Debt surges don’t cause recessions

Here’s Paul Krugman:

Second, a dramatic rise in household debt, which many of us now believe lies at the heart of our continuing depression. Here’s household debt as a percentage of GDP

What do you see?  I suppose it’s in the eye of the beholder, but I see three big debt surges:  1952-64, 1984-91, and 2000-08.  The first debt surge was followed by a golden age in American history; the boom of 1965-73.  The second debt surge was followed by another golden age, the boom of 1991-2007.  And the third was followed by a severe recession.  What was different with the third case?  The Fed adopted a tight money policy that caused NGDP growth to crash, which in turn sharply raised the W/NGDP ratio.  Krugman has another recent post that shows further evidence of the importance of sticky wages.  Forget about debt and focus on NGDP.  It’s NGDP instability that creates problems, not debt surges.

I favor all sorts of public policy changes that would reduce consumer and business debt, such as tax reform and reducing moral hazard in banking.  But the Fed needs to focus like a laser on NGDP.


Tags:

 
 
 

65 Responses to “Debt surges don’t cause recessions”

  1. Gravatar of Bill Ellis Bill Ellis
    18. July 2012 at 09:19

    “What was different with the third case? ”

    Perhaps the debt to income ratio ?

  2. Gravatar of Aziz Aziz
    18. July 2012 at 09:21

    I see something growing more and more unsustainable; as greater levels of debt are accumulated relative to income, more money relative to income goes to paying down debt, meaning that businesses and consumers have less money to spend on consumption and investment, making asset prices unsustainable. This leads to a fall in demand, a burst bubble, and debt-deflation, met by lukewarm Fed reflationism that barely stabilises asset prices as well as adding debt, before the deleveraging trap sets in again due to excessive debt levels.

  3. Gravatar of Tom Tom
    18. July 2012 at 09:25

    Please see my (dead thread?) comments on Net Worth changes to your Money is Half Macro.

    The unseen Net Worth variable would be dropping starting in 2007 or so, and perhaps still dropping or plateau 2010-2012.

    And, if excess household debt was bad, wouldn’t gov’t debt also be bad? In both cases, the badness of the debt depends on the Rate Of Return of whatever is being invested in with the higher debt. Houses that increase in (land) value are much different investments than those that decrease; or in companies like Solyndra that go bust.

    But yes, the Fed should be doing all it can to increase NGDP.

  4. Gravatar of Bill Ellis Bill Ellis
    18. July 2012 at 09:30

    I don’t think Krugman is saying the high debt caused the depression, but that it is limiting demand…so at the heart of why it continues.

    I am pretty sure he blames the collapse of the housing bubble, and the cascade of insolvency it caused in both banks and households for this depression.

  5. Gravatar of Tom Tom
    18. July 2012 at 09:30

    Also, I see that homeowners weren’t increasing their debts much to invest in the dot.com bubble of ’95-2000 (of ‘irrational exuberance’ fame), but did borrow more to invest in housing after 2001.

    I’d even guess that many decided, perhaps after small dot.com losses but still lots of good jobs, housing was a much safer investment than anything else.

    2001-2006, definite bubble. How would NGDP have reduced that?

  6. Gravatar of Evan Soltas Evan Soltas
    18. July 2012 at 09:31

    This is a great point.

    Some economists and commentators assume a “what goes up must come down” view of debt and real asset prices — but as your graph shows, and one sees this in house prices in the US and abroad, that physics aphorism is not as true in economics. That mistake leads these people to conclude that debt deleveraging and protracted real declines in asset values are appropriate or even necessary and that there is nothing monetary policy can or for that matter should do. But there’s a specious prior assumption at the core of their reasoning. A careful look at the data suggests that unit-root type changes are normal, and that it is reversion to a stationary equilibrium which is the exception.

  7. Gravatar of Tom Tom
    18. July 2012 at 09:31

    How would targeting NGDP reduce the bubble.

  8. Gravatar of o. nate o. nate
    18. July 2012 at 09:32

    “What was different with the third case? “

    Perhaps a huge nationwide housing bust – something last seen in the Great Depression? I suppose one could argue the Fed caused that too, but it seems more likely that there was something unique about the circumstances preceding what should have been a routine Fed tightening that caused the unprecedented result.

  9. Gravatar of marcus nunes marcus nunes
    18. July 2012 at 09:53

    Scott: Chris Sims thinks monetary policy effects on fluctuations are pretty weak!
    “Over the course of about 10 years, things that I did and other people followed up on managed to sort out what the effects of monetary policy changes are and distinguish those from co-movements in money and prices and income that didn’t have anything to do with policy. There’s now pretty much a consensus on how monetary policy affects the economy, and on what the size of that effect is. The general conclusion is that it accounts for maybe somewhere between zero and 20 or 25 percent of the fluctuations we see, but if you try to trace out historically, you can’t blame any recession on monetary policy”.

  10. Gravatar of Vivian Darkbloom Vivian Darkbloom
    18. July 2012 at 10:00

    Coincidentally, I ran across another chart yesterday that shows a somewhat longer time-frame (back to 1920). There is seldom one explanation for any one economic crisis; however, in this case, I think PK’s on to something:

    http://www.debtdeflation.com/blogs/2010/06/13/empirical-and-theoretical-reasons-why-the-gfc-is-not-behind-us/

  11. Gravatar of Saturos Saturos
    18. July 2012 at 10:08

    Marcus, clearly he’s never read your blog.

  12. Gravatar of Major_Freedom Major_Freedom
    18. July 2012 at 10:14

    What do you see? I suppose it’s in the eye of the beholder, but I see three big debt surges: 1952-64, 1984-91, and 2000-08. The first debt surge was followed by a golden age in American history; the boom of 1965-73. The second debt surge was followed by another golden age, the boom of 1991-2007. And the third was followed by a severe recession. What was different with the third case?

    If you look at this chart:

    http://research.stlouisfed.org/fredgraph.png?g=8QS

    The difference is that the Fed did not steepen the growth of the money supply even further. They chose (what they perceived as) saving the currency and maintaining dollar hegemony.

