2008 was a large negative demand shock – also in Canada

Scott Sumner has a follow-up post on Nick Rowe’s post about whether a supply shock or a demand shock caused the Canadian recession in 2008-9. Both Nick and Scott seem to think that the recession in some way was caused by a supply shock.

I must admit that I really don’t understand what Scott and Nick are saying. It is pretty clear to me that the shock in 2008-9 was negative aggregate demand shock.

Lets start with the textbook version of a negative aggregate demand (AD) shock). Here is how a negative demand shock looks in AS/AD model (the growth rate version):

Demand shock

So what happened in Canada? Here is a look at inflation measured by headline CPI and by the price deflator for final domestic sales.

CAD inflation

Both measures of inflation were running higher than the Bank of Canada’s official 2% inflation target when the crisis hit in the autumn of 2008.

However, it is pretty clear that inflation slowed sharply and dropped well-below the 2% inflation target in 2009 as the Canadian economy went into recession (real GDP contracted). It is hard to say that this is anything other than a rather large negative AD shock.

Obvioulsy inflation increased above 2% in 2011, but we all know that a major negative supply shock hit in 2011 as global oil prices spiked. In the case of Canada this in fact is both a negative supply shock and a positive demand shock (remember Canada is an oil exporter). That said, the rise in inflation was certainly not dramatic and since 2012 inflation has once again dropped well-below 2% indicating that monetary policy in Canada has become overly tight given the BoC’s 2% inflation target.

I might add that different measures of inflation expectations (both survey and market data) are telling the exact same story. Inflation and inflation expectations eased significantly in 2008-9 and once again in 2012.  

And we can tell the same story if we look at the price level. The graph below compares the two measures of prices (CPI and the final domestic demand deflator) with an 2% price path starting in Q3 2008.

Canada Price Level

Again the picture is clear. The price level – for both measures – are lower than a hypothetical 2% price level path – indicating that Mark Carney and his colleagues in the Bank of Canada have kept monetary conditions too tight over the past 4-5 years – maybe because of a preoccupation with the risk of “bubbles”. Mark Carney might be talking about NGDP level targeting, but he is certainly also speaking quite a bit about “macroprudential indicators” (modern central bank lingo for bubble risk).

Concluding, it is very clear that the Canadian economy was hit by a large negative demand shock in 2008 and initially the BoC has kept monetary policy overly tight and the recent tightening of monetary conditions certainly also looks problematic.

Once again it is monetary policy failure and it is certainly not a negative supply shock, which is to blame for the Canadian recession and sub-trend growth since 2008. Needless to say NGDP tells the exact same story. I should add that the size of this “monetary policy failure” is fairly small compared to for example for example what we have seen in the euro zone.

Reminding Scott about the Sumner Critique

Given the very clear evidence of a negative demand shock I find this comment from Scott somewhat puzzling:

Let’s suppose that the BOC had been targeting NGDP in 2008, when global trade fell off a cliff.  How would the Canadian economy have been affected?  Many would see the drop in global trade as a demand shock hitting Canada, as there would have been less demand for Canadian exports.  In fact, it would be an adverse supply shock.  Even if the BOC had been targeting NGDP, output would have probably fallen.  Factories in Ontario making transmissions for cars assembled in Ohio would have seen a drop in orders for transmissions.  That’s a real shock.  No (plausible) amount of price flexibility would move those transmissions during a recession.  If the assembly plant in Ohio stopped building cars, then they don’t want Canadian transmissions.  If the US stops building houses, then we don’t want Canadian lumber.  That’s a real shock to Canada, i.e. an AS shock.

I simply don’t understand Scott’s argument. A negative shock to exports obviously is a negative demand shock. From the perspective of nominal spending a negative shock to exports is a negative shock to money-velocity in the exact same way as a tightening of fiscal policy. Therefore, if the BoC had been targeting NGDP (it actually also goes for inflation targeting) the Sumner Critique would apply – the BoC would offset any negative shock to exports by easing monetary policy (increasing M to offset the drop in V). As a consequence domestic demand would rise and offset the drop in exports. And this obviously applies even if prices are sticky. Yes, the production of transmissions in Ontario drops, but that is offset by an increase in construction of apartments in Vancouver.

However, the point is that the BoC failed to offset the shock to exports and as a consequence prices have been growing slower than implied by BoC’s official inflation target.

There is absolutly nothing special about Canada – its monetary policy failure – the failure is just (a lot) smaller than in the euro zone or the US.

PS I could also have used the GDP deflator as well in my examples above. The story is the same. In fact it is worse! The GDP deflator dropped by more than 4% during 2009. The primary reason for the massive drop in the GDP deflator is that the price of oil measured in Canadian dollars dropped sharply in 2008-9. As drop in the oil price obviously is a negative demand shock as Canada is a oil exporter. The story in that sense is completely the same as what happened to the Russian economy in 2008-9. Had the BoC had followed a variation of an “Export Price Norm” as the Reserve Bank of Australia is doing then the negative shock would likely have been much smaller as was the case in Australia.

GDP deflator Canada

PPS JP Irving also comments on the Canadian story.

Leave a comment


  1. Looks like we both went right for the national accounts deflators! I think I am somewhere between you and Scott on this. I do think inflation looked a bit high in Canada, given the size of the AD shortfall and all of the apparent long run positive AS in Canada, but then I am only looking at this to distract myself from the more urgent work on synthetic NGDP expectations….

