« Canada and the Eurozone: a comparison | Main | How do we reduce our debt/income ratios? The paradox of debt. »


Feed You can follow this conversation by subscribing to the comment feed for this post.

This is yet another superb posting on WCI. Thanks to both of you!

I'm wondering to what extent changing expectations affect the shape of the SRAS curve. In particular, during the current recession it looks as if people's expectations are for a serious recession. If people's expectations are for something worse than we will be experiencing, will they start taking jobs much sooner? Will unemployment durations be lower than we might otherwise expect?

Your point about how the SRAS curve in an empirical setting will reflect the average over the sample is especially important in the current context. We can't simply extrapolate empirical estimates from "normal" periods to the current one.

Also very interesting point that the goal of endogenous fiscal policy is to stabilize and not necessarily stimulate output (if that makes sense) and so attempts to measure "shocks" will be biased towards zero. Subtle but I think quite important.

Thanks very much John!

I had to think about your question.

On the supply side, yes: if people expect it will be a long recession, the unemployed will be more willing to take whatever jobs are offered, and this means that fewer firms will face supply constraints, the SRAS will stay flatter longer, and the expansionary fiscal policy will be more effective in increasing output.

On the demand side, the story is a bit different. The worse people expect the recession to be, the more they will cut spending (via the marginal propensity to consume, accelerator effects on investment, and the effect of expected deflation on real interest rates). But conversely, if they expect fiscal policy will be effective in shortening the recession, all those effects will work in the opposite direction, and make fiscal policy more effective.

It is not impossible (and we can even take macromodels off the shelf of second-year textbooks to get this) that sacrificing a goat could work, providing enough people believe it would work. When you are stuck in a liquidity trap (Canada is not quite, but is not far off), there can be two equilibria, and expectations can be self-fulfilling.

This might explain why some people (I'm thinking more south of the border) come across like the Spanish Inquisition in dealing with fiscal sceptics. Enough agnostics in a foxhole might mean none of us reach salvation.

Thanks Brendon! But I was thinking, after I had posted, if fiscal policy is typically pro-cyclical (not the counter-cyclical it should be) because governments respond to good times and relaxed budget constraints by spending more, and cutting back in bad times, then the bias would go in the other direction, and make multipliers look bigger than they are.

OT, but I don't think you've mentioned your thoughts on Carney's rate cut yesterday, as well as his rather surprising economic projections for 2010 (3.8% growth?!?). Were you planning on posting on that topic soon? I will admit some confusion that the BoC does not decrease rates more rapidly to 'slay the bear', but I suppose that if they believe those projections, they should be hesitant to cut rates further.

Andrew: I was thinking about posting on the BoC decision, and monetary policy report. I even had the title picked out: "Some random reflections on the Bank of Canada". But I couldn't think of anything especially new or interesting to say. I think they should have cut to zero, but I had already said that. And I had already said why it's better to put too many bullets in the bear. Michael Parkin had an interesting commentary in the National Post, but I had already covered what I had to say on that topic in the Useful Austrian post. So I didn't.

Maybe we should just have an "Open Thread" for times like that, so readers can add their comments, even if Stephen or I don't want to say much?

I'm afraid that Mr. Carney is somewhat delusional about the economic outlook.

I'm also afraid that, similar to US blogosphere comments on his US government bretheren, we're eventually going to hear more media criticism about his Goldman Sachs pedigree and hubris when the Canadian economy turns out to be in significantly worse shape than his very unwise and overconfident forecast.

Related to the questionable forecast coming from the BoC - does anyone know how well their DSGE forecasting model (TOTEM) is fairing in this environment? I would guess that a lot of weight is being put on judgement vs. models these days.

Interesting question Brendon: I have just skimmed through this technical report on TOTEM: http://www.bankofcanada.ca/en/res/tr/2006/tr97-e.html .
It's certainly an impressive achievement, but it is not at all designed for handling the current problem, because:

1. Only one interest rate (the 90 day rate), so there are no (domestic) spreads in the model. They probably have to hand-massage the link between the overnight rate and the 90 day commercial paper rate, since the spreads will have changed a lot recently. There's also the question of *which* 90 day rate (which wasn't previously an issue).

2. No financial sector.

3. No monetary aggregates. I think there is one stock of "bonds" and all (Canadian) bonds are identical and (I think) 90 day bonds. So it can't handle quantitative or qualitative easing, for example.

Half the households are assumed credit-constrained "current income consumers" (the other half are permanent income consumers). Maybe they could get some action by shifting that ratio. But firms (I think) are never credit-constrained.