    The boom of 1965-1973 had the cost of 1970s stagflation.

    The boom of 1991-2007 had the cost of 2010s stagflation.

    The Fed adopted a tight money policy that caused NGDP growth to crash, which in turn sharply raised the W/NGDP ratio.

    The rate of growth of aggregate money supply would have had to increase even more to prevent the correction from taking place. If the Fed did prevent NGDP from falling, then they would have gone a step closer to vertical money printing, then when the next correction is due, a further acceleration, and so on.

    While the central bank can postpone the corrections for a time, 5, 10, 15, 20 years, at some point, no amount of money printing can prevent the corrections. This is manifested in the central bank finding itself having to accelerate the money supply so fast that the tipping point is reached where the currency is no longer used for economic calculation.

    You believe that the Fed can give us a free lunch, that everything will be OK if the Fed just printed more dollars. If you knew your history, you’d know that it never worked in the long run. If you knew economic principles, you’d know it can’t work. It is what future historians will look back on and say was our version of wanting to turn stone into bread, lead into gold, and paper money into material prosperity.

    Just like the theologians and alchemists of yesteryear were “serious”, highly regarded at the time, and thought themselves correct, the same is true for inflationists of all stripes. You are today’s version of a persistent superstition that goes way back to a time immemorial. It is derived from not knowing that you are actors.

    Krugman has another recent post that shows further evidence of the importance of sticky wages. Forget about debt and focus on NGDP. It’s NGDP instability that creates problems, not debt surges.

    No, it’s the central bank created destabilizing credit expansion (inflation) growth that NGDP is largely composed of that is the problem, not NGDP itself.

    The central bank cannot keep papering over past mistakes forever. They don’t want to lose their control over the money supply, so they typically abstain from accelerating the currency. Those that follow your advice, will inevitably end up as Zimbabwe and Weimar, only in a slightly different way, whereas instead of NGDP accelerating, NGDP growth remains fixed and money supply growth accelerates until a point is reached at which no amount of securities the Fed owns can reduce the money supply to prevent loss of control of NGDP.

    The Fed had to start paying interest on excess reserves to prevent runaway price inflation from their accelerating monetary inflation. They didn’t have enough securities to sell. But as they do buy securities, they are influencing investors to drop their other investments, and buy government debt instead. US debt is shooting through the roof.

    As Antal Feteke writes here:

    “The same thing is happening all over again. When a central bank increases the monetary base three-fold in three years, this is a clear invitation for bond speculators to move in and make a killing. But what the central bank utterly fails to understand is that, contrary to its hopes, new money is not going to the commodity market. Speculative risks there are far too great. Instead, new money is going to the bond market where the fun is. Bond speculation is risk-free. Speculators know which side the bread is buttered.”

  13. Gravatar of Tomasz Wegrzanowski Tomasz Wegrzanowski
    18. July 2012 at 10:16

    Debt is the ultimate sticky price, and high debt load makes already difficult internal devaluation even harder. Why accept a job at lower wages if you still lose your house and go bankrupt even with a job? Having big debt burden (suddenly increased by however much % NGDP fell) massively increases reservation wage, since people will much rather keep looking for a job likely to help them keep up with their increased debt burden than accept a job that won’t.

    How much it matters depends on ratio of debt to NGDP, not on its growth rate, so third situation is obviously very different. If people have low debt levels, it shouldn’t affect their reservation wage at all, regardless of when it happened.

    This can be alleviated by forced debt haircut, or quick and easy bankruptcy, neither of which happened in the Great Recession, especially not in Europe.

    In a way, this is eventually NGDP, but for any given fall in NGDP high debt levels make adjustment far more painful.

  14. Gravatar of Ryan Ryan
    18. July 2012 at 10:29

    “Forget about debt and focus on NGDP. It’s NGDP instability that creates problems, not debt surges.”

    This is the closest I have ever seen of an economist saying “It’s not the fall that kills you, it’s the sudden stop at the end.”

  15. Gravatar of Cthorm Cthorm
    18. July 2012 at 10:37

    Bill Ellis – you said “What was different with the third case?”
    Perhaps the debt to income ratio ?

    Consumer debt-to-income ratios are at 20 year lows.

  16. Gravatar of Rademaker Rademaker
    18. July 2012 at 10:39

    How would you evaluate the hypothesis that a debt overhang confers a disinflationary pressure on an economy (i.e. by making consumers divest income from consumption to debt service), such that the central bank has to raise inflationary counterpressures, which, when certain thresholds (such as hitting the zero bound) are crossed, make less “enlightened” policy analysts than yourself hesitant to take the required action to stabilize CPI inflation by making errors such as reasoning from an interest rate change or thinking the zero bound has any relevance to the effectiveness of monetary policy?

    Also can a certain variety of money illusion perhaps cause people to stop taking on debt at a certain debt-to-GDP level despite that interest rates have fallen to match? Can such a fall in the propensity to take on debt perhaps be the cause of a secular fall in NGDP that might confuse central bankers into taking insufficient counteractive action? I realize that it isn’t part of standard economics that debt growth influences aggregate demand, but there are certain heterodox theories that claim it does which may be worth entertaining here.

    Another thing to point out: in your post you look at one category of debt only. When all forms of debt are aggregated, the rise does not nearly hiccup as much as household debt alone does: http://static8.businessinsider.com/~~/f?id=49d74ad14b54372b00134f9f

  17. Gravatar of Doug M Doug M
    18. July 2012 at 10:42

    Eye of the beholder….I was too young to be actively watching economics in the 1960s. However 1965 – 1982 was looks to me to be a period of high economic instability, flat stock market, runaway inflation and multiple periods of negative real growth.

    91-93 was a recession.

    I said in a comment to an earlier post the deleveraging IS tight money. Banks create money when they lend it. Contrariwise, people create money when the borrow it.

