Statistics Canada - Statistique Canada
Skip main navigation menuSkip secondary navigation menuHomeFrançaisContact UsHelpSearch the websiteCanada Site
The DailyCanadian StatisticsCommunity ProfilesProducts and servicesHome
CensusCanadian StatisticsCommunity ProfilesProducts and servicesOther links

Warning View the most recent version.

Archived Content

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.

Canadian Economic Observer
September 2005

Feature article

Long-run Cycles in Business Investment

by Philip Cross*


Business investment in plant and equipment is a critical factor in economic growth. Investment cycles have a major influence on the overall business cycle: rarely does the economy accelerate to peak rates of growth without a surge in capital spending, while slumps in investment spending are associated with the most severe and prolonged recessions. As well, the stock of capital is a key determinant of productivity.

As a long-time student of business cycles stated, “Of all the components of aggregate demand, it is the investment spending by firms and households that is the prime mover in economic fluctuations, being by far the most cyclical and most volatile. This fact has long been known and recognized” (Zarnowitz, 1999, p. 73).

The pivotal role of investment is all the more striking because of its relatively small share of total spending. Business investment averages only about 12% of GDP, but like housing it frequently punches above its weight. At the height of expansions over the last four decades, capital spending contributed between one-quarter and one-half of all growth, while it directly accounted for half of the drop in GDP during recessions in 1982 and 1991. This is because it can oscillate between booms with annual growth as high as 27% or busts with declines approaching 20%. In contrast, consumer spending, which accounts for over half of total spending, has fluctuated in a much smaller range of -2% to +5% in the last two decades.1 This is partly because of the role played by necessities (such as food, shelter and heating), which cannot be postponed under most circumstances but do not expand proportionately as incomes rise.

Recent years have been a good example of the variability of investment. Capital spending led a surge in GDP growth in the second half of the 1990s, hitting a crescendo during the ICT boom that peaked in 2000. The unraveling of this investment boom was reflected in a sharp slowdown in overall growth early this decade. Since then, investment has begun to recover.

After a brief overview of the importance of business investment in GDP and some basic facts on the composition of capital spending,2 this paper looks at the long-run cycles in investment spending and which components drive these fluctuations. It then compares the recent upturn in capital outlays with previous investment cycles.

The structure of investment

Business investment consists of two major categories – structures and machinery and equipment. Structures include buildings (such as offices, factories and shopping malls), as well as engineering works like dams, pipelines and oil and gas wells. Machinery and equipment encompasses industrial and farm machinery, office equipment (including computers and software) and transportation equipment.

Our greater reliance on capital-intensive industries, such as mining (especially oil and gas), utilities and communications explains why business investment is particularly important to Canada, where its share of GDP is consistently higher than in the United States. The US, on the other hand, invests more in machinery and equipment. This was especially true in the 1990s, when the US led all countries in investing in ICT goods. But the US also saw a faster drop after the ICT collapse early this decade.

Figure 1

The share of GDP in Canada devoted to investment hit its all-time high of 16% in 1981. This was driven by the commodity price boom at that time, especially energy. This also marked the last time spending on structures exceeded machinery and equipment, a testament to the scale of the large engineering projects spawned by the energy crisis.

Figure 2

After that, cuts by the energy and mining sectors sent spending on structures to an all-time low of just 4% of GDP in the 1990s, and it has remained subdued since due to weak building construction. Meanwhile, machinery and equipment hit a record 8.1% of GDP at the height of the ICT boom in the late 1990s. The bursting of the ICT bubble followed over-investment in this sector, which has since gravitated towards its long-term average of 7%.

Figure 3

The share of investment in GDP often goes up during the early part of a recession, as in 1970, 1975, 1981 and 1990. This reflects the cost and difficulty of stopping capital projects once they are initiated. Instead, investment usually bottoms out about two years later, when recovery is well underway in the rest of the economy.

Figure 4

Cycles in Investment

Since 1961, five distinct cycles can be seen in Figure 5 (measured from trough-to-trough). These cycles share a number of characteristics, which this paper will examine in more depth. They are long-lasting: the shortest on record since 1961 was almost 7 years, the longest 10 years, whereas cycles in housing and exports are visibly shorter. Capital spending accelerates steadily to a peak growth rate of over 15% after about 3 to 5 years of expansion, decelerates for 2 to 3 years, and then turns down for 1 to 3 years, which normally accompanies periods of recession and during severe contractions continues into the recovery phase for the rest of the economy.

