Economic Research at the Bank of Canada, 1935-65
by David McQueen1
It is vital for readers to understand that this account of economic research at the Bank of Canada is seriously non-definitive. For a start, it focuses on the 30-year period from the opening of the Bank in 1935 to 1965. The subperiod from 1935 to the early 1950s is firmly grounded in the invaluable, published and unpublished memoirs of the late George Watts, one of the true founding fathers of economic research at the Bank.2
But from then on up to 1965, the main knowledge base consists of my personal recollections, harking back to the 13 years (1952-65) when I was a member of the Bank's Research Department. With the best will in the world, such recollections are bound to be incomplete and unbalanced to some degree - all the more so because at no time was I ever in a position to know about everything that was going on in the department. This was the case especially after it entered a new wave of expansion in the early 1960s - a wave that included future Governor Gordon Thiessen.
I can only hope that the appearance of this article will induce others to fill in gaps, correct misperceptions, mention more names, and above all extend the story beyond 1965.
The Prewar and Wartime Years
Economists were employed by the Bank virtually from its inception, and very much by design. Breaking decisively with the contemporary culture of Canadian commercial banking from which he came, Governor Graham Towers was keen to have on tap a pool of qualified economic researchers. In this, he was strongly seconded by his deputy-governor-on-loan, J.A.C. Osborne, who during his previous career at the Bank of England had headed that institution's Statistics Office.
The resulting Research Department was a new kind of animal on the Ottawa scene. It had certain parallels in a few hard-science branches of the federal government, but never before had people formally trained in economics been assembled in quite this fashion. Some flavour of the state of things in those far-off, mid-Depression years is suggested by the fact that other public servants are said to have stood in awe of the Research Department's possession of two table-top, electro-mechanical calculators.
Certain characteristics of the early Research Department appear surprising at first glance, but what is not surprising is that the early staff, whose numbers soon stabilized at about a dozen until the end of World War II, was of high quality. Jobs for economists and competent support staff were scarce in the 1930s, and the Bank could pick and choose.
What might seem more curious was the small proportion of the department's activity that was directly related to central banking. For this, there were two reasons. First, although the Bank quickly became very busy in such mandated roles as financial advisor to the federal government, monetary policy was not a notably active area during the first two decades of the institution's history.
By way of illustrating this, ex-Governor Gerald Bouey has pointed out that Bank rate was set at 2.5 percent in 1935, was lowered to 1.5 percent in 1944, then raised in 1950 to 2 percent, where it remained until 1955. Hardly a dramatic range of variation! Monetary policy was kept understandably easy during the Depression conditions of the 1930s, and subsequent wartime inflationary pressures were judged best kept in check, not by tight money but by high taxes, direct price and other controls, and the massive public sale of Victory Bonds. And while there were inflationary episodes over the first postwar decade, fiscal policy was such as to restrain them somewhat. The scope for effective monetary restriction was limited by the statutory 6 percent ceiling on bank lending rates, and there was a reluctance to take actions that tended to reduce the market price of still widely held Victory Bonds.
A second reason why the Research Department's activities often ranged quite far from central banking per se arose precisely from the department's initial novelty and uniqueness on the Ottawa scene. The Finance Department and other parts of government quickly came to appreciate the utility of this small but critical mass of economic expertise, and soon they were referring various chores to it. The Bank's senior management put up little if any resistance to this, concerned as they were both to gain Ottawa's wholehearted acceptance of their new institution and to improve the latter's collective understanding of how Ottawa worked.
Very soon, a non-refusable chore of the first magnitude landed on the Bank's and the department's doorstep: the parlous financial state in which, by the mid-1930s, the governments of the three Prairie provinces found themselves. Staggering under the one-two economic punches of Depression and drought, they teetered on the brink of debt default, and in Alberta's case actually went over the edge for a time. Research chief Alex Skelton and members of the department became increasingly involved in this situation, which in 1937 led to the appointment of the Royal Commission on Dominion-Provincial Relations (the Rowell-Sirois Commission). Skelton was appointed secretary of the commission, and Research Department economist John Deutsch and librarian Mary Rowland were also seconded to it on a full-time basis. The multi-volume report and studies, with Skelton as the chief author of the first main volume, were published in 1940.
