How conventional wisdom gets formed

Matt Yglesias recently argued:

[F]rom the latter part of 2005 through all of 2006 and 2007 and through the beginning part of 2008, the Federal Reserve is basically doing its job correctly. Inflation expectations are remaining stable. Construction workers are losing their jobs, but there’s no broader economic collapse. Many though not all of the construction workers are managing to find employment in other sectors. It’s only in the second half of 2008 that the monetary situation goes haywire, inflation expectations plunge, and the whole economy goes to crap. It then takes a looooong time for expectations to get back to the usual level and when they do so instead of delivering us a spurt of “catch-up” inflation the Fed lets them drop again. More recently, we’ve been back to a more-or-less stable level but with no catch-up. In my view, it’s all that stuff that constitutes the screw-up that the Fed should be lambasted for. The damning transcripts are the ones from late 2008 and all of 2009.

Noah Millman responded as follows:

I think this view implies one or more of the following:

– Either the Federal Reserve bears no substantial responsibility for the housing bubble, and therefore the Fed cannot be criticized for allowing the bubble to reach such catastrophic dimensions;

– Or, alternatively, the housing bubble bears no substantial responsibility for the financial crisis, and therefore it doesn’t matter whether the Fed allowed the housing bubble to reach such catastrophic dimensions;

– Or, alternatively, the financial crisis bears no substantial responsibility for the subsequent deep and enduring recession, in which case it doesn’t matter whether the Fed allowed the financial crisis to happen.

Obviously I agree with Yglesias.  But it’s interesting to consider the fact that Millman’s views are much closer to those of the mainstream, even though he writes for The American Conservative, whereas Yglesias is at  How did Matt and I end up on the fringes?

One answer is that Yglesias’s claim actually doesn’t imply that any of the three assertions are false.  I happen to believe the first and third are false, but even if true they are not inconsistent with Matt’s claim.  But first let’s figure out what happened, and why it was misinterpreted.

Go back to April 2008 when unemployment was 4.9%.  The housing market had been declining for more than two years, and the US was in the midst of a major financial crisis.  At the time, the IMF estimated the total losses to the US banking system would be a mind-boggling $945 billion.  Bear Stearns had recently been bailed out.

Today most economists believe the bursting of the housing bubble and the subsequent financial crisis caused the Great Recession.  But back then, they weren’t prediction a big recession.  I don’t recall the exact numbers, but I think they were forecasting just a slight increase in the unemployment rate, perhaps to 5% or 6%.  In fact, unemployment soared to over 10% by October 2009.  Why did this happen, and why did it catch our economists off guard?  After all, they now believe the bursting of the housing bubble and the financial crisis “obviously” caused the big recession, and we already knew about those problems in April 2008.

I’m pretty sure I know what happened, and if so the conventional wisdom was based on erroneous data.  We now know that between June and December 2008 both NGDP and RGDP plunged very sharply (according to monthly estimates from Macroeconomics Advisers.)  Almost the entire contraction occurred over those 6 months.  But this was not understood at the time.

In mid-September 2008, before economists understood that we were half way through a severe contraction, Lehman Brothers failed.  Over the next few weeks the financial crisis got much worse.  At the same time data began to appear suggesting that the recession (which had previously been very mild) was getting much worse.  We now know that the recession actually got much worse after June.  But it was only years later, after numerous GDP revisions, that this became apparent.  At the time, it looked like the financial crisis was driving the economy into a deep slump.  In fact, the deep slump was worsening the financial crisis.  Economists connected the wrong dots.

[Similarly we only found out about the 2001 recession in September 2001 even though it began in March.  And that recession ended in November 2001.  During September, guess which event was widely (and wrongly) seen as greatly worsening the recession.]

Eventually the financial crisis went away, and at that point the economy should have begun growing quickly.  But it didn’t.  So now the conventional wisdom searched out other culprits.  Such as a weak housing sector (even though that doesn’t explain the big decline in sectors such as services and consumer non-durables.)  Or various “structural problems.”  In late 2008 almost no one blamed monetary policy, because interest rates seemed very low.  They forgot Milton Friedman’s admonition that very low rates usually mean money has been tight.

