Lately there has been somewhat of a debate between some Market Monetarists and some Austrians about the so-called Cantillon effect. I have not participated in the debated – last time I wrote about the Cantillon effect was actually in my Master thesis on Austrian Business Cycle theory in the mid-1990s and to be frank I have not made up my mind entirely on this discussion. Therefore, I am happy Justin Merrill have written a guest post for my blog on the topic.
As it is always the case I do not necessarily agree with what the authors of guest posts on my blog write, but I always hope that guest posts can help further the debate about monetary policy and theory issues. I believe that Justin’s post is doing exactly that.
Please enjoy.
Lars Christensen
Guest post: Cantillon and Central Banking
by Justin Merrill
Much has been written recently on the topic of Cantillon effects. I risk alienating myself by potentially disagreeing with everyone, but I hope I can persuade others to see it my way. At the extremes there are two points of view. The rational expectations view basically asserts that money is neutral when inflation is expected, and therefore the Cantillon effects can largely be ignored. On the other end of the argument is the exploitative theory of state money, whereby the issuer of the currency, the banks, the politically connected, and the government employees and contractors benefit at the expense of everyone else. While both of these have nuggets of truth, they also have flaws that make them inapplicable to our current monetary system.
After a careful rereading of Sheldon Richman’s article, I agree with it almost entirely with the exception of one part:
“Since Fed-created money reaches particular privileged interests before it filters through the economy, early recipients—banks, securities dealers, government contractors—have the benefit of increased purchasing power before prices rise.”
Cross out “government contractors” from the above sentence and it is true. This is because contractors are paid out by the Treasury, which gets the vast majority of its funds from taxation and financing. This confuses fiscal transfers with monetary ones. Since the Treasury does not print money directly a la greenbacks, the government does not receive a 100% seigniorage from central bank operations. A lot of Rothbardians also make the mistake of assuming that all bank credit expansion represents a windfall gain to the issuing bank and the borrower, but this ignores that banks have to pay interest to their depositors and it is not costless to expand.
But are the Market Monetarists right that there are not Cantillon effects from open market operations since primary dealers are selling bonds to buy reserves? Not quite. The primary dealers do benefit by getting the privilege of selling securities at a premium to the market rate and buying at a discount. The gain might be small, but it exists; otherwise PDs would not have an incentive to participate in OMOs. This is effectively a risk free arbitrage that doesn’t arise out of entrepreneurial awareness, but political connectedness and size. Libertarians may overstate the size of this privilege, but it shouldn’t be ignored. Interest paid on reserves to banks and Federal Reserve profits turned over to the US Treasury are also direct injections of new money. The total amount is roughly $100 bil a year (Fed’s reported profits plus its deposit liabilities times .25%), not including the unknown arbitrage gains that PDs make from trading.
Imagine that we are on a gold standard and the price of gold is $1,000/oz. A gold miner may have a cost of production of $990/oz. He therefore earns $10 of purchasing power by adding to the outside money supply.
Now imagine that in our current system a PD buys a security for $990 and sells it to the Fed for $1,000, he also earns $10 of purchasing power.
In both cases of base money expansion, there are Cantillon effects, but the second one is largely due to legal privilege. Once the outside money enters the banking system, there are secondary Cantillon effects. Banks’ cost of capital is lowered, increasing their profit margins. They may increase their investments, benefiting prior holders of said investments, or they may increase their lending, putting new money in the hands of the borrower and shortly into the hands of the person selling the financed asset.
The way that monetary policy transmits can be difficult to predict through the credit channels. Different firms have varying access to capital markets. Small businesses rely on bank and trade credit, while large firms may be able to issue commercial paper at extremely low rates. In a credit crunch, small firms will be cut off from credit even if the Fed is aggressive with monetary policy. The beneficiaries, relatively speaking, will be primary dealers and issuers of commercial paper that can borrow near zero and lend through trade credit at high rates of interest.
This is one reason I am skeptical of NGDP targeting. I do not think output and asset prices can be kept in equilibrium through central banking. Even though I prefer NGDP targeting to inflation targeting, I think stable NGDP is a probable outcome of monetary equilibrium, not an end in itself. The transmission mechanism between reserve creation and output and inflation is messy and unpredictable.
In summary:
1) Don’t confuse fiscal and monetary transfers.
2) Austrians may sometimes be inarticulate at explaining Cantillon effects in our current system (a person counterfeiting money in his basement is not the same as OMOs) but it still exists.
3) Cantillon effects will exist in any system, but their magnitude and consequences are dependent on the institutions.
4) Money is not injected in a “helicopter drop” and should not be assumed neutral through rational expectations.
5) Non-neutral money can create malinvestment through the banking system and credit channels and it matters what the central bank buys and its impact on the yield curve.
References:
Inside the Black Box: The Credit Channel of Monetary Policy Transmission by Bernanke and Gertler
The Market Monetarist 