Most Recent Stories

JKH on “Modern Money Theory 101: A Reply to Critics”



This is the paper by Éric Tymoigne and L. Randall Wray:


The paper as suggested by the title is a reply to various critics who have raised some issues around Modern Money Theory (Modern Monetary Theory, MMT) in recent years.

MMT as a school of economic thought might be viewed in simple model terms as an analytic engine that processes the input of critical real world observation into the output of preferred policy orientation. The reason for the enthusiasm of its proponents and the active interest from its critics may have to do with the realistic nature of this observation process, which is in stark contrast to the inward looking analytic methods of mainstream economics. This critical input includes observed facts about central bank monetary operations and Treasury fiscal operations. MMT has taken the time to study how these things actually work at the transactional level.

Central to the design of the MMT analytic system is what the authors describe as the “consolidation hypothesis”. This is interpreted in more detail further below. But in brief, it characterizes a sovereign monetary model as one that necessarily integrates monetary and fiscal operations as a matter of “logic”, notwithstanding their operational and institutional separation in fact. Some critics are concerned about this expositional method.

It is fair to say that MMT does not have a monopoly claim on these facts of monetary and fiscal operations, and that there are competing ideas for how they best fit into a framework for explaining monetary and fiscal operations and monetary economics.

The “consolidation hypothesis” is not the only area of criticism that is reviewed in the paper, but it is an important one for MMT, and the only one that this post/essay will be concerned with.

It is important to keep this criticism in context. There is considerable agreement from some critics regarding the importance of the factual input as well as the benefit of exploring options in what MMT characterizes as “policy space”. In terms of the simple analytic model mentioned above, the criticism relating to the “consolidated hypothesis” isn’t directed primarily at either the input or the output. It’s the analytic engine and presentation that lies between. It’s the way the message is being presented en route to the feasible policy space. It’s a matter of exposition. This may seem like a minor point in context. But it’s only minor if you don’t care about the subject matter. People do have views on this.

The authors cite some of the critics of the “consolidation hypothesis”:

“Palley… has been joined by Lavoie, Rochon, Fiebiger and others in their rejection of the consolidation hypothesis (Palley 2012; JKH 2012a, 2012b; Lavoie 2013; Fiebiger 2012a, 2012b; Rochon and Vernengo 2003; Gnos and Rochon 2002). They note that this hypothesis does not describe the current institutional framework of developed countries, and claim it pushes MMT into unnecessary strong logical claims.” (TW Page 4)

The JKH reference is with respect to this post/essay:


The linked “Contingent Institutional Approach” addresses some issues relating to consolidated interpretations of monetary and fiscal operations. My post/essay here focuses on this same area once again, with reference to the Tymoigne-Wray paper and MMT more broadly. In that sense it is a continuation of the “Contingent Institutional Approach”. Among the other critics listed above, I am most familiar with the works of Lavoie and Fiebiger, and am generally sympathetic to their interpretations and criticisms of MMT as it relates to the issue of consolidation. The works relevant to their own criticisms are cited in the Tymoigne-Wray paper. I have referred to them in a general way in this essay, but have not repeated the explicit references that can be found in the Tymoigne-Wray (TW) paper.

Critical License

“We find it remarkable that MMT’s critics waste so much ink trying to criticize a simplification that is commonly made…” (TW Page 14)

The “consolidation hypothesis” is not really a “commonly made” simplification. It is a particular treatment of the issue of consolidation that is unique to MMT.

Some critics treat the issue of consolidation in quite a different way. For example, Godley and Lavoie spend 500 pages evolving combinations and permutations of monetary operation design in their book “Monetary Economics”, including whole or partial consolidating and deconsolidating treatments along the way, for government and other sectors. They start it out with a model they call SIM, which ironically is a highly stylized but carefully specified consolidation of Treasury and Central Bank operations. That is followed up by a carefully crafted sequence of explicit presentations and models. But they never suggest that the world of government or that around it operates as if captured by the logic of a permanent model SIM – far from it. The expositional destination is the virtual opposite to that of MMT’s consolidation arc as meant by its consolidation hypothesis. G&L define different institutional arrangements in sequence, according to balance sheet structures and operational assumptions. Indeed, they experiment with a variety of constraint combinations for exposition and explanation purposes. This approach achieves detailed clarity in the evolutionary exposition of monetary logic. They never gloss over appropriately described unconsolidated detail in these models. Consolidation is never positioned as an ultimate force of logic that vanquishes the accuracy and usefulness of detail. The distinction will be sharpened as we review the nature of MMT’s “consolidation hypothesis”.

In addition, Brett Fiebiger has delineated in considerable detail how the MMT consolidation paradigm portrays a Treasury function that operates as a bank, where this is not so in fact. Finally, the “Contingent Institutional Approach” covers a menu of choices for institutional design. The existing monetary system is contrasted with the availability of a “counterfactual” system or systems that do not exist currently and/or do not operate normally in the same intended fashion.

We should note that Tymoigne-Wray use fifty pages of “ink” in their response. From our perspective, this ink spilling is welcome analysis. There is some good explanatory material contained in those fifty pages, albeit with the eventual reversion of the overall story to the method that is the subject of criticism. That said, the authors do review actual deconsolidated operations in good detail as part of their process of explanation.

Those interested in this subject of monetary operations and economics are all students, and even aspiring teachers, in the sense of wanting to understand and explain it. So – the method of exposition is important in how the modern monetary system is described and taught. It warrants close examination.