  18. Gravatar of dwb dwb
    18. July 2012 at 10:43

    the causality is reversed: high ngdp growth is usually driven by surges in residential and commercial investment, which are typically financed with debt (because they are long term investments). Debt is just the sum notional amount of prior investment (that has not been paid off yet).

    when you compare debt and gdp, you are just comparing the past level of investment to the present.

    so when you think about a recession causing a “debt surge,” or vice verse, it’s a tautology: current investment is lower than in the past, by definition, so there is a debt surge.

  19. Gravatar of ssumner ssumner
    18. July 2012 at 10:44

    Bill, He’s wrong, the debt bubble didn’t cause the low demand, tight money did.

    Aziz, You said;

    “consumers have less money to spend on consumption and investment,”

    This sort of reasoning is very deceptive. Macro is hard enough as it is, one must not use terms in a vague way. Is the problem too little consumption? In that case saving and investment would rise. Or maybe it’s too little income, leading to a fall both consumption and investment. But that would reflect tight money, if we are talking about a shortfall of nominal income. If nominal income is doing fine and real income falls, that’s a supply-side, problem, not demand. I don’t mean to pick on you, 90% of commenters show similar sloppiness. One must be rigorously precise in terminology and analysis in macro, otherwise you’ll end up in vague generalities that will not convince anyone who doesn’t already agree.

    Tom, You said;

    “2001-2006, definite bubble. How would NGDP have reduced that?”

    I’m not sure it would have, but why should I care? The housing mess isn’t what caused the recession.

    Evan, Good point, But you will find that 99% of people diagree with you and me. When shown a gragh generated by a random walk process, 99% of people (including many elite economists) will see patterns. The human mind is a machine for seeing patterns. We see them even when they aren’t there.

    O. Nate. Undoubtedly there must have been SOMETHING different, otherwise the Fed would not have screwed up so badly. Nonetheless the screw-up was the problem.

    Marcus, That’s right, the 50% fall in NGDP between 1929 and 1933 can’t have had any causal role in the 30% drop in RGDP. I can’t even imagine what he’s thinking.

    Vivian, Thanks for that graph. Here’s my prediction: 99% of people will misread that graph. Most will think it shows a debt bubble before the Great Depression. In fact it shows there was no debt bubble before the Depression. Rather the debt ratio rose DURING the Depression, but only because the denominator (NGDP) fell.

    Tomasz, Debts are sunk costs and hence don’t effect the decision to work at the margin.

    Rayn, I think you misread me. I am saying it’s the fall (in NGDP) that kills you. When you have a debt bubble but no fall in NGDP, you don’t get hurt. Or maybe I misread your comment . . .

  20. Gravatar of ssumner ssumner
    18. July 2012 at 10:54

    Rademaker, You said;

    “How would you evaluate the hypothesis that a debt overhang confers a disinflationary pressure on an economy (i.e. by making consumers divest income from consumption to debt service), such that the central bank has to raise inflationary counterpressures, which, when certain thresholds (such as hitting the zero bound) are crossed, make less “enlightened” policy analysts than yourself hesitant to take the required action to stabilize CPI inflation by making errors such as reasoning from an interest rate change or thinking the zero bound has any relevance to the effectiveness of monetary policy?”

    I can’t tell whether you are making an AD or an AS argument, can you be more specific? Suppose we hit my NGDP target? Are you saying inflation would be unacceptably high?

    Regarding you second point, I’m claiming a tight money policy by the Fed caused the big drop in NGDP. I’m not seeing the Fed as an innocent bystander whose only flaw was to fail to offset a secular fall in NGDP.

    You said;

    “Another thing to point out: in your post you look at one category of debt only.”

    I used the graph Krugman used. I have no idea what sort of graph is best, as I don’t think debt is important. I’m simply using the graph provided by people who do think debt is important. If others come up with better graphs, I’ll respond to those as well. I responded to one by Vivian above, which included more than just consumer debt. It also supports my point.

    Doug, If you think 1965-73 and 1991-2007 were bad times, I’d hate to think how you regard the rest of US history!! I’m surprised we aren’t all starving to death, if those boom periods are bad times.

    dwb, I agree that debt is often procyclical, but that graph shows it is often acyclical as well.

  21. Gravatar of Ryan Ryan
    18. July 2012 at 11:07

    Prof. Sumner,

    It’s like this: If rising debt causes a decrease in NGDP, then rising debt kills you. If NGDP falls for some other reason, then it is that other reason that kills you.

    Changes in outstanding debt may cause a change in NGDP. Changes in employment may cause a change in NGDP. Changes in the money supply may cause a change in NGDP.

    But in all cases, a change in NGDP is the result of some other factor. NGDP does not simply rise and fall in a vacuum, for no discernable reason.

    So basically, what I am suggesting is that when you invite us not to “reason from” anything other than NGDP, you are correctly pointing out that it is the sudden stop at the end (NGDP) that kills us. On the other hand, you are moving us away from any analysis of why on Earth we might be plummeting toward the ground. It’s like trying to prevent suicides by analysing splatter marks on the sidewalk.

    But I do not really understand these things, so take my limited comprehension with a grain of salt.

  22. Gravatar of dtoh dtoh
    18. July 2012 at 11:11

    Net increase in debt results in an increase in consumption and investment.

  23. Gravatar of Bill Ellis Bill Ellis
    18. July 2012 at 11:12

    Scott,
    I did not say that Krugman thought the bubble caused the depression. I said he thinks (IMO) that the collapse of the bubble caused it.

    You took his post as saying the high rate of debt caused it, while I think you got that wrong, that what he was saying in that post is that the high rate of debt is limiting demand.

    Are you saying that high rates of debt don’t limit demand and can’t cause down turns (whatever causes them ) to be more severe ?

    My point is Krugman was not addressing the cause of the depression in his post at all. He was addressing why it goes on.

  24. Gravatar of Bill Woolsey Bill Woolsey
    18. July 2012 at 11:18

    Scott:

    Of course I agree with you. Consider the following:

    There is only a limited quantity of gold. This greatly limits any increase in spending on output. Sometimes, a ballooning debt can increase spending on output a while–perhaps years, even decades–but eventually, it must fall.

