The trouble with modern monetary theory

About a week ago, in response to a post I had put up opposing fiscal stimulus, a reader challenged me to consider the MMT perspective.  I initially deflected for fear of being drawn down a rabbit hole. But I later figured I would give it a shot, here.

One of MMT’s central claims is that governments issuing their own currency do not face a budget constraint beyond that imposed by the need to avoid excessive inflation.  In this post, I propose a thought experiment showing that claim breaks down or dissolves into unhelpful semantics near the limit, which is where it is relevant.

That does not necessarily mean that policy makers adhering to MMT would create a government funding crisis, preceded by a bond market collapse.  Indeed, the bond market might price a deflationary/contractionary correction, to put macro policy back on a sustainable track, out beyond the medium term.  Arguably we have seen a dynamic that rhymes with this play out in Japan over the past couple decades. [1]

The problem with MMT is that would promote bad fiscal planning, eventually a loss of confidence in the real economy and in risk asset markets, and then ultimately extreme disappointment and in some cases hardship among key stakeholders in the public sector. It is ironic, then, that MMT is often embraced by self-described “progressives.” As a liberal (why surrender the term?), I would rather see worthwhile programs anchored in a less tenuous funding base: taxation.

No public sector budget constraint?

MMT embraces a lot of ideas, many of which I find odd.  I will not be sidetracked here by what seem — in this context — to be peripheral issues, except to make the general point  that accounting identities imply co-determination but a not a particular direction of causation.  The specific idea that private savings are determined by the government budget balance confuses causation, and is wrong even as accounting.[2] For fake-balance centrism, let me also just say that I like their take on debt rollovers, which per se should never be a serious issue. US1 is not HYG.

But the sine qua non of MMT, is that a government issuing its own currency need not worry about the path of the government debt because in a pinch the debt can always be funded with base money.[3] Or to put it in MMT terms, governments cannot run out of money.

Here, to cite a prominent example, is Warren Mosler near the opening of The Seven Deadly Innocent Frauds of Economic Policy:

Deadly Innocent Fraud: #1

The federal government must raise funds through taxation or borrowing in order to spend.  In other words, government spending is limited by its ability to tax or borrow.


Federal government spending is in no case operationally constrained by revenues, meaning that there is no “solvency risk.”  In other words, the federal government can always make any and all payments in its own currency, no matter how large the deficit is, or how few taxes it collects.

… This is not to say that excess government spending won’t possibly cause prices to go up (which is inflation). But it is to say that the government can’t go broke and can’t be bankrupt.  There is simply no such thing.

I love that the Frauds are innocent. I have met Mosler, who is a real gentleman. He would never call those who disagree stupid. Stupid is willful, like Fox or pouty potus.  Fools like me are merely “innocent.” Awwwww.

That aside, I am kind of attracted to the Fraud, on practical grounds.  In fact, defending the Fraud is the main point of this post.

There is a semantical sense in which the Truth is true.  If we denote as just inflation the currency debauchment[4] that arises inevitably from using money finance in to relax the public sector budget constraint, then, sure the process is limited by inflation.  Governments don’t run out of money. It is just that printing money to finance expenditure is intolerably inflationary. Helpful?

Practically, that bit of semantics is extremely unhelpful, which we can easily see by considering a thought experiment describing how money finance looks near the limit, which is where it is relevant.

Consider the limit

I think one of the reasons there remains a market for MMT ideas is that people do not have an instinct to consider the limit.  We imagine that MMT is applied at the margin and (correctly, IMV) intuit that it would not be dangerous there.  In other words, MMT seems ok because we reflexively assume it would never actually be extended to its logical conclusion.

I realize that is a provocative claim. I don’t really demonstrate it, and nor is this bit of amateur psychology central to my argument.  But perhaps you could consider the assertion and see for yourself if it rings true. At worst, it might help you follow my argument.

With that in mind, consider this simple thought experiment.  Imagine that US fiscal policy managers (henceforth Treasury) and the Fed embrace MMT and try to relax what would otherwise be the public sector budget constraint by money financing fiscal deficits.  Imagine also, that they do not wait for a funding “pinch” to arrive but shift immediately to MMT operating procedures.  This assumption is not central to my argument against MMT.  I ask you to make it only to simplify the discussion.

Finally, imagine that the debt/GDP ratio rises to a level that would otherwise be difficult to carry if funded with even one-week treasury bills.  The reason for defining the debt level in terms of what would be difficult to finance with bills will become obvious below. For concreteness only, call that debt level 200% of GDP.

The United States could conceivably get to that level of debt within ten to twenty years of using MMT reasoning to “relax” the public sector budget constraint.  But if you prefer abstraction and generalizability, then you can call that debt level X.

