Believe those who are seeking the truth. Doubt those who find it. Andre Gide

Wednesday, September 4, 2019

A conversation with Eric Tymoigne on MMT vs SMT

There are a lot of moving parts to the MMT program. I want to focus on one of these parts today: the relation between monetary and fiscal policy. One thing I find appealing about MMT scholars is their attention to monetary history and institutional details. I've learned a lot from them in this regard. But as is often the case with details, one has to worry about whether they help shed light on a specific question of interest, or whether they sometimes let us not see the forest for the trees. And in terms of the broader picture, since I grew up in that branch of macroeconomics that tries to take money, banking, and debt seriously (i.e., not standard NK theory), I sometimes have a hard time understanding what all the fuss is about. Much of standard monetary theory (SMT) seems perfectly consistent with some of the ideas I seen discussed in MMT proponents; see, for example, The Failure to Inflate Japan
This post is devoted to better understanding a contribution by Eric Tymoigne. Eric is one of the people I go to whenever I want to learn more about MMT (if you're interested in MMT, you should follow him on Twitter @tymoignee). In this post, I discuss his article "Modern Monetary Theory, and Interrelations Between the Treasury and Central Bank: The Case of the United States." (JEI 2014). Passages quoted from his paper are highlighted in blue. The working paper version of the paper can be found here. Eric has kindly agreed to respond to my comments and let me post our conversation. We had to some editing, hopefully this did not disrupt the flow too much. In any case, I hope you find it interesting. And, as always, feel free to join in on the conversation in the comments section below. -- DA


DA: Eric, let's start with the opening paragraph:

One of the main contributions of modern money theory (MMT) has been to explain why monetarily sovereign governments have a very flexible policy space. Not only can they issue their own currency to spend and to service their public debt denominated in their own unit of account, but also any self-imposed constraint on budgetary operations can be easily bypassed. 

I'm curious to know what the contribution is here relative to standard monetary theory (SMT). In SMT, the government can also issue its own currency to spend and to service the public debt denominated in its own unit of account. So this degree of "flexibility" is already accounted for. As for "self-imposed constraints on budgetary operations," SMT takes several approaches to this issue, depending on the purpose of the analysis. One approach is to take these constraints as given and then to study their implications. But it is also common to consolidate the central bank, treasury and government into a single authority, which implies no self-imposed constraints on budgetary operations. 

Perhaps what is meant is that MMT shows how existing self-imposed constraints on budgetary operations can be (or are) bypassed in reality.  This leads us to question, however, concerning what those self-imposed constraints are doing there in the first place. Are they there by design and, if so, why? Or are they there by accident (and, if so, how in the world did this happen)?

ET: Yes consolidation is not unique to MMT as we have said repeatedly. Not only is it used quite commonly in the economic literature, but also it is a common rhetorical tool in economic talks, discourse, etc.

DA: Right, so everyone understands this (at least, they should)--it's perfectly consistent with standard monetary theory. So far, so good.

ET: Most economists, politicians and the public don’t understand this or its implications. They will interpret the above as saying that it is obvious that the government can create money but it is not a normal way to proceed and it is inflationary. MMT just pushes consolidation to its logical conclusions and shows that institutional details do back those conclusions. In a consolidated framework, the federal government can only implement spending by creating money, this is not abnormal and it is not inflationary by itself. There is no other way to find the necessary dollars to spend. Here is what consolidation means in terms of balance sheets:
For the federal government, taxes destroy currency (L1 falls) and claims on non-fed sectors falls (A1 falls) (an alternative offsetting operation is net worth of government rises). When US spends, it credits accounts (L1 rises). Similarly, bond issuance does not lead to a gain of any asset for the government; all it does is replace a non-interest earning government liability (monetary base) with an interest-earning government liability (Treasury securities).

DA: I am not going to argue against your accounting. As for bond-issuance, in SMT, an open-market operation is modeled as a swap of zero-interest reserves for interest-bearing treasuries. The interest on treasuries is explained by their relative illiquidity (another self-imposed constraint). The economic consequences of such a swap depends on a host of factors, which I'm sure you're familiar with. 

ET: Sure, in addition, self-imposed financial constraints (e.g. debt ceiling, no direct financing by the Fed, no monetary power for treasury) have been put in place at various times with the argument that they impose discipline in public finances. MMT argues, these financial constraints are not necessary and are bypassed routinely through Treasury-Central Bank coordination.

DA: Sure, the standard view is that these self-imposed constraints are designed to impose discipline in public finance. The proposition that these financial constraints are or are not necessary, however, must be based on a set of assumptions that may or may not be satisfied in reality. (The fact that these constraints may be bypassed through Treasury-Central Bank coordination does not seem relevant to me -- the conflict emphasized by SMT is between an "independent" central bank and the legislative authority (e.g., the Fed and Congress, not the Fed and Treasury). I'm not sure why a new theory is needed here. We know, for example, that if the legislative branch of government fully trusts itself (and future elected representatives) to behave in a fiscally responsible manner, the notion of an "independent" central bank (and other self-imposed constraints) makes little sense. 

