Three Business Spectator readers contacted me directly about one topic last week – bank money creation, and how bank reserves work. Following an old journalism adage that three direct enquiries about a topic from the public means that everybody’s interested in it, I’m diving into wonkdom to answer their queries in detail here. Ignore this post if the adage isn’t true for you, but if it is and you haven’t yet had your morning Java, now’s the time for that stroll to the barista.
OK, caffeinated? Here we go.
The standard story about how banks create money, and how reserves work, is the "Money Multiplier Model”. Money creation starts with the government injecting "fiat money” into the economy – say by giving a welfare recipient $100 in cash. That recipient then deposits the cash in a bank, which hangs on to a government-mandated fraction of it (the "Reserve Requirement”) – say 10 per cent or $10 – and lends out the rest to a borrower. The borrower then deposits that $90 in another bank, which does the same thing – hangs onto 10 per cent of the $90 or $9, and lends out another $81 to another borrower.
The process repeats ad infinitum, and in the end a total of $1,000 is brought into existence: the original $100 in cash, plus $900 in credit money created by the private banking sector (matched, of course, by $900 in debt).
This alleged system, known as Fractional Reserve Banking, is seen as "fraud” by Austrian economists, and by many in the public. To inflationists, because Bernanke has hit the printing presses, dramatically increasing Base Money, and therefore money in circulation will soon explode, leading to hyperinflation.
Figure 1: US Money Base
To Neoclassical economists, it’s just the way banking works: bank lending is controlled by the Fed because, "even if banks hold no reserves”, Fed control over the currency means that private banks must do what the Fed wants.
And to anyone who’s done empirical research, it’s a myth.
The most recent proof of this is in an excellent discussion paper from Federal Reserve economists Carpenter and Demiralp, entitled "Money, Reserves, and the Transmission of Monetary Policy: Does the Money Multiplier Exist?”. The first clue that it doesn’t exist is given by their abstract, which notes that, "before the financial crisis, reserve balances were roughly $20 billion”. If the textbook model were correct, the total stock of money in the USA would be $200 billion, versus the multi-trillion dollar level of even a narrow definition of the money stock. As the authors note, this makes a mockery of the textbook "Money Multiplier” model:
M2 averaged about $7.25 trillion in 2007 … bank loans for 2007 were about $6.25 trillion... if we consider the fact that reserve balances held at the Federal Reserve were about $15 billion and required reserves were about $43 billion, the tight link drawn in the textbook transmission mechanism from reserves to money and bank lending seems all the more tenuous.
I’ll stop there on trashing the conventional model – save to quote Carpenter and Demiralp’s conclusion that "the textbook treatment of money in the transmission mechanism can be rejected” – and get to the real issue: if the Money Multiplier model doesn’t really describe how money is created, and how reserves figure in this, what does?
The short answer is "endogenous money”: bank lending creates deposits, so the decisions of banks to provide loans determine the level of money, and reserves are largely irrelevant. But today I want to attempt a longer answer that actually puts reserves in the picture, so I’m going to model (take a swig of coffee) a system with 3 banks: a "Buyer Bank” where a Buyer has both a deposit account and a credit card (or line of credit); a "Seller Bank” where a Seller has a deposit account, and a Central Bank that keeps the Reserve Accounts of both banks.
It’s an incomplete and unrealistic model because the money flow goes only one way – from Buyer to Seller, and therefore from Buyer Bank to Seller Bank – and therefore Buyer Bank must ultimately run out of Reserves. But it’s still effective in showing the basics of a more complete model in which money flows both ways.
Starting at Buyer Bank, the Buyer can purchase goods either using cash – which means transferring money from his Deposit account to the Seller’s Deposit account – or by using his credit card. Breaking these processes down into discrete steps, the Cash purchase first of all involves money being taken out of the Buyer’s Deposit account (shown as +Card because, from the bank’s point of view, the Buyer’s Deposit is a liability of the banks and therefore recorded by it as a negative amount), which also reduces Buyer Bank’s Reserves (shown as a minus because Reserves are an asset of the bank, and they now fall). Notionally, the amount of money in circulation has fallen at this point (imagine that the amount of Cash is currently in the limbo of interbank settlements between Buyer Bank and Seller Bank).
How’s the coffee?
Next buying on credit is shown in two stages: firstly accessing the line of credit by the amount Card, which increases the Buyer’s Deposit account and increases the bank’s loans by the same amount. At this point, the amount of money in circulation has risen by Card. But then the purchase is acted upon, so the amount of Card is deducted from the Buyer’s Deposit account (yes it’s confusing showing it as a plus – it’s taken me a while to get used to this! – but it makes sure that all rows sum to zero) and also deducted from Buyer Bank’s Reserves. So by this point, the two transactions have reduced the amount of money in circulation by the amount of Cash.
Table 1: Account operations at Buyer bank
|Assets||Liabilities||Row Sum||Money Change|
|Buy with cash||-Cash||+Cash||0||-Cash|
|Buy with credit||-Card||+Card||-Card|
Enter Seller Bank. The amounts Cash and Card are added to both the Seller’s Deposit account, and the bank’s Reserves. These actions increase the amount of money in circulation by the amount Cash+ Card. In total, the amount of money in circulation has increased by Card: money has been endogenously created.
