Why didn’t the housing crash cause high unemployment?

A recent FT article by Vernon Smith and Steven Gjerstad discusses two big housing crashes, one occurred between 1928 and 1929, and the other occurred between 2006:1 and 2008:2.  Both were associated with substantial increases in RGDP.  Why didn’t real GDP decline during these two big shocks to a major industry?  Why was unemployment so low?

There are two reasons.  First, jobs in housing construction are not a particularly large share of total employment.  The direct loss of jobs was in the 100,000s, not millions.  In addition, many of the workers who lost jobs in construction gained jobs in other sectors.  Thus RGDP continued to grow.

The article by Smith and Gjerstad contains two of the best graphs I have ever seen for illustrating the amazing resilience of the US economy.  In 1929 (a boom year) housing output fell by roughly 30%.  The recent downturn is even more striking.  Notice the severe decline in housing for 9 straight quarters during a period where RGDP is trending upward. The ability of RGDP to grow while a major sector is contracting is quite amazing.  Yet for some odd reason they drew almost the exact opposite conclusion that I did; they argued that the housing crashes of 1928-29 and 2006-08 caused severe recessions.  Why is that?

In my view their key mistake was to misinterpret the role of monetary policy.  In each case, housing continued to decline further after the period I cited.  And in each case NGDP, which had been growing, suddenly began declining as well.  It was the decline in NGDP, not the additional fall in housing, which caused the severe recession and the job losses all across the economy.  If NGDP had kept growing at 3% to 5% after 1929, and after 2008:2, the housing downturn probably would have ended, and the economy would have avoided a severe recession.

You might ask; “Isn’t it a bit implausible that two severe recessions would be preceded by housing collapses, if those collapses had no causal role in the recessions?”  In fact, the housing collapses and the subsequent recessions probably were related, but in a very indirect fashion.  Here’s what probably happened in both cases.  As housing declined, the equilibrium “natural rate of interest” began to decline as well.  There was less demand for credit.  At some point the natural rate fell far below the Fed’s policy rate, causing monetary policy to tighten accidentally.  Because Fed officials (and many private economists) wrongly think that the level of interest rates are a good indicator of the stance of monetary policy, they failed to notice that monetary policy had tightened sharply.  But the markets noticed, and there were big stock market crashes in October 1929 and October 2008.

But that can’t be the whole problem, because Smith and Gjerstad do discuss the fall in velocity, and correctly attribute it to the decline in nominal interest rates.  And of course this fall in velocity is just another way of thinking about the fall in the natural rate of interest that I discussed earlier.  So they understand that a housing collapse can reduce interest rates, velocity, and hence NGDP.  So again, why do they reach such different conclusions?

I think in the end it has to do with their approach to monetary policy.  They view the fall in velocity as something that the Fed would have had a hard time counteracting.  And they would undoubtedly point to the fact that the Fed did in fact fail to counteract it.  In contrast, I am much more optimistic about the ability of the central bank to maintain stable expected NGDP growth in a period of financial turmoil, and believe that very little of the fall in velocity that Smith and Gjerstad cite was actually caused by the housing slump.  Instead, almost all of it was caused by two monetary policy mistakes.  One was the failure to do NGDP targeting, level targeting, which would have maintained positive longer term NGDP expectations.  And the other error was paying interest on excess bank reserves.

In his Big Think interview, Vernon Smith suggested that banks took excessive risks in the 1920s.  This seems plausible, after all, lots of banks failed in the 1930s.  But in fact banks were much more conservatively managed in the 1920s than today.  If you take a close look at this graph from The Economist, you will see that bank equity was above 10% of assets throughout the 1920s, and was close to 15% on the eve of the Great Depression.  So that wasn’t the problem.

Then what went wrong?  Why did so many banks fail in the 1930s?  The answer is simple, NGDP fell in half.  The funds people and firms use to repay loans comes from income.  If nominal income falls in half, there will be many defaults, regardless of how sound the loans seemed before the Depression began.

What about today?  It is more complicated.  This time Smith is partly right.  There were many foolish loans made during the past decade, and we know this because the sub-prime crisis occurred while the economy was still booming in 2007.  About all you can say in defense of the banks is that the fall in NGDP after mid-2008 made the losses several times worse than otherwise.  But even some of that is the banks fault, as they need to anticipate the possibility of at least a mild recession, although perhaps one can excuse them for not expecting NGDP to fall at the fastest rate since 1938.