ET: Remember that MMT emphasizes the irrelevance of financial/nominal constraints for monetarily sovereign governments (bond vigilantes, risk of insolvency of social security, etc.). One can do that by using the consolidated government (taxes don’t finance, bonds don’t finance, government spends by crediting accounts, etc.) or by using the unconsolidated government (the central bank helps the Treasury, the Treasury helps the central bank). The second method conforms to actual federal government operations but it is much less easy to use rhetorically and it waters down the core point: government finances are never a financial issue as long as monetary sovereignty applies.
Given that point, as you note, financial constraints are not only irrelevant, but also disruptive and used for political games. MMT wants to make government financial operations as smooth and flexible as possible. Once society has decided how, and to what degree, government should be involved in solving socioeconomic problems, finding the money should not be an issue when monetary sovereignty prevails. That means demystifying and eliminating financial barriers to government operations so the political debate can focus on solving real issues (environment issues, socio-economic issues, etc.). Fearmongering about the public debt and fiscal deficits makes for poor political debates and policy prescriptions.

There is a view, expressed by Paul Samuelson, that if we tell policymakers and the public that there are no financial limits to government spending, policymakers will spend like mad; therefore, economists need to lie to policymakers and the public (and themselves). This is nonsense. We ought to discuss policy choices not on the basis of Noble Lies but rather on the basis of sound and informed premises. Economists needs to make sure that policymakers focus on resource constraints.

In addition, political constraints on government should be geared toward improving the transparency and participatory aspects of government (e.g. limit role of big money in elections, limit wastes, etc.). We already have a government that passes a budget (it needs to do so for transparency and accountability purposes), we already have an auditing process, and we already have some (limited) democratic process, so aim at improving these aspects. MMT proponents are not naive, we know that some politicians are self-interested, we know that policy implementation may lead to mistakes, we know people may try to game the system (“free riders”); however we trust that a transparent and democratic government can (and does) get through these issues. MMT does not see financial constraints as helping in any ways, rather they inhibit the democratic process.

Of course, MMT proponents also have a policy agenda (Job guarantee, financial regulation based on Minsky, etc.) because we do not see market mechanisms as self-promoting full employment, price stability and financial stability. As such, as you said, MMT proponents favor alternative means to achieve these goals through direct government intervention. We don’t see the central bank as an effective means to promote price stability. The central bank should focus on financial stability through interest-rate stabilization and financial regulation (an area where the Fed has not performed well).

Finally, yes independence of the central bank is seen as a big deal but MMT disagrees for two reasons. First, MMT emphasizes the lack of effectiveness of monetary policy in managing the business cycle and, second, and probably more importantly, MMT notes that central-bank independence in terms of interest-rate setting and goal settings does not mean independence from the financial needs of the Treasury.

DA: I think it's fair to say most people want to see government operations run smoothly, and would welcome a sober debate over the issues at hand without the fear-mongering that some like to promote. The broad objective seems the same--the debate is more over implementation--how monetary and fiscal policy is to be coordinated--given human frailties. 

Having said this, I think you go too far by asserting that "government finances are never an issue as long as monetary sovereignty applies." Of course, technical default on nominal debt is not an issue (we all understand this). But SMT also recognizes the importance of economic default on nominal debt. True, a government can always print money to satisfy its nominal debt obligation, but if money printing dilutes the purchasing power of money, this is a de facto default. 

On a related issue, SMT asks "what are the limits to seigniorage?" The fact that a government can print money does not give it the power to command resources without constraint. People can (and do) find substitutes for government money (they may also substitute out of taxed activities into non-taxed activities). SMT treats the limits to seigniorage as a financial constraint. Maybe MMT has a different label for this constraint? Perhaps it is related to what I hear MMT proponents call an "inflation constraint." Maybe one way to reconcile MMT with SMT on this score is by recognizing that SMT usually assumes (sometimes incorrectly) that the inflation constraint is always binding. If this is the case, a monetarily-sovereign government does have a financial constraint, even according to MMT. 

ET: Yes, ability to create a currency does not mean ability to command resources because there may not be a demand for the currency. That is where tax liabilities and other dues owed to the government become important (cf. the chartalist theory of money, a component of MMT). That’s also why taxes, monetary creation and bond issuance are not conceptualized by MMT as alternative financing means but rather as complementary. The government imposes a tax liability, spends by issuing the currency necessary to pay the tax liability, then taxes and issues bonds. Spending may be inflationary indeed and so there is an inflation constraint; but it is not a financial constraint, it is a resource constraint.

About the “printing” of money by government, inflation and economic default. Regarding the first two, there is no evidence of an automatic relation between money and inflation. In a consolidated view, government always spends by monetary creation but controls the impact on inflation via taxes and the impact on interest rates via bond issuance. In an unconsolidated view, the central bank routinely finances and refinances the Treasury by helping some of the auction bidders and by participating in the auction.
Finally, regarding economic default, governments routinely “default” in that sense with no problems. I don’t see that as a relevant concept unless someone can show that economic default raises interest rates or generates rising inflation (it does not); here again, there is no automatic link between inflation and interest rates. That link depends on how the central bank reacts; if it does not then market participants don’t either.

