The Myth of Reagan's Debt

[Cross-posted from the new blog at jwmason.org.]

... or at least don't blame him for increased federal debt.

 

Arjun and I have been working lately on a paper on monetary and fiscal policy. (You can find the current version here.) The idea, which began with some posts on my blog last year, is that you have to think of the output gap and the change in the debt-GDP ratio as jointly determined by the fiscal balance and the policy interest rate. It makes no sense to talk about the "natural" (i.e. full-employment) rate of interest, or "sustainable" (i.e. constant debt ratio) levels of government spending and taxes. Both outcomes depend equally on both policy instruments. This helps, I think, to clarify some of the debates between orthodoxy and proponents of functional finance. Functional finance and sound finance aren't different theories about how the economy works, they're different preferred instrument assignments.

 We started working on the paper with the idea of clarifying these issues in a general way. But it turns out that this framework is also useful for thinking about macroeconomic history. One interesting thing I discovered working on it is that, despite what we all think we know,  the increase in federal borrowing during the 1980s was mostly due to higher interest rate, not tax and spending decisions. Add to the Volcker rate hikes the deep recession of the early 1980s and the disinflation later in the decade, and you've explained the entire rise in the debt-GDP ratio under Reagan. What's funny is that this is a straightforward matter of historical fact and yet nobody seems to be aware of it.

 Here, first, are the overall and primary budget balances for the federal government since 1960.  The primary budget balance is simply the balance excluding interest payments -- that is, current revenue minus . non-interest expenditure. The balances are shown in percent of GDP, with surpluses as positive values and deficits as negative. The vertical black lines are drawn at calendar years 1981 and 1990, marking the last pre-Reagan and first post-Reagan budgets.
 The black line shows the familiar story. The federal government ran small budget deficits through the 1960s and 1970s, averaging a bit more than 0.5 percent of GDP. Then during the 1980s the deficits ballooned, to close to 5 percent of GDP during Reagan's eight years -- comparable to the highest value ever reached in the previous decades. After a brief period of renewed deficits under Bush in the early 1990s, the budget moved to surplus under Clinton in the later 1990s, back to moderate deficits under George W. Bush in the 2000s, and then to very large deficits in the Great Recession.

 The red line, showing the primary deficit, mostly behaves similarly to the black one -- but not in the 1980s. True, the primary balance shows a large deficit in 1984, but there is no sustained movement toward deficit. While the overall deficit was about 4.5 points higher under Reagan compared with the average of the 1960s and 1970s, the primary deficit was only 1.4 points higher. So over two-thirds of the increase in deficits was higher interest spending. For that, we can blame Paul Volcker (a Carter appointee), not Ronald Reagan.

 Volcker's interest rate hikes were, of course, justified by the need to reduce inflation, which was eventually achieved. Without debating the legitimacy of this as a policy goal, it's important to keep in mind that lower inflation (plus the reduced growth that brings it about) mechanically raises the debt-GDP ratio, by reducing its denominator. The federal debt ratio rose faster in the 1980s than in the 1970s, in part, because inflation was no longer eroding it to the same extent.

To see the relative importance of higher interest rates, slower inflation and growth, and tax and spending decisions, the next figure presents three counterfactual debt-GDP trajectories, along with the actual historical trajectory. In the first counterfactual, shown in blue, we assume that nominal interest rates were fixed at their 1961-1981 average level. In the second counterfactual, in green, we assume that nominal GDP growth was fixed at its 1961-1981 average. And in the third, red, we assume both are fixed. In all three scenarios, current taxes and spending (the primary balance) follow their actual historical path.

 In the real world, the debt ratio rose from 24.5 percent in the last pre-Reagan year to 39 percent in the first post-Reagan year. In counterfactual 1, with nominal interest rates held constant, the increase is from 24.5 percent to 28 percent. So again, the large majority of the Reagan-era increase in the debt-GDP ratio is the result of higher interest rates. In counterfactual 2, with nominal growth held constant, the increase is to 34.5 percent -- closer to the historical level (inflation was still quite high in the early '80s) but still noticeably less. In counterfactual 3, with interest rates, inflation and real growth rates fixed at their 1960s-1970s average, federal debt at the end of the Reagan era is 24.5 percent -- exactly the same as when he entered office. High interest rates and disinflation explain the entire increase in the federal debt-GDP ratio in the 1980s; military spending and tax cuts played no role.

After 1989, the counterfactual trajectories continue to drift downward relative to the actual one. Interest on federal debt has been somewhat higher, and nominal growth rates somewhat lower, than in the 1960s and 1970s. Indeed, the tax and spending policies actually followed would have resulted in the complete elimination of the federal debt by 2001 if the previous i < g regime had persisted. But after the 1980s, the medium-term changes in the debt ratio were largely driven by shifts in the primary balance. Only in the 1980s was a large change in the debt ratio driven entirely by changes in interest and nominal growth rates.

 So why do we care? (A question you should always ask.) Three reasons.

First, the facts themselves are interesting. If something everyone thinks they know -- Reagan's budgets blew up the federal debt in the 1980s -- turns out not be true, it's worth pointing out. Especially if you thought you knew it too.

Second is a theoretical concern which may not seem urgent to most readers of this blog but is very important to me. The particular flybottle I want to find the way out of is the idea that money is neutral,  veil --  that monetary quantities are necessarily, or anyway in practice, just reflections of "real" quantities, of the production, exchange and consumption of tangible goods and services. I am convinced that to understand our monetary production economy, we have to first understand the system of money incomes and payments, of assets and liabilities, as logically self-contained. Only then we can see how that system articulates with the concrete activity of social production. [1] This is a perfect example of why this "money view" is necessary. It's tempting, it's natural, to think of a money value like the federal debt in terms of the "real" activities of the federal government, spending and taxing; but it just doesn't fit the facts.

Third, and perhaps most urgent: If high interest rates and disinflation drove the rise in the federal debt ratio in the 1980s, it could happen again. In the current debates about when the Fed will achieve liftoff, one of the arguments for higher rates is the danger that low rates lead to excessive debt growth. It's important to understand that, historically, the relationship is just the opposite. By increasing the debt service burden of existing debt (and perhaps also by decreasing nominal incomes), high interest rates have been among the main drivers of rising debt, both public and private. A concern about rising debt burdens is an argument for hiking later, not sooner. People like Dean Baker and Jamie Galbraith have pointed out -- correctly -- that projections of rising federal debt in the future hinge critically on projections of rising interest rates. But they haven't, as far as I know, said that it's not just hypothetical. There's a precedent.  



[1] Or in other words, I want to pick up from the closing sentence of Doug Henwood's Wall Street, which describes the book as part of "a project aiming to end the rule of money, whose tyranny is sometimes a little hard to see." We can't end the rule of money until we see it, and we can't see it until we understand it as something distinct from productive activity or social life in general.

Four Questions about Fiscal Policy

Earlier this fall, I spent a pleasant few days at the 12th Post Keynesian Conference in Kansas City, including a long chat over beers with Robert Skidelsky. In addition to presenting some of my current work, I took part in an interesting roundtable discussion of functional finance with Steve Fazzari, Peter Skott, Marc Lavoie and Mario Seccarechia. Here is an edited version of what I said.

* * *

We all agree that fiscal policy is effective. If output is too low, by whatever standard, higher public spending or lower taxes will cause it to rise. And we all agree that the current level of public debt has no implications for the feasibility or desirability of fiscal policy, at least in a country like the United States. In the wider world that might be a controversial statement but in this room it is not.

It’s not productive to repeat points on which we all agree. So instead, I want to pose four questions about functional finance on which there is not a consensus. These aren’t questions I necessarily have (or expect to hear) good answers to at the moment, but ones that I hope will be the focus of future work. First, the political economy question. Why does the idea of a government financial constraint have such a tenacious hold on both the policy conversation and the economics profession? What function, what interest, does this idea serve? Second, how confident are we about the level of aggregate expenditure that policy should target? Is there a well-defined level of potential output that corresponds to both full employment and price stability? Third, what is the problem that we imagine fiscal policy to be solving? Is it stabilizing of output in the face of “shocks” of some kind, or is it adjusting the long term trend? And what are the sources for the variation in private demand to which policy must respond? Finally, if functional finance means that fiscal policy replaces monetary policy as the main tool for managing aggregate expenditure, what role does that leave for the interest rate?


