PSST Watch

Kerwin Kofi Charles, Erik Hurst, and Matthew J. Notowidigdo write,

while the decline in manufacturing and the consequent reduction in demand for less-educated workers put downward pressure on their employment rates in the pre-recession 2000–2006 period, the increased demand for less-educated workers because of the housing boom was simultaneously pushing their employment rates upwards (Charles, Hurst, and Notowidigdo 2016, 2015). For a few years, the housing boom served to “mask” the labor market effects of manufacturing decline for less-educated workers. When the housing market collapsed in 2007, there was a large, immediate decline in employment among these workers, who faced not only the sudden disappearance of jobs related to the housing boom, but also the fact that manufacturing’s steady decline during the early 2000s left them with many fewer opportunities in that sector than had existed at the start of the decade.

This seems like yet another paper supporting a PSST interpretation of macroeconomic phenomena instead of an AS-AD interpretation.

15 thoughts on “PSST Watch

  1. There was never anywhere where we built too many houses. Destroying the housing market did screw these workers over. But you’re begging the question. There was nothing unsustainable about what they were doing. Where home prices were high it was clearly because there weren’t enough of them. Median rents are higher than they have ever been while housing starts remain at depression levels and homeownership collapses because the mortgage market has been hamstrung. These people should be out building houses, and if LA and New York City won’t let them build where people want to live, at least let them build in Dallas. In 2016, to act as if we still don’t need to increase construction employment – to miss the fact that the pendulum has swung far too far in the other direction – is obtuse to the point of inanity.

    • You are making a case for the construction of more housing units. But we may be better off if that construction is of units owned by landlords (including corporate landlords) and financed with 25 percent or more equity than if the construction is of units owned by households and financed with 5 percent or less equity.

      • A couple ideas as food for thought:

        1) The places where home prices are the highest, because of supply constraints, generally have low ownership rates and high equity levels. Leverage worsened the crisis to the extent that it led to more negative equity, and it probably increases homeownership, but I’m not sure it has much to do with high prices. In most of the country, the subprime boom didn’t coincide with rising prices at the low end, where credit constraints should be more important. Actually, the subprime boom didn’t even happen until after home ownership peaked and most of the price increases had already happened. There just isn’t much evidence, when viewed closely, that high prices were due to leveraged buying.

        2) By all means, let’s get rid of all the owner-occupier tax benefits.

        3) If you do want to proceed with a housing market that requires more equity and has lower ownership, and especially if you think this will lower prices, then you are basically creating an asset market where there are obstacles to ownership and those that do own have higher returns on their investment. It is a move toward more of a haves vs. have-nots context. This is the trade-off. Access with lower returns vs. less access with higher returns. Yield and price are inverse of one another, and there really isn’t a way to get around that.

        My suggestion is that we get rid of all the tax benefits to ownership but that we encourage financing methods that allow broad access to ownership. Make it possible for low income households who would benefit from control of their homestead to have that control, but create a context where they should be ambivalent about ownership by removing the arbitrary advantages. I think you would have less investment in housing, less volatility, and in that context, some leveraged owners here and there wouldn’t be so much of a problem.

        4) Leverage is mostly a timing issue – some households are leveraged when they first buy, and that leverage tends to decline over time. The supply problem was more of a factor here than the mortgage market was, too. The problem was that since low income and young owners were priced out of the coastal California markets, they had to move inland, so instead of having some highly leveraged owners scattered around older neighborhoods in San Francisco and LA, they were all concentrated in places like Riverside, so the bust became concentrated.

    • Googling “new housing starts by year” and looking at images, that doesn’t look like a smooth, linear trend-line. What do you mean?

        • “There was never anywhere where we built too many houses.”

          How do you measure that? When I look at charts of new housing starts, just as a first glance, there is a big jump in the middle of the 2000s. I would call that evidence of a period of building too many houses, relative to something, in this case the prior trend.

