Monday, September 27, 2010

It's the Housing, Stupid

[In case you weren't old enough in 1992 to get the cultural reference, "it's the economy, stupid," was a phrase used by Bill Clinton's campaign workers in the 1992 election to keep everyone on message.]

Why Krugman keeps beating a dead horse, as he does here this morning, is beyond me. The man is a little slow, he's wedded to a one-good, sticky-price, sticky-wage view of the world, or maybe he just hates Narayana Kocherlakota? You tell me.

First, it would be nice if discussion of this issue did not revolve around the idea that there are two kinds of unemployment: "structural" and "demand-deficiency" unemployment. Unemployment is unemployment. It's the estimate of the number of people in the economy who are actively searching for work, according to the Bureau of Labor Statistics. Unemployment necessarily involves a friction. If I am unemployed and know that my quest for work on a particular day is going to be fruitless, I will not search. The reason I search is that I expect to eventually come up with something, but it may take some time. Thus, by definition, unemployment (and vacancies, on the other side of the market) is always about mismatch - firms and workers are trying to make good matches, and that takes time.

Now, some people, including Narayana Kocherlakota and yours truly have argued that there is something anomalous going on in the current recession (also see Dave Andolfatto's post). The unemployment rate is unusually high, and we can find good reasons for that in the data, that have to do with the sectoral reallocation of labor, both across sectors of the economy, and across geographical areas. Long-term shifts from manufacturing to services, combined with the housing boom and bust, have acted to make the unemployment rate higher, as have the facts that the long-term shifts and the housing market boom and bust affected different states disproportionately.

However, blind Paul asserts this:
But don’t bother asking for evidence that justifies this bleak view. There isn’t any. On the contrary, all the facts suggest that high unemployment in America is the result of inadequate demand — full stop.
Then, he tries to discredit people who are thinking about sectoral reallocation as an important phenomenon.
Now, the Minneapolis Fed is known for its conservative outlook, and claims that unemployment is mainly structural do tend to come from the right of the political spectrum.
Ah, so those guys are probably just Republican hacks. No need to pay any attention to them. Now, I once worked at the Minneapolis Fed and have had plenty of interaction with the Bank over the years. Some of those people are Republicans, and some are Democrats. One thing I am sure of though, is that you would be thrown out of the seminar room for presenting ideologically-driven research. That place is all about the economics. No chicken models allowed.

Further, here's the Kocherlakota-hating Krugman at work in a blog piece called My Problem with Narayana Kocherlakota.
He has too many characters in his name! I mean, folks, the regular column has to fit into a limited physical space — and every time I mention the president of the Minneapolis Fed, there goes a large chunk of scarce journalistic real estate.

I want a new rule: people who weigh in on policy debates should henceforth all have names like Ng or Ip. Yes, you can call me Kn if necessary.

Or can we just drop all the vowels?
Now, Krugman may have thought of this as good fun, but it certainly looks racist to me. Apparently Narayana should have a nice American name, maybe something like "Paul Krugman." My name has a lot of characters too, but I'm sure Krugman wouldn't be complaining if he had to type it out repeatedly.

Krugman claims that
Well, I’d respectfully suggest that Mr. Clinton [who also thinks sectoral reallocation is important, apparently] talk to researchers at the Roosevelt Institute and the Economic Policy Institute, both of which have recently released important reports completely debunking claims of a surge in structural unemployment.
I looked up the Economic Policy Institute reference, here, and this is what Lawrence Mishel has to say:
A better explanation for high unemployment is that there are simply not enough jobs to go around. The economy is operating far below its potential output because of a shortfall in demand caused by an extreme loss of financial and housing wealth, and the reduced consumption that resulted.
Hmmm. Very helpful. Not enough jobs to go around. Sectoral reallocation thoroughly debunked.

As usual, Krugman also harkens back to the Great Depression.
I’ve been looking at what self-proclaimed experts were saying about unemployment during the Great Depression; it was almost identical to what Very Serious People are saying now.
We've seen this "Very Serious People" reference before. These are apparently those know-it-alls who think they have economics down, but unfortunately fall far short relative to the Truly Wise One. It's pathetic when Mr. Nobel Prize resorts to anti-intellectual Tea-Party rhetoric.

