The strange world of negative interest rates
This post first appeared on Coppola Comment in December 2012.
There has been considerable discussion recently about central bank interest rate policy, and in particular, what further monetary easing they can provide when interest rates are at or close to zero. Central banks have been using a range of unconventional tools, of which quantitative easing is the best known, to depress market interest rates. I have serious reservations about the effectiveness of these for the real economy. But there is evidence that central banks are starting to consider negative rates as a policy instrument, so in this post I shall consider what would be likely to happen if central banks reduced rates to below zero.
There are two types of rate that central banks could reduce to below zero. One is the policy rate itself, which is the rate at which the central bank will lend to banks against collateral. The other is the interest rate that the central bank pays on excess reserves. I shall return to the policy rate later, but first I wish to discuss the effect of reducing interest rates on excess reserves to below zero.
Banks currently have rather a lot of excess reserves, since the unconventional tools themselves have caused reserves to build up, so reducing interest rates on excess reserves looks like an attractive policy. After all, if maintaining higher balances in their central bank reserve accounts than they need to becomes expensive, banks won't do it, will they?
__________________________________________________________________Join the Financialist - The easy way to find your next job in finance
Let's consider what negative interest rates on excess reserves actually are. There are two ways of looking at them:
- Negative interest rates on reserves are a charge to banks for placing money on deposit in a fully guaranteed deposit account. They are equivalent to a safe deposit charge.
- Negative interest rates on reserves can also be regarded as a tax on deposits. As with all taxation, the burden of that payment does not fall on the bank itself. I shall discuss shortly where the burden actually falls.
The thinking behind charging banks for safe deposits at the central bank is that banks will choose to lend that money out for a positive return instead of suffering erosion of principal in return for safety. This unfortunately demonstrates the misunderstanding of lending that underlies the failure of modern macroeconomics to explain or prevent financial crises. Banks lend if they choose to - if the balance of risk versus return works in their favour. If it doesn't, no amount of reserves will make them lend. They will hoard money instead. And the fact is that the balance of risk versus return at the moment is so horrible that banks do not wish to lend except to the most creditworthy borrowers. Banks are seriously damaged and very, very scared of losses: they are already carrying high levels of risky loans against which they have insufficient loss-absorbing capital, and governments are withdrawing the implicit guarantee that enables them to maintain risky lending against insufficient capital. And the world is a risky place for banks at the moment: there is a severe credit crunch in the European periphery due to the ever-growing sovereign debt crisis there, and creditworthy borrowers are thin on the ground everywhere due to damaged household and corporate balance sheets.
Not only that, but there is evidence that creditworthy households and corporates don't want to borrow, either. Borrowing and spending is out of fashion, saving and thrift is in. The Bank of England's "Trends in Lending" report says risky borrowers such as SMEs can't get the finance they want at the rates they want to pay, but it also notes that lenders complain about low demand and lack of creditworthy borrowers.
There is general risk aversion among investors, too. Yields on highly-rated government debt are at historic lows: in some countries this is partly due to QE, but German bunds have had no QE and yet were recently trading at yields at or below zero. Traditional safe haven currencies such as the Swiss franc are under such exchange rate pressure that the Swiss National Bank has intervened to prevent the franc's appreciation damaging the Swiss economy. Deposit account interest rates are on the floor, and some banks in the US are imposing negative interest rates on large sums of money in FDIC-insured deposit accounts.
It's worth remembering, too, that reserves are created by the central bank, not by commercial banks, and that commercial banks have no power to reduce the total amount of reserves in the system. That can only be done by the central bank. So if commercial banks were discouraged by negative interest rates from holding excess reserves, but there were still excess reserves in the system, banks would look for ways of passing on those excess reserves to other banks. Excess reserves would become something of a hot potato, with no bank wanting to be caught with excess reserves at the end of the day. I suppose that might improve the velocity of money, but I could see it leading to all manner of stupid investments.
