Until a few years ago, the determination of the money supply was generally taught in terms of the monetary base/money multiplier theory, which had formed the analytical basis for a great book on U.S. monetary history by Friedman and Schwartz. In reality, this view was incorrect, which the Great Financial Crisis of 2009 made clear.
The Monetary “Theories”
Before we get to the crisis of 2009 and how it debunked the monetary base story, we need some context. Although the money multiplier is derived from an identity, the causal direction is the reverse of that normally described. A central bank normally decides on the official short-term interest rate that it wants to set, and the two key ratios – the currency/deposit (C/D) ratio and the desired reserve/deposit (R/D) rate of the banks – determine the monetary base that the central bank has to create if it is to keep short-term market rates in line with the intended official rate. In truth, it is not so much a money multiplier as a money divisor.
It’s obvious that a central bank, via its monetary policy committee – such as The Bank of England’s Monetary Policy Committee (MPC), or the Federal Open Market Committee (FOMC) – sets the interest rate, and not (prior to the attainment of the zero lower bound or ZLB) the growth rate of the high-powered monetary base. Therefore, there was a basic inconsistency in most earlier monetary economic theory. When discussing current public policy issues, monetary economists would advocate that the FOMC change the official short rate to a certain extent, whereas in their theoretical analysis they would assume that the authorities set the monetary base, thereby allowing short-term interest rates to be set by market forces to equate the demand and supply of money.
This inconsistency continued for decades. That is, until the experience of the 2009 financial crisis disproved the money multiplier theory. Once the official desired short-term interest had reached the ZLB in early 2009, there was no further need to restrain the growth of the monetary base in order to maintain a desired positive level of official short-term interest rates. Instead, to enhance the degree of monetary expansion, the monetary base was raised massively via quantitative easing. But rather than promoting a roughly equivalent rise in the broader monetary aggregates, the money multiplier collapsed. A few numerical examples are shown in Table 1. (Note: There was a somewhat similar collapse in the money multiplier in the Great Depression in the 1930s, but for rather different reasons, i.e., a sharp change in the desired C/D ratio, rather than in the desired R/D rates.)
Against this background, it has become impossible to continue believing that central banks set the money stock by varying the monetary base within a system in which there was a predictably stable money multiplier. With the monetary base/money multiplier theory discredited, a search has now started for a new, revised paradigm.
Before we move on to discuss this search, let’s ask two questions. The first question is, why have almost all central banks, almost all the time (except, for example, when at the ZLB), chosen to set interest rates and not the quantity of base money? The second question is, why did the money multiplier collapse and the R/D ratio balloon once central banks espoused quantitative easing?
The short answer to this second question, explained at greater length in one of my recent papers, is that commercial banks found themselves in a liquidity trap. Interest rates on longer-dated safe debt, whether public or private, had been lowered to such low levels that the likely next move was up, thereby involving expected capital losses. Large, safe private corporates could borrow more cheaply on capital markets. Lending to risky, smaller private sector borrowers, especially because of the now higher capital requirements, could only be done profitably on banking terms and conditions that relatively few such borrowers were able and willing to meet. So, at the margin, holding additional cash reserves at the central banks – in some cases, as in the U.S., at a small, positive interest rate, with no capital requirement and adding to liquidity – became the preferred option.
The New Paradigm
In many cases, a bank makes a term loan to a corporate or a mortgage loan to a person by simultaneously writing up both its asset book (loan to X) and its liabilities (deposit of X). That loans create deposits has been a well-known fact for a century or more.
The argument that loans create deposits and, therefore, that commercial banks themselves determine the supply of money has the advantage that it is more true than the monetary base/money multiplier theory. I initially welcomed this line of thought, but it is only partially true. Also, the policy implications that some adherents of this approach draw from it – that commercial banks’ capacity to create loans and deposits needs to be further curtailed or removed – must be treated skeptically. This is all the more important because this analytical approach appears to be receiving the imprimatur of the Bank of England.
Two recent examples are McLeay, Radia, and Thomas, and Jakab and Kumhof:
MacLeay, Radia, and Thomas:
“This article explains how the majority of money in the modern economy is created by commercial banks making loans.”
