Tuesday, June 24, 2014

What Matters More: the ZLB or Debt Deleveraging?

Must private-sector debt crisis necessarily lead to sharp recessions? Are impaired household balance sheets the reason for the Great Depression, the Great Recession, and the Eurozone crisis? For many observers the answer is an unequivocal yes. People acquired mortgages based on unrealistic expectations about future income streams from housing. They, therefore, overestimated the present value of their homes and took on too much debt. When these expectations failed them, they were forced to deleverage and the resulting drop in aggregate spending ushered in the Great Recession. This view is both intuitive and widely held. But is it complete? Or does it miss a deeper, more important story?

These are the questions I discuss in my review of Atif Mian and Amir Sufi's new book, House of Debt, in the July 7 print edition of the National Review. Here is an excerpt:
Why should the decline in debtors' spending necessarily cause a recession?

Recall that for every debtor there is a creditor. That is, for every debtor who is cutting back on spending to pay down his debt, there is a creditor receiving more funds. The creditors could in principle provide an increase in spending to offset the decrease in debtors' spending. But in the recent crisis, they did not. Instead, households and non-financial firms that were creditors increased their holdings of safe, liquid assets. This increased the demand for money. This problem was exacerbated by the actions of banks and other financial firms. When a debtor paid down a loan owed to a bank, both loans and deposits fell. Since there were fewer new loans being made during this time, there was a net decline in deposits [and thus] in the money supply. This decline can be seen in broad money measures such as the Divisia M4 measure. These developments—increase in money demand and a decrease in money supply—imply that an excess money-demand problem was at work during the crisis.

The problem, then, is as much about the excess demand for money by creditors as it is about the deleveraging of debtors. Why did creditors increase their money holdings rather than provide more spending to offset the debtors? ...Mian and Sufi do briefly bring up a potential answer: the zero percent lower bound (ZLB) on nominal interest rates.

The ZLB is a floor beneath which interest rates cannot go. This is because creditors would rather hold money at zero percent than lend it out at a negative interest rate. This creates a big problem, because market clearing depends on interest rates' adjusting to reflect changes in the economy. In a depressed economy, firms sitting on cash would start investing their funds in tools, machines, and factories if interest rates fell low enough to make the expected return on such investments exceed the expected return to holding money. Even if the weak economy means the expected return to holding capital is low, falling interest rates at some point would still make it more profitable to invest in capital than to hold money. Similarly, households holding large amounts of money assets would start spending more if the return on holding money fell low enough to make household spending worthwhile. This is a natural market-healing process that occurs all the time. It breaks down when there is an increase in precautionary saving and a decrease in credit demand large enough to push interest rates to zero percent. If interest rates need to adjust below zero percent to spur creditors into providing the offsetting spending, this process will be thwarted by the ZLB.

It is the ZLB problem, then, rather than the debt deleveraging, that is the deeper reason for the Great Recession.
Here is another way of seeing the importance of the ZLB problem. The present value of housing is affected by both the expected income streams it will earn as well as the interest rate at which they get discounted. The debt deleveraging story focuses on the fall in the expected income stream, the numerator. The ZLB problem focuses on the lack of offsetting fall in the interest rate, the denominator. Had the interest rates been allowed to reach their negative market clearing (or 'natural') interest rate level, then the present value of housing would not have fallen as much. Household balance sheets would not have been impaired so badly and there would have been less need for deleveraging. In short, there would been no Great Recession but only a ordinary, mild one. So again, the ZLB really is the deeper story here.

P.S. There are ways for policymakers to get around the ZLB and hit the natural interest rate level. So even though it was the deeper story, it is not insurmountable. But these approaches are politically difficult, especially for a central bank that has exhausted its political capital on bank bailouts and make-it-up as go along QE programs. If only the Fed had spent more of its capital on one of these approaches early on.


  1. There is absolutely no need for the bizarre solution put by Miles Kimball (see link in David’s P.S. above). In fact I’m baffled as to what the big problem is that the ZLB supposedly presents.

    If interest is at zero, and government simply prints and spends new money, demand will rise. Government can spend that money on public sector stuff, which those on the political left would like. Or they can cut taxes, which the political right would tend to favour. And the evidence is that when peoples’ after tax incomes rise, their spending rises (amazing as that might seem).