Some further remarks about Alex Skelton are indispensable to this account. Though he was actually on the premises for a relatively small proportion of his 1935-44 appointment period (most of his time being taken up, first by the Rowell-Sirois Commission, then by numerous wartime assignments), he nevertheless left an indelible mark on both the character and the legendry of the Research Department.
Traditionally, as befitting guardians of the nation's money supply, central bankers try to project at least a public facade of calm responsibility and sobriety. How piquant, therefore, that two of the major figures in the Bank's early history, Skelton and secretary Donald Gordon, were among the most colourful (as well as the most able) personalities ever to impact Ottawa.
Rising spectacularly from a very modest background of formal education, Gordon went on to subsequent fame as wartime price controller, then president of Canadian National Railways.
Skelton, by educational contrast, was a star Queen's University economics graduate and Rhodes scholar. Uncommonly energetic and impatient of bureaucracy, he was a formidable practitioner both of unusually sustained bouts of hard work and of late-night partying. In the latter sphere, during federal-provincial fiscal negotiations in the mid-1940s, he found an unlikely soul-brother in the person of Quebec Premier Maurice Duplessis. (This story, I had better note, comes from a highly reliable source - namely, the late Joseph Howes, a close Bank colleague and admirer of Skelton.)
Adversarially positioned during working hours, Skelton and Duplessis nevertheless teamed up after nightfall to explore the rest-and-relaxation offerings of Hull, which in those days served as a sort of trans-riparian safety valve for staid Ottawa - an Ottawa where, for example, dancing in the Canadian Grill stopped promptly at midnight Saturday, and no liquor was served any time.
A difficulty for Skelton was that often, the next morning, he had to consult closely with Prime Minister Mackenzie King, whose puritanical views were well known. To mask the residue of the night before, Skelton resorted to Sen-Sen breath fresheners, until one day the P.M. looked up from the spreadsheet they had been examining and said, "See here, Skelton - I don't mind if you drink, but will you please stop sucking those dreadful lozenges!"
From the Prime Minister and Governor Towers on down, Skelton was pardoned his eccentricities out of respect for his enormous energy and abilities. The crowning achievement of his tragically foreshortened career (he drowned in 1950) was undoubtedly his secretaryship of the Rowell-Sirois Commission. Far more than a standard commission secretary, he was the main driving force behind the whole exercise, and set an important precedent by bringing together, productively, academic economists and other social science researchers on a scale never before seen in Canada.
But second only to that was the running start and persisting inspiration that he imparted to the Bank of Canada's Research Department. No later member of the department could quite match his vivid personal style, but something of his spirit of irreverence used to resurface in the skits performed at the annual departmental Christmas parties, held in William Lawson's capacious apartment and attended by senior management as well as departmental staff. These satirized such strangely irrational tribal customs of the 1950s as going through double-digit drafts and many hundreds of Research Department person-hours in order to produce one slim Annual Report.
Following the end of the war, the department underwent a wave of expansion. Among the newcomers were Sterling Suggett, Edith Whyte, William Lawson, Bernard Drabble, Gerald Bouey, George Freeman, Douglas Humphreys, and Al Noble. Stephen Handfield-Jones and I, arriving in the early 1950s, should probably be considered as belonging more to this wave than to any succeeding one.
During the war, emergency after emergency had kept the small departmental staff on the hop from one area to another. Peacetime generated emergencies too, such as the 1958 Conversion Loan, but did so less frequently. This, together with larger numbers, made it possible to achieve greater stability in staff assignments.
If I may take my own case as reasonably representative, people tended to stay in specific areas for around three years - enough time to learn the job, but not enough to grow stale in it. After an initial stretch in my graduate dissertation area of international economics, I moved into public finance under Joseph Howes, then into the pressure cooker of money and banking, and finally into "general economics," including forecasting.