“Money creation in practice differs from some popular misconceptions – banks do not act simply as intermediaries, lending out deposits that savers place with them, and nor do they ‘multiply up’ central bank money to create new loans and deposits.”
“The amount of money created in the economy ultimately depends on the monetary policy of the central bank. In normal times, this is carried out by setting interest rates. The central bank can also affect the amount of money directly through purchasing assets or ‘quantitative easing.’”
Jakab and Kumhof:
In the intermediation of loanable funds model of banking, banks accept deposits of pre-existing real resources from savers and then lend them to borrowers. In the real world, banks provide financing through money creation. That is they create deposits of new money through lending, and in doing so are mainly constrained by profitability and solvency considerations. This paper contrasts simple intermediation and financing models of banking. Compared to otherwise identical intermediation models, and following identical shocks, financing models predict changes in bank lending that are far larger, happen much faster, and have much greater effects on the real economy.
(Note: also see "Where Does Money Come From?," "Can Banks Individually Create Money out of Nothing? The Theories and the Empirical Evidence" and "How Do Banks Create Money, and Why Can Other Firms Not Do the Same? An Explanation for the Co-existence of Lending and Deposit-Taking").
The first implication of this line of analysis is that if anything has gone wrong with the monetary system and the credit cycle, it is all the commercial banks’ fault (this is only partially true). The following implication is that to prevent future banking crises, such powers of credit creation should be removed from commercial banks. This latter proposal has been endorsed by those who advocate narrow banking; a good history of this set of proposals is given by Laina. Interestingly, this proposal is now apparently going to be put into practice in Iceland.
Despite being partially correct, this line of argument taken by itself as a representation of how commercial banks operate, is seriously misleading – so much so as to be almost as erroneous as the money base/money multiplier story and, as I shall show, for the same fundamental reason.
Let’s start with the instance most favorable to the claim that bank loans create deposits: that is the occasion when a bank makes a term loan, or grants a mortgage, and does so by writing up both sides of the balance sheet. What would happen if the bank client had not asked the bank to do so? Picture yourself as a bank client where, without your request, the bank increases both a claim on you and a deposit liability to you. You would be furious.
Borrowers Take the First Step
The crucial point is that the bank does not initiate the bank loan; the borrower does. Banks are in a service industry, like restaurants and hotels. Banks set out the terms and conditions under which they will grant a loan, and then wait for private sector clients to approach them.
Moreover, the borrower does not take out the term loan or mortgage in order to hold onto the deposit, at an adverse spread (at least not normally). He does so to make a payment. Quite often that payment then will be used by the recipients to reduce their own bank indebtedness. Again, note that while banks have the right to call in loans, they rarely use such powers, and early repayment of outstanding bank loans is primarily undertaken at the initiative of the borrower. So both the initiation and the conclusion of bank credit to the private sector are determined by borrowers.
Rather than claim that banks create credit and then such loans create money, it would be much nearer the truth to say that the private sector creates credit and money for itself, and that the banking sector is the medium through which private sector clients do so, on the terms and conditions set out by the banks.
This becomes even clearer when we turn from loans and mortgages to overdrafts and credit card lending. In the latter case, the initial decision is the limit of borrowing that the bank will honor. The first time that the bank is aware of the transaction is when the payment order, by check, credit card, or otherwise, comes in from the borrower. The bank then writes up the loan side of its book, while the counterpart is a debit to another bank or to another customer of the same bank. The key initial decision is that of the private sector borrower to agree to an (overdraft) facility with the bank, on the terms and conditions set by the bank. Thereafter the timing and usage of that facility are almost entirely in the hands of the borrower.
In a world in which the cash flows of private sector agents can fluctuate, such overdraft facilities are valuable. In so far as the cash flow fluctuations are random, the banks can even them out overall through the law of large numbers. Therefore, overdraft and credit card lending represents a sizable proportion of total private sector lending.
