Fiscal rules and councils

So, why fiscal rules and fiscal councils?¹

There are probably two initial considerations.

One was that democracies tend to run deficits – have a deficit bias – which may be inter-temporally inefficient. I think this is the underlying motivation for all the rules that people advocate for balanced current expenditures over the business cycle.

The other was the experience of governments borrowing for political purposes, e.g. wars, with default as the possible way out of the ballooning debt. This is an old argument, going back to the Enlightenment, e.g. Kant, and probably before. Basically, it is an instrument to limit government’s recklessness. One can think of republicanism (with democratic decision making) as the instrument to constrain authoritarian or imperial fiscal excesses.

Thus, ceilings to deficits and debts to correct for biases and grandiosity respectfully.

The old Enlightenment argument was to constrain governments constitutionally. Alternatively, by requiring the approval of democratically elected representatives (“no taxation without representation”). The relatively recent is the idea to set up fiscal councils, technocratic controllers independent from the government.

The constitutional idea is conceptually sound. The public decides via their representatives who impartially look at both sides of the budget at the constitutional convention. Rawls’ veil of ignorance is not really needed, but it can be invoked for additional insurance of impartiality of the representatives (they do not think only about spending or only about taxation). One way to do that is to ensure that it is extremely hard to change the constitution to ensure that it will be constructed to last forever, which is a thick enough veil of ignorance. So, constitutional constraints on governments are intertemporally consistent.

The fiscal council idea is intertemporally inconsistent because it assumes that the government (or the chancellery) has a deficit bias so it works via the independent fiscal council convincing the public that it should hold the government to account so that it will not over-borrow or under-tax or impose austerity as fiscal policy, i.e. that the government should abide by fiscal rules which safeguard against biases and recklessness.

Both are commitment devices. But, if democracy has a deficit bias, discretion beats rules if those are political in nature, i.e. if they are breaks on what the voters want. So, not just budgetary but also fiscal rule commitments are not binding. E.g. fiscal rules can change from election to election as can budgetary commitments. Indeed, one thinks of democracy as a system in which the electorate can change their mind every four or so years – i.e. can renew their commitments or renege on them. I am summarising a long and extensive debate on rules versus discretion here.

Sometimes an analogy with monetary policy is made, but the independence of the central bank is something that fiscal councils cannot have because they are technocratic, not policy making bodies; they just assess whether rules are being followed and recommend policies for adjustment. They have no instruments to run fiscal policy, unlike the central bank which runs monetary policy (consistent with its statute).

Now, what if one were to think of what a benevolent dictator would do? That would be the technocratic question. One possible answer would be Friedman’s 1948 proposal. With stabilisation of economic activity as the target, one, e.g. Barro, recognises that it may be efficient to time when to tax, and rely on debt in the interim, so deficits and surpluses are efficient if stabilisation of economic activity is the target.

Friedman’s argument was that the level of spending and the design of taxation are to be decided democratically, hopefully not all that often. The government would not borrow (in part not to distort the financial markets) but would run deficits and surpluses financed by the central bank at zero interest rate (again not to distort the financial markets). I put aside monetary and price level implications (the optimal deflation argument). Then he thought about the optimal supply of money, as the government may very well print as much as it wants, or not print enough money, which led him to the fixed growth rate of money supply rule. Which is a kind of fiscal rule indirectly. And then the Taylor Rule won over the Friedman Rule.

What if democracy does not replicate the benevolent dictator? With spending dominating taxation and with increasing reliance on borrowing, then what? Then, fiscal rules are the possible commitment device. Commitment to what? To the stabilisation of economic activity given that it tends to go through business cycles. This is different from the aim of full employment, which sometimes causes confusions in these debates. I will come back to that.