DA: Let me return to the manner in which the Fed/Treasury/Congress are consolidated (or not) in SMT and why this matters, in your view. In some SMT treatments, Congress decides spending and taxes, which implies a primary deficit. It's up to the Treasury to finance that deficit, with the Fed playing a supporting role (by determining interest rate and issuing reserves for treasury debt). What's wrong with this approach?
ET: That goes in the right direction with an understanding that the government really has no control over its fiscal position. All this, which relates to the implementation of monetary sovereignty, helps understand why the financial crowding out is not operative, why monetary financing is not by definition inflationary, why i > g is normal. It helps explain why the hysterical rhetoric surrounding the public debt and deficits in nonsense. I recently wrote a piece for Challenge Magazine on that topic. Surpluses are celebrated, governments implement austerity during a recession to “live within our means”, Social Security needs to be fixed to avoid bankrupting it, governments need to save more, etc. All of this is incorrect.

DA: I'm not sure why you claim SMT leads to the idea of i > g. The case i < g is perfectly consistent with SMT (see Blanchard's 2019 AEA Presidential address, and also my posts here and here). The correct criticism (I think) is that mainstream economists have assumed i > g as being the empirically relevant case (it is not).

ET: That is what I meant. MMT links that to monetary sovereignty.

DA: I think that's correct. I should like to add that mainstream economists (apart from a small set of monetary theorists) have not appreciated the role of high-grade sovereign debt as an exchange medium in wholesale financial markets and as a global store of value, which in my view likely explains a lot of the "missing inflation." But as for "surpluses being celebrated," you are now talking about individual viewpoints and not SMT per se. There were plenty of calls out there for countercyclical fiscal policy based on standard macroeconomic principles. But I do agree virtually all mainstream economists are (perhaps overly) concerned about "long-run fiscal sustainability." The view is that at the end of the day, stuff has to be paid for -- and that having the ability to print money, while granting an extra degree of flexibility, does not get around this basic fact. 

DA: I'd like to ask you about this statement you make:

In (the unconsolidated) case, the Treasury collects taxes and issues securities before it can spend. However, federal taxes and bond offerings also serve another highly important function that is overlooked in standard monetary economics. Specifically, federal taxes and bond offerings result in a drainage of funds from the banking system, and MMT carefully analyzes the implication of this fact. From that analysis, MMT argues that federal taxes and bond offerings are best conceptualized as devices that maintain price and interest-rate stability, respectively (of course, the tax structure also has some important role to play in terms of influencing incentives and income distribution; something not disputed by MMT). 

DA: Well, yes, taxes serve both as a revenue device (permitting the government to gain control over resources that would otherwise be in control of the private sector) and as a way to control inflation. I'm not sure about the idea of the Treasury offering bonds for the purpose of achieving interest-rate stability (though this may happen to some extent when the treasury determines which maturity to offer). I don't think this is the way things work in the U.S. today.

ET: Taxes and issuance of treasuries drain reserves and so raise the overnight rate. Hence, on a daily basis, a fiscal surplus raises the overnight rate and a fiscal deficit lowers it. There has been significant Treasury-Fed coordination to smooth the impact of taxes (and treasury spending) on the money market.

DA: Fine, but so what? We all understand "coordination" between Fed and Treasury exists at the operational level. 

ET: I think you are too kind to other economists and policymakers. On taxes as price-stabilizing factors, there is indeed some similarities here. On the role of treasuries for interest-rate stability, it does work like this today. It may not be obvious because of the current emphasis on treasuries as Treasury's budgetary tools, but Treasury has issued securities for other purposes than its budgetary needs. In the US, this occurred most recently during the 2008 crisis (SFP bills). In Australia, in the early 2000s, the Treasury issued securities while running surpluses in order to promote financial stability.

DA: But even if this is not the way things actually work (in my view, it's the Fed that stabilizes interest rates, possibly through OMOs involving U.S. Treasuries), I'm not sure what point is being made. I think we can all agree that monetary and fiscal policy can be thought of as being consolidated in some manner. What would be good to know is how a specific MMT consolidation matters (relative to other specifications) for a specific set of questions being addressed. There is nothing in the abstract or introduction of this paper that suggests an answer to this question.  
ET: The point being made is that in a consolidated government, tax and bond issuance lose the financial purpose they have for the Treasury but keep their price and interest-stability purposes.

DA: In standard monetary theory, tax and bond issuance keeps its funding purposes for the government and at the same time can be used to influence the price-level (inflation) and interest rates. Is this wrong? I don't think so. At some level, taxes (a vacuum cleaner sucking up money from the private sector) must have some implications for the ability of government to exert command over real resources in the economy. What we label this ability (whether "funding" or ''finance" or whatever, seems inconsequential). 

ET: Ok here comes the crucial difference between financial and real sides of the economy. In financial terms, taxes do not increase the capacity of the government to spend, i.e. the government does not earn any money from taxing; taxes destroy the currency. In financial terms, there is no reason to fear a fiscal deficit; deficits are the norm, are sustainable and help other sectors grow their financial net wealth. As such, it is not because a government wants to spend more that it must tax more or lower spending somewhere else. That is the PAYGO mentality. This mentality makes policymakers think of spending and taxing in terms of how they impact the fiscal balance instead of their impact on employment, inflation, incentives, etc. While deficits may have negative consequences, they are not automatic. If one takes a look at the evidence, deficits have no automatic negative impacts on interest rates, tax rates, public-debt sustainability, or inflation.