So there were two big mistakes.  They got the timing between the severe contraction and the severe financial crisis exactly backward, and they misjudged the stance of monetary policy.  That made it seem “obvious” that the financial crisis was the proximate cause of the recession, and the housing bubble seemed to be the cause of the financial crisis.

In fact, the housing bubble probably only caused about 1/3 of the financial crisis.  Between April 2008 and April 2009 the estimated losses (IMF) to the US banking system soared from $945 billion to $2.712 trillion, as the troubles spread to non-subprime mortgages, as well as commercial and industrial loans.  That was caused by the huge drop in NGDP growth, to 9% below trend.  To most economists it looked like the subprime crisis got worse, although what really happened is that falling NGDP dramatically increased the losses to the banking sector.

Let’s say I’m wrong, and that the Fed’s to blame for the housing bubble.  And let’s say Millman’s right that it caused a “substantial” share of the financial crisis.  Indeed let’s use my estimate of 1/3.  That still just gets us to April 2008, when the consensus view of economists was that no severe recession was likely.  You still need a big drop in NGDP.

And lets say I’m wrong about the Great Recession not being caused by the severe financial crisis of late 2008.  Even then, even granting all Millman’s assumptions, it would still be true that Fed policy errors caused the Great Recession.  Why?  Because excessively tight money caused the big drop in NGDP in late 2008, which made the financial crisis three times worse.  Which turned it from a crisis that most economists didn’t think was severe enough to cause a recession, into one that most (wrongly) think caused the Great Recession.  If the Fed doesn’t let NGDP growth fall (i.e. if they do NGDP level targeting) then the financial crisis stays much smaller and we end up with unemployment rising from 4.9% in April 2008 to perhaps 6% in late 2009, not the 10% that actually occurred.

So even if Millman is right on all 3 points, Yglesias’s is still right about monetary policy.  But Millman probably isn’t right, even though he presents a highly effective defense of the conventional wisdom.  If I was presenting a case to 12 jurors I’d much rather be in Millman’s shoes.  But if I had to guess, it’s Yglesias’s take that will eventually be accepted by monetary historians.  If you used a Great Depression analogy; Millman’s giving the Austrian perspective and Yglesias is defending the Friedman and Schwartz view.  And we know which view won out in the long run.

HT: Ramesh Ponnuru



31 Responses to “How conventional wisdom gets formed”

  1. Gravatar of StatsGuy StatsGuy
    3. February 2012 at 05:47

    I know you think Bernanke is secretly trying to target NGDP, although using a “framework” that is not explicit. His actions, however, point to him as a pure inflation targeter.

    “We are not seeking higher inflation,” Bernanke said yesterday in response to questioning from Republican Representative Paul Ryan of Wisconsin, chairman of the House Budget Committee. “We do not want higher inflation and we’re not tolerating higher inflation.”

    Bernanke replied to Ryan’s suggestion that the Fed might be prepared to allow inflation to exceed its goal to fulfill the second part of its congressional mandate, which is to promote maximum employment. The Fed last week set a 2 percent annual inflation goal after saying that it would keep its benchmark interest rate low for longer to boost the economy and push down unemployment.

  2. Gravatar of Morgan Warstler Morgan Warstler
    3. February 2012 at 06:12

    Scott at a 4.5% level target from 2000 onward, your own theory says that tightening would occur far earlier, in fact it would have made it impossible for Fannie and Freddie to have gotten us into this mess.

    It would have been OBVIOUS and immediate, because remember the NGDP paradigm is obvious and immediate, and you admit openly that paradigm would lead to smaller government.

    We’re bouncing along at 4.5% NGDP and F&F standards start to relax and BAM! the computer / futures market running the Fed raises rates.

    They do not alter their activities, and BAM BAM BAM up go rates.

    The outcry is immense, we’re going to slow down real growth, so a bunch of bad credit risks can have houses????

    “Screw the creditless!” we hear yelled through-out the land.


    Now that is the Sumner story, it is exactly, or nearly exactly what you think happens if we started your plan in 2000.