Under Howes, I first compiled, then tried to adjust for consistency, statistics of municipal revenue and expenditure, destined for inclusion in the National Accounts. This was a task, begun in Rowell-Sirois days, the whole of which "DBS" (the Dominion Bureau of Statistics, now Statistics Canada) was not yet quite ready to take on. Then Howes and I were seconded to the Gordon Commission on Canada's Economic Prospects to assist Yves Dubé of Laval University in a 25-year forecast of expenditure on housing and "social capital" (the physical kind - roads, schools, university buildings, waterworks, etc.). Other Research staffers also did work for the commission - e.g., Stephen Handfield-Jones helped pioneer flow-of-funds accounts for Canada.
Disdaining any pretensions to high-powered economic theory, Howes was a mine of accurate information about provincial and municipal finances, which he kept up partly through multifarious personal contacts. He decided that the three of us - Dubé, Howes, and myself - should tour Canada east and west to interview key municipal and educational decision makers. I questioned whether this would much improve our forecast numbers, which were for modest growth. Howes conceded the point but said that it would give us a better feel of the problems, and in any case he thought it important for Yves to see western Canada.3 And so we set off, close to half the western leg being by dome train, as Howes disliked airplanes.
As it turned out, Howes and I were both right and wrong. Urban population apart, our forecast numbers were ludicrously low, but we managed in our report to say a few useful things about the nearer-term implications of exploding urbanization. As to seeing the West, Yves told me many years later, when he was director of research for the Canadian Transport Commission, that the trip out to the Pacific and back had been one of the more memorable and significant events in his career.
Thus it would seem from this example that the Research Department continued in the late 1940s and the 1950s to be a broad-ranging facility, serving much more than just the central bank and offering to its members a practical general education in the Canadian economy, for which I remain grateful to this day.
But under the surface, things were changing. Other Ottawa departments and agencies, notably Finance, Trade and Commerce, and the future Statistics Canada, were acquiring greater research capabilities of their own. Even within the Bank, researchers began to be spread around a bit more widely, with Douglas Humphreys and Al Noble moving into the Securities Department.
Also, events were moving the Bank as a whole towards a stronger focus on its own operations and their economic effects. Starting in the mid-1950s and paralleling developments in the United States, monetary policy became at once more active and a more prominent subject for academic and general public debate. As well, the so-called "Coyne Affair," culminating in the resignation in 1961 of Governor James Coyne and his replacement by Louis Rasminsky, encouraged a view both inside and outside the Bank that it should henceforth devote more of its intellectual resources to clarifying and enunciating its own role in the economy.
The Bank and the Porter Commission
Moreover, hard on the heels of the change of governors came the appointment in 1962 of the Royal Commission on Banking and Finance (the Porter Commission), which reported in 1964. Now indeed the Bank was specifically challenged to provide a thoroughgoing, consistent analysis of its role. Robert Johnstone, who had joined the Research Department in 1957, was seconded to the commission staff, and the preparation of the Bank's own submission to the commission preoccupied a number of other researchers, including notably Gerald Bouey and George Freeman, and ex-department-members Robert Beattie (senior deputy governor) and William Lawson (executive assistant to the governors).
In its submission, the Bank concluded that its main impact on the economy came through its ability to influence variations in credit conditions - in the difficulty, expense, and risk of borrowing, and in the counterpart attractiveness of acquiring and holding various kinds of financial assets.
As to the Bank's broad objectives, the submission noted that they were those of economic policy generally, including "sustained economic growth at high levels of employment and efficiency, internal price stability and the maintenance of a sound external financial position, an equitable sharing of economic benefits and burdens, and the maintenance of a high degree of economic freedom." The Bank did feel a particular obligation to ensure that price stability was adequately emphasized, but it did not perceive any basic incompatibility between this and the other objectives.
Later in his testimony, Governor Rasminsky said he thought it "would be a mistake to incorporate into the legislation any specific instruction to the central bank to focus its policies and its attention exclusively on the achievement of any single objective."
