Information identified as archived is provided for reference, research or recordkeeping purposes. It is not subject to the Government of Canada Web Standards and has not been altered or updated since it was archived. Please "contact us" to request a format other than those available.
by Philip Cross 1
The US officially entered a recession in December 2007, with the downturn morphing from a mild to a severe slump in the autumn of 2008. This paper looks at the broad implications for Canada of past US recessions, and some of the factors that separate a severe downturn from milder slumps in Canada.
Recessions in the United States have been accompanied by a wide range of outcomes in Canada. The sharp contractions in the US during 1974-75 and 1981-82 were associated with a mild and a severe recession respectively here in Canada. The mild downturns in the US in 1990-91 and 2001 were accompanied in Canada by a severe recession in the former case and no recession in the latter.
This paper looks at some of the reasons for these different outcomes. In doing so, we recall earlier studies of what a recession is, how different sectors of the economy behave during slumps, and the relationship between GDP and the labour market during recessions. The goal is to provide an overview of how recessions compare in Canada and the US and to dispel the notion that the course of the US economy alone determines our fate (summarized by the myth that when the US economy sneezes, we inevitably catch a cold). However, it is not an exhaustive study of all the linkages between the two economies during downturns.
Since much of the analysis in this paper focuses on the behaviour of real GDP during recessions 2 , it is worth repeating that neither Statistics Canada nor the National Bureau of Economic Research relies exclusively on GDP in determining recession dates (the NBER has been the arbiter of recessions in the US since the 1920s, while Statistics Canada has filled this vacuum in Canada since 1981). Both organizations specifically eschew the rule of thumb that a recession reflects two consecutive quarterly drops in GDP. 3 For example, in both 2001 and 2008, the NBER identified recessions without back-to-back declines in GDP, as did Statistics Canada in 1975. 4 In fact, the NBER noted that the onset of recession late in 2007 was clear in all its indicators except real GDP, which it called “ambiguous” about when the downturn began.
Both organizations look at variables besides GDP, notably employment. The NBER states that, “there is no fixed rule about which other measures may contribute information in determining dates for the business cycle.” 5 The fundamental point is that, as we said in 1996, “recessions are a process, not an event” (such as consecutive quarterly drops in GDP). 6 Researchers at the Economic Cycle Research Institute noted some aspects of this process: “It is the comovement of these variables [output, employment, sales and income] that generates the cycle. Under certain circumstances, if consumers pull back on spending, businesses respond by producing less and cutting jobs, which lowers personal incomes. That is why during recessions all these measures of activity fall together, spreading and diffusing like wildfire”. 7 As well, the diffusion of losses across the economy is important.
Inevitably, some judgement is required about the root cause of any drop in economic activity. Business cycles are self-perpetuating turns in economic activity: a quarterly drop in GDP caused by poor weather, followed by another quarterly decline due to a labour dispute would not meet this criterion. Indeed, the first half of 2008 in Canada is a good example of output being disrupted more by supply problems than a slump in demand. These shortfalls of supply ranged from record heavy snow to the introduction of a new holiday in Ontario (both of which helped lower GDP in the first quarter) to extensive maintenance in the oilpatch which dampened growth in the second quarter.
Finally, Statistics Canada’s contribution to research on business cycles has been assigning dates to the periods of expansion and recession. This paper does not address the question of whether Canada is currently in a recession, since it is too early as yet to determine if the downturn will be prolonged enough to qualify as a recession nor exactly when the downturn began. These questions will be addressed when the data warrant the definitive determination of a date. Instead, the focus is on the implication of past recessions in the US for growth in Canada.
The US experienced deep slumps in both 1974-75 and 1981-82. The 1970s downturn in the US was actually slightly steeper, although it was less prolonged. Real GDP fell 3.1% from peak to trough in 1974, slightly more than the 2.9% drop in the early 1980s.
