How the Banking System Really Works

Keith Brofsky | Photodisc | Getty Images

Very few people understand how the modern banking system really works.

They have in their heads a model they learned from text books in which banks take deposits from customers, then lend out those deposits as loans. In reality, banks fund their loans by borrowing in the interbank market.

Once a bank has agreed to make a loan, it then borrows the same amount of money in the interbank market at a slightly lower rate. The lending comes first, the borrowing to fund the loan comes afterwards. This is why so many loans are pegged to LIBOR: Banks charge borrowers rates that are set to levels at some point above what the banks themselves pay to borrow.

A very similar misconception applies when the government spends and borrows. People imagine that the government must first collect taxes or borrow money in order to have funds to spend. In reality, the government just spends what it wants, and then collects taxes in order to balance out the effect the spending has had on the money supply.

In short, banks lend first, fund later. Governments spend first, fund later.

There's a great discussion of this in the Harvard International Review's interview with Bill Mitchell, the research professor in economics and the director of the Centre of Full Employment and Equity at the University of Newcastle, Australia. Mitchell is one of the founders of a school of thought called "Modern Monetary Theory" (MMT).

Here he describes the differences between what's known as the "New Keynsian" school and MMT:

Well, the New Keynesian paradigm is built upon a series of false premises that affect policy prescriptions. False premise number 1: government has to borrow to fund spending. False premise number 2: there’s a fixed supply of savings available at any point in time. False premise number 3: the government, by borrowing from that fixed supply of savings, denies private sector borrowers those funds, and competition for those funds drives up interest rates.

MMT says the following:

There is no finite pool of savings in the economy. Savings is a function of national income. When you have rising national income, you have rising savings. So if government spending stimulates economic activity, and thereby (gross domestic product) and national income, savings will rise simultaneously. That’s the first part of the story.

The second part of the story is that private sector borrowing is not dependent upon a fixed supply of savings. The concept of a bank in the New Keynesian model is that the bank sits there waiting for depositors to come with their savings, and only once the bank attracts those deposits is it in a position to lend. In other words, the New Keynesian conception is that banks are constrained by their existing reserves. In reality, however, banks always have the capacity to create loans for credit-worthy borrowers because they can always get more reserves. Banks can get reserves from a number of sources, but at the end of the evening the banks know they can cover their reserves by borrowing from the central bank. So the conception of banking in MMT is much different from the stylized treatment in New Keynesian economics.

The third story is what happens when the government runs a budget deficit. What happens in the money market is as follows: the U.S. government buys something from the private sector. It pays the manufacturer, who then pays the workers. A whole range of transactions follows from that initial government purchase. All of those transactions work their way through the system and find their way to the reserves of the banks each day. Typically—though not at the present because we are in an extraordinary situation where the central bank is paying interest on reserves—those reserves would just sit there and earn zero interest for the banks. And so typically, as I’ve explained before, banks try to get rid of those reserves, driving down the interest rate in the interbank market in the process. What you can understand from that is that budget deficits, independent of any monetary operations, drive interest rates down, not up. This is the complete opposite of what orthodox economists claim is the case, and it’s confirmed by the present combination of record low interest rates and very large budget deficits.

Your really should go read the whole thing.

What I would add here is that MMT insight can be strengthened by paying closer attention to the distorting effects of government spending. The U.S. government does not just buy "something" from the private sector. It buys particular goods and services — goods and services whose purchase is demanded by bureaucracies, special interests, and politicians. This means that the added demand introduced into the economy is not added neutrally, it is added sequentially, beginning with the politically favored enterprises. The transactions do not "work their way through the system" — they cluster around the politically favored enterprises and spread out through other enterprises in the order of their proximity to the original beneficiaries of government spending. Real resources are drawn toward them, rendered unavailable to parts of the economy that are not politically favored.

This may not be a big deal in the short term, especially when there is a lot of idle resources (unemployed people, for example), but it has long-term implications. In the first place, resources drawn to political enterprise cannot easily or immediately be drawn back into the private sector. There are various frictions that lock them in place. Second, the private sector may continue to underperform because it perceives a lack of resources to supply expansion due to the fact that government spending has prevented prices from falling to levels at which private demand can return.

Nonetheless, it is far better to think about these things clearly, and without illusionary risks of bond vigilantes and other such hobgoblins of so much economic thinking.

Questions? Comments? Email us

Follow John on Twitter @

Follow NetNet on Twitter @

Facebook us @