Sunday, July 14, 2019

The July 2019 Monetary Policy Report 2: Risks to the Bank of Canada’s Inflation Outlook


As part of the Monetary Policy Report ( pdf here), the Bank of Canada includes a section on risks, in which it includes both up-side and down-side risks.  Up-side risks are factors which could push the AD curve further to the right than the Bank had anticipated, and they’re really only risks if Canada is pretty close to potential output, in which case the output gap could close unexpectedly and inflation build up faster than the Bank anticipated.  So having a good handle on where GDP is relative to potential, by which, if we’re working with a very Keynesian AS curve, we mean how close we are to the region of GDP in which the AS curve goes from being flat to being positively sloped, and equally importantly, how quickly that positively sloped region steepens into the vertical region of the AS curve, is extremely important in judging what the Bank should be doing with interest rates.

Broadly speaking, we can categorize risks under a couple of headings.  There are external risks, which we can only assess and try to respond to, and there are policy response risks which relate to how the Canadian economy reacts to the Bank’s policy actions.


In terms of external risks the big one at the moment is the international trade risk.  Most people’s focus there is on the China/US trade war situation, but we can’t overlook the situation between China and Canada.  In the China/Canada brouhaha we’re looking primarily at an AD risk – the degree to which China is prepared to cut off its imports from Canada (which are, obviously, Canadian exports) in order to chastise Canada for not toeing the correct line.  Current evidence suggests China is prepared to cut back on what are, from the Canadian point of view, pretty significant Canadian exports.  That’ll shift the AD curve to the left, reducing GDP, for as long as China is playing the trade war game.  The Bank will have to decide whether that’s a long term risk and whether it should respond with monetary easing (and for that matter whether monetary easing would be an effective response).

The other trade risk is the China/US trade war, and here the Bank seems to have decided that the impact on Canada will be very much on the AS side – that it’ll shift the AS curve up and back to the left.  That seems to be the notion that’s behind some of Stephen Poloz’s recent remarks.  So if that trade war persists, the Bank will have to decide whether there’s anything which monetary policy can do in response – the discussion in the MPR suggests that it’ll decide that monetary expansion would be not just ineffective but downright counterproductive.  In that case the Bank'll be on the hot seat, having to explain why it’s not doing anything in the face of press and political commentary demanding that it act.

Imagine trying to explain shifting AS curves, and their interaction with shifting AD curves, in a press conference.

In terms of response risks, the big one comes out of the housing market, but perhaps in an odd way.  If the Bank decides it has to raise interest rates, how will that increase interact with current household debt levels to affect non-housing consumption spending? The Bank of Canada’s stress tests for borrowers – how well could a borrower handle a two percentage point increase in their mortgage interest rate – focused on whether households had the financial capacity to absorb a rate increase without defaulting on their mortgages.  But they’re likely to have to handle the adjustment by cutting back on their spending on other things.  So when we’re looking at the impact of the rate of interest on consumer spending, we have to take account of the current debt service ratio and, in broad terms, how the elasticity of consumer spending with respect to the rate of interest varies depending on the level of long-term household debt.  The problem here is that it’s been a long time since we’ve seen interest rates this low so it’s not clear whether we have the data to estimate the kind of response information which policy makers need.

That comes into another question which is often raised – with rates as low as they are, does the Bank have any room to respond to a downturn by cutting rates further?  When this question’s asked it’s usually prefaced by reference to the number of percentage points by which the Bank cut interest rates in response to recessions in previous years.  That size cut to policy rates today would put us among those countries with negative short term interest rates, and there’s no evidence the Bank really wants to join that group.  So it’s got to be wondering about elasticities – how sensitive are components of AD when interest rates are this low?  Again, a tricky one to tackle, given the data we have.

Taken as a group, the importance of the interest rate response risks which the Bank has to be thinking about depends crucially on where we are relative to the potential level of output.  Estimating the output gap is not easy, and different central banks use different approaches to doing it.  The big question out of the US these days is whether the unemployment rate is still a good measure of labour market tightness.  The assumption behind the notion of a stable Phillips Curve (which John Cochrane doesn't believe in) is that the relation between the measured unemployment rate and the output gap has been stable over time, so we can use the unemployment rate as an index of how close we are to potential output.  The American literature is suggesting that, in the US, the relation between measured unemployment and the labour force participation rate has changed.  In technical terms, that would change the slope of the Phillips Curve and make old assumptions about the size of any inflation/unemployment trade-off obsolete.  

From the Bank of Canada’s perspective one issue has to be whether the same kind of change has occurred in Canada.  This is one of those cases where it’s particularly risky to assume that what’s happened in the US has automatically also happened in Canada.

And finally, there’s the issue of where we actually are relative to potential output (as distinct from the technical question of how we measure the gap) and what would happen to inflation expectations if the AD curve shifts into the steep portion of the AS curve.  Inflation expectations are what anchor the inflation rate, so if people in general come to expect that inflation will be above 2%, firms will build that into their plans and the anchor rate of inflation will rise.  When the anchor rate rises the actual rate will also rise, and once that happens, pulling inflation back down can be pretty painful.  

You’ve got to think that as well as looking at full CPI inflation and at its various measures of core inflation, the Bank is trying to get a handle on how the inflation anchor is likely to change if inflation gets pushed up.  An upward blip in inflation probably won’t change the anchor, but there’s still the question of how long the inflation rate would have to stay above 2% for the anchor to shift up.  As I say, that depends on where potential output is right now relative to actual GDP, which comes back to how you measure the output gap and to how sensitive the AD curve is to changes in external factors and to the rate of interest, and also to how strongly price pressure builds up once we get into the positively sloped region of a Keynesian AS curve.  

So incredibly enough, all this stuff ties together.  And drops onto Stephen Poloz’s desk with a heavy thud.  But analyzing it's got to be keeping the Bank's research staff fully employed.

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.