Canada
Selected Issues

This Selected Issues paper on Canada examines economic development and policies. Capital ratios before the crisis have been a key determinant of bank performance during the turmoil; and Canadian banks had ample capital. Specifically, most banks with critically low capital at end-2006 later experienced dramatic equity value declines, and many had to be rescued. Regulatory and structural factors contributed to the resilience of Canadian banks by reducing their incentives to take risks.

Abstract

This Selected Issues paper on Canada examines economic development and policies. Capital ratios before the crisis have been a key determinant of bank performance during the turmoil; and Canadian banks had ample capital. Specifically, most banks with critically low capital at end-2006 later experienced dramatic equity value declines, and many had to be rescued. Regulatory and structural factors contributed to the resilience of Canadian banks by reducing their incentives to take risks.

I. Canadian Banks And The Credit Turmoil 1

1. Canadian banks have been relatively resilient to the ongoing credit turmoil. The effect of the turmoil on the Canadian financial system is, so far, milder than in other G7 economies. Interbank money markets remained functional. No injections of public capital into banks were necessary. The resilience appears particularly striking given the close economic and financial links between Canada and the United States.

2. To shed light on this resilience, this chapter searches for key variables explaining Canadian bank performance during the crisis. It considers a sample of large OECD banks and studies how pre-crisis balance-sheet structure affected bank performance during the crisis. The sample includes all large OECD commercial banks (72 institutions with assets in excess of 100 billion euros at the end of 2006). We consider three main fundamentals: the equity-to-assets ratio, the balance-sheet liquidity-to-debt liabilities ratio, and the depository-funding-to-assets ratio. Measures of performance are: equity price decline from January 2007 to January 2009, two binary variables for particularly large equity declines (greater than 70 or 85 percent), and two binary variables for government intervention (undertaken to alleviate significant financial distress).

3. Capital ratios before the crisis were a key determinant of bank performance during the turmoil; and Canadian banks had ample capital. Specifically, most banks with critically low capital at end-2006 later experienced dramatic equity value declines, and many had to be rescued (Table 1). Prior to the crisis, all Canadian banks had capital ratios (equity over assets, a leverage measure) above 4 percent, which has assured their resilience to asset shocks. Interestingly, a large number of currently distressed U.S. banks had relatively high pre-crisis capital, which was nevertheless quickly exhausted through troubled asset exposures and (in some instances) problematic acquisitions.

Table 1:

Bank Fundamentals and Performance during Turmoil

article image
Sources: BankScope and staff calculations.

Equity over total assets

Liquid assets over total debt liabilities

Depository funding over total assets

4. Compared to OECD peers, Canadian banks had slightly above-average balance sheet liquidity. Buffers of highly liquid assets allow banks to bridge temporary cash flow shortfalls, which proved critical during the rush for liquidity. Interestingly, some U.S. banks were shown to have particularly low measures of high-quality liquidity, as they were using assets such as tradable mortgage-backed securities as part of their liquidity buffers.

5. During a liquidity crisis, access to stable funding is key to survival; Canadian banks had a high ratio of retail to wholesale deposits. Retail deposits are insured and hence “sticky,” and provide a stable source of long-term funds for banks. In contrast, wholesale funds can withdraw rapidly upon minor negative news, and were a major source of vulnerability during the turmoil. The relative abundance of retail deposits seems to have been key for the resilience of Canadian banks.2 Easy access to retail deposits in Canada is explained by limited competition for household savings from non-banks, and by restrained bank asset growth that limited overall demand for funding.

6. Multivariate regression analysis confirms and extends these preliminary findings (Table 2). A high share of depository funding and a capital ratio above a critical minimum (although not the capital ratio per se) appear to be the most significant and robust determinants of bank resilience during the turmoil. Balance-sheet liquidity is less robust: it correlates with major failures, but not equity value declines (except very large ones). By considering interactions, the regressions also identify substitution between bank funding structure and capital, where, for a given risk profile, a bank with more depository funding can operate with lower capital, and a bank with higher capital can use less depository funding. In addition, larger banks have a higher probability of government intervention. However, rapid balance sheet expansion before the crisis appears irrelevant for performance during the crisis.