    • JP…good to hear you still work on that. If you did it for Canada then OIL/CAD would be the only indicator you would need…

  2. felipe

     /  March 8, 2013

    Hi Lars,

    I think the point Scott is getting at is that even if the BoC can keep NGDP on track, a recession might have occurred due to rebalancing. Imagine that there is a world wide agreement to ban car sales. Also imagine there is an economy that heavily exports car parts, but doesn’t really buy many cars. That country is experiencing a demand shock. Yet, because no amount of monetary easing will cause car parts to be demanded again, people will have to stop exporting car parts and doing something else. And that process does not come for free.

    In other words, I think one can say that a negative demand shock in a major trading partner can (also) be a negative supply shock to you.

    • felipe, that might be what Scott is thinking. But again – if the cb keeps NGDP/AD on track then all “sectoral” demand shocks will be offset. But obviously given the AS curve is not vertical in the short-run (prices and wages are sticky) then a shock to NGDP will also reduce RGDP growth, but that is not a supply shock.

      • fsateler

         /  March 8, 2013

        If a negative supply shock causes the central bank to tighten in order to keep inflation on track, can you really say that the recession was caused by an AD shock?

        NGDP targeting would have been better than IT, but there would still have been a recession due to the supply shock: the supply shock left you poorer in real terms.

  3. Hi Lars!
    1. Take a look at core CPI in this old post of mine: http://worthwhile.typepad.com/worthwhile_canadian_initi/2013/01/the-bank-of-canadas-success-and-failure.html
    Nothing happened. No sign of a recession. It kept growing at trend.
    Sure, total CPI didn’t keep exactly at trend, but it doesn’t fit the pattern of a recession either. It blips well above trend in 2008, then slightly below trend in 2009, then is back on trend in 2011. If you compare it to previous blips, even times when there wasn’t a recession, it was nothing especially big. And it’s only in late 2012/early 2013, when the Canadian economy has pretty well recovered from the recession, that we see both total and core inflation falling below target.

    2. Think about that metaphor for international trade. We produce wheat, put the wheat into ships, sail the ships out into the ocean, where the wheat gets magically transformed into cars, and the ships come home full of cars. A terms of trade shock is like a technology shock, for transforming wheat into cars. Is that AS or AD?

    3. I’m still agnostic. I can’t decide whether Scott’s AS story is correct.

    • Nick,

      1) CPI and core CPI are horrible measures of monetary inflation. CPI because of indirect taxes and imports prices and core CPI because the average core CPI inflation always is low than other measures of inflation. The deflator is the one to use – btw that is what we use in MV=PY. Btw take a look at the GDP deflator in Canada from 2003/4 to 2007/8 in Canada. Monetary policy was clearly too easy – despite CPI being on track. CPI simply is to supply side driven. We had a case of what Austrians term relative inflation. And isn’t it problematic to take out “energy prices” from the measure of prices in Canada. Afterall energy is a pretty important part of DEMAND.

      2) I generally don’t like the term “terms-of-trade shocks”. It is misleading as it is both demand and supply shocks. An increase in import prices is a negative supply shock. However, an increase in export prices is a positive demand shock.

      3) Nick, you know I am right – it was a AD shock. Recessions are always and everywhere a monetary phenomena – http://worthwhile.typepad.com/worthwhile_canadian_initi/2011/08/recessions-are-always-and-everywhere-a-monetary-phenomena.html

  4. Or: Take a second look at your second graph. If you started it a couple of years earlier, you would get a very different 2% trend line. You started that 2% trend line right from the peak.

    • That I admit. What I am just trying to say is that there is a disinflationary process. Prices growth clearly decelerates. Exactly as we would expect, but of course prices are downward sticky – other monetary policy failure would be irrelevant. But saying that the AS is downward sloping doesn’t mean that we have a negative supply shock.

  5. Lars: “1) CPI and core CPI are horrible measures of monetary inflation.”

    I would say that CPI and core CPI are horrible measures for monetary *targets*. Which is precisely what those Canadian graphs show. Because CPI and core CPI don’t “signal” the recession, and so didn’t give the Bank of Canada a clear signal to respond to (in contrast to NGDP, which provided a very clear signal). But I want to know *why* they failed so badly (in Canada, recently) as monetary targets.

    The NGDP dog barked loud and clear that monetary policy was too tight. It started barking loudly when the recession started, and slowly stopped barking as the recession slowly went away. The total CPI dog barked softly and at random, almost unrelated to the recession. The core CPI dog didn’t bark at all.

    But *why* didn’t those two CPI dogs bark at the right time?

    • Nick, I am not sure why core CPI and CPI are so bad indicators, but the fact is that they are. If we can’t win the NGDP fight then at least we can try to convince central banks to focus on the GDP deflator rather than CPI. That is what my friend Jeff Frankel is now arguing. In Canada you should also look at oil prices in measured in loonies as a measure of monetary tightness – that would be another Frankel recommendation: Look at the export price as an indicator of AD.

  6. Lars: If I understood *why* CPI and core CPI are bad indicators, in addition to seeing that they *are* bad indicators, I would have greater confidence that NGDPLPT would do better.

    Plus sometimes, we just want to know why. Because it’s interesting trying to understand the world.

  1. TheMoneyIllusion » Louis Woodhill on QE3
  2. “Shock Signatures” | Historinhas

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: