Abstract
This paper explores the factors that have led to a Canada-U.S. productivity gap using a sectoral growth accounting approach. Both fiscal and monetary policies have had significant effects on the saving rate. The Canadian dollar’s appreciation was followed by a protracted period of exchange rate weakness. This paper reviews the institutional aspects of Canada’s real return bond program. The Canadian system provides a successful model for pension reform. Free trade has helped promote the integration of U.S. and Canadian economies, but significant differences remain.
III. The Canadian Dollar: Back to Fundamentals?1
1. The Canadian dollar’s appreciation during 2003 followed a protracted period of exchange rate weakness. In particular, from late 1991 to early 2002, the Canadian dollar fell by almost 35 percent in real effective terms and reached historical lows against the U.S. dollar, despite the economy’s favorable fiscal and growth performance. The more recent appreciation of the Canadian currency—while unusually rapid—has only recouped roughly half of this earlier depreciation, which still leaves the dollar about 10 percent below its 1980–1991 average in real effective terms. Nonetheless, the recent appreciation has raised concern regarding the impact on the competitiveness of Canadian firms and short-term growth prospects.
2. These developments have underscored long-standing questions about where the Canadian dollar stands relative to underlying economic fundamentals. The analysis below suggests that while the dollar’s slide during the 1990s was broadly consistent with a number of fundamentals—including commodity prices and productivity differentials—the dollar had fallen to levels that appeared somewhat undervalued. Estimates based on a modified purchasing power parity (PPP) framework suggest that the 2003 appreciation has brought the exchange rate more closely in line with fundamentals.
A. Recent Developments
3. In 2003, the Canadian dollar appreciated sharply against the U.S. dollar and other major currencies. The dollar rose to US$0.787 in early January 2004, a 25 percent increase since late 2002 (Figure 1). Within the space of only a year, the Canadian dollar has thus recouped almost 60 percent of a decade’s worth of cumulative losses against the U.S. dollar.2 These recent gains partly reflected the U.S. dollar’s broader weakness against other major currencies (Figure 2). However, the Canadian dollar also increased in value relative to other major currencies during 2003 (with the exception of the euro), which translated into a 15 percent appreciation in real effective terms since early 2002.
Real Effective Exchange Rate
Citation: IMF Staff Country Reports 2004, 060; 10.5089/9781451806908.002.A003
Sources: Haver Analytics; and IMF’s Information Notice System.Real Effective Exchange Rate
Citation: IMF Staff Country Reports 2004, 060; 10.5089/9781451806908.002.A003
Sources: Haver Analytics; and IMF’s Information Notice System.Real Effective Exchange Rate
Citation: IMF Staff Country Reports 2004, 060; 10.5089/9781451806908.002.A003
Sources: Haver Analytics; and IMF’s Information Notice System.
Nominal Exchange Rates
Citation: IMF Staff Country Reports 2004, 060; 10.5089/9781451806908.002.A003
Sources: Haver Analytics; and IMF’s Information Notice System.Nominal Exchange Rates
Citation: IMF Staff Country Reports 2004, 060; 10.5089/9781451806908.002.A003
Sources: Haver Analytics; and IMF’s Information Notice System.Nominal Exchange Rates
Citation: IMF Staff Country Reports 2004, 060; 10.5089/9781451806908.002.A003
Sources: Haver Analytics; and IMF’s Information Notice System.4. The Canadian dollar’s weakness during the 1990s has been partly ascribed to the decline in global commodity prices (Figure 3). Empirical specifications have generally identified a strong short- to medium-run relationship between non-energy commodity prices and the dollar (Cerisola, Swagel, and Keenan, 1999; Gauthier and Tessier, 2002). The Canadian dollar is often termed a “commodity currency” along with the Australian and New Zealand dollars. This view has held despite evidence by Chen and Rogoff (2003) suggesting that the relationship between the Canadian dollar and commodity prices is not as strong as for the other two countries, and the fact that the share of commodities in Canadian exports has dropped from 45 percent in 1970 to about 30 percent in 2002.