Such was the account of itself and its place in the economy that the Bank gave after just under 30 years of existence - a period strongly influenced by both the experience and the memory of depression and war. What was to be another highly influential (one might almost say "haunting") experience - that of 1970s stagflation - still lay some years into the future.
Many readers, anxious to compare the intellectual underpinnings of the Bank's submission to the Porter Commission and those of its public statements of the early 1990s, will be eager at this point to learn about the evolution of economic research and analysis at the institution over its second 30 years of life. Unfortunately, I must disappoint them. Like the central banker of legend, I must take away the punch bowl just as the party grows particularly interesting. Having left the Bank in 1965 and ventured into other fields, I do not have the necessary detailed information to carry the main story line forward from the mid-1960s. Others better informed, in due course "recollecting in tranquillity" and freed by the passage of time from worries about upsetting the financial markets, must take up the challenge, as I sincerely hope they will.
Handling Data and Analysis in the 1950s and 1960s
But something I can do that will perhaps be of interest is to sketch how we handled economic data at the Bank in the 1950s and early 1960s, and how it gradually became apparent why we had to undertake a relative shift away from our hitherto highly judgmental approach towards greater use of formal, explicit, mathematical models. Ours was not a unique experience: other economics shops of my acquaintance, both in Canada and abroad, were also coming, with varying degrees of reluctance, to the same conclusion at about the same time.
Keynes, though famously intuitive by nature and deeply sceptical about econometrics, nevertheless said that his best ideas came to him from "messing about with figures and seeing what they must mean." That well describes a lot of what we used to do in our pre-econometrics era. Thanks in no small measure to the likes of Margaret Fitzpatrick and other members of the best senior clerical staff I have ever worked with, keeping constantly in touch with their counterparts at DBS, we did know our statistical series, with all their little quirks, discontinuities, strengths, and weaknesses. We diagrammed them assiduously, referring to the results in maritime fashion as "charts," and constantly eye-balling them for trends, turning points, irregularities, and likely candidates for revision.
I still think this an excellent way to begin an empirical investigation. It raises good questions, generates a helpful provisional separation of the significant from the trivial, and so forth. What has changed, and rightly, is the follow-up: more math and less intuition (although those who would now deny virtually any role for the latter, such as "visioning" and gap-bridging, are only deceiving themselves).
Prior to the arrival at the Bank of the econometrics revolution, various sophistications such as semi-log scales crept into our number-imaging. Don Daly and his Trade and Commerce colleagues deserve much credit for their unflagging proselytizing zeal in favour of seasonal adjustment and the use of business-cycle indicators. That involved a two-stage battle: first they had to convince us, then we had to convince our own management; but eventually the battle was more or less won.
In short-term general economic forecasting, we deprecated fancy academic talk about models, but of course our procedures implied a model. As best I can reconstruct it from memory, it was a simple, rough-and-ready Keynesian affair, strongly oriented to the demand side of the economy, such as might have been found in almost any of the Economics 101 textbooks of the period.
Our basic procedure was to gather every possible forward-looking indication of three key exogenous variables - exports, government expenditure on goods and services, and private fixed investment. Having worked out rough values for these and having also given a little special attention to the prospects for inventories and for sales of cars and other consumer durables, we would then try to construct around these numbers a plausibly consistent set of National Accounts, all the while comparing notes frequently with other Canadian forecasters.
For our fix on much the biggest part of exports, plus some other things, we depended heavily on annual or even semi-annual trips to New York and Washington, making the rounds of a first-rate visiting list of U.S. economic pulse-takers built up by Bernard Drabble. This was highly instructive, and also just plain fun, as when interviewing shrewd, salty, straight-talking characters such as James Knowles, then research director of the Joint Economic Committee of Congress.
Readers will note our heavy implicit concentration on the future positioning and shape of the aggregate demand curve. Aggregate supply curves had not yet come into common analytical and teaching use, but in retrospect our view about the chief cause of inflation signalled the presence in our model of an implicit, reverse-L-shaped supply curve, with a bit of rounding at the corner.
