Saturday, January 12, 2013

Prolegomena to Any Future Post on Fiscal Policy

Hyman Minsky famously asked, Can "it" happen again? No, he answered, it can't: A deep depression on the scale of the 1930s is not possible in the post-World War II US. One reason why not:
There is a large outstanding government debt... This both sets a floor to liquidity and weakens the link between the money supply and business borrowing.
In other words, a large government debt is stabilizing, because it means that the supply of liquidity -- assets that can serve as, and can readily be converted into, means of payment -- depends less on the state of the financial system. 

Let's take a step back. Any unit in a capitalist economy incurs various money obligations, and receives various streams of money payments. [1] If a unit cannot meet its contracted payments in any period, it must default, with whatever legal consequences that entails. To avoid this, economic units, especially banks, must manage both market liquidity (the ability to convert assets in their portfolio into means of payment) and funding liquidity (the ability to issue new liabilities.) In general, when a bank expands its balance sheet, it becomes less liquid -- that is, it increases the number of possible future states of the world in which it is unable to acquire the means of payment to meet its current obligations. This is why interest rates tend to rise in response to increased private borrowing. 

Or rather, a bank that expands its balance sheet in order to acquire private debt becomes less liquid, or is likely to find itself less liquid in a crisis. A bank that acquires public debt, on the other hand, becomes more liquid. This is why banks hold government liabilities as reserves, beyond any statutory requirements. Government bonds are, in Minsky's words, "ultimate liquidity" -- the only assets that can always be converted to means of payment as needed. This is why interest rates do not rise in response to public borrowing, even when government deficits are very large. [2]

DR is the federal deficit as a percent of GDP. It's the one that goes way up during the war. CPR and RRBR are the two main private interest rate indices for the period. They're the ones that don't.

This all respectable mainstream economic theory. But I don't think the Minskyan implications are really acknowledged in mainstream policy discussions. Do we agree that one of the main reasons that a crisis on the scale of 1929-1933 was impossible postwar was the "floor to liquidity" provided by banks' holdings of federal debt? Then perhaps we shouldn't be surprised that a crisis of that scale almost did occur in 2007, when the share of federal debt in financial-system assets had fallen to less than 5 percent, compared with 15 percent when Minsky wrote those lines.

Indeed, much of the Fed's response to the crisis was various policies to raise this ratio; and one danger of deficit reduction is that the banking system still needs more government bonds. Or as Brad DeLong says:
When the world is short of safe assets--and investors are desperate to hold them--to complain about budget deficits in rock-solid reserve-currency countries and thus about safe asset issuance is profoundly stupid.
Right on; Delong has read his Minsky.

Now, Brad, will you take the next step with me and Hyman? Can we also agree that even when there isn't a shortage of safe assets today, it's good to keep a stock on hand, just in case? Can we agree that if there's a chance that in the next decade the world economy will fly apart due to a lack of safe assets, then it's a bit foolhardy to deliberately reduce the supply of them? Can we agree that, in retrospect, those big Clinton surpluses were -- well, I won't say profoundly stupid, but maybe not the best idea?

And then we can agree that whenever anyone talks about "tackling our long-term government debt problem," what they really mean is "making future financial crises more likely."

EDIT: Obviously, this sounds a lot like Modern Monetary Theory. But while I agree substantively with MMT, I think it's better to think of government liabilities being special because they increase the net liquidity of the financial system, rather than because they can be used to satisfy tax obligations.

There's one other analytic issue, which I haven't seen dealt with satisfactorily. Government deficits operate through two channels: They increase the flow of demand for currently-produced goods and services, and they increase the stock of government debt in private hands. Now, under certain assumptions, you might say these are just two ways of describing the same phenomenon. If you think of the economy as a market with two goods, current output and bonds, then increasing the demand for one and increasing the supply of the other are logically equivalent. But this is not the only way of thinking of the economy. (Among other things, while markets certainly exist as social phenomena, describing the economy as a whole as a market is only a metaphor -- one that may be more or less illuminating depending on the questions we are interested in.) In general, the two channels are going to have two distinct effects, and it would be nice to be able to think them through separately. But almost everyone, across the whole spectrum, tends to collapse them into one.

[1] One reason I like David Graeber is that he understands that this is a better starting point for economic analysis than the exchange of goods.

[2] Obviously this claim applies only to the United States and similar countries. I am going to leave aside for now what "similar" means.


  1. You present the Minsky quote: There is a large outstanding government debt... This both sets a floor to liquidity and weakens the link between the money supply and business borrowing.

    And you paraphrase it: In other words, a large government debt is stabilizing, because it means that the supply of liquidity -- assets that can serve as, and can readily be converted into, means of payment -- depends less on the state of the financial system.

    I didn't know any of that.

