Just because a shock is a bad shock doesn't mean it should cause a recession. A recession is a monetary coordination failure. Monetary coordination failures are caused by monetary policy.
Start with a Robinson Crusoe economy. By assumption, Robinson Crusoe always allocates his resources perfectly to maximise his expected utility given his information about the constraints he faces. There are "good" shocks that raise his (expected) utility, and "bad" shocks, that lower his (expected) utility. But there are no shocks that cause a coordination failure in this economy, because Robinson Crusoe doesn't have anyone else to coordinate with. And he doesn't use money, so a "monetary coordination failure" is doubly nonsensical in a Robinson Crusoe economy.
Recessions are a bad thing, but they are not bad in the same way as a bad shock in Robinson Crusoe's economy.
Every Saturday Canadian output and employment drop. And they drop again every Sunday. Every weekend, output and employment drop for two successive quarters days. Are weekends mini recessions? I would say "no". In those things we call "recessions", we don't just see a reduction in the volume of trade, output, and employment; we see something peculiar happening in many markets where goods are traded for money. In recessions, it becomes harder to sell goods (including labour) for money, and easier to buy goods for money, relative to normal times. And the economy just doesn't seem to be working as well as it normally works, to do the best it can to satisfy our wants, given available resources and technology. The increase in the number of unemployed workers is one symptom of this.
Recessions are a coordination failure, and a specifically monetary coordination failure. Weekends don't look like a monetary coordination failure. They don't look like a coordination failure at all; they look like a coordination success, where we all take time off work at the same time so we can work together and do other things together.
A recession is a monetary coordination failure with a reduction in the volume of monetary trade. A monetary coordination failure probably means that output and employment will be lower than they would have been without that monetary coordination failure. It does not necessarily mean they will be lower than they were in the previous period.
Whether a given shock causes a monetary coordination failure depends on monetary policy. And it is perfectly possible that a shock that would be a "good' shock otherwise could cause a monetary coordination failure, for some monetary policies. For example, discovering that gold cures cancer would normally be a good shock, but if the world is on the gold standard the increased non-monetary demand for gold would mean that all other prices would have to fall to prevent monetary disequilibrium, and if prices are sticky this won't happen quickly, so we get an excess demand for money, and a recession. The net result might (or might not) be worse for human welfare than if the discovery had been suppressed.
We should not blame the shock for the monetary coordination failure it causes under some monetary policies. We should blame monetary policy instead. And I can think of no obvious reason why there should be any correlation between the goodness/badness of shocks in a hypothetical perfectly coordinated Robinson Crusoe economy and whether or not those shocks are likely to cause a monetary coordination failure.
In 1996 the Canadian economy was hit by a big shock, when the federal government tightened fiscal policy, turning a fiscal deficit into a surplus. We can argue about whether that was a good shock or a bad shock. But it did not cause a recession. It did not cause a monetary coordination failure. But there are good reasons for believing that if monetary policy had been different, if the Bank of Canada had targeted a fixed exchange rate rather than 2% inflation, that same fiscal shock would have caused a monetary coordination failure. Because the Bank of Canada allowed nominal interest rates and the exchange rate to fall, so private spending and net exports could rise, and that could not have happened with a fixed exchange rate.
In 2008 the Canadian economy was also hit by a big shock. That shock was probably a bad shock for Canada. If the customers for your goods get into difficulties, that will lower your expected utility even if you respond perfectly to that shock. But there is no reason why a bad shock should necessarily cause a monetary disequilibrium. It depends on monetary policy. The monetary policy in 2008 was the same as monetary policy in 1996. The Bank of Canada targeted 2% inflation in both cases. But the 2008 shock did cause monetary disequilibrium. It did cause a recession. Or, rather, the same 2% inflation target monetary policy caused a recession when faced with the 2008 shock, but did not cause a recession when faced with the 1996 shock.
Why?
[For Bob Murphy; I'm trying to clarify something about my previous post to try to prevent us arguing at cross-purposes.]
Worthwhile Canadian Initiative 
If gold turned out to cure cancer… I love it. Your posts are so great, thanks for writing.
-Ken
p.s. I still think sectorial inflation (due to sector-specific negative supply shocks/constraints) plus downward nominal rigidity in other markets can really mislead an inflation targeting central bank, enabling it to hit its inflation target even when aggregate demand is insufficient due to exactly the sort of monetary coordination failure you discuss in this post, and that could explain Canada’s 2008-2009 “success” of hitting its inflation target even as NGDP plunged.