    OK, this only applies to a gold standard. When we target the growth path of nominal GDP itself instead of the price of gold, then any intuition based upon an approximately (and for many, I think their intuition is based upon a literally) fixed nominal quantity of gold.

    “Balance Sheet Theorists”:

    Don’t forget about the creditors. For every creditor their is a debtor. What are the creditors doing? You can’t just say that people in debt spend less. What are those with all the assets doing? I think you are implicitly assuming that they necessarily accumulate larger money balances.

    In reality, it is possible to reduce debt and raise interest rates while at the same time increasing both the quantity of money and spending on output. Your problem is that you are assuming these things are impossible.

  25. Gravatar of Doug M Doug M
    18. July 2012 at 11:20

    I was born in the ’60s and was a child in the 1970’s, so I can’t comment from personal expirence. I look at the numbers… unemployemnt rose, the stock market moved sideways, GDP was choppy — not weak but high quarter to quarter variability, CPI rose.

    91-93 was definately a bad time, unemployment was high, I was unemployed (that is the difference between a recession and a depression), housing was weak… 94-99 was booming.

  26. Gravatar of Bill Ellis Bill Ellis
    18. July 2012 at 11:29

    Cthorm,
    Ok, but it was at its peak when the poop hit the fan…

    By way of wiki…”Household debt as a % disposable income rose from 68% in 1980 to a peak of 128% in 2007, prior to dropping to 112% by 2011.”

    If the debt had not been that high, then the down turn in demand would not have been as severe. Right ?

  27. Gravatar of Woj (@BubblesandBusts) Woj (@BubblesandBusts)
    18. July 2012 at 11:30

    The difference in the three instances reflects the aggregate amount of household debt compared with incomes. The use of credit (debt) instead of money (income + savings) has an extra cost associated with the interest payments. As the aggregate amount of debt and interest rises, the percentage of income used to pay interest costs or pay down debt also rises, lowering the amount available for consumption/investment. Fiscal and monetary policy played a role but this difference is important in the size of the hole (drop in NGDP).

    On a related note, while I don’t believe there are nominal limits to debt, I would argue that debt-to-income has a practical limit where creditors no longer believe they will be repaid in full.

  28. Gravatar of Brtio Brtio
    18. July 2012 at 12:04

    Debt surges make the economy much more sensitive to initial small changes in NGDP; which produce non performing loans, which harm bank balance sheets, which reduce lending, which reduces NGDP, which produces more non performing loans, which….

  29. Gravatar of Aziz Aziz
    18. July 2012 at 12:18

    @Scott: One must be rigorously precise in terminology and analysis in macro, otherwise you’ll end up in vague generalities that will not convince anyone who doesn’t already agree.

    Correct. It was sloppy to say “businesses and consumers”. I should have said debtors, because that is who is forced into doing the deleveraging.

    I look at this from a mixed Hyman Minsky/Steve Keen/Ludwig von Mises-ish perspective. I am interested in your perspective because I am considering the possibility of combining Keen’s idea of debt forgiveness (i.e. printing money to pay off consumer debt) with NGDP targeting and writing off debt up to a nominal GDP target. To me this will address both the problem of excessive debt overhang, as well as address any aggregate demand issues (not, in my view, the real problem).

    If you want to have a deeper look at my overall view of this, please read this, it’s a short piece I wrote:

    http://azizonomics.com/2012/07/18/why-i-still-fear-inflation/

  30. Gravatar of Brito Brito
    18. July 2012 at 12:20

    @Aziz

    To me debt overhang is the same thing, or at least causes, low aggregate demand.

  31. Gravatar of Major_Freedom Major_Freedom
    18. July 2012 at 12:26

    ssumner:

    “2001-2006, definite bubble. How would NGDP have reduced that?”

    I’m not sure it would have, but why should I care? The housing mess isn’t what caused the recession.

    The bubble made recession inevitable. If it is not brought about by the Fed abstaining from accelerating inflation, then it will brought about by an eventual abandonment of the currency.

    ———–

    Evan:

    Some economists and commentators assume a “what goes up must come down” view of debt and real asset prices “” but as your graph shows, and one sees this in house prices in the US and abroad, that physics aphorism is not as true in economics. That mistake leads these people to conclude that debt deleveraging and protracted real declines in asset values are appropriate or even necessary and that there is nothing monetary policy can or for that matter should do. But there’s a specious prior assumption at the core of their reasoning. A careful look at the data suggests that unit-root type changes are normal, and that it is reversion to a stationary equilibrium which is the exception.

    The mistake you and Sumner are making rests on a prior assumption that is at the core of your reasoning: That inflation cannot itself cause real side problems if it follows a particular “optimal” targeting trajectory, like 5% NGDPLT. As a corollary assumption then, should problems occur, they must necessarily be a result of “real side” issues that market monetarists can deny being intellectually responsible for. A careful analysis of economic reasoning and of the data however suggests that the problem is the inflation itself.

    The mistake leads you both to making policy recommendations that are more of what caused the problems in the first place; problems that manifest themselves in a sudden decline of “spending” throughout the economy, despite the Fed not stopping its monetary inflation, and certainly despite the Fed not destroying people’s money outright.

    There is no theory or economic principle than can be borne of the data ex nihilo. All understanding of historical data is entirely undrestood by a priori assumptions you brought with you. Case in point, your assumption of constancy in how economic events are related to each other, such that you conclude unit-root changes are “normal” whereas mean reversions are “the exception.” A “careful” look at the data however will show you are indeed looking inflation influenced data. One cannot look at the data, abstract away from the inflation that influenced it, and then come up with an inflationary solution, and then expect to avoid using circular logic. Have you even considered the notion that problems are caused by central bank inflation itself, in which case a “solution” of inflation will only result in a future where it again seems like there is not enough inflation?

    ——————-

    ssumner:

    Evan, Good point, But you will find that 99% of people diagree with you and me.

    Almost the entire economics establishment agrees with you that we need more inflation and that we are stagnating because of not enough inflation.