In real terms, the necessity to increase tax rates to prevent inflation, and so move more resources to the government, depends on the state of the economy and the permanency of the increase in government spending relative to the size of the economy. In an underemployed economy, the government can spend more without raising tax rates. In a fully employed economy, shifting resources to the government without generating inflation does require raising tax rate and/or putting in place other measures such as rationing, price controls, and delayed private-income payment. Here Keynes’s “How to Pay for the War” provides the roadmap. Standard economics is full-employment economics so opportunity costs are always present. MMT follows Kalecki, Keynes and the work of their followers (have a look at Lavoie’s “Foundations of Post Keynesian Economic Analysis”) and note that capitalist economies are usually underemployment and economic growth is demand driven. Put in a picture, the economy is usually at point a.

Put succinctly, the real constraint is conditionally relevant, the financial constraint is irrelevant if monetary sovereignty prevails. That is the proper way to frame the policy debates and to advise policymakers; don’t worry about the money, worry about how spending impacts the economy.

ET: Moving to another topic, consolidation of the government brings to the forefront forces that are operating in the current system but that are buried under institutional complications. Namely that a fiscal deficit lowers interest rates and treasuries issuance brings them back up, that spending must come before taxing and treasuries issuance, that monetary financing of the government is not intrinsically unsound and does not mean that tax and treasuries issuance don't have to be implemented. 

DA: The statement that "deficit lower interest rates" needs considerable qualification. Among other things, it depends on the monetary policy reaction function. As for the claim that spending *must* come before taxes, this is not a universally valid statement (even if it may be true in some circumstances. But even more importantly, who cares? Mainstream theory does not suggest that monetary financing is intrinsically unsound (seigniorage is fine, if it respects inflation ceiling). As for money, taxes and bonds not being alternative "funding" sources, I worry that this semantics. You can call X a "funding" source or not -- it's just a label. The real question is: what are the macroeconomic implications of X? 

ET: Let me emphasize where I agree. Yes, evidence shows the central role of monetary policy for the direction of interest rates, fiscal policy is at best a very small driver. And yes, one ought to focus on the real implications of government spending and we ought to forget about the financial implications. A fiscal deficit is not unsustainable nor abnormal; deficits are the stylized fact of government finances and are financially sustainable if monetary sovereignty is present. So don’t try to frame the policy debate and set policy in terms of household finances, bankruptcy, fixing the deficit, etc.
To conclude I see three reasons why the "taxes/bonds don't finance the government" rhetoric is helpful:

1- It is strictly true for the federal government (i.e. consolidation).

2- it brings to the forefront some lesser-known aspects of taxes and treasuries issuance: impacts on money market, role of central bank in fiscal policy, role of treasury in monetary policy.

3- It changes the narrative in terms of policy and political economy: government does not rely on the rich to finance itself, taxes should be set to remove the "bads" not to finance the government (e.g. one should not set tax rates on pollution with the goal of balancing the budget but with the goal of curbing pollution to whatever is considered appropriate, that may lead to much higher tax rates than what is needed to balance the budget), PAYGO is insane, one should focus on the real outcomes of government policies not the budgetary outcomes.

1. I think this is semantics.
2. Not sure how it helps in this regard.
3. I think all of these positions are defensible without the statement "taxes/bonds don't finance the government", so if this is the ultimate goal (and I think it should be), perhaps we should set aside semantic debates and focus on the real issues at hand. 

ET: 1 is not semantic. I know you have in mind taxes as a means to leave resources to the government. MMT makes a clear difference between financial (ability to find the money) and resources constraint (ability to get the goods and services) as explained above. The financial constraint is highly relevant for non-monetarily sovereign governments so it should be noted and clearly separated from the real constraint. Too many policy discussions and decisions by policymakers operating under monetary sovereignty are based on an inexistent inability to find money and the imagined dear financial consequences of budgeting fiscal deficits. 2 helps to understand how monetary sovereignty is implemented in practice. On 3, yes focus on the real issues.

DA: We agree on 3! Thank you for an interesting discussion, Eric. There's so much more to talk about, but let's leave that for another day. 

ET: You are welcome and thank you too!


  1. Great conversation. Here's just one thought.

    MMT benefits from the "20,000 foot view." They sometimes call this "the general case." That is, the theoretical entry point, or abstract model if you will, for understanding government is one that de-naturalizes the existing division of labor between the Treasury and Central Bank. From that model, you can get some crucial insights like the operational linkage between monetary and fiscal policy (eg. a government deficit for the day adds to bank reserves and decreases the interest rate, necessitating bond sales; or) that go totally overlooked when these are viewed as separate jobs.

    Then, armed with that understanding, MMT compares specific features of real-world countries to see how they stack up against this "general case." However, if there doesn't turn out to be a material difference in the ultimate results, as there isn't in the US, then the additional detail in the "specific case" contributes nothing to most discussions of policy, and so becomes extraneous and distracting.

