DeLong and Baker discuss two bubbles and two recessions
Here is an interesting (as always) cyber discussion between Brad DeLong and Dean Baker.
DeLong asserted (as he often does) that the great recession was so severe because of finance and “clogged credit channels”
Baker argued (as he often does) that the decline in house prices alone was sufficient to explain the downturn.
In particular he argues
1) that the housing bubble was larger than Brad calculates because Brad used 2002 as his base year and the bubble was already well inflated by then.
2) that the decline in construction plus the decline in consumption due to reduced housing wealth explains the decline in aggregate demand (without any need to discuss finance, underwater mortgages, or clogged credit channels).
and finally that
3) The 2001 recession which followed the bursting of the dot.com bubble was also severe if one considers employment not GDP.
For what it’s worth, I agree with Baker’s first point (and would go even further). I don’t have a firm view on the second point, but I think I agree with Baker. I don’t agree with Baker’s third claim.
I would (and do) say: yes, you have a problem. You need to rebalance, but competent policymakers can balance the economy up, near full employment, rather than balancing the economy down. And from late 2005 to the end of 2007 the balancing-up process was put in motion and, in fact, 3/4 accomplished.
There is no reason why moving three million workers from pounding nails in Nevada and support occupations to making exports, building infrastructure, and serving as home-health aides and barefoot doctors needs to be associated with a lost decade and, apparently, permanently reduced employment.
I think Dean Baker agrees with Brad Delong that competent (Keynesian) policymakers could have handled the large problems.
I comment in more detail on the discussion after the jump.
Of course your time is better spent reading their posts than reading this comment.
First my comment at Brad’s blog (I didn’t cut and paste an earlier comment and the confession that I was wrong)
Dean Baker isn’t too thrilled by Mian & Sufi household debt based stories, but household debt levels were quite different in 2001 and 2008. If, pace Baker, debt and wealth have asymmetric effects, there is less need to bring high finance (as opposed to consumer finance) into the picture.
I am not so convinced that 2001 was a big deal. Baker looks at the change of employment to prime age population from the peak. The 2000 peak was very high. The level is much lower now than the low point between 2000 and 2008. There is also the fact that some prime ages are primer than others — that’s why he uses prime age not working age, but the 21st century decline in employment to prime age population could be partly explained by more detailed consideration of demography.
Oh I have another problem with Baker. His story is about aggregate demand. It seems to me that if output is demand limited, high productivity growth would cause low employment growth. I think the standard use of GDP compared to the last peak when one is talking about aggregate demand is standard for good reason.
I’d consider other explanations for what was worse this time. The decline in construction employment was larger than the decline in .com and cable laying employment — so the amount of rebalancing (and the amount of stimulus required to keep employment up) was much greater. Also housing wealth has a much larger effect on consumption than the value of stock portfolios because much of stock is owned by the very rich.
OK now in more detail
On Baker’s point 1, I entirely agree that the housing bubble was very large. Brad only considers the increase in prices and construction during the period in which the bubble was obviously a bubble. But, especially with the benefit of hindsight, it is clear that it started sooner and therefore was larger. Baker uses 94-96 as his base period “The problem with this assessment is that 2002 was a year in which we were already well into the bubble. If we use the period 1994-1996 as a base, construction was already 10 percent above its population adjusted trend path by 2002.” I would at least consider going much much further back. Robert Shiller has argued that there has long been a widespread and erroneous belief that the ratio of house prices to the consumer price index has an upward trend. He (and many research assistants) calculate that there was no change in the relative price of houses over the 20th century as a whole. If he is right, then the bubble is as old as the belief that houses are good investments. I think this dates from the 70s when they were (due to inflation and the huge nominal rigidity of 30 year fixed nominal interest rate mortgages). Baker notes data (click on the Baker link then at “data showing that vacancy rates were already at a record high in 2002” and a spreadsheet will download) on the housing vacancy rate. It rose dramatically from 96 through 2002 (and reached an extremely high level in 2008). Empty houses are a sign of excess supply possible due to overly high estimates of relative home price appreciation. The housing vacancy rate increased from below 10% to around 13% during the period which even Baker considers pre-bubble.