Real U.S./Canada Exchange Rate and Selected Variables, 1980 - 2003 1/
Citation: IMF Staff Country Reports 2004, 060; 10.5089/9781451806908.002.A003
Sources: Haver Analytics; and Fund staff calculations.1/ An increase in the RER implies an appreciation of the Canadian dollar against the U.S. dollar in real terms.Real U.S./Canada Exchange Rate and Selected Variables, 1980 - 2003 1/
Citation: IMF Staff Country Reports 2004, 060; 10.5089/9781451806908.002.A003
Sources: Haver Analytics; and Fund staff calculations.1/ An increase in the RER implies an appreciation of the Canadian dollar against the U.S. dollar in real terms.Real U.S./Canada Exchange Rate and Selected Variables, 1980 - 2003 1/
Citation: IMF Staff Country Reports 2004, 060; 10.5089/9781451806908.002.A003
Sources: Haver Analytics; and Fund staff calculations.1/ An increase in the RER implies an appreciation of the Canadian dollar against the U.S. dollar in real terms.5. Movements in interest rate differentials and net foreign liabilities have also favored a weaker exchange rate in recent years. Fiscal consolidation, among other factors, contributed to a decline in long-term interest rates in Canada relative to the United States since the mid-1990s, which may have dampened the demand for Canadian securities by foreign investors. Moreover, a substantial increase in Canada’s net foreign liabilities in the early 1990s, which led to higher external debt service costs for most of the 1990s, has only been reversed in recent years.
6. In addition, productivity differentials relative to the United States appear to have contributed to a long-term downward trend in the Canadian dollar. The studies cited above uniformly identify the U.S./Canada productivity gap as a key factor in explaining the downward trend in the Canadian dollar over the past decade. Indeed, this productivity differential has meant that—despite the depreciation of the nominal exchange rate—the real exchange rate in the machinery and equipment manufacturing sector appreciated vis-à-vis U.S. competitors (Figure 4).3 By contrast, higher productivity growth in the Canadian service sector has led to a depreciation of the sectoral real exchange rate in that sector.
Canada vs. United States: Change in Sectoral Real Exchange Rates, 1990 - 2001 1/
Citation: IMF Staff Country Reports 2004, 060; 10.5089/9781451806908.002.A003
Sources: OECD’s Structural Analysis Database; and Fund staff calculations.1/ An increase in the RER implies an appreciation of the Canadian dollar against the U.S. dollar in real terms.Canada vs. United States: Change in Sectoral Real Exchange Rates, 1990 - 2001 1/
Citation: IMF Staff Country Reports 2004, 060; 10.5089/9781451806908.002.A003
Sources: OECD’s Structural Analysis Database; and Fund staff calculations.1/ An increase in the RER implies an appreciation of the Canadian dollar against the U.S. dollar in real terms.Canada vs. United States: Change in Sectoral Real Exchange Rates, 1990 - 2001 1/
Citation: IMF Staff Country Reports 2004, 060; 10.5089/9781451806908.002.A003
Sources: OECD’s Structural Analysis Database; and Fund staff calculations.1/ An increase in the RER implies an appreciation of the Canadian dollar against the U.S. dollar in real terms.B. Modeling Exchange Rate Fundamentals
7. To assess the contribution of economic fundamentals to the Canadian exchange rate, this paper turns to a model that incorporates recent refinements of the traditional Balassa-Samuelson approach. The model is first cast in terms of the standard decomposition of the real exchange rate:
where p (p*) denotes the log of domestic (foreign) aggregate price level, pM (pM*) denotes the log of domestic (foreign) traded goods price, and e denotes the log of the nominal exchange rate. The first bracketed term
8. The model traditionally assumes that nontraded goods are not subject to international arbitrage, and that strict PPP applies to traded goods. In that case, the second term in the equation vanishes, and the long-term exchange rate is determined only by the relative price between traded and nontraded goods. The traded goods sector is typically characterized by higher productivity growth, which leads to a decline in production costs of traded goods relative to those of nontraded goods. In this framework, productivity gains cause the relative price of nontraded goods to rise, leading to an appreciation of the real exchange rate.