    You don’t actually agree with Matty, you actually agree with Millman:

    1. The real problem was bad F&F policy.
    2. The Fed could have and should have neutered that policy – your plan does neuter that policy.

    The most damning transcripts are from 2006, not 2008.

    Here’s the point:

    UNLESS you forcefully make Matty lose what he wants to keep “poor people get houses” – Millman et al., don’t see that Matty is only right in the above under a POLICY where poor people do not get houses.

    You can’t have a NDGP target kept stable AND poor people get houses.

    They are mutually exclusive.

  3. Gravatar of Ryan Ryan
    3. February 2012 at 06:30

    Prof. Sumner, I guess my main problem with your take is that I recently came back from a trip to Bangladesh, where the end-point of all this real-estate bubble stuff actually occurs.

    You may not realize the extent of the real-estate bubble because you mostly sit in the developed world, where the “unfinished projects” that Mises warned about take on the appearance of a driveway that takes a year to complete. (That’s a real example – my home builder has not yet completed the driveway in my own house because he doesn’t have enough capital to finish the subdivision he started.)

    Yet, when you go to Bangladesh, you see the unfinished buildings just sitting there. You can’t ignore them because they are propped-up on bamboo sticks. There are gaping holes in the road, where somebody had to start digging and then had to abandon the project for lack of capital.

    And yet Dhaka city is expected to grow. The villagers keep moving there, and they continue to require housing. The expectation is of future growth, not stagnation.

    When you experience an economy where capital is far more scarce than it is in North America, the problem becomes obvious. Neither expectations nor falling NGDP are to blame (as the alarming levels of inflation in Bangladesh prove).

    Instead, what is obvious is that every patch of Dhaka’s soil has been used to build a new apartment building. They have even filled-in the rivers in order to build new buildings. But there is not enough real capital – real value – to finish the construction projects that have begun. Rather than utilizing some of Dhaka’s soil for the purposes of productive enterprise, they have used every scrap they can find for housing. The remaining projects are left unfinished. Eminent domain runs rampant.

    I am convinced that the reason developed-world economists can’t understand the truth of the Austrian perspective is that they have lost touch with how truly scarce resources are. In Bangladesh, it cannot be ignored.

  4. Gravatar of DonG DonG
    3. February 2012 at 06:31

    Looking back I recall that the economy was being driven by two bubbles*: housing and creative financials. Yes they were linked, but they are different sectors.

    Now a bubble doesn’t necessarily have to burst to cause a recession. A stagnation in housing and financials would have lowered GDP increases. Enough to be negative (recession) is a calculation for others. But, when a bubble bursts and assets broadly decrease in value [definition of bubble] that has a stronger impact on GDP and that was clearly enough to cause a recession.

    Add to that another factor, the liquidity crises. Suddenly, businesses didn’t trust each other. Every business needed to raise cash-on-hand. The only way to get a quick boost of cash is to drastically cut expenditures. That gives us 3 distinct recessionary forces. Two are extreme since they involved bubbles bursting and the third was very rapid.

    You can say the FED messed up, but I say the FED was just not agile enough to react to those forces. Even if the FED could boost M by 10% in a month, would we want that?

  5. Gravatar of bill woolsey bill woolsey
    3. February 2012 at 07:02

    The housing crisis, financial crisis, etc., made the Fed’s policy approach of slow, steady decreases in short term interest rates ineffective. Since it is never the Fed’s fault, and so, it can’t be the fault of its policy approach? The fault lies, like always, with the economy. And the housing crises and financial crises must be guilty. If they hadn’t happened, then the Fed’s approach of gradual interest rate adjustments would have been able to keep the output gap low and expected inflation steady.

    By the way, the run up in oil prices can get some blame too.

  6. Gravatar of dwb dwb
    3. February 2012 at 07:21

    in fairness to the Fed, in june-july of 2008, oil was up 80% year over year, 5 yr inflation expectations peaked at 2.7% in mid-july (by the end of august, oil up 40% YoY and 5-yr breakeven inflation at 2% ), 1st quarter real growth was reported at .9% and 2nd quarter real growth advance estimate was 1.9% which was revised up to 3.3% in late august. Unemployment was around 6%-6.5%.

    the speed with which the economy fell off the cliff from september – december of 2008 when credit markets froze was truly astonishing. I do not think it was at all clear how steep the decline was.