1. The political economy question.  We all agree that in a country like the modern United States (or the EU as a whole), public budgets are never constrained by the willingness of the private sector to hold the government’s liabilities. There are a number of routes, both logical and empirical, to reach this conclusion, which I won’t repeat here. And yet both policymakers and academic economists are, with few exceptions, committed to the idea that government does face a financial constraint. I don’t think it’s a sufficient explanation that people are just stupid. I was on an earlier panel with Randy Wray, where he quoted Paul Samuelson describing the idea of a government financial constraint as “religion” that has no rational basis but is nonetheless "scares people ... into behaving the way that civilized life requires." [1] Randy focused, understandably, on the first half of that quote, the acknowledgement that a balanced budget is desirable only on ritual or aesthetic criteria. But what about the second part? What is the civilized life that benefits from this taboo?

The political salience of the balanced-budegt myth has been particularly clear in the debt-ceiling fights of the past few years.  You read John Cassidy in New Yorker: “Every country needs to pay its creditors or face ruin.” [2] This is framed as a statement of fact, but it really describes a political project. Creditors need to threaten countries with ruin if they are going to be obeyed. The threat doesn’t have to be real, but it does have to be believed.

The most important political use of the government budget constraint today is undoubtedly in the Euro area, where it seems clear that a central part of the elite motivation for the single currency was precisely to reimpose financial constraints on national governments. This view of the euro project was forcefully expressed by Massimo Pivetti in a 2013 article in Contributions to Political Economy. As he puts it, the ultimate effect of European countries’ renunciation of monetary sovereignty has been the dismantling of the social democratic order.

What is being liquidated is but one of the most advanced experiences of civil coexistence the world has ever known—in fact, the greatest ever achievement of the bourgeois civilization. ... 

Surrender of national sovereignty in the monetary and fiscal fields subscribed by European governments produced a situation of political ‘irresponsibility’, which greatly facilitated their declining commitment to high employment and the redistribution of income, as well as the priority given to reducing inflation, the gradual dismantling of the welfare state, and the privatization drive. ...  [The euro] is an infernal machine: a machine born out of a deliberate continental project to undermine wage earners’ bargaining powers.

Wolfgang Streeck similarly argues that policies that result in rising debt are not the result of rising demands for redistribution and public services, but rather have been supported by the wealthy, precisely because “rising public debt can be utilized politically to argue for cutbacks in state spending and for privati­zation of public services.” You can find similar arguments by Perry Anderson (in The New Old World), Gindin and Panitch, and others. Financial constraint “disciplines” “irresponsible” policymakers — in other words, it makes them responsive to the interests of owners of financial assets. And I would stress the same fundamental point emphasized by Gindin and Panitch — the interest that counts here is not a direct pecuniary interest, defined within the economic system. It is the interest of wealthowners as a class in the perpetuation of a social order based on the accumulation of private wealth.


2. Next, I think we need to interrogate the notion of potential output more critically. The assumption of almost the entire functional finance literature — including my own work — is that there is a well-defined level of aggregate expenditure that policy should be targeting, which corresponds to full employment and full utilization of society's resources. In the standard formula, once we see rising inflation, we know that this target has been reached and there should be no further expansionary policy. In this respect, there is no difference between functional finance and mainstream policy thought. The difference is about the tools used, not about the goal. Most importantly, both policy orthodoxy and functional finance assume that neither inflation nor employment is affected directly by macroeconomic policy, but only via the level of output. So output, inflation and employment can be treated as three indicators for a single target. [3]

It is not obvious, though, why the goals of full employment, price stability and steady output growth should always coincide. Now, in practice it may be that they generally do, or at least are close enough that this is not a big problem. One thing Arjun and I do in our paper is examine this question directly. We compute a number of different measures of the output gap from 1953 to the present. We compare output gaps based on the deviation of current output from trend, the level of unemployment, the level of inflation, and the change in inflation, as well as a measure combining unemployment and the change in inflation that corresponds to the Taylor rule. The interesting thing is that these different measures perform very similarly. The output and unemployment measures fit especially closely, with a simple correlation coefficient of 0.94.  In other words, the Okun relationship between output and unemployment is very stable, and the Phillips curve relationship between output or employment and inflation is also fairly stable. So the statement “When output is above trend, you will see rising inflation; when output is below trend, you will see high unemployment” does in fact seem to be a reliable generalization. (See figure below.)

The figure shows measures of the difference between current output and potential based on (1) trend GDP, as computed by the BEA; (2) the deviation of unemployment from its long-term average; (3) the average of the deviations of unemployment and inflation from their long-term averages; (4) the average of the deviation of unemployment from average and the year-over-year change in inflation; and (5) the year-over-year change in inflation. 

But even if these measures agree with each other for the US over the past 60 years, that doesn’t mean they will agree in other times and places. And in fact, we see that the inflation-change measure does not agree with the others for the post-2007 period, suggesting a much smaller negative output gap. (This is because inflation has stabilized at a low level, rather than continuing to fall.) And even if these measures do generally agree with each other, that doesn’t mean they are right, or that interpreting them is straightforward. In particular, we should ask if hysteresis might not be a more general phenomenon, and that the inflation that comes with output above potential isn't better thought of as an adjustment cost. This brings me to the next question.


3. Is aggregate demand only an issue in the short run, or does it matter in the long run as well? In other words, is the problem to be solved by policy deviations of output around a trend that is determined on the supply side, or is the trend itself the object of policy?

If the former, shouldn't we have a more positive theory about what these "shocks" are that policy is responding to. This has always struck me as one of the weirdest lacunae in mainstream macro. The entire problem of policy in this framework is responding to these shocks, so you would think that a central question would be where they come from, how large they are, whether there are identifiable factors that affect their distribution. But instead the existence of these vaguely defined "shocks" is just the unquestioned starting point of analysis. Now obviously there are reasons for this. Shocks, by definition, are changes in the state of the world that are not due to rational optimization, so if that's your methodology, then "shocks" just means "things I have nothing to say about." (And I have a sneaking suspicion that there is a logical inconsistency between the existence of unanticipated shocks and the idea of intertemporal equilibrium. But maybe not.) But on our side we don't have that excuse. We shouldn't limit ourselves to showing that changes in the government budget position can offset changes in desired private spending. We should try to explain why desired private spending varies so dramatically.

And what if demand matters in the long run, thanks to hysteresis (and what I call anti-hysteresis) in the laborforce, and Verdoorn-Kaldor changes in productivity growth? [4] In that case these questions are even more urgent. And we also have to face the political question that was banished from respectable macro in the 1980s: What is the desirable tradeoff between output and inflation? More broadly, if we can't take a given path of potential output as given, how do we define the goals of macro policy?


4. What is the role of interest rate policy in a functional finance framework, given that it is no longer the primary tool for adjusting aggregate expenditure? On Thursday’s panel, we had three different answers to this question. Arjun and I say that if for whatever reason the public debt to GDP ratio is a concern for policymakers, adjusting the policy interest rate is in general sufficient to stabilize that ratio at some desired level. Peter Skott says that if we have some idea of the optimal long-run capital-output ratio (or perhaps more precisely, the optimal choice of technique), the interest rate can be set to achieve that. And Randy says that we shouldn’t worry about the debt-GDP ratio and that business investment decisions are not very responsive to the interest rate, so its main consequences are distributional. Since there is no social interest in providing a passive, risk-free income to rentiers, the nominal interest rate should be set to zero permanently.


[1] The quote is from an interview with Mark Blaug:  "I think there is an element of truth in the view that the superstition that the budget must be balanced at all times [is necessary]. Once it is debunked [that] takes away one of the bulwarks that every society must have against expenditure out of control. There must be discipline in the allocation of resources or you will have anarchistic chaos and inefficiency. And one of the functions of old fashioned religion was to scare people by sometimes what might be regarded as myths into behaving in a way that the long-run civilized life requires."

[2] The title of the piece is "Why America Needs a Stock Market Crash."

[3] This isn’t strictly correct, since an important component of functional finance in its modern UMKC form is a job guarantee or employer of last resort (ELR) policy. But given the stability of the Okun relationship, employment and output can safely be treated as a single target in practice. The problem is the relationship of employment and output, on the one hand, with inflation, on the other.

[4] As late as the 1980s, people like Tobin took it for granted that the reason that inflationary or deflationary gaps would not continue indefinitely, was that aggregate supply would adjust.

Michael Woodford on the Interdependence of Monetary and Fiscal Policy

(I started writing this post a couple weeks ago and gave up after it got unreasonably long. I don’t feel like finishing it, but rather than let it go to waste I’m putting it up as is. So be warned, it goes on for a long while and then just stops.)

I sat down and read the Michael Woodford article on monetary and fiscal policy I mentioned in the earlier post. It’s very interesting, both directly for what it says substantively, and indirectly for what it says about the modern consensus in economics.

