Unemployment and housing bubbles

As I mentioned in comments below, Casey Mulligan posted a working paper this weekend that argues this recession is consistent with shifts in labor supply. He gives evidence that this shift is due to increases in the labor wedge.

He gives two causes of increased labor market distortions. The first is talk by the IRS and politicians to be more lax in enforce against economically distressed tax payers. Of course, this would give an incentive to be viewed as “economically distressed” by the IRS. The second distortion unique to this recession is caused by the way distressed mortgages are handled. Rational banks and soft-hearted policy makers will decrease mortgage payments for “economically distressed” individuals. At some levels of income and some sizes of mortgage payments, then, there’s a 100% income tax.

Anyway, there are a million models that could generate labor market wedges. The name of the game is to find one that is also consistent with other salient facts. Mulligan’s mortgage story seems like a stretch but there is one testable implication that I can think of: The housing bubble affected different parts of the country differently. Homeowners in areas most bubbly will be more likely to be given incentives to reduce their labor supply because of offers (or expectations of offers) to decrease mortgage payments conditioned on income or employment status. Here’s a scatter plot of the change in the case-shiller index (a measure of the size of the housing bubble) by the change in the unemployment rate. These are year to year changes from November 2007 to November 2008.

Somebody much better than I at econometrics could see if this simple relation holds for reals.

(h/t Everyday Economist)

20 thoughts on “Unemployment and housing bubbles”

  1. Could also be that in the bubble areas, the decline in home values has been more severe, and this decline in wealth causes people to spend less. This reduction in spending would then reduce the labor demand for non-tradables, things like restaurants, construction, etc.

    Productivity in this recession doesn’t look thaaat much different than in past recessions. Below 2001, even with the 1990s recession, and above but probably within the MoE compared with the 70s recession…

    Real wages and productivity often rise in recessions when non-essential people are laid off. Say you’re the president of a two person firm. You’re firm makes $100k. Sales go down $10K due to recession, so you lay off your secretary/personal asssistant. Total hours worked at your firm dropped, and although sales are way down, productivity is up!

    Replace labor with capital, redo the same calculations, and you could conclude that machines have decided to take more vacation time…

    Not incorporating heterogeneity in macro models is standard and looks harmless, but show me a macro model whose conclusions are sensitive to the assumption that everyone’s the same…

  2. Casey Mulligan needs to get out of the ivory tower and take a look around. When you see companies announce hundreds of thousands of layoffs per month – that’s labor demand shifting, not voluntary unemployment. When you see depression talk in the media, when workers accept pay freezes or cuts instead of layoffs – that doesn’t sound like people angling for better mortgage terms, that sounds like plain old fear of losing your job. Finally, if Mulligan read Calculated Risk, he’d know that it is an empirical fact that mortgage modification programs have low uptake rates and high redefault rates. In the first 3 qtrs of 2008, OCC reports 320k loan mods, with redefault rates of more than 50% after 6 months – but there were 2.6 million job losses in 2008. That’s what people are going to risk their incomes for?

  3. Thorstein, maybe you should take another look at the paper. Productivity, until recently, has been counter- pro-cyclical. And what, oh what, would cause shifts up the demand curve leaving remaining workers with higher productivity/wages?

    Also, if people felt poorer, they’d work *more* and consumption would rise (counterfactual). If people think future capital is worth less, they’ll consume more today, they wouldn’t save more (counterfactual). There’s something wrong with your first paragraph.

    Also, capital is fixed in the short term, so if you did this analysis in the reverse, you wouldn’t get lazy machines. You’d get decreased demand for machines.

    The “my friend has been looking for a job for months are you going to call him lazy” is a straw man. There’s many dimensions on which people make employment decisions. One, for example, is when an unemployed person chooses their reservation wage.

  4. ssendam, those 2.6 million jobs were famously lost in the last months of the year. You’re occ link is a report of uptakes through September.

    Also, the anticipation of such programs is all that is needed to get an employment affect.

