A. Context and Motivation

1. Real interest rates in Canada, as elsewhere in the world, have been low for a remarkably long period of time. Current low interest rates are the result of cyclical and trend (i.e., structural) factors. Global factors—such as higher savings in emerging market economies and increased demand for safe assets—are largely responsible for the trend decline in interest rates (e.g., IMF, 2014; Mendes, 2014; Wilkins, 2014). Cyclical factors such as slow growth in advanced economies and lower investment rates have also contributed to keeping global demand for funds low since the global financial crisis. In Canada, real long-term rates averaged just ¾ percent in 2008–14, and short-term rates have been in negative territory since end-2008 (Chart). Also, with remarkably stable inflation expectations, anchored at the Bank of Canada’s target rate of 2 percent, the dynamics of real and nominal rates coincide.

Canada and United States: Real Interest Rates

(percent; quarterly)

Citation: IMF Staff Country Reports 2015, 023; 10.5089/9781498385305.002.A002

Sources: Bank of Canada; Consensus Forecasts; and IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Canada and United States: Real Interest Rates

(percent; quarterly)

Citation: IMF Staff Country Reports 2015, 023; 10.5089/9781498385305.002.A002

Sources: Bank of Canada; Consensus Forecasts; and IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

Canada and United States: Real Interest Rates

(percent; quarterly)

Citation: IMF Staff Country Reports 2015, 023; 10.5089/9781498385305.002.A002

Sources: Bank of Canada; Consensus Forecasts; and IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

2. The unprecedented global low interest rate environment challenges the evaluation of the appropriate monetary policy stance. More specifically, given low global real rates, it is harder to assess the degree of accommodation (tightness) of their monetary policy stance. The present Selected Issues Paper will try to answer the following questions for the case of Canada:

• How much monetary accommodation is currently present in the economy in light of secular trends in interest rates and cyclical conditions?

• Looking forward, conditional on the US monetary policy normalization, how much policy accommodation will be present in the economy given current policy rate projections?

To shed light on the above questions we introduce and estimate the neutral (natural) interest rate for Canada, which is the most direct benchmark to assess monetary policy conditions.² In particular, assessing the degree of monetary policy stimulus is important to appropriately balance current and expected costs and benefits of the monetary policy stance.

3. A neutral rate of interest is a useful benchmark against which to assess a country’s monetary policy stance. We define the neutral rate as the short-maturity (real) interest rate such that any (real) policy rate below (above) the neutral rate implies that the central bank is stimulating (restraining) the economy relative to its natural course—i.e., a course that is affected only by non-monetary factors in absence of rigidities related to the existence of money. It is also useful to think of the neutral rate as determined by a mean-zero cyclical factor and a slow moving trend—both of them affected by non-monetary factors. Indeed, various non-monetary factors that influence the business cycle will also affect the cyclical component of the neutral rate (see next section).

4. Our analysis finds that monetary policy in Canada has generally been quite accommodative in the post-recession period and the neutral interest rate is expected to increase smoothly towards its long-term value as U.S. monetary policy normalizes. More precisely, the current neutral rate plus inflation is estimated to be between 1¾ and 2½ percent and expected to converge to between 3–4 percent over the medium term.³

The rest of the paper is divided as follows: Section B will define the concept of the natural rate, Section C describes the model and methodology, Section D presents and interprets the results, and Section E concludes.

B. The Neutral Rate of Interest

Definition

5. Estimates of neutral interest rates help central banks think about monetary conditions and how to communicate them. Since neutral interest rates are unobservable and difficult to estimate precisely, they are best thought of as providing a useful conceptual framework for thinking about monetary policy (Chetwin and Wood, 2013). In this regard, it is helpful to distinguish between a short-term cyclical component and the longer-term component of the neutral rate.