(For those who don’t know, Michael Woodford is a central figure in mainstream New Keynsian macro. His book Interest and Prices is probably the most widely used New Keynesian macro text in top graduate programs.)

The argument of this article is that the question of what monetary policy rule is the best route to price stabilization, cannot be separated from what fiscal rule is followed by the budget authorities. Similarly, any target for the public debt cannot be reduced to a budget rule, but depends on the policy followed by the monetary authorities — though Woodford is not so interested in this side of the question.

This is not a new idea for readers of this blog. But it’s interesting to hear Woodford’s description of the orthodox view.

It is now widely accepted that the choice of monetary policy to achieve a target path of inflation can …, and ought, to be separated from .. the chice of fiscal policy.

Woodford rejects this view — he insists that fiscal policy matters for price stability, and monetary policy matters for the debt-GDP ratio. Most economists think that monetary policy is irrelevant for the debt-GDP ratio, he says,

because seignorage revenues are such a small fraction of total government revenues. … [This] neglects a more important channel … the effects of monetary policy upon the real value of outstanding government debt, through its effects on the price level and upon the real debt service required, … insofar as monetary policy can affect real as well as nominal rates.

There are two deeper issues in the background here, that help explain why orthodox economics ignores the importance of monetary policy for the debt ratio and fiscal policy for price stability. First is the idea, which we can trace from Wicksell through Hayek to Milton Friedman and on to today’s New Keynesians, that the “natural” interest rate in the sense of the interest rate consistent with price stability, must be the same as the “natural” interest rate in the sense of the Walrasian intertemporal rate that would exist in a frictionless, perfect-information economy that somehow corresponded to the economy that actually exists. Few economists are bold enough or naive enough to state this assumption explicitly, but it is fundamental to the project of reconciling orthodox monetary policy with a vision in which money is neutral in the long run. If you want the same interest rate to be “natural” in both senses, it’s a problem if the price-stability natural rate depends on something like fiscal policy, which is not reducible to tastes, technologies and endowments.

Second is the notion of the “long run” itself. For economists, this refers to a situation in which the endogenous variables have fully adjusted to the exogenous variables. This requires a clean (or anyway order-of-magnitude) separation between “fast” endogenous and “slow” exogenous variables; it also requires a sufficiently long time between disturbances.

Woodford, in the passage above, refers to the effects of changes in inflation and interest rates on the burden of the outstanding government debt. This is an important departure from orthodoxy, since in a true “long run” situation, debt would have fully adjusted to the prevailing interest and inflation rates. Woodford, in tune with the practical concerns of central bankers, rejects the ubiquitous methodological condition of modern macroeconomics, that we should only consider fully adjusted long run positions. His whole discussion of public debt, in this paper and elsewhere, rejects the usual working assumption that the existing levels of inflation and interest rates have prevailed since time immemorial. He explicitly analyzes changes in interest and inflation in the context of a historically given debt stock.

Woodford's attitude toward the “natural” rate is more complicated. He certainly doesn’t take it for granted that the price-stability and Walrasian “natural” rates are the same. But a big part of his project — in Interest and Prices in particular — is precisely to develop a model in which they do turn out to coincide.

Let’s continue with the paper. Most economists believe that:

“Fiscal policy is thought to be unimportant for inflation… [because] inflation is a purely monetary phenomenon,” or else because “insofar as consumers have rational expectations, fiscal policy should have no effect on aggregate demand.”

Woodford rejects both of these claims. Even if people are individually rational, the system as a whole can be “non-Ricardian.” By this he means that changes in government spending will not be offset one for one by changes in private spending. “This happens essentially through the effects of fiscal disturbances upon private sector budget constraints and hence on aggregate demand.” For this reason, “A central bank charged with maintaining price stability cannot be indifferent as to how fiscal policy is set.”

Traditionally, the orthodox view of inflation is that it is the result of the money supply growing at a different rate than real economic activity, the latter being independent of the money supply. This does allow for fiscal effects on the price level, but only insofar as public borrowing is monetized. In the familiar “fiscal dominance” scenario, the primary surplus or deficit is fixed by the budget authorities and if the implied issue of public debt is different from the desired holdings of the private sector, the central bank must finance the difference with seignorage. The resulting change in the money supply produces corresponding inflation.

As Woodford says, this is not a useful way of thinking about these issues in real economies, at least in developed countries like the United States. In reality, even when the central bank is subordinate to the budget authorities, as in wartime, this does not take the form of direct monetization of deficits. Rather, “fiscal dominance manifests itself through pressure on the central bank to use monetary policy to maintain the market value of government debt. A classic example is … U.S. monetary policy between 1942 and 1951. … Supporting the price of long-term [government] bonds seems to have been the central element of Fed policy through the late 1940s.” This policy did not affect the price level directly through the money supply, but rather because the target interest rate was too low during the war (although the resulting inflation did not show up until after 1945, due to price controls during the war itself), and too high during in 1948-1950, when the federal government was running large surpluses. In either case, Woodford emphasizes, the causality runs from interest rates, to the price level, to the money supply; the quantity of money plays no independent role.

The basic story Woodford wants to tell is the fiscal theory of the price level. If the stock of outstanding government bonds is greater than the public wants to hold at the prevailing interest rate, then the price of bonds should fall. Normally, this would mean an increase in rates. But if interest rates are pegged — or in other words, if the price of bonds relative to money is fixed — then the price of the whole complex of government liabilities falls. Which is another way of saying the price level rises. Another way of looking at this is that, if output is initially at potential and the volume of government debt rises  with no fall in its nominal price, this must

make households feel wealthier … and thus leads them to demand goods and services in excess of those the economy can supply. … Equilibrium is restored when prices rise to the point that the real value of nominal assets no longer exceeds the present value of expected future primary surpluses.

Of course, this begs the question of why government debt is voluntarily held at a positive price even when there is no reason to expect future primary surpluses. More broadly, it doesn’t explain why anyone wants to hold non-interest bearing government liabilities at all. Woodford could take a chartalist line, talk about tax obligations, but he doesn’t. But even then he'd haven’t of he wouldn't have explained why people hold large stocks of government debt, which by definition is in excess of tax burden.  The natural answer is that government liabilities are a source of liquidity for the private sector. But if he says that, the rest of his argument is in trouble. First, if demand for government debt is about liquidity, then private actors should consider the terms on which other private actors will accept government liabilities. Second, if liquidity is valuable, then real outcomes will be different in a liquidity-abundant world than in a liquidity-scarce one. This is the fundamental problem with the idea that money is neutral. If money were truly neutral, in the sense that the exact same transactions happen in a world with money that would happen in a hypothetical moneyless world, then there would be no reason for money to be used at all.

Despite these serious logical problems, Woodford uses the orthodox apparatus to make some interesting points. For example, he argues that the reason the mid-century policy of fixing a nominal interest rate did not lead to price instability, was because of adjustments in the federal budget position. It is, he says, a puzzle how

a regime that … fixed nominal interest rates was consistent with relatively stable prices for so long. … According to the familiar Wicksellian view summarized by Friedman, an attempt to peg nominal interest rates should lead to either an inflationary or a deflationary spiral. … It is striking that people were willing to hold long-term Treasury securities at 2.5% during the temporary high inflation (25% annual rate) of 1946-1947; evidently there was little fear … [of] an explosive Wicksellian ‘cumulative process.’

Woodford is right that the consistency of fixed nominal interest rates with price stability even after the removal of wartime price controls is a problem for the simple Wicksell-monetarist view. Whether his preferred solution — an expectation of continued federal budget surpluses — is right, is a different question.

Turning to the other side, the dependence of the budget position on prevailing interest rates, Woodford gives an excellent critique of the prevailing notion of a government intertemporal budget constraint (ITBC), in which government spending must be adjusted so that the present value of future surpluses just equals today’s debt. It is widely believed, he says, that government must satisfy such a constraint, “just as in the case of households and firms. It would then follow that fiscal policy must necessarily be Ricardian,” that is, have no effect on the spending choices of rational, non-liquidity-constrained private actors. “It is true,” he continues, “that general equilibrium models always assume that households and firms optimize subject to a set of budget constraints that imply an intertemporal budget constraint, though they may be even more stringent (as it may not even be possible to borrow against all … future income.”

Note the careful phrasing: I’ve noticed this in Woodford’s other writing too, that he adopts rational expectations as a method without ever endorsing it as a positive claim about the world. Of course we shouldn’t talk about intertemporal budget constraints at all, it’s a meaningless concept for private as well as public borrowers, for reasons Woodford himself makes clear.

