MAN. SASK. ALTA . B . C . NEWFOUNDLAND 0 .0 23 . 54 2 . 16 0. . 24 1 .64 3 .02 0 ,65 4 .62 4 . 27 NOVA SCOTIA 13 . 70 O .0 16 . . 74 0. 47 2 . 10 2 . 10 1 . , 2 1 6 . 26 5 .02 NEW BRUNSWICK 3 .40 14 . 27 0. .0 0. 68 1 . 16 0 .63 0. ,48 3 . 39 0 .99 QUEBEC 0. , 28 0. . 92 1 . 14 0. .0 0 . 86 0. . 38 0. , 24 0 . 93 0 . 77 ONTARIO 1 . . 10 2 . . 23 0. 72 0. 35 0 .0 1 . . 12 0. 94 2 .84 2 , . 32 MANITOBA 0. ,40 2. . 72 0. 63 0. 13 ' 1 .03 0. .0 6 . 44 7 . . 58 6 ,01 SASKATCHEWAN 0. . 22 1 . 48 0. 10 0. 05 0. . 53 2 . 97 0. 0 10. . 1 1 5 . , 24 ALBERTA 0. 35 1 . 78 0. 4 1 O. 16 0. . 70 1 . . 74 3 . ,73 0. .0 6 . ,90 BRITISH COLUMBIA 0. 68 1 . 97 0. 56 0. 13 0. . 56 1 . 29 1 . 83 4 . .34 0. ,0 Table 5.2.6 Migration Rates Adjusted for Size of the Labour Force, 1976-1981, Fine & Commercial Artists. FROM/TO NFLD N . S . N. ,B . QUEBEC ONT . MAN. SASK. ALTA . B . , C . NEWFOUNDLAND 0, .0 1 1 , .76 1 . ,59 0. ,52 2 . 23 2 . . 52 3 . 56 2 , 70. 0. , 79 NOVA SCOTIA 1 , , 68 0, .0 9 . 28 0. 65 1 , .99 2 , . 94 2 , .77 3 , 15 2 . .44 NEW BRUNSWICK O, ,0 10 . 52 0. ,0 1 . ,00 1 , . 17 0. .46 1 , .31 2 . . 39 1 , , 30 QUEBEC 0, . 26 1 .01 0. 90 0. ,0 1 . 10 0. . 57 0 . 32 0 .99 0. . 73 ONTARIO 0, , 74 1 .92 0. 96 0. .60 0 .0 1. . 74 1 .76 3 .52 2 . ,06 MANITOBA 3 . , 78 0 .98 0. 46 0. , 4 1 1 , .66 0 ,0 16 , .08 13 . . 23 3 , .89 SASKATCHEWAN 0, ,0 0. .0 0. 0 0. 43 0 .84 10 .90 0, .0 16. .01 3 , .88 ALBERTA O. , 54 3 . 15 0, 80 0. , 27 1 .06 2 . 21 10, .01 0 .O 7 .90 BRITISH COLUMBIA 0, . 39 2 .90 0. . 29 0, , 25 0 .95 1 . 37 3 .07 6 . 53 0 .0 48 Table 5.2.7 Migration Rates Adjusted for Size of the Labour Force, 1976-1981, Clerical Workers. FROM/TO NFLD N .S. N, .B. QUEBEC ONT . MAN. SASK. ALTA . B .C . NEWFOUNDLAND 0 .0 15 . 74 4 . 95 O. . 24 2 .07 1 . 54 1 . 84 9 . 84 1 .67 NOVA SCOT IA 6 . 12 0 .0 8 .92 0 . 43 1 .76 1 .88 1 .40 9 . 3 1 2 .94 NEW BRUNSWICK 2 . 42 1 1 . 29 0. .0 0. . 78 1 .07 0 . 94 0 . 99 5 . 84 1 . 44 QUEBEC 0. , 32 0. .69 0. .81 0. 0 1 . 18 0. . 34 0. . 16 1 . .85 0 .85 ONTARIO 0, ,95 1 . . 27 0. .69 0. . 35 0 •0 0 .86 0 . 75 4 . 4 1 1 .61 MANITOBA 0. .58 1 . ,48 0. 29 0. . 10 0. .67 0. ,0 7 . . 13 9, , 22 4 . 26 SASKATCHEWAN 0. 32 0. .73 0. 35 0. .09 0. . 39 4 . .69 0. .0 17 . .38 4 , .91 ALBERTA 0. 50 1 . 61 0. 79 0. 14 0. .64 1 . .81 7 . 27 0. .0 7 . ,93 BRITISH COLUMBIA 0. 24 0. 82 0. 42 0. 14 0. 56 1 . 00 2 . 35 9 . 58 0. .0 Table 5.2.8 Migration Rates Adjusted for Size of the Labour Force, 1976-1981, Sales Occupations. FROM/TO NFLD N . s. N . B . QUEBEC ONT . MAN. SASK. ALTA . B .C . NEWFOUNDLAND 0 .0 10 .59 2 . 75 0 . 23 0 .98 0 . 58 0 . 12 5 .63 0 . 50 NOVA SCOT IA 4 .51 0 .0 8 .38 0 .33 1 .37 1 .08 0, .80 5 .63 1 . 88 NEW BRUNSWICK 1 , . 13 10. . 43 0. .0 0 . 72 0. . 73 0 .42 0. 49 3 .48 0. .80 QUEBEC 0. , 12 0. 8 1 0. 85 0 .0 1 . . 28 0 . 28 0. 10 1 . .41 0, ,68 ONTARIO 0. . 94 1 . .61 0. 72 0 . 34 0. 0 1 .06 0. 61 3 .93 1 . . 5 1 MANITOBA 0. .69 1 . 02 0. 31 0. . 16 0. 79 0. ,0 7 . 30 10. . 74 4 . ,44 SASKATCHEWAN 0. 36 0. 20 0. 22 0. .06 0. 26 3 . 01 0. 0 10. .05 2 . 64 ALBERTA 0. 26 0. 90 0. 57 0. .08 0. 66 1 . ,83 4 . 60 0. 0 7 . 54 BRITISH COLUMBIA 0. 32 0. 58 0. 23 0. 14 0. 55 1 . 00 1 . 46 7 . 34 0. 0 49 T a b l e 5.2.9 M i g r a t i o n Ra t e s A d j u s t e d for S i z e o f t he L a b o u r F o r c e , 1976-1981, S e r v i c e O c c u p a t i o n s . FROM/TO NFLD N. . S . N. B. QUEBEC ONT . MAN. SASK. ALTA . B , C . NEWFOUNDLAND 0. .0 25 . 18 4 . 9 1 0. .21 2 . 46 1 . . 73 1 . .69 5 . 88 2 . 09 NOVA SCOTIA 7 . 18 0. .0 6 . 12 0. .47 o 38 2 .24 2 . 05 7 • 3.1 3 . , 18 NEW BRUNSWICK 1 . .64 6. .98 0, .0 0. .69 0. .92 1 .01 0. .62 2 .92 0. ,83 QUEBEC 0. . 25 1 . , 39 0 .98 0. 0 0. .93 0. .42 0. , 40 1 . . 48 0. .80 ONTARIO 1 . .09 2 . 49 o . 90 0. . 33 O. ,0 1 . 33 0. . 82 3 . 48 1 . .58 MANITOBA 0, .72 1 , .60 0. . 68 0. . 10 0. .86 0 .0 3. . 72 5 .88 3 . 29 SASKATCHEWAN 0. . 29 1 . . 30 0. . 55 0. . 14 0. . 40 2 .80 o: .0 7 . 77 2 . 9 1 ALBERTA 0. .55 2 . . 33 0 .63 0. , 22 0. ,75 1 .87 4 .01 0 .0 5 .82 BRITISH COLUMBIA 0, .29 1 . .96 0. .49 0. . 20 0. ,59 1 .36 1 .88 5 .81 0. .0 T a b l e 5.2.10 M i g r a t i o n Ra t e s A d j u s t e d for S i z e o f the L a b o u r F o r c e , 1976-1981, A g r i c u l t u r e . FROM/TO NFLD N . S . N . B . QUEBEC ONT . MAN. SASK. ALTA . B . C . NEWFOUNDLAND 0 .0 49 .52 28 .95 0. .0 8 . 17 2 .51 2 .53 7 .08 9 . 19 NOVA SCOTIA 8 . 25 0 .0 21 . 59 0 . 76 • 2 .02 0. .0 0 .28 4 .43 7 .20 NEW BRUNSWICK 0 .0 28 . .06 0. .0 2 , .67 2 . 47 0. .66 0 .33 4 .07 4 .21 QUEBEC 0 .0 0. .95 1 . . 78 0. .0 1 .81 0. . 14 0 .07 0 .82 2 . 23 ONTARIO 1 , . 72 3 . 08 1 . , 24 0. 44 0 • 0 0. 83 0. ,21 2 , 49 3, .94 MANITOBA 0, ,0 O. 28 0. 0 0. 03 0, . 44 0. 0 1 . ,37 1 . , 59 4 , ,06 SASKATCHEWAN 0. .0 O. 42 0 . 17 O. 0 O, . 1 1 o . 58 0. 0 2 . OO 1 . ,93 ALBERTA 0. 0 1 . 1 1 0 . 37 0 . 05 0. 44 0 . 87 1 . 84 0. 0 7 . , 19 BRITISH COLUMBIA 0. O 5 . 14 2 . 40 0. 58 1 . 26 1. 87 2 . 56 7 . 36 O. O 50 Table 5.2.11 Migration Rates Adjusted for Size of the Labour Force, 1976-1981, Primary Occupations. FROM/TO NFLD N. . s. N. ,B. QUEBEC ONT . MAN. SASK. ALTA . B . ,C. NEWFOUNDLAND 0. .0 1 .82 0. . 29 0. .09 0. .64 5 . 15 0. 90 5. . 10 0. . 30 NOVA SCOTIA 0 .87 0 .0 0. . 48 0. .03 0. .42 1 . 27 1 . 05 5 .94 0. . 78 NEW BRUNSWICK 0. . 18 0. . 82 0. .0 0. .31 0. 35 0. 95 0. 15 2 .91 0. .50 QUEBEC 0. .07 0 . 15 0. . 74 0. 0 0. 98 0. 54 0. 34 2 .04 0. .67 ONTARIO 1 , ,45 1 . 37 0. . 7 1 0. 38 0. .0 2 . 65 1 . 93 8 . 73 1 . ,73 MANITOBA 1 . .09 1. .09 0. . 32 0. . 12 1 . 96 0. .0 10, .51 22 . 39 3 . 67 SASKATCHEWAN 0. .0 0. .0 0. . 15 0. 0 0. 39 5 . 66 0. .0 30 . 15 2 . , 30 ALBERTA 0, , 5G 1. . 1 1 0. .40 0. 03 0. 53 2 . 15 10. .94 0 .0 5 , 55 BRITISH COLUMBIA 0. . 12 . 0. . 32 0. ,05 0. 02 0. 26 0. 53 0. .96 6 . 57 0. ,0 Table 5.2.12 Migration Rates Adjusted for Size of the Labour Force, 1976-1981, Processing. FROM/TO NFLD N .S. N .B . QUEBEC ONT . MAN . SASK. ALTA . B .C . NEWFOUNDLAND 0 .0 8 . 14 1 .44 0 .21 2 .67 3 . 48 4 .77 10 . 58 1 . 70 NOVA SCOTIA 1 . 68 0 .0 3 . 55 •O . 20 1 . 56 1 . 9 1 0 .0 6 .96 2 .42 NEW BRUNSWICK 0 . 29 5 .68 0 .0 0 . 78 1 . 28 3 .42 0. .0 7 . 73 1 .30 QUEBEC 0 . 17 0. .31 1 . 47 0. .0 0. . 73 0 . 29 0, . 24 1 . .72 0. .68 ONTARIO 1 , .84 1 . .99 1 . . 16 0. .31 0. .0 1 .30 1 , .08 7 . 13 1 . ,86 MANITOBA 0. .87 0. .48 0. 49 0. 06 1. . 17 0. ,0 10. 53 17 . 98 5. .65 SASKATCHEWAN 0. 0 0. 0 0. 40 0. 14 0. 52 5 . 67 0. 0 32 . 52 7 . 89 ALBERTA 0. 24 2 . 32 0. 79 0. 17 1. 03 4 . , 20 9 . 53 0. 0 1 1 . 56 BRITISH COLUMBIA 0. 08 0. 38 o. 52 0. 05 0. 39 1 . 05 1 . 62 9 . 16 0. 0 51 Table 5.2.13 Migration Rates Adjusted for Size of the Labour Force, 1976-1981, Manufacturing. FROM/TO NFLD N . S . N . B . QUEBEC ONT . MAN . SASK . ALTA . B . C . NEWFOUNDLAND 0 .0 29 . , 64 8 . 27 0. .61 6 . . 37 4 .66 2 . 34 27 . 75 6 . 33 NOVA SCOTIA 6 . 59 0 .0 14 . 64 0. . 38 2 . 52 2 . 23 4 . 44 19 . 56 4 . 44 NEW BRUNSWICK i. . 76 16 . 04 0. .0 0. .81 1 . 30 1 .66 1 . 73 13 .60 2 . 13 QUEBEC 0. . 14 0. . 56 1 . .09 0. .0 0. . 63 0. . 35 0. . 33 2 .46 0 , 75 ONTARIO 1 . .48 1 . 76 0. 89 0. 24 0. 0 1 . .02 1 . . 22 6 . 26 1 . .97 MANITOBA 0. . 74 1 . . 24 0. 62 0. . 17 0. 57 0. .0 12 . 16 15 .96 7 . 30 SASKATCHEWAN 0. 0 0. 89 1 . .73 0. 13 0. 4 1 6 . .87 0. .0 30 .99 7 . 45 ALBERTA 1 . 19 2 . 51 1 . 43 0. 17 0. 98 3 . 36 15 . 92 0. .0 15 . 27 BRIT ISH COLUMBIA 0. 26 0 . 98 0. 49 0. 22 0. 63 1 . .46 4 . 59 15 .32 0. .0 Table 5.2.14 Migration Rates Adjusted for Size of the Labour Force, 1976-1981, Construction. FROM/TO NFLD N . S . N . B. QUEBEC ONT . MAN . SASK . ALTA . B . C . NEWFOUNDLAND 0 .0 1 . 50 0 .89 0 . 14 0 .98 0 .66 1 . 68 7 .69 0 .68 NOVA SCOTIA 1 . 15 0. .0 1 . 69 0 . 1 1 0. . 72 0 . 59 1 . 34 9 .93 1 . .65 NEW BRUNSWICK 0 . 53 2 . 19 0 .0 0 . 23 0. . 46 0 . 27 0 . 39 6 .80 0 .70 QUEBEC 0. .05 0. 21 0. . 43 0. .0 0. 40 0. .07 0. . 22 2 . 7 1 0. .53 ONTARIO 0. 60 0 . 73 0. 5 1 0. . 17 0. 0 0. . 37 0. .68 6 . 36 1. 51 MANITOBA 0. 37 0 . 29 0. 14 0. 03 0. 38 0. O 6 . . 5 1 10. . 82 3 . 61 SASKATCHEWAN 0 . 0 0 . 17 0. 06 0. 05 0 . 15 1. 33 0. 0 15. .04 2. 97 ALBERTA 0 . 12 1 . 38 0 . 20 0 . 09 0 . 35 1. 07 5 . 82 0. 0 6 . 42 BRIT ISH COLUMBIA 0 . 06 0 . 27 0 . 23 0. 06 O. 28 o . 46 2 . 20 7 . 68 0. 0 52 Table 5.2.15 Migration Rates Adjusted for Size of the Labour Force, 1976-1981, Transportation. FROM/TO NFLD N. , s. N . B . QUEBEC ONT . MAN. SASK. ALTA . B . C . NEWFOUNDLAND 0. .0 5 . 31 1 . .95 0. , 20 0. .97 0 . 65 0. 0 4 .73 0 .89 NOVA SCOTIA 4 . 85 0. .0 4 . 05 0. , 20 1 . 14 1 . . 17 0. 49 6 . 84 2 . 28 NEW BRUNSWICK o". .97 5. .07 0. .0 0. 23 0. 59 0. .51 0. 23 4 . 15 0 . 44 QUEBEC 0. .45 0. .40 0. .58 0. .0 0. 76 0 .34 0\ 16 2 .08 0. .51 ONTARIO 1 . 32 1 . 49 0, .66 0, , 27 0. .0 1 . 53 0. 95 6 . 43 1 . . 37 MANITOBA 0. 65 1 . 31 0. .41 0. . 23 0. .68 0 .0 5. 77 1 1 .86 3 .82 SASKATCHEWAN 0. .0 0. .49 0. . 1 1 0. .10 0. .23 3 . 81 0. 0 14 .42 2 . 89 ALBERTA 0. .21 1 . ,50 0. ,40 0. 19 0. 86 2 . , 78 4 . 76 0 .0 9 . , 1 1 BRITISH COLUMBIA 0. . 13 0. .80 0, . 28 0. . 12 0. .46 1 .05 1 . 7 1 10 .69 0 .0 Table 5.2.16 Migration Rates Adjusted for Size of the Labour Force, 1976-1981, Other Occupations. FROM/TO NFLD N . s. N .B . QUEBEC ONT . MAN. SASK. ALTA . B .C . NEWFOUNDLAND 0 .0 9 .4 1 5 .97 0 .04 4 .82 2 .87 1 .03 8 .89 2 . 56 NOVA SCOTIA 3 . 76 0 .0 9 . 46 0. . 39 1 . 77 1 .28 1 . 38 7 .93 2 . 26 NEW BRUNSWICK 1 .66 1 1 .04 0. .0 0 . 59 1 .41 1 . 20 0 .65 8 .71 1 . 50 QUEBEC O . 25 O .67 0. 75 0 .0 0 .71 0 . 23 0, . 25 1 . .67 0 . 58 ONTARIO 1 , 64 1 . 56 1 . 00 0. 25 0 .0 1 , . 12 0, .87 4 . 58 1 . 70 MANITOBA 1 , . 9 1 1 . 00 0. 75 0. 15 0 . 88 0, 0 7 , 72 1 1 . . 75 5 .05 SASKATCHEWAN O. 0 • 0. 77 0. 49 0. 09 0. ,6 1 6 . 20 0. 0 13 . 17 3 . 63 ALBERTA 0. 52 1 . 56 0. 66 0. 06 0. 84 2 . , 55 5 . 69 0. 0 8 . 57 BRITISH COLUMBIA 0. 44 1 . 10 0. 39 0. 08 0. 49 1 . 76 2 . 55 7 . 96 0. .0 53 Table 5.2.17 Migration Rates Adjusted for Size of the Labour Force, 1976-1981, A l l Occupations. From/to NFLD NS NB QUEB. ONT MAN • SASK. ALTA. BC Newfoundland * * * * * 15. 08 4.91 .29 2. 44 2. 32 1. 58 7. 33 2. 15 Nova S c o t i a 7. 17 * * * * * 9.73 .42 1. 70 1. 46 1. 48 7. 91 2. 97 New Brunswick 2. 07 10. 66 ***** . 78 1. 04 • 91 • 84 5. 10 1. 23 Quebec • 26 • 74 .96 ***** 1. 03 • 37 • 27 1. 64 • 84 Ontar i o 1. 15 1. 78 .81 . 37 * * * * * 1. 22 1. 02 4. 78 1. 91 Manitoba • 95 1. 15 .43 .16 • 98 * * * * * 8. 23 10. 85 4 . 51 Saskatchewan • 13 . 68 . 36 . 10 • 43 4. 55 * * * * * 15. 54 3. 93 A l b e r t a • 51 1. 74 .61 .14 75 2. 11 6. 81 ***** 7. 84 B r i t i s h Columbia 26 1. 48 .44 .17 61 1. 22 2. 34 7. 73 ** * * * 54 following section will first describe the overall Canadian pattern of migration and then look at the pattern for individual occupations. A number of generalizations can be made from the tables. First , with one exception (agriculture) Quebec experienced far less inmigration than expected. The adjusted migration rates are in most cases well below 1.00. A t the same time, however, outmigration from Quebec was also general^ less than expected. This reasons for this are the cultural and political explained above. Thus French speaking Quebecers are likely to have a much lower tendency to migrate, whereas of those that do migrate to another province the majority will in all probability be English speaking. The destination of most of the migrants from Quebec is either New Brunswick (also bilingual), Ontario or Alberta , although inmigration from Quebec to Alberta was not as high as for the other provinces. Second, there was a great deal of interaction within the Marit ime provinces. Although the propensity to migrate was the highest in the Marit imes, migrants from those provinces do tend to stay within the Marit imes, i.e., move to another Marit ime province. The major non-maritime destination for the Atlantic provinces was Alberta. N o v a Scotia seemed to be the Marit ime pole of attraction. In many cases inmigration was greater than expected. Most of the migrants likely moved to Halifax, the major growth centre in the Marit imes between 1976 and 1981. Third , the attractivity of Alberta was more readily apparent in the adjusted migration rates. In only three cases was inmigration less than expected on the basis of Alberta's share of the Canadian labour force. Final ly , migration rates tend to be greater than 1.00 between neighbouring provinces and Alberta, and less than 1.00 for other provinces. The above findings are similar in two instances with those of L y c a n (1969). The Quebec pattern and the greater propensity to migrate within the Marit ime provinces also existed between 1956 and 1961. However, Alberta was not as popular a destination. In addition, the variation in migration rates was also not as great as it was between 1976 and 1981. Increased mobility leading to higher migration rates appears to be a national trend. 