Hostname: page-component-cd9895bd7-dzt6s Total loading time: 0 Render date: 2024-12-22T23:11:46.515Z Has data issue: false hasContentIssue false
  • English
  • Français

Housing Demand in Canada, 1947 to 1962: Some Preliminary Experimentation*

Published online by Cambridge University Press:  07 November 2014

Ernest H. Oksanen
Ernest H. Oksanen*
Affiliation:
McMaster University
Get access

Extract

The paucity of econometric analyses of the demand for consumer durables stands in sharp contrast to the number of studies involving the demand for non-durables, and for agricultural commodities in particular. The neglect of durables undoubtedly reflects in part the greater difficulty of obtaining the necessary data, especially data on stocks, and in part the heterogeneity of most durables, such as houses or automobiles. Econometric analysis in the post-war period has also been centred around the construction of large-scale macro-economic models. Until rather recently, however, such models were relatively highly aggregated, and seldom involved examination of particular durables markets by themselves. Moreover, where an equation for the housing market has been included, it has often been a hybrid of supply and demand forces, intended to serve in a “forecasting” role, presumably in the hope that appreciable structural change would not occur during the forecast period.

The present study will, it is hoped, cast some light on the controversy generated in the United States over the size of demand elasticities for housing with respect to income. The results of the econometric techniques employed here may also assist in evaluating the appropriateness of alternative estimating techniques for the study of demand for durables.

An econometric analysis of the demand for housing immediately raises two questions. First, is it a flow or a stock demand that is to be analysed? Here it is the demand for a long-run equilibrium stock, the “desired stock” of housing, that is the focus of concern.

La demande d’habitation au canada, 1947–62: une experience preliminaire

La Demande D’Habitation au Canada, 1947–62: Une Experience Preliminaire

Dans le domaine de la demande pour les biens durables au Canada, très peu de travail a été accompli. Pour les fins de cette étude, on a construit un modèle de stock et de flux qui est défini par rapport à l'unité familiale et on a utilisé la méthode des moindres carrés ordinaires pour l'estimation des élasticités de demande pour l'habitation. Au lieu de choisir la série d'investissement brut en dollars constants des Comptes Nationaux, on a établi une nouvelle série de l'habitation qui est basée sur un indice des coûts au pied carré (à l'exclusion du coût des terrains) des maisons unifamiliales financées sous l'empire de la Loi Nationale sur l'Habitation. Cet indice a également servi à construire en dollars constants une série de stock en début d'année (à l'aide de chiffres-repères du BFS) et à définir la variable des prix relatifs.

Les résultats des tests statistiques sont plus ou moins clairs. Les variables revenu et stock ont eu tendance à établir de fortes associations, tandis que le taux d'intérêt et le prix ont conduit à de faibles associations. Le mauvais signe est apparu pour le prix quoique la variable n'était pas statistiquement significative. Les élasticités au revenu de la demande pour le stock d'habitations ont eu tendance à être notablement inférieures à l'unité (environ 0.5 et même moins), tandis que les élasticités de la demande pour le flux d'habitations se sont avérées supérieures à l'unité par une marge considérable. Et la demande pour le stock et la demande pour le flux d'habitations furent inélastiques aux changements du taux d'intérêt et des différentiels de taux d'intérêt qui ont été utilisés comme variables dans l'étude.

Les études à venir seront consacrées à une tentative pour imaginer des variables de crédit qui tiennent compte des diverses échéances hypothécaires. On aura recours à l'estimation simultanée.

Type
Research Article
Information
Canadian Journal of Economics and Political Science/Revue canadienne de economiques et science politique , Volume 32 , Issue 3 , August 1966 , pp. 302 - 318
Copyright
Copyright © Canadian Political Science Association 1966

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

Footnotes

*

This article is based upon the author's doctoral thesis at Queen's University, and the help and guidance of the supervisors, Professors T. M. Brown and D. W. Slater, is gratefully acknowledged, as is the helpful advice given by Mr. Frank Denton of DBS and Mr. A. Goracz of CMHC. With respect to this paper, the author would like to express his thanks to Professors T. M. Brown and D. C. Smith, and to Dr. George Post of the Bank of Canada, for their helpful comments. Mr. Logie Macdonnell of Royal Military College developed the computer program used in carrying out the statistical estimation.

References

1 Harberger, A., ed., The Demand for Durable Goods (Chicago, 1960), 9.Google Scholar

2 Stone, J. R. and Rowe, D. A., “The Market Demand for Durable Goods,” Econometrica, XXV (1957), 423–43.CrossRefGoogle Scholar

3 Chow, G., Statistical Demand for Automobiles and Their Use for Forecasting,” in Harberger, , ed., The Demand for Durable Goods, 149–78.Google Scholar

4 R. F. Muth, “The Demand for Non-Farm Housing,” in ibid., 29–96.

5 This paper will not undertake to present the underlying utility analysis which can be developed to encompass the static stock-flow model or to consider dynamic extensions of this model. This analysis is treated, for simple cases, in Bushaw, D. W. and Clower, R. W., Introduction to Mathematical Economics (Homewood, Ill., 1957), 128–39.Google Scholar For the static case they show that it is a simple matter to extend the (ordinal) utility function so as to include desired stock holdings as well as desired consumption of the various commodities. With an extension of the budget constraint to encompass the fact that commodities are not only desired for current use, but also in order to add to existing stocks, constrained maximization leads to desired stock functions. The dynamic extensions are more tortuous, partly because it is no longer clear precisely how the utility maximization hypotheses should be specified. For example, is it satisfactory to assume that the consumer attempts to maximize utility as rapidly as possible? The authors analyse the implications of such an assumption, and show that these accord with those of static theory, but they question the validity of such a behavioural assumption (p. 139).