    By november-december I think it was clear (unemployment climbed 2% in 3 months) and i don’t think the Fed reacted fast enough. But, its very hard to drive in the dark with no headlights.

  7. Gravatar of Richard A. Richard A.
    3. February 2012 at 07:51

    Matt Yglesias wrote,
    “The damning transcripts are the ones from late 2008 and all of 2009.”

    I fear that these damning transcripts could be so damning that there may be attempts to suppress their release. We mustn’t let that happen.

  8. Gravatar of MP MP
    3. February 2012 at 07:54

    One question. Yglesias says “in the second half of 2008… inflation expectations plunge”. You agree with him. But don’t you also generally say the Fed is currently targeting inflation? So if inflation expectations can come unmoored from the target, why would Fed targeting of NGDP be better able to control NGDP expectations?

    Am I missing a step? Or is the broader point that the Fed isn’t communciating its target or its commitment to that target effectively, as well as having the wrong target?

  9. Gravatar of ssumner ssumner
    3. February 2012 at 08:07

    Statsguy, No I don’t think he’s secretly trying to target NGDP.

    Nor do I think he’s a pure inflation targeter, indeed he recently denied it. And the Fed cut rate sharply in late 2007, which it obviously wouldn’t have done if it was a pure inflation targeter. But I agree he does try to keep inflation near 2% over time, so in that sense he’s an inflation targeter.

    Morgan, Yes, in retrospect a 4.5% target might have been better (although the 2001 recession would have been worse.) You can make good arguments both ways. But that’s not their policy, so they shouldn’t have started that policy in 2008.

    Ryan, It’s a really, really bad idea to think that looking at unfinished houses in Dhaka, Bangladesh tells us anything about the optimal path of NGDP growth in the US.

    And Austrians aren’t the only people who understand “scarcity.”

    DonG, You said;

    “Now a bubble doesn’t necessarily have to burst to cause a recession. A stagnation in housing and financials would have lowered GDP increases. Enough to be negative (recession) is a calculation for others. But, when a bubble bursts and assets broadly decrease in value [definition of bubble] that has a stronger impact on GDP and that was clearly enough to cause a recession.”

    Obviously bubbles don’t have to burst to cause a recession. Very few US recessions are preceded by bubbles. And I’d add that bubbles can burst without causing a recession. Housing construction plummeted sharply in the 27 months after the bubble peaked in January 2006, and the unemployment rate remained nearly unchanged. A massive stock bubble burst in October 2007, and unemployment actually fell in the following years.

    So I’m afraid your facts are wrong. It takes a slowdown in NGDP growth to cause a recession (in most cases, real shocks are occasionally strong enough.)

    The Fed doesn’t have to be “agile” if we do level targeting–indeed that’s the whole point of level targeting.

    Bill, I agree.

    dwb. The Fed was looking in the rear-view mirror and past CPI inflation, instead of looking down the road at the TIPS spreads. That’s why they were so slow to react.

    The speed of the economy’s decline reflects the failure of Fed policy. With level targeting of NGDP the economy would not have fallen off a cliff in the period after June 2008. And the fall 2008 financial crisis would have been far milder–more like the Bears Stearns crisis early in the year.

    I want the Fed to turn on its headlights.

  10. Gravatar of ssumner ssumner
    3. February 2012 at 08:09

    Richard, Don’t worry, we’ll force them to release the minutes. 🙂

    MP, It was a combination of factors–NGDP gave a clearer signal than inflation, they needed to do level 9not growth rate) targeting, and they needed to target the forecast. All three failures contributed.

  11. Gravatar of M.R. M.R.
    3. February 2012 at 08:16

    I’ve been perplexed before by your argument on this … I’m still not completely getting it.

    Were we really already “half way through a severe contraction” when Lehman failed?

