Recent Posts

Wednesday, November 7, 2018

Comments On Eric Lonergan's Fiscal Rules Post

Eric Lonergan has recently written a post about the fiscal rules debate that has cropped up around the U.K. Labour Party's adoption of such a rule. As usual, his post covers a lot of territory -- if I were to attempt to cover all the points he raised when following my preferred writing style, it would be a 10 part article. Meanwhile, I am supposed to be working on finishing off my breakeven inflation book, as well as dealing with other projects. As a result, I just want to respond to a couple big picture points.

The article is interesting, but I would note that Eric Lonergan and I disagree on a subject (starts with "m," rhymes with "honey") that I want to abolish from economic discourse. Therefore, even though I like the article, I have reservations that I do not wish to pursue.

The Rule Being Discussed

I have not followed the internal Labour Party debate carefully, so I will take the liberty of using a lengthy quote from Lonergan's text.
So what is this (in)famous fiscal rule and what is the nature of the criticism it has received? Deriving a relatively clear and simple rule for fiscal policy does not imply simplistic analysis. For full context on the depth of understanding on the complexity and difficulties this paper by Wren-Lewis and Portes is a good starting point. 
Based on their extensive work, Wren-Lewis and Portes, together with the shadow treasury team, conclude that the government should balance current expenditure over rolling five year periods, with a knock-out operating at the lower bound on interest rates. The government should also aim to target a reduction in the ratio of public sector debt to trend GDP over the life of parliament. 
This looks very reasonable – and clear. It is totally at odds with austerity in the face of a deflationary recession, because it advocates suspending the rule at the zero interest rate bound, and it permits borrowing for capital expenditure, because the rolling balance is only for current expenditure. It also imposes a constraint on public capex, because there is a long-term target for the ratio of net debt to trend GDP.
If I were to put my "conventional economist" hat on, I would not see anything in that text that warrants particular interest. If the government attempts to accomplish nothing with fiscal policy during an expansion -- which is a good summary of the post-1990 behaviour of developed countries -- the debt-to-GDP ratio would tend to decline in the expansion. The "zero interest rate bound" is just trendy mainstream academic waffle covering up the fact that recessions will happen, and the debt-to-GDP ratio will rise in a recession.

In other words, a pointless and arcane debate that left-wing political parties love to tear themselves apart over.

But if I put my "economic theorist" hat on, this is an interesting debate -- even if it is pointless and arcane.

Optimal Fiscal Policy

I did not look at the article by Wren-Lewis and Portes referenced by Lonergan; I already gave preliminary comments on their work in this article.

If I put on my "academic" hat on, I would note that they have stepped into what was part of my domain of expertise: optimal control theory. My doctorate was in the area of robust control theory -- which was built on the rubble of the failed optimal control theory paradigm. Optimal control theory was a fad in the 1960s, based on mathematical elegance. It had some useful results for path-planning problems; getting a spacecraft to the moon was probably its greatest practical achievement.

The problem with optimal control was that it failed completely in the face of uncertainty. Calculating the macro trajectory to the moon was difficult, but there is very little uncertainty about the physics. However, controlling for small deviations from that trajectory is more difficult, as it is harder to model the sources for such deviations.

When we move to a more uncertain environment -- such as controlling an aircraft -- optimal control failed completely. Optimisation leads to degenerate outcomes: control rules exploit properties of the baseline model, whereas the actual system may not have such properties as a result of uncertainty. When used in such applications, optimal control theory had catastrophic results, and had to be abandoned.

Robust control was built around the concept of model uncertainty, following on from a landmark article by the late George Zames of McGill University (my undergraduate alma mater). The approach has mathematical similarities to optimal control, but the control laws are not degenerate: they do not exploit assumed model properties to the maximum.

The optimal fiscal policy approach leads to a degenerate solution: the optimal net debt-to-GDP ratio is negative. Why? The optimisation problem to be solved wants to "fund" a target level of government consumption with the lowest possible tax rate. The way to do that within the model framework is for the government to accumulate claims on the private sector (a negative debt position). The technical problem is straightforward: is it reasonable to believe the model used would even slightly approximate reality if the government net debt level were negative?

If we think about this from first principles, we would realise that the truly optimal way for the government to fund target expenditures with the lowest possible tax rate is for it to seize the means of production, and use the profits to "fund" the target expenditures. From what I recall of the Cold War era, that used to be a big debate in economics. I do not believe that picking a obscure mathematical optimisation problem is the best way to settle that debate.

The Debate Labour Should Have

I generally stay clear of offering policy advice, but I would suggest that it is straightforward that a political party should focus its debate on what policy changes are to be made to steer the real economy in a certain direction.

The debt to GDP ratio tells us nothing about this task. Like other floating currency sovereigns, Britain's debt-to-GDP ratio adds almost no information for the guessing the probability of default, as that is a political decision. Even for other economic issues, it provides little guidance. For example, the debt-to-GDP ratio tells us very little about growth rates, other than reverse causality that low nominal growth rates lead to elevated debt-to-GDP ratios.

Lowering Debt-to-GDP is "Easy"

As an accounting identity, we know that the amount of central government debt is equal to the holdings of government debt by the "non-central government sector" (which includes the external sector). Governments can attempt to control the flow of debt (the deficit), but the behaviour of the stock needs to take into account the behaviour of the other sectors.