My guess is somewhere along the same lines as Majro and Ken (who are saying something similar in different ways).
I think you hit on it with the gold/cancer hypo shock. And price stickiness. It’s when money myths become too prevalent in the economy in question and the particular shock plays against those myths. The basic myth being I suppose that money itself is a good. Because you can’t eat it, live in it, nor (really) heat with it. Another myth I suppose is “I should be able to get a fixed 10% return when I invest my money and I will shop, cajole, and/or manipulate until I get it”. (Let alone the Krugman policy myths: confidence fairy, bond vigilante.)
Surely there are other alternatives besides frictionless optimum and monetary coordination failures? For instance, maybe information problems cause labor/goods/credit markets to fail to function optimally, and these market failures worsen during recessions. So there is a coordination failure, but it is not monetary in nature, or correctable through conventional monetary policy.
Jeff: well, money is a good, but not in the same way as all the other goods. It’s a good that helps us coordinate our trade in all the other goods. And sometimes it coordinates well, and at other times it coordinates less well.
jonathan: yes. But it looks to me like recessions are a specifically monetary coordination failure. Because it is money that gets harder to buy and easier to sell in recessions. Plus (AFAICT) barter trade actually expands in recessions.
In the Austrian model it seems that people come to the realization at some point in the trade cycle that they have been busy producing the wrong kind of things.
All the factoriess producing the useless stuff shutdown and the workers want to go back to their old jobs as wheat farmers. But they can’t becasue the factories producing farming equipment all switched to producing useless stuff as well. They restart farm equipment production but it will take a while to get the equipment ready. In the mean time there is unemployment and a drop in RGDP. This would happen even if prices were perfectly flexible.
The Austrian model may be flawed – but why would they be wrong to see this as an example of a recession caused by non-monetary reasons ?
Nick: It is true that during a recession it becomes harder to sell labor and goods (for money), and harder to obtain loans (of money). But it’s not clear to me that this is essential to the failure, or just reflects that in an economy that uses money as a medium of exchange, every transaction involves money.
This is especially true in a liquidity trap in which money and T-bills are equivalent at the margin. You could equally well say that the problem is that it’s hard to get T-bills, since under current conditions one can immediately swap one piece of government paper for another.
(I’ve also never heard that barter increases during a recession. Are you thinking of unemployed people exchanging favors?)
Nick, I love the bold face you often use: it amounts to a handy tl;dr version.
Nick, I think the answer is quite clear if you introduce the level of debt. Higher level of debt make monetary policy less efficient, independently that you change inflation for GDP level objective.
See http://www.miguelnavascues.com/2015/08/1996-y-2008-por-que-no-funciono-la.html
@Fiscalist:
In the Austrian view, why would a recession be consistent with falling prices? In the 2008 recession, the Bank of Canada did what it described as monetary easing in order to increase inflation and inflation expectations. In the counterfactual of “no change in the monetary instrument,” we would have experienced a collective fall in the price level and very arguably worse real output.
The Austrian view of recessions makes a lot of sense for “stafglation” recessions like that of the 1970s oil shock, where there really is a factor of production that becomes scarce. What happened in 2008 in Canada that would have been comparable?
@Miguel:
‘In the Austrian view, why would a recession be consistent with falling prices?’
Good question. I suppose the easy answer is “Hayek’s secondary deflation” – a recession caused by an Austrian supply shock turns into a recession worsened by monetary dis-equilibrium.
jonathan: “(I’ve also never heard that barter increases during a recession. Are you thinking of unemployed people exchanging favors?)”
To my mind, this is a crucial bit of evidence that supports the monetary disequilibrium theory, and is rather hard to explain otherwise. Unemployed people exchanging favours would be one example. I wish I had more data on this. Here is one old post I did linking to some anecdata (with links to more posts in that post).
Tom: I stole the bold face idea from blogger Razib Khan (who is a brilliant blogger BTW, even if a lot of stuff he writes goes way over my head).
jonathan: another old post on empirical evidence for countercyclical barter. Plus, see the post by Alex Tabarrok that I link to there.
