“Practical men, who believe themselves to be quite exempt from any intellectual influence, are usually the slaves of some defunct economist.” – JM Keynes
The Keen/Krugman debate has largely devolved into bickering, but that doesn’t mean the message isn’t still incredibly important. My post the other day summarized the basic points, but I see the debate has moved into different territory in what appears like an attempt to distract from the original point. That original point was rather simple. Banks make loans and find reserves later. Banks are never constrained in their ability to make loans by the amount of reserves they currently carry. In a floating exchange rate system the central bank will always supply the necessary reserves in order to maintain a smoothly functioning payments system. The central bank has no choice. In other words, the money multiplier no longer applies.
Anyhow, I have no interest in getting involved in the back and forth that has resulted in little progress, but I wanted to leave the reader with an old note from JP Morgan that someone passed on to me today. It covers this point quite thoroughly. I’ll happily help readers with any questions they might have in the comments. Hopefully, some of this sinks in for people and results in a slightly better understanding of the way the monetary system works. Via JP Morgan’s Michael Feroli:
“In an increasing number of countries today, the required reserve ratio is zero. To take one example close to home, Canadian banks are not required to hold reserves. According to the simplest application of the above logic, the money multiplier there should be infinite. So if there are no currency withdrawals, a one-looney increase in reserves could lead to an infinite expansion of the Canadian banking system! Of course this is nonsense. Anyone with even a vague familiarity with modern banking would see that capital requirements would quickly bind to prevent an outsized expansion of the banking system.
This highlights a major limitation of applying the BrunnerMeltzer reasoning to current circumstances: capital ratios are not considered a constraint on bank balance sheets. When they were writing in the 1960s, this may not have been such an oversight: reserve requirements were large and meaningful while capital requirements, pre-Basel, were patchwork and less meaningful. When reserve requirements are the binding constraint on bank balance sheets, an increase in central bank reserves allows the aggregate banking system to grow. When capital is the binding constraint on bank balance sheet, an increase in central bank reserves does not affect equity capital in the banking system and therefore does not increase the system’s ability to lend.
In a country like Canada, clearly reserve requirements will never be the binding constraint. But what about in the US? While there are still formally legal reserve requirements in the US, over time they have become so watered down that they will almost never be a meaningful constraint on the ability of banks—individually or in the aggregate—to expand.
In the US, banks are currently required to hold reserves equal to 10% of transaction deposits. Cash that banks have in their vaults and ATMs counts against this requirement. Throughout the 1980s and 1990s, the scope of reserve requirements was narrowed to exclude nonpersonal or time deposits, but perhaps the most important change was in 1994 when the Fed began allowing banks to use retail deposit-sweeping programs. These software programs allowed banks to temporarily reclassify high-reserve transaction deposits into low- or no-reserve nontransaction deposits. These programs were so effective that for many banks reserve requirements fell to levels that could be entirely met by vault cash that banks would normally keep on hand anyway. In other words, many banks are no longer constrained by reserve requirements at all. While reserve requirements still exist officially in the US, the system is effectively very similar to the Canadian one. And for that reason, the money multiplier is no longer meaningful.
Another, more subtle, assumption in the money multiplier story is that reserves earn no return and so are dominated by other assets with the same risk and duration profile—such as T-bills. Therefore banks would want to shed excess reserves in favor of those competing, return-earning assets. That was true when reserves earned no interest, as was the case when Brunner and Meltzer were writing, but no longer holds now that reserves earn interest from the Fed. Of course, banks generally aren’t in the business of holding a portfolio of T-bills, but for the theory to make sense the return on excess reserves should be dominated in an apples-to-apples comparison with other similar assets. As with the effective elimination of reserve requirements, the establishment of the payment of interest on reserves is an institutional change that overturns a key assumption in the money multiplier analysis.
These institutional changes aren’t going away, and so there is no reason to expect the money multiplier to spring back. If anything, the rotation from reserve requirements to capital requirements as the more relevant limitation on bank balance sheets will become even greater in the future, as Fed policymakers have at times expressed an interest in following the global trend of dropping reserve requirements altogether, and as Basel policymakers are increasingly stressing higher capital requirements.”