Ok, so we wind up with the monetary base (reserves plus currency in circulation) equal to 200% of GDP.  This has resulted from “printing money” to finance past and recurring deficits. At some risk of starting a fight with dictionary writers, I am going to call the monetary base the liability side of the Fed’s balance sheet.

What is formally on the asset side does not really matter, recognizing that the Fed and Treasury are part of the consolidated public sector, and that we should pierce the veil, as MMT strongly urges us to do. It could be Treasury bonds. It could be a loan to the US government. Or it could be bonds or loans that have been forgiven. As Warren Moser would say, “debits and credits.”  (It matters that there is no asset generating income to the consolidated public sector, but I will bracket that discussion for now.[5])

Ok, so the consolidated public sector, with Fed as agent, is sitting there with outstanding monetary base in the amount of 200% of GDP.  To see what comes next, I think we need to consider two scenarios, one where the Fed is paying a market rate of interest on reserves and one where it is not.

Case I: Fed pays market rate of interest

I find it hard to imagine how the narrow money stock could be 200% of GDP without the Fed paying interest on the reserves stock. The equilibrium demand for currency is about 7 to 10% of GDP, and outside liquidity trap, reserves demands is tiny at interest rates below the market rate.[6]  So let’s start by considering the case where the Fed is paying a market rate of interest on reserves, so that they remain “bottled up”, and then go to the more complicated and less realistic set up where reserves earn no interest.

From the perspective of debt dynamics, this first case looks a lot like the Treasury carrying a debt of 200% of GDP financed entirely with one-week bills.  The effective interest rate on the debt is that on reserves, which will be very close to the bill rate, so the resulting debt dynamics are pretty similar.

If the rate of interest is durably below the nominal GDP growth rate, then that it a mercy. If it is above, then that is a problem. The thing is, from a debt dynamics perspective, there is not much difference between carrying the debt as weekly bills or as reserves.

If this were all there were to the story, then we could retire the MMT thesis straight away.  Bills financing would do just fine, just as Larry Summers has proposed in a less imaginary scenario, and to loud guffaws.

But I think implicit in the MMT story is a view that reserves financing is somehow different.  I think the intuition is that there is no risk of default on reserves and that reserves creation can continue indefinitely, even if it corresponds to an unlimited rise of government debt.  (See footnote 5 for the relevance of that condition.)

But that intuition seems mistaken. As Reinhart and Rogoff document in Chapter 7 of This Time is Different, Russia and the axis powers defaulted on base money after the Second World War, because base money was “functionally” a form of government debt that proved not worth honoring. R&R’s details are scarce, and I have not chased the details down.  That sounds to me like a fun project. For now, I am taking it on their authority.

Base money defaults are rare among developed economies since the late 1940s and early 1950s, but this reflects that base money has not been used as an important means of government finance, and not that defaulting on base money is inherently outré.

To repeat, I suspect that the reason this is not obvious to people is that our recent experience has not put us in contact with the limit or logical extension of the MMT perspective.  Our intuition is guided by the experience of conventional public finance, which MMT would overturn.

So, if we imagine a government debt level high enough to be difficult to carry in the form of weekly bills, then we can easily imagine that it might be difficult to carry in the form of reserves as well.  To wit, markets could easily begin to worry about default on base money, and with all the trouble we associate with rising default risk within a conventional public finance perspective.

The Fed and Treasury would then face a choice: course correction or inflation. The first option would involve, first, raising the interest on excess reserves, possibly detaching that rate from those prevailing elsewhere in the money market, crazy as that may sound.

Precisely how that might look probably deserves a separate post. It seems wildly counterintuitive when default risk sends people away from “money” and that what we would normally think of as the “policy” rate needs to go up because of that. Perhaps this is an application of the Lucas critique.  The reserves market is default free, until MMT tries to exploit that.

Anyhow, turbulence in the reserves market, which has effectively become the government debt market, forces to the Treasury to impose steep spending cuts and/or tax increases, to slow the required growth of the reserves stock and drive down the perceived risk of default.

So, this first option looks familiar. No?  It is just how governments conventionally respond, once unsustainable debt accumulation begins to cause trouble.  It arises in the MMT framework, just as in the conventional framework, although I concede the timing could be slightly different if markets were to believe that defaulting on reserves would be more costly (to the economy and politicians) than defaulting on bills.  That complication is relevant. But one thing we do know: how it plays out has nothing to do with the accounting identities and financial sector plumbing so beloved by MMT adherents.

Back now to the choice. The second option within it is that is that the Fed and Treasury  just collectively decide to let inflation rip.  Once that choice is made, it is limited only a disappearance of the currency.  More likely, before that happens, learning encourages the participants involved to abandon inflationary finance, i.e. MMT.

