Friday, 6 June 2014
Sacred cows and the demand for loans
Scott Sumner, on a recent post, asked me to explain what I meant by "loan demand". This got me thinking. Exactly what DO we mean by "loan demand"? Who is demanding, and what do they really want?
By "loan demand", we usually mean the demand from households and corporations for the credit provided by banks. But actually this makes no sense. What households and corporations actually want is not loans. It is money.
Households that have money generally do not borrow. They buy their houses, cars, yachts, holidays to Bermuda with money they already have. It is households that DON'T have money that borrow. They do so in order to buy the houses, cars, holidays to Ibiza (perhaps not yachts so much) that they don't have the money to afford. They would really like to buy these things from money they already have, but there isn't enough of it right now, and in the case of houses there won't be for at least 25 years even if they save assiduously. They do not "want" loans. They want money. If someone gave them the money to buy their house, car, holiday, they would gratefully accept it - and that would be one potential borrower lost to the banking system. If everyone had enough money for everything they needed, we wouldn't need banks.
Clearly, therefore, in an economy in which people don't have enough money to buy all the things they want, there is no such thing as lack of demand for credit money. People still want these things. Only if everyone had unlimited money would there be no demand for credit money.
So it is actually meaningless to talk about changes in demand for (credit) money. What we should be talking about is the supply of loan assets. But it is not banks that supply loan assets. What we usually call "demand for loans" is actually the supply of loan assets by households, corporations and governments to banks and investors. What we usually call the "supply of loans" is actually the supply of money by banks and investors in return for loan assets from households, corporations and governments.
This is more obvious when loan assets are supplied in the form of bonds. A lender to the UK government receives a pretty piece of paper which in days gone by used to have a golden edge (that's why UK government bonds are known as "gilt-edged securities"). This piece of paper IS the loan asset.*
Loan assets are claims on the future income of households, corporations and governments. Lending is always a bit of a gamble: future income is by definition uncertain. Default happens when income in reality does not match the expectations against which the loan was advanced. The interest payments on a loan are both compensation for the opportunity cost to the lender of not using the money (although in the case of banks which create money when they lend, the existence of this opportunity cost is debateable) and, more importantly, a consideration or surety against possible future default. The higher the likelihood of future default, the higher the interest payments. Payday lenders such as Wonga operate a model in which very high interest payments more than compensate for the losses due to default: they make money not by ensuring that loans are repaid, but by rolling them over and continuing to charge interest, often at ever-higher rates, when they AREN'T repaid.
Traditionally, we model demand for money as if money itself is interest-bearing - so demand for money reduces as the interest rate rises. But fiat money in the form of currency does not bear interest, and it is exchangeable at par with credit money: credit money therefore cannot bear interest either. It is the loan assets associated with credit money that bear interest. The amount of money borrowed is discounted by the interest rate over the period of the loan: we can regard this as a "haircut". The longer the duration of the loan, and the less creditworthy the borrower, the deeper the haircut. In practice, of course, the amount of money borrowed is fixed, and the amount paid to the lender varies (at least for long-term loans such as mortgages). The haircut can be reduced by guarantees or collateral that reduce the risk to the lender.
Clearly, a lender that wants a high return will lend to less creditworthy borrowers and for longer periods of time. The ability to recall money lent out reduces the risk to the lender and therefore the price of the loan. We see this most obviously in the case of call deposits at banks, where the ability to withdraw money on demand means that depositors earn practically nothing: those who are prepared to tie up their money for longer periods of time earn more. But it equally applies to longer-term lending by banks and investors. We call the higher rate paid for longer-term loans the "illiquidity premium".