Table 2:

Bank Fundamentals and Performance: Multivariate Regression Results

article image
Source: Staff calculations.Notes: A sample of large OECD banks (assets above 100 billion euro at end-2006). The dependent variables are three dummy variables for imminent bank failure, stock market value decline by >85%, and stock market decine by >70%, as well as the absolute value of stock market decline (stock decline measured January 2007 to January 2009). Capital ratio is the equity-to-assets ratio; balance sheet liquidity is the liquid assets-to-debt-liabilities ratio; depository funding is the deposits-to-assets ratio; all taken at end-2006. The regressions involving dummy variables are estimated based on Probit, with coefficients transformed to be interpreted as probability change (0 to 1). T-statistics are reported in parenthesis. ***, **, * indicate statistical significance at the 1%, 5%, and 10% level.

7. Regulatory and structural factors contributed to the resilience of Canadian banks by reducing their incentives to take risks. Canadian capital requirements are significantly more stringent than Basel minima (national targets of 7 percent for tier 1 capital and 10 percent for total capital, versus 4 and 8 percent prescribed by the Basel Accord). Banks are also subject to a maximum assets-to-total-capital multiple of 20 (corresponding to a leverage ratio of 5 percent). Besides providing an enhanced cushion, stringent capital requirements have beneficial incentive effects: they impede rapid balance sheet growth, restrict wholesale activities, and limit foreign expansion to niches where banks have clear competitive advantage not related to low cost of capital. Notable structural factors in Canada include high franchise values, a mortgage market characterized by prudent underwriting, and an overall prudent and conservative culture in the financial sector. Limited exposure to U.S. assets was a key additional factor behind the resilience of Canadian banks to the crisis.

Conclusions and policy implications

8. Canadian banks appear well positioned to weather the turmoil. A combination of strong capital and robust funding, in the context of sound regulation and supervision, has lent resilience to the banking system. With a severe recession underway, credit losses are likely to continue to climb, particularly on exposures to highly-leveraged households. But with banks stable and macroeconomic policies supportive, financial instability appears to be a tail risk.

II. Canadian Residential Mortgage Markets: Boring But Effective?1

1. Canada’s financial system has often been criticized for being “too conservative” or “not dynamic enough”. Indeed, when compared to the United States, Canadian banks seem to offer fewer loan options, in particular in the residential mortgage area. This could mean that households are underserved and that there is wide room for welfare improvements via increased financial innovation. However, Klyuev (2008) concluded that this is not the case and housing finance is highly advanced and sophisticated in Canada. Nevertheless, the same paper finds thatfinancing options were somewhat limited, particularly at terms longer than five years.

2. This chapter concurs with previous research documenting the sophistication of Canada’s financial system, but suggests that regulations have limited the supply of some products. In particular, the paucity of longer-term loans is caused by a five-year maturity cap on government-guaranteed deposit insurance, and a prepayment penalty limit on residential mortgage loans in the Interest Act. The chapter also argues that for prime borrowers, the availability and cost of residential mortgages are comparable to those in the United States.

3. The Canadian predominance of shorter terms is driven by the more important role (versus in the United States) of retail deposits to fund mortgages, a feature driven by regulation. Deposits longer than five years are not popular because Canadian Deposit Insurance Corporation (CDIC) guarantees do not cover longer terms. Hence, Canadian banks have no natural funding for cost effective longer-term mortgages.