Reliance on unused overdraft facilities in the U.K. differs markedly between sectors. The Bank of England publishes statistics on committed credit facilities and on credit (lending) outstanding. The difference between the two is a good proxy for unused (or undrawn) facilities. These statistics are broken down by U.K. industries and credit to individuals (the latter includes credit card facilities). The data is available on the Bank of England’s website. The overall figure derived (total facilities outstanding less total lending outstanding) for the latest date, July 2015, comes to about £417 billion, or 18.2% of total loans outstanding. However, reliance on unused overdrafts differs greatly between sectors, as shown in Table 2. It is very low in lending to other financial intermediaries (except for insurance companies and pension funds) and to individuals on mortgages (below 5%). It is much higher for lending to manufacturing (nearly 120%), utilities (138%), trade (74%), transport and information (101%), construction (42%), and other (non-mortgage related) loans to individuals (62%). In short, if one strips out the network of intrafinancial loans and the mortgage business, then the use of committed but undrawn overdraft facilities is a central feature of the British banking system.
Unused overdraft facilities are as much immediate liquid purchasing power as bank deposits, as Keynes noted in his Treatise on Money. In principle, one should consider aggregating the two together, but until quite recently one could not do so because the U.K. banks were unwilling to provide the data. When the government tried to control bank lending by quantitative ceilings, prior to the Competition and Credit Control reform of 1971, one claim that the banks made was that they could not control such lending because it was precommitted via overdraft facilities. So, as documented, the Treasury then (in the late 1960s) asked the banks for data on such facilities, but the banks refused, correctly fearing that the authorities would seek a quantitative limit on such facilities. However, the authorities – that is, the Treasury and the Bank – did not press the case then, possibly on the grounds that any limit on formal, legal borrowing facilities might be avoided by their transformation into informal agreements.
Be that as it may, the timing and activation of bank borrowing is done at the initiative of the borrower. Fluctuations in such borrowing are far from random. Thus, after the bankruptcy of Lehman Brothers in September 2008, there was quite a large increase in bank lending to the private sector in the U.K. and eurozone despite the sharp decline in activity (see Table 3). This was to some large extent driven by borrowers’ fears that banks might either become insolvent (and therefore be unable to honor their overdraft commitment) or would seek as soon as possible to lessen the scale of their facility,19 or to withdraw it entirely. (Note: The author’s personal overdraft limit was summarily reduced by his bank in November 2008 from £10 thousand to £2 thousand). It then became temporarily worthwhile to protect access to liquidity by paying the adverse spread, while shifting the resulting deposit to a safer bank. So in this quarter (Q4 2008), there was the remarkable coincidence of a temporary surge in bank lending and deposits exactly at the time when the banking system was trying to cut back on facilities and new lending agreements as fast as it could. From the viewpoint of overall liquidity and purchasing power, it was the latter effect that mattered. The banks were drastically tightening the terms and conditions of lending in Q4 2008, and the underlying liquidity was falling sharply.
More generally, turning points in the cycle are frequently related to unexpected declines in sales and increases in inventories at the start of the downturn, and the reverse at the start of the recovery. So there is likely to be an unexpected shortfall in cash flow in the former case, and vice versa in the latter. Then the unexpected cash flows may be quite largely met by calling on used overdraft facilities. This helps to explain why the contemporaneous relationship between changes in GDP and in bank lending is often countercyclical, although the relationship between GDP growth and lagged bank credit is strongly positive.
To summarize, the commercial banking sector acts as a service industry, setting out the terms and conditions on which it will provide its financial services. Given these, its private sector clients then determine the timing and amount of bank credit provision. The key variables are the banks’ choice of such terms and conditions and the private sector’s appetite for borrowing on those terms. Seen in this light, the claim that bank credit is the genesis of money creation without any mention of the private sector’s key role amounts to a misrepresentation.