Now the MMT adherent, if I read him correctly, might respond to this criticism by reminding us that the process is limited by inflation and not solvency concerns. But that is pointless semantics because the inflation risk arises directly out of those solvency concerns.

In this story, the government ignores the trajectory of the debt/GDP ratio, default worries creep in and the risk of a public debt spiral manifests.  That sure looks like a solvency issue to me, although you can call it just inflation if you like.  The guy who loses his Social Security or sees his taxes surge won’t fight you on the semantics.

Case II: Reserves carry no interest, like currency

So that’s the relatively realistic case in which the Fed pays interest on excess reserves, while being temporarily blinded by MMT. Now let’s consider the purer but much-less-realistic scenario in which the Fed does not pay interest on reserves.

The first thing to note about this case is that it is almost impossible to imagine, at least within the thought experiment I have proposed.  To wit, there is almost no chance that the monetary base could reach 200% of GDP.  Instead, attempts to lift the monetary base would make the denominator there, nominal GDP, extremely upwardly mobile!  So the trouble I imagine sets in long “before” the monetary base reaches 200% of GDP.

I don’t want to get into this scenario too deeply because it is so unrealistic and because it is quite a bit more complicated than I am smart.  But we can imagine two ways in which this case quickly goes intolerably inflationary, one following a conventional logic and one following the fiscal theory of the price level. I will be brief – and yes, conclusory – on each.

From the conventional perspective, the failure to pay a market rate of interest on reserves initially forces the overnight interest rate to zero or slightly lower, because the supply of reserves is heavily surplus to demand — and initially there is no risk of default on reserves, because the process goes self-limiting before that risk arises.

If an interest rate of zero is too low relative to how the economy is doing, then we get inflation as the Fed falls increasingly “behind the curve”, by virtue of not being able to move.  I will stop there because the story is very familiar to all of you, and I have nothing to add to it.

The fiscal theory of the price level is both more complicated and more controversial.  Here I will just defer to an authority, such as it is,  except to say that I think this intuition works, if not necessarily as a description of normal times then at least at the limit that MMT invites us to consider.

The Grumpy Economist, John Cochrane has emerged as a leading light of the Neo-Fisherian school, which – contra the conventional Keynesians – insists that pegging the funds rate to zero would ultimately lead to lower inflation, not higher. But within that framework, expectations and the institutional set-up are absolutely central, and Cochrane offers as a Caveat that Neo-Fisherianis does not hold when monetary policy is made subservient to fiscal policy, as MMT urges:

Raise interest rates to raise inflation? Lower interest rates to lower inflation? It’s not that simple.

A correspondent from an emerging market wrote enthusiastically. His country has somewhat too high inflation, currency depreciation and slightly negative real rates. A discussion is going on about raising rates to combat inflation. Do I think that lowering rates in this circumstance is instead the way to go about it?

As you can tell, posing the question this way makes me very uncomfortable! So, thinking out loud, why might one pause at jumping this far, this fast?

Fiscal policy.  Fiscal policy deeply underlies monetary policy. In my own “Fisherian” explorations, the fiscal theory of price level is a deep foundation. If the government is printing up money to pay its bills, the central bank can do what it wants with interest rates, inflation is coming anyway. 

Undermining the funding base is hardly progressive

So the central tenet of MMT, at least as regards the public sector budget constraint, seems to me to fall somewhere between flat wrong and a very unhelpful bit of semantics.  Following MMT would result in something that very closely resembles the solvency issues that conventional public finance associated with reckless fiscal planning.  The short run cure for that would be higher interest rates and the medium-term cure would be abrupt fiscal contraction.

So, to say that this process is limited by inflation only is extremely misleading. The “inflation” risk rises predictably from the forlorn attempt to escape the public sector budget constraint, which ultimately is binding, with or without MMT.  Sure you can get around that with hyperinflation, but conventional public finance offers the same “option.”

The point here is not to be alarmist about the bond market or what Trumpian fiscal stimulus might do to it. So long as the MMT guys are kept far away from the levers of power, their take on the current set of arrangement does not seem that far off to me.  Fiscal stimulus would probably put more upward pressure on interest rates and have a lesser effect on growth and employment than the MMT guys assume, for reasons I argued here. But they seem right now, as they certainly were in 2010, to insist that a funding crisis is not currently a relevant concern.

With that in mind, here is my sense of how things might play out if policy makers were to adopt the MMT perspective.  In speculating on this, I assume – perhaps oddly – that throughout people outside the government retain the “correct” view of things.  Government is on the wrong path, but the private sector at least knows it.