4. Also, regulations cause lenders to pass on the higher cost of hedging prepayment risk for longer mortgages in the form of higher interest rates (Figure 1). Section 10 of Canada’s Interest Act effectively fixes the prepayment penalty for most mortgages with a term to maturity greater than five years at three-months of interest, which is likely less than the penalty charged during the first five years of mortgage terms. Offsetting this to some degree is the portability of Canadian mortgages.2

Figure 1:
Figure 1:

Canadian Bank Mortgage Rates (February 20, 2009)

(percent)

Citation: IMF Staff Country Reports 2009, 163; 10.5089/9781451807097.002.A001

5. Despite different regulations, mortgage costs are broadly similar in the United States and Canada. Even though at first sight mortgage rates are higher in Canada than in the United States, “posted” rates overstate actual transacted rates in Canada. Canadian five-year conventional rates have averaged about 100 basis points above the U.S. thirty-year conforming rate (Figure 2, in which both rates are normalized by their respective interest rate swap comparators).3 However, the Canadian rates are “posted” rates that overstate actual transacted rates, typically by more than 100 basis points. The Canadian Association of Accredited Mortgage Professionals (CAAMP) estimates that, on average, recent posted rates have exceeded transacted rates by 159 basis points (CAAMP, 2008).

Figure 2:
Figure 2:

Residential Mortgage versus Interest Rate Swap Rates

(basis points)

Citation: IMF Staff Country Reports 2009, 163; 10.5089/9781451807097.002.A001

6. In addition, the apparently lower U.S. thirty-year conforming rates reflect the payment of upfront points, which effectively prepay interest. For example, on February 19, 2009, the posted conforming rate was 5.04 percent with 0.7 points upfront, which is equivalent to 5.34 percent (plus 30 bps) with zero points. Moreover, there is an incremental term premium embedded in U.S. rates, which reflect the longer term of U.S. loans.

7. Anecdotal evidence suggests that the non-interest costs of originating and refinancing mortgage loans is significantly cheaper in Canada, which would also contribute to closing any apparent gap between the costs in both countries. Canadian borrowers pay about C$2,000 in upfront fees and taxes for a new loan, and on a refinancing about C$1,000 plus a prepayment penalty of about C$3,000 on the old mortgage.4 On the same loans (new loans and refinancings), U.S. borrowers pay origination fees of $1,000 to $3,000, plus about $1,000 of costs and fees, and local government taxes of about $1,000.

8. Payment affordability criteria for prime borrowers are broadly similar in both countries. For example, in order to qualify for mortgage insurance in Canada, gross debt service should usually not exceed 32 percent of gross household income, and total debt service cost should usually not exceed 40 percent (versus 28 and 36 percent to qualify for Fannie Mae and Freddie Mac insurance).

9. However, the approval criteria for adjustable-rate loans in Canada are usually based on the three-year fixed-term rate, which is usually the highest fixed rate inside of the five-year term, whereas U.S. practice is to use the current floating rate.5 Canada also has a small “Alt-A” market aimed mainly at self-employed people who have difficulty documenting their stated income.

10. Down payment requirements are roughly in line with those in the United States. Canadian federally-regulated deposit-taking institutions have been able to underwrite insured mortgages with loan-to-value ratios as high as 95 percent since 1992, and occasionally before then.

11. There are no limits to the size of individual loans that the Canada Mortgage and Housing Corporation (CMHC) and other mortgage insurers will insure, which minimize risks to banks’ balance sheets. In the United States, Fannie Mae and Freddie Mac insurance is only available on loans up to the “conforming limit”, which vary by geographic areas, but in 2009 is $625,500 for loans on single-family homes in “high-cost” areas.

Conclusions and Policy Implications

12. The availability and costs of Canadian residential mortgage loans to prime borrowers are comparable to those in the United States. Moreover, even though there are clear institutional differences, homeownership in both countries is virtually identical at about 68 percent of all households. This said, some aspects of Canada’s mortgage market can be improved. To encourage the development of longer-term mortgage markets, the government might consider dropping the five-year cap on CDIC deposit insurance and the five-year prepayment penalty fixed in the Interest Act. Until that happens, rates on fixed-term residential mortgages beyond the five year term will remain uneconomical for most borrowers. The opening up of longer fixed-rate terms would help households to better manage financial risks.

