Three new bazookas
Recent shifts in policy direction by global central banks may prove to be profound. The Bank of Japan and the ECB appear to be conceding that further bond purchases and lower policy rates are self-destructive. In the ECB’s case bond-buying has achieved its primary objective: consigning sovereign contagion to history. Its secondary objective, strengthening the fiscal positions of European sovereigns and underpinning their accounting solvency, has also been achieved. If ever proof was needed that Europe’s sovereign panic was induced by Trichet’s ECB rejecting QE in 2009, Mario Draghi provided us with the counterfactual. When the central bank with its printing press stands behind the government there is no credit risk, when it leaves them standing alone they are subject to runs. Such is the difference between money and debt.
It has also been increasingly clear for several years that bond-buying QE and ever lower interest rates were not just running the risk of having counter-productive effects on consumer spending, but are threatening financial stability and the banking sector (as the Swedish Riksbank is at pains to point out). It is true that more government debt (and inflation) can be a solution to a private sector debt problem – but it cannot be the case that more private sector debt is the way out for an over-leveraged private sector. There are a good reasons why credit growth has been anaemic despite the collapse in its cost. Austerity tripped on this basic fallacy.
Worse still, the empirical and theoretical basis of lower official interest rates raising spending is weak at the best of times. But in Europe and Japan, where high GDP per capita and relatively even distributions of income prevail, angst about the future – rising health care costs, and shrinking labour forces – may well cause income effects to dominate. The effect is the opposite to that intended: lower rates raise desired savings.
Did we really need to wait until yield curves were flat and short-rates negative to realise that a permanent attrition of banks profits was unlikely to be a solid basis for cyclical strength? Who, after all, was paying for those central bank profits?
Better late than never. Finally a consensus may be emerging that bond-buying QE, flat yield curves and ever-lower interest rates are the policies of the past. We must all concede that the hedge fund manager with his glib analogies had a point, after all. If you stuff you face with enough jelly-doughnuts you will get sick. Alfred Marshall was less vulgar: marginal utility diminishes.
So where next for monetary policy? Governor Kuroda deserves a great deal of credit. He knows inflation is the central bank’s responsibility, and temporary failure is no grounds for jumping ship. He is testing contemporary macro to the limit: and has committed to print money until Japanese CPI rises above its previous 2% target (see para 2).
I am inclined to believe him. I don’t think many people outside a semi-religious core of modern theoretical fantasists really believe that central banks can, armed with determination alone, simply raise ‘inflation expectations’ at will, when all the evidence suggests abject failure to achieve more modest objectives. Nonetheless, the clever economists at the BoJ and ECB have been innovating busily. They have already moved far away from anything you will find in a monetary textbook.
Consider two wonderful innovations: tiered reserves and TLTROs. Part of their brilliance resides in obfuscation. Economics struggles to define ‘money’, let alone the difference between fiscal and monetary policy, or asset pricing. It appears an inevitable feature of modern democracy that central banking must operate behind a veil of ignorance – analogous perhaps to the machinations of the secret service. In this instance, the barrier is cognitive effort.
The inventor of tiered reserves deserves a nobel prize. As does the genius behind the TLTRO. What both have in common is in decoupling monetary stimulus from market interest rates, a point Martin Sandbu and Simon Wren-Lewis have observed.
Tiered reserves make it explicit that the monetary base and t-bills, while sometimes close substitutes, are in fact qualitatively different. And not just for Greece. Bills are issued into a market at a price – an interest rate. The ‘interest’ paid on reserves or IOR, is an interest rate only in name – it should correctly be named a payment (or deduction) to the holders of reserves. Central banks can set market interest rates by setting an interest rate on a fraction of ‘excess’ reserves. If they want to, they can remunerate the balance at a premium, or discount. In future, it is conceivable that required reserves could be shrunk using negative ‘interest rates’ and excess reserves paid a premium. Some economists like to dismiss all this as fiscal policy – via the banking sector. This is not just analytically lazy, but disingenuous. Differential interest rates on reserves is very clearly monetary policy – just not what we are familiar with. The institution is the central bank, and money is being created and destroyed. That’s about the clearest definition of monetary policy you will find.