Well, I think that Paul Krugman is a Very Unserious Person. Paul, get off the pot and show us where this demand-deficiency unemployment is coming from. According to you, the right model is IS-LM. In that framework, the demand deficiency is due to sticky prices, but I'll give you sticky wages too, if you want. How does that explain what is going on in the economy currently? Where are the sticky wages and prices? Are unemployed construction workers in Nevada unemployed because the prices of houses in Nevada are sticky, because their nominal wages are sticky, or what? Did firms go out of business during the recession because they could not bear to lower their prices? The IS-LM I was taught did not have sub-prime mortgages and financial crises in it. Maybe Krugman was taught a different IS-LM? Can we employ workers in Nevada by creating government jobs in Washington, D.C.? I'm all ears.


  1. Professor Williams,

    I enjoy your blog. But I think you should write less about Professor Krugman. It is getting tiresome.

  2. You seem to particularly out of sorts today.

    First, where is the data part of your response to Krugman?

    Second, what a cheap shot calling Krugman racist! Do you know who he is married to? Sorry Steve - your little gambit to tar him with that label just won't fly. Anyway, does being a `serious modern micro founded macro person' mean they strip of your sense of humor?

  3. If I am unemployed and know that my quest for work on a particular day is going to be fruitless, I will not search ... Thus, by definition, unemployment (and vacancies, on the other side of the market) is always about mismatch

    Oh, please. This is beyond reaching. If I'm unemployed, I will of course keep searching because I have no choice as I need to find some way to support my family, and in any case I need to provide evidence of doing so in order to collect unemployment insurance. And if I give up looking and am no longer counted among the unemployed, most people would still consider this an economic problem. No one is going to accept this weak attempt to win an argument by redefining the terms.

    If you cannot engage Krugman's arguments rationally, you would really be better off leaving them alone.

  4. Anonymous 1: It's Williamson, not Williams. It's a dirty job, but someone has to do it.

    Anonymous 2: So, who I am married to can give me license to make racist comments? A name is a very personal thing, obviously. Canadians are very much on the tolerant side, but I'm sure that, growing up in Winnipeg, Narayana had to endure some flack because of his name. The fact that you appear to think Krugman's post was harmless fun is appalling. To everything there is a season. I can be funny too, but some things are not funny.

  5. "No one is going to accept this weak attempt to win an argument by redefining the terms."

    All I did was tell you the actual BLS definition (roughly) of what unemployment is. Apparently you are reading something into the fact that I actually want to correctly define the terms we are arguing about. Once we know what unemployment is, then we can ask what we do about it, which is another question entirely. Krugman thinks the answer to the question is obvious. I don't.

  6. What if there is both structural & cyclical unemployment?

  7. I like it. But, one question, don ´t you think that perhaps a bit of more QE would reduce the frictional caracter of Labor market? I mean that an inflation rate of 1% (and abating), is not a good context to create more jobs, I think.
    in any case, I like it.

  8. The term unemployment is already defined quite clearly by the BLS. It requires only that one have no job and be looking for one. There is no requirement that the searched-for job actually exist. You may believe that unemployment is always a matching problem. You may even have evidence, that you have not shown, that this is true. But there is nothing in the definition that requires it.

  9. ` So, who I am married to can give me license to make racist comments? A name is a very personal thing, obviously. Canadians are very much on the tolerant side, but I'm sure that, growing up in Winnipeg, Narayana had to endure some flack because of his name. The fact that you appear to think Krugman's post was harmless fun is appalling. To everything there is a season. I can be funny too, but some things are not funny.'

    The sanctimonious tone really is too much. Whether you are appalled or not, the comment IS funny. There is nothing racist in it whatsoever. To say that it is puts you in the ranks of the PC police and the other left wing crazies. Stick to economics.

  10. "What if there is both structural & cyclical unemployment?"

    As I said, it's all unemployment. I think the question you want to ask is: How much of unemployment can we explain as being due to: (i) sector-specific factors; (ii) aggregate factors, including aggregate productivity, government intervention through monetary and fiscal policy, and financial factors; (iii) sticky prices and wages.

    "There is no requirement that the searched-for job actually exist."

    Of course. But the person searching has to think that the thing exists, otherwise he or she would not search.

    "The sanctimonious tone really is too much. Whether you are appalled or not, the comment IS funny."

    So you might like someone having fun with your name? Maybe you think Barack Obama, Franco Modigliani, and Albert Pujols have pretty funny names as well?

  11. So you might like someone having fun with your name? Maybe you think Barack Obama, Franco Modigliani, and Albert Pujols have pretty funny names as well?

    My goodness. In the future we send all attempts at humor through you to make sure they are ok. Just admit that you screwed up by reaching too far and move on. Confess and move on. Is it that you don't have the goods to definitively to refute the main substantive hypothesis that Krugman raised? If you do, let's see the data and the argument. (Saying you are modern macro guy and Krugman isn't won't cut it.)