That said, the banks’ influence on the rate of growth of both loans and deposits, via shifts in setting their terms and conditions, can still be large and significant. Although the level of short-term interest rates is set by the monetary authorities, the spread between deposit and lending rates is controlled by the commercial banks; this spread is generally found to be a significant factor in determining the growth rate of both loans and deposits. (Note: If the spread goes to zero, it is rational for borrowers to both borrow and hold extra deposits until the marginal utility of extra liquidity goes to zero. If the spread rises indefinitely, borrowing will fall to zero). Similarly, changes in banks’ risk aversion can trigger sizable fluctuations in the provision of lending limits, both overdraft and credit card, and in required terms for collateral and margin. But, partly because time series data on unused facilities and lending terms are scarce, the actual contribution of banks to the growth of loans and deposits has not been satisfactorily modeled or even properly analyzed.
Bank Finance of the Public Sector
A purchase by a commercial bank of public sector paper – a bond or a bill – creates money in exactly the same way as does a bank loan to the private sector. The bank buys the bond, say, and pays for it by crediting the seller with a check written upon itself. As noted earlier, the recipient of a loan is no more likely to keep the proceeds on deposit with the same bank than is the seller of the government bond to the bank.
But we tend not to think of bank lending to the public sector as being the initiating factor in creating money, in the same way as many think of bank lending to the private sector as creating money. This is because banks are the passive residual providers of finance to the public sector. When the government runs a fiscal deficit, not covered by debt sales to the non-bank public, the excess spending is financed in the first instance by creation of monetary base, which mostly ends up in larger commercial bank deposits at the central bank. The commercial banks, then, effectively have the choice of how to distribute their portfolio of claims on the public sector between deposits at the central bank, Treasury bills, short-dated bonds, etc. This latter portfolio choice will depend on expected rates of return, perceived (interest rate) risk, liquidity requirements, and so on.
In our financial system, commercial banks can’t prevent themselves from acting as a buffer to absorb any excess financing need of the public sector. In dealing with the private sector, commercial banks do have an extra degree of freedom. They can vary the terms and conditions on which they will extend loans and financing facilities to private clients. But one key element in such terms and conditions is the interest rate offered, and this latter is primarily under the control of the central bank, not of the commercial banks.
Just as the commercial banks provide a financial buffer to the public sector, so do they do so for the private sector, responding to fluctuations in the private sector’s need for both credit and money, at an interest rate chosen by the authorities.
Comparison of the Various Approaches
Both the monetary base/money multiplier and the commercial banks as creators of the money story share a common failing. They ascribe to the central bank and to the commercial banks a degree of activism in setting quantities – the monetary base and the volume of bank loans, respectively – which these institutions in practice generally don’t display. Instead, both the central bank and the commercial banks set interest rates as well as terms and conditions, such as the requisite collateral. Having set such terms and conditions, they then react in the short run, essentially passively, to the applications of both the private and public sectors for credit and money. (Note: Does this passivity enable one to describe banks as “intermediaries,” a description hotly denied by Jakab and Kumhof. This is a semantic issue that I prefer not to address here). Of course, a significant change in the volume of bank credit and deposits may make both the central bank and the commercial banks respond by adjusting their terms and conditions for making loans. But then again, the significant change in volumes may not call forth such a response, as, for example, when the central bank focuses entirely on its inflation target rather than the monetary aggregates, and when commercial banks are unconstrained by leverage or other required ratios. Indeed, commercial bank margins (relative to the official rate) and other terms and conditions tend to be, often quite strongly, procyclical, rather than contracyclical – i.e., they lower their margins in a boom to sustain market share, and vice versa in a bust.
But it is not enough just to establish the facts, though theory ought to be consistent with fact (though often it is not so in monetary economies). One ought to dig deeper to ascertain why both central banks and commercial banks act in this way. The answer is that this approach enables the central bank and the commercial banks both to achieve their own objectives and to respond flexibly to the desires of their clients.
A central bank can vary its official short-term interest rate to hit its macroeconomic objective, while, at the same time, by meeting all lending requests at that rate, handling unpredictable fluctuations in the demand for base money. The European Central Bank’s full allocation response to the eurozone banking problems is a case in point. Recall that the Federal Reserve System was set up to establish “an elastic currency.” There are many operational problems with monetary base control. Among these are that it makes the future time path of interest rates inherently less predictable and much more volatile, and that it cannot cope with the massive, unforeseeable surges in the public’s desire for liquidity during crises. Recall that the 1844 Bank Act was meant to impose an early version of monetary base control on the U.K., but then had to be temporarily lifted in each of the following crises. Remember also that the monetary base in the United States rose sharply during the Great Depression, but not enough, because of the public’s shift out of deposits and into currency.