    Put another way, asking the public to consider "how will the government get the money" is a harmful venture, whereas asking them to consider where real resources should be diverted from (or newly employed) is a beneficial goal. The advantage of MMT here is that it naturalizes the latter and denaturalizes the former, while SMT does the opposite most of the time.

    This is primarily about rhetoric and framing, not economics per se, but it feeds back into economic theory in that framing guides us to conclusions (eg., how many people still subscribe to the quantity theory of money, simply because it "makes so much sense" in the conventional metalist paradigm?)

  2. David I'm impressed by your continued efforts at building bridges and starting high-quality discussions with MMTers, like this one with Eric about his paper. I found many useful nuggets within it, like Eric's explanation that MMT wants the central bank to focus on financial stability and not price stability - this helped me understand some of the MMT-proposed reforms a little better. And his comparison of SMT full-employment thinking with MMT following the post-Keynesian view (that we are typically demand-constrained) was also a useful reminder.

    Where I struggled most was when he dismissed bond vigilantes altogether. I see this often in online discussions, but have never really understood the view. Here I'm hoping that we can find ways to bring in a third perspective - students of fiat with a lineage like that of Paul Krugman (Dornbusch, Mundell). I'm thinking of people like Cullen Roche, Sri Thiruvadanthai (who put out a discussion paper last year on sudden stops). I haven't seen either of them predict a sudden stop for the US, but this is not based on the simple fact that the US borrows in its own currency. With Krugman, I think his mockery of debt scolds is based more on back-of-the-envelope calculations about debt service costs and their sustainability.

    I know that MMT does treat the question, describing it as the "external constraint". I didn't see it discussed in your conversation above. I would have liked to see an acknowledgement of it when Eric made this statement:

    "A fiscal deficit is not unsustainable nor abnormal; deficits are the stylized fact of government finances and are financially sustainable if monetary sovereignty is present."

    It would be nice if we had good policy tools to manage international flows more effectively, so that domestic policy actions didn't spill over, and so that we could actually have the degree of sovereignty that Eric discusses with you above. Apparently such tools are limited due to rules against "currency manipulation" in the current international monetary order. But this means that our thinking and discussion about the external constraint remains muddled, as people issue warnings about national debt, without a good understanding of how an actual crisis would play out, and how close or far we are from having one.

  3. Thanks for publishing the above conversation. I’d like to answer a question put by DA since I’m not happy with the answer given by ET. DA’s question:

    “Let me return to the manner in which the Fed/Treasury/Congress are consolidated (or not) in SMT and why this matters, in your view. In some SMT treatments, Congress decides spending and taxes, which implies a primary deficit. It's up to the Treasury to finance that deficit, with the Fed playing a supporting role (by determining interest rate and issuing reserves for treasury debt). What's wrong with this approach?”

    What’s wrong (at least according to many MMTers) is that the rate of interest should be fixed more or less permanently at zero. Ergo “determining the rate of interest” is out.
    That of course means that demand is adjusted purely by fiscal means (or some combination of fiscal and monetary means, like creating more base money and spending it when stimulus is needed). And that in turn raises the question as to how to deal with irresponsible excess spending by politicians.

    The answer to the latter question is an idea to which Bernanke gave an approving nod, namely to have the central bank decide the SIZE OF the deficit, while purely political decisions (e.g. deciding the % of GDP going to public spending, and how that is split between education, defense, etc) remains with politicians.

  4. Well, this is one hell of a blog post, and I will have to read it a few more times.

    A few quickie observations:

    1. Mark Carney, BOE, and several others at the recent Jackson Hole shindig, said that r* is now set globally. So we must evaluate the actions of any lone central bank as taking place in a globalized capital market (one that is presently glutted with capital and about $350 trillion in scale, including bonds, equities and property). So the Fed can wiggle-waggle short-term rates a little--so what? Long-term rates are set globally.

    2. With $350 trillion in global capital markets, what means a few trillion in Fed QE? So what? That may, or may not, budge long-term rates much. 20 basis points? Big whoop. (The interesting aspect is the QE seems to allow a nation to monetize debt with inflationary effect.)

    3. Recently, BlackRock, Pimco and Ray Dalio have called for some version of MMT. This is amazing---Wall Street is gravitating towards MMT, while academia and central bankers are stuck in the theoretical mud.

    4. This BIS official just said monetary policy is a pop-gun, and you have to go fiscal....

    So...I think money-financed fiscal programs are the way to go.

    MMT wins the day, but egads, MMT proponents need to shed their liberal baggage. MMT is not programmatic. Japan used MMT to sidestep the Great Depression, but sadly used money-financed fiscal programs to build up a military. As a macroeconomic policy, it worked.

  5. Bond vigilantes…
    There are none with any power in a government bond market absent a rogue central banker willing to take down the entire financial system.
    You have to remember that the government banks with the central bank and the central bank is the ultimate clearing house for interbank transactions and payment settlement. Any massive demand for interbank reserves, such as to buy government bonds, would cause the central bank to intervene to defend its interest rate targets (hence the US FED & Treasury coordinate treasury auctions). Likewise, a massive selling frenzy of government bonds would also cause the central bank to intervene to defend its interest rate targets as the demand for exit liquidity would impact its target rate. The only way a bond vigilante could exist is if it is the central bank. But denying the government indirect access to liquidity would also require denying the entire financial system access to funds to clear and settle payments. It would only be a matter of days, if not hours, before the whole financial system collapse do to illiquidity and panic. That’s why you don’t see bond vigilantes in currency sovereign government bond markets.