This is a nice part of the paper, Woodford’s treatment of the “transversality condition.” This, or the equivalent “no-Ponzi” condition, says that the debt of a government — or any other economic unit — must go to zero as time goes to infinity. The reason mainstream models require this condition is that they assume that in any given period, it is possible for anyone to borrow without limit at the prevailing interest rate. This invites the question: why not then consume an infinite amount forever with borrowed funds? The transversality condition says: You just can’t. It is still the case that at any moment, there is no limit on borrowing; but somehow or other, over infinite time net borrowing must come out to zero. This amounts to deal with the fact that one’s assumptions imply absurd conclusions by introducing another assumption, that absurd outcomes can’t happen. Woodford sees clearly that this does not offer a meaningful limit on fiscal policy:

What kind of constraint upon fiscal policy does this [theory] require? A mere commitment to “satisfy the transversality condition” is plainly unsuitable; this would place no constraints upon observable behavior over any finite time period, so that it is hard to see how the public should be convinced of the truth of such a commitment, in the absence of a commitment to some more specific constraint that happens to imply satisfaction of the transversality condition.

What Woodford doesn’t see, or at least doesn’t acknowledge, is that the transversality condition is equally meaningless as applied to private actors. Which means that you need some positive theory about what range of balance sheet positions are available in any given period — in other words a theory of liquidity. And, that the intertemporal budget constraint is meaningless, has no place in any positive economics.

But in any case, even if we accept the intertemporal budget constraint for private units, it is not applicable to sovereign governments since, (1) they are large relative to the economy and (2) they are not maximizing consumption. All that is needed is that someone ends up voluntarily holding the government’s debt. Even if the government is optimizing something, it is not doing so with respect to fixed prices — or fixed output, though Woodford never considers the possibility of unemployed real resources, a rather major limitation of all his work I’ve read.

Woodford notes, reasonably enough, that if the government issues more liabilities than the public wants to hold at the prevailing prices, then the price of government liabilities will fall; “but this is a condition for market equilibrium given the government’s policy, and not a precondition for the government to issue” new liabilities. In this respect, he suggests that the government is in the same position as a company that issues stock, or, more precisely, a company that repurchases shares rather than issuing dividends. (The formal argument that dividends and repurchases are interchangeable, for which Woodford cites Cochrane, is relevant for my “disgorge the cash” work.)

The advantage of this analogy [between the government a share-repurchasing corporation] is that it is clear in the stock case that the equation is an equilibrium condition that determines the share price, … and not a constraint on corporate policies. There is no requirement, enforced in financial markets, that the company generate earnings that validate whatever market valuation of its stock may happen to exist.

The government is different from the company only because prices happen to be “quoted in units of its liabilities.”

As a positive argument this is not useful, for two related reasons. First, it is still essentially a monetarist account of inflation, except with total government liabilities replacing “money.” And second, he deliberately leaves out any discussion of real effects of inflation. This means that he doesn’t give any explanation for price stability is important. More broadly, he doesn’t have any account of the inflation process that links up to real-world discussions. The article purports to be about a central bank following a Taylor rule, but the word “unemployment” does not occur in it. Nor does the word “liquidity”, inviting the question of why anyone holds money in the first place. As I mentioned earlier, this is a larger problem with the whole idea that money is neutral. In this case, Woodford suggests that one can fully explain inflation in a framework in which inflation is costless, and then introduce costs (to motivate policy) without the positive analysis being affected. Interest and Prices does not have this problem — it is carefully constructed precisely to ensure that conventional monetary policy is both welfare-optimizing, and produces an outcome identical to a Walrasian world without monetary frictions. But this isn’t general — the book’s central model is custom-designed to produce just that result.

The question raised by the article is: If both price stability and debt sustainability are functions of both the government budget and monetary policy, why do we have such a strong consensus in favor of stabilizing output solely via the interest rate, and adjusting the government budget position solely in view of the government debt? Woodford admits that in principle, price stability can be achieved in a “bond price-support regime” in which the government budget responds to shifts in private expenditure and the central bank is responsible only for interest payments on government debt. Formally, Woodford acknowledges, this type of regime should work just as well as the conventional “independent” central bank setup. The problem is that in practice

the nature of the legislative process in a democracy makes it unlikely that government budgets can subjected to the same degree of discipline as monetary policy actions. A nontrivial degree of random variation in the equilibrium price level would be inevitable under the price-support regime, both as a result of random disturbances to fiscal policy that could not be prevented, and as a result of inability to adjust fiscal policy with sufficient precision to offset the consequences of other real disturbances. 

There it is. The only argument for central bank independence is an argument against democracy. Woodford continues:

Controlling inflation through an interest-rate rule such as the Taylor rule represents a more practical alternative, both because it is more politically realistic to imagine monetary policy being subordinated wholly to this task, and because it is technically more feasible to “fine-tune” monetary policy actions as necessary to maintain consistency with stable prices.

The claim that interest rates can be adjusted more quickly than budgets is worth taking seriously. Though of course, one way of taking it seriously would be to contemplate arrangements under which taxes and spending could be adjusted more quickly. But look at the other point: The selling point of orthodoxy, says the pope of modern macroeconomics, is that policy can “subordinated wholly” to “controlling inflation.” Look at Europe today, and tell me they aren’t reaping what they sowed.

Alvin Hansen on Monetary Policy

The more you read in the history of macroeconomics and monetary theory, the more you find that current debates are reprises of arguments from 50, 100 or 200 years ago.

I’ve just been reading Perry Mehrling’s The Money Interest and the Public Interest, which  is one of the two best books I know of on this subject. (The other is Arie Arnon’s Monetary Theory and Policy Since David Hume and Adam Smith.) About a third of the book is devoted to Alvin Hansen, and it inspired me to look up some of Hansen’s writings from the 1940s and 50s. I was especially struck by this 1955 article on monetary policy. It not only anticipates much of current discussions of monetary policy — quantitative easing, the maturity structure of public debt, the need for coordination between the fiscal and monetary policy, and more broadly, the limits of a single interest rate instrument as a tool of macroeconomic management — but mostly takes them for granted as starting points for its analysis. It’s hard not to feel that macro policy debates have regressed over the past 60 years.

The context of the argument is the Treasury-Federal Reserve Accord of 1951, following which the Fed was no longer committed to maintaining fixed rates on treasury bonds of various maturities. [1] The freeing of the Fed from the overriding responsibility of stabilizing the market for government debt, led to scholarly and political debates about the new role for monetary policy. In this article, Hansen is responding to several years of legislative debate on this question, most recently the 1954 Senate hearings which included testimony from the Treasury department, the Fed Board’s Open Market Committee, and the New York Fed.

Hansen begins by expressing relief that none of the testimony raised

the phony question whether or not the government securities market is "free." A central bank cannot perform its functions without powerfully affecting the prices of government securities.

He then expresses what he sees as the consensus view that it is the quantity of credit that is the main object of monetary policy, as opposed to either the quantity of money (a non-issue) or the price of credit (a real but secondary issue), that is, the interest rate.

Perhaps we could all agree that (however important other issues may be) control of the credit base is the gist of monetary management. Wise management, as I see it, should ensure adequate liquidity in the usual case, and moderate monetary restraint (employed in conjunction with other more powerful measures) when needed to check inflation. No doubt others, who see no danger in rather violent fluctuations in interest rates (entailing also violent fluctuations in capital values), would put it differently. But at any rate there is agreement, I take it, that the central bank should create a generous dose of liquidity when resources are not fully employed. From this standpoint the volume of reserves is of primary importance.

Given that the interest rate is alsoan object of policy, the question becomes, which interest rate?

The question has to be raised: where should the central bank enter the market -short-term only, or all along the gamut of maturities?

I don’t believe this is a question that economists asked much in the decades before the Great Recession. In most macro models I’m familiar with, there is simply “the interest rate,” with the implicit assumption that the whole rate structure moves together so it doesn’t matter which specific rate the monetary authority targets. For Hansen, by contrast, the structure of interest rates — the term and "risk” premiums — is just as natural an object for policy as the overall level of rates. And since there is no assumption that the whole structure moves together, it makes a difference which particular rate(s) the central bank targets. What’s even more striking is that Hansen not only believes that it matters which rate the central bank targets, he is taking part in a conversation where this belief is shared on all sides.

