¹ Rosenbluth, G., “Changes in Canadian Sensitivity to United States Business Fluctuations,” this Journal, XXIII, no. 4, 11, 1957, 480–503 Google Scholar (hereafter referred to as “Changes in Sensitivity”).

² See review in ibid., section I.

³ Ibid.

⁴ See “Changes in Sensitivity,” section IV and Appendix B.

⁵ Bank debits or clearings, excluding the figures for New York City, are frequently used to indicate fluctuations in the total value of transactions. By excluding data for New York City, the bulk of the transactions that represent only transfers of financial assets, not economic activity, are excluded. See Garvy, George, The Development of Bank Debits and Clearings and Their Use in Economic Analysis (Washington, 1952), esp. 72–92.Google Scholar

⁶ All monthly time series used have been corrected for seasonal variations. For the methods used to process time series and to derive the results reported in the text and in Tables I, II, III, and V, see “Changes in Sensitivity,” Appendix B. For all Tables, sources of data are given in the Appendix to this paper.

⁷ Our record of exports, imports, and clearings goes back to 1896. In the period 1896–1902, not discussed in the present paper, none of the seven turning points in U.S. imports or the three turning points in U.S. clearings was matched by turning points in Canadian exports. The change to fairly close agreement in 1903–14 deserves further analysis, which is, however, not within the scope of this paper.

⁸ For a detailed description of the procedure and its rationale, see “Changes in Sensitivity,” section II and Appendix B.

⁹ The Gross National Product figures, being quarterly, are less sensitive to minor fluctuations in U.S. output. Moreover the period preceding 1921 is excluded. These factors may account for the higher correlation.

¹⁰ Very low ratios and average deviations are found in the period from 1896 to 1902 which is not included in the present paper. Thus, conformity of Canadian export fluctuations to U.S. output was poor before 1902, both in terms of amplitude and in terms of the proportion of turning points matched. Cf. n. 7 above.

¹¹ This paragrapn is based on changes in the percentage distribution of exports of Canadian produce, including non-monetary gold, computed from data compiled by D. W. Slater from D.B.S., Trade of Canada. For descriptions of the shifting trade patterns see also Gibson, J. D., ed., Canada's Economy in a Changing World (Toronto, 1948), chaps, v, viii.Google Scholar

¹² In Canada this fact has been obscured by the exceptional collapse of wheat exports and prices in the downswing of 1929–33.

¹³ Data in Mitchell, W. C., What Happens During Business Cycles (New York, 1951)Google Scholar, Table 42, indicate the differences in the typical cyclical amplitudes of fluctuations in the output of consumer goods, capital goods, wheat, newsprint, and certain lumber and metal products.

¹⁴ Percentages based on domestic exports including non-monetary gold (bullion included 1935–8). Computed from data compiled by Slater, D. W. (up to 1950), and from Canada Year Book, 1958 (for 1951–1954).Google Scholar

¹⁵ See Chambers, E. J., “Canadian Business Cycles since 1919,” paper presented at the annual meeting of the Canadian Political Science Association in Ottawa, 06 15, 1957.Google Scholar

¹⁶ See section V below for discussion (and rejection) of the possibility of an influence through capital exports.

¹⁷ “Reference dates” are dates of peaks and troughs in general business conditions, located through study of a large number of monthly time series. See Burns, A. F. and Mitchell, W. C., Measuring Business Cycles (New York, 1947), pp. 71 ff.Google Scholar Recent and revised dates (unpublished) have been obtained from the National Bureau of Economic Research. The Bureau's reference dates for Great Britain are available only to 1938. Annual data suggest a downswing in 1945–6 but there are not sufficient monthly data for the location of reference dates. Monthly series from 1946 to 1954 show that the only general downswing is from 1951 to 1952. A peak in June, 1951, and a trough in August, 1952, have been provisionally determined by the author by inspection of a number of deseasonalized monthly or quarterly series relating to production, prices, trade, and finance.

¹⁸ Turning points are “matched” if they are a year or less apart and there is no intervening “opposite” turning point in either country.

¹⁹ Cf. Tinbergen, J. and Polak, J. J., The Dynamics of Business Cycles (Chicago, 1950), pp. 60–2, 81–2.Google Scholar Some investigators, of course, have found better support for this theory by using different procedures for identifying “cycles.”

²⁰ But exports to Britain fell less and recovered earlier than those to the United States, as shown by the following figures (computed from Canada, Dept. of Trade and Commerce, Monthly Report), on the change in the Canadian exports of merchandise, including non-monetary gold, from 1910 to corresponding quarters in 1911 ($ million):

²¹ Exports to the United Kingdom were well maintained in 1949, during the brief recession of exports to the United States. In 1950, however, when exports to the United States were rising, exports to Britain collapsed as a result of devaluation and import restrictions introduced late in 1949. This fact helps to explain the weakness of the Canadian response to the early stages of the U.S. upswing.

²² The time series were processed as described in “Changes in Sensitivity,” Appendix B. Based on eighteen cyclical swings in the period 1919–38 and 1946–54, the analysis yields the regression equation C = 0.84S + 0.11K — 0.67, where C is the amplitude of a cyclical fluctuation in Canadian manufacturing production estimated by the equation, S stands for the amplitude of the corresponding fluctuation in United States manufacturing production, and K for the corresponding British amplitude. The units are percentage points of amplitude. The multiple correlation coefficient is 0.92, which is only fractionally higher than the coefficient obtained if the British data are omitted. The average deviations of actual Canadian amplitudes from the values estimated by the regression equation are: for 1919–29, + 2.1; for 1929–39, – 1.7; and for 1946–54, – 0.7. The corresponding average deviations that were obtained when U.S. fluctuations were the sole independent variable followed the same trend (“Changes in Sensitivity,” Table V); for 1919–29, + 3.2; for 1929–39, —2.7; and for 1946–54, —1.0. For sources of U.K. data see Appendix; for Canadian and U.S. data see “Changes in Sensitivity,” Appendix C.