    Looking at the quarterly data, it looks like the major disruption was 2008Q4. I take it that the Macro Advisers monthly data shows something different (you’ve presented this data in a prior post that I can’t find).

    Just trying to understand the right way to think about this.

    Also you say “Eventually the financial crisis went away, and at that point the economy should have begun growing quickly. But it didn’t.” I don’t think the failure to resume fast growth necessarily implies that the financial crisis didn’t trigger the recession. After a heart attack ends, the patient doesn’t go for a run that afternoon.

  12. Gravatar of Ryan Ryan
    3. February 2012 at 08:16

    “t’s a really, really bad idea to think that looking at unfinished houses in Dhaka, Bangladesh tells us anything about the optimal path of NGDP growth in the US.”

    I’m not looking at an NGDP path at all. Instead, I am making the claim that looking at NGDP paths do not tell us anything about economic conditions that pertain to the supply and available of real capital.

    You compared Millman’s view to the Austrian perspective, and I am giving you the real Austrian perspective on the real estate bubble. My argument is that you’re underestimating the severity of the bubble because you’re only looking at the way it popped in North America, and potentially Europe. You don’t see what the “pop” looks like in the developing world.

    It was the destitute countries that bought up the majority of US credit expansion, so the Austrian argument indicates that it is there where you will find the majority of the unfinished projects referred to in Mises’ “Causes of the Economic Crisis.” I am telling you that I went there and my observations confirm what Austrians expect to see.

    In Sumnerian language, the “conventional wisdom” holds true if you trace capital flow beyond US borders.

  13. Gravatar of Ryan Ryan
    3. February 2012 at 08:18

    I might add that inflation occurring in such places also does not bode well for North America…

  14. Gravatar of Morgan Warstler Morgan Warstler
    3. February 2012 at 08:22

    “Morgan, Yes, in retrospect a 4.5% target might have been better”

    Scott, this is your EXACT policy agenda!

    This is your entire LIFE! Every blog post, every boring egghead filled conference, pure economic pilpul just to get this one thing!

    And since you don’t want to ADMIT Matty gets no houses for poor people, you just used the word “might”

    Your friggin personal reputation and any claim you have on future awards is based on the promise that IF WE HAD TARGETED 4.5% NGDP level – we wouldn’t be in this mess now.

    Own up Scott.

    Matty getting houses for the poor DOES NOT HAPPEN under your monetary policy regime.

  15. Gravatar of Benjamin Cole Benjamin Cole
    3. February 2012 at 09:18

    Another terrific post.

    I also wonder why our policymakers do not have “more better” data. Sheesh, with online info it seems to me we should know very accurately how the economy is performing.

    To not react for six months after an economic collapse is criminal. It bespeaks of ritualists who believe in the power of their theories and insights more than what is happening on the ground.

  16. Gravatar of OGT OGT
    3. February 2012 at 09:46

    Scott- This actually gets me back to your EMH post a few days ago. In the comment section of that post you said that EMH had been useful to you in determining that monetary policy was too tight in 2008. I made the assumption that you were referring to the stock market, which you said was incorrect, mentioning the need for a NGDP futures market. Though that begs the question how did you ‘use EMH’ in 2008 without a NGDP futures market?

    This post indicates that the stock market crashed well over half way through the worst of the recession. And that suggests the stock market, at least, was not effective as a forward looking indicator, but suffered a Wily E Coyotte moment along with the Fed, Hank Paulson and the rest of the world. Interestingly the residential housing market, in the US, at least, historically is a much better forward indicator of recessions, leading to the claim that ‘Housing is the Business Cycle,’ and to an extant certainly was this time too. Though that only leads to the question what plausible conclusions can be drawn about the housing market’s NGDP expectations in 2005-2006?

  17. Gravatar of Jon Jon
    3. February 2012 at 10:06

    Scott, great exposition on our common views here. If only the op eds in late 2008 had been as clear. Alas back then, it was only clear that the IOR policy was creating tight money. It was hard to argue effectively against the financial crisis causing panic line of reasoning.