Thanks Nick this is awesome. I’m not being lazy, but I think this article I wrote after my debate with Karl Smith still perfectly captures how I would respond to your post here. If you get time to check it out, I’d love to hear your feedback (here in these comments). If you were willing to do that, I will try to do something at my blog reconciling our two perspectives. I think it’s one of those “blind economists describing an elephant” situations.
Miguel: I had a read of your post. Who/what is “PP”? Did Spain do a big fiscal tightening in 1996, just like Canada? And did it also result in no recession? But Spain had its own central bank back then, right? Did it (de facto) keep inflation on target? And in 2011, with Spain in the Euro, there could be no monetary offset of fiscal tightening, of course, so it’s not surprising if things got worse for a bit.
In Canada, federal government debt/GDP was lower in 2008 than 1996. Not sure about private debt.
Nick: The examples of barter cited are interesting, but here is a counterpoint: if the problem is a shortage of money as medium of exchange, then why is the T-bill rate zero? Given that T-bills do not serve as a medium of exchange in transactions, and given that the interest rate on T-bills should equal the liquidity premium of money over T-bills, I would expect that if the economy were suffering from a medium of exchange shortage, there would be a positive interest rate. People would have to be compensated for holding T-bills instead of money.
Instead, it seems that T-bills are an asset that is in demand more than usual relative to money. This fits in well with another sort of shortage — a shortage of assets that can serve as a safe store of value. In this case, T-bills are serving as stores of value, and at the margin are equivalent to money. I consider this strong evidence that the problem is not a shortage of money for transaction purposes, but of stores of value. Or equivalently, excessive demand for future goods relative to current goods, and the intertemporal price (real interest rate) can’t fall.
Market Fiscalist-The Austrian theory is a monetary theory of recessions. People produced all the wrong things because of an earlier monetary expansion.
You’re reducing the Austrian story to a supply shock because you’re preoccupied with the bust. The Austrian story is a theory of the boom, not the bust.
In AD/AS terms, a rightward shift in the AD causes a leftward shift in LRAS but we temporarily move to the right on the SRAS curve. A demand shock causes a supply shock.
From the sound of it you may have ABCT confused with Kling’s PSST.
jonathan: suppose we used cows as money. An increased demand for milk would cause a recession, because people would stop spending their cows as quickly. And if goats produce milk too, but are not used as money, we would not expect the price of goats to fall.
If there is expected to be an excess demand for money in future too, so NGDP is expected to rise more slowly than usual, that will cause equilibrium nominal interest rates to fall.
The IS curve slopes the “wrong” way, not because interest rates have the “wrong” sign on investment demand, but because dI/dY + dC/dY > 1.
Low nominal interest rates do not mean money is loose. They mean money is expected to be tight in future.
Andrew: “From the sound of it you may have ABCT confused with Kling’s PSST.”
I sometimes get them a bit confused too.
Bob: I had a read of your post Karl Smith article. And this morning I read your other old Mises post on ABCT (linked in your potpourri). You are good at explaining it. But I am still not convinced it is a coherent theory.
But assume for the sake of argument ABCT is both coherent and correct. A collapse in foreign demand for your goods will (generally) require a change in the time-structure of production. A collapse in government demand for your goods will also (generally) require a change in the time-structure of production.
If one believes in the musical chair model, it is pretty clear that nominal income, not consumer prices, are at the heart of monetary coordination failures. Negative growth in nominal income can be especially problematic.
Given that real gdp growth can, for all kinds of reasons, be smaller than -2%, I don’t think that it’s surprising that 2% cpi inflation (which may or may not be related to the production of a small open economy such as Canada) was observed amidst the 2008 recession.
Nick-The main difference is that PSST has no special role for money. Money plays the central role in the Austrian theory. Which is why I found it odd that Market Fiscalist was talking about it as if it was about recessions unrelated to money.
Something I wrote from a long, long essay I should post online at some point:
“I suggest we should consider a recession as a time when a mismatch occurs between the individual economic actors’ plans for production and consumption, in such a way as to be unfavorable both to producers and consumers simultaneously.