55 Table 5.3.1 Regression of Labour Force Adjusted Migration Rates with Distance and Intervening Opportunities, by Province of Origin. Province Newfoundland Nova Scotia New Brunswick Quebec Ontario Manitoba Saskatchewan Alberta British Columbia Canada* R .578" .313" .460:' .317" .695 * .591 .746:' .734:' .787* .333" R z .334 .098 .211 .101 .483 .349 .557 .539 .619 .111 ln(D) -.577* -.306* -.444* -.317* .693* -.221* -.432* -.485* -.559* -.234* IO .553* .304* .350* .252^ -.048 -.163* -.009 .155:l .211 .043 : significant at 10% * * : significant at 5% * * * : significant at 1% x : does not include Prince Edward Island, Yukon and Northwest Territories 56 Although Alberta was the major region of attraction for every occupation, this attractivity was more pronounced in the blue collar and resource-related occupations such as primary occupations, processing and manufacturing. In general, however, blue collar workers do not seem to have as high a propensity to migrate as white collar workers. The reason for this is likely to be found in the social and cultural characteristics of blue collar workers. One of these is that they are more conservative and thus more deeply rooted to their region of origin. Nova Scotia tends to attract migrants from white collar occupational groups. Again, this is most likely related to the development of Halifax as the gateway to eastern Canada. Halifax's fastest growing industry, the port, requires a large number of service industries. These service industries have in turn been able to attract many other, non port-related activities. Apart from the above observations, labour force adjusted migration rates do not show any notable variation between occupational groups. 5.3 Migration Rates Adjusted for Labour Force, Distance and Intervening Opportunities. The next step in the analysis is to determine the influence of spatial factors (distance and intervening opportunities) on the adjusted migration rates, and in turn adjust the rates according to equation (4.2.4). Table 5.3.1 shows the results of the multiple regression. Overall distance and intervening opportunities explained 11.1% of the variation in migration rates, significant at 1%. This is very much lower than the result of Lycan's (1969) study, where the multiple r showed that 49.8% of the variation was explained by spatial factors. As was to be'expected, however, the explanatory powers of the two spatial factors vary considerably between provinces of origin, and are, in most cases (Nova Scotia and Quebec excepted), higher than the coefficient based on all provinces. Generally the multiple R tends to increase in an east to west direction. In all cases the sign of distance was negative, except for Ontario where there was a highly significant positive relationship (62.7% of migrating Ontarians moved to Alberta or 57 British Columbia). However, intervening opportunities had a positive sign in all provinces save Ontario, Manitoba and Saskatchewan. For the eastern provinces the relationship between distance and intervening opportunities is one of a mirror image. In addition, the partial regression coefficients are almost equal to the multiple regression coefficient. The major reason that intervening opportunities show a positive correlation is that, as mentioned earlier, migrants from the Marit ime provinces tend to go to Alberta and British Columbia if they do not stay within the Marit imes. A s the majority of the non-Maritime moves are therefore destined to provinces west of Ontario, it is not surprising that the sign is positive. The largest proportion of Canada's labour force is in the industrial heartland of Ontario and Quebec, which tends to be "skipped" by migrants from the Marit imes, thus explaining the positive partial correlation coefficient. Overall the amount of variation explained by spatial factors for migration to the Marit imes was relatively low, compared to the rest of Canada. Again , this is accounted for by the fact that there are only two major destinations: neighbouring provinces or Alberta. For Quebec the coefficients were comparatively low. This can be explained by the same reasons as above: the small amount of migration that does occur is either to New Brunswick, Ontario or Alberta. For Ontario the partial correlation of intervening opportunities was negative, but totally insignificant. This was to be expected as Ontario itself is the industrial heartland of Canada (along with Quebec). The strong positive relationship with distance can be explained by the fact that a good proportion of the moves are over long distances. The multiple r for the western provinces is relatively high compared to the rest of Canada, ranging from a low of .349 (Manitoba) to a high of .619 (British Columbia). While distance is highly significant in all cases (greater than 1%), intervening opportunities do not account for as much of the variation and are generally not as significant. The results of the multiple regression confirm that the use of an economy-wide equation would lead to errors. The explanatory power of spatial factors vary quite considerably, as do the signs of the partial correlation coefficients. 58 Table 5.3.2 Migration Rates Adjusted for Size of the Labour Force, Distance and Intervening Opportunities, 1976-1981, Managerial, Administrative. FROM/TO NFLD N .s. N .8 . QUEBEC ONT . MAN . SASK . ALTA . B . C . NEWFOUNDLAND 0 .0 1 .01 0. . 68 -0 . 29 - 0 .31 - 1 .46 - 1 . 32 0 . 24 - 1 .21 NOVA SCOTIA 2 . 29 0 .0 1 . .60 - 1 .40 -0 . 15 - 1 .04 -0 . 23 0 . 75 -0 . 34 NEW BRUNSWICK 1 . .91 1 . 53 0. .0 - 0 . .47 - 0 , , 4 1 -0 .84 - 0 , . 14 1 , . 16 - 0 . . 1 1 QUEBEC -o, , 19 0, .50 0. 67 0. ,0 0. ,49 - 0 .95 - 1 , .26 1 , .00 0 . 34 ONTARIO -o. , 75 0. .43 0 . 17 0. .02 0. ,0 - 0 , .20 - 0 , .44 0, .70 - 0 , ,42 MANITOBA 0. 26 0. 55 - 0 . 74 -2 . ,02 - 0 . 68 0, .0 0. .62 1 . . 20 1 . ,02 SASKATCHEWAN - 0 . 13 1. 24 0 . 29 - 1 . 98 - 0 . 14 - 0 , 38 0. 0 0. 97 1 . 27 ALBERTA 0 . 07 0 . 55 0 . 23 - 1 . 96 0 . 51 - 0 . 29 0. 14 0. 0 0. 89 BRITISH COLUMBIA - 0 . 07 0 . 95 - 0 . 43 - 1 . 15 0 . 95 0. 03 0. 40 0. 22 0 . 0 NOTE: POSITIVE VALUES INDICATE THAT MIGRATION IS GREATER THAN EXPECTED BASED ON POPULATION, DISTANCE AND INTERVENING OPPORTUNITY Table 5.3.3 Migration Rates Adjusted for Size of the Labour Force, Distance and Intervening Opportunities, 1976-1981, Natural Sciences FROM/TO NFLD N . S . N. .B . QUEBEC ONT . MAN. SASK. ALTA . B . C . NEWFOUNDLAND 0, .0 0 .97 0. . 10 0.61 O. . 27 - 0 .98 -1 , .29 1 .42 - 0 , . 4 1 NOVA SCOTIA 2 , . 24 0. .0 0. .81 - 1 .20 0. .07 - 1 . • 1 ^ - 1 , . 24 1 . 55 0, ,45 NEW BRUNSWICK 1 , . 83 0. . 66 0. .0 - 0 . 9 7 - 0 . . 48 - 1 . .00 0, . 55 1 .21 - 0 . ,48 QUEBEC O, .60 0. . 74 0. 02 0 .0 0. . 4 1 - 0 , ,49 - 0 . . 17 . 1 . 30 0. , 50 ONTARIO - 0 . 76 O. 46 - 0 . 59 0 .02 0. 0 - 0 . . 23 - 0 . , 44 0 . 94 - 0 . , 47 MANITOBA 1 . 24 0. 59 0. 32 -1 .41 - 0 . 79 0. 0 0. ,55 1 . 15 0. 80 SASKATCHEWAN 0.. 99. 1 . 37 - 0 . 02 - 1 . 49 - 0 . 37 - 0 . 54 0. 0 1 , . 12 1 . ,01 ALBERTA o . 6 1 O. 87 - 0 . 81 - 1 . 35 0. 20 - 1 . 07 - 0 . 34 0. .0 0. , 49 BRITISH COLUMBIA 0 . 0 0 . 77 - 0 . 49 - 0 . 8 2 0. 51 - 0 . 54 - 0 . 24 0, .06 0. 0 NOTE: POSITIVE VALUES INDICATE THAT MIGRATION IS GREATER THAN EXPECTED BASED ON POPULATION, DISTANCE AND INTERVENING OPPORTUNITY 59 Table 5.3.4 Migration Rates Adjusted for Size of the Labour Force, Distance and Intervening Opportunities, 1976-1981, Social Sciences. FROM/TO NFLD N, ,S. N, ,B. QUEBEC ONT . MAN. SASK . ALTA . B . ,C . NEWFOUNDLAND 0. .0 0. . 55 -o, ,09 - 1 . . 38 0, . 13 -0, .09 -0. .59 0. .20 - 1 . .62NOVA SCOTIA 2 . , 7 1 0. .0 0. ,63 -o, , 78 -0, ,03 -0. .90 -o. 66 1 ,  28 0. . 12 NEW BRUNSWICK 2.. , 57 O. , 78 O. ,0 -0, ,81 -0. .51 - 1 , , 18 -0. 48 1 ,  12 0 .01 QUEBEC 0. . 29 0, , 32 0. .44 0. ,0 0. , 14 -0. , 20 -0. 42 0. , 7 1 0 . 46 ONTARIO -0. , 52 O, ,67 -0. 24 0, , 17 0. ,0 0. . 15 -0. 18 0, ,96 -0. .09 MANITOBA 0. ,63 1 . ,09 -0. . 15 -2 . , 24 -0. 35 0, ,0 0. 16 1 . , 20 1 . . 13 SASKATCHEWAN 0. .0 1 . ,92 0. 66 -2 . ,67 -0. , 26 -0. . 74 0. 0 0, . 54 0 .94 ALBERTA 0. ,80 1 . ,01 -0. 35 - 1 , .73 0. 49 0. ,08 -0. 33 0. ,0 0. . 72 BRITISH COLUMBIA 0. , 15 1 . 25 -1. ^46 - 1 . ,31 1 . 01 0, . 35 0. 44 -0. . 30 0. ,0 NOTE: POSITIVE VALUES INDICATE THAT MIGRATION IS GREATER THAN EXPECTED BASED ON POPULATION, DISTANCE AND INTERVENING OPPORTUNITY Table 5.3.5 Migration Rates Adjusted for Size of the Labour Force, Distance and Intervening Opportunities, 1976-1981, Teachers FROM/TO NFLD N. . S . N . ,B . QUEBEC ONT . MAN . SASK. ALTA . B ,C . NEWFOUNDLAND 0 .0 0. . 12 -0. , 59 -0.43 -0. . 78 -0. .61 - 1 . 77 0. . 20 0. .05 NOVA SCOTIA 2, .08 0. .0 0. ,62 -1.18 -0. . 37 -0. .60 -0. 25 1 , ,02 0 .08 NEW BRUNSWICK 0, ,03 O, , 70 0. .0 -0.55 - 1 . .01 -0. . 79 -0. 25 0. .90 -0. .66 QUEBEC 0. .29 0. . 59 0, , 50 0.0 0. . 27 -0. 24' -0. 63 0. . 70 0. . 59 ONTARIO -0. . 5 1 0, .51 -0. . 34 0.52 0. .0 0. 00 -0. 10 0. . 47 -0. . 1 1 MANITOBA 0. . 7 1 0. .86 -0. 80 -1.31 -0. 85 0. O -0. 20 0. . 36 0. .89 SASKATCHEWAN 0, . 27 0. . 27 0. 0 - 1 . 94 -0, .06 -0. 87 0. 0 0. . 58 1 .  27 ALBERTA 1 . .06 1 . 25 0. 21 - 1 . 36 0. 52 -0. 61 -0. 21 0'. 0 0. . 76 BRITISH COLUMBIA 0. .83 1 . .56 -2 . 05 -0. 48 1. 1 1 -0. 1 1 0. 13 -0. .21 0. 0 NOTE: POSITIVE VALUES INDICATE THAT MIGRATION IS GREATER THAN EXPECTED BASED ON POPULATION, DISTANCE AND INTERVENING OPPORTUNITY 60 T a b l e 5.3.6 M i g r a t i o n R a t e s A d j u s t e d f o r S i z e o f the L a b o u r F o r c e , D i s t a n c e a n d I n t e r v e n i n g O p p o r t u n i t i e s , 1976-1981, M e d i c i n e a n d H e a l t h . FROM/TO NFLD N .S. N .B . QUEBEC ONT . MAN . SASK . ALTA . B , c . NEWFOUNDLAND 0. .0 1 . 37 - 0 , .40 - 0 , . 67 0 . 20 0 .04 - 1 . 33 0. .92 1 .06 NOVA SCOTIA 2 . 37 0 .0 1 , 28 - 1 . 23 0. .02 - 0 . . 22 - 0 . . 74 0. .96 0 . 74 NEW BRUNSWICK 1 . ,62 1 . . 16 0, .0 - o . 75 - 0 . . 28 - 0 . . 87 - 1 . .04 1 . .02 - 0 . . 13 QUEBEC - 0 . .21 0 , . 54 0 . ,61 0. ,0 - 0 . ,07 - 0 . . 52 - 0 . .93 0 . 51 0. . 37 ONTARIO - 0 , ,42 0. . 77 - 0 , . 19 - 0 . .06 0 . ,0 - 0 . .08 - 0 . .46 0 . .43 - 0 . ,02 MANITOBA - 0 . .03 1. .59 - 0 , ,01 -2 . , 22 - 0 . . 77 0. .0 - 0 . .01 0. 92 1 . . 34 SASKATCHEWAN 0 . , 20 1. .66 -1 , . 20 -2 . . 37 - 0 . , 23 - 1 . 05 0 . .0 0 . 64 1 . .56 ALBERTA - 0 . 1 1 1. 12 - 0 . 43 - 1 . 50 0 . 31 - 0 . 45 - 0 . 67 0 . 0 0 . 80 BRITISH COLUMBIA 0 . 97 1. ,57 0 , ,23 - 1 , .30 0 . .64 0 . .04 - 0 . 19 - 0 . 42 0 . 0 NOTE: POSITIVE VALUES INDICATE THAT MIGRATION IS GREATER THAN EXPECTED BASED ON POPULATION, DISTANCE AND INTERVENING OPPORTUNITY T a b l e 5.3.7 M i g r a t i o n R a t e s A d j u s t e d fo r S i z e o f the L a b o u r F o r c e , D i s t a n c e a n d I n t e r v e n i n g O p p o r t u n i t i e s , 1976-1981, F i n e a n d C o m m e r c i a l A r t i s t s . FROM/TO NFLD N . s . N .B. QUEBEC ONT . MAN . SASK. ALTA . B .C. NEWFOUNDLAND 0 .0 0 .68 - o .71 0 . 17 0 . 18 - 0 . 72 - 0 .05 0 . 17 - 0 .73 NOVA SCOTIA 0 . 27 0 .0 0 . 69 - 0 . . 88 - 0 . 13 - 0 .05 - 0 .04 0 . 20 - 0 .00 NEW BRUNSWICK 0 .0 0 .85 0 .0 - 0 . 36 - 0 . 4 1 - 1 . 38 - 0 .21 0 . 58 0 .09 QUEBEC - 0 . . 28 0 .64 0 . 38 0 . .0 0 . . 18 - o . . 18 - 0 . .66 0 . 59 o . . 36 ONTARIO - 0 . .80 0 . 63 0 . 1 1 0 . 49 0 . .0 0 . 36 0 . 16 0 . .63 - 0 . ,15-MANITOBA 2 . 30 0 . 66 - 0 . 20 - 1 . 01 - 0 . 29 0 . 0 0 . 90 1 . . 44 0 . 86 SASKATCHEWAN 0 . 0 0 . 0 0 . 0 - 0 . 23 0 . 23 0 . 25 0 . 0 1 . 10 1 . 26 ALBERTA O. 30 1. 66 0 . 18 - 0 . 94 0 . 78 - 0 . 18 0 . 32 0 . 0 0 . 93 BRITISH COLUMBIA 0 . 40 1. 94 - 0 . 47 - 0 . 6 1 1. 26 0 . 17 0 . 36 - 0 . 01 0 . 0 NOTE: POSITIVE VALUES INDICATE THAT MIGRATION IS GREATER THAN EXPECTED BASED ON POPULATION, DISTANCE AND INTERVENING OPPORTUNITY 61 Table 5.3.8 Migration Rates Adjusted for Size of the Labour Force, Distance and Intervening Opportunities, 1876-1981, Clerical Workers. FROM/TO NFLD N S . N B . QUEBEC ONT . MAN. SASK . ALTA . B C . NEWFOUNDLAND O 0 0 97 O 43 -0 62 0 49 -0 98 -0 6 1 1 52 0 07 NOVA SCOTIA 1 1 5G 0 0 0 65 - 1 32 -0 14 -o 43 -0 69 1 30 0 20 NEW BRUNSWICK 1 28 O 93 0 O -0 61 -0 36 -0 59 -0 44 1 50 0 22 QUEBEC -0 08 0 26 0 27 0 0 0 25 -0 72 - 1 40 1 16 0 45 ONTARIO -0 56 0 20 -0 24 -0 04 0 0 -0 34 -0 69 0 86 -0 39 MANITOBA 0 38 1 03 -0 71 -2 42 - 1 19 0 0 0 09 1 09 0 96 SASKATCHEWAN 0 60 0 96 0 07 - 1 77 -0 55 -0 59 0 0 1 18 1 49 ALBERTA 0 23 0 99 0 19 - 1 64 0 25 -o 39 -o 00 0 0 0 93 BRITISH COLUMBIA -0 07 0 69 -0 09 - 1 26 0 70 -0 17 0 09 0 38 0 0 NOTE: POSITIVE VALUES INDICATE THAT MIGRATION IS GREATER THAN EXPECTED BASED ON POPULATION. DISTANCE AND INTERVENING OPPORTUNITY Table 5.3.9 Migration Rates Adjusted for Size of the Labour Force, Distance and Intervening Opportunities, 1976-1981, Sales Occupations. FROM/TO NFLD N s. N B. QUEBEC ONT . MAN. SASK . ALTA . B c. NEWFOUNDLAND 0 0 0 58 -0 20 -0 76 -0 23 - 1 55 -2 93 1 23 -0 95 NOVA SCOTIA 1 26 0 0 0 59 - 1 59 -0 37 -0 85 - 1 12 0 89 -0 19 i NEW BRUNSWICK 0 51 0 85 0 0 -0 69 -0 70 - 1 23 -0 98 1 10 -0 28 | QUEBEC -1 1 1 0 42 0 31 0 0 0 32 -0 82 - 1 79 0 93 0 26 I ONTARIO -0 58 0 44 -0 19 -0 09 0 0 -0 13 -0 90 0 75 -0 45 MANITOBA 0 50 0 59 -0 7 1 -2 02 -1 03 0 0 0 1 1 1 27 1 03 SASKATCHEWAN 0 70 -0 35 -0 39 -2 16 -0 96 -1 03 0 0 0 63 0 88 ALBERTA -0 37 o 46 -o 09 -2 18 o 25 -o 40 -0 46 0 0 0 88 BRITISH COLUMBIA 0 23 0 38 -0 64 - 1 22 0 63 -0 2 1 -0 41 0 1 1 0 0 NOTE: POSITIVE VALUES INDICATE THAT MIGRATION IS GREATER THAN EXPECTED BASED ON POPULATION, DISTANCE AND INTERVENING OPPORTUNITY 62 Table 5.3.10 Migration Rates Adjusted for Size of the Labour Force, Distance and Intervening Opportunities, 1976-1981, Service Occupations. FROM/TO NFLD N s. N B . QUEBEC ONT . MAN . SASK. ALTA . B C . NEWFOUNDLAND 0 0 1 44 0 38 -0 95 0 6 1 -0 42 - 0 29 1 29 0 54 NOVA SCOTIA 1 72 0 0 0 27 - 1 27 0 16 - 0 1 1 - 0 18 1 16 0 35 NEW BRUNSWICK I O 87 0 44 0 0 -0 74 -0 46 - 0 31 -0 7 1 0 97 -0 19 QUEBEC - 0 35 0 93 0 46 0 0 0 00 -0 42 - 0 39 1 00 0 44 ONTARIO - 0 42 0 88 O 02 -0 1 1 0 0 0 09 -0 60 0 63 -0 40 MANITOBA 0 52 1 03 0 04 -2 46 -0 95 0 0 -0 56 0 66 0 72 SASKATCHEWAN 0 50 1 53 0 52 - 1 34 - 0 52 - 1 •1 1 0 0 0 38 0 97 ALBERTA 0 37 1 4 1 0 04 - 1 14 0 38 -0 38 - 0 60 0 0 0 62 BRITISH COLUMBIA 0 16 1 61 0 17 -0 85 0 70 0 1 1 - 0 15 - 0 12 0 0 NOTE: POSITIVE VALUES INDICATE THAT MIGRATION IS GREATER THAN EXPECTED BASED ON POPULATION, DISTANCE AND INTERVENING OPPORTUNITY Table 6.3.11 Migration Rates Adjusted for Size of the Labour Force, Distance and Intervening Opportunities, 1976-1981, Agriculture. FROM/TO NFLD N S . N B . QUEBEC ONT . MAN. SASK. ALTA . B C . NEWFOUNDLAND 0 0 2 12 2 26 0 0 2 90 1 74 1 50 1 58 1 32 NOVA SCOTIA 1 86 0 0 1 53 - 0 70 0 28 0 0 - 1 79 0 66 0 94 NEW BRUNSWICK 0 0 1 84 0 0 0 62 0 76 -0 27 -1 01 1 2 1 1 08 QUEBEC 0 0 o 6 1 1 06 0 O 0 67 - 1 34 -2 04 0 21 1 12 ONTARIO - 0 01 1 05 0 32 0 17 0 0 -0 38 -1 93 0 36 0 60 MANITOBA 0 0 - 0 98 O 0 -3 84 - 1 62 0 O - 1 56 - 0 44 1 27 SASKATCHEWAN O 0 0 40 - 0 69 0 0 - 1 80 -2 68 0 0 - 0 98 0 57 ALBERTA 0 0 o 67 - 0 54 -2 82 -0 42 - 1 34 38 0 0 0 84 BRIT ISH COLUMBIA 0 0 2 42 1 54 -0 1 1 0 97 0 02 0 00 0 1 1 0 0 NOTE: POSITIVE VALUES INDICATE THAT MIGRATION IS GREATER THAN EXPECTED BASED ON POPULATION, DISTANCE AND INTERVENING OPPORTUNITY 63 Table 5.3.12 Migration Rates Adjusted for Size of the Labour Force, Distance and Intervening Opportunities, 1976-1981, Primary Occupations. FROM/TO NFLD N . S . N. .B. QUEBEC ONT . MAN. SASK . ALTA . B .C. NEWFOUNDLAND 0 .0 - 1 . 