  12. Obviously we have different opinions about where you draw the line on humor. If you are interested in the substance, read this and see what you think:

  13. alright, i think it is better to stop. anyone has met Steve in person cannot deny he is a respectful gentleman. obviously he just feels Kocherlakota is not respected and said a few words for him.

    let have a look on previous posts. there are somethings substantive enough for resuming the discussion on economics. should we go back monetary policy or interest on reserve?

  14. Compliments are certainly always welcome. Yes, exactly. Kocherlakota is a thoughtful, intelligent, first-class economist. His graduate training was at Chicago, but he was also an undergraduate at Princeton, and took principles from Krugman's colleague Alan Blinder. He has particular principles concerning how economics should be done, but he is not ideological. Anyone who wants to turn him into a bad guy doesn't know him or understand him.

  15. I don't think the joke was meant to be offensive. The main point of Krugman's column is more worrisome, though. Reading Krugman lately, it seems that he has gone crazy. I am worried about him. Maybe he has a tumor in his head. I don't know, but I'm something strange is happening to him.

  16. We can't know intent. However, for some reason Krugman wants to think of anyone associated with Chicago or Minnesota as another kind of being, and those others are callous types who we really don't have to be concerned about. The troubling part is that Krugman knows that the issues he is dealing with are complicated and contentious, but he writes as if everything is abundantly clear. This doesn't teach the lay person anything about economics and how we do it.

  17. Great post, Steve. Krugman also says WW2 shows boosting AD solves everything and nothing is structural - but in WW2 24% of the workforce was shipped overseas, and 2% of it was killed or wounded. Wouldn't that fix structural as well as cyclical unemployment?

  18. "I think the question you want to ask is: How much of unemployment can we explain as being due to: (i) sector-specific factors; (ii) aggregate factors, including aggregate productivity, government intervention through monetary and fiscal policy, and financial factors; (iii) sticky prices and wages."
    What's your answer to the question I want to ask?

  19. @David Barker, Actually Krugman was careful to show that the economic benefits of the war build up were apparent BEFORE major deployments. and
    Right now the government could employ at least 5 to 10% of the workforce with out significantly "crowding out" the private sector. Why not?

  20. Stephen,

    Though I'm not a religious reader of the blog I've read several posts. All tend to be well reasoned but they all also seem to focus on what won't work, or why Krugman is wrong.

    I'm struggling to see any policy prescription in what you write. Do you have any prescriptive recommendations for policy makers? (Fiscal or monetary).

    I mean suppose we agree that a lot of the unemployment is "structural", does that really mean nothing can or should be done? It's perfectly fair for you to say you think money is neutral or there really is nothing for policy makers to do (sort of Prescott style) but you never quite say that either.

    Supposing I agree with you on why all these ideas of Krugman are wrong, which ideas do you think are right?

  21. To David Barker (as well as Stephen),

    You get at a crucial thing here, that I've wanted to write about when I've had more time to gather some things:

    "Krugman also says WW2 shows boosting AD solves everything and nothing is structural"

    First, Krugman doesn't say none of the unemployment is for structural reasons. See amongst other posts:


    I've never seen Stephen acknowledge that the current high unemployment could be any portion due to low aggregate demand cyclical reasons. Stephen, you seem to imply that it's 100% for structural reasons, even though I've never seen anywhere near the job offers from industries that can't find workers with the skills to come close to equaling the number of unemployed today (minus about the amount you'd expect due to normal job applying time, etc.)

    So, Stephen, could you please just straight out say -- even very approximately -- what percentage of the high unemployment you think is is due to structural reasons (100%?) and want percentage you think is due to insufficient aggregate demand, cyclical reasons (0%?)?

  22. On Krugman and WWII:

    I wrote about this here:

    Krugman in some point in the piece I was commenting on said this about WWII:

    "Deficit spending created an economic boom — and the boom laid the foundation for long-run prosperity."

    He's saying that WWII was all about deficit spending, but I see it, as David Barker does, as a huge sectoral reallocation story. Resources (some of which were unemployed of course) are moved out of the private sector and into the government sector, then back again, human and physical capital accumulation is temporarily held up, etc. It's a huge disruption, not some story about how deficit spending makes things better.

    "I'm struggling to see any policy prescription in what you write. Do you have any prescriptive recommendations for policy makers? (Fiscal or monetary)."

    You should read more of my posts. I've been quite explicit about both the monetary and fiscal policy. For example, in this piece:

    The policy prescription comes at the end:

    "I don't see a particular fiscal role that comes out of the sectoral reallocation story, other than a long-term re-evaluation of the adequacy of our unemployment insurance system and our system of public education."

