« Because it's not summer if you don't go swimming in Lake Couchiching | Main | Asset price bubbles, monetary policy, and the Lucas Critique »

Comments

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

"Did a change in monetary policy cause the house price bubble? In Canada, absolutely not... The Bank of Canada hit its inflation target almost exactly..."

I'm not arguing that monetary policy in Canada alone caused a price bubble, it's obvious from the charts that price fluctuations can be a local phenomenon. We have the same monetary policy in Calgary and Ottawa, yet the charts are completely different. But my answer to your statement above is, "So?". How can you conclude monetary policy wasn't wrong simply because CPI was a certain level nationally? How can you know that level was appropriate for a given local market? The same argument for the US, too. California vs. Texas are two different worlds.

[sorry if this shows up twice, I attempted to post it before but it hasn't shown up]

Hi Nick

What of the argument (approximately that made by HB in the comments to your previous posting) that the natural rate is not necessarily, by definition, the rate which causes the Bank to meet its target? If the natural rate is actually defined as the rate consistent with a non-accelerating rate of inflation (along the same lines as NAIRU) then the definition of inflation becomes important in determining the natural rate, and thus in seeing whether the policy rate was correct.

If the "correct" rate of inflation - if such a thing were definable - is the one including owner imputed rent, then perhaps the policy interest rate was less than the natural interest rate. Because of inter-sectoral changes such as technological progress and the availability of cheap tradable goods from China, the distribution of price changes was uneven: core CPI stayed low but the monetary impact wound up in house prices instead.

This is not to say that the Bank of Canada was not doing its job - it met its target after all - but that its job was wrongly defined.

I am not expert enough in this area to say how strongly I support that argument, but it seems plausible. And even if it is correct, I don't know whether higher interest rates would have been the right response - there are other tools available to the central bank.

If this argument is correct, then perhaps the "cause" of the house price rise was not a change in monetary policy but a failure to change it in response to a change in external factors. Scott has discussed on his blog the philosophical implications of whether this is actually "causal" but I don't think we need to get into that for this purpose.

The questions might then become:
1) was a change in the price of houses relative to other goods inevitable
2) if it was, would it have been appropriate to allow deflation in other goods to keep inflation-including-houses at 2%?
3) even if the change in relative price were inevitable, presumably it overshot (that's what a bubble is, after all). What tool could the central bank have used to control this overshooting?

[pointbite also makes an interesting related point since my first attempt to post: can or should the central bank attempt to mediate between differential local rates of inflation?]

From your previous post:

"anon@6.35: I'm not sure if I follow you there."

You wrote:

“If the central bank sets an interest rate below the natural rate, inflation will rise relative to expected inflation ... It is economic nonsense to blame a central banker for keeping interest rates too low, unless you mean to blame him for keeping interest rates too low relative to the natural rate.”

This is a contradiction.

The argument is that Greenspan kept rates too low and therefore caused a housing bubble. You suggest that the effectiveness of monetary policy relative to the natural rate is tested by the CPI result. But CPI does not include housing prices. It includes owners’ equivalent rent, which is VERY different in behaviour. There are all sorts of examples of negative correlation between housing prices and OER.

Therefore, it is perfectly reasonable to blame monetary policy for housing prices even though monetary policy is effective relative to a natural rate which in turn is dependent on the measure of the CPI inflation indicator. Those who blame Greenspan are obviously inferring that he should have changed monetary policy to account for housing prices, and not followed a “natural rate” guideline where the natural rate interpretation is a function of an OER inclusive CPI.

The comparison with Canada misses the point. The essence of this is the whole debate about whether central bankers should include asset prices more formally in their monetary policy formulation, Canada or the US, or whether they should attempt to pop asset bubbles. Then the observation of housing prices becomes part of the equation, regardless of country specific reasons for housing price behavior.

pointbite: yes, good point. We could check provincial CPI data to see if it backs you up. I suspect it might. But then given we have one money in Canada, there's nothing the bank of Canada could have done about it. But then I would reframe the debate by asking what the real shocks were that caused house prices to rise in some provinces and not in others. And I would further add, even if we did have 10 monies, and provincial central banks (like in the Eurozone), would those real shocks still have caused *real* house prices to rise by the same amount.

In other words, just because I cannot prove that the real BC house price increase would have been the same even with a made in BC monetary policy, neither can anyone else prove the opposite. And AVERAGE house prices did increase in Canada, even though AVERAGE monetary policy was correct.

Given that house prices increases were much smaller in Canada than in the United States, the Canadian example would in fact argue for a big role for CRA.