III. The Impact of Global Shocks On Canada—What Do Macro-Financial Linkages Tell Us?1

1. The paper analyzes how Canada’s strong macro-financial ties with the United States affect its real activity. It builds on related literature,2 by focusing on the role of financial conditions in driving Canadian growth. For instance, in the current global crisis, Canada has been affected not only by declining external demand but also by global financial shocks, the latter inducing domestic financial strains. Notably, credit conditions have remained tight notwithstanding aggressive monetary policy rate cuts. This paper sheds light on how these external and domestic shocks bear on real GDP growth.

2. These issues are assessed by building a macro-financial conditions index for Canada using a Bayesian VAR (BVAR) estimation model, which uses “informative priors” about the steady-state values of the variables to limit loss in estimation precision caused by the heavy parameterization of VARs.3 This is especially important given the short sample size, as data on Canadian banks’ lending standards begin in 1999. The model is then used for scenario analyses—i.e., projecting Canada’s growth outlook for alternative external and domestic conditions—and building a macro-financial index for growth. The “macro” part of the index accounts for the impact of external demand conditions, while the “financial” part accounts for external and domestic financial conditions.

3. Banks’ non-price lending standards in the United States and Canada are used to measure credit availability for Canadian businesses. Canadian lending standards (from the Bank of Canada’s Senior Loan Officer’s Survey (SLOS)) are strongly correlated with U.S. SLOS for large corporations (Figure 1). Both measures are negatively correlated with Canada’s contemporaneous and future real GDP growth, which is not surprising given that Canadian businesses raise one-quarter of their funding in the United States.4 For instance, the deceleration in Canadian growth in the early 2000s and more recently coincided with a sharp tightening of one or both measures of lending standards.

Figure 1.
Figure 1.

Banks’ Lending Standards for Corporations and Real GDP Growth

Citation: IMF Staff Country Reports 2009, 163; 10.5089/9781451807097.002.A001

Source: Haver Analytics.

4. The baseline model comprises nine macro-financial variables: (i) U.S. real GDP growth (expressed as year-over-year percent change), (ii) oil price growth, (iii) U.S. three-month t-bill rate, as a proxy for U.S. monetary conditions, (iv) U.S. SLOS, (v) U.S. high-yield spread over the 10-year treasury yield, (vi) Canada’s real GDP growth, (vii) Canadian SLOS, (viii) Canada’s three-month t-bill rate, and (ix) growth in Canada’s real effective exchange rate. The U.S. variables are allowed to be exogenous with respect to the Canadian variables. The data cover the period 1999Q2–2008Q4 at a quarterly frequency. While the steady-state priors were drawn from existing literature or historical averages, the posterior values from the model are close to the priors, implying that the chosen priors are reasonable.

5. Impulse-response functions confirm that financial shocks have a significant effect on growth (Figure 2). A one-percent standard deviation (s.d) shock to the U.S. SLOS—a net tightening of 9.5 percentage points (pp)—reduces Canadian growth by about 0.4 pp in six quarters, while the impact of a similar tightening in the Canadian SLOS is weaker but works with a shorter lag—reducing growth by ¼ pp in three quarters. The larger effect on growth of a tightening in the US SLOS than the Canada SLOS could reflect indirect effects of a higher U.S. SLOS on Canada—tighter U.S. lending standards slow U.S. growth (and hence external demand) while also spilling over to Canadian credit conditions. Also, a one s.d. shock to the U.S. high-yield spread (129 basis points) shaves off 0.15 pp from growth over three quarters.

Figure 2.
Figure 2.

Impulse Response Functions

(Impact on Canada’s real GDP growth of one standard deviation shocks to financial indicators)

Citation: IMF Staff Country Reports 2009, 163; 10.5089/9781451807097.002.A001

6. Looking ahead, Canada’s economic outlook would be firmly tied to prospects for external activity as well as also domestic financial conditions. Conditional forecasts show that if U.S. growth in 2009 were 1 pp lower than that projected by the unconditionalmodel, then Canada’s growth would also be 1 pp lower (Figure 3), while if Canadian SLOS were 10 pp tighter in 2009, then Canada’s growth would be 0.4 pp lower (Figure 4).