In any case, it’s not about blame, it is about figuring out where we go from here.  And despite the fact that I diagnose the problem slightly differently from Smith and Gjerstad, I reach almost identical policy conclusions from those discussed by Smith in his recent Big Think interview:

1.  We should require much more collateral on loans and derivatives

2.  We should raise the price level 6% (although I would substitute NGDP for the price level.)

PS.   I couldn’t copy the FT graphs, but this one from The Economist shows just how much of the housing downturn had occurred before the recession even began.  Starts had fallen from over two million to roughly one million in late 2007.  In the early part of the recession, starts actually leveled off at close to one million, but then fell to 500,000 when NGDP declined sharply.

PPS.  Tyler Cowen links to a graph showing the effect of “recalculation.”  Oddly, a few weeks back I linked to a similar graph arguing that it showed recalculation was a minor factor in the current recession.  BTW, the graph he links to has an error; unemployment rose between July 2009 and November 2009, whereas it shows a decline.  But I shouldn’t throw stones as a commenter named Tom found errors in my graph as well.

I was born in Michigan, grew up in Madison, and went to grad school in Chicago.  These three areas encapsulate how I think about the recession.  Chicago is a cross-section of America, with a highly diversified economy.  Its unemployment rate is 10.3%, close to the national average.  Madison is blessed with unusually acyclical industries, and I don’t recall it ever experiencing high unemployment.  Because its economy is dominated by state government, college education, insurance, biotech, and dairy, it has only 5.5% unemployment.  At the other extreme is Detroit, with 15.4% unemployment.  Detroit has two problems.  First, heavy industry is unusually cyclical, and thus steel, autos, machinery, etc, will suffer more job losses when AD falls, even if there is no recalculation.  Of course the auto industry is the main problem in Detroit.  It is not true that the US auto industry is in a long term state of decline, but the Big 3/UAW auto industry is in a long term state of decline.  So Detroit’s unusually high unemployment rate is due to both cyclical factors and structural (recalculation) factors.

HT:  Mike Belongia


Tags:

 
 
 

23 Responses to “Why didn’t the housing crash cause high unemployment?”

  1. Gravatar of Bill Stepp Bill Stepp
    27. January 2010 at 16:30

    Scott,

    Lots of good food for thought here, as usual. Wasn’t a key reason for the massive number of bank failures in the 1930s (~8000) due to the fact that banks were too thinly capitalized with dodgy portfolios thanks to unit banking rules and other state-imposed restrictions? The inability to branch out led to too many small banks, and made it difficult and perhaps even impossible to achieve sufficient portfolio diversification to withstand economic gales of a certain force. You are right to point to the severe decline in NGDP as the precipitating factor, but in a world of free(er) banking, surely fewer banks would have failed, both in absolute terms and the percentage of failures.

    Your point about the bearish consequences of the natural rate falling below the Fed’s target rate is also a good one.
    I do think your macro focus might be supported by zeroing in on the effects of interest rate changes on investment values and the valuation process used by capitalists. When rates decline below a certain level (in relation to the natural rate) long-dated cash flows lead to security (and private equity) overvaluation, which are corrected when rates rise relative to the natural rate. This process could fill in some micro aspects of the NGDP focus you have. I would suggest reading Alfred Rappaport’s _Creating Shareholder Value_, 2nd ed. for more on the topic.

  2. Gravatar of Marcus Nunes Marcus Nunes
    27. January 2010 at 17:34

    Scott
    Between 06.2 and 08.2, NGDP grew at 5.1% on average (YoY) and the lowest it got was 3.6%. Likewise, RGDP growth over the period averaged 2.2%. The fall in Res Inv/GDP after the housing crisis was “compensated” by the rise in Non Res Inv/GDP and Exp./GDP. Unemployment remained relatively stable and there was a net gain in employment . With the steep fall in NGDP growth (to negative territory) after mid 08 nothing “compensates” anything and “all hell brakes loose”.

  3. Gravatar of Marcus Nunes Marcus Nunes
    27. January 2010 at 17:58

    From the FT aricle:
    “The lesson is not just that the Fed failed to anticipate the collapse in the bubble: it also didn’t foresee its devastating consequences. This is reflected in the candid comment last year by Fed Vice Chairman Donald Kohn (Cato Journal): “I and other observers underestimated the potential for house prices to decline substantially, the degree to which such a decline would create difficulties for homeowners, and, most important, the vulnerability of the broader financial system to these events.”
    And then you have from Alan Blinder:
    “… while between mid 2007 and mid 2008 Fed actions fell short, following the Lehman blow-up “the Fed deserves extremely high marks for its work since then. It has hit the bull’s-eye regularly under very trying circumstances”.