Obviously it would make little difference what maturities were purchased or sold if any change in the volume of reserve money influenced merely the level of interest rates, leaving the internal structure of rates unaffected. … In the controversy here under discussion, the Board leans toward the view that … new impulses in the short market transmit themselves rapidly to the longer maturities. The New York Reserve Bank officials, on the contrary, lean toward the view that the lags are important. If there were no lags whatever, it would make no difference what maturities were dealt in. But of course the Board does not hold that there are no lags.

Not even the most conservative pole of the 1950s debate goes as far as today's New Keynesian orthodoxy that monetary policy can be safely reduced to the setting of a single overnight interest rate.

The direct targeting of long rates is the essential innovation of so-called quantitative easing. [2] But to Hansen, the idea that interest rate policy should directly target long as well as short rates was obvious. More than that: As Hansen points out, the same point was made by Keynes 20 years earlier.

If the central bank limits itself to the short market, and if the lags are serious, the mere creation of large reserves may not lower the long-term rate. Keynes had this in mind when he wrote: "Perhaps a complex offer by the central bank to buy and sell at stated prices gilt-edged bonds of all maturities, in place of the single bank rate for short-term bills, is the most important practical improvement that can be made in the technique of monetary management. . . . The monetary authority often tends in practice to concentrate upon short-term debts and to leave the price of long-term debts to be influenced by belated and imperfect re- actions from the price of short-term debts." ' Keynes, it should be added, wanted the central bank to deal not only in debts of all maturities, but also "to deal in debts of varying degrees of risk," i.e., high grade private securities and perhaps state and local issues.

That's a quote from The General Theory, with Hansen's gloss.

Fast-forward to 2014. Today we find Benjamin Friedman — one of the smartest and most interesting orthodox economists on these issues — arguing that the one great change in central bank practices in the wake of the Great Recession is intervention in a range of securities beyond the shortest-term government debt. As far as I can tell, he has no idea that this “profound” innovation in the practice of monetary policy was already proposed by Keynes in 1936. But then, as Friedman rightly notes, “Macroeconomics is a field in which theory lags behind experience and practice, not the other way around.”

Even more interesting, the importance of the rate structure as a tool of macroeconomic policy was recognized not only by the Federal Reserve, but by the Treasury in its management of debt issues. Hansen continues:

Monetary policy can operate on two planes: (1) controlling the credit base - the volume of reserve balances- and (2) changing the interest rate structure. The Federal Reserve has now backed away from the second. The Treasury emphasized in these hearings that this is its special bailiwick. It supports, so it asserts, the System's lead, by issuing short- terms or long-terms, as the case may be, according to whether the Federal Reserve is trying to expand or contract credit … it appears that we now have (whether by accident or design) a division of monetary management between the two agencies- a sort of informal cartel arrangement. The Federal Reserve limits itself to control of the volume of credit by operating exclusively in the short end of the market. The Treasury shifts from short-term to long-term issues when monetary restraint is called for, and back to short-term issues when expansion is desired.

This is amazing. It’s not that Keynesians like Hansen  propose that Treasury should issue longer or shorter debt based on macroeconomic conditions. Rather, it is taken for granted that it does choose maturities this way. And this is the conservative side in the debate, opposed to the side that says the central bank should manage the term structure directly.

Many Slackwire readers will have recently encountered the idea that the maturities of new debt should be evaluated as a kind of monetary policy. It's on offer as the latest evidence for the genius of Larry Summers. Proposing that Treasury should issue short or long term debt based on goals for the overall term structure of interest rates, and not just on minimizing federal borrowing costs, is the main point of Summers’ new Brookings paper, which has attracted its fair share of attention in the business press. No reader of that paper would guess that its big new idea was a commonplace of policy debates in the 1950s. [3]

Hansen goes on to raise some highly prescient concerns about the exaggerated claims being made for narrow monetary policy.

The Reserve authorities are far too eager to claim undue credit for the stability of prices which we have enjoyed since 1951. The position taken by the Board is not without danger, since Congress might well draw the conclusion that if monetary policy is indeed as powerful as indicated, nonmonetary measures [i.e. fiscal policy and price controls] are either unnecessary or may be drawn upon lightly.

This is indeed the conclusion that was drawn, more comprehensively than Hansen feared. The idea that setting an overnight interest rate is always sufficient to hold demand at the desired level has conquered the economics profession “as completely as the Holy Inquisition conquered Spain,” to coin a phrase. If you talk to a smart young macroeconomist today, you’ll find that the terms “aggregate demand was too low” and “the central bank set the interest rate too high” are used interchangeably. And if you ask, which interest rate?, they react the way a physicist might if you asked, the mass of which electron?

Faced with the argument that the inflation of the late 1940s, and price stability of the early 1950s, was due to bad and good interest rate policy respectively, Hansen offers an alternative view:

I am especially unhappy about the impli- cation that the price stability which we have enjoyed since February-March 1951 (and which everyone is justifiably happy about) could quite easily have been purchased for the entire postwar period (1945 to the present) had we only adopted the famous accord earlier …  The postwar cut in individual taxes and the removal of price, wage, and other controls in 1946 … did away once and for all with any really effective restraint on consumers. Under these circumstances the prevention of price inflation … [meant] restraint on investment. … Is it really credible that a drastic curtailment of investment would have been tolerated any more than the continuation of wartime taxation and controls? … In the final analysis, of course,  the then prevailing excess of demand was confronted with a limited supply of productive resources.

Inflation always comes down to this mismatch between “demand,” i.e. desired expenditure, and productive capacity.

Now we might say in response to such mismatches: Well, attempts to purchase more than we can produce will encourage increased capacity, and inflation is just a temporary transitional cost. Alternatively, we might seek to limit spending in various ways. In this second case, there is no difference of principle between an engineered rise in the interest rate, and direct controls on prices or spending. It is just a question of which particular categories of spending you want to hold down.

The point: Eighty years ago, Keynes suggested that what today is called quantitative easing should be a routine tool of monetary policy. Sixty years ago, Alvin Hansen believed that this insight had been accepted by all sides in macroeconomic debates, and that the importance of the term structure for macroeconomic activity guided the debt-issuance policies of Treasury as well as the market interventions of the Federal Reserve. Today, these seem like new discoveries. As the man says, the history of macroeconomics is mostly a great forgetting.


[1] I was surprised by how minimal the Wikipedia entry is. One of these days, I am going to start having students improve economics Wikipedia pages as a class assignment.

[2] What is “quantitative about this policy is that the Fed buys a a quantity of bonds, evidently in the hopes of forcing their price up, but does not announce an explicit target for the price. On the face of it, this is a strangely inefficient way to go about things. If the Fed announced a target for, say, 10-year Treasury bonds, it would have to buy far fewer of them — maybe none — since market expectations would do more of the work of moving the price. Why the Fed has hobbled itself in this way is a topic for another post.

[3] I am not the world’s biggest Larry Summers fan, to say the least. But I worry I’m giving him too hard a time in this case. Even if the argument of the paper is less original than its made out to be, it’s still correct, it’s still important, and it’s still missing from today’s policy debates. He and his coauthors have made a real contribution here. I also appreciate the Hansenian spirit in which Summers derides his opponents as "central bank independence freaks."

Functional Finance in Rome and Kansas City

Arjun and I have continued to work on our project on fiscal and monetary policy, which develops the simple -- but strangely overlooked [1] -- point that both the level of output and the trajectory of the public debt-GDP ratio are jointly determined by both the government budget balance and the interest rate set by the monetary authority. (An early stage in our thinking on this was the subject of a post on this blog last year.) Part of our argument is that the fiscal space metaphor is backward -- that the case for countercyclical fiscal policy gets stronger, not weaker, when debt ratios are already high. I'm hoping there will be a working paper version of this soon. But in the meantime, the work is getting presented at various places -- by me at the Eastern Economic Association this past spring, by both of us at the International Economics Association in June, by Arjun at an OECD conference in Rome earlier this week, and by me at the University of Missouri at Kansas City tomorrow. If you're interested, here are our current slides.


[1] This paper by Michael Woodford, which I haven't yet had time to read properly, seems to have a similar starting point but ends up somewhere quite different.

Long Run Growth and Functional Finance

Tom Michl has some comments -- insightful as always -- on the previous post on functional finance. The key point he raises is that we can't treat the long-run growth rate as given exogenously. Policy that targets current output will also have effects on the long run trajectory that have to be taken into account. He writes:

I've become convinced that the real interest rate belongs in the investment equation, g(r), which means that g-r is not something we can just set greater than zero to solve the problem of fiscal deficits. Also that fiscal policy, e.g. the debt ratio, affects the long run position of the IS curve (assuming interest payments go to rentiers who spend them on consumption), so it affects the r that stabilizes inflation.  