Not that I really think CRA had a big impact. Instead the most important explanation to why house price increases were much more modest in Canada is that the Fed pursued a more inflationary monetary policy than the Bank of Canada by cutting rates more aggressively

Leigh: After Friedman won his battle over the natural rate of unemployment, some of his opponents grudgingly conceded defeat, but renamed it the NAIRU (non-accelerating-inflation rate of unemployment).

There were some good reasons for this:
1. "natural" implies it is both God-given and good, and the natural rates are neither.
2. NAIRU seems a simpler definition.

But I resist this change of names.

First, NAIRU is not really simpler. "Natural" rates are inherently theoretical constructs, and cannot be defined independently of a class of model. A model in which real variables are independent of certain aspects of monetary policy is a natural rate model, and natural rates of unemployment, interest, etc., are only meaningful within that class of models.

Second, NAIRU is a highly misleading concept too. In 17something David Hume showed that long run real variables are independent of P. In 1967 Milton Friedman showed that long run variables are independent of Pdot (the derivative of the price level with respect to time, i.e. the rate of inflation). As others have remarked, it took the economics profession 300 years to slip one derivative! But once we had seen the slip from P to Pdot, and added rational expectations, it didn't take as long to realise that the same could be said of Pdoubledot, or Ptrebledot, or indeed any predictable time path for P, however convoluted. (Yet some are still stuck on what I call "the fetish of the first derivative").

Now, given neutrality, and super neutrality, and super-duper neutrality (as many dots as you like after the P), we can use exactly the same argument to say that in the long run, when prices are flexible and expectations adjust, to show that the choice of price index to target is also irrelevant. If real wages are rising at 3% per year (say), then targeting 0% P inflation is the same as targeting 3% W inflation, and since targeting 0% P inflation is the same as targeting 3% P inflation (by superneutrality), then it follows that targeting 3% P inflation is the same as targeting 3% W inflation. So, by extension, it doesn't matter in the long run which price index you target, when all prices are flexible and expectations adjust.

To get around that argument, so it does matter what price index you target, you need to argue that some prices are more flexible than others, or that people are better informed about some prices than others, or that some relative prices are more subject to real shocks, or something like that. And on those sorts of grounds, asset prices generally are not good candidates for consideration. They are usually very flexible, easy to get information on, and very volatile in real terms. Maybe house prices are an exception to that.

How would this work? Mortgage rates come down, say. Now, for some people, it will indeed be prudent and wise to buy a house when mortgage rates come down. However, it will not take long for this prudent investment to become dicey if home prices rise faster than other prices. Apparently, Glaeser says that Interest Rates can only account for at most 20% of the housing bubble. Even that seems high to me.

So, at the beginning of a bubble, there will exist a certain amount of prudent and sensible investing. Can we really call that the cause of a bubble? It certainly comes first in time, and often leads to home prices going up. But I can't, in good conscience, call this prudent behavior a cause of the bubble.

I think that Angelo Mozilo, yes, that Angelo Mozilo, gave my view of the bubble when he testified before congress:

http://oversight.house.gov/documents/20080307121803.pdf

"In June 2004, however, the Fed commenced what turned out to be 17 consecutive
interest rate increases. The combination of increasing interest rates and higher
home prices initially prompted a still higher spike in demand, as many borrowers
rushed to buy homes for fear of getting priced out of the market."

There's no way to make this prudent. In fact, it takes a cock and bull story to justify it. In my view, the story went like this:

1. You cannot count on the govt for retirement.
2. You cannot seem to save for retirement.
3. You can buy rather than rent, and turn a cost into an investment.
4. Housing prices are not coming down, so, if you don't but now, then you never will.

It's a story of fear and panic, the fertile grounds where fraud and foolishness can plant themselves. Oddly, Mozilo gets closer to the truth than most economists.

anon@12.07: Thanks. OK, I understand you now.

I understand that owners equivalent rent does not capture house prices fully (though I wish i understood a bit better exactly what it does do).

And it might be a good idea to broaden the definition of inflation in the inflation target to capture house prices better. But it is not clear to me that a different definition of inflation for the inflation target would have made a difference. This is the same point I made to Leigh. Sure, if we changed the definition of inflation we would change the Bank's response to inflation as currently defined. But we would also change expected inflation as currently defined. And expected inflation ought to change by exactly the same amount as actual inflation. So it ought not change the gap between actual and expected inflation, and it is that gap which defines monetary stimulus, and creates the short run effects of monetary policy.

I'm just invoking the Lucas Critique in a different context. If we change the target, expectations will change too. So we cannot say that monetary policy would have been tighter in the face of the house price bubble, and real interest rates higher.