From that definition, three sorts of recession that might hypothetically occur can be gleaned: 1. Unexpected, unfavorable reductions in the productive capacity of industry in general, such that consumers may have wanted more production, but adequate production proved physically or practically impossible. The real supply of goods is reduced 2. Producers produce an excess of goods in general, greater than the quantity consumers intend to consume. Products lie unsold on shelves. 3. The goods producers have produced are inconsistent with what consumers actually wish to consume, that is, there is not an excess supply of goods in general, but an excess supply of some goods and a shortage of others. Superficially recessions of this type resemble those of type two, as some products lie unsold on shelves. But other goods lie unproduced entirely, despite the desire of consumers to purchase them.”
Type one are supply shocks, and I would say you could have a recession due to, say, an asteroid impact or something that would effect all production at once. A supply shock for a single individual good would not cause a recession. Either way these are rare to non existent in modern, diverse economies, or at least they’re less important than types two and three.
Type two are General Gluts, they imply people in general desire higher real money holdings at the expense of nominal demand for goods.
Type three are the busts that follow Austrian style booms. Technically I suppose you could say PSST looks like this to. But I have a hard time understanding why a PSST style recalculation would become necessary without invoking a distortionary role for money, in which case PSST starts to become very difficult to distinguish from ABCT-but not impossible.
Nick, Interesting post. Suppose Canadian manufacturing produces components for US firms. The US goes into a deep recession (a la 2008). There is a big drop in demand for Canadian components, and no feasible price change for those components can maintain output, when the final goods factories in the US (say car factories) are shut down.
So it’s a re-allocation problem. The empirical question is how much smaller is the loss of jobs under monetary equilibrium (say NGDP targeting) as compared to monetary disequilibrium (say inflation targeting, or interest rate pegging or exchange rate pegging)
The experience of the US from January 2006 to April 2008 suggests that re-allocation shocks are not important at the macroeconomic level. US housing construction fell by more than 50%, and yet unemployment barely budged (from 4.7% to 5.0%) And even that overstates things, as NGDP growth slowed.
On the other hand the global recession of 2008 may have hit Canada’s economy harder (in a real shock sense) than the housing collapse hit the US. Housing was 6% of GDP at the peak, whereas manufacturing is much larger. So I have an open mind as to the extent to which monetary stability would have shielded Canada from recession in 2008-09. My guesstimate is that the slowdown (under NGDPLT) would have been only 1/4 to 1/3 as large as it actually was, which is of course much better.
Check out the bicycle video at moneyIllusion
@Majromax
Your example is of a supply shock. Price level targeting (with higher price levels in the future) is designed to deal with negative demand shocks and downward sticky prices. This is part of the argument for NGDP targeting.
@Andrew_FL
“The Austrian theory is a monetary theory of recessions”: I don’t doubt it. I was deliberately focusing on the co-ordination failures that occur during the bust phase of ABCT and ignoring the cause of the initial boom to see how that fitted with Nick’s views that all recessions are the result of monetary disequilibrium. I think Bob does a similar thing in his Gnome story that he links to above. Of course at heart ABCT is a special case of monetary dis-equilibrium theory but that wasn’t my point.
Nick says “A monetary coordination failure probably means that output and employment will be lower than they would have been without that monetary coordination failure”. I think this applies to ABCT and Bob’s Gnome theory.
In fact Bob’s gnome story is interesting in that it has a random supply shock totally unrelated to monetary policy.
Bob then draws the conclusion “Any attempt to short-circuit this[recovery] process with a wave of fiat inflation or deficit spending would simply screw up price signals and make the adjustment process that much longer.”.
So we have 2 polar views:
Nick: A monetary coordination failure probably means that output and employment will be lower than they would have been without that monetary coordination failure.
Bob: Any attempt to short-circuit this process with a wave of fiat inflation or deficit spending would simply screw up price signals and make the adjustment process that much longer.
For Nick you need monetary policy that consciously strives to avoid monetary dis-equilbrium, for Bob any attempt to do this just makes a bad situation worse!
Scott Sumner, I believe you are right.
In the case of Canada, it was not only manufacturing (cars in particular) that were hit, but also commodities. Case in point, exports crashed from 34.5% of GDP in 2008 to 28.4 % in 2009. (Imports decreased by a smaller margin 32.7% to 29.9%). That’s a very sudden fall in net exports (3% of GDP). I think that the fact that Canada is an open economy played a big role. It’s not impossible that the huge depreciation in the CAD in mid-2008 was enough to lift CPI inflation to the 2% trend, but was not enough to fully accommodate this huge exogenous shock.