The question of exo/endogenous growth is important because a given growth target puts constraints on the feasible fiscal policy, or given fiscal policy, on the ability of a monetary authority to set the inflation-neutral interest rate. 

This is something I've heard from other smart people in response to these arguments. [1] I certainly agree with Tom and everyone else that a complete story cannot just take g as given. And I agree that there should be some systematic relationship between the liquidity conditions that we summarize as the interest rate and demand conditions, and the long term growth of income. But it doesn't seem so straightforward to show that this relationship will be a constraint on the fiscal position.

There are two (sets of) channels: The one Tom mentions here, from financial conditions to investment to growth, and the other, on the supply side from the output gap via the labor supply (hysteresis) or technological change (Verdoorn's law) to growth. I think the second kind of channel is very important, but it doesn't create any issues for a functional finance position. It just means that we should define a higher level of output as "full employment" or "price stability." So let's focus on the first channel.

We think of investment as additions to the capital stock. Then we have g = s/c - dk, where g is the growth rate, s is the average savings rate, c is the incremental capital-output ratio, d is the depreciation rate and k is the average capital-output ratio. This is just accounting. As we know, this accounting relationship is often used to develop the idea of knife-edge instability. But that's never seemed right to me (and I don't think it's what Harrod intended with it.) s is the average savings rate, so in a Keynesian framework it will be a negative function of output. So what this equation is telling us is that if we need to achieve a given growth rate, and the capital-output ratio is fixed, then the output gap will have to adjust so as to get s to satisfy this equation. This is the adjustment that policy is allowing us to avoid. Fiscal policy raises or lowers average s at a given level of income. Monetary policy perhaps raises or lowers s also; more conventionally it is supposed to change the desired capital-output ratio c. 

So from my point of view, it is not quite correct to say that growth is a function of the interest rate. Rather, variation in the interest rate allows us to reconcile full employment with our chosen growth rate, whatever it may be. 

Now, if we think that interest rates act through the capital-output ratio, then we need that as an additional degree of freedom if we want to combine full employment, our chosen growth rate, and a stable debt-income ratio. As it happens, Peter Skott and Soon Ryoo presented a paper at the Easterns where the requirement to achieve a target capital-output ratio meant that monetary policy was not available to close the output gap, requiring the use of fiscal policy. The additional degree of freedom is supplied by allowing the debt-GDP ratio to evolve freely.  This isn't a problem in this case. In their model, fiscal policy works through its additions to the stock of assets available to private wealth-owners, not through the flow of demand for currently produced goods and services. So the public-debt GDP ratio will automatically converge to whatever level satisfies the private sector's demand for net wealth above the capital stock.

So Peter's paper, I think, addresses Tom's point. Yes, fiscal policy affects the IS curve. Namely, it moves the IS curve to wherever it needs to be to get full employment at a given growth rate, as long as we are willing to let either the ratio of either capital or debt to output vary endogenously.

The bottom line is that when you move from the short run to the long run you do have to think about growth but that does not necessarily impose any additional constraint. First, if monetary policy operates through the savings rate, then the price stability target already implicitly means price stability at a given growth rate. So the model works the same regardless of what you think the growth rate is or should be. If monetary policy works through the capital-output ratio, then we can no longer say that, but in that case the capital-output ratio itself provides an additional degree of freedom. Only if we impose both a given growth rate and a given capital-output ratio do we possibly foreclose the option of setting r at whatever value is consistent with price stability and a constant debt ratio. And even then, I emphasize "possibly," because it depends on what you think happens to maintain the constraint. It seems to me the most natural answer is that at some point the desired capital-output ratio becomes insensitive to the interest rate, so, assuming that savings are also insensitive, then full employment cannot be reached at our target growth rate through monetary policy alone. In that case, fiscal policy becomes mandatory. This is where Keynes ended up, more or less, and also the implication of Peter's paper. But this doesn't give any argument for why interest rates cannot stay arbitrarily low, it just says that even very low interest rates won't be sufficiently expansionary to get us to full employment at low growth rates and that fiscal policy or some equivalent will be required as well. Which, as they say, is where we came in.

(I realize that this post probably will not make sense unless you're already having this conversation.)

[1] Last year's functional finance post is, thanks to my coauthor Arjun Jayadev, evolving into an academic article; I presented a version at the Eastern Economic Association a few days ago. This was one of the main comments there.

A Response to Tom Palley

Tom Palley wrote a strongly worded critique of Modern Monetary Theory last year, which got a lot of attention int he world of heterodox economics. He has just put up a second piece, MMT: The Emperor Still Has No Clothes, reiterating and extending his criticisms.

Palley is a smart guy who I've learned a lot from over the years. But this is not his best work.

By way of preliminaries: MMT consists of three distinct arguments. First is chartalism, the claim that the value of money depends on its acceptability to settle tax obligations. This goes back to G. F. Knapp. Second is functional finance, the claim that government (conceived of as a consolidated budget and monetary authority) seeks to adjust the budget balance to achieve full employment, it can never be prevented from doing so by a financing constraint. This goes back to Abba Lerner and Evsey Domar. And third is the employer of last resort (ELR), a proposal for a specific form of spending to be adjusted under the functional finance rule. This seems to be an original contribution of Warren Mosler goes back to Hyman Minsky.

Personally, I find it useful to set aside the first and third of these arguments and focus on the second, functional finance. My own attempt to restate the functional finance claim in the language of contemporary textbooks is here. In my view, the essential difference between functional finance and orthodoxy is that the assignments of the interest rate and budget instruments are flipped. Instead of setting the interest rate to keep output at potential and setting the budget balance to keep the debt on a sustainable path, we assign the budget balance to keeping output at potential and the interest rate to debt sustainability.

Palley wants to knock down all three planks of MMT. What are his objections to functional finance?

(1) In the absence of economic growth, government deficits will lead to inflation regardless of the output gap. This claim is asserted rather than argued for. [1] It's not clear what the relevance of this point is, since he agrees that deficits are not inherently inflationary when there is positive growth. Perhaps more importantly, this is a rejection not just as of MMT but of almost all policy-oriented macroeconomics, mainstream and heterodox. Whether you're reading David Romer or Wendy Carlin or Lance Taylor, you're going to find a Phillips curve that relates inflation to current output. There are plenty of disagreements about how expected inflation comes in, but nobody is going to include the budget balance as an independent term. In his eagerness to debunk MMT, Palley here finds himself reasserting Milton Friedman-style monetarism.

(2) If we assume an arbitrary floor on spending and an arbitrary ceiling on taxes, then it may be impossible to achieve both full employment/price stability and a sustainable debt path with interest rates fixed at zero. Yes, this is true. But it proves too much: If we impose arbitrary constraints on tax and spending levels then there is no guarantee that we can achieve debt sustainability and price stability with ANY interest rate. At any given interest rate, there is minimum primary balance that must be achieved to keep output at potential. There is also a minimum primary balance that must be achieved to keep the debt-GDP ratio constant. There is no a priori reason to think the first balance is higher than the second. So Palley's argument here could just as well be a proof that there is nothing government can do to prevent public debt from rising without limit. In any case, these arbitrary limits on taxes and spending are not a feature of standard macro models (including Palley's own models elsewhere), so it's hard to justify bringing them in here.

(3) MMT lacks a theory of inflation. On the contrary, MMT has exactly the same theory of inflation as orthodox macro: High or rising inflation is the result of output above potential, disinflation or deflation the result of output below potential. In other words, MMT is consistent with a standard Phillips curve of the same kind Palley (and almost everybody else) uses. To be fair, the Wray and Tymoigne piece Palley is responding to is not as clear as it might be on this point. But Palley is supposed to be writing a critique of MMT, not just of one particular article. And people like Stephanie Kelton state unambiguously that MMT shares the orthodox output-gap story of inflation; see for instance slide 13 here.

(4) MMT doesn't work in open economies because it requires persistent interest rate differentials between countries. Palley claims that the idea that you can hold interest rates in a given country at zero indefinitely is inconsistent with covered interest parity, a "well-established empirical regularity" that "states there is no room for systematic arbitrage of cross-country interest rates." It seems that Palley has confused covered and uncovered interest parity. CIP is indeed well established empirically, but it only says that there is no arbitrage possible between the spot and forward markets for a given exchange rate. It does not rule out interest-rate arbitrage in the form of the carry trade, and so does not have any implications for the viability of MMT. If UIP held, that would indeed rule out a persistent zero interest policy in the absence of an equally persistent currency appreciation. But UIP, unlike CIP, gets no empirical support in the literature. Japan has sustained the near-zero interest rates that Palley says are unsustainable for 15 years now, and in general, persistent interest rate differentials without any offsetting exchange rate movements are ubiquitous. Furthermore, if financial openness rules out a policy of i=0, then it equally rules out the use of interest rates as a tool for demand management. The best thing you can say about Palley here is that he is parroting orthodoxy; otherwise he is thoroughly confused.