9. The traditional model has failed to hold up empirically, however. Work by Rogoff (1996) and Engel (1999) finds that PPP fails to hold even for traded goods (i.e., the second term in the equation does not vanish) and that, in fact, changes in relative prices between traded goods account for the bulk of exchange rate fluctuations. These results led to work that augmented the traditional model with factors explaining changes in relative prices between traded goods.
10. More recent approaches have sought to augment the traditional PPP model to include explanations for price differentials among tradable goods.4 They relax the assumption that strict PPP holds for tradables, and allow for the fact that traded goods are imperfect substitutes, reflecting both physical characteristics and asymmetry of tastes. As the result, the relative price between traded goods—the second term in equation (1)—is affected by economic fundamentals that affect relative supplies and demand. In the framework put forward by Bayoumi, Faruqee, and Lee (2003), relative prices between traded goods are driven by three factors:
Relative supply. An increase in the relative productivity of the domestic traded goods sector contributes to a depreciation of the real exchange rate, as the resulting increase in the relative supply of traded goods can only be absorbed through a decline in relative traded goods prices. This effect, which is proxied empirically by the difference between real manufacturing output at home and abroad (RYM), is separate from (and of opposite sign to) the Balassa-Samuelson effect based on the productivity differential between the traded and nontraded sector.
Commodity prices. A long-term increase in net commodity export revenues boosts the purchasing power of a country, thereby increasing the country’s demand for the traded goods it produces.5 Under imperfect substitutability between traded goods, this demand increase leads to an appreciation in the real exchange rate, owing to a rise in the price of traded goods produced by the country itself. Commodity effects are proxied by the terms of trade (TOT), here calculated as the difference between prices of commodity exports and imports, scaled by the ratios of commodity exports and imports to manufacturing imports.
Net foreign assets. Similarly, an increase in the level of net foreign assets (and thus net income flows on foreign assets) will tend to lead to real exchange rate appreciation. In the estimation, net income on foreign assets (NFA) is captured by multiplying foreign assets and liabilities with the average rate of return for each country. The difference between receipts and payments, scaled to manufacturing imports, is the measure used in the estimation.
11. This conceptual framework leads to an exchange rate equation that includes the Balassa-Samuelson effect together with the three factors explaining the relative price of traded goods. The estimated equation is of the following format:
where REER refers to the CPI-based real effective exchange rate, TNT to the Balassa-Samuelson effect, and other variables are as defined in the preceding paragraph. Allowing for country-specific fixed effects, the equation was estimated for a group of 10 major currencies covering two decades worth of data. Estimates were obtained using panel dynamic OLS—a panel extension of the method proposed by Stock and Watson (1988) and advocated by Mark and Sul (2001) among others. This approach corrects for the finite sample bias generated by endogeneity among some of the variables, and provides corrected standard errors. The estimated parameters represent a co-integration relationship among non-stationary variables.
C. Estimation Results and Assessment
12. The estimation results are largely as predicted by the theoretical model, but differ somewhat depending on which variable is used to capture the Balassa-Samuelson effect. The literature suggests two alternative measures: (1) the ratio of the consumer price index (CPI) relative to the wholesale price index (WPI); and (2) the ratio of the GDP deflator to the manufacturing deflator. The first measure provides greater homogeneity in cross-country analysis, but the second measure may better reflect the domestic productivity differential.6 Therefore, the model was estimated separately for each of the two variables. The results of both specifications are reported in Table 1.