Making elevated transfers on tiered reserves in excess of money market rates is a Friedmanite dream. Monetary policy is never out of ammunition, just raise rates on required reserves and use ‘excess reserves’ to set market interest rates. There is the small matter of who receives this bonanza – but requiring banks to pass it on to the corporate or household sector is an obvious next step.
The TLTRO has similarly broken new ground. Previously, Bagehot’s dictum was sacrosanct – central banks only ever lent to the bank sector at ‘punitive’ rates, or at least at a premium to what they paid on reserves. This was only ever convention. It has died. Since the ECB announced this year that the TLTRO would be extended to banks at the same rate as reserves were being remunerated, a whole new avenue for monetary policy opens up.
The ECB has made explicit that there are in fact three axes along which monetary policy can be still be eased: duration, credit risk, and price. The TLTRO programme seems far more radical than QE, because there is literally no limit along any of these axes. So far, TLTROs have been extended for four years, at negative interest rates (subject to certain restrictions), and in the form of collateralised loans.
But there is nothing stopping the ECB from making the TLTRO programme its principal policy tool. The next step is to decouple the interest rate on TLTROs from the interest rate the ECB pays on reserves. And before anyone panics that this exposes the ECB to losses and unwarranted risks, that barrier has already been hurdled – the risk is no different to buying long-dated bonds at very low or negative rates.
Many commentators, including the Bank of England’s Ben Broadbent, seem to believe that engaging in policies with a negative net interest income to the central bank are for some reason taboo. This is a spurious form of mental accounting. QE programmes, although currently net income positive, expose central banks to the prospect of huge future net income losses (unless they use tiered reserves) – a point Chris Sims has correctly made. Furthermore, how can it make sense that central bank operations must make the treasury a profit, and never a loss? That is an entirely arbitrary asymmetry.
The Bank of Japan and Bank of England also have TLTRO programmes, with different nomenclature. A logical next step is for the duration and pricing of these directed lending programmes to be extended aggressively. There will be shrill voices crying ‘subsidies’! This too is mental accounting – what is the ‘wealth effect’ of QE if not a ‘subsidy’ to the owners of financial assets. On this logic, negative interest rates are a bailout of irresponsible borrowers. Every change in interest rates has distributional consequences, whether we like it or not. Every economic recovery is a ‘bailout’ in austerianese.
Directed lending via the banks to the private sector provides central banks with a new bazooka. If Kuroda wants, he can deliver his new target – and Japanese real GDP may be 10% higher, if he does.
The BoJ has also innovated in targeting the equity risk premium directly. Other central banks’ focus on purchasing government and corporate bonds epitomises the very ‘risk aversion’ they are trying to combat and lacks intellectual rigour. The great irony is that they are far more likely to make huge balance sheet losses on their bond portfollios, and BoJ officials will be high-fiving over their equity gains. The implied equity risk premium in Europe, Japan and the UK is currently enormous.
I do worry that buying what is currently a cheap asset on the scale that government bonds have been bought might rapidly create an equity bubble, but it is a far more logical way to raise investment spending than targeting interest rates on supposedly ‘safe’ assets, damaging the pensions, insurance and banking industry in the process.
The abandonment of ever lower bond yields by the Bank of Japan, and its attempts to raise demand by new means should be welcomed. So too should the rumours that the ECB is abandoning further QE. This is a recognition that these policies, once relevant, are now counterproductive. Our economies do not need lower government bond yields, or lower official overnight interest rates – that is surely obvious. Fortunately, the BoJ and ECB have also revealed new tools of unlimited monetary power. The next step will be a measure of their courage to use them.
“QE programs, although currently net income positive, expose central banks to the prospect of huge future net income losses (unless they use tiered reserves) – a point Chris Sims has correctly made. Furthermore, how can it make sense that central bank operations must make the treasury a profit, and never a loss? That is an entirely arbitrary asymmetry.”