So, if - say - the ECB imposed negative interest rates on excess reserves in a world which is both risky and risk-averse, how would banks behave? I can't see any reason at all why they would increase lending. They would be more likely to look for other "safe" investments as an alternative to parking deposits at the central bank. And they wouldn't have far to look. The debt of countries like Germany, the US and the UK is explicitly backed by government guarantee just as central bank reserve accounts are. If the yield on these bonds were higher than the negative interest rate charged by the ECB, banks would purchase these with their surplus deposits. There might also be round-tripping from banks seeking to arbitrage the difference between sovereign debt yields and central bank negative interest rates. The inevitable effect would be downwards pressure on the yields on "safe" government debt in much the same manner as QE.
But negative interest rates on reserves could have an even more toxic effect too. For this we need to look at the experience of Denmark, which introduced negative rates on excess reserves in response to the ECB's cutting of rates to zero. Denmark was forced to do this because it is maintaining a currency peg to the Euro, and interest rate parity with the Eurozone would place that peg under pressure due to capital flight from the stressed areas of the Eurozone. Denmark's banks complained that their margins were being squeezed. The Danish central bank's rather unsympathetic response was to suggest that Danish banks could simply raise their rates to borrowers.
Now, Denmark is experiencing some inflationary pressure due to capital inflows from the Eurozone, and its currency is under pressure, so raising interest rates to borrowers helps to dampen this down - which is why the Danish central bank was unimpressed by the banks' complaint. But consider what would happen if an economy experiencing deflationary pressure introduced negative interest rates. The squeeze on the margins of already-damaged banks would inevitably lead to higher rates to borrowers and reduced lending volumes. This is monetary tightening, not easing, and the effect would be contractionary. It would make the recession worse.
This is also consistent with my description of negative rates on excess reserves as a tax on deposits. I've described it as monetary tightening, but it could also be seen as a fiscal tightening since central banks are part of government: one of the strange effects of very low interest rates is that the distinction between monetary and fiscal policy is becoming ever more blurred. And the Danish experience shows where the incidence of that tax would fall. It would not fall on banks, which would find a way of passing on that cost to people: nor would it fall on savers. It would fall initially on borrowers, as lending rates were raised: then, as lending volumes fell in response to higher rates, it would fall on shareholders in the form of lower returns, and employees in the form of lower wages. And the overall effect on the economy would be deflationary, due to reductions in both credit and income for businesses and households.
So negative interest rates on excess reserves would be counter-productive. Fortunately the Fed seems to be wise to this and has ruled them out, at least at present: but unfortunately the classical macroeconomic bias of the Eurosystem leadership appears to have rendered them incapable of understanding the contractionary effect of negative interest rates on reserves, even though they have an example on their doorstep. The head of the ECB, Mario Draghi, has appeared to suggest that negative interest rates might come soon. Heaven help the Eurozone if that goes ahead.
But what about cutting the policy rate itself to below zero? This would in effect mean central banks would pay commercial banks to borrow from them.
A negative policy rate would be a direct subsidy to banks. And if it actually worked to improve lending volumes and bring down rates for riskier borrowers, it might be a worthwhile subsidy, too. But in a world that is both risky and risk-averse, I don't think it would have that effect. Remember that banks only lend if the risk versus return profile works for them. And people and corporates only borrow if they wish to spend - which at the moment they don't, much. The failure of QE as an economic stimulus to the real economy demonstrates that throwing money at banks doesn't make them lend. Why would paying banks to borrow be any more effective?
A negative policy rate could be effective as a bank recapitalisation. The central bank lends against collateral - typically its own sovereign debt and sometimes other sovereign debt too. Suppose the central bank pays banks to borrow against collateral of the sovereign bonds they already hold, on which they are receiving interest. They then use that money to purchase more sovereign bonds to repo back at the central bank for more funds. And they receive interest both on the borrowing and the purchases, all of it effectively paid by taxpayers (since central bank payments are guaranteed by government). That is state-funded bank recapitalisation by the back door. Nice.
There is no doubt that banks need recapitalising. But why on earth do it by such a roundabout route? And what is to stop them paying out their earnings from this sovereign round-trip in the form of shareholder dividends and employee bonuses?