19 -2 . 80 -2 .47 - 1 . 27 1 . 38 -0. .05 2 . 1 1 -0, .31 NOVA SCOTIA -0 . 38 0. .0 -2 . 28 -4 . 26 - 1 .62 -0. . 37 -0. .49 1 . 34 -0, .62 NEW BRUNSWICK - i . .47 - 1 . .70 0. .0 - 1 .52 - 1 .40 0. .08 - 1 . .62 1 .49 -0, . 1 1 QUEBEC - 1 . ,81 - 1 , . 33 0. , 18 0 .0 0. .06 0. ,07 -0. . 28 1 .61 0. .59 ONTARIO -0. 12 0. . 28 -0. 21 0. .03 0, ,0 0. . 78 0. 25 1 . 54 -0, , 33 MANITOBA 0. .83 0. 48 -0. 95 -2 . 51 -0. . 12 0. 0 0. 48 1 .98 0, .79 SASKATCHEWAN 0. 0 0. 0 -0. 77 0. .0 -0. .55 -0. 40 0. 0 1 . 73 0. 74 ALBERTA 0. 54 0. 88 -0. 16 -2 . 91 0. 07 -0. 22 0. 41 0. .0 0. 58 BRITISH COLUMBIA -0. 48 0. 08 -1 . 77 -2 . 86 -0. 02 -0. 78 -0. 79 0, ,00 0. 0 NOTE: POSITIVE VALUES INDICATE THAT MIGRATION IS GREATER THAN EXPECTED BASED ON POPULATION, DISTANCE AND INTERVENING OPPORTUNITY Table 5.3.13 Migration Rates Adjusted for Size of the Labour Force, Distance and Intervening Opportunities, 1976-1981, Processing. FROM/TO NFLD N . S . N .B. QUEBEC ONT . MAN . SASK . ALTA . B ,C. NEWFOUNDLAND 0. .0 0 .31 -0 .92 -0 .96 0 . 14 -o . 13 0 .47 1 . 76 O .51 NOVA SCOTIA 0. . 27 0. ,0 -0 . 27 -2 . 1 1 -0. .40 -o. . 37 0. .0 1 . 10 0. . 15 NEW BRUNSWICK -0. .89 0. . 24 0, .0 -0. .61 -o. .31 0. . 77 0. .0 1 .90 Oi .31 QUEBEC -0. 73 -0. , 55 0, .87 0. .0 -0. . 24 -0. . 75 -0. .86 1 .23 0. .42 ONTARIO 0. 12 0. 67 0. .30 -0. . 17 0. 0 0. .07 -0. . 32 1 . 34 -0, . 27 MANITOBA 0. 79 -0. 12 -0. 23 -3. 05 -0. 63 0. O 0. .48 1 . 75 1. . 19 SASKATCHEWAN 0. 0 0. 0 0. 22 - 1 . .32 -0. 25 -0. 40 0. 0 1 . 81 1. 97 ALBERTA -0. 49 1 . 40 0. 25 - 1 . 38 0. 74 0. 46 0. 27 0 .0 1. 31 BRITISH COLUMBIA - 1 . 13 0. 03 0. 25 -1 . 99 0. 4 1 -0. 06 -0. 24 0. .33 0. 0 NOTE: POSITIVE VALUES INDICATE THAT MIGRATION IS GREATER THAN EXPECTED BASED ON POPULATION. DISTANCE AND INTERVENING OPPORTUNITY 64 Table 5.3.14 Migration Rates Adjusted for Size of the Labour Force, Distance and Intervening Opportunities, 1976-1981, Manufacturing. FROM/TO NFLD N . S . N . B . QUEBEC ONT . MAN. SASK . ALTA . B . c . NEWFOUNDLAND 0 .0 1 .60 0 . 98 0 . 38 1 . 37 - 0 . 37 - 0 .83 2 . 18 1 .21 NOVA SCOTIA 1 . 64 0 .0 1 . 15 - 1 . .42 0. . 13 - 0 . . 4 1 0. .32 1 .93 0 . 54 NEW BRUNSWICK 1 .42 1 . 28 0 .0 - 0 . . 57 - 0 . . 27 - 0 . .21 - 0 . .06 2 . 19 0. . 52 QUEBEC - 0 . .87 0 .06 0 . 56 0, ,0 - 0 . . 38 - 0 . .75 - 0 . . 73 1 .41 0. . 33 ONTARIO - 0 . . 1 1 0. . 54 0. .03 - 0 . .44 0. 0 - 0 . 17 - 0 . .21 1 . .21 - 0 . . 20 MANITOBA 0 . . 70 0. .94 0. . 14 - 1 . 84 - 1 . 36 . 0 . 0 0. 62 1 . .63 1 . .45 SASKATCHEWAN 0 . 0 1. 16 1. .69 - 1 . 38 - 0 . 50 - 0 . 21 0. 0 1 , , 76 1 . 91 ALBERTA 1 . 04 1. 37 0 . 7 1 - 1 . 50 O. 69 0 . 24 0 . 78 0 . .0 1 . 59 BRITISH COLUMBIA - 0 . 05 0 . 82 0 . 02 - 0 . 80 0 . 87 0 . 27 0 . 79 0 . .85 0 . 0 NOTE: POSITIVE VALUES INDICATE THAT MIGRATION IS GREATER THAN EXPECTED BASED ON POPULATION, DISTANCE AND INTERVENING OPPORTUNITY Table 5.3.15 Migration Rates Adjusted for Size of the Labour Force, Distance and Intervening Opportunities, 1976-1981, Construction Trades. FROM/TO NFLD N. . s . N , B . QUEBEC ONT . MAN. SASK. ALTA . B . , C . NEWFOUNDLAND O .0 - 1 . .38 - 1 . 38 - 1 . 37 - 0 . 20 - 1 . 23 - 0 , . 16 1 . 75 - 0 . , 52 NOVA SCOTIA - 0 . . 10 0. O - 1 . .01 - 2 . 68 - 1 .00 - 1 . 40 - 0 . .55 1 .53 - 0 . 27 NEW BRUNSWICK - 0 . , 27 - o . . 72 0. .0 - 1 . 84 - 1 . 1 1 - 1 . 57 - 1 , . 14 ' 1 . 88 - 0 . 33 QUEBEC - 1 . .91 - o . 94 - 0 . . 36 0 .0 - 0 , .84 -2 . 17 - 1 , .00 1 .62 0. 02 ONTARIO - 1 . 03 - o . 35 - 0 . .54 - 0 . . 79 0 .0 - 1 , . 18 - 0 . . 78 1 . 23 - 0 . 45 MANITOBA - 0 . 16 - 0 . 70 - 1 . 6 1 - 3 . .80 - 1 . , 75 0 .0 - 0 . ,00 1 . 26 0 . 83 SASKATCHEWAN 0 . 0 - o . 52 - 1 . 62 -2 . 35 - 1 . .49 - 1 . .85 0 . ,0 1 .04 ' 0 . 99 ALBERTA - 1 . 09 0 . 92 - 1 . 07 -2 . .09 - 0 . , 37 - 0 . , 93 - 0 . 23 0, .0 0 . 72 BRITISH COLUMBIA - 1 . 38 - 0 . 34 - 0 . 59 -2 , ,03 - 0 . 05 - 0 . 99 - 0 . 01 0, . 16 0 . 0 NOTE: POSITIVE VALUES INDICATE THAT MIGRATION IS GREATER THAN EXPECTED BASED ON POPULATION, DISTANCE AND INTERVENING OPPORTUNITY 65 Table 5.3.16 Migration Rates Adjusted for Size of the Labour Force, Distance and Intervening Opportunities, 1976-1981, Transportation. FROM/TO NFLD N.  s. N. . B . QUEBEC ONT . MAN. SASK. ALTA . B .C . NEWFOUNDLAND 0 .0 - 0 . 12 - 0 .58 - 1 .04 -0 .59 - 1 .52 0 .0 1 .02 - 0 . 30 NOVA SCOTIA 1 . . 33 0. .0 -o . 14 -2 . 1 1 - 0 . .•65 - 0 . . 79 - 1 . .62 1 .08 0 .03 NEW BRUNSWICK o! ', 34 0. 12 0. .0 - 1 .84 - 0 . .98 - 1 . .02 - 1 . . 72 1 .30 - 0 .81 QUEBEC 0. . 23 - 0 . 31 -0 .06 0. .0 - 0 . 20 - 0 . .60 - 1 . 28 1 . 38 0 .06 ONTARIO - 0 . 23 0. 37 - 0 . . 28 - 0 . . 32 0. .0 0. 23 - 0 . 45 1 . 24 - 0 . . 56 MANITOBA 0. 45 0. 85 - 0 . .46 - 1 . .67 - 1 . 18 0. .0 - 0 . 12 1 . 36 0. .86 SASKATCHEWAN 0. 0 0 . 55 - 1 . 04 - 1 . .69 - 1 . 05 - 0 . 80 0. 0 1 .00 0. .96 ALBERTA - 0 . 60 0 . 97 - 0 . 44 - 1 . 26 0 . 53 0 . 02 -o. 43 0 . 0 1. .07 BRIT ISH COLUMBIA - 0 . 66 0 . 72 - 0 . 39 - 1 . .30 0. 48 - 0 . 14 - 0 . 22 0. 49 0. 0 NOTE: POSITIVE VALUES INDICATE THAT MIGRATION IS GREATER THAN EXPECTED BASED ON POPULATION, DISTANCE AND INTERVENING OPPORTUNITY Table 5.3.17 Migration Rates Adjusted for Size of the Labour Force, Distance and Intervening Opportunities, 1976-1981, Other Occupations FROM/TO NFLD N . S . N . B . QUEBEC ONT . MAN. SASK. ALTA . B . C . NEWFOUNDLAND 0 .0 0 .46 0 .56 -2 . 57 1 . . 24 - 0 . . 30 -1 . .07 1 . 47 O .58 NOVA SCOTIA 1 . .08 0 0 0 . 7 1 - 1 . 4 1 - 0 . . 15 - 0 . .78 - 0 . .65 1 . 18 - 0 . .03 NEW BRUNSWICK 0. . 88 0 .90 0 .0 -0 .88 - 0 . . 10 - 0 . . 3 1 - 0 . .81 1 .93 0. 30 QUEBEC - 0 . . 3 1 0 . 23 0 . 20 0. .0 - 0 . . 26 - 1 . .08 - 0 . 92 1 .08 0. 10 ONTARIO - 0 . 02 0. . 4 1 0. . 14 - 0 . ,40 0. 0 - 0 . .08 - 0 . 54 0 .90 - 0 . 34 MANITOBA 1. 57 0. .62 0. . 22 -2 . .04 - 0 . .93 0. .0 0. 17 1 . 34 1 . 13 SASKATCHEWAN 0. 0 1 . .00 0. .4 1 - 1 . ,80 - 0 . 10 - 0 . 31 0. 0 0 .90 1 . 19 ALBERTA 0 . 29 0. 98 0. 02 -2 . 44 0. 51 - o . 06 - 0 . 25 0. .0 1 . 01 BRIT ISH COLUMBIA 0 . 54 1 . 01 - 0 . 13 - 1 . 78 0 . 57 0 . 39 0. 17 0. . 19 0 . 0 NOTE: POSITIVE VALUES INDICATE THAT MIGRATION IS GREATER THAN EXPECTED BASED ON POPULATION, DISTANCE AND INTERVENING OPPORTUNITY 66 Tables 5.3.2 to 5.3.17 show migration rates by occupation adjusted for labour force size, distance and intervening opportunities, which are simply the residuals of the multiple regression as given in equation (4.2.2). The question is whether the explanations for the above patterns still hold, or whether the patterns were removed and accounted for by distance and intervening opportunities. The tables show that the removal of the influence of distance and intervening opportunities has resulted in both the emergence of new patterns and the enhancement and/or removal of old ones. Alberta still shows the highest inmigration rates. Nova Scotia now has the second highest rates, followed by British Columbia, which has lost much of its original attractivity. In the Marit imes much of the increased interaction between other Marit ime provinces has been removed (although one can still see a pattern), whereas in some occupations it has been reversed, that is migration is less than expected. Some examples of these negative residuals are primary, processing and construction occupations. For Newfoundland and New Brunswick the residuals are also in most cases positive, i.e., inmigration is greater than expected, for most most occupations, but not quite as pronounced as Nova Scotia. In general the Maritimes do not do as badly as one would have otherwise expected. In fact inmigration rates are generally higher than those of Ontario. Outmigration rates, on the other hand, tend to average out to zero. The fact that the Maritimes performed comparably well can be explained by three factors: (1) much of this inmigration is most likely to be return migration; (2) social security benefits are in some cases probably easier to obtain than in other provinces; and (3), an explanation already mentioned in the discussion of labour force adjusted migration rates, Halifax's role as a growth pole. The tendency found above to migrate to a neighbouring province has now largely been removed. Instead, a new pattern has emerged where migration rates to the eastern neighbours are generally negative, whereas to western neighbours they are usually positive, indicating a "pull" to the west. Quebec, Ontario, Manitoba and Saskatchewan all perform quite poorly in 67 attracting migrants due to other than spatial or labour force related factors compared to the rest of Canada. The above results are again quite different to those of L y c a n (1969). First , the variation in the size of the residual was much lower (the correlation was also greater). Second, no significant east to west movement was found. Quebec and Saskatchewan, however, were also poor attractors of migrants between 1956 and 1961. F r o m an occupational point of view three distinct groups appear when observing the signs of the residuals: (1) occupations that are relatively reluctant to migrate (mostly negative residuals), (2) occupations that show a tendency towards "hypermobility" (mostly positive residuals), and (3) occupations which do not show a tendency in either direction. The first group of occupations is comprised of sales (58% of the residuals are negative), construction (81%) and transportation (60%). Salespeople, although they do travel quite frequently, are not required to move in order to gain access to jobs. The company a salesperson works for can be located anywhere in Canada (or the United States). Companies prefer to assign regions to people that are on their "home turf". Construction- and transport-related occupations are highly unionized, as mentioned earlier, which would partly explain the reluctance to move. Hypermobile occupations are fine and commercial artists (63% of the residuals are positive), service occupations (60%), agriculture (67%) and manufacturing (65%). For artists many jobs are short term contracts, thus often requiring a move when a new contract is won. In the agricultural sector of the economy (this concerns not so much the farmers, but more the agricultural workers) there are constant up- and downswings, and job opportunities are consequently not as stable as in other occupations. Thus agricultural workers have traditionally been nomadic people. In the service related occupations we find the hypermobile "Yuppies". F o r manufacturing it is difficult to find an explanation for the relatively high tendency to migrate, especially considering the heavy unionization in those occupations. Perhaps increased automation and the resulting relative instability of jobs could provide a partial explanation. 68 5 . 4 Wages, Unemployment and Adjusted Migration Rates. The purpose of this section of the chapter is to test the influence of wage and unemployment rates on the residual migration not explained by the size of the labour force and spatial factors. According to the neoclassical hypothesis we should thus find that if the wage rate is higher in the province of destination than it is in the province of origin, and unemployment rates are lower, then the residual should be positive. Consequently, as stated earlier, the partial coefficient should then be positive for wage rates and negative for unemployment rates. Table 5.4.1 shows the results of the multiple regression broken down by province of origin. In general the correlation coefficients are not very high and in most cases quite insignificant. Differences in wage and unemployment rates explained between 0.7%' (New Brunswick and Ontario) and 8.1% (British Columbia) of the variation. Overall (Canada) the r was 0.5%. The partial regression coefficients were in most cases not of the expected sign: for unemployment rates only in three out of nine cases was the sign negative, whereas for wage rates only four out of nine were positive. Intuitively this suggests that wage and unemployment rates do not influence migration rates in the manner they are expected to according to the neoclassical hypothesis. However, before such a judgement can be made, data problems and individual cases (provinces and occupations) have to be considered. The negative partial correlation coefficients for wage rates in the Maritimes indicate that individuals from those provinces tend to go to low wage regions, in this case other Marit ime provinces (although a large portion of the intra-Maritime migration pattern was removed after migration rates were adjusted for spatial factors). For Ontario and Saskatchewan the multiple correlation coefficients (.082 and .100, respectively) were also very low. To explain this phenomenon it is necessary to go back to the labour force and spatially-adjusted migration rates tables. For Saskatchewan the residuals (in terms of outmigration) are generally positive for the Maritimes and Alberta, and negative for the remainder of the 69 Table 5.4.1 Significance of Interregional Differences i n Wage and Unemployment Rates on Residual Migration, by Province of Origin. Province R R 2 Unempl . Wages Newfoundland .137 .019 -.083 -.126 Nova Scotia .122 .015 .079 -.079 New Brunswick .081 .007 -.024 -.081 Quebec .214* .046 .089 .213** Ontario .082 .007 .077 .054 Manitoba * * .247 .061 .053 * * .245 Saskatchewan .100 .010 -.100 -.039 Alberta .102 .010 .055 .100 Brit ish Columbia * * * .285 .081 * * .145 ** -.188 . C a n a d a x .068* .005 * * .066 .036 : significant at 10% * : significant at 5% * * : significant at 1% : does not include Prince Edward Island, Y u k o n and Northwest Territories. 