Figure 3.
Figure 3.

Canadian Real GDP Forecast for 2009 Under Alternative paths for US Real GDP Growth

(percent)

Citation: IMF Staff Country Reports 2009, 163; 10.5089/9781451807097.002.A001

1/ The path for US real GDP is assumed to be 1 percentage point lower than the unconditional forecast.
Figure 4.
Figure 4.

Canadian Real GDP Forecast Under Alternative Conditions for Canadian SLOS

Citation: IMF Staff Country Reports 2009, 163; 10.5089/9781451807097.002.A001

1/ The path for Canadian SLOS is assumed to be 10 percentage point tighter than the unconditional forecast.

7. The variance decompositions for Canada’s real GDP growth confirm that foreign shocks are the most important source of variation in Canada’s growth over the long run, with U.S. growth accounting for close to 42 percent, U.S. financial shocks another 28 percent, and oil prices 14 percent (Figure 5). The contribution of domestic financial conditions to Canadian growth increases from 0 percent in the short-term to over 8 percent in 12 quarters. Oddly enough, the contribution of Canadian growth to its own variance declines from 43 percent to a little over 7 percent in the long run—this could reflect the fact that the model is based on a recent sample period (since 1999), when the openness of the Canadian economy to external volatilities, especially vis-à-vis the United States has increased markedly. Indeed, a simple monetary BVAR model estimated starting in the early 1990s (which excludes data on financial market indicators, i.e., lending standards and high yield spreads) attributes a larger role to domestic growth shocks.

Figure 5.
Figure 5.

Variance Decomposition of Canadian Real GDP Growth

(percent)

Citation: IMF Staff Country Reports 2009, 163; 10.5089/9781451807097.002.A001

8. A macro-financial condition index built from the coefficients of the baseline model tracks real GDP growth well, and shows that tightening in the Canadian SLOS and effects of past real appreciation have played a key role in the deceleration of the Canadian growth rate in the run up to the recent crisis. However, U.S. economic and financial conditions will increasingly bear on growth in the near term.

Figure 6.
Figure 6.

Contribution to the Macro-Financial Conditions Index

Citation: IMF Staff Country Reports 2009, 163; 10.5089/9781451807097.002.A001

Conclusions and policy implications

9. The paper underscores the need for a normalization of U.S. and domestic financial strains for Canada’s recovery. Tight U.S. and domestic financial conditions depress Canada’s growth by intensifying credit strains for businesses, although monetary easing helps ameliorate these strains somewhat. Thus, stability in U.S. financial conditions would be critical for a sustained pick up in Canadian economic activity.

IV. Is The Canadian Housing Market Overvalued? A Tale of Two Regions1

1. Canadian house prices have increased significantly between 2003 and early 2008, with a marked downward trend since mid-2008 (Figure 1). House prices rose by around 60 percent in nominal terms (45 percent in real terms) from 2003 to the peak, before falling by around 10 percent (both in real and nominal terms) in the latter part of 2008. The decline is particularly acute in the west given the collapse in commodity prices, although modest declines are occuring elsewhere.

Figure 1.
Figure 1.

Real House Prices

(1990=100, NSA)

Citation: IMF Staff Country Reports 2009, 163; 10.5089/9781451807097.002.A001

Source: CREA, Finance Canada, and Haver Analytics.

2. This chapter summarizes estimates of the gap between actual house prices and their equilibrium levels for five large Canadian provinces (Alberta, British Columbia, Ontario, Quebec, and Saskatchewan). The price measure used is the existing home price from the Canadian Real Estate Association’s Multiple Listing Service (MLS) database, and is deflated by each province’s CPI.2 We examine current valuations against economic fundamentals using quarterly regional data—such as disposable income, demographic developments, and mortgage credit. The analysis is based on an error correction model, which combines the long-run, cointegrating relationship among the levels variables and the short-run relationships among the first differences of the variables.