    As an aside, a lot of the talk at that time was disappointedly about shaping an election year narrative. We’ve never managed to purge the mendacity of that period…

  18. Gravatar of Brendan Brendan
    3. February 2012 at 11:50

    Scott, its been so much fun to watch the stream of evidence accumulate over the last few years in favor of your views. Truly unbelievable.

    Hetzel’s new book is out in less than 2 months- any chance that is a major catalyst for further acceptance of your views?

  19. Gravatar of Floccina Floccina
    3. February 2012 at 13:30

    Either way the fed made mistakes. Which leads to the question: How could a system that rests on the knowledge of a few bankers (the Fed’s open market committee) be expected to work any better than it does?

  20. Gravatar of Shane Shane
    3. February 2012 at 13:30

    Just to play devil’s advocate, since I agree with your interpretation, I think the problem with advancing the NGDP account is that then people say: but what caused the fall in NGDP? Your answer is passive tightening. Then people say: OK, the Fed didn’t stop the economy from falling off the cliff, but what actively caused it to head toward that cliff? But there is no active cause at this point. The best we can say is that the economy just drifted off course, and that changes in the financial system and consumer indebtedness exacerbated that drift. It’s like a drunk who falls and breaks his leg: alcohol didn’t really cause the broken leg, but people see it as the cause from a moral purpose.

    Kant separated empirical causality from moral causality and perhaps something like this is needed here: the empirical cause of the crisis was random drift, but the moral cause was that finance and indebtedness helped create conditions that magnified this drift such that Fed inaction would create more havoc than usual and Fed action would be unusually ineffective.

  21. Gravatar of david david
    3. February 2012 at 17:39

    By the way, you do not need to invoke Austrian capital theory to explain half-completed construction in Bangladesh – you need only observe that, being a very poor country, it receives a generous amount of infrastructure aid. It is unfortunately also a very corrupt country, one of the most corrupt in Asia. Western agencies dispense the cash, the right palms are greased, a grand ground-breaking ceremony is staged, and then in the nature of aid agencies, the project is forgotten. Therefore: half-completed buildings.

    You cannot deduce that a abandoned project has negative value because even a positive-value project in Bangladesh is unlikely be viable without generous bribery.

    (the falseness of the malinvestment story is more evident in Southeast Asia circa 1997, where many stopped projects were resumed once the crisis receded)

  22. Gravatar of ssumner ssumner
    3. February 2012 at 19:11

    MR, I don’t think the heart attack is a good analogy. RGDP grew really fast after the early 1933 banking crisis, which was very severe. Indeed the crisis was still going on in the early stages of that fast growth.

    Quarterly GDP is very misleading, which is why monthly data is better. The Macroeconomics Advisers data is consistent with the quarterly data, but more accurate. Example: Suppose RGDP is flat up to the very last day of the second quarter, then falls at a 8% annual rate for the rest of the year. The third quarter will show a 4% decline and the 4th quarter will show an 8% decline (both at annual rates), even though output is declining at exactly the same rate in both quarters. Most people don’t realize that. Quarterly growth rates are quarter over quarter, not beginning to end of quarter.

    Ryan, You said;

    “It was the destitute countries that bought up the majority of US credit expansion”

    This isn’t even close to being accurate, the developing countries were net lenders to the US.

    Morgan, But it’s not better until the Fed learns how to deal with the zero bound. Do that and I’ll take 4%, maybe 3%.

    Ben, I wondered about that too.

    OGT, I wouldn’t say I put zero weight on the stock market, but rather a small weight. You look at many market indicators and try to infer what an NGDP market would show if we had won. Maybe 50% on TIPS spreads, and another 50% on a mix of real estate prices, stock prices, commodity prices the foreign exchange value of the dollar, and private sector consensus RGDP forecasts. All indicators were suggesting recession and disinflation. If only the stock market shows that (say late 1987) then you put much less weight on it.

