Recent Posts

Sunday, July 19, 2020

Changing Budgetary Procedures: Outline Of The MMT Approach

Although most discussion of reforms associated with MMT are in the areas in upcoming sections of this chapter, the first thing that must be understood is that MMT proponents argue that the framework for analysing fiscal policy needs to be changed. The title of the article “Replacing the Budget Constraint with an Inflation Constraint” by Scott Fullwiler summarises the change perfectly.

NOTE: This is a very preliminary unedited draft from a section of my manuscript. I am putting it up somewhat early as I wanted to get feedback. I have spent very little time cleaning up the text, and so it will probably be tweaked heavily.

Functional Finance suggests that inflation constrains fiscal policy for a floating currency sovereign, not financing concerns. As such, that is how budgets ought to be analysed.

This is a non-trivial change, and given the difficulty of modelling inflation, people will disagree over details. Nevertheless, it is better to attempt the correct thing rather than do the wrong thing because it is easier.

Another reason to emphasise this topic is that it nullifies a significant portion of MMT criticism (which will be waded into within Chapter 5). The premise of MMT fiscal policy analysis is that the effect on inflation is the correct methodology. By implication, it makes no sense to complain that “implementing MMT” is somehow inflationary by itself. The only way that critique makes sense is that the critic and MMT proponent analyse inflation risks differently – which is almost certainly the case. That said, we need to look at the competing methodologies, and see which appears to be more useful for that task (a step that MMT critics generally do not do, since that would require reading primary research).

Fullwiler Article

The Fullwiler article was in response to analysis by Tim Worstall. The following text summarises the argument.
As argued bazillions of times, the real point MMT is making is that the government’s budget constraint is the wrong constraint—the correct constraint is whether or not a particular budget position will raise inflation beyond an official target rate (say, 2%, which seems to be the choice of most central bankers).
Let me explain to Mr. Worstall and others how this could work rather easily—just as the CBO and OMB now evaluate government budget proposals regarding their effects on the budget stance, the CBO and OMB could instead shift focus on evaluating these proposals against the inflation target (I argued the same thing here [link to SSRN paper]). Much like how policy makers supposedly take estimates of effects on the budget position rather seriously in making budget conditions, they could replace these with projections of inflationary effects. An inflation constraint provides more fiscal space than a budget constraint, but in no way does it provide unlimited fiscal space (again, as we’ve always argued).
Fullwiler notes that expertise to do such functional finance analysis no longer exists. (This was a feature of wartime planning.) That reflects the decisions of the neoclassical consensus, which were a mistake in the eyes of MMT proponents.

Copyright laws constrain my ability to quote from that article, and so I will just offer an outline of my personal views on the matter. However, if one is interested in the academic consensus, that article would be the place to start.


Although it seems clear that excessively loose fiscal policy poses inflation risks in the absence of demand management policies that were implemented in World War II, drawing the line where “excessive” starts is not at all obvious. If we look at the post-1990 policy environment, we see that no implemented policies caused a large deviation from inflation target levels (Section 2.5).

Therefore, under current institutional conditions, there is clearly a buffer area where fiscal policy changes do not have a measurable effect on inflationary outcomes. There are many competing explanations for this possibility, and I would argue that not everyone agrees on which one(s) are correct.

Another thing to keep in mind is that observed outcomes are the result of decision-making in both the private and public sector. If the private sector launches a delusional fixed investment bubble, it can cause observed inflation to rise above target – even if governmental policy in the absence of the bubble left inflation on target. All forecasting approaches must contend with fundamental uncertainty.

The post-Keynesian theory of inflation suggests that it is a far more complex process than suggested by aggregated neoclassical models. I do not expect that it will be easy to write down an abstract approach to functional finance that will work under all circumstances.

Another practical issue is that each government approaches this task differently. The American budget process is convoluted, and reforming it from within is extremely difficult. Conversely, a country with a largely dictatorial Prime Minister's Office like Canada might be able to ram a framework through, and it would be up to the civil service to figure out the details. (That assessment is coming from someone who has zero interest in Canadian parliamentary procedures, it could easily be very wrong.)

Factors in Favour of the Approach

Although I agree that an abstract approach to Functional Finance that applies to any budgeting situation is an extremely complex task, the government has the means to make its life less difficult. In particular, it does not need to be price taker (Section 4.3). A major reason why the Old Keynesians helped drive the inflationary cycle was that they wasted efforts on aggregated analysis, and they did not understand the causes of inflationary pressure. Modern neoclassicals follow the same aggregated approach, and based on comments by critics, they assume that MMT proponents do the same.

If we look at the Monetary Monopoly Model discussed in Chapter 4, we see that the government can fix the price it is willing to pay for goods and services, and let the quantities adjust. To a certain extent, this is partly done: salaries and transfers are set as nominal quantities, although there is often (partial) indexation of these payments.

The idea is that the government needs to set the price it is willing to pay in other contexts. That is a key feature of the Job Guarantee (Section 3.3) – which explains why people like myself are not concerned about the long-term inflationary effect of the policy without even needing to dig into budget methodologies.

This principle can be extended, and departmental budgets could be set with conditions on prices paid. There will obviously be cases where the government will have no choice to be a price taker, but these are likely for expenses that represent a small percentage of total spending. (For example, some pirate-like software supplier might raise the price of its software.)

Most of the time, governments are not making radical changes to its expenditure patterns. It seems very likely that back-of-the-envelope analyses will give adequate results for mild fiscal policy changes.

For larger changes, governments will need to dig into the real resource requirements of the policy, such as the materials used as well as personnel requirements. It is hard to set out the exact parameters of such an analysis that covers all possible policies; realistically, each programme needs to be judged on its own merits.

Another approach to deal with the inflationary aspects of governmental policy is to use administrative measures to control inflation. That is a controversial subject, and discussed in Section 5.6.

Concluding Remarks

The real concern about the side effects of fiscal policy changes is the effect on inflation. The fact that “hyperinflation” appears in most conventional criticism of MMT shows that this is a consensus view. Meanwhile, analysing inflation risks is difficult, and anyone offering a magic analytical solution – like the Friedman k% rule – is most likely incorrect. This is where policy debates ought to focus, instead of juvenile musings about “bond vigilantes.”

Chapter 5 digs into criticism of MMT. Although I am not impressed with almost all of the critiques I outlined, I want to emphasise that getting inflationary analysis correct is non-trivial. As such, I am perfectly willing to accept that a particular inflation analysis by a MMT proponent is incorrect. That said, the MMT proponent will correctly be focussing on what matters – which cannot be said of many conventional analyses.

References and Further Reading

(c) Brian Romanchuk 2020


  1. Perhaps the JG wage target should be a wage designed to pull workers from the (pick) industry, thereby decreasing (pick) output and causing price increases. The remaining (pick) workers would probably get higher wages, (pick) prices would probably rise, and we should be able to judge whether our (2%?) inflation goal has been reached.

    Setting JG wages would be part of the budget process. If government is redesigned to micro-manage the economy, choices like this would need to be made.

  2. This comment has been removed by a blog administrator.


Note: Posts are manually moderated, with a varying delay. Some disappear.

The comment section here is largely dead. My Substack or Twitter are better places to have a conversation.

Given that this is largely a backup way to reach me, I am going to reject posts that annoy me. Please post lengthy essays elsewhere.