3. The error correction model postulates that the growth rate of real houses is explained by a combination of the following factors (depending on the province considered) (Table 1):3

  • Past growth rates of real house prices. For most provinces, we find that the current growth rate is positively correlated with the past growth rate.

  • Reversion to fundamentals implied by the long-run equation. We find that only for Ontario, the growth rate of house prices shows long-run reversion to the equilibrium prices derived from the model, implying that prices would tend to fall when they are out of line with fundamentals.

  • Economic fundamentals. For most provinces, we find that the growth rate of house prices is positively affected by (per capita) real income growth—as this increases households’ purchasing power and borrowing capacity—and positively affected by mortgage credit growth (higher rates indicate that households are less credit rationed), and population growth (as a proxy for the growth rate of households).

Table 1:

Determinants of House Prices in Selected Canadian Provinces

(Summary of Em pirical Results,1992-2008) Dependent Variable: Real House Price (growth)

article image
Source: Staff estimates.Note: T-statistics are listed within brackets.

Conclusions and policy implications

4. Results reveal that prices in the west are above the levels implied by the model, although prices remain close to or slightly below equilibrium in the east (Figure 2). Indeed, the econometric model indicates that most of the recent surge in Canadian home prices, even in the west, reflects a catch-up from stubbornly undervalued levels following the housing collapse in the early 1990s, rather than a housing bubble per se.4 While resource-rich western provinces continue to have house prices above the model prediction, their prices have diminished significantly in the last year. In contrast, Quebec and Ontario appear to be close to equilibrium, or slightly below the prices implied by the model, indicating a divergence between western and eastern provinces in house price dynamics.

Figure 2.
Figure 2.

Canada: House Price Over/Undervaluation

(In percent)

Citation: IMF Staff Country Reports 2009, 163; 10.5089/9781451807097.002.A001

Sources: Finance Canada, CREA, Fund staff calculations.

5. While a crash in the national housing market appears unlikely, a correction in western housing markets could have national implications. With the west accounting for 35 percent of Canada’s GDP and around 30 percent of Canada’s labor force, an abrupt correction in its housing market, could cause adverse spillovers to the rest of Canada; with housing assets and mortgage debt at record ratios of disposable income, household balance sheets are particularly exposed to house price dynamics. Similarly, given that most mortgages are originated by banks (55 percent of which do not carry mortgage insurance but have a loan-to-value ratio below 80 percent),5 this could also somewhat impact the banking sector, affecting the future provision of mortgages and credit in general, imposing additional downward pressure on spending, incomes, wealth and thus house prices. Last but not least, house prices directly impact headline and core inflation, thus affecting inflationary expectations as well.

V. How Do Commodity Prices Affect Economic Slumps In Commodity Exporters? 1

1. The financial crisis has been accompanied by a sharp fall in commodity prices, with deleterious effects on commodity exporters. Commodity exporters, including Canada, have been subject to twin shocks during the ongoing crisis—on external demand and on commodity prices (Figure 1). The latter could affect the duration and severity of Canada’s ongoing recession, but current research sheds scant light on the relationship between commodity prices and business cycle characteristics in commodity exporting countries.2

Figure 1.
Figure 1.

Commodity Price Index

(2005 = 100) 1/

Citation: IMF Staff Country Reports 2009, 163; 10.5089/9781451807097.002.A001

Source, World Economic Outlook, Statistics Canada1/ World price expressed as simple average of the following five categories: food, agriculture, metals, gold, and crude oil.

2. This chapter analyzes whether the length and depth of recessions in commodity exporters is affected by the commodity-price cycle. Commodity prices can have offsetting effects on business cycles of commodity exporters. On the one hand, commodity booms raise real incomes and support domestic demand. On the other hand, an appreciation of the real exchange rate in response to the upturn in commodity prices can affect external competitiveness and stifle real exports, depressing external demand. Both of these effects were experienced by Canada through the first half of 2008. This analysis takes into account these offsetting effects.