Ok, come the adoption of MMT, the first thing to go wrong would be that decent fiscal planning is (even further) undermined. Just to cite one example, we would stop worrying about the sustainability of the Social Security System.  Not only would we no longer make any effort to ensure that that the Social Security System is fully solvent from an internal accounting perspective, but we toss also  idea that it needs at least to be supportable out of general taxation.

Two things would result from this. First, most people would have no idea if Social Security would be there for them. It would be like the conservatives now (perhaps wrongly) claim it is, but much worse. Second, a lot of people would be too optimistic and would put themselves at extreme risk of being disappointed in the future.

Assuming you get that, now broaden it out to all areas where fiscal planning is relevant.  Not good. Wrongly convincing people that there is no such thing as a public sector budget constraint (except for the purposes of managing inflation risk) is extremely unhelpful.  And how promoting this deception could be characterized as “progressive” is beyond me. It would be far more “progressive” to insist on stable financing of public spending programs we think are worthy. The last thing a proper liberal would want to do would be to put these programs in an extreme unfunded status and thereby expose them to conservative starve the beast. (Starve the beast is the old Reaganite idea that we first cut taxes  and later claim that debt worries require spending cuts.)

Under MMT, the second thing to go wrong might be an erosion of confidence among actors in the real economy and risk asset markets.  I am not worried about the bond market, per se, because it might price that the MMT mistake would ultimately be resolved with a return to conventional finance and an abrupt fiscal tightening, which would be contractionary/deflationary on impact. Arguably, a story similar to this has been playing out in Japan, although Japan is not “on” MMT and merely has the debt trajectory that one might associate with that.

Rather, a recognition that fiscal policy is on an unsustainable track and needs to be corrected would push down confidence, push up risk premia, etc., all of which would tend to slow demand growth and maybe raise the risk of returning to the zero bound.  Obviously, I am speculating wildly here. My only high conviction point is that there would likely be trouble at some point and that we need not necessarily look for it in the bond market – or take comfort from the bond market’s complacency in Japan.  I agree with Ken Rogoff when he makes this point in a different context.

Finally, the thing we fear would arrive.  Predictably choosing away from inflation, the Fed and Treasury would collectively deliver a major fiscal contraction, which would risk dramatically slowing growth and – more to the point – would lead to extreme hardship among various stakeholders in the public sector. I have chosen to highlight future Social Security recipients, but you can broaden that out to a lot of others as well, from the aged, to the sick, to the disabled, to even your friendly defense contractor.

MMT does not actually relax the public sector budget constraint.  Believing it could would simply undermine planning and put progressives values at even greater risk than they seem to be now.  So being a good progressive would not require you to support it. Certainly, it is not some forbidden truth that austrians are trying to keep from the people.

[1] I am not sure we are supposed to extrapolate the experience in Japan.

[2] In a closed economy, the public sector financial balance would be equal to the gap between savings and investment, not to savings itself.  A reader tells me that is what the MMT guys really mean. But I cannot address what they really mean, when pressed. I know only what they write. See Mosler’s Third Innocent Fraud for an example of what they write.

[3] A much less bold claim would be that a country issuing its own currency and not involved in an exchange rate peg faces little risk of being pushed into a fiscal crisis by a self-fulfilling prophecy.  Conventional Keynesians and MMT both insist on this point, which I think is important. It has the advantage of not including other, implausible baggage.

[4] Debauchment is such a grumpy-old-man term and I never figured I would use it except in jest.  ZIRP and QE do not debauch the currency.  But it seems appropriate when addressing MMT, something I have never done before. Plus, I am getting older and need to start practicing get off my lawn. Joking aside, I do not allow still-Republicans to visit my property on a social basis.

[5] A comparison with QE seems obvious here. When the Fed did QE, it took government bonds out of the market and replaced them with base money, specifically interest bearing reserves.  The purpose of this policy was not to relax the public sector budget constraint, but to drive bond yields down.  Leaving aside whether it succeeded in that, as a first approximation the policy was not causally linked to a steepening of the trajectory of the debt/GDP ratio.  In contrast, under MMT, the same mechanical operations would be designed to relax the budget constraint and could be seen then as causing – by eliminating a restraint against – a rise of the public debt level. Accordingly, these two operations are not analogous.  Obviously, getting into that would be a tangent.  But this is just another example of where following the mechanics only can lead to confusion.  Incidentally, that principle applies in spade to a consideration of h money, which the monetary side of which is a promise, not a mechanical operation.

[6] When the Fed shifts to paying interest on reserves, as it did at the dawn of QE, the reserves stock arguably becomes debt and not money.  So I am not sure the term “monetary base” has had a stable meaning in recent years.  But I don’t care about semantics and so will still with the conventional language.