6. In the short term, the neutral rate varies with the cycle. Following the New Keynesian tradition, the neutral rate is defined as the one prevailing in absence of nominal rigidities and other frictions associated with the existence of money.⁴ In other words, the neutral rate is affected by all shocks that are deemed non-monetary (i.e., real shocks) such as technology shocks, time preference shocks (which may capture demographic shifts), and external shocks. For example, even in the absence of money, a demand shock—possibly induced by expectations of higher income growth due to faster technological progress in the near future—may increase demand for funds and, in turn, the neutral rate. The advantage of keeping track the neutral rate over the cycle is that it can be used to gauge the monetary policy stance on a high-frequency basis. It may also be useful to assess how much stimulus is present in the economy when nominal interests are constrained by the zero lower bound.⁵

7. Over the medium to long term, the neutral rate is the real policy rate consistent with output at its potential level and stable inflation. Over the long term, we expect the neutral rate to be determined solely by its trend component which is the rate prevailing after the effects of all cyclical shocks have dissipated. Some authors have defined the neutral rate as just its trend component (e.g., Mendes, 2014). The advantage of this definition might be in terms of communication with the public and in part methodological. The two concepts, however, are clearly complementary. In the current paper we will focus on the first but we use knowledge from IMF (2014) and Mendes (2014) to calibrate the range of where the neutral rate is expected to be in the long term.

8. A simple illustration. Formally, we can write the neutral rate as follows:

$r_t^n={\rho }_t+r_t^{n,c},$

where the first term is a slow moving process or drift term while the second term is a mean-zero cyclical process. A monetary policy rule could be expressed as:

$i_t=\overline{\pi }+r_t^n+{\phi }_{\pi }({\pi }_t-\overline{\pi })+{\phi }_x{x}_t$

where the short-term policy rate responds to inflation’s deviation from target $({\pi }_t-\overline{\pi })$ and the output gap x_t, and moves one-to-one with the natural rate, $r_t^n$. Over the cycle, monetary policy can turn out to be tight, $i_t>\overline{\pi }+r_t^n$, even though the policy rate is below its long-term value $i_t<\overline{\pi }+{\rho }_t$ (Chart).⁶ Finally, in the presence of a trade-off between inflation and output gap stabilization, setting the policy rate to the neutral rate $i_t=\overline{\pi }+r_t^n$ in general is not an optimal choice.⁷ In other words, the optimal policy rate may deviate from the neutral rate over the cycle depending on the state of the economy.

Neutral Rate and Policy Rate

(Percent)

Citation: IMF Staff Country Reports 2015, 023; 10.5089/9781498385305.002.A002

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Neutral Rate and Policy Rate

(Percent)

Citation: IMF Staff Country Reports 2015, 023; 10.5089/9781498385305.002.A002

• Download Figure
• Download figure as PowerPoint slide

Neutral Rate and Policy Rate

(Percent)

Citation: IMF Staff Country Reports 2015, 023; 10.5089/9781498385305.002.A002

• Download Figure
• Download figure as PowerPoint slide

C. Methodology

We perform a Bayesian estimation of a small open economy (SOE) model (Adolfson and others 20014, and Justiniano and Preston, 2010) using Canadian and U.S. data. For parsimony, the model abstracts from the commodity sector.⁸

9. The model economy. The core model has non-commodity tradable consumption and investment goods that can be imported from (and exported to) the United States. Capital and labor are factors into the production function; capital accumulation is subject to investment adjustment costs.⁹ Four Phillips curves capture nominal rigidities in wages, domestic prices, import prices, and export prices which result in relative price dispersion. The central bank responds to lagged interest rates, 1-year ahead expected CPI inflation and to a weighted average of the model-consistent output gap and the Bank of Canada’s estimate of the output gap.¹⁰ Financial capital is perfectly mobile, the exchange rate is free to adjust, and a home and foreign bond can be traded globally.

