Recent Posts

Sunday, November 11, 2018

Do Central Governments Need To Issue Bonds (Again)?

The old "should the government issue bonds" debate has come up again. I would point the reader to this article at Mike Norman Economics, as well as the Richard Murphy article it refers to. I would argue that there is limited room for debate. The Treasury of the central government certainly can stop issuing bonds, conditioned on there being changes to the legal/regulatory framework for the central bank. The more important question is whether such a policy is a good idea. My argument is that doing so would run into a variety of consequences, and other policy decisions would need to be rethought (mainly the structure of pension provision).

I have written about this before, and will therefore keep this article short. In particular, I discussed this in Section 6.7 of Understanding Government Finance. I just want to respond to this statement by Richard Murphy:
In the light of my blog on modern monetary theory today and the comment I made in it that the government must act as the borrower of last resort I think it appropriate to republish it. I do so knowing it contradicts modern monetary theory. Political judgement and the needs of financial markets suggests that doing so is appropriate for the reasons I note.
Whether or not the Treasury issues bonds is not a "contradiction" of Modern Monetary Theory (MMT). Certainly that is a policy proposal that has been put forth. However, I would argue that it is a secondary policy issue. (I certainly have a bias in this matter - I just published a book on inflation-linked bonds, which are almost entirely issued by central governments.) There would be ramifications of such a policy shift, and we would need to address those issues at the same time.

Is it Possible? Yes!

It is not hard to find people who argue that suspending bond issuance is impossible because it would run afoul of some particular operating rule under current law. Which is a remarkably silly response. Abolishing bond issuance by the Treasury is a policy proposal. One amazing empirical regularity of policy proposals is that they invariably seem to propose changing policies.

(As a technical note, this discussion does not apply to all "government" bonds; sub-sovereigns would probably have to issue bonds.)

The easiest way to get there is for the Treasury to switch to running an unlimited overdraft at the central bank. Whether or not it pays interest on that overdraft is not material: all the interest income gets swept back to the Treasury anyway!

This does not mean that there will not be central government securities. Currently, there are entities that need default risk free assets, and only the central government can supply them. The ugly solution is to allow entities to bank directly with the central bank. However, this puts the central bank in direct competition with private banks for providing financial services to the non-bank sector. A simpler solution is to issue bills on a fixed price basis. That is, sell unlimited amounts of bills at a fixed yield. This is similar in concept to savings bonds, but they are securities that can be traded in the secondary market. The central bank could be the entity issuing the bills.

Some readers may have concerns about "money printing." These concerns are entirely ideological; we cannot differentiate the proposed system from the existing system within most mathematical models of the economy. The only difference is the intra-governmental accounting, which has no effect on the behaviour of entities outside the central government.

Policy Impact

There would be a number of side effects of such a policy.

  • Prudential financial regulations that refer to Treasury bonds would need to be revised. Since banks can hold settlement balances ("reserves") at the central bank, the banking system should largely be unaffected.
  • Non-bank entities that require risk free assets will need some mechanism to directly hold government liabilities. It will be inefficient to force them to use bank intermediaries to get safe assets, since the patterns of banker behaviour are well documented.
  • From a Minsky-ite perspective, the loss of Treasury bonds will be dangerous for private sector portfolios. The fact that Treasury bonds increase in price during a financial crisis is a key factor propping some entities' balance sheets, giving them the buying power to intervene and stabilise the markets in private sector liabilities.
  • We live in environment where pension provision has been pushed onto individuals. Taking away the only easily understood source of safe assets will make personal pension planning even harder. (Pension and insurance funds need safe assets, and Treasury bonds are extremely useful for their portfolios, as noted in the previous point. However, one might hope that they would have the sophistication to find "safe" private assets, although 2008 showed the limits of such "sophistication.")
  • As a technical addendum to the previous, almost 100% of the supply of inflation protection comes from central governments (Section 4.6 of Breakeven Inflation Analysis).
  • Abolishing bond issuance would largely imply a loss of control of the risk-free interest rate, unless the central bank starts issuing long duration instruments. Although this is not a major concern of MMTers -- who mainly are argue that the effects of interest rates on the economy are mixed -- the reality is that a significant majority of economists (and market participants) believe that interest rate policy is crucial. Losing that policy lever would be a massive political fight, with extremely limited gains. The compromise I would push for is to dump interest rate control in the hands of the central bank, and let them take the political heat for their mistakes.

One might hope that the private sector can sort out the safe asset issue (I have serious doubts), but pension provision is an extremely important question. We have a very large cohort of people in retirement needing guaranteed cash flows. (It would have been a lot easier to muck around with pension policy in the 1960s-1970s, when the population was weighted towards youths.) Although I am not a fan of the policy trend to push pension provision into the hands of individuals, there is no obvious policy fix at present. I see serious political or implementation issues with almost any proposal at this point. Obviously, there can be improvements, but those improvements will likely be highly jurisdiction-dependent.

Concluding Remarks

There is no doubt that central governments can stop issuing Treasury-backed bonds; the question is how to deal with the side effects of the policy.