    You said;

    .”And that suggests the stock market, at least, was not effective as a forward looking indicator, but suffered a Wily E Coyotte moment along with the Fed,”

    Actually stocks had already fallen well below their 2007 peaks. But the bigger problem I have with this is it makes the Fed seem like an innocent bystander. The stock market crashed because they correctly saw that the Fed was adopting a tight money policy (higher IOR) just when easy money was needed.) I think they had wrongly assumed (as I has wrongly assumed) that Helicopter Ben wouldn’t allow NGDP to crash, and then stay down. But they were wrong, and found that out in the second half of 2008, especially October.

    Jon, Yes, I really wish I’d had my blog up and running at that time.

    Brendan, I can’t wait until the Hetzel book comes out.

    Floccina, Good question.

    Shane, I think that’s because people really don’t have a clue as to how monetary policy works. Consider the big foreign demand for dollars after the Soviet bloc collapses. Suppose the Fed is passive. There’d be a depression and I guarantee people would have blamed the Fed. That’s because the initial impact would have been higher rates. This time the increased demand for dollars was associated with a credit crash, and hence lower interest rates. So people didn’t notice. But it’s exactly the same phenomenon.

    BTW, If one wants to make this active/passive distinction, then one could argue the Fed caused the initial contraction in late 2007, because the monetary base grew at a much slower rate than normal (about 1%) between May 2007 and May 2008. That’s active tightening according to those who look at things from a money supply perspective.

    David, Good points.

  23. Gravatar of Shane Shane
    3. February 2012 at 20:09

    So in this analogy does housing/financial crisis (+ maybe oil price surge) = collapse of Soviet Bloc? That’s a genuine question as I’m never quite sure if you believe, regardless of the Fed response, that the housing bubble actually did cause money hording in some way, or if it was other causes, or if it was just a random movement in animal spirits, like when a herd of Buffalo spontaneously changes direction.

  24. Gravatar of Morgan Warstler Morgan Warstler
    3. February 2012 at 20:32

    “Morgan, But it’s not better until the Fed learns how to deal with the zero bound. Do that and I’ll take 4%, maybe 3%.”


    Under a targeted level NGDP 4.5% from 2000 (even 5%), Matty does NOT get poor people getting homes loans.

    Your agenda = poor people better off because they don’t get voter gravy, and progressive Dems are worse off for same reason.


  25. Gravatar of david david
    3. February 2012 at 22:36

    Warstler suffers from a really bad case of mood affiliation syndrome.

  26. Gravatar of M.R. M.R.
    4. February 2012 at 05:22

    Scott, fair points. I hear you on quarterly growth rates being quarter over quarter. Nevertheless …

    I found your older post with the Macroeconomic Advisers data, here:

    That post was on 7/16/11. Two weeks later, the BEA released its revised (quarterly) GDP estimates for 2008. That released show a much more significant contraction in 2008Q4 and 2009Q1 than previous estimates. You say above that “the Macroeconomics Advisers data is consistent with the quarterly data.” I wonder whether Macroeconomic Advisers has revised monthly figures. (I confess I haven’t checked – I don’t even know whether they release this publicly.)

    Also, when I look at the Macro Advisers graphs from your older post, I’m not sure I agree with your interpretation. It looks like in August 2008, NGDP was over $14.4tn and RGDP was $13.2tn. If that’s right, then we weren’t “half way through a severe contraction” when Lehman failed. It’s hard to line up the data on the x-axes, but that’s my reading.

    This seems like your main empirical basis for saying that the financial crisis couldn’t have been responsible for the recession. Am I reading it wrong?

    Finally, growth *did* resume pretty much right when the financial crisis ended: middle of second quarter 2009, when the stress tests came out, the big banks raised a ton of capital, and short-term funding spreads got back to pre-crisis levels. Looking at your Macro Advisers graphs, that’s exactly when growth resumed.

    It wasn’t “fast” growth in the sense of catching up with the prior GDP trajectory. But so what? Why should the failure to close this gap be treated as evidence that the financial crisis didn’t bear primary responsibility for the recession? I don’t think that follows.

    Obviously, I’m part of the “conventional wisdom” you’re referring to in the post. It has always seemed apparent to me that the financial crisis bore primary responsibility for the recession. I’m trying to stay open-minded …

  27. Gravatar of David Pearson David Pearson
    4. February 2012 at 07:54

    Don’t confuse a decline in housing activity with a decline in house prices. The latter caused the GFC and the Great Recession (through the broader credit channel). The former was just a side-show.