Nick, I refer to private debt. It’s my fault not to have precise it. In any case, I agree with you except in you monetarist model. You say that the crisis of 2008 has change your mind, and prefer a NGDP objective rather than inflation rate. Good. But I think it is not very different, because the problem, the big problem, is that inflation or NGDP, both are ineffective when the debts have risen.
Market Fiscalist, the only way your contrast between Nick and me–which is clearly designed to make Nick come out as the better man–works is if you define a bunch of gnomes rearranging all the capital goods as “a monetary coordination failure.” That’s the whole point of my example, to show that the Keynesian (and by extension on this issue, Market Monetarist) framework forces you to diagnose drops in real output as due to bad monetary policy when, by construction, they have nothing to do with monetary policy.
Robinson Crusoe can take some time to think.
“The social function of liquidity is that it gives time to think.” Hicks, 1974, p.57.
Bob,
My comment certainly wasn’t intended to “make Nick come out as the better man” but merely to try and state the differences in your views of what is the appropriate monetary policy in the face of a supply shock
I do not think it true to say that the only way my contrast works “is if you define a bunch of gnomes rearranging all the capital goods as “a monetary coordination failure.”. Rather, I think that given a random supply shock (such as your Gnome example which is clearly not a monetary shock) Nick thinks that monetary policy aimed at preventing monetary disequilibrium will optimize the recovery process, while you appear to believe that any tinkering with the money supply (“wave of fiat inflation”) will mess with prices and de-optimize the recovery process.
Am I am mis-stating your views ?
BTW just to avoid confusion, since the casual reader might think Andrew_FL and I are contradicting each other:
In canonical Austrian business cycle theory, it is loose monetary policy that pushes interest rates below their “natural” level and misleads entrepreneurs into making unsustainable investments. This creates a temporary boom period that is an illusion.
Eventually the boom ends, typically when the central bank gets nervous about rising price inflation and so cuts back on the injection of new money. Interest rates spike and a bunch of businesses realize they are in trouble. The recession ensues.
However, the critical point is that once malinvestments occur during the boom period, the economy is now in a physically unsustainable configuration. The bust is inevitable at that point. So the boom is caused by monetary factors, but the bust is a “real” phenomenon. Once you are in the midst of a bust, pumping in more money isn’t going to fix it. At best (worst), it will keep the illusion of the boom alive for longer, allowing even worse malinvestments to occur so that an even more severe reallocation is now going to be necessary.
Thanks Bob.
At the risk of actually disagreeing, can you comment on whether you think “typically when the central bank gets nervous about rising price inflation and so cuts back on the injection of new money.” isn’t the central bank compounding one error with another? It’s too late for the central bank to undo the policy which put the economy in an unsustainable configuration, and real goods are scarcer than they appeared to be during the boom, and they’re also scarcer relative to money than they appeared to be during the boom. If the central bank “does nothing,”-or more accurately, stops doing something-prices should rise as output falls. If they don’t, it would make sense to say that the central bank compounded the problem, not because it tried to re-inflate (which I agree would be a mistake) but because it tried to reverse price inflation it had already caused?
That’s Hayek’s secondary deflation, I think.
@Market Fiscalist, you write:
“For Nick you need monetary policy that consciously strives to avoid monetary dis-equilbrium, for Bob any attempt to do this just makes a bad situation worse!”
Now how can we go about telling how it is for reality? (I’m not into this post-modernist “everybody’s reality is equally valid” nonsense) ;^)
Andrew_FL I’m talking about what usually makes the boom come to an end. Prices start rising too quickly, so the central bank slows down on its asset purchases, the monetary base stops rising so rapidly, and interest rates end up rising. Then the boom turns to a bust.
I think you are talking about something else, where they try to reverse the price inflation of the boom years.
Scott: “On the other hand the global recession of 2008 may have hit Canada’s economy harder (in a real shock sense) than the housing collapse hit the US. Housing was 6% of GDP at the peak, whereas manufacturing is much larger.”
In the 1990’s Canada went from a 5% budget deficit to a 2% budget surplus, which is 7% of GDP. That’s quite big, though it changed a little more slowly than the drop in US housing construction.
Nick,
I’m glossing over all this because I find the economics profession’s specific use of the term “co-ordination” to be inherently vague and mind-bogglingly amorphous. That may well be because I’ve spent little time looking at it.