    On monetary policy, I have talked about instruments the Fed should or should not use, whether we need more inflation or not, how to get it, etc. I'm critical, but constructive too. Go read more, and enjoy.

    "What's your answer to the question I want to ask?"

    Maybe you want the solution to global warming too? This is a blog for God's sake. A serious answer to that question would take a few months of careful research. We need to construct a serious model that can incorporate all these factors and then try to quantify everything. I wrote this piece:

    That did the best job I could of assembling available evidence on sectoral reallocation and the current state of the world in the morning I allotted to the job. I seems to me that there is something unusual and interesting going on here, and it has something to do with how resources get shifted over time among sectors of the economy. What are my conclusions given what I know about economics today and the empirical evidence I have seen? I think that sticky prices and sticky wages do not help at all in understanding what is going on. I think that sectoral reallocation does. I don't think a real business cycle model helps me understand what is going on - it doesn't help me much in understanding the current recession to think of it as arising from an aggregate productivity shock. There are obviously important financial factors at work, and these may interact with sectoral reallocation in important ways that I do not understand completely. In any event, stories about "deficient demand" arising from sticky wages and prices are completely unconvincing to me. I think that way of thinking is dead as a doornail. I'm open to thinking about this, but no one is giving me any convincing evidence. When I said "I am all ears" above, I meant it. If Keynesians were right about how the economy works, that would be a wonderful thing, but I don't think it works that way. Try to convince me otherwise. I'm really not wedded to anything.

  23. Steve (or anyone else) - how would we identify structural unemployment? My impression is that the data is not sufficient to allow much confidence in a conclusions: breakdown by industry for NFP data (establishment survey of the unemployment report) only seems to be available since 2000, and breakdown by region for JOLTS data is not very granular.

  24. You need a model to do it. In principle, you could take the whole history of unemployment and decompose how much is due to what shocks and what mechanisms at each point in time. One approach is to back out the shocks that replicate the data exactly, then see what happens when you "turn off" each type of shock in turn, to get the counterfactuals. It doesn't seem to me that what you are calling "structural unemployment" (and I do dislike that word) is something you can measure directly from the data.

  25. The current unemployment rate is 9.6%, which is 14.9 million people.

    Normally unemployment is about half that, so we need to explain excess unemployed of about 7.5 million people.

    If that excess unemployment is 100% for structural reasons, then we need to have about 7.5 million job offers currently where the employers can't find workers with the skills necessary to fill those jobs, 7.5 million offers where there's a matching problem, where there aren't workers with the skills to match what's needed in those offers for jobs.

    Can you show me the data for those 7.5 million job offers, or anything close?

    Because if not, then there must be something else big that's causing the very unusually high unemployment.

    And so far, none of the industries mentioned as having matching problems that I've heard come anywhere close to having 7.5 million unfilled job offers.

  26. I'll give you a hint.

    Currently there are about 5 unemployed people for each job opening (BLS data, at:

    Even if every job opening in the country were unfillable for structural reasons, due to a matching problem, that would still only explain 1/5th of the total unemployment and about 2/5ths of the excess unemployment.

    So at the absolute most, inability of workers to match what's need for the job openings is about 2/5's of the problem. The other minimum of 60% of the problem must be something else.

  27. Richard,

    I know you want this to be simple, but I'm afraid it's not. The labor market is constantly churning. Today's unemployed are not the same as yesterday's unemployed; today's vacancies are not the same as yesterday's vacancies. Sometimes a match gets made and an unemployed person and a vacancy disappear; sometimes an unemployed person gets discouraged and drops out of the labor force. Sometimes a firm gets discouraged and stops posting a vacancy. Difficulties a particular sector has in matching vacancies with unemployed may have spillovers to other sectors. What's the "normal" unemployment and vacancies about? Sectoral reallocation is always going on. What's the normal amount of sectoral reallocation? You have to solve the whole general equilibrium problem to figure out what is going on.

  28. I do intend to get back to this, but among other things, out of town guests.

    I will say this for now. Look at the BLS graph. Job offers to unemployed have been at about 1 to 5 for at least a year. How can you construct a story of churning unemployment and job offers that says the abnormal unemployment is for 100% structural reasons that fits this data, and the other evidence from the world?

    You could say that 1 unit of job ads go up, and two months later they say, oops, none of the unemployed qualify, or live near enough. So then they take down the ads and give up. Then a totally new crop of 1 unit of job ads go up, and at the end of two months they give up, and take the ads down, and stop trying again, and so on, so that after 10 months, there have been 5 units of job ads and 5 units of unemployed, so it's 100% for structural mismatch reasons, but at the current time you only see 1 to 5, because 4 of the units of job offers gave up after a few months.