  6. What is your view on this paper that claims loans create deposits. Thus no hyperinflation from printing money:

    "In the modern economy, most money takes the form of bankdeposits. But how those bank deposits are created is oftenmisunderstood: the principal way is through commercialbanks making loans. Whenever a bank makes a loan, itsimultaneously creates a matching deposit in theborrower’s bank account, thereby creating new money.The reality of how money is created today differs from thedescription found in some economics textbooks:• Rather than banks receiving deposits when householdssave and then lending them out, bank lending createsdeposits.• In normal times, the central bank does not fix the amountof money in circulation, nor is central bank money‘multiplied up’ into more loans and deposits"

  7. Good discussion. ET does an effective job of explaining MMT and DA does an effective job of pushing back on certain logical frailties (in my view) of the MMT framework (i.e. the founder’s “paradigm”).

    I’ve long thought that MMT’s most significant conceptual weakness is reflected in its obsession with this notion of “the government spends first”. In my view, this is symptomatic of several major logical category errors in its “consolidation view”.

    The first error has to do with MMT’s failure to clarify whether the logical framing of the consolidation view is that of a view of a consolidated financial result of existing institutional arrangements or a view of a counterfactual institutional arrangement for monetary operations that does not currently exist in reality. And that problem is galvanized in this: MMT in effect describes the second version but seems to claim to describe the first. I.e. it purports to have an insight into a consolidated view of existing operations that in fact is a consolidated view of counterfactual operations. I think recognition of this problem is partially embedded in DA’s push back. At the same time, it’s not clear to me that SMT deals effectively with the kinds of accounting and operational issues that MMT attempts to do but fails to do because of some logical mangling of standard accounting.

    The standard MMT consolidated view (do we need SMMT as another acronym?) employs descriptive language that from an institutional perspective seems very much to reflect the idea of a central bank that has assumed Treasury fiscal powers into its arsenal of operational capabilities. In such an institutional context, it is easy (and somewhat fun) to visualize the ideas of “easing” and “tightening” as standard central banking functions that have been expanded to include the notion of spending (fiscal) money into existence and withdrawing it by imposing taxes and bonds. Such a consolidated entity when released from other constraints can do anything it wants by way of money creation or destruction at the margin and in any temporal order (hence the “spends first” option). And it can do so in the mode of either standard central banking operations or newly acquired powers of fiscal spending and taxing/borrowing operations. But such a consolidated institution does not exist in fact.

    Accordingly, I think there are two logical category errors inherent in the MMT “consolidation view”

    … continued …

  8. … The first category error results from the fact that there is an enormous difference between a view of the consolidated financial results of existing separate institutional operations and a view of an imagined consolidated institution that operates as I’ve described above but which simply does not currently exist in fact – a counterfactual institution – but whose imagined operation effectively produces a similarly framed consolidated financial effect. Conversely, the consolidated financial result view of existing factual operations is a useful depiction of reality that provides insight into the way in which separate institutions interact in their consolidated financial effect. The consolidated institution view on the other hand is a very interesting depiction of an alternative institutional architecture that does somewhat the same thing – with more flexibility - but that simply does not exist in reality. The logical framing problem for MMT is that it depicts the counterfactual institution view while claiming to depict unusual insight into the combined effect of the existing institutional arrangement.

    The second category error is embedded in the fact that even this imagined consolidated central bank concept with fiscal easing and tightening powers is one that conceptually still operates within a framework in which central bank reserves are the fuel of private sector banking capitalism. It is elemental to MMT that it depends on framing its perception of monetary operations in a way that is very bank reserve centric. And there is therefore a larger framing context in which the operation of the entire monetary system is oriented toward the idea of private sector banking capitalism. After all, the ultimate purpose of bank reserves is to facilitate competition for assets and liabilities amongst a set of commercial banks – i.e. private sector banking capitalism. The unintended result of this implicit framing is that there is a much larger context for this famous “spends first” mantra of MMT. This is logically problematic for MMT in that its methodology inherently cherry picks this characteristic for the purpose of assigning unique characteristics to the state in the sense of its “spend first” mantra – characteristics which in fact are not unique at all.

    Because in fact we ALL “spend first” in this context, and we ALL spend by crediting bank accounts. This is evident because we all provide instructions to our banks to credit bank accounts (e.g. cheques, pre-authorized debits). And we ALL pay the piper in terms of the debits that must offset our instructed credits. I see no difference between the state spending function and my own in this categorical context – the context in which the state’s fiscal operations are fully integrated into a private sector banking configuration. The state instructs credits and in the existing monetary architecture pays with debits to a particular Treasury account at the central bank (or perhaps in some cases for expediency and system reserve management coordination its complementary accounts at the commercial banks) In this larger sense, the way in which a check in its first contact reflects spending with corresponding reserve credit and then clears back to the payer’s bank for corresponding expense through reserve debit is indicative of the kind of temporal causality that MMT seems to covet as a matter of presumed unique logic in the case of government operations. But it’s an empty observation in the context of the existing banking architecture - i.e. in the context of something other than an imagined counterfactual consolidated institution.