Wren-Lewis and Portes use a simple equation to discuss fiscal rules that might be proposed. They are mostly concerned with taxes because they assume that rising taxes come with efficiency losses. So, assume that decisions on the structure and the level of public spending has been made and taxes have been designed to pay for the desired spending. Then taxes follow the equation which combines the discount rate on future taxes, β, with the interest rate, r, (corrected for the growth rate, g), 1+(r-g). Thus, taxes τ (as a share of GDP):

          τ_t=β(1+(r-g))τ_{t+1                                                                                             (1)}

They take β(1+(r-g))=1. If it were less than 1, given the spending commitments, debt would rise until it would have to be inflated away; if it were higher than 1, taxes would decline to zero because debt would be repaid and spending would be financed out of profits from government investments.

With tax returns increasing and decreasing with the aggregate income, public debt absorbs shocks and cyclical movements. That is the argument for fiscal rules which target budget deficits with no concern for the level of public debt. At least from the stabilisation point of view.

How to determine the level of debt? If an economy is dynamically efficient, that would require the rate of return to capital to be above the growth rate of the economy. That might require that not all the profits are reinvested, but some may be saved in safe assets, e.g. government bonds. That would finance public, mostly current, spending. Then, the level of debt will be determined by the efficiency of the economy. While shocks would induce a random walk increases and decreases of that level over the business cycle.

(Safe interest rate is connected with the rate of return to capital by the appropriate valuation of risk, which was Friedman’s argument for monetary policy staying away from the banking business and banks staying away from the money business, thus zero rate of interest on money. In general, if risk is properly evaluated, interest rate on short-term government bonds, which is the instrument of monetary policy, is equivalent to the risk weighted rate of return on capital. This has to be corrected for regional or cross-border differences in capital to labour ratios with economic and political unions having different returns to capital with common monetary policy. This basically works only within monetary unions, exchange rate policies not being an adequate substitute. See my paper Delaying Integration.)

In 1948 Friedman assumed, if I remember correctly, that dynamic inefficiency, i.e. too high capital to labour ratio with the rate of return to capital below the growth rate of the economy, is unlikely. So, public bonds are not needed. Alternatively, taxes on profits and wealth might substitute for public debt at least when it comes to keeping the capital to labour ratio on the golden path. So, there is no straightforward way to determine the public debt level. But that is not necessary if the deficit bias is eliminated. If for instance β=1, i.e. people behave as if they live forever albeit through their descendants, then only the difference between the interest rate and the growth rate, if it is positive, needs to be repaid.

So, one would expect that more developed economies would have higher and secularly growing public debt levels. Even with the balanced budget over the business cycle rule, e.g. over five years, which is what Wren-Lewis and Portes advocate.

If, however, there is a deficit bias which cannot be corrected for by rules which limit the level of taxation or spending, due to intertemporal inconsistency, public debt rule may do that. Which is obvious, but is the point also made by Lucas and Stokey 1983. But the commitment to public debt level is as much intertemporally inconsistent as the balanced budget rule.

Rather than giving a theoretical argument for that, as it is somewhat trivial, one can look at the public debt ceilings in the USA and the EU. In both cases, those ceilings are not really binding because they are supposed to bite post factum. Now almost routinely the US Congress threatens not to increase the debt level, but that is unfeasible because the spending commitments have already been made. Similarly, in the EU the debt ceiling becomes binding once it has been crossed, so commitments to spending already made need to be reneged on or taxes need to be increased, which tends to be unpopular (this unpopularity is captured by the discount factor β).

The rules mostly apply to current spending, capital spending being different. Perhaps the easy way to see that is to take the example of public and private partnership. Take two examples.

One is infrastructure investment. Assume that it is profitable, but the profits cannot be captured by the private investor due to external effects. So, it is socially but not privately profitable, or rather there is a difference between the social and the private returns. Then, it makes sense to partner a budget centre, taxing authority, and profit centre, equity investors. The budget centre can guarantee profits to the profit centre which will make sure that the social returns to the planned investments are indeed positive, e.g. through proper project selection and cost control. With that, infrastructure investment should grow the economic activity and bring in more tax returns. So, when it comes to capital spending, they should very well pay for themselves. And the government need not take an equity position in them.

(That does not come out from the accounting identity for GDP. It is not that investment will increase GDP ceteris paribus. It may very well reduce it if it is misguided. The accounting identity will be of course always satisfied.)