    But the story looks very unlikely given the data and evidence we have. Low labor quality has plummeted as being cited as the single most important problem by businesses in the National Federation of Independent Businesses survey. It's been the exact opposite of what you'd expect if there were a giant mismatch problem; it's been far lower than normal, not far higher. This is consistent with the inadequate aggregate demand explanation, not the mismatch explanation. And the fact that low sales has skyrocketed in the survey is really consistent with the inadequate demand explanation.

    Plus, why would employers be giving up and dropping their offers after only a few months. Ads are cheap, and it's even free to keep an offer open with an employment agency as you don't pay the agency unless you decide to hire someone.

    Another attempt to claim it's 100% for structural reasons is you could say that five million people just very suddenly became unemployable to anyone at any wage higher than the minimum wage (since were ruling out stickiness of wages and prices); again looks very very implausible given the evidence.

    So, as I said, it's just hard to think of a story that's very plausible, given the evidence, that the abnormal unemployment is 100% for structural/mismatch reasons, or close to 100%, so that no other reason is significant in force.

    It looks very strongly that there are other, or additional, substantial factors causing the very abnormally high unemployment.

  29. I think the clearest piece of evidence is Canada. Look at the time series of unemployment rates (or other labor market data), real GDP, and whatever sectoral activity you want, for the US and Canada, but pay particular attention to residential construction and housing starts. See if you can reconcile that with an Econ 101 aggregate-demand-management view of the world.

  30. My first impression is the point you're making is that the US had a housing bubble and Canada didn't (or had a far smaller one), and the US unemployment soared and Canadian unemployment didn't. So as you titled this post, it's the housing stupid.

    But the question is what did the housing do? It could have been a catalyst for a plunge in aggregate demand, or it could have been what caused a large structural shift in what businesses wanted to do and thus what kind of workers they wanted to employ, or some combination of both.

    The Canadian example shows that housing made a difference, and was a factor in the abnormally high US unemployment. But it doesn't necessarily show HOW it made a difference, whether it was a catalyst that led to financial crisis and fear among consumers and businesses, that led them to spend and invest a lot less, causing a cycle of lower aggregate demand leading to layoffs leading to even lower aggregate demand, etc. Or, it caused businesses to reevaluate what they wanted to do and thus who they wanted to hire. Or, a combination of substantial amounts of each.

  31. Richard,

    You're supposed to be entertaining your guests. Soon they will be coming to drag you away from the computer. I thought this was fairly obvious. Typically the unemployment rate as measured in Canada is about 2 percentage points higher than here. The real GDP paths during the recession in Canada and the US were similar - the depth was about the same, though in Canada the recession was a little shorter and the recovery a little stronger. However, the increase in the unemployment rate was much larger in the US - so much larger that the Canadian unemployment rate is now at 8.1% and the U.S. rate is 9.6% If this were an "aggregate demand" recession, then the Canadian unemployment rate should be up at about 11.6%. The big difference that is staring us in the face is that Canada did not have our mortgage market problems, there was not a severe housing bust, and the housing sector has bounced back there.

  32. You're right, this wouldn't be the first time I'd be in the dog house with my wife over blogging!

    I'll have to think about your Canadian example and maybe try to see the data. Still, looks like at least it's a substantial combination of the factors. Canadian employment didn't drop as much with GDP as US employment did, but it still dropped a lot. Maybe the additional drop over Canada was for structural reasons, but the Canadian amount could have been due largely to low aggregate demand.

    In any case, you can always get more employment with government spending (when there's slack). You may say better to wait (years? decades?) and have the free market do it more efficiently, but you know I think the pure free market does it a lot less efficiently, and total utility maximizing, than a combination of government and market. And I think we're way underspending on high return government investment anyway, so now's a great time to rectify that. For more on this see:

  33. It looks like there are two parts to your argument.

    1. In the long run, you want a larger government. There is nothing wrong with that, as long as you can make the cost/benefit arguments. Of course, you have to deal with the political realities. I don't see much enthusiasm for long-run tax increases.

    2. In the short run part of your argument is for government spending now while the opportunity cost is low. Again, there is nothing wrong with that idea. You have to be careful about it, of course, and that is where the sectoral issues come in. I don't think it is socially productive at the moment to be investing in public transit systems in Flint Michigan, or highways in Las Vegas. Improving the sewers in New York City could be useful though.