Chartalism, in broad strokes, is the idea that government creates the demand for the currency it issues through the imposition of taxes. While perhaps debatable as the primal force of currency acceptance, it is a model of some logical elegance. The authors and MMT believe in it, but are open to persuasion from alternative theories:

“Critics conflate the logical argument that taxes are sufficient by jumping to the conclusion that MMT believes there can be no other possibility. In truth, MMT is agnostic as it waits for a logical argument or historical evidence in support of the belief of critics that there is an alternative to taxes (and other obligations). We have not seen any plausible alternative.” (TW Page 10)

An interesting aspect of MMT is how it expands this strategic idea of currency acceptance to a connected idea at the operational level:

“The injection of government currency (through expenditures or advances) into the other sectors must occur before the destruction of the government currency (through tax enforcement and repayment of advances). In an economic system in which a sovereign government operates through its own monetary system, spending (or lending) must occur before taxing. In addition, taxes are not a funding source in that logic. (TW Page 5)

We will have more to say about this operational perspective later in this essay. It is a theme repeated often by MMT. It is congruent with Chartalism. It recognizes the fact that the central bank creates and issues commercial bank reserve balances in order to facilitate payments among the institutions that hold those balances. And while the central bank makes reserves available so that the private sector can make tax and bond payments to the government, the same thing is true for all private sector payments that result in settlements between competing commercial banks. Given this general fact, and notwithstanding the specific case of government transactions, it is not immediately obvious how this should be a crucial element for a “consolidation hypothesis” that focuses on Treasury and the Central Bank. The fact that a US taxpayer’s bank or a bond buyer’s bank pays the Federal Reserve with reserve balances doesn’t seem uniquely pertinent for what is “logical” in describing the government nexus for monetary operations. Thus, while the central idea of Chartalism is important for MMT, and understandable in that sense, we don’t see that its natural extension to this operational level should be an issue of major concern. This lesser perspective will be revisited later in this essay. We should emphasize however that Tymoigne-Wray view this additional operational level interpretation as important and inherent to the “consolidated hypothesis”:

“A third logical conclusion is that taxing, “borrowing”, and monetary creation are not mutually exclusive methods of funding a government. Indeed, all occur but at different stages of the circuit.  (TW Page 6)

We will also argue later in this essay that “the circuit” as applied to the consolidation hypothesis is a model that is livelier in the case of a specified counterfactual institution in which greater funding flexibility by design encourages a more natural interpretation of funding options that can be described pragmatically in the context of such circuit stages.

An Overview of MMT Consolidation Logic

Once again, consider a crude model for an economic “school of thought”, consisting of three process components: input, analysis, and output.

The input consists of the raw material that goes into the analysis. Here MMT warrants high marks. Unlike mainstream economics, MMT considers an understanding of real world monetary and fiscal operations to be a priority input. The importance of understanding the banking system is preeminent. MMT shares this trait with a wider group within post Keynesians economics at least.

The output consists mostly of policy conclusions, taking the form of specific recommendations and/or available options. MMT refers to this as “policy space”. This seems like a good way of thinking about it.

That leaves the middle analysis component. This connects the observation of real world monetary operations to the exploration of policy space. That analytic exposition is what we are mostly concerned with here. The general criticism pertains to MMT’s boilerplate framework or “paradigm” for the explanation of how the monetary system works – the teaching method, so to speak. Thus, the criticism is not directly aimed at either the input or the output of MMT (according to our simple model above). It is the middle part that is at the heart of the question.

“The essential point here is that the consolidation hypothesis is a theoretical simplification that makes sense once one understands the logic of the interrelations between the central bank and the Treasury, and between the government and nongovernment.” (TW Page 11)

“The consolidation hypothesis does not aim at describing current institutional arrangements, rather, it is a theoretical simplification to get to the bottom of the causalities at play in the current monetary system.” (TW Page 13)

Thus, with the consolidation hypothesis, MMT claims not to be describing the existing system, while invoking the underlying logic of that system, but not referring to anything other than the existing system. This seems conceptually gymnastic.

A subsequent statement refers to the combination of two existing separate institutions into something referred to as one “entity”:

“The logic of the argument is about a government sector that combines the central bank and the Treasury into one entity that issues currency … the consolidation hypothesis and ensuing conclusions are not descriptive, they are logical conclusions.” (TW Page 15)

If the consolidation thesis defines the logic of the system, but does not describe it, it would seem to follow that a description of that same system (i.e. as opposed to its logical characterization) must reflect something that is inherently at least less logical if not illogical. The MMT consolidation paradigm rests on the idea that government sector consolidation is ideally logical – to the point where the logic of consolidation somehow overrides the descriptive usefulness of actual non-consolidated operations. Critics question this duality of perception, as the notion that such a “combined entity” exists in logic but not in factual operations leads to descriptive or logical ambiguity.

One can argue there are better ways of organizing monetary and fiscal operations, but it’s not clear that the existing organization is illogical. There are at least degrees of argument involved in such a perspective, depending on the specific design feature being analyzed.

For example, while bond borrowing can be circumvented in a counterfactually designed monetary system, it isn’t necessarily illogical in the context of existing design. The logic of government bond issuance is a special, separate topic, extending beyond the observation (e.g. by MMT) that bank reserves are arguably as good as bonds from a government finance liquidity perspective. The full risk analysis is more complex than that. (I hope to write something on that sometime in 2014.) On the other hand, the argument is probably stronger in the case of a debt ceiling rule that is a constraint notwithstanding separate budgetary authorizations that by inference should allow the necessary deficit financing.

We suggest that the category separation should not be between “descriptive” and “logical”, but rather between factual and counterfactual. Factual observations and counterfactual constructions both admit description. Most importantly for analysis, they both admit their own logic. Instead of assuming away or ignoring operational details as standardised simplification, one might instead subject them to separate debate as issues that arise from an examination of monetary operations in the factual case.

It is inherent in the nature of the criticism here that MMT should be able to pursue its objectives in “policy space” without fashioning a consolidation thesis that purports to be exclusively insightful on the logic of monetary operations. It is possible to segment analysis between factual and counterfactual designs, both of which can be both descriptive and logical, and both of which allow further investigation of the policy space.

In response to criticism, Tymoigne-Wray present copious illustrations of pro forma T-account consolidations and deconsolidations. This is good by way of explanation, but it misses the core nature of the criticism of the “consolidated hypothesis”. The authors may believe that critics aren’t familiar with consolidated financial statements. Or perhaps they believe that critics believe that MMT doesn’t understand accounting for central bank and Treasury operations. Neither of these inferences is true. Both MMT and its critics know monetary operations and their accounting (and consolidated versions of it).

In order to address the concern of critics, the paper goes out of its way to depart from the normal MMT “consolidation hypothesis” by delineating existing Treasury and central bank operations in some detail. But again, in responding to critics, MMT underestimates their understanding of consolidation as financial accounting or as an institutional design concept, and the difference between those two ideas.