70 country. Thus it is possible to make the same kind of arguments above with respect to the low r squares in the Marit imes (note that the partial correlation coefficient for wage rates is also negative). For migrants from Ontario the residuals seem to be randomly distributed. With the exception of Alberta , Ontarians do not seem to favour any particular destination. British Columbia was the only province of origin where residual migration rates showed a significant correlation with wage and unemployment rates, although in both cases the partials were of the wrong sign. Again , it is necessary to refer to the residual migration tables to provide a partial explanation for this finding. The residuals tended to be positive for Ontario, Nova Scotia, and to certain extent also for New Brunswick (all ranking lower than British Columbia on the wage and unemployment rates scale), and random for other destinations (including Alberta). Return migration could again partially account for the positive residuals for Marit ime destinations and Ontario. Manitoba had the second highest correlation (.247, significant at the 5% level). Residuals tended to be positive for western destinations and negative for eastern destinations. In the provinces to the west of Manitoba wage rates are relatively low (positive partial). For Quebecers the two variables explained the third highest proportion of the variation. While unemployment rates were insignificant, wage rates showed a positive relationship with residual migration (significant at the 5% level). Table 5.4.2 shows the results of the regression of wage and unemployment rates broken down by occupation. Overall the multiple correlation coefficients were slightly higher than obtained in the previous regression, but still insignificant. White collar occupational groups tended to have a slightly higher r squared than the blue collar groups, although the lowest was in managerial and administrative occupations (0.4%). In only one case (teachers) the partial correlation coefficients of unemployment rates was of the expected sign. For wage rates the signs tended to be negative for the white collar groups and positive for the blue collar occupations. A s with the previous regression, the results show that the residual migration pattern does not seem to fit the neoclassical hypothesis. 71 Table 5.4.2 Significance of Interregional Differences in Wage and Unemployment Rates on Residual Migration, by Occupation. Occupation R R 2 Unempl . Wages Managerial .063 .004 .062 -.014 Natura l Sciences .250 .063 .213* .131 Social Sciences .202 .041 .024 -.202* Teachers .252 .064 -.076 -.229* Medicine & Health .272* .074 .211* -.070 Artists .227 .052 .165 * -.204 Clerical Workers .159 .025 .063 -.126 Sales .282* .080 * * .234 -.000 Service .163 .027 .162 .009 Agriculture .146 .021 .024 .136 P r i m a r y .143 .021 .103 .129 Processing .214 .046 .170 .212* Manufacturing .172 .030 .042 .172 Construction .219 .048 .088 .197* Transportation .162 .026 .160 .043 Other Occupations .188 .035 .186 .051 A l l Occupations * .068 .005 * * .066 .036 : significant at 10% : significant at 5% 72 5.5 Net Migration and Changes i n Wage and Unemployment Rates. This final part of the analysis will examine the effects of migration on wage and unemployment rates between 1976 and 1981. The neoclassical hypothesis here is that net inmigration will increase unemployment rates while reducing wage rates, whereas net outmigration will reduce unemployment rates and increase wage rates. It is assumed that these changes will bring the system back to a state of equilibrium. Table 5.5.1 shows the relationship between net migration and wage and unemployment rates broken down by province. While net migration did have an effect on changes in wage and unemployment rates between 1976 and 1981 (although significant only in a few cases), in most cases the direction of the sign was not as expected. The relationships were strongest in Quebec, Alberta and New Brunswick. The reversed signs suggest a process of cumulative causation rather than an equilibrating process, particularly in Alberta, where, as we know, net inmigration was the highest in Canada. Broken down by occupation the relationship becomes clearer (see table 5.5.2). The correlation coefficients are generally higher for wage rates and lower for unemployment rates. In all but one case the signs are reversed. For unemployment rates we find that net migration had a stronger and more significant effect in the upper occupational groups, while for wage rates there was little variation. Only in two cases (fine and commercial artists and other occupations) there was a significantly strong relationship. This latter regression suggests more strongly that net inmigration promotes a. process of cumulative causation. This implies that, rather than increasing unemployment and reducing wage rates, net inmigration causes the opposite effects. Thus the neoclassical hypothesis on the effects of labour migration can not be substantiated. 73 Table 5.5.1 Significance of Net Migration on Changes in Wage and Unemployment Rates, 1976-1981, by Province. Unemployment Wages Province R R 2 R R 2 Newfoundland .076 .006 .414 .171 Nova Scotia .168 .028 .239 .057 New Brunswick * * -.535 .286 .314 .098 Quebec * * .524 .274 .465* .216 Ontario -.153 .023 -.463* .215 Manitoba -.083 .007 .321 .103 Saskatchewan .108 .021 .489* .239 Alberta * * -.498 .248 .408 .167 Brit ish Columbia -.012 .000 -.277 .077 C a n a d a x * * * -.529 .280 . 1 4 0a .019 * : significant at 10% : significant at 5% * * * : significant at 1% x : does not include Prince Edwards Island, Y u k o n and Northwest Territories 74 Table 5.5.2 Significance of Net Migration on Changes i n Wage and Unemployment Rates, 1976-1981, by Occupation. Unemployment Wages Occupation R R 2 R R 2 Managerial * -.666 .443 .046 .002 Natura l Sciences -.572 .328 .273 .074 Social Sciences -.661* .437 .205 .042 Teachers * * -.574 .666 .000 .000 Medicine & Health * * * -.816 .666 .008 .000 Artists * * -.635 .403 * * .672 .451 Clerical Workers -.565 .319 .250 .063 Sales ** -.697 .486 .370 .137 Service -.489 .239 .117 .014 Agriculture -.457 .209 -.071 .005 Pr imary -.546 .298 .144 .021 Processing -.560 .314 .297 .088 Manufacturing -.328 .108 .312 .098 Construction -.538 .290 .013 .000 Transportation -.539 .291 .123 .015 Other Occupations -.513 .263 .599* .358 A l l Occupations * * * -.529 .280 * .140 .019 : significant at 10% * : significant at 5% : significant at 1% 75 5.6 Summary of Findings. Once the effects of the size of the labour force and spatial factors have been removed from the gross migration data, we find that Alberta and Nova Scotia seem to attract the largest proportion of the residual migration. While wage and unemployment rates in general perform poorly in explaining migration rates, it is not possible to conclusively reject the neoclassical hypothesis. The argument might be made that the factors affecting an individual's decision to migrate consist of more than just wage and unemployment rates. They might include climate, political factors, and other such considerations. If this is the case an individual may still maximize their utility even though the destination province has lower wage rates and higher unemployment rates than the province of origin. Nonetheless it still can be argued that the standard notion that wage and unemployment rates are the major determinants of migration patterns is highly questionable, at least in the case of Canada. A stronger case can be made against the neoclassical assumption about the effects of migration. Here we do not have the data problems encountered in the first part of the analysis. Furthermore, in almost every single case the outcome was the exact opposite of the predicted result. Instead of reducing regional disparities (by equalizing wage and unemployment rates), migration seems to have increased them. 76 6. ACCOUNTING FOR FAILURES OF T H E NEOCLASSICAL MIGRATION MODEL. The previous discussion has thus far raised three different kinds of doubts about the neoclassical approach to migration analysis: (1) there is considerable concern about the validity of some of the basic assumptions of neoclassical migration theory; (2) the methodologies employed in neoclassical migration models are questionable; and, (3) the above data analysis, while not disproving the neoclassical hypotheses, has at least raised some serious questions about their validity. This final chapter will try to account for some of the ostensible problems with the neoclassical scheme. Specifically, it will be argued that the major reason why the neoclassical model fails is that the neoclassical assumptions are often too simplistic and do not reflect reality. In doing so we will discuss the maximization principle, temporal considerations, social constraints, labour markets, government intervention and the notion of spatial equilibrium. 6.1 Maximization. This section of the chapter will deal with the maximization principle, more specifically with the associated concepts of utility, the ceteris paribus principle and the assumption of perfect information flows. In chapter 2 we saw that the concept of utility is a key pivot around which much of neoclassical theory revolves. In chapter 3 it was further shown that neoclassical theory contends that individuals' utilities are the reason why people move from one place to another: if for any reason it is possible to increase one's utility by migrating, then such a move will always occur. Having already discussed the theoretical aspects of utility above we now turn to the practical issue of empirically testing the utility-maximization thesis. We have seen that in neoclassical labour migration theory peoples' utilities are frequently measured in terms of wage and unemployment rates. What we have not been able to prove, however, is that these socioeconomic variables are the only significant measures of utility. Such a proof, though, is 77 necessary to demonstrate that the neoclassical model is invalid. For if it cannot be proved that wage and unemployment rates are the only variables affecting utility, then it could be argued by a neoclassicist that people are still maximizing utility, but they are maximizing variables other than wage and unemployment rates, for example, climate, cost of living or quality of life. The problem with this counterargument is that if each person measures his or her utility in terms of different variables, then, in order to test the hypothesis, case histories and psychological profiles of each migrant would be needed. Thus , in light of the absence of sufficient data (i.e. individual case histories) and the fact that the notion of utility is a crucial ingredient in the neoclassical theory on labour migration, the entire hypothesis would be untestable. Thus the crux of the problem is that if neoclassical economic geographers reject the claim that utility can be measured by wages and unemployment, then they are left with a theory that cannot be empirically tested. A further fundamental aspect of neoclassical labour migration theory and the utility maximization principle that underlies it is the so-called "ceteris paribus" principle, or the premise of "all other things being equal". Not only is its existence crucial in that without this condition the neoclassical theory would not hold, but it is also used to show that a neoclassical world is in fact a desirable one. F o r only if all things are equal in all regions is it possible for all individuals to maximize their utilities. If, for example, there were barriers to movement in one region, then individuals living in that region would not be able to maximize their satisfaction by migrating. Therefore, the argument goes, it is desirable to move the real world closer to the neoclassical one so that all individuals are able to maximize. Thus we encounter the term in many neoclassical arguments. Within the context of labour migration, we learn that if in any region wage rates are higher and unemployment rates are lower than in other