    In any case, its fine to tear apart a theory. Its a lot of fun. Its much harder to come up with stories that fit the data better than the torn up theories. If you don’t think labor market frictions are the problem in this recession, then explain lower consumption/employment but higher productivity. If you think the frictions story works, but you don’t like Mulligan’s story, then come up with an alternative friction (one that is consistent with a positive correlation between housing bubbles and unemployment).

  5. If there is a job (at an above-minimum-wage wage) and I don’ take it then unemployment is not “involuntary”? What if there are absolutely no jobs advertised?

  6. Another dimension people make employment decisions is industry. Unemployed construction workers often look for construction jobs. This makes sense, but he is implicitly making the decision to lengthen his unemployment by not looking at other sectors.

    Today on craigslist there were seven listings made for healthcare workers in the sacramento area. Two of those listings were for multiple nursing jobs. There were also seven listings (for seven jobs) for accountants. I even saw a couple ads for real estate loan officers.

  7. I’d expect some friction (is that the right term?) in career swapping. Your unemployed carpenter will have some difficulty passing an interview for a nursing position. The carpenter can learn, sure, but that’s a big investment of time and money that will shake off a lot of potential career hoppers.

    What about geographical mobility? Would your construction worker prefer to move to a region that’s hiring?

  8. Yeah, these are labor wedges (frictions). The industry shift makes sense to me, but I wonder about geography. For there to be a net friction people have to be where there not needed on average. Why would construction workers in chicago need to move to arizona to become nurses while the construction workers in arizona need to move to chicago? I know old people stereotypically live in certain places but there’s old people everywhere.

    This particular sectoral shift (from construction and finance to education and health care) requires credentialing. I wonder how much that’s playing a role in this recession.

  9. The friction I’m looking at in my dissertation is job-specific human capital. More experienced workers don’t like to switch jobs because experienced people are better at their jobs and higher paid. Going to a new type of job means you lose that experience (and the higher pay).

    I don’t think this is an important friction in this recession. I don’t know if there’s much of an experience premium in finance or construction, but I would doubt it.

  10. I’m just asking whether it’s easier for an unemployed construction worker in Phoenix to move to New Orleans to do drywall instead of going to night school in Phoenix to become a hospice worker. Training is expensive!

    Not that that solves everything. If there really is a fall in demand, you can’t hide it by pushing it around on your plate like unwanted brussel sprouts.

  11. So, if labor wedges are a significant cause of unemployment, then how could we have full employment with sizeable labor wedges from payroll taxes? Do you think that if tomorrow we eliminated all labor wedges we would have full employment?

    “Also, if people felt poorer, they’d work *more* and consumption would rise (counterfactual). If people think future capital is worth less, they’ll consume more today, they wouldn’t save more (counterfactual). There’s something wrong with your first paragraph.”

    There is the wealth effect though, which can reduce consumption. Have you not seen the large increases in the savings rate recently? The massive use of home equity loans seems to imply that if people consumed more when they felt wealthier and when they felt their capital would be more valuable. This is what’s wrong with your paragraph.

    Thorstein’s story is completely consistent with higher productivity and lower employment and consumption. It is well known that the Solow residual is noisy and reflects factor utilization. If factors are used less intensively, it looks like same inputs, fewer outputs, thus the solow residual decreases. The RBC literature blindly uses Solow residuals literally as productivity shocks, though the Solow residual is just a measure of our ignorance.

    Search for: Sectoral Solow Residuals for a good NBER paper on this topic.

    There are millions of workers trying to get jobs that cannot. Their productivity has not changed significantly- but the demand for the products they produce has. I thought the world had realized this from Keynes, but apparently the Chicago school has still not learned the lessons most sane people realized decades ago.

  12. “Do you think that if tomorrow we eliminated all labor wedges we would have full employment?” No.

    “…This is what’s wrong with your paragraph.” Their endogenous response to a wealth shock would be to work more thus increasing consumption. The shock itself would decrease consumption. What would happen in net, who knows?