My guess? 90% judgment. (But then it always was way more judgement than model anyway, I think, because however fancy your model, it still has all the exogenous variables which end up getting extrapolated.

This is really worthy of a post, but I don't really trust my understanding of big models.

My own guess is that the TOTEM model is what led the Bank to sharply revise down its forecast. Our terms of trade was what kept us above the fray until 2008Q4: at least half of the real income growth we've had since 2002 came from the terms of trade, so when the terms of trade went south, forecasting a recession was a pretty easy call.

And it may also explain the short-and-sharp recession story. Futures prices are indicating an increase in oil prices, and if China spends a chunk of its mountain of reserves on a fiscal stimulus, it's not unreasonable to think that oil and commodity prices have overshot and are due for a sharp rebound two or three quarters from now.

I'm former student of yours at Carleton and I recently came across your blog. Reading it makes me feel like I am sitting in lecture again in Southam. I'll have to go through your old posts to catch up on your thoughts.

While reading your comments on what theory normally tells us, I instantly flashed back to two things you said about theory in class. One is that all theories are [nonsense]. Two, while that may be the case they are our best attempt to explain the world around us because without them we couldn't do much in economics.

My question is about the interest rates hitting the zero floor and thus AD actually being able to increase in a small open economy such as Canada's because the currency will not appreciate. When the BoC sets interest rates it is the overnight lending rate if I am not mistaken? At the end of the day the rubber hits the road with the big five banks dealing with everyday Canadians and businesses. The theory if I am not mistaken is that as the BoC lowers interest rates banks should correspondingly lower theirs as well. Obviously they will never charge zero interest because they wouldn't make a profit. But presumably it'd be BoC Rate (zero percent) + Bank Profit rate. One complaint I've read (I don't know how true it actually is) is that the banks are not lowering their interest rates correspondingly to changes in BoC rates. If this is true and Banks are not lowering their rates as much as the BoC is, how, if at all, will this effect your speculation about the money multiplier?

Regarding TOTEM - it is very impressive, but I just never trust models that complicated or that large. I prefer Murchison's earlier work, the NAOMI model used at Finance. I've adapted it for my own forecasting needs and find that it has very nice properties. I also like Doug Laxton's small QPM's, he is doing some very interesting work with them at the IMF. Though you are right, all models come down to judgement and add-factors.

Hi Mark! Glad to see you have remembered stuff. (But you meant to say "depreciate", not "appreciate"!) Yep, you've a lot of reading to catch up on. Have fun with it!

When the BoC cuts the overnight rate, this causes all interest rates to come down (but longer rates never came down as much, unless the market expected the BoC overnight rate to stay down at its new lower level for a long time), and this stimulates consumption and investment demand. (And the exchange rate would depreciate as well, causing net exports demand to rise.)

Recently, as interest rates approach the lower bound of zero, and as risk and liquidity spreads have widened, different sorts of loans are not as close substitutes as they once were, so a given cut in the BoC overnight rate has a smaller effect on the whole interest rate curve(s), so is less powerful than it would otherwise be. That means the BoC must be more aggressive to get a given effect on AD.

But they still have some room left. And they can also follow unorthodox policies, by doing Open Market Operations (remember them?) in longer bonds, and even private bonds, or even stocks, (rather than just the usual very short term government bonds, alias T-bills).

Stephen, Brendon: The TOT part of TOTEM is nice. The thing I really like about it is that it manages to build in inflation inertia, and not just price level inertia. Everybody uses the Calvo (changes prices at random) model, because it's tractable. But Calvo gives only P-level inertia. The Taylor assumption (fixed period for price changes, so most of the prices that change are the oldest prices) gives inflation inertia, which makes a very big difference to how the model responds to monetary policy, because it's not just the level, but the rate of change of prices, which can't jump. Unfortunately, they couldn't get the computer to solve it fully Taylorised, and had to do a bit of a fudge to get sort of the same effects.

I don't know NAOMI. Why do you prefer her to TOTEM brendon?

Interesting insight - i'm going to have to go back and review the Calvo and Taylor models with that in mind.

I prefer NAOMI because she's a simple girl, nothing flashy, a little more my speed.

Brendon: here's one way to look at Calvo/Taylor I find useful: in both models, 1/n firms changes its price each period. In Calvo since it's random, the price that's changed is the average price, while in Taylor it's the oldest price. So:

Calvo Pdot = (1/n)(NewP - P)

Taylor: Pdot = (1/n)(NewP - OldestP)

The comments to this entry are closed.

Search this site

  • Google

Blog powered by Typepad