regions, then, ceteris paribus, it will experience inmigration which in turn through the law of supply and demand will bring wage and unemployment rates back to equilibrium levels. "Ceteris paribus" can in this sense be interpreted as presupposing that all of the neoclassical assumptions (perfectly competitive market, perfect information, maximization of utility, 78 freedom of movement, etc.) apply. If any one of these conditions does not exist, then the theory does not hold. Neoclassicists thus argue that if the results of an empirical test of a neoclassical hypothesis prove to be inconclusive, then initial (real life) conditions may not be the same as the initial assumptions. Therefore, because it is desirable to bring the system back to equilibrium - which would occur only if the initial conditions are equal to the initial assumptions - something is wrong with the system. One should design policies that bring the real world closer to a neoclassical one. Another element of the maximization thesis upon which the neoclassical theory of labour migration is based is the assumption that information flows freely across space. There are no natural or artificial barriers to the flow of information. Inherent in this assumption is that all individuals have equal access and ability to process that information. If they did not, then all individuals would not be able to maximize their utilities. Thus at any given point in time each and every person has perfect knowledge. This has very important implications not only for the efficiency of the labour adjustment process, but also for the resulting tendency for the spatial economic system to move back towards a state of equilibrium. A s soon as an individual receives information that he or she can make him or herself better off by moving to another region (taking into consideration the cost and the friction of distance), then they will do so. This move will then be destined to a region where the net gain in satisfaction will be maximized. If, however, any one of the assumptions about information (flow, access and ability to process) does not hold, then the neoclassical theory will not work because people make wrong decisions. The subject of information has been the subject of much research, notably within the general field of job search theory (for some examples see Pissarides, 1975; Azariadis , 1981; Clark & Whiteman, 1983). Like many other areas of regional analysis and economic geography, research in job search theory has also been dominated by neoclassical economic conceptions of individual behaviour and macro-equilibrium-oriented adjustment (Clark, 1986). 79 This dominant school of thought recognizes the existence of imperfect information and uses it to explain inefficiencies in the allocation of labour (Holt, 1970; Phelps, 1970). Phelps' (1970) "island parable" is a typical example of a neoclassical explanation of this problem. In it the author likens the spatial economic entity to a system of islands between which information flows are costly, i.e., workers must forgo wages by travelling to other islands to learn of the wages there. Assuming a constant labour supply, homogeneous labour with respect to the techniques of production and perfect inter-island competition, then, if for some reason relative wages fall on the local island, the worker will be induced to search for a better wage offer, either locally or on some other island. Because searching costs money (in terms of lost wages), there is a constraint on the number of searches and the geographical pattern of search (Clark, 1986). No worker could afford to visit all islands, which results in few workers reaching peripheral islands. However, as the existence of an inter-island equilibrium depends on the efficiency of workers' searches (which in turn depends on the flow of and efficiency with which workers process the information), a limited number of searches and an inefficient information transfer will lead to an inefficient labour market adjustment and thus a persistent disequilibrium. The neoclassical argument concludes that workers' search efficiency should therefore be improved by training and education, while the flow of information should be improved by some government sponsored information gathering and dissemination agency. Holt (1970), for example, argues that a national computer matching network could greatly improve the efficiency of the labour adjustment process. Clark (1986) maintains that this is in fact the logic behind the Canadian federal labour market information system and the United States job search program. There are many examples within the geographic literature of alternative views on the subject of job search theory and the spatial inefficiency of labour markets. While many economists believe that information is distinctly aspatial (for some examples see Pissarides, 1985 and Evers & V a n der Veen, 1985), geographers have argued that information on job and 80 wage offers is in fact systematically dispersed among regions as a result of the spatial differentiation of the economy (Clark, 1986). Greenwood (1981), for example, argues that empirical evidence suggests that not only are workers' searches subject to a friction factor of distance, as the Phelps island parable suggests, but also that migrants tend to have specific search paths which are directly linked with their origin. This is in accordance with the findings of this thesis. It was shown that there is a wide range among provinces of origin in the sensitivity of migrants to the influence of distance and intervening opportunities. Even after these influences are removed, distinct interaction patterns can still be observed. Recall, for example, the tendency for migrants from a Marit ime province to stay within the Maritimes. One of the reasons for this observation could be that Marit imers limit their area of job search to other Marit ime provinces. The importance of imperfect information and its impact on the spatial economic system is shown by Maier (1985). He recognizes that the standard neoclassical job search and migration models based on the assumption of perfect information are too restrictive to permit insights into the migration decision process. M a n y hypotheses derived from these assumptions in fact contradict empirical observations. Once the assumption of perfect information is dropped the task of modelling job search behaviour becomes more complex. H e goes on to argue that under the assumption of imperfect information, strategies which seem absurd or suboptimal under conditions of perfect information can actually be preferable. He constructs a model which analyses job search strategies based on information about the wage rate distribution. The information is imperfect and accumulated through the search process. Individuals "buy" information about wage offers in other regions. The decision to migrate thus depends on four key variables: each person's knowledge about the wage offer distribution; the cost of searching for the information; the actual cost of the information; and the cost of migrating to another region. Although the model is essentially neoclassical with modified assumptions, Maier argues that it can explain many phenomena used in polarization theory, 81 most notably effects of cumulative causation caused by past migration. This directly contradicts the neoclassical hypothesis on the effects of migration. While at first glance Maier's model seems to offer a relatively elegant solution to the problem of imperfect information, it still does not get around some of the other fundamental problems of neoclassical migration models. In chapter 3 it was mentioned that most contemporary migration models do not consider time as a variable, yet it has been shown that time can be a vital factor (Clark & Bal lard, 1980; a more detailed discussion of temporal considerations will follow in a subsequent section of this chapter). A further drawback of his model is that wage offers constitute the sole source of information on the basis of which a potential migrant will base his or her decision to move. Final ly , Maier offers no empirical evidence to support his contentions. C u r r y (1985) suggests a different (non-neoclassical) approach to handling the problem of information flows in determining migration. He strongly questions the standard neoclassical notions on how the spatial economy is supposed to work. Furthermore, he argues that past attempts at ameliorating spatial inefficiencies in the operation of labour markets have more often than not produced the opposite of the desired results. The reason for this, he argues, is that workers do not exploit opportunities in an optimum manner because they are in many cases ignorant of them. This ignorance can drastically affect the economic landscape. Because information is acquired locally, and there are large regional differences in the quantity and quality of information available (which in turn is a result of the spatial differentiation of the economy), each region and occupation has unique expectations and specific degrees of sensitivity and selectivity with regards to jobs and wage offers. Therefore, because each region will react differently to changes in wages or other variables, central policies cannot work and in fact often produce adverse results. Decentralized policy instruments, specifically geared towards individual regions, would on the other hand serve to decrease inefficiencies in the labour market. 82 A further alternative approach to dealing with information is that of "indeterminate information" (Clark, 1986), as opposed to imperfect information. It is argued that if information is indeed indeterminate, i.e., it is heterogeneous, incomplete, and/or contingent upon other factors, then the likelihood that information will, as the neoclassical theory postulates, lead to an efficient reallocation of labour would appear to be remote. Clark maintains that in the extreme case indeterminate information might mead to a total collapse of the spatial labour market. Clark recognizes, however, that not all circumstances need be characterized by indeterminate information. In some cases well defined, albeit spatially contingent information channels do exist. One example is the rather well developed and very reliable information network between the upper midwest of the United States and California, which is based on previous migrants' experiences and growth trends in the two areas. In most cases, however, reliable information channels do not exist. For example, empirical evidence suggests that as a result of their inability to develop reliable information channels about employment prospects in other parts of the country, many midwest auto workers preferred to remain unemployed during the recession rather than risking their remaining assets (Clark, 1986). This in turn suggests that the influence of information is far more complex than Phelps made it out to be. Clark argues that it is not possible to circumvent the existence of uncertainty by simply putting a price tag on information. Neoclassicists frequently contend that while indeterminacy of information is well within the realm of possibilities, its source is government policy (intervention), rather than problems with the internal logic of the neoclassical model (Clark, 1986). Note that the implication of this, as has previously been argued and will be argued in subsequent sections, is for policy makers to change reality to suit the neoclassical model, as opposed to respecifying the model to accomodate reality. A typical example of this philosophy is the notion of rational expectations based on available information (Lucas, 1981). Lucas argues that while all individuals essentially make correct (rational) decisions, spatial disequilibrium persists because 83 of misinformation about other regions originating from government policy. In addition, local policies such as unemployment benefits and min imum wages prevent the labour allocation process from working efficiently. A final alternative approach to job search theory that is closely linked to the notion of indeterminate information are contract models (Clark, 1986). Within these models information is believed to be inherently indeterminate for two reasons: first, many events can not be anticipated, and second, it is impossible to completely rely on an unanticipated event, no matter how high the likelihood of its occurrence. Consequently there will always be a degree of uncertainty. In order to reduce the level of uncertainty and to protect the worker (and the firm) from unanticipated events, contracts are drawn between employers and employees (Clark, 1986). These can be of an explicit nature (union - labour) or implicit (company policies). Although obviously not every worker has a contract (such as the self-employed), contracts will in the aggregate hinder short term adjustments of labour to changing market conditions because of their long term rigidity. Both workers and management will not immediately renegotiate contracts in response to fluctuations in the space economy, rather they will gather as much information as possible about both inter and intra-regional differences until the contract expires. In summary, utility maximization, the ceteris paribus principle and the neoclassical assumption labour information flows are all conditions that have been highly criticized, for the most part because they