    Thorstein’s story is consistent with Mulligan’s story. To move up the demand curve, you need to have a shift in supply. The question is what is the nature of the shift in supply.

    “There are millions of workers trying to get jobs that cannot.” This is a labor wedge story. Behavioral evidence for it would be appreciated.

  13. ““Do you think that if tomorrow we eliminated all labor wedges we would have full employment?” No.”

    So then why are so so preoccupied with this story of labor wedges? You admit it will not solve the entire unemployment story. Then we should be looking at the real culprits, i.e. a lack in aggregate demand.

    “Their endogenous response to a wealth shock would be to work more thus increasing consumption.” There are no jobs. I doubt that employers are willing to pay for more labor hours. This is the problem, a lack of demand for labor.

    And we know what is happening to savings- it is increasing. Just look at the BLS numbers for 2008, they are unambiguous: link. So we do know what happens on net.

    My take on Thorstein’s story is a shift in demand. I do not think labor wedges are a significant factor in this recession. This is the classical-type position, I do not take it.

    just a quick question: if a worker does not want work, then is he unemployed or a discouraged worker? He is a discouraged worker. A decrease in willingness to work does not increase unemployment, it simply shrinks the labor force. This is just a definition the BLS uses.

  14. Can you give your interpretation of the chart you link to? Specifically, why does this show there’s been a shift in the demand for labor (which I take to be your position)?

    BTW stereotypically, classical analysis has no labor market frictions so its awkward to label labor wedge stories “classical”. RBC theory, for example, had productivity wedges only, no labor wedges. As far as I’m aware, there’s no politics in the data. A labor wedge is an attribute of the data.

    What is your story for increased productivity and decreased quantity by shifting the demand curve? I genuinely want to hear a story like this if it can explain the data.

  15. The chart clearly shows a rise in savings. That is what I aimed to show. You stated that consumption might rise-

    “Also, if people felt poorer, they’d work *more* and consumption would rise (counterfactual). If people think future capital is worth less, they’ll consume more today, they wouldn’t save more (counterfactual). There’s something wrong with your first paragraph.”

    If people consume less, they save more, simply by definition. Now the problem with the simple C+I+G+NX accounting is that I don’t really think the savings is going into investment, so where is it going? Maybe to take capital losses? Obviously if savings flowed into I we would have no problems, aggregate demand would not fall.

    “BTW stereotypically, classical analysis has no labor market frictions so its awkward to label labor wedge stories “classical”.”

    Ok, new classical then.

    “What is your story for increased productivity and decreased quantity by shifting the demand curve?”

    I don’t buy that story, I think it’s about layoffs of least productive workers and capacity utilization. What parts of the economy have gotten less productive in the past few months? A massive equity bubble in housing popped, seriously wounding the financial sector. People are losing jobs, can’t take out home equity loans, etc. Firms fire the least productive workers. Quantity declines and productivity appears to increase. What is you explanation for an increase in productivity during a recession?

    “A labor wedge is an attribute of the data.” Sure, there are labor wedges in all economies, during full employment and massive recessions like the one we are experiencing. What caused the labor wedges to become so large in this recession? There’s no FDR to blame this time.

    I am still a little fuzzy on how mortgage forgiveness causes unemployment. If you don’t want to work to qualify for this forgiveness, you are a discouraged worker, not unemployed. So how does this increase unemployment? If you don’t want a job, why look for one?

  16. I think you’ve decided the labor wedge has increased and is the most important phenomenon to explain in this recession. You agree, then, its not drops in productivity (demand) and its not wealth or substitution effects. Good. We (you, me and Mulligan) are on the same page reading the data.

    Mulligan has a story for *why* the labor wedge increased so dramatically. Underwater mortgage holders stay unemployed or they drop out of the labor market to depress their earnings and improve their chances at getting bailed out of their mortgage. This is fully rational because at some payment levels and some income levels, the implicit tax rate is 100%. The chart I posted in this post give some evidence consistent with his story. More bubbly economies, where there would be more underwater mortgages and so more people with these incentives, tend to have higher increases in unemployment.