Stefan: I'm not sure whether Canadian house price increases were so much less than in the US. Stephen has a graph showing both here: http://worthwhile.typepad.com/worthwhile_canadian_initi/2009/06/the-housing-market-the-nonstory-of-the-canadian-recession-.html
I think it depends on the time period. Canadian data don't cover the whole country as well either.

But in general I accept your point. There may well be multiple causes, with different degrees of relative importance in different countries.

Don: yes, there is indeed the whole separate question of the precise mechanism whereby setting interest rates below the natural rate cause a bubble. You can see why it would cause real effects, and inflation to accelerate relative to expected inflation, but why precisely a house price bubble?

Central banks try to set the rate of interest equal to the natural rate. Since they don't observe the natural rate, they will almost always get it wrong. 49% they set it above, 49% they set it below, and 2% of the time they get it just right. We ought to see housing bubbles 49% of the time. And crashes the other 49%. If the economy really were that fragile, we would all be dead long ago.

Nick, isn't looking at the "average" monetary policy the equivalent of looking at debt-to-gdp figures? I think you told me that chart was "worse than worthless" or something similar, because only the de-aggregated figures matter.

Nick,

might it not be that the housing bubbles in both the US and Canada (and the UK) where all caused by the monetary policy of one country?

That country of course is China.

Basically, OER inflation correlates closely with the level of interest rates themselves - which means the more successful CPI based monetary policy is in general, the lower and more stable interest rates are, and the lower OER inflation is. This is a real problem in that low interest rates mean cheaper leverage and higher house prices. The OER measure is counterproductive for house inflation control.

Central bank interest rates would have been a lot higher if they were intended to control house price inflation. This is a big problem in discussing house prices and monetary policy.

pointbite: when we aggregate over agents, take the average, and look at the effects of average monetary policy on the average person, we do indeed lose some information. But when we look at average debt, we lose all information, because the average agent has zero net debt. So I don't think it's the same.

Adam: I don't know. The accumulation of forex reserves by the Bank of China is (as far as I can tell) a farm on national savings, and is really more like what you and I would call fiscal policy, rather than monetary policy, even though it's done by a central bank. And then, if we add in all the other countries that have high national savings, you get back to the standard "savings glut" theory, according to which the natural rate was very low, and a low natural rate means a high equilibrium Price/Earnings ratio for assets, which in houses translates into a high Price/rent ratio.

But then that's a theory of a high *equilibrium* price for houses, as opposed to a disequilibrium bubble. Though maybe one could lead to the other.

But maybe that's not the theory you had in mind?

Should read: "...a form of national savings.."

anon@4.48: Aha! that makes sense. Thanks for clarifying. In which case, OER has a rather weird property, in that a low or falling OER may be a symptom of loose or expected loose monetary policy. At an extreme, if the Bank of Canada were targeting only OER, for example, the result might be unstable. Lower interest rates cause falling OER, which the Bank responds to by lowering interest rates further...

Yes, OER has a very weird property. Its essentially the same property that precludes central banks from simultaneously controlling CPI and asset price bubbles - be they houses or equities. Success with CPI based monetary policy contributes to asset bubbles because the result of success is lower and lower interest rates and discount rates. Until the bubble bursts from deflation risk.

The Fed did raise interest rates. Isn't that what they bubble bursters wanted the Fed to do? I suppose that you can argue that they didn't do it fast or far enough, but look at the economic view of the head of the largest home lender in the US. He claims that higher interest rates accelerated demand and caused housing prices to climb. For three or four years!

You can't separate the price of homes from the lending practices and demand. The only way to deal with the housing bubble was to Increase down payment requirements as the price of homes rose relative to other goods. That's prudence. The Fed is a leaner of last resort. I wish you all luck trying to explain the housing bubble using Monetary Policy. You've got a lot of explaining to do.

One thing that should have been a sure sign that something was amiss was the widespread belief that owning a house was a free lunch. In other words asset appreciation would forever out pace inflation + debt servicing + maintenance. Thus all you had to do to be rich was own a house. Of course, there are no free lunches. Makes me wonder if there was perhaps an information problem in the housing market. Is there transparent way (i.e. not using crazy derivatives)to sell the housing market short?

And Mozilo is wrong. It wasn't the fear of being priced out of the market for shelter that drove the mania, it was the fear of being left of the apparent gravy train.

Nick,

I was thinking currency manipulation as a form of monetary policy.

There was a time when scarcity of resources determined the economic level of prosperity. During the era of Big Cheap Oil, expectations regarding progress (in the broadest possible sense) increased like never before in economic history. Have we factored in a steadily decreasing supply of especially those resources that fueled (no pun intended) this expansion? Not if that is seen as undermining confidence in our seemingly innate ability to "create wealth".