are never met in reality. If this is the case (in fact if only one of the conditions does not hold), then the neoclassical hypotheses on labour migration will not hold. 6.2 Temporal Considerations. It has been argued repeatedly that neoclassical migration models largely ignore temporal aspects of the migration process. Adjustments are usually assumed to occur instantaneously. Clark (1982) has argued that the structure of the neoclassical model itself cannot deal with the dynamics of migration, despite the fact that migration is perceived as a 84 flow - a term that in itself implies change over time. Because of the failure to consider temporal factors, neoclassical models thus essentially analyze migration in terms of comparative statics, rather than dynamics. But simply comparing the economy between one point in time to another could lead to the failure to recognize migration cycles and to the over-or underestimation of the sensitivity of migrants to origin and destination characteristics because of the neglect of time lags in the decision-making process. Clark (1982), Clark & Ballard (1979) and Greenwood (1970, 1981) have proposed models that do incorporate time as a variable. In order to illustrate the significance of time we will review some literature that does account for the temporal dimension. Greenwood (1970) analyzed lagged responses in the decision to migrate. He argues that migration from one place to another increases geometrically. The reason for this is that the more people migrate from i to j , the more information is available: friends and family who have moved to j will send back information about the success of their move. Thus the flow of migrants from i to j during a particular time period is not only a function of current socioeconomic indicators, but also related to the cumulative number of migrants that have moved from i to j in the previous periods. Consequently parameter estimates of most socioeconomic variables tend to obscure the true relationship with migration because they also indirectly influence migration through their effect on past migrants. But these past migrants will in turn affect future migration by relating their experiences, for example. This hypothesis was tested on inter-state migration in the United States between 1955 and 1960 and compared to a standard model not considering past migration as an independent variable. The results of the analysis showed that the explanatory power of the model was greatly increased when incorporating time (the average r increased from .77 to .93), while at the same time individual (socioeconomic) variables showed a decline in significance. Greenwood (1981) puts forth a further argument concerning the problems with atemporal approaches to migration modelling, particularly with respect to employment opportunities. The dilemma is that migration must be measured over some finite time interval. 85 During this time interval, however, migration is itself influencing the observed growth of employment opportunities. This could lead to a bias in the parameter estimates of single equation, multiple regression models of migration over one extended time period. Clark (1982) maintains that neoclassical economists present labour migration as a static process. The question for neoclassicists is how to allocate labour between regions given the current labour supply so that spatial equilibrium may be reached. Thus the issue is allocative efficiency from one point in time to the next. The two points in time are analyzed in terms of comparative statics, not by modelling the process of adjustment from one time period to the other (this is one of the reasons that net migration flows are generally preferred to gross flows in neoclassical models; while net migration fits the comparative statics framework, the adjustment properties of gross migration do not (Clark, 1982)). However, while it is possible to evaluate the effects of migration on marginal changes in independent variables, typical neoclassical models employing a comparative statics framework (such as the popular Lowrey model) cannot analyse the temporal sequence of adjustment. Clark & Ballard (1979) have shown that gross flows of both in- and outmigration are indeed quite sensitive to short-run fluctuations in economic factors (including wage and unemployment rates). They conclude that labour' migration, like many other regional processes, is a cyclical phenomenon. The implication of this is that as the economy moves through a sequence of fluctuations, migration patterns will also vary in a similar fashion. Given that these temporal fluctuations can be quite significant, macro-adjustment models using census data compiled at five year intervals will miss the true pattern of migration. 6.3 Social Constraints to Labour Migration. One of the assumptions of the basic neoclassical hypothesis is that people are free to move anywhere and anytime. It has been shown, however, that in reality there are many barriers to migration. In the data analysis section, for example, it was argued that only few people migrated to and from Quebec because of social and cultural (language) barriers. Other 86 impediments are family ties and, although more of a short run hindrance, job contracts (see Clark, 1986). Housing is considered to be a further important barrier to labour mobility (Cullingworth, 1969; Johnson et al, 1975). Although there is a general agreement on the connection between housing and labour mobility, the nature of the relationship has been highly debated. Housing is in itself a complex phenomenon when regional differences, preferences of various social classes and their accessibility to housing finance are considered (Johnson et al, 1975). The white collar occupational groups, for example, will, because of higher wages, have easier access to owned housing and mortgages than the blue collar groups. Thus the availability and type of housing is in turn related to other socioeconomic variables. The most popular view, however, is that workers owning houses show the highest mobility rates and tend to migrate further. Publicly owned and rented housing is generally associated with much lower levels of mobility and shorter moves (Cullingworth, 1969). 6.4 Labour Markets. A further argument for barriers to labour migration is provided by the dual labour market theory (Sloane, 1985). The basic version of the dual labour market theory suggests that there are two distinct sectors: a primary market where "good" jobs and majority workers predominate, and a secondary market where "bad" jobs and minority workers are dominant. While Sloane argues that the existence of these two markets hinders inter-market mobility, it can also be argued that it hinders interregional mobility. In contrast to the primary labour market, wages are generally low, unemployment high, and the often hazardous jobs in the secondary labour market are mostly occupied by minority groups (blacks, hispanics, etc.). If, for example, a country has two regions that traditionally offer most employment opportunities in the secondary labour market (such as the southern United States), then a move to a region other than those two would thus often also involve moving up the occupational hierarchy, i.e., a promotion into the primary labour market. Empirical studies have shown, however, that, because of racial discrimination, lack of education and training and other factors this is 87 extremely difficult (Sloane, 1985). Therefore minority groups and workers in the secondary labour market in general are faced with fewer choices of destinations when deciding to move. M c K a y & Whitelaw's (1977) study of the role of large private and government organizations in generating interregional migration is one example of the dual labour market approach to constrained choices of destinations within the spatial labour market. They point out that there are a multitude of formal and unwritten rules of entry into many occupations imposed by both employers and trade unions. It is argued that occupations in the primary labour market are more mobile due to the fact that an increasingly large proportion of interregional moves are a result of company policy, i.e., a transfer from branch to another (this in many cases also applies to government employees, in particular teachers and health workers). Individuals in the secondary labour market, on the other hand, were found to be considerably less mobile, in part because local trade unions make entry into the job market extremely difficult for newly arrived migrants. 6.5 Government Intervention. A s was shown in chapter 2, neoclassicists generally argue for a "laissez faire" type of approach to public policy. This implies that, because natural forces exist that drive the spatial economic system back to equilibrium, governments should desist from enacting policies that serve to reduce regional disparities. The neoclassical hypothesis on labour migration assumes that a perfectly competitive market exists: there are no outside influences on migration patterns. A number of recent studies have shown, however, that government action such as fiscal policy does in fact have a pronounced effect on migration patterns (Riew, 1975; Kosinski , 1981; Peek & Standing, 1982; Winer & Gauthier, 1982; Clark, 1983). Thus migration is no longer just a function of individuals' choices and motivations, as it is assumed in neoclassical or "human capital" type of models, rather a highly significant proportion of interregional migration is directly and indirectly caused by government policies. 88 One of the best examples of a federal program directly affecting migration is to be found in Canada (Kosinski, 1981). The Department of Employment and Immigration (DEI) administers a number of programs, the most important of which is the Canada Manpower Mobility Program ( C M M P ) . The following grants are made available under this program (Kosinski, 1981): - exploratory grants to search for jobs elsewhere; - relocation grants for workers and their families; - special travel grants to obtain manpower services not offered locally; - travel grants for temporary employment; - travel grants to temporary employment; and - travel grants to attend training courses offered under the Canada training program. Although only some of the grants provide funds for moves across provincial boundaries, their impact is nevertheless quite significant. Table 6.4.1 shows the size of migration flows generated by C M M P grants between 1966 and 1979 and their share of total interprovincial migration in Canada. Whereas overall the impact is not very significant (the range is from 1.6% to 15.3% of total interprovincial migration), its importance does differ quite considerably between provinces of origin (see table 6.4.2). Between 1976 and 1978 34.7% of all outmigration from Newfoundland was generated by C M M P grants, followed by Quebec (20%) and Ontario (19.4%). The lowest share was in Alberta (1.2%) and in Prince Edward Island (1.6%). Data for the destinations of C M M P migrants were not available. Although C M M P is the most important program affecting labour mobility, it is not the only one. Others include the Canada Manpower Consultive Service, the Local Employment Assistance Program, the Canada Works Program and the Opportunities for Youth Program (Kosinski, 1981). Winer & Gauthier (1982) analyzed the indirect impact of the Canadian fiscal structure on interprovincial migration (in terms of both in- and outmigration) between 1968 and 1977). The independent variables were provincial changes in the total sum of unemployment insurance benefits paid (Ul), unconditional grants ( G U : comprised of the sum of equalization 89 Table 6.5.1 Table 3. CMMP generated migration as compared to the total internal migration in Canada, 1966-1976. Financial 1 2 3 4 year Number of Estimate of the total Total interprovincial Proportion of relocation grants flow supported by CMMP^ migration in Canada CMMP migration 1966-67 2,138a 6,400 406,000 1.6 1967-68 5,757 17,300 382,000 4.5 1968-69 6,591 19,800 365,000 5.4 1969-70 7,460 22,400 414,000 5.4 1970-71 6,382 19,100 404,000 4.7 1971-72 9,026 • 27,100 399,000 6.8 1972-73 10,653 32,000 390,000 8.2 1973-74 11,019 33,000 436,000 7.6 1974-75 11,361 34,100 417,000 8.2 1975-76 12,786 38,400 375,000 10.2 1976-77 20,338 61,000 398,000 15.3 1977-78 10,206 . 