    Plus the above policy increases the private sector’s stock of cash or base money to be exact. And what do people do when their stock of cash rises, e.g. when they win a lottery? Well this is the non-revelation of the century, but their spending rises!!! Amazing.

    The above is of course all second nature to MMTers, and anyone with some common sense. Why the majority of the economics profession don’t get it, I can’t understand. I sometimes wonder whether so called “professional economists” can tie their own shoe laces.

  2. Ralph, the reason 'the majority of the economics profession don't get it" is because they understand such helicopter drops (or OMOs for that matter) do not matter if they are not permanent. Japan tried your approach for many years and still had deflation. There had to be regime change, one that allows the monetary base injection to be permanently higher will it work. Japan has done that via a higher inflation target.

    Think of it this way. Fiscal policy does a helicopter drop that starts to spur aggregate demand growth. Monetary policy, however, fearing higher inflation steps in and sterilizes that drop by tightening policy.

    1. David,

      If helicopter drops don’t work unless they are permanent, then the solution is to make them permanent, as indeed you suggest. So that solves that problem.

      However the idea that the private sector will want to hold a COMPLETELY CONSTANT stock of base money relative to GDP (or a constantly rising stock) till the end of time is frankly bizarre. I would expect to see fluctuations in the desired stock, an “expectation” or assumption that MMTers make.

      As for the idea that a “higher inflation target” is needed as part of the solution, I flatly disagree (assuming the existing inflation target is somewhere around the standard 2% rather than something very unconventional like 0% or minus 5%).

      If inflation is say 1%, government can start helicopter dropping till inflation reaches 2% and in the process, unemployment will decline somewhat. Problem solved. At that point, the economy is at NAIRU. Unemployment at that point may be higher than one would like, but that’s just tough. Labour markets are not 100% efficient, which is why there will always be some unemployment.

      Re your 2nd paragraph (monetary policy steralizing fiscal policy etc), I don’t see the big relevance of that point. Of course, the central bank has the power to wreck the efforts of well thought out fiscal policy (where fiscal policy really is well thought out rather than irresponsible). Likewise the US military (or Russian military) have the power to wreck the US economy by nuking US cities. That doesn’t prove that efforts by the US government and Fed to maximise growth etc are flawed.

      Ergo, I’m sticking to my point, namely that the ZLB is no problem. It can be solved by creating and spending new money into the economy (as advocated by Keynes, MMTers, etc).

    2. Ralph, I did not say the private sector will hold a constant share of monetary base. I only said that (1) it has to believe the there will be a permanent increase and (2) either the central bank has to follow through and/or base velocity has to increase given this belief (and thus not requiring a large permanent injection).

      Also, a higher inflation target is only one solution and is not my ideal solution. All that is needed a temporary bout of inflation and that can be achieved by a level target, a floating exchange rate between cash and deposits, or a higher inflation target. I am not a fan of inflation targeting, but used it here to acknowledge it would at least accomplish the objective of temporarily higher inflation.

  3. "Recall that for every debtor there is a creditor. That is, for every debtor who is cutting back on spending to pay down his debt, there is a creditor receiving more funds"

    If people choose to borrow less, then creditors face the prospect of lower revenues in the future.

    The last thing a bank wants is for everyone to pay off their loans and not take any more loans out. That would completely destroy their business model. For a bank, less loans means less profit.

    So you are wrong to argue that a reduction in borrowing is offset by creditors receiving more funds.

  4. James,

    I think you are missing the point.

    For every debt payment there must be a recipient. If the recipient is a bank its response is different from a non-bank, but the ultimate consequence in both cases is the same. The payment to the bank, which is not wanting to extend more loans because of the crisis, results in a decline in deposits and therefore a drop in the money supply. The payment going to non-banks creditors, who also are concerned about the crisis, ends up being parked in treasuries, money market accounts, and other safe assets.In both cases, it is the recepient's response that is making matters worse. So there is both less money and greater money demand.

    Banks could be making more loans and non-banks could be spending their funds as an offset. But they are not and this is the issue.

  5. Well, I'd love to see the Fed actually try to hit zero. Then we can talk about possible limitations. From what I can see, rates asymptotically approach zero as the monetary base / NGDP increases.

    The base could go infinite, and never see absolute zero. Infinite base would tend to boost NGDP. The ZLB is a chimera.

    Still, I'm really unclear about what the "money" is here: demand deposits are the flip side of lending; reserves are exclusive to the banking system and rate-setting; while currency in circulation is the ultimate deliverable for all financial obligations in the real NGDP economy.