Housing prices should be ignored as a measure of confidence and faith in economic conditions. Better indicators would be:
how closely the education system fits in with employers' expectations, how much energy and non-renewable materials are consumed in critical processes such as transportation, agriculture, manufacturing, medicine, education, etc.

A frugal and judicious use of resources may avoid a "return to the stone age". The scaling down of our expectations will happen whether we choose to collaborate or not.


BTW, China had little to do with the US housing bubble.

The bubble was caused by low "teaser" mortgage rates, which later exploded into high reset rates.

The teaser rates were short term rates and a direct function of Federal Reserve policy.

This appears relevant:
http://www.ritholtz.com/blog/2009/07/michael-lewis-the-man-who-crashed-the-world/

"I'm convinced that our input into the system led to a substantial portal of the increase in housing prices. We facilitated a trillion dollars in mortgages, just us." - Former AIG employee

"The bubble was caused by low "teaser" mortgage rates, which later exploded into high reset rates."

There were housing bubbles all over the world, suggesting that the cause had to be either behaviour by a central bank in a country that could influence the whole world (only the U.S., I suspect), some sort of action taken by many/most central banks around the world (targetting core cpi and ignoring asset prices, maybe) or else some sort of global psychological effect (which seems unlikely given that it missed a few areas like Germany and Canada east of Toronto).

"Did a change in monetary policy cause the house price bubble? In Canada, absolutely not. There was no change in monetary policy in Canada."

Interest rates were at 20% about 28 years ago, and have generally been coming down ever since. They were extremely low (2.25%) when Canadian housing prices (in the West, anyway) entered into their steepest, bubbliest ascent. Rate increases back up to 4.75% started prices back downwards, but the subsequent drop of interest rates to 0 seems to have arrested and perhaps reversed this decline.

You can argue that monetary policy has no impact on housing prices in Canada, but I can think of at least half a dozen instances of people who bought in the last month or two solely due to the interest rates having come down so low (and I don't know that many people). See here for a sad but typical case. Something doesn't connect here.

From the article:

"What really helped? The 2.75-percent interest rate they were offered. It ultimately allowed them to move from a $1,800-a-month apartment into their own home.

'But we don’t have a lot of [wiggle] room,' Morettie said. 'We can go up to four percent, but then we’re done.'"

Nick:
Friedman may have popularized this concept, but you can read a virtually identical exposition in Mises's Money and Credit.

Nick you write:


Did a change in monetary policy cause the house price bubble? In Canada, absolutely not. There was no change in monetary policy in Canada.

I don't find this claim very convincing. Its reasonable ceteris paribus, but such an assumption is begging the question. In particular, you describe a policy based on a particular CPI statistic not the theoretical notion of inflation--the latter being the actual variable of interest under the theory. The CB targets a proxy only. Thus, its quite plausible that even as the bank nailed its CPI target, inflation was swinging wildly.

*Indeed* I can argue that it was surging uncontrollably and without a clear future time-path. I need only define house-prices to be my inflation metric. Not fair? Well you need to defend your CPI metric--and I think that's an activity that will fail. CPI's clear utility is as a standard-of-living measure, but its use as a proxy for establishing a natural-rate policy is vulnerable to monetary pressures escaping in unmonitored variables or being masked by other price-changes.

Jon: Agreed on Mises. I did say "It is true that Mises and Hayek for example had previously made essentially the same point, but Friedman made it more clearly, and had a much greater influence on our way of thinking." But I expect clarity is in the eye of the beholder.

But the CPI statistic is not in this case arbitrary; it is not just the most widely cited index, it is the very price index that the Bank of Canada has a very public commitment to target. It is therefore the price index that people expect, and should expect, to have grown at exactly the rate it did grow. So if you define a change in monetary regime as a deviation of policy from a prior commitment, and hence a violation of legitimate expectations, there was no change in monetary policy.

Now, would things have been different if the Bank of Canada had had a different definition of target inflation, and people had formed their expectations around that different definition? I don't know. But the argument that including asset prices in the index would have lead to a "tighter" monetary policy is invalid, because expectations-formation decision rules would have changed too under that different regime. It is subject to the Lucas Critique. That's what I was driving at in my later post.

Declan: but much of the fall in interest rates over the last 30 years has been a fall in nominal, not real interest rates. And are changes in real interest rates caused by the Bank of Canada, or are they just the result of the Bank of Canada correctly following the natural interest rate down?

But yes, that "first-timers" article is worrying!