So if both MMT and critics alike understand T-account construction and deconstruction and consolidation, what is the problem?

I would suggest that the logical underpinnings of the consolidation hypothesis are somewhat fragile. There is no persuasive case that points to a requirement to view the consolidation of Treasury and the Central Bank as ultimately revealing of an exclusive logic to monetary and fiscal operations. The logic of the monetary system is what the design of the monetary system determines it to be. And vice versa. For example, the separation of Treasury from the Central Bank is quite logical, if the government’s policy objective for the financial system is to restrict most money supply creation to the activities of the commercial banking system. The design of the system fits the logic of the objective. One may question this objective, but that is not to say that the design of such a system is illogical. The logic of the system is what you want it to be, but in order to make that the logic, you must design the system to fit the logic. This is the inherent theme of the “Contingent Institutional Approach”. Put another way, the assertion by MMT that the “consolidation approach” reveals the underlying logic of the system really amounts to an exogenous injection of MMT’s own preferred logic into its view of the monetary system. It is an option. There are other ways of looking at it, and perhaps more accurate ways of characterizing the logic of the existing system as opposed to a counterfactual system. We will develop this idea further below.

It sometimes seems that MMT’s “consolidation hypothesis” facilitates the quickest route to a desired policy space. In short, the hypothesis reflects the positioning of a logic that leads to the most flexibility in the explanation of policy space. That is a tactical advantage in the case of an assumed policy preference.

We will go one step further and suggest that MMT’s “consolidation hypothesis” is itself falsifiable as the most revealing logic of existing monetary operations – it actually does not represent the logic of existing monetary operations. The detailed case for this will be presented in the next and subsequent sections of this paper.

Monetary Operations and the MMT Consolidation Hypothesis

We can now drill deeper into the operational detail of the MMT “consolidation hypothesis”. We’ll use the specific case and related terminology of the US Treasury and Federal Reserve. The focus will be on the normal operations of the Federal Reserve – i.e. prior to the financial crisis and without the reserve management complications associated with its various programs of quantitative easing (QE). The case of QE can be handled separately.

The central bank offers deposit account facilities to both the commercial banks and the Treasury department. The MMT consolidated hypothesis deals with this configuration by consolidating the cash flow activity of the central bank in its principal management operations together with Treasury in its principal management operations. In other words, it fuses the cash flow activity of the central bank in its own right with that of Treasury as a depositing “customer” of the central bank.

The design logic of the existing system positions Treasury and the commercial banks pari passu in their operational interface with the central bank balance sheet. They all have deposit accounts on the liability side of that balance sheet. But the MMT consolidation hypothesis treats Treasury in a very special way that separates it from the other commercial bank customers of the central bank. We think this special treatment amounts to a logical inconsistency in the construction of the consolidated argument. We will now unpack the analysis underlying this contention.

When the Fed conducts monetary policy in order to control the short term interest rate (in pre-2008 mode), it needs to watch the behavior of the banking system in responding to the distribution of reserve balances. The Fed supplies excess reserves in finely tuned amounts and it changes that supply in accordance with the behavior of commercial banks who are operating in a reserve supply zone where the reserve demand curve is extremely inelastic. The fed funds rate is very sensitive to changes in system reserve balances and/or their distribution within this inelastic demand zone. Hence the Fed pays careful attention to bank demand behavior in that zone. Moreover, the degree to which the Fed may have to adjust supply depends very much on the distribution of reserves across different banks, as aberrations in distribution may cause banks to engage in money market transactions that can put pressure on the funds rate in either direction.

But the banks aren’t the only ones that hold money balances with the central bank. Treasury’s main operating account – the Treasury General Account (TGA) – is with the central bank. And so the Fed watches more than just the distribution of commercial bank reserves – it watches the distribution of the totality of bank reserves plus TGA balances held at the Fed. If TGA balances change, the total distribution changes. If JP Morgan balances change, the total distribution changes. If Citibank balances change, the total distribution changes. Such shifting distributions happen regularly on an intra-day and overnight basis.

Here is the critical point. In the example above, all three institutions – Treasury, JP Morgan, and Citibank – operate using the same basic approach to cash management discipline, which is that of a bank customer managing a cash position within the normal objectives of institutional cash management – which is the matching of cash outflows with cash inflows. The underlying motivations for the discipline may be different, but the operational objective is not. The banks operate according to the pure economics of cash management, given the opportunity costs associated with sloppy cash position outcomes. Treasury operates according to a government-imposed constraint of cash management discipline. Its objectives include the same type of avoidance of unbalanced cash position outcomes. In both cases, it is the expectation of cash management discipline that assists the central bank as a separate institution in achieving its own objectives of reliable interest rate targeting according to the policy setting for the Fed funds target rate. In both cases, in this specific sense, the banks and Treasury are rational participants in the clearing system run by the Fed. The effect on the participants is the same in the sense of this basic discipline. For example, if JP Morgan or Citibank or any of their customers have received payments from Treasury, those banks will take steps to deploy their excess balances and Treasury will take steps to correct its shortfall in balances, other things equal, each according to their own disciplines. Thus, it is a function of existing institutional arrangements that Treasury ranks pari passu with the commercial banks with respect to its monetary interface with the central bank.

But there is more to this than just the interaction of TGA balances and bank reserves.

Suppose JP Morgan makes a payment that reduces its reserve balance. Whether that payment is made to Treasury or Citibank becomes irrelevant to the question of how the totality of reserve and TGA balances are distributed. In either case, it is possible that JP Morgan can put upward pressure on market rates including the Fed funds rate as a result of its follow up efforts to restore its reserve position. And that may happen whether it is Citibank that is temporarily long the corresponding reserves or the TGA that is long balances that have been temporarily withdrawn from the reserve pool. The distribution issue regarding the totality of TGA and reserve balances is the same in either case. If the payment is made to Citibank, the fact that bank reserves sufficient to correct that imbalance may still exist somewhere in the system (i.e. Citibank in this case) doesn’t necessarily mean there won’t be upward pressure on the funds rate. That depends on the precise pattern of bank cash management behaviors as they act collectively in response to the momentary distribution of balances at that point in time. The money market is very efficient, but not perfectly efficient in correcting aberrant balance distributions in this way. Indeed, if perfect efficiency were the case, the Fed would never have to change the level of excess reserves in the system in order to control the fed funds rate. But we know that is not the case. So there are effects on the funds rate that emanate from distribution within the banking subset, just as there are in the distribution between the banks and Treasury. The Federal Reserve may intervene in any of these cases in order to control the funds rate.  So it may intervene in a similar way in this example above, whether the initial JP Morgan payment is made to Treasury or to Citibank. The Fed may respond to an aberrant distribution of the totality of reserves and TGA balances in a similar way in either case. In this sense, the response may be similar whether or not Treasury is involved.