6 “The Demand for Non-Farm Housing,” 40.

7 Estimation of relationships not reported here was carried out for the period 1926 to 1962, the years 1942 to 1946 excluded, but household data were not avauable for the earlier period. The use of family rather than household data for the post-war period thus assured consistency with the long period estimation.

8 See, for example, Duesenberry, J. S., Business Cycles and Economic Growth (New York, 1958), 138–9Google Scholar, for a list of variables which he views as determinants of housing demand. While Duesenberry regards the number of families as a determinant of stock demand, he seems to view changes in this variable as relevant to flow demand (263).

9 Muth, , “The Demand for Non-Farm Housing,” 95–6.Google Scholar

10 A Priori Information and Time Series Analysis (Amsterdam, 1962), 21–47.Google Scholar

11 Ibid., 29.

12 Ibid., 44.

13 Ibid., 28–9.

14 Ibid., 44.

15 See, for example, Duesenberry, , Business Cycles and Economic Growth, 139–40.Google Scholar

16 Guttentag, J. M., “The Short Cycle in Residential Construction 1946–59,” American Economic Review, LI (1961), 275–98.Google Scholar

17 See, for example, Suits, D. B., “Forecasting and Analysis with An Econometric Model,” American Economic Review, LII (1962), 114–15.Google Scholar Also see n. 26 below.

18 “Interest Rates, Contract Terms, and the Allocation of Mortgage Funds,” Journal of Finance, XVII (1962), 63–80.Google Scholar

19 Ibid., 64.

20 Johnston, J., Econometric Methods (New York, 1963), 106–12Google Scholar, presents an excellent concise exposition of the assumptions sufficient for the estimators to be “best linear unbiased,” along with a demonstration of these properties.

21 Muth, , “The Demand for Non-Farm Housing,” 42–6.Google Scholar

22 Johnston, , Econometric Methods, 275–95Google Scholar, reviews Monte Carlo studies of distributions of parameter estimates and, on the basis of a “root-mean-square error” criterion which encompasses both variance of parameter estimates and bias, OLS does not appear to fare terribly badly, especially when specification error is allowed for (see p. 294 especially).

23 Ibid., 187–9.

24 Ibid., 204–6.

25 See, for instance, Fisher, , A Priori Information and Time Series Analysis, 24–5.Google Scholar

26 If the equation were designed, not as an estimate of a demand function, but as a hybrid supply-demand equation designed to explain construction activity, a positive sign on (iN − i) would not be unexpected.

This variable was formulated in the theoretical literature as a supply of credit variable whose effect was viewed as being chiefly operative on the supply of housing. In order to rationalize it in a demand context, suppose that the gross yield on housing investment is given by i + r where i = “pure” rate of interest, given by the government bond yield, and r = premium to cover risk and administrative charges. The net yield is then (i+r)−iN , where iN is the cost term represented by the NHA rate. This can be rewritten as r−(iN −i), and from this it follows immediately that if (iN −i) rises, the net yield from housing investment falls, and hence, everything else given, investment in housing decreases. A similar argument holds for a fall in iN −i, and thus investment is inversely related to (iN −i) in a demand model.

27 Brown, T. M., “Habit Persistence and Lags in Consumer Behaviour,” Econometrica, XX (1952), 355–71.CrossRefGoogle Scholar

28 Strictly speaking, the housing component of the government spending term should have been deducted, since it is in I and H, but since this is only a very small component of G (2 per cent in 1947, rapidly falling to well under 1 per cent in recent years), allowing for it would not affect the results reported above.

29 A technique for obtaining such a “combination” is discussed in Kloek, T. and Mennes, L. B. M., “Simultaneous Equations Estimation Based on Principal Components of Predetermined Variables,” Econometrica, XXVIII (1960), 45–61.CrossRefGoogle Scholar

30 “The Demand for Non-Farm Housing,” 72.

31 Lee, Tong Hun, “The Stock Demand Elasticities of Non-Farm Housing,” Review of Economics and Statistics, XLVI (1964), 86–7.Google Scholar

32 “The Demand for Non-Farm Housing,” 49.

33 “The Stock Demand for Elasticities of Non-Farm Housing,” 85–7.

34 For a recent discussion of counter-cyclical aspects of housing policy see Binhammer, H. H., “The Fiscal Implications of a Housing Program,” this Journal, XXIX (1963), esp. 344–7.Google Scholar