There is one thing that Palley is right about, which is that substantially all of MMT can be found in the old Keynesian literature. This isn't news -- in the same Stephanie Kelton slideshow linked above, she goes out of her way to say that there is nothing "modern" about MMT. And so what? There's nothing wrong with updating insights from the past. In my opinion, most useful work in economics is about pouring old wine in new bottles.

I don't write this from a position within MMT. I tend to feel that the genuine insights of Lernerian functional finance are obscured rather than strengthened by basing them in a chartalist theory of money. It's fine if Tom Palley disagrees with our friends at UMKC and the Levy Institute. But he needs to put down the blunderbuss.


[1] In fact it's a bit hard to understand what Palley is claiming here. First he says that "money financed deficits" must lead to inflation in a static economy, even with a zero output gap. He adds in a footnote that money-financed are not inflationary in a growing economy; in that case, for price stability "the high-powered money stock must grow at the rate of growth." Then when he writes down a model, this has become the condition that government budget must be balanced, which is something different again. Also, I must say I can't help wrinkling my nose a little, when I read about the "stock of high-powered money," at the smell of a musty antique.


EDIT: Thanks to Daniele Santolamazza for correcting me on the origins of the ELR proposal.

Functional Finance and Sound Finance

Introduction

Anyone who who has been following debates on fiscal policy over the past few years will have noticed that, among those who think fiscal policy can be effective, there are two distinct camps. There is a minority who think that fiscal policy is not subject to a budget constraint; that is, that as long as a government borrows in its own currency, its existing liabilities never limit its ability to adjust taxes and spending to bring the economy to full employment. And there are the majority who think that governments are subject to a binding budget constraint; that is, that while adjusting spending and taxes can in principle be used to bring about full employment, it may be impossible or undesirable to do so when the level of government debt is already high. In this view, maintaining full employment should be left to monetary policy. Following Abba Lerner, I call the first position "functional finance" and the second position "sound finance."

I believe there are important differences between these two positions. But I also believe that these differences have not been clearly articulated, and as a result these debates between them been unproductive. It is my view that there are no important differences in terms of economic theory between the two positions. A perfect application of a functional finance policy rule and of a sound finance policy rule are indistinguishable. The difference between the camps is with respect to policy errors -- which errors are most likely, and which are most costly.


Alternative Policy Rules

The starting point is the idea of instruments, which are variables directly controlled by the policymaker; and targets, which are the variables the policymaker wants to set at some level but cannot control directly. When the target variable is not at its desired level, the policymaker adjusts one or more instruments to try to bring it there. This creates relationship between the current level of the target and the chosen level of the instrument. We call this relationship a policy rule. Both functional finance and sound finance represent policy rules in this sense. Tinbergen's Rule says that for policy rules to be successful (in the sense that all targets converge to their desired levels), there must be at least as many instruments as targets. One policy lever cannot be relied on to achieve two separate outcomes.

We have two instruments in macroeconomic policy: the government budget balance, and the central bank-controlled interest rate. What are our targets?

At first glance, full employment and price stability appear to be two separate targets. But in fact, both Lerner's functional finance and the sound finance of modern textbooks agree that inflation is the result of demand-determined expenditure departing from a technologically determined level of potential output. Less than full employment means falling inflation, or deflation; overfull employment means high or rising inflation. So full employment and price stability are not two separate targets, they are two ways of describing the same target.

Both camps agree that we can identify a unique target level of output, and they agree on what that target should be. They also agree that output rises with higher government deficits, and falls with higher interest rates. So when interest rates are too high, or budget deficits too small, we will see unemployment (and perhaps deflation); when interest rates are too low or deficits are too large, we will see inflation (and perhaps bottlenecks and rising relative prices of factors in inelastic supply).

This consensus is shown in Figure 1. The full employment locus shows all the combinations of interest rates and fiscal balances that are compatible with full employment and price stability. A fall in private demand will require a rise in the deficit and/or a fall in interest rates to maintain full employment, so it will shift the full employment locus down and to the left. Similarly, a rise in private demand will shift the locus up and to the right. But for any level of private demand, with two instruments and only one target, there are an infinite number of combinations that achieve full employment.

(It is convenient to think of the fiscal balance on the horizontal axis as the primary balance, that is, the balance net of interest payments. So we are implicitly assuming that interest payments do not raise aggregate demand. It is also convenient to think of the interest rate as the real rate, that is, net of inflation. It would be straightforward to incorporate the effects of interest payments and inflation into the story, but would not change it in any interesting way.)

The first point of disagreement is what to do at a point like a. Output is below potential, but which instrument should be used to raise it? Functional finance says, the fiscal balance: government spending should be raised (or taxes should be lowered), moving the economy to the left, until we reach the full employment locus. The modern sound-finance consensus says that the interest rate should be lowered, moving the economy downward to the full employment locus. Both agree that government should do something to raise output. The disagreement is over which instrument to use.

Whichever instrument is used to keep output at potential, there is one instrument left over for some other target. The logical candidate is the sustainability of government debt.

We've discussed the math of government debt dynamics quite a bit on this blog. (See here and here and here and here.) The important thing for our purposes is that the long-run trajectory of the debt-GDP ratio depends on the primary balance, the interest rate on government borrowing, and the growth rate of GDP. If we write the ratio of government debt to GDP as b, and the primary deficit as a share of GDP as d, then for a given deficit, the equilibrium condition is b=d* 1/(g-r), where g is the average or expected growth rate of GDP over the period of interest. So for a given debt-GDP ratio b, the primary deficit required to hold it constant is d = b(g-r). (This is all just accounting; it does not depend on any economic assumptions.) It's evident that, if we take the growth rate as exogenous, then for any given debt-GDP ratio there is a set of r, d combinations for which the debt-GDP ratio is constant. We can represent these values graphically in Figure 2. The dotted horizontal line is the growth rate. The diagonal line is the constant debt ratio locus. With a deficit or interest rate above the diagonal line, the debt-GDP ratio will rise; below, it will fall.



Note that the slope of the diagonal depends on the starting debt-GDP ratio -- the higher it is, the shallower the slope will be. With no government debt, the line is vertical at the primary balance = 0 axis. So in any period in which the economy is at a point above the debt-sustainability locus, the diagonal rotates clockwise; in any period in which the economy is below the debt-sustainability locus, the diagonal rotates counter-clockwise.

What happens if the economy is off the constant-debt locus? It depends. In the area marked A (everything above the heavy line), the debt-GDP ratio rises without limit. In B, the debt-GDP ratio rises but converges to a finite value. In C the ratio falls to a finite value. In D, the debt-GDP ratio falls to zero and the government then accumulates a positive asset position, which eventually converges to a finite fraction of GDP. Finally, in area E the debt-GDP ratio falls to zero and the government then accumulates a positive asset position that rises without limit as a share of GDP. (If you are unconvinced we can go through the math.) Since the government budget constraint is normally taken to be the condition that debt-GDP ratio not rise without limit, we can ignore the distinctions between cases B through E and regard the heavy line as the government budget constraint.

We then combine this constraint with the full employment locus to give Figure 3.



Now we have two instruments and two targets. Or rather, one and a half targets: Since there is nothing special about the current debt-GDP ratio, we don't need it to stay constant; we just need it not to go to infinity. So we don't need to be on the debt-sustainability curve, we need to be on or below it. Point b, which satisfies the budget constraint exactly, is fine, but so is anywhere on the full employment locus below and to the right of b.

The functional finance-sound finance divide is just this: Functional finance says the fiscal balance instrument should be assigned to the full employment target and the interest rate instrument should be assigned to the debt sustainability target. Sound finance says the interest rate instrument should be assigned to the full employment target and and the fiscal balance instrument should be assigned to the debt sustainability target.

Functional finance and sound finance agree that the economy should be at a point like b. If policy were executed perfectly, the economy would always be at such a point, and there would be no way of knowing which rule was being followed. Since both target should always be at their chosen levels, it would make no difference -- and be impossible to tell -- which instrument was assigned to which target. The difference between the positions only becomes apparent when policy is not executed perfectly, and the economy departs from a position of full employment with sustainable public debt.