  9. David, very good piece. You're open-minded and fair, but also firm where push back is helpful and warranted.

    JKH touches on the primary problems here. This is really a battle over the power of monpol vs fispol. I tend to think that, in a low inflation environment, MMT has the upper hand here arguing for the power of fispol and against the power of monpol. Monpol just isn't a very powerful tool when inflation is low. But this seems to change when inflation is high so beware of what you wish for.

    Tymoigne is a great representative of the MMT community. Fair and objective. But as JKH said also, he loses me when he transitions into the "spend first" narrative. There are many contradictions within the MMT narratives that confuse their "general case" with the "specific case". In my opinion it is unhelpful to use the general case in such a way that gives the appearance of being the specific case. Eg, we have a very specific institutional arrangement in the USA in which private banks compete to make most of the money in the economy. That system is specifically supported and cleared by a Central Bank that essentially leverages state authorities to help support a private system. When you consolidate the Fed into the Tsy and treat govt taxation as "destroying" money then you in essence treat the private banking system as a public entity when in fact the entire purpose of the Fed and its independent status is to support a specifically private segment of the economy. The MMT description not only misconstrues this, but actually undermines a recurring theme in many MMTers writing about bank nationalization and financialization. The very existence of the Central Bank exposes a potential weakness and/or flaw in a monetary system that is very much built around private money and creates a govt dependence on banks.

    Tymoigne further loses me here:

    "In financial terms, taxes do not increase the capacity of the government to spend"

    Taxes reflect income. Taxes reflect policy space. When a govt has high domestic income and the ability to tax it has higher credibility and more ability to leverage its spending. This is a basic tenet of endogenous money and credit. Warren Buffett has a high credit line because he has high income and assets to support that credit. The same basic tenet is true of any govt. The govt relies on funding its credit in much the same way that the rest of us do even though the state obviously doesn't declare BK on itself. So there seems to be a very basic contradiction in the MMT narrative that taxes don't fund spending, but inflation constrains spending. These ideas can't both be true since domestic income constrains the spending limits....

    Of course, these sorts of institutional and descriptive elements are high level nerdery and probably of little interest to most. But I think it's worth shedding some light on these issues because MMTers have a tendency to claim higher ground on the basis of describing "operational reality" when in fact they confuse many of the operational facts.

    I think where this gets most interesting is in the application and context of monpol vs fispol in certain environments. MMT (and Keynesians in general) has gained an upper hand in recent decades because inflation has been low and MonPol hasn't appeared to work well. That won't always be the case though which is why I've always maintained the view that we shouldn't be so quick to dismiss MonPol. Sure, it is weak in this environment, but should we encounter a different environment I'd rather have my arsenal filled with tools that are also blunt instruments.

  10. “In a consolidated framework, the federal government can only implement spending by creating money, this is not abnormal and it is not inflationary by itself. There is no other way to find the necessary dollars to spend."

    This sort of explanation hinges on the fact that the Treasury account at the Fed is typically not labelled semantically in the same way as bank reserves or bank deposit liabilities - i.e. as comparable "money".

    But once again, this is a category error relative to the organization of a group of substantially similar characteristic deposit accounts in the actual monetary architecture as it exists now.

    The Treasury account at the Fed is functionally the same thing as a bank reserve account. In that view, it becomes obvious that the so-called money creation process is no different than it is for private sector spending. We all create money when we spend in that sense.

    If my spending would cause a flow of reserves from JP Morgan to Citibank, by writing a check for example, the first contact is a credit to Citibank reserves, which results in the temporary “creation of money” in the same way as would a government check written to a Citibank customer.

    The other side of that transaction in both cases is a neutralizing debit to the source account – a reserve account in my case and the Treasury Fed account in that case.

    While the spending side in both cases creates a temporary gross increase in "money", the netting of the clearing system produces a net zero effect in both cases. And the semantics of account designation as between Treasury and the banks are seen to be irrelevant to the substance of the similarity of these cases - in both gross and net terms.

    Again, the existing institutional arrangement simply doesn’t allow for the option of not clearing that Treasury transaction back to the Treasury account at the Fed.

    The essence of the thing is that Treasury deposit accounts DO exist in the actual monetary architecture. In a specified counterfactual architecture with true institutional consolidation, there would be no need for such a deposit account at the central bank. In that case, Treasury spending and taxing would resemble an extension of CB easing or tightening to fiscal operations, as is implicit in the MMT style of explaining these things. But in fact that’s not the actual world that we live in now.

  11. After second reading of an excellent discussion:

    I come down on the side of DA in terms of the incisiveness of the questions. I still see frailties in the MMT experiment. To be blunt, at this stage and after years of watching, I find MMT to be at least mildly Orwellian in the language and paradigmatic thinking it employs to develop its “theory”.


    “MMT just pushes consolidation to its logical conclusions and shows that institutional details do back those conclusions.”

    The first part is vague (as always) to the point of not being meaningful The second part is simply false. The consolidation of balance sheets does not change the underlying operational steps – and MMT as far as I know has NEVER specified changes in those steps as a different conceptual component in its “consolidation view”. If it does mean to suggest changes, the exposition is weaker than I assumed. If it doesn’t, that’s the end of it.