          • There was a jump in single family homes built for sale. All other categories were down, some of them significantly – multi-unit, built by owner, manufactured homes. When you take them all into account, even in 2005, total starts wasn’t that far above trend.
            Here’s a post on it:
            http://idiosyncraticwhisk.blogspot.com/2016/03/housing-part-122-housing-starts.html
            If you compare housing units to population over 15, the ratio has been declining since the early 1990s and only briefly recovered in 2003-2005.
            This is the core distinction that turns an interpretation of the period on its head. It looks like there were two things happening at once: more starts and higher prices, and since mortgages were growing, too, it seemed like this had to be the source of the trend. But, when you disaggregate the country, really no place looks like this. There were places with ample supply and moderate prices and places with very limited supply where prices (and rents!) were shooting skyward. Rent inflation has been above general inflation since the mid 1990s and it was all coming from the supply constrained cities until we broke the mortgage market in 2007.
            In places like Dallas, homes sell for something close to cost. There was no reason for prices there to collapse. In places like coastal California prices are unmoored to cost because limited supply is the problem.
            In the bubble cities, there is this idea that we suddenly were building millions of extra homes, the problem was actually that the supply constraints in the problem cities became so severe that migration to the bubble cities spiked. Home prices then spiked too, because those cities didn’t respond with a matching spike in housing starts. When new home inventory began to build in 2006 & 2007 its because population growth in the bubble cities suddenly downshifted as the crisis hit to well below any previous trend.
            Here’s a graph of housing permit rates in Phoenix:
            https://research.stlouisfed.org/fred2/graph/?g=4zjj
            The crazy part is after 2006. There is no way that the rate of new building in 2004 and 2005 was suddenly creating a bunch of unnecessary houses. Especially since at the time there was a surge of migration out of the high cost California cities that was boosting the local need for houses. The speculators moved in because there was a shortage.

    • I looked at the nice Housing Start graphs (from comment link) but most important are those in the paper on employment.

      They use Bureau of Labor Statistics, but I don’t trust these numbers to accurately include illegal workers. Where are good estimates of how many illegals were employed in construction from the ’95-2005 boom times?

      Part of the reason the Recession is so bad is that the illegals have suffered in Depression type levels of unemployment, but are not fully counted.

      Claims that there was no house construction bust (after some kind of bubble) are almost silly. There was a housing boom / bust, relative to an imaginary “stable” level of construction that would decline from a 1995 total of 1600k to 1200k in 2015 due to other social changes, perhaps especially new net total family formations (no good reference)

      The article title is apt: The Masking of the Decline in Manufacturing
      Employment by the Housing Bubble

      One big social change is the decline in manufacturing employment. Not discussed are the increased sales of larger homes of baby boomers whose children have left and their desire to cash out / trade down for smaller homes; or, after one spouse dies, to sell and move or to open to children who thus do not have to buy or rent outside.

      The final paragraph of the paper (not all read) notes the policy implications:
      Our results shed light on the question of how much of the recent decline in employment rates are the result of cyclical factors, and how much from structural factors like the long-term decline in manufacturing and associated losses in routine jobs. The answer to this question is crucially important because of its implications for policy response to the falling employment. Traditional monetary and fiscal policy tools, such as temporary interest rate cuts, tax rebates or ncreases in government spending, are designed to provide a temporary boost to labor demand. These tools can thus offset temporary declines in hiring arising from cyclical factors like short- lived tightening of credit markets or transitory increases in uncertainty, temporarily boosting employment until the economy returns to its normal level. By contrast, there is little reason to suppose that these traditional monetary and fiscal tools can satisfactorily address employment decline arising from structural factors. <<

      Traditional monetary & fiscal policies that have worked since WW II won't work so well anymore.

      What's not noted is the huge increase in regulations and restrictions, which make any adaptation more difficult.

  2. While I agree lowering lending standards was/is not sustainable, though under the lower interest rates that this policy attempted to avoid much of this was sustainable as evidenced by current housing prices, that is not to say those workers would not have been employed on something more sustainable. It may have been difficult for them to earn as much in other areas and we may have experienced slower growth as a result, but not either wholly.

  3. So I’m not sure there’s a contradiction between AS-AD and PSST. AS, for example, depends in part on the skills-weighted labor supply. PSST simply says that the weighting is changing–different skills are required.

    So this article suggests that AS is declining because the skills of some of the population are no longer valuable. PSST claims that their skills may someday become valuable again as entrepreneurs optimize resources. At that point AS will rise again.

    Or, in more traditional terminology, we’re experiencing a supply shock.

    What’s wrong with this analysis?

    • “PSST claims that their skills may someday become valuable again as entrepreneurs optimize resources”

      Just the opposite, the old skills are, in PSST telling, permanently obsolete.

      • That implies that entrepreneurs will simply ignore a big slice of the workforce. PSST says, as I understand it, by trial & error entrepreneurs will find a way to use productive capacity.

        So that suggests that truck drivers displaced by robots will *eventually* be put back to work doing something else. Of course “eventually” may be a very long time–and indeed it might be the truckers’ children who are reemployed.

  4. Looking at Raj Chetty’s data comparing working class incomes in 1996-2000 to their kids’ incomes in 2011-12 by county across the country, one thing that leapt out at me was how much the bursting of the housing bubble hammered forested states, like the Carolinas, while leaving treeless states, like the Dakotas, largely untouched.

    • Does the Establishment Survey separate out logging from the rest of mining and logging?

Comments are closed.