30,600 405,000 7.6 1978-79 12,302 36,900 n.a. -i Notes: a Including 1 287 grants and 851 loans; ; b Based on an assumption that one worker supports two dependents (see CMCS program). Assumption that ' one-third are singles and two-thirds support three dependents would yield similar results; c Based on assumption that all CMMP supported moves are interprovincial. Table 6.5.2 Table 4. CMMP generated migrations as compared to the total internal migration by provinces, 1976-77 and 1977-78. — - - • • 1 2 3 4 Relocation grants Total interprovincial Percentage of CMMP Provinces authorized in Estimate of outmigration by pro- generated migration 1976-77 & 1977-78 total flows vinces 1976-77 and as compared by provinces by provinces 1977-78 to the total flow Newfoundland • 2,821 8,500 24,500 34.7. Prince Edward Island 42 120 7,700 1.6 Nova Scotia 1,162 3,500 46,400 7.5 New Brunswick 1,378 4,100 36,800 11.1 Quebec 8,425 25,300 126,700 20.0 Ontario 12,790 38,400 197,600 19.4 Manitoba 490 1,500 64,600 2.3 Saskatchewan 354 1,100 49,500 ' 2.2 Alberta 485 1,500 122,300 1.2 British Columbia 2,497 7,500 112,300 6.7 Yukon & the N.W. Territories 100 300 13,300 2.3 Total 30,544 91,600 802,800 11.4 Notes a Based on an assumption that one worker supports two dependents; b Based on an assumption that all CMMP supported moves are interprovincial Source: Kosinski (1981) 90 payments, subsidies, adjustment grants, and revenue guaranteed by the federal government to the provinces) and natural resource revenues (NRR). A l l dependent variables were measured in terms of changes from 1971 to 1977. The model was tested on high, middle and low income groups of individuals. It was found that the fiscal structure indeed did have a significant impact on both in- and outmigration, although it varied systematically with the geographical composition of the migration flow and was generally more significant (as is intuitively to be expected) in the lower income class. The fiscal variables performed considerably better on outmigration from the Atlantic provinces and on inmigration to Alberta and Brit ish Columbia. Recall that in the data analysis (chapter 5) it was found that Marit imers tended to either migrate to another Atlantic province or move to Alberta or Brit ish Columbia. Furthermore, Alberta was by far the greatest pole of attraction for migrants from all Canada. Compare these findings to table 6.4.3, which shows Winer & Gauthier's (1982) results for outmigration from the Atlantic provinces in the lower income class. Changes in the total sum of unemployment insurenace benefits paid out had a positive influence on intra-Maritime moves, whereas they tended to be negative for other destinations (with only few exceptions). For example, outmigration from Newfoundland to New Brunswick was 52.9% higher than it would have been in the absence of any changes to the unemployment insurence system. Unconditional grants, on the other hand, tended to have a negative impact on outmigration, i.e., induce potential movers to stay, although the impact was generally not as pronounced as for unemployment insurance benefits. Changes in natural resource revenues had virtually no impact on outmigration save for the destinations of Alberta, Saskatchewan and British Columbia. For all three Atlantic provinces migration to Alberta was about 63% higher than it would have been had Albertan natural resource revenues remained at the 1971 level. This variables also had the most important impact on inmigration from other, non-Maritime provinces, for both Alberta and Brit ish Columbia. These results very concur with our findings in chapter 5 in that they in many ways fit the description of the residual migration patterns shown in tables 5.3.2 to 5.3.17. 91 Table 6.5.3 The Effects of Selected Changes in Fiscal Structure on Out-Migration Rates From the Atlantic Provinces Including Intra-Atlantic Moves, Lower Income Class, 1968-1977. Origin Destination U l G U N R R Newfoundland N o v a Scotia 0.385 0.095 0.000 New Brunswick 0.529 0.041 0.000 Ontario 0.021 -0.045 0.000 Manitoba -0.091 0.109 0.000 Alberta -0.131 -0.045 0.632 British Columbia 0.116 -0.045 0.106 Nova Scotia Newfoundland 0.438 -0.083 0.000 New Brunswick 0.555 -0.044 0.000 Quebec 0.459 -0.063 0.000 Ontario 0.038 -0.125 0.000 Manitoba -0.076 0.017 0.000 Saskatchewan -0.189 -0.228 0.592 Alberta -0.116 -0.125 0.631 British Columbia 0.134 -0.125 0.106 New Brunswick Newfoundland 0.325 -0.037 0.000 Nova Scotia 0.297 0.054 0.000 Quebec 0.344 -0.017 0.000 Ontario -0.043 -0.081 0.000 Manitoba -0.149 0.068 0.000 Saskatchewan -0.253 -0.189 0.591 Alberta -0.186 -0.081 0.630 British Columbia 0.045 -0.081 0.105 Source: Winer & Gauthier (1982) 92 The above empirical evidence directly contradicts the "conservative" (neoclassical) contention that interregional migration patterns are dominated by long run trends which cannot be directly manipulated by government policy (Clark, 1983). Government action, by providing for the existence of the neoclassical assumptions, can only indirectly influence migration. A s we have seen, however, government policy indeed can have quite a profound direct influence on migration, not only in the long run, but also in the short run. In the long run changes to the social welfare system, such as regional increases or decreases in the amount of unemployment insurance benefits paid out, will induce or inhibit migration. O n the other hand, manpower travel grants can induce a worker to search for a job in another region in the short run, which he or she could not afford to do in the absence of such policy. 6.6 Spatial Equilibrium. The neoclassical notion that the pr imary effect of labour migration is to lead the spatial economic system back to a state of equilibrium has been strongly questioned, even by some neoclassical theorists (for an example see Maier , 1985). The consensus seems to be that rather than leading to equilibrium, labour migration has a cumulative causatory effect. However, there are also many other aspects of the consequences of migration not normalty considered by neoclassicists. This section of the chapter will concentrate first on alternative theories advocating cumulative causation, and then proceed to look at other alternative notions about the effects of migration. The hypothesis that interregional migration promotes polarization, or cumulative causation, was put forward as early as 1957 by M y r d a l . His argument was that since migration is a selective phenomenon, it is usually the most productive and educated people that are attracted away from declining regions. Thus , because the inflow of skilled workers increases the attractivity of growth regions and the outflow of skilled workers from the poorer regions decreases their attractivity, cumulative causation will result. A similar argument is made by Webber (1972), when he suggests that where agglomeration economies are strong, 93 skilled labour is more likely to be attracted to growth centres because of the wider range in wage differentials that are likely to exist there. Kaldor (1970) maintained that while the neoclassical model predicts equilibrium in the long run, in reality regional economic disparities are generally reinforced and that divergence rather than convergence is the rule. His argument was later extended and interpreted as the "Verdoorn effect" (Clark, 1983). The Verdoorn effect basically implies that regions are able to maintain their dominant positions through increasing returns to scale. Given that there are increasing returns to scale in high wage regions, inmigration from low wage areas reinforces the dominance of those regions. Clark & Ballard (1981) tested the neoclassical assumptions about the effects of migration on United States data in a short term time-series model, taking into consideration both the demand and the supply of labour. It was found that the effect of migration on wages was mixed: while in some cases it increased them, in others it caused them to decline or had no effect at all. While it is commonly thought that migration in one way or another affect wage rates, it is frequently not realized that migration also results in redistributions of income, both in the region of origin and the region of destination (Greenwood, 1981). This implies that in both areas some individuals are made better off, while others find themselves in a worse situation. F o r example, in a depressed region outmigration in the short run will result in less competition for the few job opportunities that do exist, so that people who would otherwise be unemployed now earn those wages that the migrants formerly earned. O n the other hand, older workers in the receiving region will find it more difficult to gain employment as young immigrants enter the local job market. Also, if outmigration leads to a significant decline in the population base of a depressed region, then the average tax burden of the remaining population could increase or, because of the smaller tax base and the declining economy of scale, the government might have to practice constraint and cut back on existing programs, thereby reducing overall social welfare. 94 F r o m the above discussion it can be seen that many authors believe, and furthermore back up such beliefs with empirical evidence, that the standard neoclassical assumptions and determinants of labour migration do not hold. However, it cannot be said that migrants, particularly those from depressed regions, do not react to economic opportunities. Rather, the way in which they react to those variables is different: there is little evidence that migrants maximize their utilities in terms of economic variables, they rarely, if ever, have equal and perfect information about them, and they do not always have the freedom to move to the destination of their choice. 95 7.SUMMARY AND CONCLUSION. The task of this thesis was to empirically test the neoclassical theory of labour migration in light of Canada's experience during the period 1976-1981. In carrying out this task, it was argued in chapter 2 that much of contemporary economic geography stems from neoclassical economic theory, and, specifically the concepts of utility and resource allocation. The six major characteristics of neoclassical theory were then outlined and their applications illustrated using some examples of geographic models. Chapter 3 linked the broader neoclassical theory to labour migration. It was shown how neoclassical principles and assumptions are applied in formulating the two standard neoclassical migration models: the micro- and macro-adjustment approaches. Furthermore, it was argued that two central hypotheses stem from these two neoclassical approaches: first, that labour will flow from regions of low wages and high unemployment to areas of high wages and low unemployment; and second, that the effect of this migration will be to reduce interregional differences in wage and unemployment rates. Final ly , it was argued that despite a growing body of literature critical of the two approaches and neoclassical assumptions and hypotheses, neoclassical migration modelling is still very much alive. Chapter 4 presented the data and methodology employed in testing the two neoclassical hypotheses. Chapter 5 interpreted the results of the analysis arguing that Canadian patterns of labour migration between 1976 and 1981, as well as provincial changes in wage and unemployment rates, do not fit either one of the neoclassical hypotheses. Specifically, there is no uniformity in the direction of the flow of migrants in relation to interprovincial differences in wage and unemployment rates. Furthermore, rather than reducing regional differences in those variables, labour migration seems to have increased them. While the findings of this thesis do not entirely justify a rejection of the hypotheses (in part because of data problems), they do raise doubts as to their validity. Final ly , chapter 6 attempted to account for the failures of the neoclassical model. It was argued that the main reason for its failure are the overly simplistic and unrealistic 96 assumptions of neoclassicism. The real world is far more complex than the neoclassical one. All other things are not always equal. An additional oversimplifying assumption is that of the utility-maximizing individual and, by extension, the argument that all individuals behave and act in the same way. It is usually recognized by neoclassicists themselves that wage and unemployment rates are not the only variables that people maximize. Even so, it must be strongly questioned whether it is at all possible to measure utility. If this is not the case, then, because they are based on the assumption of the utility-maximizing individual, it is also impossible to empirically test the neoclassical hypotheses about labour migration. It is perhaps time to rethink migration theory. Although much work has been done on individual aspects of the migration process, a comprehensive model - one that is flexible, less rigid, more realistic in its assumptions and that encompasses the entire migration process -has yet to be formulated. Thus, future research should focus more closely on the causes and effects of migration. While it has been established that migrants are sensitive to regional variations in socioeconomic variables and that migration does have an effect on the spatial economy, it is still unclear how the actual process works. Before a comprehensive alternative theory of labour migration can be put forward, it is thus necessary to learn more about individual decision-making behaviour and the dynamics of the adjustment of socioeconomic variables to migration (and vice versa). Clearly, my thesis has not provided that comprehensive alternative theory. Nevertheless, it is hoped that it has made some contribution by raising questions about the orthodox approach to labour migration. For in order to get the right new theory, one must first know what is wrong with the old one. 97 8. BIBLIOGRAPHY. Alexandersson, G . & Norstroem, G . (1963): "World Shipping), John Wiley & Sons, New Y o r k Alonso, W . (1964): "Location and L a n d Use", H a r v a r d University Press, Cambridge (1970): "Equilibrium of the Household", in Dean, R . D . , Leahy , W . H . & McKee , D . L . (eds): Spatial Economic Theory, The Free Press, N e w Y o r k Azariadis, C . (1981): "Implicit Contracts and Related Topics: A Survey", in Hornstein, Z . , Grice, J . & Webb, A . (eds): The Economics of Labour