Figure 5

Analysts have long been aware of the extended duration of investment cycles and have advanced several explanations. While a certain amount of investment is needed on an ongoing basis to offset normal wear-and-tear on capital, most investment is driven by the need to expand capacity or exploit investment opportunities brought about by either new methods of production or new markets (see Gordon, p. 27). New investments require long lead times to identify, arrange financing, order capital goods and put in place. But once undertaken, these large projects usually continue to completion, partly because of the need to get them operational to start repaying the investment: one rarely sees construction of a large office building halted half-way through. A more recent example was the Millennium oilsands project: even when oil prices crashed to $10 a barrel in 1998, investors decided to proceed.

The stamina of investment cycles is evident in both the volume of structures and machinery and equipment (Figure 6). This is somewhat surprising, as firms could more easily decide to start or stop spending on items such as office equipment or vehicles than on buildings or large engineering projects. This may reflect the importance of the long-range planning horizon in capital projects and that both types of investment complement each other: the oilsands, for example, require a great deal of preparation of the terrain (which is engineering work) and then huge investments in equipment to refine the crude bitumen.

Figure 6

The amplitude of cycles in both structures and machinery and equipment is also similar. In recessions, both fell at about the same rate, except for a much sharper drop in structures during the early 1990s. In four out of the five cycles, machinery and equipment peaked at a slightly higher growth rate than structures. Little difference was seen in the timing of recoveries or recessions, with spending on both categories generally changing direction at about the same time. A notable exception is the current upswing, which began earlier for machinery and equipment. On the heels of its resilience to the downturn in the early 1990s, this may represent an increasing immunity for machinery and equipment from investment downturns and more sensitivity to movements in the exchange rate.

Figure 7 shows the breakdown of investment in structures into its two components, buildings and engineering. The cyclical movement of spending on structures usually is driven by buildings, even though engineering is twice as large. Buildings dominate because movements are longer and more accentuated. Since 1981, there have been four long movements in building. Construction rose until 1991, and then slumped for five years. Expansion resumed from 1996 until 2000, after which it retreated until early 2005.

Engineering tends to follow its own cyclical pattern – especially after 1987 when it has alternated between about two years of growth and one year of decline. This partly reflects the ebb and flow of mega-projects such as the Hibernia offshore platform and the Confederation Bridge.3

Figure 7

Long cycles in building have been noted by analysts since early in the 20th century (Long, 1939, p. 385). The explanation partly lies in the accelerator principle, which relates investment during expansions to the change in income growth. Since income growth invariably slows before a downturn, building also typically turns down before a recession. Building then lags in recoveries, because plant utilization or office vacancy rates have to reach a certain level before new construction will begin. Related to this is the extreme durability of buildings, which can last decades. Another factor is the long lag in the construction process – EnCana’s new 60-story headquarters in Calgary announced in July 2005 will not be completed until 2009.

The cyclical ups and downs of spending on machinery and equipment are driven mostly by industrial machinery and transportation equipment (mostly cars and trucks). Office and telecommunications equipment rose steadily through most of the 1980s and 1990s, turning down only after the ICT bubble burst in 2001.

Figure 8

Once overall investment spending starts to gather pace, it typically accelerates steadily to a peak annual growth rate of around 15% after four years. The only two exceptions were in 1987 and 1996, when the upswings in business investment were briefly interrupted due to drops in structures (see Figure 5). Both these events reflected sharp cuts in spending on the oilpatch. The drop in 1987 followed the crash in oil prices to below $10 a barrel in 1986, while the 1996 slump reflected the winding down of the Hibernia project (as well, the Confederation Bridge to PEI was near completion, which is why all of the weakness was in the engineering component). But none of these events pulled down overall investment in the short-run, nor stopped the investment cycle from resuming its normal trajectory the following year. The investment cycle is remarkably resilient.

As noted by Gordon (p.28) investment booms typically “arise because a new composition of output may have become appropriate” or “changes in technology … call for a different composition of the capital stock.” Examples of the latter include the booms triggered by the adoption of railroads, automobiles, electric power and, most recently, ICT technologies. These cycles tend to persist for years as firms grapple with how best to integrate these technologies into their operations.