Consider a point somewhere above b, where we are have high unemployment but the debt-GDP ratio is rising without limit. What to do? Both orthodoxy and Lernerism want to get the economy back to a point like b, but they disagree on how.

In the sound-finance view, the interest rate instrument is committed to the output target. This means we must use the fiscal balance instrument free to meet the debt sustainability condition. This is how policy is normally discussed: An unsustainable upward trajectory in the debt position requires the government balance to move  toward surplus. In this case, that means that the government must cut spending or raise taxes, despite the fact that demand is already too low. Under Lernerian functional finance, on the other hand, the fiscal balance is committed to the output target, so the rule calls for higher deficits even though the debt position is already unsustainable. It is then the responsibility of monetary policy to adjust to maintain debt sustainability.

These alternatives are shown in Figure 4. The right-hand trajectory from c to b is the orthodox path. The left-hand trajectory is the Lernerian path. Implicit in the orthodox path is the idea that deficits must be brought down first, meaning a substantial period of high unemployment and output below potential; only once debt is on a sustainable path can interest rates be reduced to move back toward full employment. While the Lernerian path says in effect: If government debt is rising out of control, the central bank should intervene to force interest rates down to a level where the debt is sustainable. Then, if the resulting liquidity raises expenditure above the full employment level, you can subsequently raise taxes or cut transfers to bring demand back down.



Orthodoxy says that budget problems must be addressed fiscally. But this is true only on the implicit assumption that the interest rate is not available as an instrument to target debt sustainability. Sound finance's policy rule is a Taylor-type rule for monetary policy, combined with a long-term government budget position that satisfies the debt-sustainability constraint at that interest rate. Functional finance's policy rule: (1) fix the interest rate at a level at or below the expected growth rate (maybe even zero); (2) adjust transfers and taxes until output is at the full employment/stable prices level. The claim that fiscal policy must be subject to a budget constraint, comes down to the claim that the central bank cannot or will not keep r sufficiently low to make the full-employment fiscal position sustainable.

Why is there such disagreement about which instrument should be assigned to which target? It seems to me that the most important argument from the sound finance side is that elected governments cannot be trusted with the instrument of discretionary fiscal policy. They will not set taxes and transfers to bring aggregate demand to the full employment level, but will choose a higher, inflationary level of demand. Only independent central banks can be trusted to bring output to its socially optimal level. In this sense, the functional finance-sound finance divide is not a debate about economic theory, but about politics and sociology.

There are also more specifically economic disagreements. The sound finance side is more confident than the functional finance side about how quickly and reliably a change in interest rates will affect output. If there is a long lag between the change in the instrument and its effect, hitting the target requires accurate prediction of the state of the economy farther into the future. The existence of the ZLB reinforces this concern, since it is really just a special case of interest-inelasticity. (The statement "output does not respond strongly to any feasible change in interest rates" is equivalent to the statement "the interest-rate change needed to achieve a strong output response is not feasible.") The functional finance side also tends to see a greater social cost in falling below full employment than rising above it, while the sound finance side tends to see the costs as symmetrical.

That is the framework. Now consider some modifications and special cases.


Extensions

A natural objection to the functional finance view is that it may not be possible for the central bank to maintain interest rates low enough to keep debt sustainable. If we live in a world of high capital mobility and our government's liabilities are close substitutes for liabilities elsewhere in the world, then the private sector will not hold them if their yield is too much lower. In this case -- which is not unrealistic for small, open countries -- the interest rate ceases to be a policy variable. This is shown in Figure 5, where r* is the exogenous work interest rate.



At a point like d in the figure, the public debt is stable but output is below potential. A move toward a primary deficit would raise output but put the debt on an unsustainable path. This case is not inherently implausible -- one would need to think carefully about the concrete assumptions it embodies -- but it is important to recognize that it rules out sound finance as well as functional finance. If the interest rate is set exogenously at the world level, it cannot be used to stabilize public debt or to stabilize output. An additional instrument is needed; the exchange rate is the natural choice. Since the exchange rate cannot straightforwardly be used to achieve debt sustainability, in this case there is a natural argument to switch the assignment of fiscal policy to debt sustainability and achieve full employment via the exchange rate.

Another possibility, which has been getting increasing attention recently, is that very low interest rates are destabilizing for the financial system. (I have criticized this idea before, but I don't think it can be ruled out definitively.) Then we have another condition to satisfy, a asset price stability condition. Like the debt sustainability condition, this is asymmetrical, it doesn't have to be satisfied exactly. But this one is a floor on interest rates rather than a ceiling. The is shown in Figure 6. Here, the asset price stability constraint does not initially prevent achieving both the other targets: As in Figure 3, point b initially satisfies all the constraints, as does any other point along the full employment locus below and to the right of it, down to the dotted line.


But what if a fall in private demand shifts the full employment locus far to the left? Here there is an important difference between the sound finance and functional finance rules. The functional finance rule says that the fall in private demand requires the government budget to move toward deficit. That is, we move left from b to the new full employment locus. This may in turn require a fall in interest rates, if the higher deficits would otherwise put the public debt on an unsustainable path. But public debt sustainability never requires an interest rate below the long term growth rate. So, since it is not plausible that the minimum interest rate compatible with asset price stability condition is greater than the growth rate, the possibility of asset bubbles should not limit the application of the functional finance rule.

The sound finance rule, on the other hand, says that the response to a fall in private demand should be a reduction in the interest rate. In other words, faced with the fall in private demand shown in the figure, we should move downward from point b  to the new full employment locus. Now there is the possibility that the required interest rate is incompatible with asset price stability. (In some views, this is precisely what happened a decade ago, setting the stage for the housing bubble.) This becomes an argument for setting interest rates higher than the conventional policy rule implies, even at the cost of higher unemployment.

Formally, the ZLB is identical to the asset price stability condition: both set floors to allowable interest rates. It is curious that, while concern with the ZLB and with the destabilizing effects of low interest rates often come from opposite political positions, they are -- at least in this framework -- equivalent in their implications for policy. Both are arguments for a reliance on fiscal policy to offset fall in private demand in general, rather than waiting for the floor to be reached -- that is, for some form of functional finance.

Finally, consider the case where the fiscal balance is exogenously fixed, as shown in Figure 7. I think this is the case most critics of functional finance have in mind. If the budget authority, for whatever reason, is committed to tax and spending policies corresponding to a primary deficit, there may be no interest rate that can deliver both debt sustainability and full employment. The central bank must choose one. If it chooses debt sustainability, we have a situation known in the literature as "fiscal dominance." The central bank must increase its liabilities as needed to finance the government deficit, even if that results in aggregate demand rising to inflationary levels. This is the situation at point e.



It is important to stress that Figure 7 is not what is advocated by functional finance. There is an understandable but unfortunate confusion between the claim "deficits can be at whatever level is needed to reach full employment" and "deficits can be at whatever level you want." Functional finance says the former, not the latter. A functional finance rule would call for the government to raise taxes or cut spending at a point like e -- not to balance the budget, but to eliminate the inflation. The practical problem for functional finance supporters is to convince skeptics that such a rule will be followed by an elected government.


Conclusion

Advocates of functional finance say that a government that borrows in its own currency never needs to adjust its taxes or spending on account of its current deficit or accumulated debt. The fiscal balance can always be set at whatever level is needed to achieve full employment. Their sound-finance critics reply, "It's true that a deficit will raise current output. But over the long run you need a primary surplus to ensure that the government stays on its budget constraint. If the central bank is forced to monetize the debt instead, you will have runaway inflation."

The critics are correctly describing the situation in Figure 7, where it is true that the government budget position has been set without regard for debt sustainability, the central bank is monetizing the debt (this is simply another way of describing holding interest rates low enough to maintain a stable path for government liabilities), and there is uncontrolled inflation. But the inference the critics draw from this possibility -- that fiscal policy must target debt sustainability -- is not correct. The correct inference is that at least one of the two instruments must target debt sustainability, and at least one must target full employment. The problem in Figure 7 is that budget balance is being set without regard for either condition -- that is, it is in violation of both policy rules. Either the sound finance rule, or the functional finance rule, or any linear combination of the two, would ensure that the economy does not remain at a point like e but instead converges to one like b in Figure 3.

The debate between sound finance and functional finance cannot be resolved as long as they are framed in terms of what kind of rule is feasible in principle, and what outcome results when it is followed exactly. The disagreement is about what kinds of rules policymakers can be expected to adhere to in practice, and about the relative costs of different policy errors.


[This post was inspired by this talk by Brad DeLong, and by some comments by Nick Rowe which I cannot locate now.]