    When it is said that the Fed regularly got it right at the same time that AD “melts down” something must be very wrong with people’s perceptions.

  4. Gravatar of David Beckworth David Beckworth
    27. January 2010 at 18:15

    Scott:

    Bill Stepp makes an important point. The unit banking laws meant there was little-to-no branch banking back then. As a consequence, (1) bank assets were not well diversified since they would be concentrated geographically and (2) they could not call upon branches of their bank for reserves in the event of a bank run. This made banks particularly susceptible to any negative shock.

    The importance of this point can be seen when comparing the Canadian banking system with the US banking system during the Great Depression. Both countries had steep declines in NGDP during this time. Canada, however, had zero bank closings while the United States had around 9000. This is not to say the collapse in NGDP was inconsequential to the banking crisis, but it does suggest structural flaws in the banking system may have been more important in explaining the severity of the banking crisis.

    See http://mjperry.blogspot.com/2008/09/great-depression-not-single-canadian.html for more.

  5. Gravatar of scott sumner scott sumner
    27. January 2010 at 20:19

    Bill you said;

    “Lots of good food for thought here, as usual. Wasn’t a key reason for the massive number of bank failures in the 1930s (~8000) due to the fact that banks were too thinly capitalized with dodgy portfolios thanks to unit banking rules and other state-imposed restrictions?”

    Yes, but this is a separate issue. Even during the 1920s, lots of small rural banks failed, because they weren’t well diversified. It did not cause any systemic problems for the US economy.

    Even in the Great Depression, the banking system did OK until the Depression got quite bad in 1931 and 1932. So even with the very real problems you cite, our banking system would not have produced the sort of crisis we just saw, unless NGDP had fallen dramtically, far more then it fell this time around.

    Marcus, Yes, that data is very revealing. Thanks. And the Blinder quotation is bizarre. The Fed made a wise move not cutting the ff target from 2% after Lehman failed. With liberal Keynesians like that, who needs conservatives?

    David, That is a good point about the Canadian system. That was a point Friedman and Schwartz correctly emphasized.

    When I said our banks were conservatively run, I should have added they were very poorly regulated. The branch banking laws were a disaster.

  6. Gravatar of cucaracha cucaracha
    27. January 2010 at 20:38

    Unfortunately neither you nor Krugman are being listened as one can learn from the last FOMC meating.

    They are all sitting with the ghost of Herbert Hoover and Montagu Norman…

  7. Gravatar of cucaracha cucaracha
    27. January 2010 at 20:40

    “Meeting”

    (but “meating” could eventually suit better)

  8. Gravatar of Joe Calhoun Joe Calhoun
    27. January 2010 at 21:22

    Scott,

    Another point that is overlooked is that the real estate crash was basically in just four states until August 2008. CA, NV, AZ and FL were ground zero and the rest of the country didn’t really participate much in the boom or the bust (at least until the rest of the economy came tumbling down). I argued at the time that if the crash was contained to those states it shouldn’t cause anything more than a mild recession. I looked pretty foolish for having said that by the time October rolled around so this at least makes me feel a little better.

  9. Gravatar of peter peter
    27. January 2010 at 21:23

    Lay offs are not the only result of a housing bust. Are you ignoring this bit of the Smith and Gjerstad article?

    “In both crises as house prices and homeowners’ equity fell, first consumption and then production and income declined, net borrowing turned negative, and monetary policy began “pushing on a string”. Banks reduced new lending to build reserves against losses.”

  10. Gravatar of Greg Ransom Greg Ransom
    27. January 2010 at 21:53

    There is strong evidence that Hoover’s tariff really took down the banks — with the tariff hitting hardest just where you find the first banks going down.

    See this EconTalk discussion:

    http://www.econtalk.org/archives/2010/01/rustici_on_smoo.html

    The Hayek theory is simply — during the artificial boom leverage increases, people take on bigger debt loads, there is a pyramid effect of expanding credit, expanding leverage, expanding money (relative to productivity cost declines and output gains), and expanding profits, income, consumption, etc.