I think the low construction post 2006 might be the new normal based on realistic assessments of relative house price appreciation. Now Shiller et al might be wrong — others argue that house price inflation has really been greater than consumer price inflation. Also if he is right, the illusion might return. My point (if any) is that models of bubbles, including models with irrational agents, always include the assumption that people know the true history of asset prices. If they don’t then a bubble can last forever without growing to an extreme level and bursting. The assumption that those who buy assets know the time series of prices makes sense for stocks and bonds, but not for houses.
On Baker’s point 2 I think his calculations make sense. He gets to his conclusion with simple estimates (no finance included) using data from before the great recession. He has a problem with the timing of the recession which was very mild until Lehman collapsed then very severe. I think he can argue that this was a short run fluctuation with effects which didn’t last and that the overall decline would have been about the same without the financial crisis. As I typed, I don’t know if I agree with him. I do think the current gap between GDP and its old trend is largely due to low residential investment (which I ascribe to low expected home price appreciation which demonstrated by survey data) and low government consumption and investment. I think it is necessary to appeal to the liquidity trap, but I’m not convinced that credit channels are still clogged.
On the other hand, on his point 3,I don’t think it makes sense to look at employment when discussing aggregate demand. US aggregate demand is a number of dollars. Baker says the 2001 recession was severe, but high productivity growth caused decent GDP growth. Keynes would say that the GDP of an economy with insufficient aggregate demand is determined by that aggregate demand and higher productivity growth would just cause lower employment growth. There isn’t any particular reason why high productivity growth would cause high nominal aggregate demand. Inflation was slightly higher then than in the great recession. So I think one has to argue (as I do above) that the dot com sector (including hardware especially fiber-optical cables) was smaller than the housing sector and that the wealth effect on consumption of stock prices is smaller than the wealth effect of house prices (because shareholders are on average richer than homowners).
Even if we kept employment high, we would still have had problems from bad loans and falling housing prices, though one lower interest rates would have had some chance of solving with aggressive monetary action. This was already quite bad by the time of Lehman and employment was holding up only due to completion of work in progress.
Employment contraction wasn’t that impressive considering the over employment by 2000. Matter of fact, going by the mean, it was the lightest since 1970.
I would argue there wasn’t a true recession in the early 00’s. Residential investment kept on growing before expanding majorly in 2002 and consumer spending was fairly robust for such a decline in non-residential spending. The surge in private debt was the culprit and it began in the late 90’s. A large chunk of the bubble had already blown up by March 2001 and definitely by March 2002. I think that was a large degree behind the corporate recapping of the 00’s. In the end, they likely overcut spending and gave the illusion of recession.
Housing busts are cause worse slumps than stock busts , and are especially bad if associated with a big run-up in household leverage.
Schularick and Taylor have been all over this over the last few years , but the IMF was aware of the phenomenon at least as far back as 2003 – early enough to send out warnings. Alas….
Baker may not like the Mian and Sufi story , but I’d highly recommend looking at their work and making your own decisions. This video is a good place to start :
Andrew Crockett Memorial Lecture, delivered by Amir Sufi, Professor of Finance, University of Chicago, on the occasion of the Bank’s Annual General Meeting, Basel, 28 June 2015
BTW , these days the “wealth effect” is , for the most part , bullshit. Rich people own all the stock and don’t sweat a decline because they know the Fed always has their backs. For normal people , housing wealth has a big effect on consumption but it’s mainly due to the collateral effect , not because they’re suddenly feeling “wealthy”. They get a chance to inhale , for a change , and they go for it.
I’m not sure what you mean by “these days”. I think that, if anything, more people are affected by stock prices as defined benefit pension plans have been replaced by 401(k)s .