    When shown a gragh generated by a random walk process, 99% of people (including many elite economists) will see patterns. The human mind is a machine for seeing patterns. We see them even when they aren’t there.

    A random walk process is inferred in, not borne out of, data.

    Every single uptick and downtick of a stock price, is the result of purposeful, not random, acts of investors. They purposefully selected a purchase or sell price, and they purposefully clicked buy or sell. Or what is more common today, computer algorithms are designed to bid and ask hundreds of times a second.

    Just because you can’t mimic physicists and produce an equation whereby a variable’s value is dependent on past values of that same variable, it doesn’t mean you are witnessing a random process. Humans just don’t act according to constancy, that’s all.

    When I see a random walk process in economic data, I see a pattern every single time. I see a pattern of non-constancy in human action.

    ———————-

    Vivian, Thanks for that graph. Here’s my prediction: 99% of people will misread that graph. Most will think it shows a debt bubble before the Great Depression. In fact it shows there was no debt bubble before the Depression. Rather the debt ratio rose DURING the Depression, but only because the denominator (NGDP) fell.

    One can’t say that for post-1980.

  32. Gravatar of Brito Brito
    18. July 2012 at 12:28

    “Almost the entire economics establishment agrees with you that we need more inflation and that we are stagnating because of not enough inflation.”

    That’s not true at all.

  33. Gravatar of Woj (@BubblesandBusts) Woj (@BubblesandBusts)
    18. July 2012 at 12:33

    “Debt surges make the economy much more sensitive to initial small changes in NGDP; which produce non performing loans, which harm bank balance sheets, which reduce lending, which reduces NGDP, which produces more non performing loans, which….”

    Brito – Exactly right. I just expanded on this with my own response (http://bit.ly/OaOn1I).

  34. Gravatar of Doug M Doug M
    18. July 2012 at 12:43

    “Debt surges make the economy much more sensitive to initial small changes in NGDP”

    NGDP is a metric for the state of the economy. To say the economy is sensitive to changes in NGDP is like saying that the location of my car is sensitive to changes in the odometer.

  35. Gravatar of polymath polymath
    18. July 2012 at 13:08

    To me this leads to the question: “If we had NGDP level targeting, how would that fact change household debt as a percentage of NGDP?” And then to the question of whether that would be good or bad.

    Worse than instability is a false promise of stability (e.g. derivatives with unknown counterparty risk during times of stress). Worse than a false promise of stability is a true promise of stability with unfortunate consequences.

    I’m pretty sure that a true promise of NGDP stability would result in a large increase in household debt — in fact in all debt as a percentage of NGDP. At the same time, it would be “obvious” to everyone that loan loss reserves could go down dramatically because empirically loan defaults would drop and stay dropped. I think those consequences are very easy to foresee. Other consequences would be more difficult to foresee.

    Would that be a good thing? I don’t think debt is evil, but it doesn’t seem like an obviously good thing either.

  36. Gravatar of Brito Brito
    18. July 2012 at 13:08

    Doug M, NGDP is to some extent auto-causitive, changes in NGDP in one quarter can in itself affect NGDP in the next quarter. I mean I literally just described a causation chain in the next sentence.

  37. Gravatar of To really understand the depression we have to stop “pulling red herrings from the hat” | Historinhas To really understand the depression we have to stop “pulling red herrings from the hat” | Historinhas
    18. July 2012 at 13:15

    […] Scott Sumner counteracts: What do you see?  I suppose it’s in the eye of the beholder, but I see three big debt surges:  1952-64, 1984-91, and 2000-08.  The first debt surge was followed by a golden age in American history; the boom of 1965-73.  The second debt surge was followed by another golden age, the boom of 1991-2007.  And the third was followed by a severe recession.  What was different with the third case?  The Fed adopted a tight money policy that caused NGDP growth to crash, which in turn sharply raised the W/NGDP ratio.  Krugman has another recent post that shows further evidence of the importance of sticky wages.  Forget about debt and focus on NGDP.  It’s NGDP instability that creates problems, not debt surges. […]

  38. Gravatar of Steve Steve
    18. July 2012 at 13:22

    As Krugman might say, every dollar of debt is an asset somewhere else in the economy. If you want to pay down debt, you need the owners of debt to reduce their debt ownership, too. So, who owns all the debt:
    1) Foreign central banks
    2) Wealth Americans (directly, or indirectly, through bond funds, pensions, CDs, IRAs, annuities, etc.)
    3) Retired Americans (through the same as #2)

    If you think there is too much debt, then ipso facto you want those three groups to reduce their fixed income holdings. Surprise surpise, the Tea Partiers and the Financial Repressers are one and the same!!!

    I suspect the underlying demographic motivation for debt is poorly understood. As a growing share of the population is retired, they need to get income from somewhere. It could be: tax transfers, interest income, or equity income. In order to support retirees, the sum of those three need to be adequate. One reason for higher debt/income than in the past is well-to-do retirees saving (lending) to generate an income stream in retirement. And the need for debt stock only goes higher if you move to reduce the social safety net, hence the relentless fall in interest rate.

    Ultimately the debate is whether a higher debt/income ratio is destabilizing, or *necessary* to support the incomes and of wealthy and retirees.

  39. Gravatar of Steve Steve
    18. July 2012 at 13:29

    As a follow up, I predict that if the Fed moved to “normalize” interest rates, to say 1.5%, the yield curve would invert faster than you could say “boo!” But I’m just one of those ‘stupid’ market people Scott wrote about earlier :)

  40. Gravatar of Austin Austin
    18. July 2012 at 13:30

    It seems to me that the recovery could be different this time because it’s a financial crisis, and not just a typical recession.

    There is a difference between just withholding spending to get through an uncertain time and being *unable* to spend because a vast amount of wealth has just -poof- disappeared.

    Maybe this explains why recoveries have been so strong in the past, because consumers make up for their delay in spending. But that obviously can’t be the case this time because consumers have to rebuild their wealth, which takes a long time.

    And it doesn’t help to have an administration doing everything in its power to extract the little wealth people do have left and destroying any attempts at creating new wealth.