Long-Run Equation for the Real Effective Exchange Rate, 1980-2002
(Panel Data, N=10, T=23)
Long-Run Equation for the Real Effective Exchange Rate, 1980-2002
(Panel Data, N=10, T=23)
| CPI-based REER | Relative Deflators | |
|---|---|---|
| Relative price of traded to non-traded goods: | 1.10 | |
| - Measured by ratio of CPI to WPI | (3.77) | |
| - Measured by ratio of GDP deflator to manufacturing deflator | 0.45 (2.36) | |
| Relative supply effect | -0.10 (-0.60) | -0.55 (-2.74) |
| Commodity price effect | 0.72 (2.51) | 0.78 (2.56) |
| Net income on foreign assets | 0.77 (1.91) | 1.10 (2.25) |
Long-Run Equation for the Real Effective Exchange Rate, 1980-2002
(Panel Data, N=10, T=23)
| CPI-based REER | Relative Deflators | |
|---|---|---|
| Relative price of traded to non-traded goods: | 1.10 | |
| - Measured by ratio of CPI to WPI | (3.77) | |
| - Measured by ratio of GDP deflator to manufacturing deflator | 0.45 (2.36) | |
| Relative supply effect | -0.10 (-0.60) | -0.55 (-2.74) |
| Commodity price effect | 0.72 (2.51) | 0.78 (2.56) |
| Net income on foreign assets | 0.77 (1.91) | 1.10 (2.25) |
When using the CPI/WPI ratio to capture the Balassa-Samuelson effect, the coefficient is close to one and statistically significant. The coefficient for the relative supply effect is of the expected sign, but numerically small and statistically insignificant. By contrast, the coefficients on commodity prices and net income on external assets are significant and suggest that a one percent increase in either variable would, ceteris paribus, result in a 0.7–0.8 percent appreciation of the real exchange rate.
The alternative specification affords greater weight to the relative supply effect. The Balassa-Samuelson effect, as captured by the relative manufacturing deflator, is less than half the size compared to the first specification. On the other hand, the coefficient for relative manufacturing output is statistically significant and larger than the Balassa-Samuelson coefficient. Coefficients on commodity prices and net income on external assets are almost unchanged.
13. The results indicate that productivity gaps and the Balassa-Samuelson effect appear to have been the dominant factor weighing on the Canadian exchange rate over the past decade. Although the two estimated equations differ somewhat in the weights allocated to Balassa-Samuelson and relative supply effects, they suggest that these two productivity-related factors have driven the fundamental level of the exchange rate down over the past two decades (Figure 5). From a theoretical viewpoint, the results of the second specification are more intuitive, since they confirm the presence of a relative supply effect; are based on a better measure of the Balassa-Samuelson effect; and are consistent with the observed absolute improvements in Canadian productivity. Even in the second specification, however, the long underlying downward trend in exchange rate fundamentals caused by Balassa-Samuelson is clearly evident.
Contributions to Changes in Real Effective Exchange Rate, 1980 - 2003
Citation: IMF Staff Country Reports 2004, 060; 10.5089/9781451806908.002.A003
Contributions to Changes in Real Effective Exchange Rate, 1980 - 2003
Citation: IMF Staff Country Reports 2004, 060; 10.5089/9781451806908.002.A003
Contributions to Changes in Real Effective Exchange Rate, 1980 - 2003
Citation: IMF Staff Country Reports 2004, 060; 10.5089/9781451806908.002.A003
14. Although the contributions of commodity prices and net external income have been relatively minor in comparison, they have helped arrest the downward trend in recent years. The drop in Canada’s net foreign liabilities from the high level that emerged during 1985-1995 has reduced debt service costs and helped to return the current account balance into surplus, but this factor offset at most about a third of the impact of the Balassa-Samuelson effect. As for commodity prices, the contribution of the terms-of-trade effect explains a large share if not most of the 2003 increase in Canadian exchange rate fundamentals.
15. The results also suggest that the recent appreciation of the dollar has brought the exchange rate closer to underlying fundamentals. Figure 6 plots the real effective exchange rate against the estimated trends that correspond to the two measures of the Balassa-Samuelson effect. Both charts suggest that the exchange rate had been considerably undervalued relative to fundamentals since the early 1990s, but that the gap has either closed or narrowed significantly in 2003. Indeed, with the model based largely on data through the first half of 2003, the appreciation of the exchange rate since then may have taken it above fundamentals.