We just need to flip around the question: how do we reduce the government bond holdings of the other sectors? I would argue that there are three main culprits.
  1. The external sector. Shall we attempt to break out of trade agreements in order to return the current account to surplus? (U.K. trade policy is currently in disarray, and I have no idea what is going to happen on that front.)
  2. The wealthy and corporations. Income redistribution has led to large increases in wealth by the upper income quantiles. One may note the economic experiment conducted by the Trump administration -- cutting taxes on the rich did little to raise inflation, nor growth, but raised deficit levels. Could the reverse of such a policy be engineered, particularly given that the wealthy have the best paid accountants? 
  3. Pension (and insurance) assets. The buildup of pension and insurance assets represents a significant nominal demand drain. Contributions represent a tax on households, which slows growth. In turn, the government deficit widens -- creating financial assets that can be purchased by pension funds. However, we need to be careful for what we wish for. From what I have seen of demographic projections, savings flows will be reversing. This will give a boost to nominal demand -- and inflationary pressures. This may replace one set of problems with another.
The above are areas of political interest. However, I think any policies suggested by the points above (e.g., tax rate changes) need to be debated on their own merits, and not the effects on the debt ratio.

What does Functional Finance Suggest?

Eric Lonergan offers one simplified summary of the MMT position as follows:
Let’s consider ‘hard’ MMT – should fiscal policy target stable inflation or full employment, which some advocates appears to argue. Or to put it less forcefully, should the government run a deficit as large as it can until inflation starts rising, at which point it should tighten fiscal policy, which seems very close to what Stephanie is advocating?
He later discusses more nuances with MMT position, but I would suggest that this is possibly the wrong starting point. From my perspective, it is unclear that Functional Finance suggests that we should even attempt to get close to the "inflation barrier." I will immediately note that this is my personal opinion, and may or may not reflect the "MMT position." (This is my between-the-lines interpretation of various MMT texts, but it is not based on explicit statements by various MMT academics.)

The reason is straightforward: post-Keynesians do not claim to have a model that predicts inflation exactly. Yes, Functional Finance argues that inflation is the constraint on fiscal policy -- however, there is no claim that we know exactly where that constraint lies. We know that an extreme fiscal policy stance will almost certainly be inflationary -- modulo policies like wartime rationing -- but what about relatively moderate policies, which is actually what are being advocated?

This is very much in contrast to mainstream macro theorists, who claim to have models that can predict the effects of policy on inflation to the tenths of a percent. One may note that although such models are used at central banks, they are scorned by the people in finance who actually trade inflation-linked products. (At least they did; perhaps there has been a renaissance in DSGE modelling at hedge funds since I left the industry.) A mainstream economist may be able to claim to be able to know how to calibrate fiscal policy in such a fashion to remain exactly at the inflation barrier. (My reading of the literature is that even mainstream DSGE models do make such claims, but someone could invent a new DSGE model that has that property). Conversely, most post-Keynesian economists will not state they they know how to calibrate policy in such an exact fashion.

It comes down to policy preferences. Some MMTers may prefer to explore the limits of fiscal policy, to see what can be gotten away with. However, I would be much more cautious; having an immediate inflationary blowup would be easiest way to get a policy reversed. In my view, we are long way away from the "inflation barrier" in most developed countries, so there is room to loosen policies, but I would not want to push my luck.

We need to keep in mind that the Job Guarantee is being marketed as a scheme to both achieve full employment and price stability based on a non-discretionary policy. Turning around and running the economy at the "inflationary limit" with discretionary policy is not exactly consistent that message.

In any event, the inflationary impact of policies depends upon the policy mix, not the level of the fiscal deficit. I would instead worry about the real effects of policy changes, and let the fiscal deficit take care of itself.

Concluding Remarks

 In order to advance, MMT needs to debate these issues with the mainstream. The issue is that debates keep being framed in a way that assumes that the underlying axioms of mainstream economics are correct. In this case, the whole debate rests upon the assumption that the nominal fiscal deficit is a meaningful economic variable, without any reference to the real economic transactions behind the nominal deficit figure.  Once we question those axioms, we see that we are debating the wrong topic. The fact that we keep having to debate axioms is why I am not incredibly optimistic about the state of knowledge actually progressing.

(c) Brian Romanchuk 2018

4 comments:

  1. Hi Brian, great post.

    I wrote on this a while ago you said it in a much better way.

    https://martinfreedman.wordpress.com/2018/02/28/issues-with-issues-in-the-design-of-fiscal-policy-rules/

    ReplyDelete
    Replies
    1. Thanks, it was a good explanation of the logic of the paper.

      Delete
  2. The ideal fiscal rule was set out by Keynes long ago: "Look after unemployment and the budget will look after itself". As to what the debt should be, that question was answered by Milton Friedman and Warren Mosler: it should be zero, except perhaps in emergencies.

    There you are. I managed to sort out a problem in fifty words that Lonergan fails to sort out in a thousand words.

    ReplyDelete
  3. This comment has been removed by a blog administrator.

    ReplyDelete

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.