  41. Gravatar of Major_Freedom Major_Freedom
    18. July 2012 at 13:45

    Austin:

    It seems to me that the recovery could be different this time because it’s a financial crisis, and not just a typical recession.

    “The basic point is that the recession of 2001 wasn’t a typical postwar slump, brought on when an inflation-fighting Fed raises interest rates and easily ended by a snapback in housing and consumer spending when the Fed brings rates back down again. This was a prewar-style recession, a morning after brought on by irrational exuberance. To fight this recession the Fed needs more than a snapback; it needs soaring household spending to offset moribund business investment. And to do that, as Paul McCulley of Pimco put it, Alan Greenspan needs to create a housing bubble to replace the Nasdaq bubble.” – Paul Krugman, Aug 2002

    Just replace “housing bubble” with “bond bubble”, and “Nasdaq bubble” with “Housing bubble”, and you’ll be like Krugman talking to Bernanke.

  42. Gravatar of commonman commonman
    18. July 2012 at 14:20

    Prof.Sumner,
    Don’t understand NGDP, but I guess the 3rd debt surge ended with recession because there was no specific new economic activity.
    What I mean by new economic activity is some new product, related to anything, which people value and would want. (ipad, kindle etc or better way of health service etc).
    In-fact the recession was in fact building from 2000. Till 2000, there was always something new product in the market , latest(94 to 2000) being technology and automation. In countries like USA, anything new in the market is quickly optimized, automated and absorbed to daily life in process quickly scaling down on wages and employment.(outsourcing/off shoring) and as such, people need to move on to new product available quickly. Unfortunately this has not happened from 2000, instead existing activities were scaled up(housing) though the pay-down was nonexistent( there was nothing in market for creating jobs).
    May be better regulation could have avoided this illusionary growth and could have guided the country onto something new, but unfortunately it didn’t happen. Instead now we have more debt and same situation as of 2000. As such it will be doubly painfull for this nation. Already we can see the dawn of protectionism.

    Thanks

  43. Gravatar of Austin Austin
    18. July 2012 at 14:53

    Major Freedom,

    Not sure if you were making a point or not, but while I agree with Krugman on the nuts and bolts of the 2001 recession, I personally wouldn’t advocate for doing anything.

    I was simply pointing out what I see as a fundamental difference between this recession and previous ones. If one were expecting a full and robust recovery I would remind them it’s a different kind of recession.

    It’s going to take a long time to recover and there are no magical bullets to make it quickly all okay. The best we can do is recognize the nature of the problem and not do anything to retard an already slow recovery.

  44. Gravatar of Tom Hickey Tom Hickey
    18. July 2012 at 15:24

    You’ve read Minsky, right?

  45. Gravatar of Brito Brito
    18. July 2012 at 15:30

    Cullen Roche responds: http://pragcap.com/failing-to-connect-the-boom-to-the-bust

  46. Gravatar of Greg Greg
    18. July 2012 at 16:20

    @ Brito, Cullen is sharp, but isn’t he basically agreeing with SS?

    “So I wouldn’t say that rising debt levels always cause recessions, but rising debt levels certainly make it easier for economic agents to act irrationally and irresponsibly thereby substantially increasing the odds of a boom and a bust.”

  47. Gravatar of mbk mbk
    18. July 2012 at 17:03

    Re: rigor.

    The one thing missing in the entire debate is an independent metric for “tight money”, other than either tautology (tight money=decreasing NGDP) or qualitative description. The “correlations” between NGDP, RGDP, and the various financial indicators are riddled with autocorrelations. Comparing graphs is visually effective in blogs but it’s just not a scientific way to explain anything. And therefore it also won’t convince anybody who isn’t already convinced.

    The thought models here are all valuable but without some kind of convincing quantification into independent and dependent variables, and causes and effects, we’re all no better than stock market chartists. I’m sorry but none of all this could sell as science (my own commentaries of course included).

  48. Gravatar of Peter N Peter N
    18. July 2012 at 18:32

    @Doug M

    “”Debt surges make the economy much more sensitive to initial small changes in NGDP”
    NGDP is a metric for the state of the economy. To say the economy is sensitive to changes in NGDP is like saying that the location of my car is sensitive to changes in the odometer.”

    Your odometer is sensitive to it’s short term history. It’s called rate of speed. Distance = Rate * Time. Given the limit on acceleration to, say, 3G, speed Changes take appreciable time.

    More to the point past NGDP and expectations of future NGDP influence demand, as do estimates of future income and expenses.

    Deb

  49. Gravatar of Skepticlawyer » Debt and boom Skepticlawyer » Debt and boom
    18. July 2012 at 19:02

    […] post is partly provoked by this post by Paul Krugman responded to by Scott Sumner and by Marcus […]

  50. Gravatar of Eric G Eric G
    18. July 2012 at 20:09

    “The first debt surge was followed by a golden age in American history; the boom of 1965-73. The second debt surge was followed by another golden age, the boom of 1991-2007. And the third was followed by a severe recession.”

    The first debt surge had tax rates more than twice as high as they are now for the rich, so thank you for pointing out that higher taxes on the rich isn’t a deterrent in growth.

    The 2nd and 3rd “booms” are bubbles.

  51. Gravatar of Lorenzo from Oz Lorenzo from Oz
    18. July 2012 at 21:08

    But public debt surges might help create surges in prosperity–if central banks don’t get in the way.

  52. Gravatar of Vivian Darkbloom Vivian Darkbloom
    19. July 2012 at 00:20

    “What do you see? I suppose it’s in the eye of the beholder, but I see three big debt surges: 1952-64, 1984-91, and 2000-08. The first debt surge was followed by a golden age in American history; the boom of 1965-73. The second debt surge was followed by another golden age, the boom of 1991-2007. And the third was followed by a severe recession. What was different with the third case? The Fed adopted a tight money policy that caused NGDP growth to crash, which in turn sharply raised the W/NGDP ratio”

    Certainly, it is in eye of the beholder. What was different in the third case? For one thing, at the peak of the first two debt surge (1964, if you follow Scott’s deliniation, which definitely is in the eye of the beholder), debt peaked at about 45 percent of GDP and in the second peak (1991), it was 60 percent of GDP. In the third case it peaked at nearly 100 percent of GDP.