    The economy becomes mal-coordinated across time (you need to study some heterogeneous production / interest theory to really get this — something almost no economist ever learns). At some point over consumption collides with malinvestment collides with evaporated profits collides with rising unemployment collides with de-leveraging collides with domino-effect insolvency.

    The collapse of housing in 1929-1931 was part of a larger picture — NO ONE argues that there is a “single” cause to the Great Depression — the explanation of this episode is causally “overdetermined” and the size of the thing is due to multiple explanatory elements.

    The collapse of housing was a bell weather of a faltering artificial boom — and the solvency, de-leveraging and credit consequences of the housing collapse are as central to the Hayek causal picture as unemployment is to the Keynesian picture. And note well, the fetish like focus on employment and unemployment is as much an artifact of the Keynesian toy model as it is anything in reality — macroeconomists think of little besides labor and money because, well, there isn’t much else in their toy construct — certainly the macroeconomist’s aggregate “K” doesn’t exist in the real world, and has nothing to say about coordinated heterogeneous production processes across time.

  11. Gravatar of sraffa sraffa
    27. January 2010 at 23:32

    I’d love to hear your thoughts on this, as an economic historian and monetarist:

    Canada had no central bank until 1935, experienced similar output and price declines as the USA, yet suffered no (zero) bank failures. Are the benefits of branch banking more than the benefits of a central bank? Does something else explain this? I was shocked when I learned this about Canada.

  12. Gravatar of Bill Stepp Bill Stepp
    28. January 2010 at 03:50

    Scott,

    Contrary to legend bank failures in the GD weren’t confined to rural banks. There were many urban bank failures, including here in the People’s Republic of New York. And just because the big cluster of bank failures didn’t happen for a year or so into it doesn’t mean that they lacked problems. It would be like saying that an obese person who shows no visible signs of heart disease until the day of his heart attack doesn’t have a health problem. He clearly does and probably long before.
    I didn’t mention in my other post, but will now, that a decline in NGDP (or any other macro measure) must be given a micro explanation. This of course is true for the boom that precedes it as well. Hayek (among others) recognized this and tried to provide a micro-oriented theory for the boom and bust. Science proceeds by breaking macro stuff into micro elements (a working knowledge of molecules preceded the discovery of the atom, which preceded the discovery of the electron, etc.). Tell it to the Keynesians, who are among the anti-scientists of the modern era.

  13. Gravatar of OGT OGT
    28. January 2010 at 06:37

    The Smith/Gjerstad article is indeed very good, particularly the graphs, thanks for linking to it. I am torn as to who’s interepretation to believe as it all seems to come down to one’s faith in the power of monetary policy, however.

    One thing I’d note is that your explanations seems to rely almost entirely on loan supply factors and the stance of the banks. My interpretation has always been focused as much on the demand side, that of credit worthy borrowers. As they write”

    In both crises as house prices and homeowners’ equity fell, first consumption and then production and income declined, net borrowing turned negative, and monetary policy began “pushing on a string”. Banks reduced new lending to build reserves against losses. Although the monetary base has exploded in the current recession, its turnover rate – what economists call the velocity of money – shrank drastically. Keynesians call this the “liquidity trap,” failing to convey understanding of its origin in accumulated balance sheet deterioration.

    The fact that monetary policy is transimitted to the economy through the credit channel makes this a hard issue to avoid. I realize that the expectations of inflation and NGDP growth can partially offset this, but once near term default fears have been heightened I have doubts about the true efficacy of that.

  14. Gravatar of Mark A. Sadowski Mark A. Sadowski
    28. January 2010 at 06:39

    Scott wrote:
    “You might ask; “Isn’t it a bit implausible that two severe recessions would be preceded by housing collapses, if those collapses had no causal role in the recessions?” In fact, the housing collapses and the subsequent recessions probably were related, but in a very indirect fashion. Here’s what probably happened in both cases. As housing declined, the equilibrium “natural rate of interest” began to decline as well. There was less demand for credit. At some point the natural rate fell far below the Fed’s policy rate, causing monetary policy to tighten accidentally.”

    I just want to make sure we’re on the same page here. In my view the demand for credit fell (and the equilibrium natural rate of interest dropped) because of the decline in net real estate wealth (the wealth effect). The effect of the decline in residential investment on the natural rate, on the other hand, was trivial at best. Because the Fed was obsessed with current inflation, and too backward looking, it was caught off guard and things quickly spiraled out of control in late 2008.