Like negative rates on reserves, negative policy rates could actually have a toxic effect on the real economy. Across Europe, including the UK, many loan rates - especially mortgages and business loans - are tied to the policy rate. So if the policy rate were cut to below zero, lenders would find their margins squeezed on existing lending: they could even find themselves receiving negative returns on these loans. Realistically they cannot cut their deposit rates to savers to below zero (savers would stuff mattresses instead), so their only option is to RAISE lending rates to new borrowers, widening credit spreads. Exactly the same effect as negative interest rates on reserves, in fact - and the same effect as QE. Existing floating-rate borrowers would of course find their borrowing costs eased: but would they spend this money? Many of them are highly indebted, and many more of them are suffering erosion of their savings due to very low real interest rates. I suspect they would use the windfall from lower borrowing costs to pay off debt, or to top up their savings, rather than spend. In which case the overall effect on the real economy of a negative policy rate would be contractionary, not expansionary.
So cutting policy rates to below zero would be as counter-productive as cutting interest rates on reserves. And it would be unpopular. I can't imagine any electorate, wounded as they are by the behaviour of banks, tamely accepting bank funding being subsidised while interest rates to new borrowers soared and the economy crashed.
But suppose, in a world gone mad, central banks decided to cut both interest rates on reserves and the policy rate to below zero. This is where things become very, very strange. Government (via the central bank) would pay banks to use state funds to settle lending while taxing them if they attempted to self-fund. The effects would be extraordinary.
Firstly, banks would be actively discouraged from holding excess reserves at the central bank. But they would have an increased need for government securities. So initially they would try to reinvest excess reserves in government securities, driving down yields. Thereafter, any further need for collateral could be met by using central bank funding to buy securities.
Both the unsecured interbank market and the repo market would die, because it would be far cheaper for banks to borrow from the central bank than from funding institutions - even with the cost of holding increasingly expensive safe assets for collateral. Banks would become completely dependent on central bank funding.
Rates to commercial depositors would crash, as banks would no longer have any incentive whatsoever to seek deposits. In fact as they would be charged to hold them, either directly through the tax on excess reserves or indirectly through negative yields on assets purchased with those deposits, they would actively discourage them by cutting interest rates. As commercial deposit rates approached zero, savers would start hoarding cash instead, or investing in physical assets such as gold. There might also be a flood of money into National Savings if the interest rate were favourable. And as returns on various kinds of wealth fund including pensions, already low, turned negative due to crashing yields on safe assets, they would close their doors and savers would be forced to turn to cash or physical assets.
Banks' aversion to holding deposits could have very weird consequences for the real economy, as the Fed notes. For example, if they imposed negative interest rates on ordinary deposits including current accounts, people would stop using banks and we would return to a cash-based economy: or they might even turn to alternative currencies such as e-credits on mobile phones. I have no doubt that e-credit suppliers would be only too happy to see their currencies take the place of sovereign currencies for day-to-day transactions.
Theoretically, recapitalisation of banks coupled with low-cost funding should encourage banks to lend. But their balance sheets are still stuffed full of risky loans. Even if they had more capital and cheap funding, they wouldn't want to take on more risk lending. No, they would go for safe assets and cash. Nowhere in this bizarre scenario is there ANY incentive for banks to increase risk lending.
Theoretically, too, deposit rates crashing to zero or below and falling rates on floating-rate loans to existing borrowers should encourage households and businesses to spend instead of saving. But not when they are highly indebted and suffering erosion of principal on existing savings. No, they would pay off debt and increase their savings. Nowhere is there any real incentive for households and businesses to spend instead of saving.
And it gets worse. As safe collateral became scarce, central banks would inevitably extend the range of acceptable collateral to other bank assets, such as mortgages.....eventually all forms of lending would be pledged to the central bank in return for funding. Banks would no longer carry the risk of that lending: in the event of default, the loss would rebound to the central bank to whom the asset was pledged, and the state - not the bank - would take the loss. Central banks are already a long way down this path even without negative rates.
So the final outcome of a complete negative-rate policy would be a state-recapitalised banking system, funded entirely by the state and with lending risk borne by the state. Effectively, that is nationalisation of the banking system - but without the control that would force banks to use state funding productively to benefit the real economy. It would be the worst of all possible worlds.....except one in which the sovereign currency was rejected in favour of private-sector alternatives because of the cost of holding it. If the sovereign currency is rejected because people have to pay to use it, how long will government and the rule of law last?
Summers on bubbles and secular stagnation forever - FT Alphaville
The costs and benefits of repealing the zero lower bound - Miles Kimball