The 1974 recession in the US started as a mild dip, but output and jobs fell rapidly after mid-1974. The sudden worsening of the economy after September 1974 when “it looked as if the economy was on the verge of a great depression: as if the sky were about to fall” is attributed by Minsky to the insolvency of the Franklin National Bank on October 8, 1974. 8 This bears some resemblance to the impact of the Lehman Brothers failure in September 2008. The epicentre of the mid-1970s recession in the US was concentrated in the housing and auto sectors. Whether the latter was due to just rising gas prices and interest rates or also reflected the impact of the Arab oil embargo remains controversial. 9 What is important for Canada is that exports fell by over 12% due to the drying up of key markets for our lumber and autos (which then accounted for nearly one-third of exports, versus just 13% in 2008). 10 This was by far the largest impact of any US recession since World War II on Canada’s exports, over twice as severe as the export decline in 1981-82. The prolonged drop in Canada’s exports in the second year of recession in 1975 was concentrated in forestry products. The lingering slump in US housing partly reflected “a classic speculative bubble” originating in the Real Estate Investment Trust (REIT) industry, which financed construction of multifamily housing, condos and commercial property. 11
Despite the large impact of the US recession on exports in 1974-75, the Canadian economy emerged relatively unscathed. Real GDP posted only one quarter of mild decline. Conversely, the 1981-82 recession was much more pronounced in Canada, with a drop in real GDP of 4.9%.
In Canada, the shallowness of the 1975 downturn and the severity of 1981-82 reflect the course of domestic spending, not exports. In 1981-82, domestic demand fell 5% in volume, driven by sharp declines in business investment, housing and consumer spending. Conversely, domestic demand rose nearly 3% in the year after the onset of the US recession in 1974. However, many factors link Canada’s economy to the US besides direct exports, including financial markets and confidence, which seem particularly relevant in the current environment.
The US experienced mild recessions in 1990-91 and 2001. The 1990s recession began when oil prices soared after the Iraqi invasion of Kuwait at a time when inflationary pressures were already high. Some also blame the savings and loan crisis, but this does not explain the exact timing of the recession in the US nor the greater impact in Canada. 12 The 2001 slump followed the bursting of the ICT bubble in both the stock market and business investment spending (partly due to the overhaul of many corporate computer systems to head off potential problems with the Y2K changeover at the turn of the millennium).
Both these recessions were relatively minor in the US, lasting about a year and with peak-to-trough GDP declines of 1.3% and 0.4% in 1990 and 2001, respectively. The response of Canada’s economy was strikingly different. The recession starting in 1990 was more pronounced, with real GDP falling 3.4% by early 1991, while employment receded steadily into 1992. Conversely, there was no recession overall in Canada in 2001, despite huge losses in some high profile ICT industries (ICT manufacturers, for example, slashed output nearly in half, although ICT services grew steadily). In fact, at its worst late in 2001, the 10.5% year-over-year drop in manufacturing output was comparable to its 11.8% contraction at the nadir of the 1991 downturn. However, the weakness was largely confined to manufacturing in 2001, which is why the overall economy avoided a recession. The drop in ICT exports led to a 5% loss in Canada’s overall exports during 2001, a retreat typical of exports during a recession in the US.
Again, the different response of Canada’s economy to these mild downturns in the US reflected the pattern of domestic demand. Inflation was more entrenched in Canada than in the US entering 1990. As a result, interest rates were already much higher at the start of the year. While interest rates fell gradually in the US in 1990, they rose in Canada even as the economy slowed, partly out of concern that the introduction of the GST (which shifted the tax burden from exports and investment goods to consumer goods and services) might trigger higher wage demands.
The introduction of the GST on January 1, 1991 at the height of the Gulf War helped precipitate a 2.3% drop in consumer spending in the first quarter of 1991, its largest quarterly decline on record back to 1947. The prospect of lower taxes on investment goods also led firms to postpone some purchases from 1990 to early 1991. This temporarily interrupted the overall slide in business investment. However, for the whole 1990 to 1992 period, business investment fell nearly 10%, one of the largest declines in a recession.
Canada escaped a recession in 2001 partly because the ICT sector was a smaller share of our economy than in the US. Business investment in equipment and software hit a record 9.4% of GDP in the US in 2000 (before retreating to 7%), versus a peak of 8.1% in Canada (which fell to just under 7%). As a result, while ICT output and investment receded sharply in both countries, this had less of an effect on Canada’s overall GDP. As well, American households suffered a much larger drop in household wealth in 2001, as their stock market holdings are a greater share of household wealth than in Canada. Household spending on big-ticket items and business investment grew steadily in 2001, except for a small interruption associated with the 9/11 terrorist attacks.