And so the idea that the central bank responds uniquely in combination with Treasury as a sort of joint efflux or reflux effect in order to correct aberrant distributions of such total balances is incorrect. The Fed can act due to an effect from net JP Morgan transactions in the same way it can act due to an effect from net Treasury transactions, in order to respond to aberrant distributions of the totality of bank reserves and TGA balances. If a particular institution is short or long balances, relative to its cash management discipline, it makes little difference who is that is temporarily long or short balances on the other side. Whoever it is must “return serve” relative to deployment or sourcing of balances in order to be rational in its cash management discipline. And that holds for Treasury and the commercial banks.

The issue for the central bank is not just how the quantity of bank reserves is affected, but how the distribution of bank reserves plus Treasury balances are affected, given pari passu operating arrangements for Treasury and the banks through Fed accounts. The central bank expects rational cash management behaviors from all participants, which includes Treasury and the commercial banks. Given existing institutional arrangements, it is a category error to think they imply a unique “consolidation hypothesis” and that Treasury and the central bank must be conjoined to reveal the “logic” of the existing system. The existing system operates logically as it is. This is not a question of what is logical and what is not. It is a question of deliberate design – and a system that functions with its own logic according to that design.

Consolidation by Aphorism

“Once the consolidated balance sheet is understood, it makes sense to say that “sovereign government neither has nor does not have money” (Wray 2011).”

It is awkward to make a point about a mere aphorism. But this one is iconic within MMT. It seems like a harmless enough expression. It is clever and pithy and memorable. Yet even while it serves as a symbol of MMT’s insight into monetary operations, it doesn’t really say what MMT would like it to mean. And while this no big deal, it becomes interesting to examine because of this property. So no harm done in having a closer look.

The intended effect seems evident. Financial accounting consolidation eliminates a common entry (the Treasury General Account (TGA) balance held at the Federal Reserve) between two government units. A neophyte student of accounting and finance would know that this item can disappear on financial statement consolidation, but would be delusional to believe that this means disappearance in operational fact. Accounting consolidation is not operational consolidation, even when its description is displaced by logical claims of the “consolidation hypothesis”.

The more substantial irony is that the expression is logically challenged. According to basic rules of logic, it cannot be true in the case of either specified factual or counterfactual operations. In the case of actual institutional operations, and notwithstanding the logical overlay of the MMT consolidated interpretation, a specific department of government still does have money at the operating level – Treasury through its TGA balances – and Treasury is part of government. And in the case of counterfactual operations where Treasury and the central bank are indeed a single consolidated institution and the TGA balance disappears from operations, no government department has money (at least in normal cumulative deficit modes), so the characterization is false there as well (see section below on SOMA/TGA fusion, where the TGA balance disappears by intended construction). At this point, I’m not sure whether to say the aphorism fails to display the intended logic of the “consolidation hypothesis”, or succeeds at conveying the nature of its challenged logic.

Financial consolidation does not mean actual operational consolidation, unless it is accompanied by a specified counterfactual consolidation of actual Treasury and central bank operations. The consolidation of the financial statements of separate institutions does not negate the fact that the government still holds money as an asset and issues money as a liability – through two separate operating units. On the other hand, an explicit operating consolidation that whose financial statement representation looks similar to financial statement consolidation in the deconsolidated case means that the government issues money (not TGA balances but bank reserves) as a liability but it does not normally hold any form of money as an asset.

(As noted briefly above, a conceivable exception to this is the aberrant case of a cumulative government budget surplus, or at least a cumulative deficit that is smaller than the required central bank level of bank reserves and banknotes. In that case, a consolidated institution might begin to hold deposits with the commercial banking system (among other possible claims on the private sector) as an outlet for what would be forced asset acquisition in essence. But these sorts of profiles are extraordinary, given the typical cumulative deficit position of government.)

More generally, in a functioning logical paradigm, “has or doesn’t have” is a binary partitioning of all possible events and non-events. It is therefore impossible that neither can occur. This is clear when factual and counterfactual cases are specified. It is murky when they are not.

Saying that the government neither has nor doesn’t have money is like saying that the world neither has nor doesn’t have financial assets. In fact, it has them, but financial consolidation can set them aside in appearance, revealing underlying real activity while submerging the financial claims that constitute the complex network of underlying ownership and indebtedness. Thus, the method of the consolidation approach is ironically analogous in the bigger picture to the case of ignoring banking and other financial system conduits when attempting to understand the economy as a whole. Somebody has already tried this – neoclassical economics.

We also note that the pure idea of financial consolidation in the case of government has nothing to do with its status as the monopoly issuer of its own types of liabilities. Corporate conglomerates including banks consolidate financial statements without rejecting the nature of separate operations that have separate purposes in fact. Those entities are also monopoly issuers of their own liabilities. Consolidation as “logic” or otherwise is not an analytic technique that is generically unique to government.