    There were no net new public assets created with QE -- only a bond- for demand-deposit swap. This is only a portfolio rearrangement into cash, and only at market prices. Some bondholder creditors went into cash, that's all (perhaps because they felt that bond present values were going to be hurt by QE: i.e. that the Fed would be effective.)

    Reserves were created, but are dead - held in excess form - because the imputed demand for lending at near-zero interest rates is nil. No need to shift reserves into required form (while IOR also provides additional dis-incentive).

    Insofar as reserves are demanded into currency form, base money reflates NGDP. Currency production at the ZLB is an essential feature, not a bug!

    But, there is a huge wall between getting base money out of reserves and into currency: obtaining cash at the bank is by appointment only, and under legal suspicion. It's too bad, because cash is valuable, almost $1.05 on the dollar: you can get discounts because of credit card fees.

    Negative rates, really? Consider defaults on corporate debt: these could make corporate bond interest rates effectively negative. If negative rates are such a panacea, then bond defaults ought to be manna.

    Try actually hitting zero rates while lifting restrictions on currency usage: problem solved.

  6. "Think of it this way. Fiscal policy does a helicopter drop that starts to spur aggregate demand growth. Monetary policy, however, fearing higher inflation steps in and sterilizes that drop by tightening policy. "


    Wasn't the whole point of your article that the monetary authority wanted higher growth/inflation? And if rate increases eventually were necessary, wouldn't the increase in rates from the monetary authority be the exact escape from ZLB that we are looking for. I'm not following you.

    To be clear, you're saying that if the government sent everyone a cheque for $10,000 every month, we wouldn't see an increase in inflation?

    1. Anonymous, yes, if the government sent everyone a $10,000 check and the Fed was committed to its inflation target the checks would not matter. The Fed would effectively sterilize the helicopter drop.

      The point of my article is that there is an institutional constraint that is keeping markets from clearing. However, in the PS link above, I note that there are ways around it: (1) allowing a floating exchange rate between deposits and currency and (2) adopting some kind of level targeting. Both of these, though, would be regime changes since they would imply some temporarily higher inflation. This would be a big departure from current monetary policy.

    2. "The Fed would effectively sterilize the helicopter drop."

      I was assuming they would 'sterilize' it by raising rates, thus solving the zlb problem, did you have some other mechanism in mind?

    3. Anonymous, remember we are talking about a depressed economy where the natural interest rate is negative. So changes in the target interest rates need to follow sustained economic expansion and a rising natural interest rate. It should not precede them or it will choke off any recovery. I may be wrong, but what I see you proposing is to move the target rate before the natural interest rate rises.

      Another way of saying this is that if the central bank sterilized the helicopter drop by raising rates while the natural rate was still negative (or below the rise in the target interest rate) the economy would fall further into a slump and the rate hike would have to be reversed. This is what happened to the ECB when it raised interest rates twice in 2011. These actions further weakened the economy and it had to reverse itself this year.

  7. "Anonymous, yes, if the government sent everyone a $10,000 check and the Fed was committed to its inflation target the checks would not matter. The Fed would effectively sterilize the helicopter drop."

    But if inflation is below target before the helicopter drop the fed will reach its inflation target due to the heli drop as opposed to being under. The fed would only sterilize the heli drop to ensure inflation doesnt go above target right? It wouldn't sterilize to ensure inflation is below target.

    1. dannyb2b, I agree. But the magnitudes we are talking about here are not very big and probably not sufficient to catalyze a robust recovery.

    2. The U.S. gov't is the largest creditworthy borrower. And they have huge borrowing needs. Whenever the commercial banks buy gov'ts from the non-bank public they create new money & required reserves (but don't incur new capital requirements). The remuneration rate has dis-incentivized the CBs from buying short-term debt obligations to shore up its liquidity & minimize its non-earning assets during the economic contraction. This policy has changed the way the CBs always operated between 1942 & Oct 2008.

    3. David Beckworth

      "dannyb2b, I agree. But the magnitudes we are talking about here are not very big and probably not sufficient to catalyze a robust recovery."

      So just increase the magnitudes if they're not big enough. Increase the frequency and size of the heli drops if they are insufficient.