So when we talk about "falling demand for loans", what we are actually saying is that banks and investors are demanding a deeper haircut on money lent out than households, corporations and governments are prepared to accept. When banks and investors won't lend at all, the haircut is 100%. And the reason for this is that banks and investors perceive the risks to have increased. Whether this perception is accurate is not the point: it may be because banks are damaged and investors fearful, rather than any real change in the circumstances of the borrowers. Or it may be because the value of borrowers' collateral has fallen (house price collapse, for example), or borrowers have suffered falls in their real income that make the price of loans unaffordable. Whatever the reason, the effect is that borrowers - whose demand for money, remember, is still the same - decide to go without their houses, cars, holidays rather than accept what they consider a disproportionately large claim on their future income. They postpone their purchases and may save up for them instead of borrowing. This reduces the supply of loan assets. And since what banks and investors provide in return for loan assets is MONEY - and banks create money when they lend - it therefore reduces the supply of money in circulation.
This creates a considerable problem for economic theories involving the demand for money versus the demand for credit. If we look at the standard IS-LM model, for example, it shows the demand for money increasing at the expense of the demand for riskier & less liquid loans (and other assets) when output and interest rates fall. But this isn't right. The demand for money does not change. What changes is the willingness or capacity of banks and investors to meet that demand by lending. This remains the case whatever action is taken by central banks to increase the actual supply of monetary base. If banks and investors remain risk-averse, and households, corporations and governments remain unwilling or unable to improve the quality of the loan assets they offer, the increased monetary base goes nowhere: it sits on bank balance sheets as reserves and in investor portfolios as safe assets.
And this raises serious questions about the conduct of monetary and fiscal policy when the supply of quality loan assets by households, corporations and governments falls - whether that is caused by declining creditworthiness on the part of loan asset suppliers, or risk aversion on the part of banks and investors. If we wish to improve activity in the economy, we need to find ways of enabling households, corporations and governments to spend.
One approach is of course to encourage those who have money to spend it rather than lend it: this is the principal objective of QE, but it founders on the problem that people who are rich enough to have large holdings of financial assets tend already to have bought their houses, cars, yachts etc. We say that they have a low "marginal propensity to consume", but we could just as easily say that they have a low demand for money. If we exchange some of their assets for money, all they do is reinvest that money in other financial assets: admittedly, if these are corporate loan assets there might be some benefit to the economy, but the evidence seems to be that far too much of the QE money went to blow up unproductive asset bubbles, not into productive lending to the real economy.
Another approach might be to make it easier and cheaper for people to borrow money, perhaps by offering guarantees for certain types of loan asset: Help to Buy is an example of this, as are the extensive systems of loan guarantees that most governments offer to small businesses and exporters. A third would be to reduce the real demand for money by simply giving people more of it: this is the famous "helicopter drop", or if the money is specifically used to redeem loan assets, debt jubilee or "monetization"**. And a fourth would be to encourage corporations to pay higher wages and governments to pay higher benefits, though this could be viewed as "robbing Peter to pay Paul", since governments must recover this from taxes and corporations from profits: it would only be beneficial if multiplier effects meant profits and tax incomes rose more than the cost of higher wages and benefits***.
Whatever approach is taken, the objective is clear. Lack of spending in an economy (shortage of aggregate demand) is caused by distributional scarcity of money, not by lack of loans. It is not necessary to restore lending in order to encourage spending. It is necessary to replace the money that is not being created by banks - this is the job of central banks. And it is necessary to ensure that the new money created by central banks goes to where it is needed. We are doing pretty well on the first of these, but are failing abysmally on the second. A whole herd of sacred cows, with names like "debt monetization causes hyperinflation", "higher wages cause inflation" and "no-one should get something for nothing", are getting in the way. We need to shoot them.
Rediscovering IS-LM - Pieria
* Well, ok, in these days of electronic settlement and custodian banks the lender probably wouldn't receive the actual security - it would be held in custody somewhere. But it would physically exist.
** Debt jubilee in this context means central bank monetization of private sector debts.
*** Yes, I know I haven't considered inflation. As direct stimulus of this kind would be a response to low aggregate demand, inflationary effects may be a sign that it is working. I find it very disturbing that inflation is always regarded as a bad thing: out of control inflation is extremely damaging, but mild inflation can be a necessary consequence of recovery from a slump and in my view should not routinely be squashed.