There has been discussion on various blogs (for example, here, here and here) concerning the choice, frequently made by expositors of MMT, to present the theory in terms of a sovereign currency issuer not subject to self-imposed constraints (what Scott Fullwiler calls the ‘general case’) rather than grounding discussion in the particular operational realities of whatever monetary system happens to be the subject of the presentation (the ‘specific case’). Some commentators have also questioned the appropriateness of giving priority, in theory, to the ‘general case’ rather than the ‘specific case’.
My own view is colored by what attracts me to MMT in the first place. For me, what is most striking about MMT is its clear depiction of the social possibilities that are inherent in a flexible exchange-rate fiat-money system. MMT gives transparency to the various policy options under capitalism. It may also indicate ways in which fiat money could facilitate more fundamental change, including a transformation to a system other than capitalism.
From my perspective, I think it would have been a mistake for the leading MMTers to cast their theory narrowly. Instead they have developed an overall coherent framework that enables analysis of particular instances within it. The purpose of macroeconomic theory is not to describe every minute detail, but to achieve an understanding of the whole and how the parts relate to the whole. Certainly it should be possible to mine down to the smallest detail without violating the logic of the overall framework when such details are of interest. If this could not be done within the general framework, there would be something wrong. But I don’t think this is the case.
A comparison with neoclassical economics may illustrate my viewpoint. In neoclassical economics, the theorist begins by asking what constraints must be imposed on the model to achieve the desired result. Once these constraints have been identified, the theorist attempts successively to relax constraints to see how few can be imposed on the model without contradicting the desired result.
In MMT, the approach is different. By observing the actual system, the theorist tries to discern its key features. An attempt is then made to represent these features as simply as possible without missing anything essential. Many details will undoubtedly be ignored in the simplest models in an effort to illuminate the most important facets of the system in a clear way. Successive complications, encountered in particular (real-world) cases, can then be introduced to determine how the results of the model are modified, how the key processes operating within the system are attenuated or reinforced, and so on.
So in neoclassical theory, the result of the model is predetermined. The theorist then sees how many constraints on the model can be relaxed without altering the result. In contrast, in MMT the result of the model is not predetermined, and the constraints that are added in examining a particular instance of the base model reflect the actual system being studied.
The basic methodology of MMT can be observed in other Keynes influenced approaches. For example, a Keynesian might start with the simple two-sector closed economy model. The motive for doing so is that this simple model is thought to capture certain fundamental features of a market economy. It involves businesses and households. It shows how leakages endogenously adjust to exogenous injections, and how autonomous expenditures determine income through a multiplier process. It is not completely “realistic” and does not capture every detail, but the central processes it highlights continue to hold true when more institutional detail is introduced. In more complex models, leakages still adjust to injections. Autonomous demands still determine income. And so on.
In a similar way, the exposition of MMT sometimes starts with a two-sector model involving the government and non-government. In this simple form, some fundamental aspects of a fiat-money system are made clear. The non-government cannot net save unless the government is in deficit. The non-government is in no position to dictate the rate of interest government pays on its own liabilities. The government need not issue debt. If it chooses to do so, it is a self-imposed constraint. In considering particular cases, these constraints can be added to the model to consider what, if any, ramifications they have for the government’s policy capacity. If such constraints alter fundamental insights of the simple model, the simple model would need modification. Otherwise it could not be regarded as capturing the essence of the reality it is designed to explain.
Along similar lines, I think the distinction between the ‘general case’ and the ‘specific case’ described by Scott Fullwiler makes sense, as does the MMT judgment that the general case should refer to the position of a sovereign currency issuer prior to any self-imposed constraints. Since the number and strength of such self-imposed constraints are various – depicted, for example, in Fullwiler’s ‘strong form’, ‘semi-strong form’ and ‘weak form’ – attempting to organize the theory around one particular case – e.g. current U.S. monetary operations – would seem to deprive the approach of generality. As long as the basic monetary operations are playing the same function in the specific case as they would in the general case, the claim to generality seems well founded. Fullwiler gives the following example:
Having said that, MMT’ers are keenly aware that governments can and do write laws that their treasuries’ operations be legally bound in certain ways. For instance, the Fed is constrained by law to only purchase Treasury securities in the “open market,” is thereby forbidden from directly lending or providing overdrafts to the Treasury. In other words, “specific” cases can and do differ from the “general” case MMT’ers describe for a sovereign currency issuer under flexible exchange rates in the sense that self-imposed constraints specify particular operations. But, this does not mean that the operational function of the Treasury’s bond sales to aid the Fed has changed—to the contrary, with or without legal prohibition of overdrafts for the Treasury’s account, either the Fed or Treasury must offset flows to/from the Treasury’s account to achieve the Fed’s target rate (with the caveat that interest on reserve balances can potentially eliminate this necessity).
This is not to say that analysis of the specific case is unimportant. It will be appropriate, for example, when formulating policy proposals that enable the government to adhere to its self-imposed constraints. Analysis of the specific case also enables a consideration of what, if any, policy capacity is given up when particular self-imposed constraints are in place.
However, even when analysis of the specific case is appropriate, an understanding of the general case remains relevant. For the very reason that these constraints are self-imposed, they will not necessarily be observed when it comes to the crunch. So it makes little sense to lose sight of fundamental relations of authority and hierarchy within the system, even when analyzing the specific case. To quote Fullwiler again:
The overarching point here is to recognize who sits at the top of the hierarchy of money for a given monetary regime. Since under flexible exchange rates it is the currency issuing government, self-imposed constraints are simply that—self-imposed and not operational. … Indeed, it is the very fact that such self-imposed constraints can be and have been disregarded in the past when it has been deemed desirable (e.g., the law requiring that the Fed only purchase Treasury obligations in the open market has been periodically relaxed) that demonstrates who is in charge …
The fact that a constraint is self-imposed is in this sense a strong reason for not including it in the general case. Since such a constraint may be removed or replaced, and may not be adhered to by other sovereign currency issuers, it is not a general feature of a flexible exchange-rate fiat currency system.
Of course, at a still more general level, a flexible exchange-rate system is itself just a special case of all fiat-money systems, which in turn are a special case of all monetary systems. The level of generality desired is a function of the type of analysis we wish to undertake. As mentioned at the outset, my own preference, in this respect, is shaped by what I find most interesting in MMT.