(Originally published 27 April 2016)
The proposition that the central bank (CB) should grant the government the funds with which to finance fiscal stimulus is, as discussed in the last blog piece, one which has been frequently advanced in recent times. The proponents of such “debt free fiscal stimulus” view it as a more effective and equitable version of the failed quantitative easing (QE) policies pursued to date.
Proponents of debt free fiscal stimulus believe that the fiscal constraints upon government can be eased if CB grants replace borrowing is the means of funding fiscal stimulus. The problem is that this is not true. If we compare CB grants with the alternative option of government borrowing accompanied by the same degree of monetary expansion, debt-free funding does not – notwithstanding the lower level of debt and the interest payments which are thereby avoided – put the government in a position where it is in future less financially constrained and able to spend more or tax less.
The key point to bear in mind in order to understand this is that interest earned by the CB on its holdings of government bonds is paid back by the Bank to the government.
If the CB simply grants X billion to the government, it is creating X billion of new money. The government ‘saves’ the interest it would otherwise have paid. By contrast, if the government borrows X billion from the public, no new money is created. It has to pay interest on the additional debt, and is apparently therefore worse off. However, to deliver the same degree of monetary expansion as under the grant scenario, the CB would need to purchase X billion of bonds (government or private) from the public. The CB would then receive interest on those additional bonds, and this interest income would subsequently be paid over to government, increasing government revenue commensurately. The increase in interest income transferred by the CB to government would be equal (in present value terms) to the interest ‘savings’ derived from debt-free finance, so that the deficit would be the same under either scenario (apart obviously from the treatment of the stimulus spending itself). The government would find itself no better off because it has received money in the form of a grant rather than loan. (This logic does not change if the government borrows some portion of the X billion from the CB rather than the public.)
The government is better off under either scenario than if nothing had happened – i.e. if there had been no stimulus spending and no monetary expansion. However, the financial benefit is entirely due to monetary expansion – in this case, the injection of X billion of new money into the economy – and has nothing to do with whether government is funded by borrowing, or by a grant from the CB.
The point is reinforced if we consider the scenario of debt cancellation which is not accompanied by monetary expansion. If the CB simply cancelled X billion in government debt which it already holds, all that would happen is that the interest no longer paid by government on the cancelled bonds would no longer be repaid to it by the CB. Cancellation of a government’s debt by its own CB, unaccompanied by monetary expansion, is of no benefit whatsoever to a country.
How can it be that the government is no better off if the CB gives it money instead of the government borrowing it, given that government debt would be lower? The reason is that lower government debt is exactly matched by lower CB financial asset holdings. If the CB creates X billion of new money by the conventional route of buying bonds (whether from the public or from the government), the bonds which it acquires can at a later stage be sold to unwind the monetary expansion. The monetary expansion is matched by an increase in CB financial assets. If, however, the CB creates the X billion by simply giving it to the government, it does not acquire any financial assets which can subsequently be sold to withdraw money from the economy.
Debt avoided by taking a CB grant rather than borrowing may for this reason resurface in the future as money which the government has to find to re-capitalise the CB. The greater the use made of CB grants, the more certain it is that future CB re-capitalisation will be required – at least if one is to avoid going down the road to uncontrolled monetary expansion and hyperinflation à la Zimbabwe. (The possible need for government to recapitalise the bank is, by the way, not based on the naïve proposition that CBs cannot be allowed to go into negative equity – i.e. that they can become ‘insolvent’ in the sense of an ordinary firm.)
Not all of the money given by the CB to government today will necessarily need to be returned to it down the track via recapitalisation. If the CB grant is limited and one-off, it may be that none of it will ever need to be. However, the part of the CB’s grant which the government does not return to it subsequently via re-capitalisation is an illusory benefit to government, because – as discussed above – it corresponds to forgone future government revenue from central bank interest income.
In summary, X billion received by the government as a gift from the central bank must, holding the quantum of money creation constant, result in X billion lost (in present value terms) to the government in foregone future revenues and/or CB re-capitalisation requirements. Other things being equal, X billion in deficit and debt avoided by the government today because it has received grant funding from the CB means increased future deficits of an equivalent magnitude (i.e. with a present value of X billion).
The only way government revenue from the CB can be increased (in present value terms) is if the CB creates more money. The benefit to the government is the so-called ‘seigniorage’. The extent to which it government can exploit seigniorage, however, severely constrained because excessive money creation must ultimately generate severe inflation. Argentina – with inflation rates in excess of 30 percent – is another recent example of this.
The difference between CB grants and government borrowing is therefore one of appearance rather than substance – the appearance that fiscal stimulus has been funded without increasing the deficit and with no increase in debt. There is no real impact on the government’s financial position.
The macroeconomic impact will also be the same in either scenario, assuming that the composition of the fiscal stimulus package is identical. The only plausible way in which this might change is if the appearance of lower debt in the grant scenario created the illusion that government finances were in better shape, and this had a positive effect the economy via market psychology. However, this might be outweighed by the adverse impact on confidence of the perception that government is engaged in unsound and risky financing practices.
Finally, the above analysis makes clear, that cancelling government debt cannot make government better off irrespective of whether the debt cancellation is seen as creating scope for additional fiscal stimulus. It is therefore strange that people who should know better should suggest that the contrary is true. Adair Turner, the former chairman of the UK Financial Services Authority is one such, with his strange assertion in his 2015 book Between Debt and the Devil that cancellation of UK government debt by the Bank of England would “reduce the required pace and severity of fiscal consolidation”.
 And, for that matter, on any private sector bonds, such as corporate bonds or mortgage bonds, that it holds.
 In the context of a shared CB like the European Central Bank (ECB), cancellation of the debt of one member country (e.g. cancellation by the ECB of the Greek government bonds it holds) would certainly benefit that country, but only because it would be receiving an implicit transfer from other EU countries.
 The point is instead that the more use is made of CB grants to government, the more the CB’s holdings of government bonds and other financial assets dwindles, and a problem arises when those holdings run out. If the CB needs, in order to implement it preferred monetary policy, to withdraw more money from the economy than it has financial assets left to sell, it can, of course, issue its own bonds to the public. But it has to pay interest on those (which creates money) and will therefore need to issue further bonds in order to absorb that money. This can turn into a Ponzi game requiring unsustainably explosive growth in the issuance of CB bonds. The limits this imposes are what Buiter (2011, 2015) refers to as the non-inflationary comprehensive loss absorption capacity (NILAC) of CBs. It might be objected that the CB has the alternative of raising bank reserve requirements without paying interest on them, but this is simply a tax increase in disguise.
 Given that government will not need to tax any more heavily in future under the borrowing scenario, it would be hard to argue that that Ricardian equivalence (RE) effects would make the debt-free funding option more effective in stimulating the economy. RE assumes if the government adopts fiscal policies which will oblige it in future to increase taxes, private agents will see this and will cut their consumption spending now in order to save the money they will later need to pay in higher taxes. This would reduce aggregate demand and could potentially depress economic activity. In the context under discussion here, the only way RE could come into play would be if private agents misperceived the situation and thought that the choice of the borrowing route implied higher future taxes than the CB grant route. However, to base an RE effect on misperception of this type would be paradoxical given that RE typically assumes that private agents are endowed with something close to perfect foresight.