In effect, we paid little attention to this curve, assuming it to be pretty well stable in the short run. Inflation was held to be chiefly caused by excessive demand pressure on available productive capacity - by an aggregate demand curve intersecting the stable aggregate supply curve somewhere significantly up the latter's vertical range. Any scenario involving, say, a sudden quadrupling of world oil prices, and a consequent leftward shift of the short-run aggregate supply function, would have seemed to us the wildest fantasy.
The Beginnings of Econometrics and Modelling at the Bank
Long before the 1973 oil price shock, however, disquiet about our methodology was growing in our minds. The burgeoning of econometrics in the academic world, greatly assisted by advances in computer technology, must surely have something practical to offer us! But what really brought matters to a head for me was not a perception that we were consistently forecasting gross national product (GNP) much worse than anybody else (which I don't recall that we were), but rather a query by Governor Rasminsky at a regular Friday morning meeting. The Canadian dollar had recently been devalued to a new fixed rate of 92.5 cents U.S. How had this affected our forecast? What amount and kind of allowance had we made for it?
It was a reasonable question, to which, humiliatingly, I had no good answer, beyond a mumble that the devaluation had moved us up into a more optimistic zone of our GNP range. Devising a better answer, however, would be a formidable task, requiring explicit estimation of relevant elasticities and multipliers, in a situation where much of our past data came from a floating-exchange-rate regime, whereas Canada was now back on a fixed rate. Responses to fluctuations, such as the one that had just occurred, could now well be different, with a no-doubt complex lag structure.
At about this time, Sidney May at Trade and Commerce had built an econometric model of the Canadian economy from which, in principle, could be drawn some of the parameters we needed. But May's model generated fixed investment endogenously, throwing up figures much lower than those of the annual surveys of investment intentions, for whose performance we had acquired a degree of experience-based respect.
In any case, cherry-picking parameters from somebody else's interactive model seemed to me improper. Either you got to know and work with the whole model, or you built and tested one of your own, forcing yourself to spell out explicitly its multiple interdependencies and otherwise putting yourself through a major, on-going learning process.
Elsewhere in the Research Department, others were similarly coming to a conclusion that our research technology needed updating. A first reaction was the mounting of a couple of part-time courses for staff in elementary econometrics. Taught by William Hood, then George Post, these certainly helped to open our eyes to possibilities, but much more was needed. What was required was nothing less than a fresh wave of new-generation staff, more fully trained and immersed in these mysteries.
And that is what the department got in 1966, with the arrival of Ian Stewart and his assembly of a mixture of suitably qualified full-time staff and part-time academics to begin the construction of the first econometric model in the RDX series. No instant miracles were expected or achieved, but at least progress could now be made towards developing better answers to the sort of question that Governor Rasminsky had asked.
Subject to much enlargement and modification through time, notably to improve their projection capabilities, the RDX models continued to be used extensively at the Bank until the early 1990s. More recently, there has been increasing emphasis on a two-model system called the Quarterly Projection Model (QPM), described as "calibrated," as opposed to estimated. Other models, including QPM satellites, are also used for various purposes.
In looking back over the 1935-65 period that is the main focus of this article, I believe that the shift towards a more high-tech and in some ways more narrowly focused research operation at the Bank was an inevitable development, given the evolution of surrounding circumstances. If I have any reservation, it is that there are indeed no free lunches and that all progress comes at a cost - in this case, a loss of breadth. Retrospectively, I find that I still put a high value on some of my earlier, low-tech learning experiences at the Bank, and I would urge that a deliberate effort be made to ensure that new generations get some exposure to the same sort of benefits.
Related to this, it seems to me important also that the independence of the central bank (to whatever degree) not mean isolation. Both high-tech research and high finance can be deeply fascinating and absorbing, attracting clever minds, who in turn attract still further clever minds, and all this comes to seem more and more the centre of the universe. It needs to be remembered that Main Street - not Bay Street, or Wall Street, or Lombard Street - is where the real, goods-and-services economy happens. Bay Street (and Wellington Street in Ottawa!) exist to serve Main Street, not the other way round. A healthy research career should include some timely reminders of this.