10. The natural equilibrium is defined as the one where prices and wages are free to adjust. A trend growth shock (a combination of a labor augmenting and investment-specific technology shock), a time preferences shock, shocks to U.S. variables’ block, and shocks to expectations to export demand and investment profitability (news shocks) are deemed fundamental shocks and, thus, affect the actual and the natural equilibrium alike. News shocks affect expectations on export demand and the profitability of investment up to 4 quarters ahead.¹¹

11. Habit formation, uncovered interest parity, and new shocks have a major role in the framework. The core behavioral equations (common to both the actual and natural equilibrium) relates the real interest rate of maturity j to the time-preference shock ξ and expected growth in household consumption, c. Habit formation, captured by the parameter h, implies that consumption accelerations (rather than increases) positively affect the real rate in the short run (Campbell and Cochrane, 1999):¹²

The stationary but persistent exogenous process ξ captures slow movements in the neutral rate. Over the balanced-growth-path the neutral rate has a one-to-one relation with consumption growth. Domestic and foreign interest rates are linked through uncovered interest parity conditions, holding at 3-month and 1-year horizons. Expressed in real terms we have:

where ${\mathbf{\Delta }}^{\mathbf{\text{j}}}{\mathbf{\text{e}}}_{\mathbf{\text{t+j}}}$ is the bilateral real exchange rate depreciation over the bond maturity period and ϵ_t is a non-fundamental shock. This formulation allows U.S. interest rates to exert a strong effect on Canadian interest rates, consistent with stylized facts (Chart). Finally, news shocks that affect export demand and investment profitability allow us to gauge the degree of pessimism (optimism) present in the economy relative to the smaller information set available to the rational expectations agent in the model.

Correlation between Canadian and U.S. Interest Rates

(percent; rolling correlation between changes in nominal yields)

Citation: IMF Staff Country Reports 2015, 023; 10.5089/9781498385305.002.A002

Sources: Bank of Canada; and IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Correlation between Canadian and U.S. Interest Rates

(percent; rolling correlation between changes in nominal yields)

Citation: IMF Staff Country Reports 2015, 023; 10.5089/9781498385305.002.A002

Sources: Bank of Canada; and IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

Correlation between Canadian and U.S. Interest Rates

(percent; rolling correlation between changes in nominal yields)

Citation: IMF Staff Country Reports 2015, 023; 10.5089/9781498385305.002.A002

Sources: Bank of Canada; and IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

12. The empirical strategy: We first estimate the exogenous U.S. block, which is summarized by a 4-variable VAR(2) in GDP growth, CPI inflation, the Fed Funds rate, and the 1-year Treasury bill real yield. We then proceed to estimate the parameters of the exogenous processes and some of the parameters of the core model by Bayesian methods. Finally, we back up the smoothed estimates of the neutral rate (and other non-observable variables) using the Kalman filter.¹³¹⁴

D. Main Results: Assessing Monetary Policy Stance and the Determinants of the Neutral Rate

Neutral Interest Rate: Baseline Scenario

(Percent, nominal 3-month T-bill rate)

Citation: IMF Staff Country Reports 2015, 023; 10.5089/9781498385305.002.A002

Sources: Bank of Canada; and IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Neutral Interest Rate: Baseline Scenario

(Percent, nominal 3-month T-bill rate)

Citation: IMF Staff Country Reports 2015, 023; 10.5089/9781498385305.002.A002

Sources: Bank of Canada; and IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

Neutral Interest Rate: Baseline Scenario

(Percent, nominal 3-month T-bill rate)

Citation: IMF Staff Country Reports 2015, 023; 10.5089/9781498385305.002.A002

Sources: Bank of Canada; and IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

13. In the post-crisis period monetary policy has been substantially accommodative. The neutral rate dropped dramatically during the global recession in 2009. However, it recovered quickly alongside growth prospects in the aftermath of the crisis. Overall, monetary policy has been quite accommodative from 2010 to today as highlighted by the average interest rate gap over the period being -0.9 percent per quarter.¹⁵ Only for a short period in 2011 did the interest rate gap record positive values. The neutral rate has remained strictly positive throughout this period, consistent with the fact that the zero lower bound on the policy rate has not been an issue in Canada during the recovery. In the last 3 quarters the interest rate gap has been below -1 percent, suggesting continued accommodation in the Bank of Canada’s policy stance.