²³ The analysis is based on thirty-nine cyclical swings, both expansions and contractions, covering the periods 1896–1914, 1919–39, 1946–54. The regression equation is C′ = 0.61S′ 40.22K′ — 2.33, where C′ is the amplitude of a cyclical fluctuation in Canadian exports, estimated by the equation, S′ is the corresponding fluctuation in U.S. imports, and K ′ the corresponding fluctuation in U.K. imports. The units are percentage points of amplitude.

The multiple correlation coefficient is 0.81, only slightly higher than the coefficient of 0.80 obtained when U.S. imports are the only independent variable (Table II). The average. deviations of actual Canadian amplitudes from the values estimated by the regression equation are: for 1903–14, + 3.7; for 1919–29, — 2.8; for 1929–39, + 2.6; and for 1946–54, —2.1. These averages follow the same pattern as those obtained when U.S. imports are treated as the sole independent variable (Table III, last row).

²⁴ An exact comparison of these two factors is not possible because of the difficulty of obtaining comparable data. The statement in the text is based on the fact that the ratio of exports, including non-monetary gold, to G.N.P. rose from about 17 per cent to about 22.5 per cent (Table IV), diat is, by about 32 per cent, while the ratio of average amplitudes of export fluctuations to fluctuations in U.S. debits fell from 247 per cent to 162 per cent (for “all fluctuations,” Table III)—a decrease of 34 per cent. If these two figures were accurate measures of the change in export elasticity and export-output ratio, their joint effect would be to reduce the Canadian output elasticity by 13 per cent (100 –66 X 132/100).

²⁵ Report of the Royal Commission on Dominion-Provincial Relations (Ottawa, 1940) Book I, Table 28, and p. 115.Google Scholar

²⁶ See “Changes in Sensitivity,” section I. For a good recent discussion of the implications of United States control see Blyth, C. D. and Carty, E. B., “Non-Resident Ownership of Canadian Industry,” this Journal, XXII, no. 4, 11, 1956, 449–60, esp. 458–60.Google Scholar

²⁷ See Blyth, and Carty, , “Non-Resident Ownership of Canadian Industry,” 451 ff.Google Scholar

²⁸ In attempting to measure the sensitivity of Canadian investment to United States business fluctuations one cannot, of course, distinguish between these direct influences and the indirect influences, by way of the fluctuation of exports, on investment in export-oriented industries. Bodi the direct and the indirect effects will be reflected in the actual fluctuations of Canadian investment.

²⁹ Even accurate annual figures would obscure short fluctuations. The figures used here, being estimates derived from very incomplete records, are subject to additional error.

³⁰ The twenty-one pairs of amplitudes yield a correlation coefficient of 0.90. The regression equation is I = 1.66 V — 1.07, where I is the amplitude of fluctuations in Canadian investment estimated by the equation and V is the amplitude of fluctuations in U.S. Gross National Product.

³¹ “Long cycles” in the rate of growth have received considerable attention from economists interested in problems of long-run development, particularly in the United States. For a recent summary of research in this field (United States and Canada) see the reports of Abramovitz, Moses and Buckley, Kenneth in National Bureau of Economic Research, Thirty-Seventh Annual Report (New York, 1957), 72–5 and 80–1.Google Scholar For a British contribution see Thomas, Brinley, Migration and Economic Growth (Cambridge, 1954).Google Scholar

³² It is still an open question whether the retardation of growth apparent in the thirties was a (or the) cause of the severity of the depression, or whether it was merely its statistical reflection. See Abramovitz, in N.B.E.R., Thirty-Seventh Annual Report, 73.Google Scholar

³³ See “Changes in Sensitivity,” Table V.

³⁴ For the purpose of the present discussion the marginal propensity to consume is defined as having a change in G.N.P. (not disposable income) as its denominator.

³⁵ Percentages in this and the preceding paragraph are computed from data in Canada, Dominion Bureau of Statistics, National Accounts, 1926–1950 (1952)Google Scholar and National Accounts, 1950–1955 (1956). “Fully and partially anti-cyclical transfer payments” include unemployment insurance benefits, relief payments, rehabilitation allowances, veterans' allowances, old age pensions and mothers' allowances.

³⁶ “Changes in Sensitivity,” p. 482.

³⁷ The ratio of absolute fluctuations in imports to absolute fluctuations in output is equal to the ratio of percentage fluctuations multiplied by the ratio of import to output levels.

³⁸ See special reclassification of import statistics in Slater, D. W., A Study of Canada's Imports (Ottawa: Royal Commission on Canada's Economic Prospects, 1958).Google Scholar

³⁹ For a discussion of the changing structural factors influencing imports in the thirties and after the war see ibid., chap. III.

⁴⁰ The post-war recessions are obscured by the use of annual data in Table VIII.

⁴¹ This summary is based on the comments of Dr. D. J. Daly who is, however, not responsible for my interpretation of his views.

⁴² Royal Commission on Canada's Economic Prospects, Preliminary Report (Ottawa, 1957), 13.Google Scholar

⁴³ Ibid.

⁴⁴ Canada, D.B.S., Canada's International Investment Position, 1926–1954 (Ottawa, 1956), 23–7, 30–9.Google Scholar