In summary, the noted aphoristic premise is unfortunately:

a)     Not true in the case of deconsolidated operations

b)     Not true in the case of the financial consolidation of deconsolidated operations

c)      Not true in the case of specified consolidated operations and/or their financial representation

Aphorisms are sometimes a way of dumbing down the subject matter. In the case of this particular subject matter, they can be words that suggest the feasibility of a predetermined policy orientation. For example, the government can spend its own currency into existence by “crediting bank accounts” in a counterfactual world where by contrast the normal constraint of bond issuance is lifted. But in the world that already exists, this happens with no greater logic than does household spending from a credit card or business spending from a standby line of credit, in the sense that those things act to expand the creation of broad money. The constraint structure is different, but the crediting of bank accounts is pretty similar. Indeed, it happens to an even lesser degree for Treasury with respect to the expansion of bank reserves in the existing system, according to the intention of the constraint structure. The actual rules for Treasury are even more onerous than for the other two. All sectors spend with bank accounts credited on the counterparty side. And in any event, it is a commercial bank and not the government that in the first instance credits bank accounts with the proceeds of government expenditures. We can understand the option of lifting such a constraint from regular operating procedures as a reference point for potential monetary design and/or operating mode, but not as a “logical” basis for describing a system that when characterized hangs somewhere between factual and counterfactual.

The Language of Accounting and Finance

MMT uses accounting as a critical input to its analytic process, but transforms the usual language of accounting and finance as part of its expositional technique. This language construction is an innate feature of the MMT paradigm. A typical example is the notion that taxes do not constitute a source of revenue, or financing, or funding for government – or words to that effect. This is an abrupt change to standard accounting and finance terminology.

One does not need to make a claim suggesting that taxes are not a source of government revenue in order to highlight the policy space that is available via deficit financing. It is enough that mainstream economics for the most part has largely failed to incorporate solid accounting foundations (especially for banking and money) into economic theory. That shortcoming doesn’t need to be accentuated by a revision of standard accounting principles and terminology in the heterodox case, particularly given the importance that accounting is typically accorded so rightly as an input to the analytic process for heterodox analysis of monetary operations in general. Accounting is a precious resource for economics. MMT’s awareness of its importance on the input side is too valuable to abandon it with unnecessary language transformations on the analytic processing and policy output sides. Taxes constitute revenue for government in accounting terms. Nothing in the arsenal of MMT arguments for opening up deficit spending policy space requires a rejection of this standard.

Part of the motivation for this MMT treatment may be an inclination to describe balance sheet management in flow of funds language, while crowding out standard linkages to corresponding revenue and expense items where they exist. But the coherence of balance sheet, flow of funds, and income accounting as a set of interconnected financial statements is very appropriate to maintaining consistency when adding up all sectors taken together, including the government sector, as captured in national income and macro flow of funds accounting. The opening up of deficit spending policy space can be handled easily in this standard accounting context, by adding in functional finance and Godley sector balance analysis as appropriate.

“Taxes cannot be a source of revenue in the consolidated balance sheet. They do not add monetary assets, they reduce liabilities.” (TW Page 15)

This statement is also unfortunate. The test of revenue is not necessarily the creation of an asset. The application of revenue to the reduction of liabilities is common place, and not just in government. A simple example is the payment of interest on a commercial bank loan by debit to a deposit held at the same bank. Moreover, it is not categorically the case that taxes can’t add to monetary assets in the case of government. It is possible in theory for government to run a cumulative surplus, in which case taxes do add to monetary assets in the form of necessary monetary claims on the private sector. While this is unusual to say the least (and arguably undesirable), the point is that such an expositional technique avoids normal accounting, ostensibly on the basis of policy preference. There is nothing to criticize in the sense of the policy preference itself – it is quite rational. But this preference should not be used as the basis for an analytic technique that really should be neutral with respect to accounting norms.

And here it extends to bond offerings:

“It also makes sense to state that bond offerings are voluntary and taxes do not finance spending.” (TW Page 12)

We see a slippery slope. In fact, bond offerings are not voluntary in the actual institutional arrangements that exist. To deny this is like saying that highway speed limits are voluntary. Bonds are only voluntary in a counterfactual construction. A counterfactual monetary system can provide this kind of ultimate flexibility. But it is the institutional design itself that is the essential choice – not the bond offerings under an existing institutional design that specifies them as part of normal operations. Bond offerings are only voluntary in an institutional system that is designed to make them voluntary.

Self-Imposed Constraints

Self-imposed constraints as characterized by MMT are inherent in a number of operational rules associated with current institutional arrangements. Examples include prohibition of overdrafts in the TGA, prohibition of direct bond buying by the central bank, and the requirement to fund deficits with bonds.

According to MMT, self-imposed constraints are considered to be a nuisance that only obscures access to feasible and desirable policy space. This is an understandable interpretation. And it may help explain the way in which such constraints fall away in the exposition of the “consolidation hypothesis”.

Tymoigne-Wray note that workarounds may be available for the evasion of such constraints even in the current institutional configuration. Such workarounds should still be considered as institutional contingencies to the degree that they are not intended as part of the normal design and functioning of the monetary system. For example, the TGA overdraft rule might be eliminated while still retaining institutional separation between Treasury and the Central Bank, but presumably only with the approval of Congress. This would allow for significant operational change in deficit financing, with Treasury being supplied TGA balances without having to issue debt, thereby expanding bank reserves in the process of deficit spending. The inspired trillion dollar platinum coin idea (courtesy of Carlos Mucha) is another example of this sort of change.

The Contingent Institutional Approach specifies such temporary or even permanent lifting of normal constraints as a type of counterfactual institutional scenario. Operations will change and balance sheet accounting will change as a result of any of these types of adjustments. And such changes might invite a more comprehensive rethink of the existing deconsolidated institutional situation. Potential for a more formal integration of policy formulation is then worthy of serious investigation. Congress always has the option of changing the existing deconsolidated configuration of Treasury and the Central Bank in a fundamental way, effectively combining them into a consolidated operation. Full blown institutional fusion of Treasury and the Central Bank is the comprehensive option.

The category of self-imposed constraints constitutes a broad space in total. A monetary gold standard may be the mother of all self-imposed constraints. A permanent zero interest rate policy by design (an idea favored by some MMT advocates) could be viewed as a self-imposed constraint, and could well end up being as brittle as gold in its effectiveness and its endurance. The same might hold for NGDP targeting. All of these are examples that pose critical points of rigidity in the design of monetary policy.