    (Incidentally I was one of the few people in my department who anticipated a major recession well before it happened. I was arguing my case based largely on the above net real estate wealth scenario.)

    This also raises the following issue. Although I hate the FOMC hawks who keep raising the idea that monetary policy should take into account asset prices, I personally do believe that asset prices matter in the conduct of monetary policy. This is because it’s important to anticipate where the natural rate is going. This matters especially if you’re going to use interest as the policy instrument, but I suspect it may still matter somewhat with expected NGDP level targeting (one can never be foreward looking enough).

  15. Gravatar of Kevin Bob Riste Kevin Bob Riste
    28. January 2010 at 09:02

    First time commenter, and really on a marginally related note. I was wondering what implications there are for the EMH that investors are all looking for different returns on a certain stock price. It seems like traders looking for a 1% return have a very different conception of the future of the stock than those buying at the same price looking for a 10% return. Or is this all somehow “priced in” in the sense that anyone who sees a profit opportunity of sufficient size will buy the stock? I was wondering what a market for shares would look like if you had to make an bet equal to your investment in shares that predicts your return on the stock. This information could then be aggregated to give a sort of translation of what the current stock price “means.” Does this make sense? I imagine someone has thought of this already, but I wonder if they’ve tried it in practice. Or is there a market where this is already happening, or something similar is happening? I haven’t been able to find any information on this, and your knowledge of how to accurately explain and interpret the EMH makes me think you will be able to help.

    P.S. I absolutely love your blog, and am constantly trying to bring fellow economics students to it (and some professors). It has helped me understand and be able to explain to others many important economic concepts that I think are frequently misunderstood or waved off in a Krugman-like way. (Though I do very much like Paul Krugman. I just wish he had a touch of your reserved nature when it comes to dismissing others’ arguments.)

  16. Gravatar of Arash Arash
    28. January 2010 at 09:22

    Hi Scott,

    I am a first time commenter even though I follow your blog for a while. I am a German grad student and became soon convinced by your interpretation of the crisis: it was a fall in NGDP due to unanchored expecations: inflation targeting allows for base-drifts and that is a major shortcoming! I am long in favor of NGDP level path targeting and think that your idea of NGDP futures is simply fantastic: efficient and simple! I share your concerns with the Joan-Robinson-fallacy, that is, associating high (low) nominal interest rates with high (low) real interest rates und thus with restrictive (expansive) monetary policy. I just weigh in because I disagree with your description of the indirect effect of the housing crisis. Even though you hint a RDGP growth, you introduce a falling natural rate. Of course, the sectoral rate of return on housing fell, but the average rate of return did not … otherwise RGDP would not have grown. Velocity did not fall because of a fall in expected returns on investment, but because of a monetary regime that did not target spending levels. Velocity shocks are a direct result of unachored expecatations right because markets are efficient. I think this view is more consistent with your general perspective than your indirect effect.

  17. Gravatar of D. Watson D. Watson
    28. January 2010 at 09:33

    Some linking notes:

    Your search box doesn’t work and hasn’t ever worked. I was trying to find some things to explain to a colleague but it’s impossible to search through it all by hand. (In particular, an explanation for WHY banks are hoarding all the reserves when the interest rate is a piddly 0.25%)

    The FAQ is still attached to your old blog. I clicked on the FAQ link, then on the title of the blog to get back to the main page, but found myself at the old one.

    Reading through the FAQ again was fun in that I noticed you cite 1933, rather than 1937 or 1929. The reason for picking that date is so clear when I saw it, and it got me wondering: Have you recently discussed the proper role of monetary policy in those other two years?

  18. Gravatar of Nick Rowe Nick Rowe
    28. January 2010 at 11:01

    The search box works for me. Click on “search term”, and those words disappear, then type in your search term, then hit Enter, and it should work.

  19. Gravatar of ssumner ssumner
    28. January 2010 at 14:36

    cucuracha, Yes, that is the problem.

    Joe, You and I were right, and those who predicted a severe recession (like Mark below) just got lucky. There was no reason for a severe recession as of mid-2008. That’s why neither the consensus forecast of economists, nor the markets, expected a severe recession in mid-2008.

    Peter, Yes, but unless you have layoffs you don’t have a recession. I was asking why the housing crashes didn’t cause recessions. I concede that other things happen. But those other things also didn’t cause a recession, indeed didn’t even cause economists or markets to forecast a recession.