    On the other hand, I have a hard time believing this mortgage-based labor incentive story explains much of the increase in the wedge. It may be true, but its not likely to be the most important part of the story. This is why I’m on the look-out for other stories that explain an outsized labor wedge.

    What is your story for the increasing labor wedge in this recession?

  17. I think I have a different understanding of labor wedges than you do, but I admit I am not very familiar with the literature. I would not say its accurate to say I agree with Mulligan at all.

    I don’t think that productivity is what’s driving demand, it’s a massive collapse of bubble-fueled consumption and a lack of credit due to deleveraging from banks, credit rationing, and the financial accelerator. These changes may be isomorphic with productivity, but I don’t see anything to do with productivity. These can definitely be wealth effects, but it seems to be something more.

    RBC models may be able to replicate business cycle characteristics with productivity shocks, but that doesn’t prove any type of causation from producitivity to business cycles. It’s like asking why the chicken crosses the road. The RBC theorist would say: an exogeneous electric shock can explain why the chicken’s muscles contracted and propelled it across the road. But that doesn’t explain why at all.

    I don’t see how there is any unemployment in an RBC model. It seems the negative productivity shock simply reduces MPL, which decreases the equilibrium amount of labor the representative agent chooses. But if you are simply choosing more leisure rather than labor, this simply means you are not in the labor force or working fewer hours. This still has nothing to do with unemployment. Unemployment is wanting work and not getting it. So maybe this labor wedge idea is useful, but it seems very broad.

    Keynes said “we have magneto [alternator] trouble.” I think this is the right way to think about crises like this. If you think this is a labor wedge, then so be it. I think the bubble is the cause of the crisis, but I don’t yet have a good explanation of how. I think credit rationing of some form is a useful avenue, like Stiglitz-Weiss 1981 or Bernanke 1983, but it needs to be better understood. Obviously the housing bubble led to massive overconsumption which is now rapidly contracting. But what determines the recession’s bottom? The Keynesian cross is unfortunately the best model I am seen(hopefully there are other better models I am just not familiar with), but this still doesn’t explain how fast the economy contracts, nor really how the economy can recover absent government intervention. Obviously we don’t see permanent slumps even absent active fiscal and monetary policy, so something must eventually bring the economy back, so what breaks us out of the contractionary spiral? I have not seen much discussion of this.

    My point in all this was that Mulligan’s story is too much of a stretch to be really serious. Plus unwillingness to work doesn’t explain unemployment at all. If anything, unemployment is an *over* willingness to work, if you think about it.

  18. The trouble, as always, isn’t coming up with ideas… its taking those ideas to the data.

    The era of “my idea is plausible therefore its operative” is over in macroeconomics and it has been for at least 30 years. Show me how your story, your set of causal mechanisms, reproduces the movements in macro data and is consistent with micro evidence. Otherwise, I’m not interested.

    Mulligan’s story is plausible. You can’t dismiss its quantitative effects by thinking about it while scratching your beard nor will an “aha” moment in the shower help you discover it has a small effect. It takes data to dismiss a model and data alone.

  19. “The trouble, as always, isn’t coming up with ideas… its taking those ideas to the data.” Yes I am fully aware. You asked me for my story, did you not? I admit I do not have a rigorous model nor data to support it. You are free to be uninterested, but please realize that a story is different than data.

    I still do not think Mulligan’s story is plausible. I would appreciate an explanation of how unwillingness to work explains unemployment, and not discouraged workers. This is why I do not think it is plausible.

    If you would like some data to explain why I think the downturn is related to housing and consumer demand:


    “Basically about a third of the [California’s] losses (over the past year) have been in consumer-oriented industries,” Callori said. “Another third have been in housing and housing-related industries like construction and financial services.”

    They somehow failed to mention and job losses from reduced desire to work. I wonder why?

Comments are closed.