Government Default

“The point is that, in that extreme case where nobody wants to buy bonds from the Treasury, the central bank will intervene, or the Treasury will finds ways to avoid having no funds in their coffers.” (TW Page 27)

Thus, MMT assumes away the possibility of default in its consolidation hypothesis. That is consistent with its determined unshackling of self-imposed constraints as a fundamental part of that hypothesis. It is also consistent arguably with Congress needing to act to avoid voluntary default in certain circumstances. That said, voluntary default was indeed on the table as an “unthinkable” operational possibility during the recent US debt ceiling impasse.

There is one type of “insurance” against involuntary default that is worth examining in more detail. The point is often made that the primary dealers are essentially obligated to buy Treasury bonds at auction, and that the Fed is always capable of supplying sufficient reserves in order to impel the dealers to buy the bonds. For practical purposes, this is plausible to a point. However, one should be careful with the logic of such an assumption, for several reasons.

The first thing to point out is that there has been a common misunderstanding of what has been meant historically by the term “bond vigilante”. The critical risk in question in the first instance is not that dealers would necessarily perceive default risk on government books – but that they would fear the potential effect of interest rate risk on their own capital positions. That is the origin of the 1980’s “bond vigilantes” characterization – not direct default risk. In a market in turmoil, dealers may substantially back down their pricing on bids for bonds at auction, essentially predicting that the Fed will have to raise short rates at some point in the future. This is their way of avoiding capital losses, based on their own expectations for future Fed rate policy. Hence their “vigilance” over what they think the Fed may be forced to do with future monetary policy in conditions that warrant such concern. That vigilance is in effect a democratic call for Fed tightening as much as an expectation of it. Heterodox observers therefore should be careful in announcing that there is no such thing as “bond vigilantes”, based solely on their analysis of the usual dimensions of default risk and “affordability” on government books. There always have been and there always will be bond vigilantes, although it may be quite some time before they show up in full force again, given current zero lower bound interest rate conditions. So bond vigilantism is all about interest rate risk, and the effect that interest rate risk can have on private sector equity capital positions. And one should consider further that at least in theory, primary dealers may choose to give up their status rather than expose their capital to horrific interest rate risk in a nightmarish interest rate scenario.

Furthermore, dealers will not force themselves to buy bonds of longer and riskier duration just because the Fed has supplied excess reserves in the system. Taking bonds into position requires the allocation of capital for interest rate risk. Interest rate risk is similar to credit risk in this sense of capital requirements. We know that banks don’t “lend reserves” to take credit risk. Neither will they or their dealers necessarily assume interest rate risk simply because of the presence of excess reserves. And in the nightmare scenario, when dealers at least consider withdrawing in order to protect their own capital positions, a technical default scenario is more imaginable, unless the central bank steps in with extraordinary measures that go well beyond providing dealers with reserves. Like bank lending officers, dealers lend (i.e. buy bonds) based on risk to capital. And in the case of bank-owned primary dealers (i.e. most of them), that decision is not centered rationally on excess reserve availability.

All defaults by sovereign monetary operations are self-imposed in the sense that extraordinary institutional action can avoid it. But we should be cautious about making absolute claims regarding such a buyer of last resort function for primary dealers when it comes to the risk of default. Bond vigilantes focus directly on interest rate risk, and no private sector investment dealer will commit to buy bonds unconditionally under any and all circumstances. In putting private capital at risk, a dealer may reach a tipping point where it chooses to sacrifice its dealing license or seek extraordinary stop-loss assurances from government – an institutional/operational change of a sort.

“Thus, insolvency and bankruptcy of this government is not possible. It can always pay.” (TW Page 5)

The key technical difference between the US and Greece monetary systems obviously does not come down to the primary dealer function. That does not make a lot of common sense. The fundamental difference lies in the pure technical structure of monetary operations and policy, together with the power of the sovereign (the US Congress in this case) to shape institutional arrangements – in other words the power of the US Congress to invoke contingent institutional change when absolutely necessary to avoid default. The Greek government has no comparable capacity, having surrendered that to the auspices of the European Central Bank.

Policy Independence versus Operational Co-ordination

“Ultimately, the financial operations of the Treasury and the central bank are so intertwined that both of them are constantly in contact to make fiscal and monetary policy run smoothly. The Treasury gets involved in monetary policy and the central bank gets involved in fiscal policy. As such the independence of the central bank is rather limited and it must ultimately financially support the Treasury in one way or another (Tymoigne 2013).” (TW Page 26)

The core idea behind central bank independence is responsibility for the formulation of monetary policy – not freedom from the expectation of rational co-operation in the implementation of it. In the case of the United States, this policy responsibility has been delegated by the US Congress to the Federal Reserve. The US Treasury does not determine FOMC policy. And the fact that another department of government co-operates in operational implementation of policy does not mean that the central bank has lost its policy independence. As the authors point out, Treasury assisted the Fed in the early part of the financial crisis, prior to the Fed’s receipt of legal approval to pay interest on bank reserves. That the Fed was able obtain Treasury’s operational assistance during this temporary gap period was a sign of the Fed’s position of policy responsibility, not the opposite.

Alternative Consolidation Techniques

“Finally, we note that the technique of consolidating the central bank and Treasury is frequently adopted outside MMT, for purposes quite similar to our purposes. Even Lavoie has used this technique in the textbook he co-authored with Wynne Godley.” (Page 14)

MMT’s “consolidation hypothesis” is unique. Consolidation as developed in Godley and Lavoie’s “Monetary Economics” bears little resemblance to what MMT does with the same concept. MMT stamps its consolidation hypothesis with its own trademark logic that purports to be comprehensive with respect to the existing monetary system. Godley and Lavoie do no such thing. They define their consolidated model SIM as the first in a long series. The logical flow of the book is from simplicity to complexity. There is no critical path that ends with the reduction of detailed complexity to “logical” simplicity. Every one of G&L’s complex models is logical and complete in its construction – not just model SIM. Virtually every one of them probably has some counterfactual element or another relative to the full detail of today’s modern system – save perhaps the final most comprehensive one, where the authors assemble as much as is realistically possible to model the actual economy, keeping within a pragmatic scope for the aim of the book as a whole. They set aside the consolidation characteristic of their initial model SIM in developing the book’s subsequent differentiated models. The purpose of G&L’s SIM model thus is to provide an analytical start point for the evolving complexity of subsequent models. There is no suggestion that this complexity is inherently less logical than simplicity. That is in contradistinction to MMT’s consolidation hypothesis. Moreover, recognizing the fact that Marc Lavoie is one of the critics to whom Tymoigne-Wray reply with their paper, it should not be surprising that that G&L use the idea of consolidation differently than what is revealed in the “consolidation hypothesis”. So this is not the best comparison if it is meant to suggest a similarity of approach with the “consolidation hypothesis”.