Markets , H M S O , London Ballard, K . P . & Clark, G . L . (1981): "The S h o r t g u n Dynamics of Inter-State Migration: A Space-Time Economic Adjustment Model of In-Migration to Fast Growing States", Regional Studies, 15:213-228 Barnes, T . J . (1984): "Theories of Agricultural Rent within the Surplus Approach", International Regional Science Review, 9:125-140 (1985): "Theories of Interregional Trade and Theories of Value", Environment & Planning A , 17:729-746 & Sheppard, E . (1984): "Technical Choice and Reswitching in Space Economies", Regional Science and U r b a n Economics, 14:345-362 Blaug, M . (1973): "Was There a Marginal Revolution?", in Black, R . D . , Coats, A . W . & Goodwin, C . D . W . (eds): The Marginal Revolution in Economics, Duke University Press, D u r h a m Bodenhofer, H . J . (1967): "The Mobility of Labour and the Theory of H u m a n Capital", Journal of H u m a n Resources, 2:431-439 Brennan, M . J . (1965): " A More General Theory of Resource Migration", in Brennan, M . J . (ed): Patterns of Resource Behaviour, Brown University Press, Providence Carline, D . , Pissarides, C . A . , Siebert, W . S . & Sloane, P . J . (eds) (1985): "Surveys in Economics: Labour Economics", Longman, London Cebula, R . J . (1979): "The Determinants of H u m a n Migration", Lexington, M A 98 & Vedder, R . K . (1969): " A Note on Migration, Economic Opportunity, and the,Quality of Life", Journal of Regional Science, 13:205-211 Clark, G . L . (1982): "Dynamics of Interstate Labour Migration", Annals of the  American Association of Geographers, 72:297-313 (1983): "Interregional Migration, National Policy, and Social Justice", Rowman & Allanheld, Totowa & Bal lard, K . P . (1979): "Modelling Out-Migration from Depressed Regions: The Significance of Origin and Destination Characteristics", Discussion Paper D79-17, Dept. of City & Regional Planning, H a r v a r d University, Cambridge & Bal lard, K . P . (1980): "The Demand and Supply of Labour and Interstate Relative Wages: A n Empir ica l Analysis", Economic Geography, 56:95-112 Clark, G . L . & Gertler, M . (1983): "Migration and Capital", Annals of the American  Association of Geographers, 73:18-34 Clark, G . L . & Whiteman, J . (1983): "Why Poor People Don't Move: Job Search Behaviour and Disequilibrium Amongst Local Labour Markets", Environment & Planning A , 15:85-104 Courchene, T . J . (1970): "Interprovincial Migration and Economic Adjustment", Canadian Journal of Economics, 3:550-576 Cullingworth, J . B . (1969): "Housing and Labour Mobility", O E C D , Paris C u r r y , L . (1972): " A Spatial Analysis of Gravi ty Flows", Regional Studies, 6:131-147 (1985): "Inefficiencies in the Geographical Operation of Labour Markets", Regional Studies, 19:203-215 Deane, P. (1978): "The Evolution of Economic Ideas", Modern Canbridge Economics Series, Cambridge University Press, Cambridge Dicken, S . N . (1955): "Economic Geography", Prentice-Hall, New Jersey Diehl, W . D . (1966): " F a r m - N o n - F a r m Migration in the South-East: A Cost-Returns Analysis", Journal of F a r m Economics, 48:1-11 99 Dobb, M . (1973): "Theories of Value and Distribution Since A d a m Smith", Cambridge University Press, Cambridge Evers , G . H . M . & V a n der Veen, A . (1985): " A Simultaneous Non-Linear Model for Labour Migration and Commuting", Regional Studies, 19:217-229 Fields, G .S . (1976): "Labour Force Migration, Unemployment and Job Turnover", Review of Economics and Statistics, 58:407-415 (1978): "Place-to-Place Migration: Some New Evidence", Review of Economics and Statistics, 60:21-32 Foot, D . & Milne, J . (1982): "Net Migration Estimation in an Extended, Multiregional Gravi ty Model", Institute for Policy Analysis , Toronto Gordus, P . G . , Jarley, P. and Ferman, L . A . (1981): "Plant Closings and Economic Dislocation", W . E . Upjohn Institute for Employment Research, Kalamazoo Grant, K . E . & Vanderkamp, J . (1976): "The Economic Causes and Effects of Migration: Canada, 1965-1971", Economic Coucil of Canada, Ottawa Greenhut, M . L . (1956): "Plant Location in Theory and Practice", University of North Carol ina Press, Chapel Hi l l Greenwood, M . J . (1970): "Lagged Response in the Decision to Migrate", Journal of  Regional Science, 10:375-384 (1975): "Research on Internal Migration in the United States: A Survey", Journal of Economic Literature, 8:397-433 Harvey , D . W . (1982): "Limits to Capital", Chicago University Press, Chicago (1985): "The Urbanization of Capital", John Hopkins University Press, Baltimore Henderson, J . M . (1958): "The Efficiency of the Coal Industry: A n Application of Linear Programming", H a r v a r d University Press, Canbridge Hicks, J . R . (1932): "The Theory of Wages", MacMi l lan , London 100 Holland, S. (1976): "Capital vs. the Regions", M a c M i l l a n Press L t d . , New York Hollis, M . & Nell , E . J . (1975): "Rational Economic M a n " , Cambridge University Press, Cambridge Holt, C . C . (1979): "How C a n the Phillips Curve be Moved to Reduce Both Inflation and Unemployment?", in Phelps, E . (ed): Microeconomic Foundations of Employment and Inflation  Theory, W . W . Norton, New York Hoover, E . M . (1978): "The Partial Equil ibrium Approach", in Deane, R . D . , Leahy , W . H . & McKee , D . L . (eds): Spatial Economic Theory, The Free Press, New Y o r k Howey, R .S . (1973): "The Origins of Marginal ism", in Black, R . D . , Coats, A . W . & Goodwin, C . D . W . (eds): The Margina l Revolution in Economics, Duke University Press, D u r h a m Isard, W . (1956): "Location and Space Economy", John Wiley & Sons, New Y o r k (1960): "Methods of Regional Analysis: A n Introduction to Regional Science", M I T Press, Cambridge & Ostroff, D . J . (1960): "General Interregional Equilibrium", Journal of Regional Science, 2:67-74 Jevons, W . S . (1871): "The Theory of Political Economy", MacMi l lan , London Johnson, J . H . , Salt, J . & Wood, P . A . (1975): "Housing and Migration of Labour in England and Wales", Saxon House, Farnborough Johnston, R . J . (1973): "On Friction of Distance and Regression Coefficients", A r e a , 5:187-191 Kaldor, N . (1970): "The Case for Regional Policies", Scottish Journal of Political  Economy K i n g , L . J . (1979): "On Neoclassicism in Economic Geography", in Hamel in , L . - E . & Beauregard, L . (eds): Retrospective 1951-1976, Canadian Association of Geographers, Montreal 101 Kosinski , L . A . (1981): "Federal Programs Directly Affecting Migration in Canada", in Webb, J . A . , Naukkarinnen, A . & Kosinski, L . A . (eds): Policies of Population Redistribution, Geographical Society of Northern Finland, Oulou Kottis, A . (1972): "Mobility and H u m a n Capital Theory: The Education, Age, Race and Income Characteristics of Migrants", Annals of Regional Science, 6:41-61 Kriesberg, E . M . & Vining, D .R . (1978): "The Contribution of Out-Migration to Changes in Net Migration: A Time-Series Confirmation of Beale's Cross-Sectional Results", Annals of  Regional Science, 12:1-11 Lansing, J . B . & Mueller, W . (1967): "The Geographic Mobility of Labour", Institute for Social Research, A n n Arbor Lefeber, L . (1958): "Allocation in Space: Production, Transportation and Industrial Location", North-Holland, Amsterdam Loesch, A . (1954): "The Economics of Location", Yale University Press, New H a v e n Lowrey, I.S. (1966): "Migration and Metropolitan Growth: Two Analytical MOdels", Chandler, San Francisco Lucas , R. (1981): "Studies in Business-Cycle Theory", M I T Press, Cambridge L y c a n , R. (1969): "Interprovincial Migration in Canada: The Role of Spatial and Economic Factors", Canadian Geographer, 13:237-254 M c K a y , J . & Whitelaw, J . S . (1977): "The Role of Large Private and Government Organizations in Generating Flows of Interregional Migrants: The Case of Austral ia", Economic Geography, 53:28-44 Maier , G . (1985): "Cumulative Causation and Selectivity in Labour Market Oriented Migration Caused by Imperfect Information", Regional Stidies, 19:231-241 Massey, D . B . (1973): "Towards a Critique of Industrial Location Theory", Antipode, 5, #3:33-39 Menger, C . (1871): "Grundsaetze der Volkswirtschaftslehre", Erster allgemeiner Tei l , Braumueller, Wien 102 Milne, W . J . (1981): Migration in an Interregional Macroeconometric Model of the United States: Wil l Net Outmigration from the Northeast Continue?", International Regional  Science Review, 6:71-83 Moses, L . N . (1978): "The General Equil ibrium Approach", in Deane, R . D . , Leahy , W . H . & McKee , D . L . (eds): Spatial Economic Theory, The Free Press, New Y o r k M y r d a l , G . (1953): "The Political Element in the Development of Economic Theory", Routledge & Kegan Paul , London Nelson, P. (1965): "Migration, Real Income, and Information", Journal of Regional  Science, 2:43-74 Olsson, P. (1965): "Distance and H u m a n Interaction: A Review and a Bibliography", Philadelphia (1970): "Explanation, Prediction, and Meaning Variance: A n Assessment of Distance Interaction Models", Economic Geography, 46:223-231 Peek, P. & Standing, G . (1982): "State Policies and Migration", Croom Helm, London Pred, A . R . (1967): "Behaviour and Location: Foundations for a Geographic and Dynamic Location Theory", L u n d Studies in Geography, Series b: H u m a n Geography, #27, L u n d Pissarides, C . (1975): "Job Search and the Functioning of Labour Markets", in Carline, D . , Pissarides, C , Siebert, W . S . & Sloane, P . J . (eds): Surveys in Economics: Labour  Economics, Longman Group, London Plane, D . A . (1981): "Estimation of Place-to-Place Migration Flows from Net Migration Totals: A M i n i m u m Information Approach", International Regional Science Review, 6:33-51 , Rogerson, P. & Rosen, A . (1984): "The Cross-Regional Variat ion of In-Migration and Out-Migration", Geographical Analysis , 16:162-175 Rabianski, J . S . (1970): "Real Earnings and the Present Value of Future Earnings in a Theory of H u m a n Migration", unpublished P h . D . Thesis, University of Illinois at Urbana-Champaign Renshaw, V . (1970): "The Role of Migration in Labour Market Adjustment", Ph.d . dissertation, M I T , Cambridge 103 Richardson, H . W . (1978):"Regional and U r b a n Economics", Penguin Books, New York Riew, J . (1973): "Migration and Public Policy", Journal of Regional Science, 13:65-76 Robinson, J . (1962): "Economic Philosophy", Penguin Books, London (1973): "Collected Economic Papers", Vol . I V , Basil Blackwell , Oxford Roseman, C . C . (1977): "Changing Migration Patterns Within the United States", Resource Paper #77-2, Association of American Geographers, Washington Rowthorn, B . (1974): "Neo-Classicism, Neo-Ricardianism, and Marx i sm", New Left  Review, 86:63-87 Sayer, A . (1976): " A Critique of U r b a n Modelling: F r o m Regional Science to U r b a n and Regional Political Economy", Progress in Planning, 6:187-254 Schwartz, A . (1976): "Migration, Age and Education", Journal of Political Economy, 701-719 Shaw, R . P . (1975): "Migration Theory and Fact", Bibliography Series #5, Regional Science Research Institute, Philadelphia Simmons, J . W . (1982): "The Stability of Migration Patterns: Canada , 1966-1971 and 1971-1976", U r b a n Geography, 3:166-178 Sjaastad, L . (1962): "The Costs and Returns of H u m a n Migration", Journal of Political  Economy, 70:80-93 Sloane, P . J . (19850: "Discrimination in the Labour Market", in Carline, D . , Pissarides, C . A . , Siebert, W . S . & Sloane, P . J . (eds): Labour Economics, Longman, New York Stock, R. (1981): "Monitoring Migration in the Prairie Provinces", Canadian Plains Report #5, Canadian Plains Research Centre, University of Regina, Regina Stone, L . O . (1978): "The Frequency of Geographic Mobility in the Population of Canada", Statistics Canada, Ottawa 104 (1979): "Occupational Composition of Canadian Migration", Statistics Canada, Ottawa Stouffer, S .A . (1960): "Intervening Opportunity and Competing Migrants", Journal of  Regional Science, 1:1-26 (1962): "Social Research to Test New Ideas", New York Streissler, E . (1973): "To What Extent Was the Austrian School Marginalist?", in Black, R . D . , Coats, A . W . & Goodwin, C . D . W . (eds): The Marginal Revolution in Economics, Duke University Press, D u r h a m Vanderkamp, J . (1968): "Interregional Mobility in Canada: A Study of the Time Pattern of Migration", Canadian Journal of Economics, 1:595-608 (1976): "The Role of Population Size in Migration Studies", Canadian Journal of Economics, 9:508-516 Walras , L . E . M . (1874): "Elements d'economie politique pure", Corbaz, Lausanne Walsh , W . & G r a m , H . (1980): "Classical and Neoclassical Theories of General Equil ibrium", Oxford Webber, M . J . (1972): "Impact of Uncertainty of Location", M I T Press, Cambridge Weber, A . (1909): "Ueber den Standort der Industrien", Tuebingen; Engl ish Translation by C a r l J . Friedrich (1929): "Alfred Weber's Theory of the Location of Industries", Chicago Winer, S . L . & Gauthier, D . (1982): "Internal Migration and Fiscal Structure", Economic Council of Canada, Ottawa Young, J . W . (1984): "Recent Changes in Interprovincial Migration Within Canada", unpublished paper, Concordia University, Montreal (1985): "Labour Market Structures and Occupational Differentials in Mobility of Canadian Male Workers", unpublished paper, Concordia University, Montreal 105 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0097140/manifest

Comment

Related Items