    Leverage, public or private, can certainly help to expand the economy—until it doesn’t. I would argue, and I think that this is likely the essence of Krugman’s case (as well as Rogoff, Reinhardt, et al), that there are limits to the amount of leverage an economy can bear. At a certain point, leverage reaches a breaking point. In the larger picture, the measure that is most revealing is the combination of public and private debt. But, when one breaks the threshold in one sector, debt likely needs to be transferred to the other sector (what we are now seeing) and this causes a temporary economic disruption. This transfer can also happen in the reverse through financial repression. If one does not think that this is necessarily the case, I suspect that one is likely (knowingly or not) in the MMT camp.

    One can certainly disagree about the relevance of these differences, but I don’t think one can reasonably argue that in this respect the third case is not different from the first two.

    @Steve

    “As Krugman might say, every dollar of debt is an asset somewhere else in the economy.”

    Krugman might say that, but he’s being very political when he does because he’s pushing for fiscal stimulus which he hopes will result in a permanent expansion of federal government. He brushes aside the fact that those debt assets are not necessarily in “the economy” if, by “the economy” we are talking about the good ole US of A. We are very, very close to the point where more than one-half of that Treasury debt is owed to non-US persons and when that happens he will be more wrong than right. That is also a difference between these three “debt peaks which is not reflected in the charts. We are now witnessing a substantial reduction in the private sector debt load. Part of that is due to actual repayment and part is due to a transfer to the public balance sheet. Financial repression might also be a partial answer in the short run. I don’t think that the economy will experience a vigorous recovery until that deleveraging process is completed. (That is what led me to seek out that chart in the first place).

    Speaking of PK, this reminds me of his rejoinder to Ben Bernanke about the cause(s) of the Great Depression. Bernanke has apologized for the Fed’s role in that, and PK has replied by drawing the distinction between *causing* the Depression and failing to prevent it or pull us out of it. That distinction might well be relevant here, too.

  53. Gravatar of Quibbler Quibbler
    19. July 2012 at 00:23

    You say Debt but you are only using a graph of domestic household debt

  54. Gravatar of Phil Phil
    19. July 2012 at 00:52

    It seems to me that a debt to nominal GDP graph would be more sensible, as the burden of debt is reduced by an increase I nominal wages. Alternatively add real wage growth to your chart. There is a version on my blog. If you look at such a graph you will notice a few features,

    Debt grew in line with GDP until the end of the Breton woods agreement. This seems like a feature of the monetary policies of that era. Since 1985 the level of debt growth has been larger than wage growth. Shown such information the proper inference is that on average interest rates have fallen over this period. They did and ngdp growth was steady as the debt was essentially an expansion of the money supply.

    I believe it is correct that debt has grown inline with falling interest rates. This is how monetary policy was conducted. I believe that the fed could have prevented this outcome by expanding the money supply through QE rather than by lowering interest rates. QE and intest rates should be used together to hit GDP targets without unnecessarily high debt:M0 ratios.

    The story of our financial crises is that expanding the money supply through lowering interest rates only works if the debts can be traded like money in he shadow banking markets. The sub-prime crises suddenly turned liquid assets into illiquid one, ie the money supply dropped, so there was a demand for money and we were in a tight money situation despit low interest rates.

  55. Gravatar of ssumner ssumner
    19. July 2012 at 07:33

    Ryan, You said;

    “But in all cases, a change in NGDP is the result of some other factor. NGDP does not simply rise and fall in a vacuum, for no discernable reason.”

    The Fed determines NGDP.

    dtoh, No, NGDP drives RGDP, not debt.

    Bill, You said;

    “Are you saying that high rates of debt don’t limit demand and can’t cause down turns (whatever causes them ) to be more severe ?”

    Yes, I am saying if the Fed stabilizes NGDP, debt fluctuations don’t matter.

    Bill Woolsey, I agree.

    Doug, I am old enough to remember those years, I think I know the difference between good and bad times.

    Woj, You said;

    “As the aggregate amount of debt and interest rises, the percentage of income used to pay interest costs or pay down debt also rises, lowering the amount available for consumption/investment.”

    This is simply factually wrong. Every debt payment is money received by someone else.

    Brtio, No, monetary policy determines NGDP, not debt. NGDP growth during 2001-07 was unusually SLOW for an expansion.

    aziz, Thanks for the link.

    polymath, I favor public policy changes to sharply reduce debt.

    Steve, You of course are an exception to the rule (i.e not stupid.)

    Austin, The Fed determines NGDP, debt is not a factor.

    Commonman, I’d recommend my National Affairs article, you can google it.

    Tom, No, I haven’t read him.

    mbk, That’s why I keep insisting that the Fed is criminally negligent in not setting up and subsidizing trading in an NGDP futures market. That’s the policy indicator we need.

    Eric, Yes but they were falling. The top rate on labor income (which is what you are presumably referring to, as cap gains rates were far lower) was 90% in 1963 and 50% by the early 1970s. That gradual reduction surely helped growth.

    Lorenzo, I’m not a big fan of public debt. What’s the Aussie net public debt/GDP ratio?

    Vivian, The key difference is the big fall in NGDP. If the Fed keeps NGDP rising at 5%, then the debt overhang gradually gets reduced, without a big recession.

    Quibbler, I explained why above.

    Phil, You said;

    “Debt grew in line with GDP until the end of the Breton woods agreement.”

    The graph shows exactly the opposite.

  56. Gravatar of Vivian Darkbloom Vivian Darkbloom
    19. July 2012 at 07:56

    “Vivian, The key difference is the big fall in NGDP. If the Fed keeps NGDP rising at 5%, then the debt overhang gradually gets reduced, without a big recession.”