    House prices (Case Shiller) peaked in mid-2006 and began a shallow decline that lasted through August of 2007. This corresponded to the initial period of rising subprime delinquencies; it culminated in a global run on the ABCP market. Credit conditions then tightened considerably and the GFC began in earnest. House price declines accelerated; a broader swath of shadow bank collateral (AAA-rated) become impaired, and the general economy started to suffer. A paper by the NY Fed (sorry, no link) described how Fed liquidity programs “backstopped” the shadow banks during this time, effectively delaying the impact on the broader credit channel. “Backstopping” worked until the political costs became too high and insolvency (rather than illiquidity) became the issue for Bear, the mortgage insurers (AMBAC, MBIA), Lehman, Citi, and the GSE’s. By August we lost the GSE’s and the Fed lost the ability to control the GFC through liquidity measures. A full-blown run started that culminated in Lehman and the near-failure, in October and November (the period you describe as having “no news”!) of Citi, Morgan Stanley, Merrill Lynch, and an assortment of European universal banks.

    In short, read Gorton.

  28. Gravatar of ssumner ssumner
    4. February 2012 at 19:04

    Shane, The housing crisis would slightly lower the interest rate, and hence slightly increase real money demand. But other factors were much more important in the actual increase in money demand.

    David and Morgan, I have to agree with David. You are the antithesis of Tyler Cowen on the mood affiliation issue.

    Morgan just needs to accept the fact that I’m not a team player. That might have hurt me at various points in my life, but that’s the way I am. I’m not angling for a government job.

    MR, Someone should check to see if MA has revised their monthly estimates based on the new Fed data. As I recall the Fed lowered it’s estiamtes of Q3 growth for 2008, which strengthen’s my case.

    I probably misspoke when I talked about Lehman being 1/2 way through the crash. I was thinking more in terms of the severe systemic crisis, which occurred in October, and was in my view partly triggered by rapidly falling asset prices, which were in turn caused by falling NGDP expectations.

    We should also recall than even in June the economy had already been in mild recession for 6 months.

    I had thought the financial crisis eased much before mid-year, but that’s relying on memory.

    You said;

    “Obviously, I’m part of the “conventional wisdom” you’re referring to in the post. It has always seemed apparent to me that the financial crisis bore primary responsibility for the recession. I’m trying to stay open-minded …”

    This was also the original view of the Great Depression. Only later was it blamed on monetary policy. I agree that it seemed like the financial crisis was to blame. But think about how you believe RGDP would have evolved had the Fed kept NGDP rising at a steady 5% annual rate. Also think about what the financial crisis would have looked like in that case.

    When I raise that counterfactual many people say “of course that would have prevented a severe recession, but the Fed wouldn’t have been able to do that.” In fact, we know from monetary history that a suitably expansionary policy can cause rapid NGDP growth in the midst of a severe banking crisis.

    David, I don’t agree with your timeline. Unemployment was still quite low in April 2008, so I don’t see August 2007 as an important turning point for the real economy, although I agree it was an important turning point for the financial crisis. The key turning point for the real economy occurred in mid-2008, when NGDP started falling. That was the big problem. The subprime crisis was a small problem, which is why economists quite sensibly did not predict a big recession in the spring of 2008. The big mistake (M-policy) had yet to occur.

  29. Gravatar of Shane Shane
    4. February 2012 at 20:45

    Will you do a post on these “other factors”?

  30. Gravatar of ssumner ssumner
    5. February 2012 at 06:16

    Shane, I have in mind slow NGDP growth which led to low interest rates. And also the interest on reserves program. I see those as the two big factors.

  31. Gravatar of Weaving a narrative through the facts | Historinhas Weaving a narrative through the facts | Historinhas
    5. February 2012 at 18:29

    […] recent posts: Matt Yglesias, Noah Millman and Scott Sumner, present a discussion of the “foundations” of the crisis. Scott Sumner also takes on […]

Leave a Reply