The Contingent Institutional Approach

The US Federal Reserve in its current arrangement is institutionally separate from the US Treasury. This arrangement is represented symbolically as {CB, TR}. An important defining feature of this separation is that the US Treasury has a bank account with the Federal Reserve. From an accounting perspective, it is an asset of the US Treasury and a liability of the Federal Reserve. From an operational perspective, the Treasury writes cheques on or otherwise instructs payments from that bank account. Importantly, the US Treasury is not itself a bank in the context of this arrangement.

An institutional choice different from that which exists now can be classified as a counterfactual construction. One such choice consists of the institutional fusion of the existing US Federal Reserve and the existing US Treasury function. Two previously defined institutions now become one. We referred to this in the Contingent Institutional Approach as CTRB – the Central Treasury Bank, an institutional consolidation of {CB, TR}. An important defining feature of this arrangement is that for purposes of institutional design and definition, there is no longer a bank account issued by the Fed and held by Treasury – because these are no longer separately defined institutions. In other words, the TGA disappears from operations. An elimination that manifested previously only for purposes of financial accounting consolidation (and in the case of MMT, its “consolidated hypothesis”) now becomes an elimination in actual operation and in the underlying account structure that defines the external boundaries of the new institution.

The CTRB is in effect a bank. It is an unusual bank because its equity is structurally negative, given the standard net liability profile created by the US debt. From an operational perspective, the CTRB writes cheques or otherwise instructs payments like any sovereign central bank does – “from thin air”. Because no higher monetary authority exists, there is no external bank account from which to make payments. Because of its monopoly position in its own liabilities, it creates its own balance sheet as it goes along, through net “efflux” of the medium of exchange.

There is a rough similarity in basic form between the CTRB structure of the Contingent Institutional Approach and the model SIM from Godley and Lavoie’s ‘Monetary Economics’. Both are specified as institutional design choices. With respect to the relationship of those two models to the MMT “consolidation hypothesis”, there is no similar claim that a consolidated operation is necessarily more “logical” than a deconsolidated one.

Contingent Institutional Design – SOMA/TGA Fusion

The CTRB (Central Treasury Bank) institutional structure fuses the prior existing SOMA (the Fed’s System Open Market Account) and the TGA (the Treasury General Account at the Fed). The result allows a different interpretation of consolidation than what is presented in the MMT consolidation hypothesis.

We should set this up by reviewing our earlier discussion once more. Starting with the existing institutional set up, Tymoigne-Wray describe a required co-ordination between Treasury and the central bank in the management of aggregate bank reserves so as to target interest rates. That is correct as far as its goes. But as noted earlier, the central bank does that in the larger context of also monitoring the distribution of the totality of bank reserves and TGA balances, including distribution effects that may not necessarily involve net TGA quantity effects. Given that all depositors with the central bank practice proprietary cash management disciplines, including Treasury, the central bank must be vigilant over the behavior of all those deposits, as the distribution of those balances at any moment in time may have temporary effects on short term interest rates. This vigilance combined with responsive action when required is critical to the Fed’s control over the interest rate level at which fed funds trade (i.e. pre-2008). Again, a redistribution on its own won’t change the totality of TGA balances and reserve balances. (Note that such redistributions include self-directed transfers of funds between TGA and Treasury tax and loan accounts (TT&L), since any such transaction moves money from bank reserves to TGA balances or vice versa.) It is central bank principal operations intervening from the asset side of the balance sheet that change the totality of these balances – which may be required as a matter of interest rate management. From an institutional perspective, there is no reason to argue that the efflux from the Central Bank and Treasury is any more conjoined as a source of the Fed’s interest rate control motivation than the efflux from the Central Bank and JP Morgan in an analogous flow of funds situation (i.e. with JP Morgan producing net reserve account outflows from its account instead of Treasury producing net TGA outflows from its account). Both JP Morgan and Treasury practice cash management disciplines aimed at controlling such net effects through their accounts with the Fed. Thus, the US Treasury, JP Morgan, and Citibank rank pari passu operationally in this regard, and in the context of existing institutional arrangements, it is misleading to attach a consolidation logic to any single pairing over the other because of cash flow patterns.

Thus, individual commercial bank reserve balances and Treasury TGA balances rank pari passu with respect to flow distribution effects requiring central bank responsive action. This characteristic applies to Treasury balances, reserve balances, and the sum of Treasury and reserve balances. Tymoigne-Wray note the aggregate quantity effect for reserve balances as a subset of total balances, but play down or omit the unifying aggregation characteristic of the distribution effect across all balances at the Fed. That distribution may be highly volatile, affecting the Fed funds trading rate in similar ways whether or not the source of the disruption is due to a redistribution involving TGA balances. The potential for these disruptions is tempered by the standard cash management disciplines used by both the commercial banks and Treasury. Indeed, this cash management discipline on the part of Treasury constitutes what MMT might classify as a “self-imposed constraint”. But this is an important feature of actual monetary operations. And accordingly it constrains the appropriate logic that might otherwise be more persuasive in framing the “consolidation hypothesis”.

MMT’s consolidated description treats the issuance of Treasuries as a seamless reserve drain rather than a normal course separate TGA operation. Given the collection of competing proprietary cash management disciplines involved, this is comparable to suggesting that JP Morgan conducts its own reserve/TGA drains from the rest of the system whenever it receives net long cash inflows in its reserve account. Yet we don’t speak of JP Morgan doing cash additions and drains to non JP Morgan balances held at the central bank. Extending the comparison down the money hierarchy, we don’t describe individual households doing M1 drain or add adjustments with counterparties in the universe of remaining bank customers.