Since domestic demand is an important factor in whether Canada experiences a mild or a severe recession (or none at all), it is worth looking at some of the components of domestic demand. Using a classification developed in our 2001 paper on post-war business cycles, 13 Figures 7 and 8 show that consumer spending is remarkably resilient in both mild and all but the most severe recessions. In fact, consumer spending advanced steadily in all the mild recessions, with the exception of 1951 (which reflected rationing introduced at the start of the Korean war). But consumer spending rose even during a majority (3 out of 5) of severe recessions. The two exceptions were 1981-82, when interest rates soared above 20%, and 1990-91, when the introduction of the GST and the Gulf War led to a sharp drop in spending in the winter of 1991 (otherwise, consumer demand was little changed during that recession). Housing is always more sensitive to downturns in the economy, falling even during mild recessions, but recovers quickly.
Business investment is usually a key determinant of the severity of recessions in Canada. Investment rose slowly but steadily in almost all the mild recessions after World War II (Figure 9). Business investment in plant and equipment fell by nearly 10% or more in the deeper recessions of 1954, 1957, 1981-82 and 1990-91 (although the last was interrupted by a spike in investment spending after firms delayed purchases from 1990 to take advantage of the GST-induced drop in the price of capital goods after January 1, 1991).
The key role played by business investment in determining the depth of a recession reflects a number of factors. Investment decisions by firms are much more complex than the household decision to buy a home or a car, involving as they do assessments of the current and future course of demand, inventories, credit availability, prices and profits. And once taken, investment decisions are more difficult to reverse.
This is not to say that domestic demand is not affected by events beyond Canada’s border. Firms in Canada raise a significant portion of their funds in US financial markets, which are deeper and more liquid. 14 Business investment in Canada is heavily influenced by commodity prices set on global markets. This is particularly true of the current cycle after energy and mining rose to nearly half of all investment outlays during the recent boom. Spending by manufacturers also is always very sensitive to export demand. Investment spending in energy and manufacturing (the sectors most exposed to export demand) together accounted for over half of investment in 2008, and almost single-handedly account for the drop in business investment plans for 2009. In turn, layoffs in the resource and manufacturing sectors could hamper consumer spending. And the global financial crisis in 2008 shows the increasingly inter-related nature of the modern economy, which has already led to sharp drops in purchases of homes and autos by nervous Canadian households late in 2008.
Every business cycle has its unique characteristics. One of the most marked differences between the current cycle and the previous two in Canada is the course taken by interest rates. In both 1981-82 and 1990-91, interest rates were well into double-digit levels as the economy turned down. Moreover, short-term interest rates rose further even as the economy began to slump (and in the spring of 1982, rose again to abort the first stirring of a recovery in the US). 15
The contrast with 2007-08 is striking. In Canada, the prime rate was about 6% through most of 2007, two percentage points below the US rate. This was a reversal from their relative position in the previous two recession cycles. Government interest rates fell sharply in both Canada and the US as the global credit crisis began to unfold in August 2007, with the prime rate reaching 4% in both countries after the crisis intensified following the bankruptcy of Lehman Brothers in September 2008.
However, while central banks did not raise their interest rates during the current slowdown, the turmoil in credit markets nevertheless led to a sharp hike for many borrowers. The most dramatic example was in the US. Figure 11 compares the federal funds rate set by the Federal Reserve Board with the 3-month US dollar LIBOR (London Interbank Offered Rate) charged by banks — roughly speaking, a comparison of what the Fed wanted interest rates to be and what many borrowers actually paid. Normally, the two are very close. A gap first appeared in August 2007, when the onset of the credit crisis (marked by the suspension of redemptions at three investment funds at BNP Paribas) sent LIBOR up, which the Fed countered with cuts to its rate. The two interest rates were realigned by March 2008, when the collapse of the Bear Stearns investment bank kept LIBOR high even as the Fed further lowered its rate.
Finally, the Lehman Brothers bankruptcy on September 15, 2008 sent the LIBOR rate soaring despite drastic cuts to the federal funds rate. The increase in LIBOR accompanied even sharper hikes in other rates paid by borrowers (at the extreme, bonds rated C grade, or junk status, hit 30%) or the complete freezing of markets such as for 90-day commercial paper. 16 The net result was a severe tightening of credit conditions in the midst of a recession that had begun in the US nine months earlier.