Actual and Fitted Real Effective Exchange Rates, 1980 - 2003
Citation: IMF Staff Country Reports 2004, 060; 10.5089/9781451806908.002.A003
Source: Fund staff calculations.Actual and Fitted Real Effective Exchange Rates, 1980 - 2003
Citation: IMF Staff Country Reports 2004, 060; 10.5089/9781451806908.002.A003
Source: Fund staff calculations.Actual and Fitted Real Effective Exchange Rates, 1980 - 2003
Citation: IMF Staff Country Reports 2004, 060; 10.5089/9781451806908.002.A003
Source: Fund staff calculations.References
Bayoumi, T., H. Faruqee, and J. Lee, 2003, “Fair Exchange”, (unpublished, Washington: International Monetary Fund).
Benigno, G., and C. Thoenissen, 2002, “Equilibrium Exchange Rates and Supply-Side Performance”, Bank of England Working Paper No. 156, pp.5– 42 (London: Bank of England).
Cerisola, M., P. Swagel, and A. Keenan, 1999, “The Behavior of the Canadian Dollar”, in Canada: Selected Issues, IMF Staff Report SM/99/3 (Washington: International Monetary Fund).
Chen, Y., and K. Rogoff, 2003, “Commodity Currencies”, Journal of International Economics, 60, No.1, 133– 60.
Engel, C., 1999, “Accounting for U.S. Real Exchange Rate Changes”, Journal of Political Economy, 107, No. 3, pp.507– 38.
Gauthier, C., and D. Thessier, 2002, “Supply Shocks and Real Exchange Rate Dynamics: Canadian Evidence”, Bank of Canada Working Paper 2002–31 (Ottawa: Bank of Canada).
Isard, P., and H. Faruqee, 1998, “Exchange Rate Assessment: Extensions of the Macroeconomic Balance Approach”, IMF Occasional Paper No. 167 (Washington: International Monetary Fund).
Lane, P., and G.M. Milesi-Ferretti, 2002, “External Wealth, the Trade Balance, and the Real Exchange Rate”, IMF Working Paper WP/02/51 (Washington: International Monetary Fund).
Lee, J., 2004, “The Return of National Price Levels”, (unpublished, Washington: International Monetary Fund).
MacDonald, R., and L. Ricci, 2002, “Purchasing Power Parity and New Trade Theory”, IMF Working Paper WP/02/32 (Washington: International Monetary Fund).
Mark, N., and D. Sul, 2001, “Nominal Exchange Rates And Monetary Fundamentals: Evidence From a Small Post-Bretton Woods Panel”, Journal of International Economics, 53, No. 1, pp. 29– 52.
Rogoff, K., 1996, “The Purchasing Power Parity Puzzle”, Journal of Economic Literature, 34, No. 2, pp.647– 68.
Stock, J.H., and M.W. Watson, 1988, “Testing for Common Trends”, Journal of the American Statistical Association, 83, pp. 1097– 107.
Prepared by Jaewoo Lee (RES) and Martin Muhleisen.
The last major peak against the U.S. dollar was reached at $0.89 in October 1991.
As explained more fully in Chapter 1 in this Selected Issues paper, the different classification of the IT producing sector required combining the manufacturing and electronic goods producing sector for the purpose of sectoral comparison. The strong productivity increase in this sector in the United States has driven the sharp appreciation of the sectoral real exchange rate.
Recent work on exchange rate determination includes, for example, Benigno and Thoenissen (2002), Lane and Milesi-Ferretti (2002), MacDonald and Ricci (2002), Chen and Rogoff (2003), and Lee (2004).
This result assumes a (generalized) home bias that causes a relatively larger increase in demand for domestically produced goods in response to favorable income and other shocks (Lee, 2004).
It is common practice in the literature to use manufacturing output and prices to approximate those of traded goods.