    About a year back, you told me: Arjun and I had quite a few discussions about the best way of measuring leverage, and I'm certainly open to the idea that we ended up with the wrong answer. But it's not at all clear to me why debt/M1 would be a better measure.

    My reply now: What Minsky said.

  2. The other day I was talking with my old professor Jim Crotty -- the guy who really taught me Keynes -- and he observed that nobody agrees on the exact meaning of liquidity. There's no standard definition.

    In the current issue of JPKE, there's an article (no free version, unfortunately) that defines liquidity as the sum of liabilities of the banking system and the central bank, which was a new one to me.

    Leverage isn't quite so bad, but there are similar questions, and probably there's no one defintion that is best for all purposes.

    Have you ever looked at the Minsky essays? You might like them.

  3. JW, in the late 1970s there was M1 and M2 and maybe they were experimenting with M1A and M1B. Now M3 has come and gone, and we have MZM and all sorts of financial innovation acronyms.

    I always go back to the footnote from Chapter 13: Without disturbance to this definition, we can draw the line between "money" and "debts" at whatever point is most convenient for handling a particular problem.

    If we can draw the line anywhere, it doesn't matter much that there is no *one* definition of liquidity. But you can't make headway paying down debt if you pay with credit.

    When people show me Minsky stuff I do like it.

    One other simple thought on your post:
    If government debt is a source of liquidity,..
    And if "In general, when a bank expands its balance sheet, it becomes less liquid"...
    Then the greater the level of private debt, the greater the need for public debt.
    Double-sided, like everything else in the economy.

    Thanks JW.

  4. Thought you might find this interesting, some semi-anecdotal empirical support:

    And, with less confidence in the thinking:

  5. Steve,

    Thanks for this. It's reassuring to see I'm not crazy. (Or at least not alone in craziness.) A few points in response:

    As usual, I like the point being made by MMTers (Wray in this case) but wish they would make it slightly differently. Government debt is not supplying money to the private sector, it is supplying safe, liquid assets to the financial system. Framing the issue as the supply of money leads to an overly mechanical conception of the role of fiscal policy (and makes MMT sometimes seem like the mirror image of monetarism, only with monetary and fiscal policy transposed.)

    Following this, I don't think the way to think of it is that the economy cannot grow without a deficit to supply additional means of payment. The banking system is the source of liquidity for the nonfinancial economy, and can expand means of payment just fine without any additional government debt. The problem, rather, is that an expansion of bank balance sheets with a fixed base of federal debt (or similarly stable outside liquidity) leaves the banking system vulnerable to panics, runs and crashes.

    So the logic (IMO) is *not* falling federal debt --> lack of money -> depressed incomes. It is falling federal debt --> increasing bank reliance on inside liquidity --> vulnerability of financial system to crashes --> financial crisis --> depressed incomes. That last link *might* take the form of a fall in the money supply, as in the early 1930s, but it can also take the form of a fall in the credit supply, or fall in private wealth, etc. One important consequence is that we don't need to wonder "how did the economy keep growing in the early 00s without federal deficits to increase the money supply?" as you do in your second post. Banks can increase the money supply without federal debt just fine. The problem rather is that the financial system becomes increasing fragile as it does this.

    We also have to think about the demand side. Both the expansion of the financial system relative to the rest of the economy, and the role of dollar assets as global liquidity, mean that a higher debt-GDP ratio is needed for financial stability.

    The frustrating thing from my point of view is that the New Keynesian orthodoxy basically accepts all these premises, but then assumes the need for safe assets is a unique feature of the crisis that we can ignore when thinking about the long term. Just today, DeLong tells us that "a lower debt-to-GDP ratio would be a good thing in the long run" as if a shortage of safe assets will never be a problem again. Or Krugman -- if he followed the logic of the ISLM model he's constantly using, he would have to agree that we would have been better off if the federal deficit had been larger during the boom. But that seems to be an unthought. Instead they keep compulsively repeating, "Lower debt good!" It's annoying.

    1. This also goes to the point that we need to distinguish the effect of deficits on the flow of demand for currently produced output, from the effect on the stock of federal debt outstanding. A movement of the budget toward surplus directly reduces current demand, but that can be compensated for by any other source of demand, like IT investment in the late 90s or housing investment and housing wealth-spurred consumption during the 00s. The change in the stock of debt is separate from this (except in certain specific models where they are equivalent) and doesn't directly affect current output, but rather the liquidity position of the financial system.

      Again, though I think Wray is right on the big question here, I think he sometimes writes in a way that blurs the two effects, so that the *only* importance of federal deficits is the resulting change in the stock of federal liabilities.

  6. Also, I think that calling some part of GDP "false" because of what was happening on the financial side, can only lead to confusion.