    Consolidation of balance sheets is a standard financial analysis technique used in ALL sectors. It doesn’t mean that underlying deconsolidated operations that in fact have led to the consolidated balance sheet result have been altered. That instead would require the explicit assumption of an alternative sequence of operations – a different sequence that would also correspond to the consolidated balance sheet result. But you can’t just make up changes in underlying facts merely because you consolidate end of period of financial statements.

    What does “view” really mean as in “consolidated view”? It’s meaningless, unless you define whether “view” means the consolidation of balance sheets or the modification of underlying transaction procedures that lead to that consolidated balance sheet result, or something in between. Is the “view” a view of facts on the ground, or it is a view of something other than “facts” (as in a counterfactual construction of underlying operations). MMT has never explained this clearly. And I suspect this is because it wants a wide and friendly interpretative zone that permits a “view” that is suitable to the “spends first” mantra. It’s all vague, it’s meaningless, and it’s Orwellian. There’s just no substance there in the word “view”.

    And as I said above, I’ve never found a reference to actual operational changes in the “view”.

  12. 2.

    “The government imposes a tax liability, spends by issuing the currency necessary to pay the tax liability, then taxes and issues bonds.”

    That’s not how it works and no amount of balance sheet consolidation alone can change it. The world actually works as follows, and this should be reflected in any “view” of how it works – if that view is based on facts:

    Treasury must attract money that has been created by the private sector banking system into its account at the Fed – before it can spend. The notion that the Fed provides the reserves to enable the banks to process such a transaction at the inter-institutional level is a diversion from the fact of where the private sector money comes from originally. Moreover, the Fed provides reserves to the banks to process ALL transactions – not just the movement of money from private sector bank accounts to the Treasury account. Indeed, the Treasury account is functionally equivalent to a bank reserve account. The intra-day gathering of money in the Treasury account is no different in procedure than the intra-day gathering of money in the JP Morgan account. Either can happen, for various reasons, including bond issues in either case. In both cases, the Fed can respond by providing intra-day overdraft limits to the paying banks or by easing the markets with OMO. In both cases, the recipients take further action to clear the account more or less by the end of the day – JP Morgan by undertaking market transactions (e.g. buying money market securities) and Treasury by undertaking off-market transfers to its accounts with the commercial banks. In BOTH cases, the motivation is the same. To restore balance in their respective accounts and therefore to restore balance in the system and assist in doing so with the smooth operation of the system in aggregate. And the system is designed to encourage BOTH types of transfers through economic motivation – JP Morgan gets a better rate in the money market and Treasury gets a better rate in its commercial accounts. The system is designed this way.

    So there’s a lot of symmetry in the operation of the system in this way. The architecture is designed deliberately to position Treasury alongside the banks with minimal disruption or at least minimally unique disruption to the reserve management effectiveness of the system as a whole. Positioning Treasury as some sort of asymmetric participant in the bank clearing system just doesn’t correspond to the facts or the intention of the architecture.

    In short, the commercial banks create the money that is required for the private sector to pay taxes or buy bonds and the central bank just facilitates the corresponding inter-institutional payment. Commercial banks create most of the money in the system, and that’s the type of money that is the sine qua non for private sector tax payments and bond purchases.


  13. 3.

    “Taxes and issuance of treasuries drain reserves and so raise the overnight rate. Hence, on a daily basis, a fiscal surplus raises the overnight rate and a fiscal deficit lowers it. There has been significant Treasury-Fed coordination to smooth the impact of taxes (and treasury spending) on the money market.”

    A relatively minor point, but that’s just not correct as a rule in the current IOR regime. MMT in developing its theory failed to foresee that an IOR regime would change this. Another category error.

    "In financial terms, taxes do not increase the capacity of the government to spend"

    This is wrong. It mistakes the effect at the margin as if it were the effect in total. If the government has fiscal room because it is currently on the good side of the inflation constraint, then it will have even more room if it increases taxes. That’s not an entirely unreasonable step within an iterative targeted tax strategy. This is also the reason why MMT is wrong to condemn management accounting as in “what funds what”. Such accounting doesn’t necessarily preclude the existence of a larger strategy for a spending path that is aimed consistently toward the inflation constraint boundary – although we know that governments unfortunately in the real world would typically not necessarily be perfectly intelligent in their strategic planning. Still, management accounting is a constructive thing if employed intelligently. Conversely, to much consolidation effort in this dimension is also another logical category error by MMT.

    DA – “As for money, taxes and bonds not being alternative "funding" sources, I worry that this semantics. You can call X a "funding" source or not -- it's just a label.”

    It’s worse than just semantics, as I think I alluded to above.

    “I think all of these positions are defensible without the statement "taxes/bonds don't finance the government"

    Totally agree, at least with respect to defensibility of what is correct. In general, there’s obviously a lot of merit in not confusing real resource constraints with financial constraints.


  15. "If the government has fiscal room because it is currently on the good side of the inflation constraint, then it will have even more room if it increases taxes."

    I should have added that this assumes that an increase in taxes reduces the demand for real resources, which is pretty reasonable.