    Greg, The tariff did hurt the economy, but mostly by making people more pessimistic about monetary cooperation.

    You said;

    “The collapse of housing in 1929-1931 was part of a larger picture — NO ONE argues that there is a “single” cause to the Great Depression — the explanation of this episode is causally “overdetermined” and the size of the thing is due to multiple explanatory elements.”

    I was talking about the crash of 1928-29, which didn’t cause a recession. The crash of 1929-31 was of course a symptom of the GD. In one sense the Great Contraction has a simple explanation, NGDP fell in half. But the factors causing the fall in NGDP were very complex.

    You said;

    “The collapse of housing was a bell weather of a faltering artificial boom”

    The 1920s boom was not artificial, it was the most fundamentally sound economic boom in all of American history. Prices were stable, taxes were low, unions were weak, the government ran budget surpluses–we even had trade surpluses, for those of you who care about that sort of thing. No wonder stocks did well!

    sraffa, Yes, almost everyone thinks branch banking explains the difference. And I’ve always agreed. But one tiny note of uncertainty. Even the big diversified US banks did much worse than the big diversified Canadian banks in this crisis. So other factors might have also played a role.

    As far as no central bank; the US would have been much better off if the Fed had not been created. Not better off today (I think the Fed now does as well as the pre-1913 gold standard), but much better off during 1918-41.

    Bill, Yes, many urban banks failed in the 1930s, but most survived. I wonder if a single bank would survive a 50% fall in NGDP today. Banks were much more conservatively managed back then. There is no indication that the banking system had major problems even 14 months into the Depression, by which time it was already much worse than the low point of this recesssion.

    OGT, I think it is a myth that monetary policy works through the credit channel. In my view it works through expectations of the excess cash balance mechanism, which impacts future expected NGDP, and current asset prices.

    The pushing on a string metaphor they use is not at all helpful, and usually indicates the person who uses it doesn’t really understand monetary policy. Money and credit are two very different things. Zero rates don’t stop monetary stimulus.

    Mark, You said;

    “I just want to make sure we’re on the same page here. In my view the demand for credit fell (and the equilibrium natural rate of interest dropped) because of the decline in net real estate wealth (the wealth effect). The effect of the decline in residential investment on the natural rate, on the other hand, was trivial at best.”

    I disagree. I think the wealth effect is very weak (note the 1987 crash, which had no impact on consumption.) Rather, real interest rates tend to be positively correlated with investment. When housing construction declined, so did real rates.

    Asset prices matter only to the extent that they help us forecast NGDP. Since they do help somewhat, they help monetary policymakers for that reason. They are not important in and of themselves, but only to the extent that they forecast NGDP.

    Thanks Kevin, I am afraid that is over my head (I never took any courses in finance.) I am pretty sure that a finance expert would answer your question by looking at derivative prices, such as puts and calls. Those tell us something about not just the expected future price, but also the expected volatility.

    Arash, That is a good point, and you might be right. But I am not sure you are. It seems to me that the RGDP growth can be unchanged even if the nature of production moves from investment goods like housing to consumer goods and services. If people decide to buy fewer houses and eat out more in restuarants, wouldn’t that tend to lower the real interest rate? Surely the equalibrium real interest rate is positively related to RGDP, but just as surely it is not a perfect correlation–the mix between consumption and investment must also matter.

    Having said all of that, it is possible that the mild slowdown in NGDP growth in late 2007 and early 2008 began to reduce the natural rate, as the economy slowed. So you may be correct even if your theoretical argument is not airtight.

    Thanks for tuning in from Germany. Most of what you say seems right on the mark. Maybe someone else can offer an opinion on your argument.

    D. Watson, I don’t know why this conversion is taking so long. I have something like a 150 comments waiting to be approved at the old site, but I can’t get to my dashboard to approve them. I’d like to just have one blog site.

    If you google “themoneyillusion three octobers” you may find a post I did on 1929, 1937 and 2008.

    If not, let me know and I’ll look for it. It was back around February or March.

    Thanks Nick.

  20. Gravatar of Mark A. Sadowski Mark A. Sadowski
    28. January 2010 at 21:38

    Scott,
    You wrote:
    “I disagree. I think the wealth effect is very weak (note the 1987 crash, which had no impact on consumption.) Rather, real interest rates tend to be positively correlated with investment. When housing construction declined, so did real rates.”