    Sorry, Scott, you are not convincing, nor do I think you are responding to the argument. By definition, a “debt overhang” is when you’ve reached the point at which you can’t take on additional debt. When people suddenly stop borrowing (for whatever reason) aggregate demand is going to slow or decline. The bigger the overhang, the larger the effect and the deeper and longer a recession (or depression) is going to be.

    I could perhaps buy an argument that if there is a debt overhang, increasing NGDP at 5 percent might then accelerate the de-leveraging process and therefore shorten a recession, but I doubt it is going to prevent one in face of a large existing debt overhang.

    Perhaps what you are meaning to argue is that a consistent NGDP monetary policy would prevent the debt overhang *in the first place*, but this is certainly not to say that “debt surges don’t cause recessions”, particularly when a “debt surge” is defined here as a percentage of GDP.

    I’m at the point where I think I agree that NGDP targeting might help prevent these types of recessions, or help us get out of one, but I don’t have that much invested in the idea to think that this aspect of monetary policy is the *only* thing that is relevant to preventing or alleviating recessions. I sometimes think that you have so much invested in this one idea and area of macro economics (which by your own admission does not constitute *all* of macro) that you reject out of hand any other explanation for economic phenomena.

  57. Gravatar of The True Role Of Debt In Boom And Bust Cycles – Business Insider « mawillits The True Role Of Debt In Boom And Bust Cycles – Business Insider « mawillits
    19. July 2012 at 18:36

    […] Sumner has a post up today that says “debt surges don’t cause recessions”.   I think this is a lot like saying that eating a lot of food doesn’t make you fat.  Of […]

  58. Gravatar of Bond Vigilante Bond Vigilante
    20. July 2012 at 14:27

    Agree, debt surges don’t cause recessions. But debt need to be serviced. (=interest payments). And rising debts require rising interest payments. The trouble begins when a debtor needs an ever increasing part of his/her income to service the debt. That leads inevitably to a disaster down the road.

  59. Gravatar of e e
    20. July 2012 at 16:29

    Scott,

    Sorry I’m late to this but I don’t think the empirical evidence of sticky wages you and Krugman have been pointing to is very convincing. You’re really just saying that 0 or inflation plus 0 comes up way more frequently than you would expect using a normal distribution, but that just means wage changes arent normally distributed. Specifically Krugmans first chart (I really couldnt see anything in the second) just shows a very kurtotic distribution. But there are lots of kurtotic distributions in the world, and they have nothing to do with stickyness in one direction.

    What you should be looking for is skew, i.e. is the average down change larger in magnitude than the average up change. In other words people won’t lower wages unless they’re going to severely cut them b/c of the unpleasantness of lowering wages. This is the phenomenon you see in micro situations that most people accept do have one directional stickyness. The best example is loss aversion, people care about having “wins” in addition to making money so they tend to lock in winners and let losers run. I think that the empirical evidence (its been a long time since college) shows pretty dramatic skew relative to the fair distribution of the game, bet being studied. And here it is pretty clear that people’s preferrences are actually downwardly sticky at 0. Also, I’m not sure this works but I think you can interpret Becker’s test for discrimination the same way, so that people are upwardly sticky against discriminated groups and you see higher pay for those groups. Now that I think about it I’m not sure that works but I’m going to keep it b/c i think its clever.

    Anyways I still want NGDP targeting but I think the actual evidence for sticky wages isn’t terribly convincing and I would be pretty uncomfortable reasoning from them to a conclusion that doesn’t have as much evidence as NGDP targeting behind it.

  60. Gravatar of ssumner ssumner
    22. July 2012 at 08:31

    Vivian, You said;

    “When people suddenly stop borrowing (for whatever reason) aggregate demand is going to slow or decline.”

    I can’t imagine why a sudden stop in borrowing would cause NGDP to fall if the Fed is pegging the price of NGDP futures contracts. What’s the mechanism?

    e, You are misunderstanding the argument. The argument I’m making is that zero is special, not inflation plus zero. There are lots of wage changes at zero percent, far fewer at minus 1%, and that’s true whether inflation is minus 1% or plus 1% or plus 3%. It shouldn’t be true at a variety of inflation rates.

  61. Gravatar of e e
    22. July 2012 at 17:27

    Ok, I see what you’re saying. FWIW I wasn’t saying you thought I + 0 was special I was proposing that as a mean of the distribution. That said wouldn’t you expect 0 to be the most likely outcome of the distribution even without one directional stickiness? Certainly over the short run.

    At least eyeballing Krugman’s first chart I saw a normal distribution with a single super likely point, but that seems more consistent with two directional stickiness (or just too short a time frame) than the kind of one directional stickiness that you see with loss aversion.

    If you tell me there is a lot of empirical evidence of the kind I’m describing I’ll believe you, but it did seem strange to me that you’re both seemed to focus on the high likelihood of no change.

  62. Gravatar of e e
    22. July 2012 at 17:32

    Looking at Krugman’s chart again there does seem to be some skew if you take the 0 point out and redraw the mean over the new peak. So I guess the evidence is stronger than I thought, but not nearly as strong as you would think if you put a lot of focus on the probability of 0 change.

  63. Gravatar of Major_Freedom Major_Freedom
    23. July 2012 at 09:36

    ssumner:

    dtoh, No, NGDP drives RGDP, not debt.

    Not when inflation is carried out by debt purchases, and by bank credit expansion. Then NGDP drives debt.

  64. Gravatar of Major_Freedom Major_Freedom
    23. July 2012 at 09:44

    ssumner:

    Vivian, The key difference is the big fall in NGDP. If the Fed keeps NGDP rising at 5%, then the debt overhang gradually gets reduced, without a big recession.

    Historically it’s been the exact opposite.

    Looks like debt leveled off along with the big drop in NGDP, whereas prior debt kept increasing.

  65. Gravatar of Uncertainty in China | Economics 274 Winter 2015 Uncertainty in China | Economics 274 Winter 2015
    5. April 2015 at 18:33

    […] if too much private debt does not cause recessions, as economists Scott Sumner and Paul Krugman both firmly assert, an excess of public uncertainty is also probably not a good […]

Leave a Reply