(c) Brian Romanchuk 2018

9 comments:

  1. Hi Brian, I put some reasons at the link below for thinking a "zero government debt" regime would not be a problem for the pensions industry.

    https://ralphanomics.blogspot.com/2018/11/richard-murphy-tries-to-claim.html

    ReplyDelete
    Replies
    1. You argued that pension schemes are pay-as-you-go. That is literally impossible for individuals, and a key point I made was that we have pushed pension provision onto individuals, either by defined contribution, or tax-advantaged accounts. Even where there are defined benefit funds, many defined benefit private funds are in run-off mode.

      Delete
  2. "The easiest way to get there is for the Treasury to switch to running an unlimited overdraft at the central bank. Whether or not it pays interest on that overdraft is not material: all the interest income gets swept back to the Treasury anyway!"

    Agreed 100%, and that is something that most people don't get, inside and outside the MMT community.

    "This does not mean that there will not be central government securities. Currently, there are entities that need default risk free assets, and only the central government can supply them"

    Or, like in the UK, the central bank could simply start remunerating bank reserves, and then the banks would exchange all their bonds to bank reserves. That would have no major impact in the economy. Maybe it would required some small law adjustments, depending on the country.

    ReplyDelete
  3. Brian, thanks for this post. I'm very interested in MMT, but without formal economics background (chemist by training), so currently just reading as much as I can (including your book). But can you expand a bit on your remark, "Whether or not it pays interest on that overdraft is not material: all the interest income gets swept back to the Treasury anyway!"? Perhaps I'm being stupid, but it's not immediately clear to me why such income to bond holders would automatically return to the Treasury. Thanks for any insight you can provide.

    ReplyDelete
    Replies
    1. (My $&@& phone won’t let me log in; this is Brian).
      I didn’t supply details. The Treasury runs an overdraft at the CB. If it pays interest, the CB will just sweep its own payment back to the Treasury.

      Delete
    2. Just to elaborate on what I think Brian is saying- In the context of a currency issuing government, when the central bank lends money, it creates that money out of nothing, and incurs no cost. If it charges interest on the loans it makes, that interest income is paid back to the national currency issuing government. So if the laws were changed, and if the central bank of the US charged interest on overdraft loans it made to the US federal government, those interest payments by the government would be refunded to the government anyways. So they pretty much don't matter at all.

      But it could matter quite a bit if the government is a "sub-sovereign" and does not issue its own currency. I don't know where the interest the ECB receives from Italy's government (for example) goes. I think that might be a different story?

      Delete
    3. Rolf S. Arvidson,

      "'Whether or not it pays interest on that overdraft is not material: all the interest income gets swept back to the Treasury anyway!'? Perhaps I'm being stupid, but it's not immediately clear to me why such income to bond holders would automatically return to the Treasury. Thanks for any insight you can provide."

      Well, things are complex, and I recommend you the book "Modern Money Theory: A Primer on Macroeconomics for Sovereign Monetary Systems" by L. Randall Wray.

      I don't know if I can explain as well as Wray does, but I will try.

      The thing is that today, in most countries, the Treasury cannot go negative on its reserves. So, if it wants to spend a lot, it will have to issue bonds first. It will either issue to the market (if the market desires it), or to the central bank. Usually the Treasury is not allowed to issue bonds directly to the central bank, but it isn't a big deal: some banks will act as intermediaries between the Treasury and Central Bank, so problem solved.

      Then a bizarre thing happens. The Treasury will pay interest on its bonds to the Central Bank. But, from time to time, the Central Bank remmits its profits back to the Treasury. So, basically, the Treasury is paying interest to itself - the net effect is zero.

      So the interest that the Treasury pays to the Central Bank doesn't matter - not at least if you consider the whole cycle of the Treasury paying interest and receiving it back.

      Also, when you look at the government as a whole (central bank + treasury) it is clear that such a bizarre structure is the remain of a gold standard time that is not real anymore. It doesn't make much sense today. In one way or the other, the whole government is simply issuing currency out of thin air. It's actual financial limit is given by the budget, not the bonds market. (In the US, there is another thing indirectly limiting the amount of spending: the debt limit, although it is risen when the need arrives).

      And I'm talking about self-imposed financial limits, of course. All of those can be legislated away. The limit that cannot be legislated away is the real one. No matter how financial capacity it has, the government will not be able to colonize Venus if there are not real resources available to do that...

      Delete
  4. If the currency-government guaranteed that all households and firms could access government-guaranteed term deposits, wouldn't this remove the need for government bond issuance? Why would the public sector need tradeable government bonds when they could access guaranteed term deposits instead? The term deposits would be a zero-risk interest bearing asset for households and firms. That's all they really need, surely? Why would the asset need to be structured as a tradeable bond?

    ReplyDelete
    Replies
    1. (1) If the term deposit cannot be traded, it's not useful for liquidity management.
      (2) The lack of capital gains means that they do not act as a counter-cyclical crisis buffer.

      Delete

Note: Posts may be moderated, and there may be a considerable delay before they appear.

Although I welcome people who disagree with me, please be civil.

Please note that my spam comment filter appears to dislike long "anonymous" posts. I get no warning about this, and only go through my "spambox" infrequently. The best bet it to keep comments short, and if you think the spam filter struck, let me know with a short comment.