Back during the dot com bubble, there was an anomalously high ratio of consumption to disposable income. I regressed the ratio of consumption to disposable income on (among other things) the ratio of the S&P index to disposable income. I’m not sure if one can trust the t-statistic, but it was 6.6.
The regression and a graph of the dependent variable and fitted values are in this pdf (the graph is figure 7 the regression right above figure 7)
I think the wealth effect of stock prices is economically significant.
Housing wealth effects appear to be present in the data even back before home equity loans were common. I think it is partly a perceived wealth effect. Note high house prices make people who don’t yet own a house but want to buy one poorer. Totally aside from the insanity of house prices, the apparent wealth effect makes no sense. US residents in general have the same number of houses no matter what their price and can’t sell unless someone else buys. There shouldn’t be a wealth effect. However, I think there is one (and again not just the very important effect of home equity loans).
These days , more people may own stocks , but the vast majority own trivial amounts. Look up Edward Wolffs stuff on wealth distribution and try to estimate how much stocks would have to go up to tease out any meaningful consumption out of the bottom 90% of the wealth distribution.
The dotcom bubble was unusual in that from the mid-90s to just after the crash decent income gains were achieved across the distribution , something that hadn’t been seen for a couple of decades. That may have contributed to any apparent consumption anomaly , as more-than-usual income flowed to high MPC households.
You can find people who can econometrically manufacture stock market wealth effects of 5-10 cents on the dollar , but they’re almost always those of a certain “persuasion” , politically speaking , i.e. mouthpieces for the 1%. Greenspan was a famous example.
Other asset prices certainly respond to a stock market bubble , like art and other collectibles , but that doesn’t do much for the economy either.
Shiller has never believed much in the stock market wealth effect , and I tend to think he’s on the right track. A couple pennies on the dollar in the US , maybe :
Housing wealth is more potent , I’m sure , but when you back out the collateral-enhanced borrowing increase , I doubt that it amounts to more than a penny or two attributable to wealth “animal spirits”.
Finally , ‘splain this :
If there’s any kind of generalized wealth effect , it should be going gangbusters right now , bigger even than the dotcom or subprime booms. That’s hard to square with this economy’s performance , which has limped along right through the wealth boom. Maybe some would argue that without the wealth boom we’d be entirely dead , but my feeling is we’ve designed the economy to generate wealth instead of gdp. In that sense , we’re doing great !
While you may be able to shoehorn the 2008 decline into a “decline in housing welath” theoretical scenario, the facts on the ground were and are quite clear: the huge decline of late 2008 was a credit event, not a housing event:
Indeed the ratio of personal consumption expenditures to personal disposable income is the highest its been since 1950 except for 2005 2006 and 2007 (the height of the housing bubble).
Note in the discussion that Brad DeLong, Dean Baker and I all agree that housing wealth has more effect on aggregate demand than stock market wealth. I argued as you do that the wealth of the rich has little effect on their consumption (which is I think limited by 24 hours in a day not a budget).
I have no idea why it is that people who assert there is a stock price effect on consumption tend to be right wing. They often argue that promoting saving is very important (hence capital income shouldn’t be taxed). In general they argue that consumption is too high not too low (and that it crowds out investment). Thus they should argue that causing low consumption is a good thing about low stock prices.
In fact, I think that usually (when the economy is not in a liquidity trap) lower consumption would be better. This is one of many reasons why I would like to effectively confiscate part of the value of stock by taxing dividends. It is exactly the wealth effect that makes the optimal tax on capital income (as correctly calculated using the standard model used by critics of capital income taxation) very high.
In any case, I don’t think one should decide what is true by group affinity for people who say one thing or another. Rather I think it is better to look at data (as I did following your absolutely correct albeit rhetorical gangbusters prediction).
GDP is way below trend because of low residential investment and low government consumption and investment. Consumption is high — much higher than one would guess with the most empirically successful model with no wealth effect (which is the paleo Keynesian consumption function).
I think I am pulling this discussing up to the main blog.
Is it possible that we have cause and effect reversed? Could it be that, as people have more to spend, they bid up the housing prices?