I believe this analysis is consistent with the critique of Brett Fiebiger, whose work is cited in the TW paper. Fiebiger in effect has also questioned the logic of conflating SOMA and TGA efflux/reflux – SOMA in the course of financial asset acquisition, and TGA as a result of goods and services acquisition or transfers. Treasury and the commercial banks are all depositors with the central bank in existing institutional arrangements. It is a category error to presume that Treasury is inherently different in this regard, given existing institutional arrangements. Redistributions between TGA balances and bank reserves are only a subset of all relevant redistributions.

Thus, a consolidated reserve add/drain function unique to Treasury and the central bank doesn’t really characterize the logic of existing monetary and fiscal operations.

But it could describe a counterfactual CTRB operation as suggested in the Contingent Institutional Approach, where the TGA is actually eliminated as an operating account because of the explicit design of institutional structure. So we can definitely consider an institutional counterfactual case that includes formal SOMA-TGA account consolidation with associated operational unity. And we can now follow through on this to show that there is indeed a difference in the overall interpretation of consolidation when considering this explicit counterfactual.

With the explicit design fusion of SOMA and TGA, the cash management and market operations of CTRB take on a distinct Open Market Operations (OMO) character that occurs in the existing unconsolidated operations only in the case of the Federal Reserve SOMA account. From an operational perspective, there is now a seamless efflux of reserve creating expenditures in the case of both goods and services and asset acquisitions, with reflux of reserve destroying inflow in the case of taxes, bond issuance, and asset disposals. There is no longer a central bank operating in reaction to both commercial bank and Treasury activity. The Central Bank is Treasury and Treasury is the Central Bank – the CTRB – and the consolidated institution now reacts in a unified way to the behavior of the banking system in the context of the effect of reserve distribution on the CTRB’s interest rate control function.

Implicit in the concept of a CTRB structure is the idea that there would be associated flexibility in a single liability management function – so that the institution may have more options open to it in terms of the mix of bank reserves and debt that are issued. That would depend on the asset-liability management flexibility it was granted by Congress.

In this CTRB context, the following statement takes on more practical relevance:

“Treasury operations do affect the volume of government currency when the consolidation hypothesis is used. If the Treasury really were to spend trillions” then in the first instance that would create “trillions” of bank reserves, but these would be drained by “trillions” of sales of treasuries.” (TW   Page 35)

Of course, this description was intended to apply to the MMT consolidation hypothesis, but is really more apt to a specified counterfactual arrangement. As pointed out earlier, the deconsolidated reality of existing institutional operations includes a TGA cash management discipline in which gross volumes of outflows and inflows are managed to be in balance as part of that process, similar to the commercial banks. One rationale for the CTRB structure is that CB/TR liability management becomes operationally integrated according to design intention – so it would be natural to open up policy to the possibility of more discretion in the mix and timing of bank reserves and bonds issued by the CTRB. The “trillions” in focus above would be of more realistic interest, given the options that may open up more naturally in a process of fully integrated liability management. And the operational window through which that would be executed would be an expanded SOMA function.

The CTRB institutional structure is interesting to consider in the context of our earlier point on the distinction between policy independence and operational co-ordination. In the consolidated CTRB counterfactual, each previously separate function has a natural leadership role in blending each of policy and operational responsibilities within the new institution. The new institution absorbs the previous monetary policy independence of the Central Bank. The unified institution would be responsible for comprehensive asset and liability management policy formulation and implementation, as well as the execution of budget spending authorizations, tax collections, and bond financings (where desired), with Congress authorizing fiscal spending as well as broad parameters for CTRB asset and liability management.

It is somewhat ironic that existing central bank operations incorporate a consolidated view – not of Treasury and the Central Bank in logical terms – but of Treasury and the commercial banks in operational terms. It is the counterfactual CTRB that represents Treasury and Central Bank consolidation in both logical and operational terms, and which then has a singular operational focus on the commercial banks, whose reserve balances form the exclusive deposit liability category of the CTRB.


The consolidation of financial statements is not unusual in accounting practice. Corporate conglomerates routinely consolidate subsidiaries in financial reporting. At the same time, they provide descriptions of subsidiary operations in reporting to shareholders. There is nothing particularly unique about the case of Treasury and the Central Bank in the aspect of consolidation alone. A balance sheet consolidation of two separate existing institutions sheds light on the joint effect of operations, but obscures the operational route that gets us there.

MMT analysis uses high quality input from a factual standpoint, with an emphasis on central banking and fiscal operations. That input is transformed according to the consolidation hypothesis to an expositional platform for policy options and recommendations.

The existing monetary architecture of separate institutions should not an impediment to the exploration of that policy space. It can be described in its current condition, with various adjustment options that might be available if arguments for greater policy freedom and constraint relaxation are legitimate in their own right.

Policy doesn’t need to be argued on the basis of assumed institutional change that has not yet been normalized. Everything MMT wants to achieve should be achievable – if it can be argued effectively from a policy perspective – without the need to fade the factual separation of existing monetary and fiscal operations. This is an unnecessarily indirect method for tackling the challenges of exploring the policy space. Why submerge the current reality as if it has already been changed in such a way? This question applies to the full continuum of such change – from the lifting of self-imposed constraints that are currently normalized within the prevailing institutional structure to the consideration of a more formal institutional consolidation of Treasury and the Central Bank.

Finally, I wish to emphasize once more that the purpose of this essay has not been to criticize the nature of MMT’s policy orientation. There are many interesting ideas associated in some way with that part of MMT – things like the shortening of bond financing duration, or the outright elimination of bond financing, or the permanent imposition of a central bank zero interest rate policy, or the Employer of Last Resort (ELR)/Job Guarantee (JG) program. Those ideas should be evaluated in their own right, separate from what has been examined here. Nor has the purpose been to deny MMT’s high quality input of real world observation of monetary and fiscal operations in the way in which it approaches the subject matter of monetary economics in general. The observation here is about the thing that lies between those two stages – the template characterization of the existing monetary system.

Presentation style is a choice, and target audience is a consideration. Preferences for presentation style are understandable, as they mold strategies for reaching target audiences, including policy makers. The point simply is that there are different ways of doing this – ways that don’t necessarily conflict with a desired exploration of policy possibilities and ultimate policy influence, and that conceivably may even facilitate achieving those objectives.

Comments are closed.