3. The sample comprises 15 advanced and emerging market countries with varying degrees of dependence on commodity exports (Table 1). While commodities comprise close to 10 percent of total exports in most of these countries, ten are net commodity exporters—Australia, Brazil, Canada, Chile, Denmark, Mexico, New Zealand, Norway, Russia, and South Africa. The other five—Netherlands, Sweden, Switzerland, the United Kingdom, and the United States—form a control group of net commodity importers. In the absence of data for country-specific commodity price indices for most of the sample countries, commodity export prices are constructed as weighted averages of the global commodity price indices (food, agriculture, crude oil, gold, and metals), where the weights are given by the export share of each commodity in the total commodity exports of each country (2002–04 averages), provided by the IMF’s World Economic Outlook.

Table 1:

Sample of Countries

article image
Source: International Monetary Fund (WEO)

4.The sample includes 120 spells of economic downturns during 1981–2008 with considerable differences in experiences across the countries. An economic downturn is defined as the period from peak to trough in (year-on-year) real GDP growth. The duration of the downturn is measured in quarters, and a spell is considered only if it lasted at least three quarters. The data do not reveal any particular trend in the nature of downturns (Figure 2)—while the average length of a spell was about 5½ quarters, some commodity exporters experienced longer recessions averaging 6½ quarters (New Zealand, Norway, South Africa). At the same time, Sweden, a net commodity importer, experienced the longest downturns on average (7 quarters), while Brazil and the United States experienced the shortest (4.2 quarters). For commodity exporters, almost all spells were associated with a fall in commodity prices—either in the run up or during the downturn—signaling their importance in explaining business-cycle fluctuations in these economies.

Figure 2.
Figure 2.

Average length of Economic Downturns

(in quarters)

Citation: IMF Staff Country Reports 2009, 163; 10.5089/9781451807097.002.A001

5.Estimation results confirm that large declines in commodity prices significantly exacerbate ensuing economic downturns in commodity exporters. A fall in pre-crisis commodity prices induces longer and sharper recessions; the latter effect measured by the difference between peak and trough real GDP growth rates.

  • Recessions are relatively longer for commodity exporters that experience a pre-crisis fall in commodity prices (Table 2. Column A). In particular, a 10 percent decline in commodity prices prior to the downturn increases the length of the recession spell by 3/5th of a quarter, on average. However, this result is statistically insignificant for the overall sample including commodity importers.

  • Commodity prices also affect the severity of the ensuing recession, whereby a 10 percent decline in pre-crisis commodity prices reduces real GDP growth at the trough by 0.4 percentage point, a result that is significant for both the overall and restricted samples (Table 2, Column B).

Table 2.

Fixed Effect Regression of Length and Depth of Economic Slump 1/

article image

Economic Slump is length of spell from peak to trough. Explanatory variables represent levels in the quarter before the start of the slump. Growth rates are expressed in year-on-year percentage changes.

* and ** refer to significance at 5 percent and 10 percent levels (p-value in parentheses).

Includes net commodity exporters only.

6. The results also show that longer and deeper downturns are positively associated with the pre-crisis real GDP growth. The higher the real GDP growth before the crisis, the longer it takes for the ensuing recession to “bottom out” and the lower is real growth at the bottom. The other results are not significant—while the regression also controls for the impact of the real exchange rate, the latter does not have a significant effect on influencing duration or depth of the economic downturn.

Conclusions and policy implications

7. Commodities play a key role in affecting the severity and length of economic downturns in commodity exporters. The possible ameliorating effect of falling commodity prices on the real exchange rate and exports is likely more than offset by the negative effect on domestic incomes and activity. On average, a 10 percent decline in commodity prices (y-o-y) prolongs the ensuing downturn by 0.6 quarters and deepens it by 0.4 percentage point in real GDP growth. While these results do not indicate how much worse it could be for countries that continue to experience falling commodity prices after the downturn is underway,3 they stress that a sustained rebound in commodity prices is crucial for Canada’s economic prospects.

Canada: Selected Issues
Author: International Monetary Fund