    A number of papers have found the wealth effect of real estate to be substantially greater than corporate equities. For example John D. Benjamin, Peter Chinloy, and G. Donald Jud found that the marginal propensity to consume MPC out of real estate and corporate equity wealth to be 8% and 2% respectively:

    http://www.uncg.edu/bae/people/jud/Real_Estate_Versus_Financial_Wealth.pdf

    From August to December of 1987 the loss of corporate equity wealth was less than $500 billion out of a GDP of $4.7 trillion or about 10% of GDP. Thus the decline in consumption due to the wealth effect of the stock market crash was perhaps 0.2% of GDP.

    More importantly increases in other forms of wealth more than compensated for this decline. In fact net household and non-profit wealth increased from 357% to 362% of GDP between 1987 and 1988. The stock market crash of 1987 was not much more than a blip on the national wealth radar.

    On the other hand between 2006 and 2Q 2008 (before all hell broke loose) net household and non-profit sector wealth decline from 482% to 413% of GDP, or a decrease of 69% of GDP. Of this 39% represented a decline in net real estate wealth, 13% was a decline in corporate equity wealth and 17% was other forms of wealth. Assuming that the MPC from other forms of wealth is the same as corporate equity wealth, then using the estimated the MPCs in the above paper, this decline in wealth would have resulted in a decline in consumption equal to 3.7% of GDP.

    Gross private investment as a percent of GDP declined from 17.4% to 14.9% between 2006 and 2Q 2008, or a decline of 2.5% of GDP. The vast majority of this decline was in the form of residential investment of course.

    Naturally there is a positive correlation between investment and the demand for credit. But there is also a negative correlation between wealth and the supply of savings. And it takes both to determine the equilibrium natural rate of interest.

    Perhaps I was lucky that I saw what was coming long before it occurred, but I don’t think so.

    P.S. I reiterate that declining net real estate wealth didn’t make a great recession inevitable. It was the Fed’s obsession with current inflation and its lack of foresight that made that happen.

  21. Gravatar of OGT OGT
    29. January 2010 at 08:14

    Sumner- I quite agree that money and credit are largely decoupled, every empircal paper I’ve seen points to this. Asset prices, however, are far more closely tied to credit than base money. And expectations and asset prices, are of course, positively correlated, dangerously so in fact.

    The degree of control that the Fed has on expectations is very debatable. Under the nebulous ‘dual mandate’ I think they are necessarily weak, whatever other levers the Fed might pull. The fact that credit is decoupled from base money only makes this worse (and is clear indication of Fed’s lack of control of expectations).

    I am all for a change in Fed policy with more forward looking stance and clearer objectives, but, in my view, politiburo-esque maneuverings of the FOMC are never going to end ‘long and variable lags’ as long as the congressional dual mandate is around.

  22. Gravatar of D. Watson D. Watson
    29. January 2010 at 12:34

    Nick – Thanks, I never tried clicking there.

    Sumner – Thank you.

  23. Gravatar of ssumner ssumner
    29. January 2010 at 14:37

    Mark, You said;

    “P.S. I reiterate that declining net real estate wealth didn’t make a great recession inevitable. It was the Fed’s obsession with current inflation and its lack of foresight that made that happen.’

    I think this is the key. I accept your data that the wealth decline was much bigger than in late 1987. But it din’t have much affect on consumption until NGDP fell. So I still think monetary policy was the key.

    It will be interesting to see what happnes next time wealth falls by 69% of GDP. Last time you predicted a recession but the markets didn’t. Will the markets predict one next time? And will the Fed note what it did wrong in late 2008, and be much more aggressive next time? In other words, will the market be wrong again, but for the opposite reason?

    OGT, You said;

    “Sumner- I quite agree that money and credit are largely decoupled, every empircal paper I’ve seen points to this. Asset prices, however, are far more closely tied to credit than base money. And expectations and asset prices, are of course, positively correlated, dangerously so in fact.”

    Yes, but the base is not a good indicator of monetary policy The base rose a lot in the early 1930s and in late 2008, but money wasn’t actually easy either time.

    That’s a good point about the dual mandate, it’s vagueness causes big problems. I think NGDP is the only single mandate that is politically acceptable. The Dems would never accept a price level-only mandate, and a real mandate is obviously impossible. So what other choice do we have, if we want to make monetary policy work better?

Leave a Reply