Table 2: Account operations at Seller Bank
|Assets||Liabilities||Row Sum||Money Creation|
|Deposit Cash Revenue||+Cash||-Cash||0||+Cash|
|Deposit Card Revenue||+Card||-Card||0||+Card|
However, this creation of money has to be intermediated by the Central Bank. So how do these operations look on its Balance Sheet? From the Central Bank’s point of view, it simply has to affect a transfer of Reserves from Buyer Bank to Seller Bank, by debiting Buyer Bank’s Reserve account and crediting Seller Bank’s.
Table 3: Account operations at Central Bank
|Assets||Liabilities||Row Sum||Reserve Creation|
|Loans||Buyer Bank Reserves||Seller Bank Reserves|
|Buy with Cash||Cash||-Cash||0||0|
|Buy with Card||Card||-Card||0||0|
But what if Buyer Bank doesn’t have enough Reserves – if it’s at its Reserve Requirements limit already, or worse still, if its reserves are zero? Will the Central Bank refuse to transfer funds that Buyer Bank doesn’t really have?
I hope the answer to that question is now obvious: of course it won’t. The Central Bank will either give Buyer Bank time to find the Reserves, or lend them to it. To do otherwise – to refuse to transfer Reserves from Buyer Bank to Seller Bank – would void the purchase made by the Buyer from the Seller (and note that this could happen with the Cash purchase just as easily as with the Card one). The system of commerce would break down. We’d have an interesting social system the instant after the Central Bank did such a thing, but it wouldn’t be called capitalism.
I hope this also explains why, in every country in the world where Reserve Requirements exist (and that’s not every country – Australia, for one, doesn’t have them), they are backward-looking: they depend on the level of deposits existing in the previous reporting period "and thus after banks have extended the credit demanded by their customers”. It should also explain the comment made by Alan Holmes over half a century ago that the Fed has "little or no choice” about doing this:
- In the real world, banks extend credit, creating deposits in the process, and look for the reserves later. The question then becomes one of whether and how the Federal Reserve will accommodate the demand for reserves. In the very short run, the Federal Reserve has little or no choice about accommodating that demand.
Finally, what about "looking for reserves later”? There are two obvious sources: Buyer Bank can either borrow them from Seller Bank, or from the Central Bank itself.
The Central Bank has an obvious motivation to lend: it doesn’t want to destroy capitalism. But why should Seller Bank lend to Buyer Bank in this case? Because it now has excess reserves relative to its Reserve Requirement: the transfer of funds from Buyer Bank increased its Reserves by precisely as much as Seller’s Deposit account rose, and since the Reserve Requirement is a fraction of deposits (10 per cent of household deposits only in the USA – there is no reserve requirement for corporate deposits), it now has a lot more Reserves than it needs. Better to lend them at the interbank rate to Buyer Bank than to get no return (or a very low return) on the Reserves themselves.
On the Buyer Bank’s accounts, the two operations look like this:
Seller Bank sees it this way:
And the Central Bank records the action this way:
Notice that I’m showing these reserve operations as not changing the amount of money, whereas the private bank operations did. That’s because money is the liabilities of the banking sector to the non-bank sectors of the economy.
Money could be created – and demand increased – if Buyer Bank decided to lend some of these liabilities to the Non-Bank public (or the government), as illustrated in Table 7 – and that’s why the Fed has been trying to cause inflation via its "printing press”.
But for that to actually happen, there have to be willing borrowers out there – and with the massively overindebted private sector we now have, willing borrowers are few and far between (as, for that matter, are willing lenders too – better to earn a few shekels from the Fed than risk losing money with the public). And that’s why those excess reserves are just sitting there.
Figure 2: Excess Reserves data from the St Louis FRED database
I am sure I will be told that actual bank operations are far more complex than this, and that is true. But at a very parsimonious level, I think this post captures the basic principles of how banks endogenously create money, why the Central Bank has no choice but to accommodate that behaviour, and why as a result it will be a cold day in Hell before Ben Bernanke’s printing press causes inflation. Other factors are far more likely to do that long before the Federal Reserve’s pushing on a string has any impact.
Carpenter, S. B. and S. Demiralp (2010). Money, Reserves, and the Transmission of Monetary Policy: Does the Money Multiplier Exist? Finance and Economics Discussion Series. Washington, Federal Reserve Board.
ECB (2012). Monetary and Financial Developments. Monthly Bulletin May 2012. Brussels, European Central Bank.
Holmes, A. R. (1969). Operational Constraints on the Stabilization of Money Supply Growth. Controlling Monetary Aggregates. F. E. Morris. Nantucket Island, The Federal Reserve Bank of Boston: 65-77.
Krugman, P. (2012). "Banking Mysticism, Continued." The Conscience of a Liberal HYPERLINK "http://krugman.blogs.nytimes.com/2012/03/30/banking-mysticism-continued/" http://krugman.blogs.nytimes.com/2012/03/30/banking-mysticism-continued/.
O'Brien, Y.-Y. J. C. (2007). "Reserve Requirement Systems in OECD Countries." SSRN eLibrary.