I saw this post from David Glasner about one of your posts on the subject:
He says:
“Thus, the condition for macroeconomic coordination is that all agents have correct expectations of all currently unobservable future prices. When they have correct expectations, Walras’s Law is satisfied, and all is well with the world… actually, Nick admits that coordination failures can be caused by factors other than an excess demand for money, but for some reason he seems to think that every coordination failure must be associated with an excess demand for money. But that is not so. I can envision a pure barter economy with incorrect price expectations in which individual plans are in a state of discoordination.”
I also skimmed Murphy’s post, and that last sentence from Glasner reminded me of it.
Question:
Can you define the term “co-ordination” and/or the term “monetary co-ordination” according to how you see them interpreted in the broad economic context?
comment in spam I think
JKH: I fished you out of spam.
Fair critique. “Coordination failure” gets used in many ways.
It can mean where you have multiple Nash equilibria, some better than others, and we get stuck in a bad equilibrium.
I’m using it in a loose sense here. If you read Hayek’s essay on the use of knowledge in society (which is well worth reading quite apart from this post) you will get a sense of the fuzzy sense in which I am using it here.
Those particular coordination failure we call “recessions” seem to have a peculiarly monetary character, for reasons I state above. It’s easy to buy goods for money and hard to sell goods for money. So I call them “monetary coordination failures”.
I think I have a part of the answer.
Canadian Total Government Deficit started shrinking as of 1992:
1992 -8% of gdp
1997 ~0.8% of gdp
The average rate of tightening was 1.6% of gdp per year. The 1996 tightening was about 2.2% of gdp. A bit above average, but still in the same ballpark. In other words, austerity in 1996 wasn’t that large, in the context of the 1992-1997 overall provincial+federal fiscal tightening.
Nick, from Uneasy money, cited by JKH,
“…In a monetary economy suffering from debt deflation, one would certainly want to use monetary policy to alleviate the debt burden, but using monetary policy to alleviate the debt burden is different from using monetary policy to eliminate an excess demand for money. Where is the excess demand for money?”
So an excess of debt can make ineffective al MP guided by. Inflation objective or by a NDGP one.
Thanks for the clarification Bob, yeah, we are talking about slightly different things.
What was different about 2008/09? A possible hypothesis for consideration:
There was a coordination failure in the Asset Backed Commercial Paper market when that market effectively froze solid – ABCP being something of a shadow banking system member. ABCP was a popular place for firms to park short-term cash for the obvious reason – higher returns. Monetary policy that acts on banks had no impact on the ABCP market. There were signficiant funds frozen in the ABCP market in 2009 – take Nav Canada as an example where money that was targeted for capital projects was suddenly frozen causing capital projects to be delayed causing a drop in economic activity.
Companies caught in the ABCP freeze were not about to go to their bank to finance these activites as they were facing the possibility of losing their capital investment in ABCP – risk management would have said to retrench.
It tooks months for the ABCP market to get settled. Long enough to do real damage to economic activity.
Possibly there were several shocks hitting Canada at the same time: ABCP market freezing plus a recession in our largest trading partner.
@Nick, Thanks for responding, and agreed money is a “meta-good”, not to be confused with “goods and services” in the usual sense. More oil than gas, in an ICE metaphor.
Jeff: It’s an older metaphor than that: “Money… is none of the wheels of trade: it is the oil which renders the motion of the wheels more smooth and easy.”
David Hume’s essay is well worth reading.
Nick, you said “But I am still not convinced it is a coherent theory.” Why don’t you write a post on why you think ABCT might be incoherent?
Keshav: it’s an idea. But I’m afraid my post would just be “I can’t figure out how ABCT is supposed to work”, and would be followed by lots of comments telling me to read XYZ, or trying to explain it to me.
If I can figure out a more constructive post, I might do it.
Kathleen: “It tooks months for the ABCP market to get settled. Long enough to do real damage to economic activity.”
Suppose Canada never had an ABCP market. Presumably that would be worse for Canada, but it wouldn’t have put Canada in a permanent recession with higher unemployment. Maybe the sudden disappearance of the ABCP market caused search unemployment, while workers reallocated to different firms. But the same thing would have happened in 1996, as workers needed to reallocate to the export sector.