  8. This is just utterly stupid. Unless money expands at least at the rate prices are being pushed up, output can’t be sold and hence jobs & incomes will be lost (the 1st qtrs. of 2011 & 2014 also prove this). Bankrupt U Bernanke turned "safe assets" into "impaired assets". He drained legal reserves for 29 consecutive months (which means he drained the money stock for 29 consecutive months beginning in Feb 2006 or beginning at the top of the Case-Shiller housing index). Don't confuse "peak debt". with peak prices.

    Then Bankrupt U Bernanke unknowingly conducted a contraction money policy starting July 2008. The rate-of-change in monetary flows (the proxy for real-output), turned negative during the 4th qtr of 2008. This was the trajectory that was predicted in December 2007. If that wasn't bad enough, the 3rd strike was the payment of interest on excess reserve balances that destroyed non-bank lending/investing.

  9. David, know this post is from a bit dated but hope you check back. I see reaching the ZLB as a strictly financial event, not directly related to deleveraging in the context of the post.

    Seems like on the surface, the idea of cash lenders and borrowers is being taken too literally, especially in terms of the ZLB on nominal rates affecting the real economy. The credit crisis impaired financial markets to such an extent that liquidity vanished from overnight markets and the system simply did not function. It was not economic conditions that drove O/N rates to the ZLB over a prolonged economic contraction but an attempt to forestall a financial collapse rapidly during the peak of the crisis in late 2008. This point is critical because in financial markets the ZLB simply represents the level against which funding transactions are priced relative to.

    To make my point in very simple terms, the idea that cash lenders would not invest and “hold” funds at a rate near, equal to, or LESS than 0 is not applicable and in fact does not make economic sense. On numerous occasions in the last 5 years, the implied rate for treasury general collateral has been negative- the persistence of which can only be the result of less visible conditions in securities finance/lending markets. It is true that “for every debtor there is a creditor”, however when a creditor is reinvesting the “free credit balance” from lending specific longer term securities, the dynamic is altered.

    This leads to a complex discussion on the relationship between short and longer term rates and how/when/where the “rehypothication” of securities and effective ‘market value leverage’ at the ZLB “comes back” to household deleveraging in the real economy and the availability of consumer credit.

    Here, I note that in a comment from someone on this post QE is described as- ‘only a bond- for demand-deposit swap. This is only a portfolio rearrangement into cash, and only at market prices. Some bondholder creditors went into cash’…

    While far from what “is actually” happening, this view is in line with what many believe but, once again takes the relationship between the buyer and seller too literally. When the Fed buys a bond with 7 years remaining until maturity, the “market price” will reflect a discount of the future interest and principal cash flows. It held by the Fed till maturity, the full “par value” will either be redeemed by the system or rolled over into new issues at auction- the difference in cost to what is realized being very relevant. Prior to maturity with duration decreasing over time, the security will be lent or pledged from the SOMA securities lending facility or RRP facility at increasing values, effectively draining a higher level of reserves relative to immediate cost. Regardless, it is incorrect to state that any creditor is swapping for a bank deposit during the LSAP program because only primary dealers participate in the operation, and they are VERY far from creditors and bank deposit holders “per se”.

    LSAP’s or QE operations are more like a series transactions with the desk at the NY Fed “auctioning off” non-borrowed reserves at the “highest price”. Some may imagine the desk just bidding across a bunch of securities until a dealer is finally willing to sell when the price is high enough but what actually occurs is the opposite; each security in the maturity range of an operation has an offer price and the desk buys the most economically practical up to the quantity stated- In a sense, the fed knows more about the future they are shaping and buy the cheapest. From 2009 to mid-13’, yields at the longer end steadily declined meaning that the price of every bond sold by dealers did nothing but increase in market value from where it was when originally purchased by the Fed. Why was this good for dealers/"TBTF evil banks" again??? To really see it would mean observing the term structure of the “fungible” Treasury Interest Only STRIP market….a bit much for an already long comment.

  10. I think there's a third explanation, but it's related to the balance sheet idea. Yes, it's true that every debt payment is an income for somebody else. But before anybody starts paying off debt there is the moment in time when everyone realises "these debts can't be paid". At this instant, the debtor feels that their 'real-terms' debt has increased. But the creditor doesn't necessarily see a corresponding mirror image increase in their 'real-terms' assets. This is because the creditor also has to factor in the increased risk of default.

    This is the interesting asymmetry I think. Increased default risk is bad for creditors, but it doesn't contribute an equal-but-opposite advantage to debtors.