Chris Sims is a deeply thoughtful economist and his paper at the recent Jackson Hole conference has rightly received a lot of attention, notably from Gavyn Davies, in the FT, and Paul Krugman.
It is simultaneously deeply insightful and, I believe, fundamentally flawed. The main insight is that dramatic central bank balance sheet expansions under QE run the risk of generating very large ‘losses’, or large wealth transfers to the private sector. This makes clear that rejecting policies which generate a negative net interest income for the central bank because they are loss-making, and have magically become ‘fiscal’, while simultaneously engaging in large-scale QE, makes no sense. Examples of this stark inconsistency are the ECB’s arbitrary rule that it will not buy bonds yielding less than its deposit rate, or the Bank of England’s Ben Broadbent stating that if interest rates on the TFS went sharply negative and below base rates – ‘we would ourselves be making losses’, as if this was relevant.
One clear implication from Sims’s observations is that the dividing line between QE and helicopter drops is not the impact on CBs balance sheets. The risk of real balance sheet losses under large QE programmes is very high, in part because the impact on final demand is negligible. The dividing line between helicopter money and QE is the form of transfer and the probability of success.
Think of it like this: instead of the central bank writing checks to households, in the closest approximation to Friedman’s parable, they decide to buy bonds issued by households at two or three times their market values – immediately incurring mark-to-market losses and allowing households to buy back their bonds, making net gains in their cash holdings. Overpaying for assets is economically equivalent to making a transfer.
So Sims is absolutely right to point out that if your concern is central bank balance sheet losses – you should oppose QE. By this standard, QE is far riskier than helicopter drops – which would limit ‘losses’ and would be far smaller in scale. If tax rebates are the analogue, 3% of GDP in cash transfers to households is likely to have more impact than 30% of GDP in QE at the near-zero bound.
This focus on the probability of losses on CBs balance sheets is lucid and important. But I think Sims then makes three related errors: 1) he worries about these ‘losses’; 2) he assumes fiscal dominance; 3) because he assumes fiscal dominance he thinks expectations about fiscal policy are holding us back.
I have written before about why CB balance sheets don’t matter – and I don’t agree that tiered reserves, reserve requirements, and prudential measures can simply be dismissed as ‘taxing’ banks – because they are specifically aimed at affecting the relationship of base money, credit growth, and inflation. Sims, like many others, would benefit from outlining a clear distinction between monetary and fiscal policy. In summary, the central bank’s accounting ‘equity’ never matters – what (always) matters is whether the CB can control the stock of reserves, and the effect of reserves on inflation. There are many effective tools for doing so, with or without ‘equity’.
But what is most striking is that the entire premise of Sims’s argument is an institutional contingency: that the fiscal authority dominates the monetary authority. This dates back to Sargent and Wallace, fathers of FTPL. Fiscal dominance is not a conclusion Sims draws, it is an assumption he makes. We know that fiscal dominance is contingent because the ECB is politically, legally, economically dominant in Europe. None of Sims’s framework applies to Greece: when the ECB changes interest rates, the rate of interest on Greek government bonds do not move in lock-step. Perceptions of Greek solvency determine its cost of debt.
Sims correctly asserts that monetary policy has ‘fiscal effects’, but then falls for the Bundesbank Fallacy. A fiscal ‘effect’ is not a fiscal ‘policy’ – almost all economic activity has a fiscal effects. The fiscal effects of conventional monetary policy – to the extent that it stabilises growth, and limits the severity of recessions – are vast.
There is a clear dividing line between fiscal and monetary policy. Monetary policy involves base money and is implemented by the central bank. It has an institutional component and a qualitatively distinct, theoretical, component – money. Part of the problem with the fiscal theory of the price level is it collapses the distinction between money and government debt, emboldened by the widespread adoption of interest on reserves (and ignoring the implications of tiered reserves).
But money is not debt, as Simon Wren-Lewis and Willem Buiter point out. Governments in the Eurozone have a real budget constraint – if they threaten the ECB’s mandate they can be forced into actual default. The ECB has an inflation-targeting mandate, and it is dominant. It also has powers far beyond simple open-market operations (a restriction Sims takes as given), and can create money to meet its inflation target whatever governments do.
Now, the Eurozone is one extreme at the far end of a spectrum of degrees of central bank independence. But it is precisely because money and debt are distinct that we want central banks to operate independently. We want governments to have budget constraints, and central banks to produce the right amount of money, distributed in the right way, to maintain full employment and price stability. The fiscal theory of the price fundamentally conflates this distinction.[1]
To be clear, none of this amounts to a preference for monetary over fiscal policy. I am firmly in the Brad DeLong and Simon Wren-Lewis camp – when economies are cyclically weak, we need rational, aggressive doses of both. But there should be no doubt – central banks are to blame if inflation is too low.
[1]For similar reasons, the Bank of England seeking an ‘indemnity‘ for its QE programme was a major strategic error, based on flawed reasoning. The sole achievement was to compromise its independence.
Postscript
On reflection, I can’t see very much that is different in Sims’s perspective from that of Sargent & Wallace’s unpleasant monetarist arithmetic. in fact, on re-reading Sargent & Wallace, they are significantly clearer (sorry Prof Sims, at least you’re in good company!). They point out on page two, that if monetary policy dominates, the central bank controls inflation, if fiscal policy dominates (which really means if the fiscal authorities run monetary policy!), then the fiscal authority controls inflation.
It is also interesting that the main example Sims sights as supporting evidence, is that of Brazil in the 1980s. He argues, correctly, that it is possible that raising interest rates contributed to rising inflation because of the increase in government interest payments (which of course was accommodated by money base growth). But this is no way illustrates the need, as he argues, for coordination between fiscal and monetary authorities. It is very clear that raising interest rates in certain circumstances can be inflationary. For example, if the central bank were to set the interest rate on required reserves at 1000% and leave the rate on excess reserves at zero – i.e. if interest rates are simply used as a transfer from the central bank to the public or private sector. But the reason why raising interest rates in Brazil was plausibly inflationary was only because the central bank was willing to accommodate any fiscal deficit. To my knowledge, there is no precedent for hyperinflation coinciding with a contraction in the monetary base. In other words, if the central bank is independent, it can control any inflation by contracting base money – and essentially imposing a budget constraint on the government. The example of Brazil is a case of assuming what you want to prove: yes, if the fiscal authorities dominate, they can determine inflation!
In fact what this discussion reveals is there may be a difference between raising interest rates using the IOR, and raising interest rates via contracting reserves, refusing to allow reserves to grow, or imposing reserve requirements. What is not in doubt is that if a central bank refuses to accommodate a given expansionary fiscal profile, it can impose a constraint on government. It is also very clear from tiered reserves, TLTROs, TFS schemes etc., that independent central banks can always and everywhere create demand, and inflation, if they want to.