But while the ICT-driven investment boom of the late 1990s resembled earlier periods of adapting a new general purpose technology, the current cycle is the more classic need to change the composition of output. For most of the last two decades, firms have shifted away from the resource sector due to the secular decline in commodity prices. For example, the share of natural resources in employment fell from 7% in 1981 to less than 4%. Investment was so low in metal mining that it did not keep up with depreciation, leading to an outright drop in the capital stock in mining over the last two decades – the only industry to record a contraction.

The surge in commodity prices so far this decade has left firms scrambling to reverse the long-run slide in the primary sector. The coincidence of rising prices, burgeoning corporate financial surpluses and strong equity and bond markets has created a perfect storm of conditions to fuel an investment boom in the energy and mining sectors.

Recent Trends

The preceding analysis of investment cycles has some interesting implications for the current state of business investment. After its post-2000 downturn, business investment has been slowly but steadily picking up over the last two years. This growth has been driven by the boom in the resource sector, especially energy. Investment by the energy sector has risen from $30 billion in 2000 to a planned outlay of $46 billion at the start of 2005, largely due to exploration and development of oil and gas. This has offset continued weakness in areas like manufacturing.

Figure 9 compares the recent recovery of business investment to the average of the past five cycles. It shows that the current upturn is typical for the early stages of an upswing in the investment cycle, with annual growth approaching 8% in the second year of an upturn.

Figure 9

Will investment continue to accelerate to double-digit rates of growth as consistently occurred in previous cycles? The spate of new projects announced in the spring is encouraging. Non-residential building permits jumped to a record level in April. Contract awards also soared, driven by large engineering works in the oilsands such as the $10 billion Horizon project. Overall, the oil industry alone currently has tabled plans for over $64 billion of investment over the rest of the decade. And work has begun on a slew of projects in BC related to the mining boom, the need for more infrastructure to support trade with China and the 2010 Vancouver Olympics. Central Canada too has benefited, with Toyota’s plan to add a new plant in Ontario.

Figure 10

A sudden crash in oil prices could temporarily interrupt this rosy outlook, as occurred in 1987. But the odds of this are reduced by the steady rise in far future oil prices and the long-term planning horizon of these mega-projects.


The recent strengthening of business investment has been a major factor in sustaining GDP growth, helping to offset the slowdown in exports after the Canadian dollar rose. Based on past upturns and the recent surge in building permits, it is reasonable to expect investment to remain a driving force in growth for the foreseeable future. Another implication of this strengthening of business investment is that workers will have more capital to work with. Recent data has shown much of the post-2000 slowdown in labour productivity was due to the slack in business investment. When the capital stock starts to rise rapidly, labour productivity should begin to improve.


R.A. Gordon: “Investment Behavior and Business Cycles.” In The Review of Economics and Statistics, Vol. 37, no. 1 (Feb. 1955).

C.D. Long Jr: “Long Cycles in the Building Industry.” The Quarterly Journal of Economics, Vol. 53 no. 3 (May 1939).

V. Zarnowitz: “Theory and History Behind Business Cycles: Are the 1990s the Onset of a Golden Age?” in Journal of Economic Perspectives Vol. 10 no. 2 (Spring 1999).

Recent feature articles


* Current Analysis Division, (613) 951-9162.
1 Statistically, the standard deviation of year-over-year growth in real business investment since 1961 is 8.1, comparable to the 10.1 for housing but much more than the 2.0 for consumer spending.
2 Investment by governments is excluded, partly because it often has a counter-cyclical pattern, rising when business investment fell in 1982 and 2002 and falling during the investment surge of the late 1990s.
3 The lumpy nature of spending on projects at the firm level is reflected in aggregate investment. This was noted by research into US manufacturing investment, which found that “investment spikes account for a large fraction of the investment of [individual] plants. Furthermore, they show that the number of plants undergoing primary spikes exhibits strong positive correlation with aggregate investment cycles.” See R. Caballero, E. Engel, and J. Haltiwanger: “Plant Level Adjustments and Aggregate Investment Dynamics.” Brookings Papers on Economic Activity, 2: 1995, p. 3.

Home | Search | Contact Us | Français Return to top of page
Date Modified: 2008-11-21 Important Notices
Contents Tables Feature article Economic events Current economic conditions Charts User information PDF version