Several points:
a) This is a matter of risk assessment. Don’t underestimate central bank net interest margin staying power in the face of rising rates. Interest rates on reserves are likely to rise slowly, such rising rates will likely confirm QE asset exit strategies (which will shrink the balance sheet and proportionately some of the negative margin exposure), and zero interest currency will start growing normally again as a proportion of the balance sheet (favorable for margins). So there is risk, but there are powerful mitigating factors against the size of the risk. This is interest rate risk management. Commercial banks do it, with surprisingly similar risk considerations. And central banks do it, although they are in the interesting position of pulling the trigger on the realization of short rate risk.
b) Central bank interest margin management should be considered over long cycles. This should include the consideration of cumulative profit that has been previously disbursed to the fiscal authority. That’s a favorable information input to a more comprehensive risk assessment, considered as a long run issue.
c) The Fed has anticipated such risk to capital. It actually has the authority to create a negative liability – instead of negative capital – in the event of net interest margin losses due to QE. This again recognizes the long cycle of central bank interest margin management. It means that if such losses materialize, they will be offset by future profits, as sensibly anticipated. Those future profits would be used first to eliminate such a cumulative negative liability – instead of being distributed to Treasury. Future profits are almost certain in a post-QE environment with a more normal balance sheet. So therefore losses may be realized if necessary for long term management (which is not necessarily inconsistent with your view, although seemingly more qualified). But an expected pattern of permanent losses is not in the cards (which I assume is not consistent with your view, or you would have said so). So I think that this risk framing can’t be used as a justification for more central bank fiscal risk adventures that are more extreme and that constitute a more obvious breach of monetary policy authority granted to the central bank.
JKH – you may be interested in Draghi’s recent press conference remark on this issue. He was very clear that monetary policy actions are not, and should not, be motivated by central bank ‘profitability’. I think he is absolutely correct. There is a distribution of losses (gains) associated with competing policies. QE is far riskier than direct transfers because of the size of the tail of the distribution. If bond prices fall precipitously, the stock of reserves would vastly exceed the mark to market value of assets, making reserve contraction through asset sales impossible. By contrast, a more effective stimulus (such as a direct transfers) would expose the CB to smaller, but more certain, losses. Remember, driving up asset prices to unsustainable levels is a form of transfer. Potentially a very large one.
Eric,
There is absolutely no need for a central bank to sell bonds in exiting QE. They can mature the bonds, in which case potential marked to market losses evaporate. Indeed, that is the plan of the Fed, announced long ago.
QE exit is a long term proposition. Central banks do not need to panic in that context. The entire effect is taken up within interest margins over the long term. This is the context in which I said that a central bank has interest margin staying power, with the beneficial effects I noted. There is no good reason to sell bonds at a loss. In particular, the very fact that central banks are not profit maximizers eliminates that sort of private agent economic argument as a reason to sell bonds at a loss. Central banks are not day traders in this context. Indeed, this is why central banks don’t mark their sovereign issued Treasury bonds to market in their financial accounts.
Conversely, if there are sensible opportunities to occasionally sell some of their bonds in the process of exit, incurring losses that are limited in the context of overall interest margin performance, that approach may be reasonable from time to time. We may well see the Fed do that in the event of exit. But I think in the case of the Fed at least, those who stoke the fear of unrealized marked to market losses simply haven’t run the numbers on long term interest margin expectations and risk.
P.S.
I only scanned the transcripts of the last few Draghi press conferences, but couldn’t see anything on the question of central bank profits. Would you have a link on that?
The reasons for selling at market to market losses has nothing to do with P&L. The important point is that if CBs have large mk-to-mkt losses on their bond holdings they cannot withdraw reserves by reversing QE. This is the equivalent problem with having created reserves without corresponding assets. In fact, it is potentially a much bigger problem given the ineffectiveness of QE and negative rates.