  7. About the creation of money: the relation between M-3 money and events in the economy is spelled out in this publication of the ECB:

    Especially chapter 5 and p. 147 is interesting. Very clearly, banks can create money and legal tender at that, as the central bank guarantees that bank money can be changed 1:1 into banknotes while there also is a 1:1 exchange relation between reserves provided by the central bank (bank reserves) and 'M-3' money created by banks (I stick to M-3 money as it clearly shows the 'best' cyclical behaviour). But the point is of course that banks can only create money when they accept debts emitted by the private sector. And when these debts turn sour, possibly because the collateral turns sour (i.e house prices drop, mortgage credit is the single most important kind of credit around and houses are therewith the most important collateral for our stock of money. The idea that we have a 'house' standard instead of a gold standard or a fiat standard is not entirely ridiculous) the problems start. And banks would be glad when the government would emit debt to the public to buy the mortgages on their balances. It's, in a way, happening in Spain, it happened in a way in Ireland and it happened in the USA.


  8. I've had this inkling that government debt is a hidden government program (as in Mettler's submerged state) but I haven't been able to get my head around it. Your angle helps a lot. The thing about MMT and monetarism being mirror images really clicks too.

    How would you go about determining the optimal level of government debt?

    I could be wrong, but I'm pretty sure Sweden, Denmark, and Finland have had low public debts and more than their share of severe banking crises. That would back you up.

    1. chris
      "the optimal level of government debt" varies with the level of everyone else's debt. This follows from JW's restatement of the Minsky quote.

    2. Does it? Isn't the idea that sufficient government debts will prevent phony safe assets and the financial crises they lead to?

    3. Chris-

      Right. There's an argument that MBSs were developed precisely as a substitute for Treasuries when the demand for Treasuries outstripped the supply. I've linked to this VoxEU piece before -- it makes the point really well (and froma respectable mainstream soure):

      The entire world had an insatiable demand for safe debt instruments – including foreign central banks and investors, but also many US financial institutions. This put enormous pressure on the US financial system and its incentives (Caballero and Krishnamurthy 2008). The financial sector was able to create micro-AAA assets from the securitisation of lower quality ones...

      The massive buying of US government paper by emerging market central banks displaced other investors whose preference previously had been for safe, short-term, liquid assets. ... A piece of a pool of mortgages represents a longer-term asset ... The excess demand for short-term, safe, liquid assets created by emerging economies’ accumulation of reserves could not have been satisfied by the securitisation of US mortgages (and consumer credit) without massive credit and liquidity “enhancements” by the banking system. ...

      From a systemic point of view, this new found source of triple-A assets was much riskier than the traditional single-name highly rated bond. ... [Once the crisis began], the underlying structural deficit of safe assets worsened as the newly found source of triple-A assets from the securitisation industry dried up and the spike in perceived uncertainty further increased demand for these assets. Safe interest rates plummeted to record low levels.

      Though he doesn't quite say so, this makes it very clear that the structural precondition for the crisis was that the lack of sufficient Treasury bonds. Without that, there would not have been the pressure to innovate a substitute safe asset.

    4. That bit about "Knightian uncertainty" sounds a lot like Keynes. That guy should come out of the closet.

      I know SRW disagrees with you on the supply of safe assets. I thought he'd show up by now.

      Is there some underlying problem that should be addressed instead? Who needs all these safe assets? It's not households. If it's pension funds I've got an easy solution for that. If it's China and oil exporters, that's what capital controls are for.

    5. The underlying problem is the desire for liquidity, which is the result of organizing society around the accumulation of money claims.

    6. Slightly more concretely, you really can't organize the development of long-lived assets on financial criteria. Too much of the returns depend on conditions in the distant future. Keynes believed the fundamental solution was the shift of substantially all investment activity away from profit-making enterprises to public or, better, quasi-public bodies. The best discussion of this aspect of Keynes' thought is in Jim Crotty's article "Was Keynes a Corporatist?" One incremental step in that direction would be to roll back the shareholder revolution.

  9. JW - Great post as always. I had thought Art might press upon this point, but I will mention it instead.

    You say "The banking system is the source of liquidity for the nonfinancial economy, and can expand means of payment just fine without any additional government debt. The problem, rather, is that an expansion of bank balance sheets with a fixed base of federal debt (or similarly stable outside liquidity) leaves the banking system vulnerable to panics, runs and crashes."

    I agree entirely. My concern arises with the role played by fiscal (and, to a lesser degree, monetary) policy in subsidizing the expansion of inside liquidity. As the banking system's balance sheet expands at ever-higher rates, pressure builds for increasing outside liquidity even though many macroeconomic variables remain strong. The Treasury/Fed are often reactive, only providing the necessary liquidity after a crisis has begun.

    So although I agree that an expansion of federal debt (or outside liquidity) is not necessary for growth, it may be necessary to maintain growth given a higher desired level of private bank balance sheet expansion. My suggestion is to limit the subsidies to inside liquidity expansion in order to reduce the burden on outside liquidity in future crises.