14. The export sector has negatively affected the neutral rate until recently. Weak non-energy exports reduced the neutral rate on average during the recession and post-recession period mainly through weak investment and labor demand, which, in turn, have negatively affected household income prospects. In particular, prospects of a strong export recovery were disappointed in 2010, and expectations turned pessimistic in 2011 before recovering by the end of the year. Export expectations turned again pessimistic, on average, between the end of 2012 and early 2014; especially before recovering sharply on the backdrop of a surprise contraction of U.S. GDP in 2014Q1 (Chart).¹⁶

15. Perceived low investment profitability has negatively affected the neutral rate. After a quick turnaround at the end of the recession, labor productivity and, especially, the investment-specific shock (which drives investment profitability) have been underperforming until recently, reducing the demand for funds and, thus, preventing the natural rate from rising. In particular, news shocks to investment profitability turned particularly overly pessimistic by the end of 2011 even though labor productivity has been slowly improving.

16. U.S. real rates have a strong influence on the Canadian neutral rate. A typical¹⁷ 100 basis point (bps) permanent increase in the (short-term) U.S. real rate translates into about a 50 bps increase in the neutral real rate over one year, and 100 bps over the long run, by assumption.¹⁸

17. Looking forward, the neutral rate is expected to slowly converge to about 3.5 percent by the end of 2017. Conditional on the U.S. Fed Funds rate that begins increasing in September 2015 and converges to 3.75 percent in 5 years, the Canadian neutral rate will start increasing by the end of 2015, crossing 3 percent in the second half of 2016. The Canadian neutral rate is quite sensitive to the U.S. interest rates. Indeed, if U.S. rates converged faster, two years earlier in 2016, the neutral rate in Canada would cross 3.5 percent already in 2016 (Chart).

Neutral Interest Rate: Faster U.S. Normalization Scenario

(Percent, nominal 3-month T-bill rate)

Citation: IMF Staff Country Reports 2015, 023; 10.5089/9781498385305.002.A002

Sources: Bank of Canada; and IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Neutral Interest Rate: Faster U.S. Normalization Scenario

(Percent, nominal 3-month T-bill rate)

Citation: IMF Staff Country Reports 2015, 023; 10.5089/9781498385305.002.A002

Sources: Bank of Canada; and IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

Neutral Interest Rate: Faster U.S. Normalization Scenario

(Percent, nominal 3-month T-bill rate)

Citation: IMF Staff Country Reports 2015, 023; 10.5089/9781498385305.002.A002

Sources: Bank of Canada; and IMF staff estimates.

• Download Figure
• Download figure as PowerPoint slide

E. Conclusions

18. A substantial degree of monetary policy accommodation is currently present in the Canadian economy. Estimates of the short-term neutral rate for Canada suggest that monetary policy has been very supportive of the economy during the post recession period on average. Current estimates indicate the policy rate is more than one percent below the neutral rate, suggesting that a large degree of monetary policy accommodation is present in the economy.¹⁹

19. Disappointing non-energy exports and labor productivity have been the main domestic factors that have kept the neutral rate low in recent years. Excessively pessimistic expectations on non-energy exports and on investment profitability and disappointing labor productivity performance have prevented the neutral rate from rising substantially in the post-recession period and have prevented monetary policy from becoming too accommodative.

20. U.S. monetary policy is key to understanding future developments in the Canadian neutral rate. Looking forward, conditional on the projected U.S. monetary policy rate path, the neutral rate is expected to start increasing toward the end of 2015 toward its long-term level.