The other example is the so-called entrepreneurial state. Assume that innovation is not just risky, but indeed an uncertain activity. So, it may make sense for spirited anima, but not for profit seeking investors (the distinction which is at the heart of Keynes’ investment function). They could diversify, but that might not do away with the uncertainty. The public partner, being a budget and not a profit centre, can invest with the view of making up for the losses with the successful projects even if in the aggregate investments in innovation have negative returns. Because social returns to successful innovations may make up for the failed projects and can still be positive and grow the economic activity and thus tax receipts even if private returns in the aggregate remain negative (which I think was Knight’s point).

So, assuming that the projects chosen are not socially wasteful in the aggregate, public investments need to be treated differently from the current spending. They pay, in taxes, for themselves.

One could argue that public investments might crowd out private investments rather than complement them if public investments were directed to competitive industries. There is no point for the government to go after profits, except if there is no other way to reduce the monopolisation of an industry or activity and taxation proves unhelpful (an argument made by Malinvaud long time ago: if there is oligopolistic competition in an industry, government could set up shop and sell at prices equal to marginal costs thus transforming the industry into one with perfect competition). Otherwise, it should complement and thus even crowd in private investment. Even if the government were to borrow commercially.

As taxes are scaled by GDP, the growth rate of GDP is relevant and thus appears in equation 1. If the r-g term is negative, the budget constraint would be lax. I have argued, against stubborn misunderstanding, that this is the case with developing and catching-up economies. Blanchard and Krugman appear to be saying now that this is the normal case in developed economies, like the US. This certainly is the case in the EU and in the euro area at the moment (perhaps with a couple exceptions). That this is the normal state of affairs is debatable, but I will put that aside.

If that is the case, i.e. r being smaller than g, fiscal rules that target taxation or expenditure (they come out to the same thing if treated properly) are inappropriate. The government and the public would still want to spend whatever it is that they think is appropriate, but that might not determine the level of taxation because it makes sense to finance public spending with borrowing. If indeed it were normal that the government can borrow at rates below the growth rate of economic activity, taxation would not be needed at all.

That, however, appears to be an inconsistent claim. The cost of government’s borrowing is as low as it is because it is considered safe. And it is safe because the government can tax, i.e. can coerce people to transfer money to the government. Once it does not tax, its borrowing costs should be no different from those of other debtors – and if the debt is large compared to aggregate income, costs to borrowing will become a constraint. By backward induction, costs of borrowing will be higher already at the beginning of the process. So, the government will not be able to run a Ponzi scheme.

Finally, if the usual case is that of β(1+(r-g))=1, which is to say that taxes change with the elasticity of 1 with GDP (while expenditures have 0 elasticity), no fiscal rule is needed, that being the proper fiscal rule by itself. So, fiscal rules are not needed to constrain the public, but they might be needed to constrain the government. It may suit the government to rely on debt financing rather than on taxes and it may also suit it to pursue a policy of austerity, i.e. of spending cuts in order to run down the debt without raising taxes. So, fiscal rules might be used to guide the game between the public and the government, with the fiscal council as a referee. The problem is that the players themselves are setting the rules and can agree to disregard them or to change them whenever they want, i.e. the time inconsistency problem remains.

Ironically, the case Wren-Lewis takes to be empirically the normal one is the one a Swabian housewife would take: pay-as-you-go and honour your debts. So, in essence, the fiscal rule and the fiscal councils rely for their desirability and implementability on the public holding a Swabian morality. Which also does not see public debt as the one we owe to ourselves as it disregards the distributional issues which are centred on the question of how much each of us owes to others.

Peščanik.net, 29.11.2019.

1. These comments were prompted by Wren-Lewis’s note on Labour vs. Tory on fiscal rules (check the linked paper by him and Portes). We once had a discussion on fiscal rules in theory and practice with Kopits. I forgot what the motivation was for holding the workshop or whatever it was. Anyway, Wren-Lewis is an authority on the subject.