Similarly, the overdraft system and committed facilities allow the private sector to be reasonably confident both about the future availability of funds and the likely path of future interest rates. No one knows the future, and forecasts are always wrong to some extent. But we tend to believe that we do know, even if only roughly, the authorities’ reaction, and we make educated guesses at the future path of rates. Indeed, the monetary authorities have become keen to provide “forward guidance” on this. Such partial confidence about future financial conditions would dissipate under a “narrow banking” system.
If the availability of bank loans became restricted, either by direct controls or dependent on the ability of long-term funding institutions to raise longer-term funding (as in the “narrow bank” proposals), then not only would access to such loans for the individual borrower become far more uncertain but the going rate would become more volatile and less predictable. In short, the quality of financial services provided to the general public by the banking system would deteriorate. Banks’ raison d’être is to provide attractive services to their clients, and any “reform” that worsens such services will be opposed both by banks and their clients.
Those who propose that “bank credit expansion creates money” tend to place the emphasis on the role of the banks as the active element. Here, I have emphasized the role of private sector borrowers in initiating the timing and, up to a limit, the amount of such borrowing, given the terms and conditions of borrowing. These terms and conditions are important, but a key term, the level of the short-term interest rate, is primarily determined by the central bank’s official rate.
For the rest – e.g., the margin over bank rate and other requirements, such as collateral – there are two main determinants: the extent of competition in banking and the extent of risk aversion. The greater the competitive pressures, the lower margins and collateral requirements will be, and the more expansionary (in a boom) the banking system. While more competition – e.g., from challenger banks – is good for customers, it adds to systemic risk. One reason why Australian and Canadian banks did comparatively well in the recent financial crisis was that their home banking markets were comfortably oligopolistic. While the promotion of challenger banks for the benefit of customers is all very well, a failure (as in the Vickers Report) to recognize its trade-off with systemic risk could be unfortunate.
No doubt the risk aversion of bankers has been procyclical, and this played some considerable role in the financial cycle. But it is not so clear that the procyclicality of bankers’ risk aversion was that much more marked and culpable that that of private sector borrowers, or of the authorities and of the regulators.
To conclude, the claim that banks create money when making loans is partially correct, but is only one step in a longer process. It ignores altogether the far more important prior step when the private sector initiates its decision about how much it wants to borrow, given the commercial banks’ terms and conditions, which are themselves strongly influenced by the central bank’s official short-term interest rate and the regulatory structure. Next, by placing undue emphasis on the banks’ responsibility for cycles in credit and money, this line of reasoning suggests that all would be well if only banks’ credit-creating powers were cabined and constrained – or even eliminated altogether. This ignores the fact that the operational structure of banking has developed in order to provide valuable services to clients, and that effective reform should cover other financial markets, such as housing finance, as well as banking, and also should involve a larger role for the authorities to counter the financial cycle.
About the Author: Charles A.E. Goodhart, CBE, FBA, is Emeritus Professor of Banking and Finance with the Financial Markets Group at the London School of Economics, having previously been its Deputy Director, 1987-2005. Until his retirement in 2002, he had been the Norman Sosnow Professor of Banking and Finance at LSE. Before that, he had worked at the Bank of England for 17 years as a monetary adviser, becoming a Chief Adviser in 1980. Earlier he had taught at Cambridge and LSE. Besides numerous articles, he has written a couple of books on monetary history; a graduate monetary textbook, Money, Information and Uncertainty (2nd Ed., 1989); two collections of papers on monetary policy, Monetary Theory and Practice (1984) and The Central Bank and The Financial System (1995); and a number of books on financial stability. His latest books include The Basel Committee on Banking Supervision: A History of the Early Years, 1974-1997 (2011), and The Regulatory Response to the Financial Crisis (2009)..