And none of this is to say that stupidly loose lending standards (zero percent downpayment in a country like the US with non-recourse mortgages!!!) were not both stupid, but also had nothing to do with house prices.

Nick:
I disagree, just because people's expectations about CPI are meet does not mean that their expectations about 'inflation' are met. This is precisely the problem and the insight that cuts the Gordian knot.

Part of people's expectation is that CPI is a good proxy for 'inflation'. When this is violated, policy has real effects. In practice this is easy to violate because the components of CPI are highly wage dominated and thus narrow.

Jon: I can't quite clarify my thoughts on this to explain properly why I think you may be partly right but mostly wrong. Maybe I will succeed in getting my head clearer, and write a post on it if I do.

Nick: Please do.

A short time later, White argued for his model once again in a working paper titled "Is Price Stability Enough?" Low inflation rates are not a sign of normalcy, he warned, and central banks should not allow themselves to be led astray by low rates. Both the LTCM bankruptcy and the collapse of the stock markets in 2001 occurred "in an environment of effective price stability." (page 5)
...
White made one last, desperate attempt to bring the central bankers to their senses. "Virtually no one foresaw the Great Depression of the 1930s, or the crises which affected Japan and Southeast Asia in the early and late 1990s, respectively. In fact, each downturn was preceded by a period of non-inflationary growth exuberant enough to lead many commentators to suggest that a 'new era' had arrived," he wrote in June 2007 in the BIS annual report. (page 5)
...
As an adviser to German Chancellor Angela Merkel's group of experts, White helped to shape the basic tenets of the new order. And the 79th annual report of the BIS, published in Basel last week, also reads like pure White. It lists, as the causes of the crisis, extensive global imbalances, a lengthy phase of low real interest rates, distorted incentive systems and underestimated risks. In addition to improved regulation, the BIS argues that "asset prices and credit growth must be more directly integrated into monetary policy frameworks." (page 6)
...
He is familiar with human nature, and he knows how to handle it. White is more concerned about the things he doesn't understand. New Zealand is a case in point. Interest rates were raised early in the crisis there, and yet the central bank was unable to come to grips with the credit bubble. Investors were apparently borrowing cheap money from foreign lenders.

This is the sort of thing that worries him. "That's when you have to ask yourself: Who exactly is controlling the whole thing anymore?" (page 6)

http://www.spiegel.de/international/business/0,1518,635051,00.html
"THE MAN NOBODY WANTED TO HEAR - Global Banking Economist Warned of Coming Crisis"

"The Bank of Canada did not set interest rates below the natural rate, and yet still had a big increase in house prices. "

Well no that's not true at all. Our interest rates were artificially low.

And while we had no CRA, we did have tax incentives to use our registered retirement savings as down payment's on houses. Down payment rules dropped from 25% down to 5% down in a matter of 2 years, while mortgage rates where pushed down by the this action and the BoC dropping the rate (to fight inflation) so that mortgage rates dropped from about 9% to about 4 to 5% for consumers. In a matter of two years. We even flirted with 0% down for about 6 months.

The result was a housing boom like the US, only not as hot and not as risky, but a central bank induced bubble nonetheless.

Sorry, but you are simply wrong.

Mike: Oh dear! Sorry, but I am NOT *simply* wrong. (I might be wrong in a complicated way of course!)

The Bank of Canada's interest rate was lower than normal, yes. So is the price of computers, or fresh asparagus. That doesn't mean it was "artificially" low. (What do you mean by "artificial" anyway? I spent the whole post arguing that BoC interest rates were NOT artificially low, and explained what I meant by that.

"Ours the Task Eternal".

(Actually, it was the previous post, the one I linked to, where I argued that BoC interest rates were not artificially low: http://worthwhile.typepad.com/worthwhile_canadian_initi/2009/07/greenspan-and-his-critics-again.html )

Interesting post. But I never understand the concept of natural rate unemployment, natural rate of interest, etc. These are very subjective and additionally these keep changing constantly. Beacuse this concept is so fluid, it becomes difficult to even give a narrow range to these rates.

Thanks Sudhanshu. If you have an economic model in which long-run equilibrium values of real variables are independent of some aspects of monetary policy, then you can call those long run equilibrium values "natural rates". They do vary over time, when real shocks hit, and cannot be observed (except ex post, and then only roughly), but they are not really "subjective", though they are theory-dependent. I think the key is to try and find a way to conduct monetary policy that does not depend on our being able to observe natural rates in real time.

The comments to this entry are closed.

Search this site

  • Google

    WWW
    worthwhile.typepad.com
Blog powered by Typepad