Another potentially important difference in the current cycle is the sharp drop in the terms of trade for Canada (the ratio of export to import prices) from their record high set in June. Already, the terms of trade had fallen 11% by December, as commodity prices continued to slump through year-end. In the recessions starting in 1974, 1981 and 1990, the terms of trade were relatively stable, with average declines of 4.9%, 3.8% and 3.7%. This implies that the terms of trade will play a more important role in the current downturn, as already seen in investment plans for 2009.
Some readers may be surprised by the economic pain experienced during even mild recessions: after all, a fractional drop in output of less than 1% does not sound unduly harsh. But the largest cost of mild recessions is not the actual drop in GDP, but the loss of potential output gains. Instead of expanding by 3% (the actual average real output gain from 2003 to 2007), the absence of growth represents $49 billion of foregone incomes in 2008. By comparison, the difference between no growth and a 0.5% decline is equivalent to only $8.2 billion. 17 From this point of view, the debate about whether the Canadian economy was in recession in the first half of 2008 misses the more fundamental point that growth had slowed from 3% to nearly nothing. Of course, a severe recession where GDP falls 3% (or more) implies an actual contraction of over $49 billion from the current level of GDP, on top of $49 billion of foregone potential output.
That most of the costs associated with a mild recession are incurred due to the slowdown of growth rather than the outright contraction of the economy is reflected in jobs and the unemployment rate. Below 1% growth, jobs no longer proliferate enough to keep up with the normal increase in the labour force, and unemployment rises. Figure 12 shows that recessions are consistently accompanied by marked increases in the unemployment rate, usually more than one percentage point. In fact, the onset of recessions are more easily identifiable in graphs of the unemployment rate than plots of real GDP. One caveat is that the unemployment rate tends to lag the business cycle early in a recovery. Employment and unemployment both react quickly to a downturn in GDP. While employment lags only slightly at upturns, the recovery of the unemployment rate is delayed by an influx of people into the labour force looking for work as employment recovers. This is why both Statistics Canada and the NBER look at employment but not unemployment when dating recessions.
In severe recessions, the year-over-year increase in the unemployment rate hit 5 percentage points in 1982 in Canada and half that in 1982 in the US and in 1991 in Canada. Milder recessions saw increases of 1 to 2 points (such as 1980 and 2001 in the United States). The sharp hike of over 3 points in the unemployment rate by early 2009 in the United States is symptomatic of a severe recession.
The close correlation of employment with recessions in recent business cycles is also consistent with recent research that recessions impact employment as much as output. As noted by Robert Hall, chairman of the NBER committee 18 that dates recessions in the US, “One of the important features of the modern recession is that productivity does not decline as it did in earlier ones.” 19 The notion that employers ‘hoard’ workers during downturns no longer seems to hold. 20 This finding supports Statistics Canada’s decision in 1996 to look at both GDP and employment when determining recession dates.
Recessions in Canada and the US are often, but not always, closely synchronised. Moreover, there is a large disparity in the magnitude of the downturns since the severity of a recession in Canada is usually determined by the course of domestic demand, not exports. Of course, domestic spending is influenced by global trends, especially the impact of commodity prices on business investment. The record drop of our terms of trade late in 2008 suggests that the global recession will play a determinant role in domestic spending.
Every business cycle has its unique characteristics. The current downturn in the US has already had several remarkable features, most notably the magnitude of the shock to the global financial system and the speed and breadth that its repercussions were felt in the global economy. Another is the marked easing of monetary policy from the very onset of the credit crisis. For Canada, there already has been the largest drop in the terms of trade on record, as commodity prices reversed course. Another is the very quick response of firms in the US to lower employment from the onset of the recession, even before GDP started to recede decisively. Employment in Canada began to fall almost a full year after the US. It remains to be seen whether the path of the two economies will continue to converge in 2009.
September 1947 to March 1948
February 1949 to July 1949
June 1951 to December 1951
April 1953 to April 1954
April 1957 to January 1958
February 1960 to March 1961
March 1970 to June 1970
January 1975 to March 1975
February 1980 to June 1980
July 1981 to October 1982
April 1990 to April 1992
February 1945 to October 1945
November 1948 to October 1949
July 1953 to May 1954
August 1957 to April 1958
April 1960 to February 1961
December 1969 to November 1970
November 1973 to March 1975
January 1980 to July 1980
July 1981 to November 1982
July 1990 to March 1991
March 2001 to November 2001