Tuesday, March 2, 2010

Is the Fed Undereporting the Money Supply?

I have received a number of emails pointing me to the analysis by James Turk of the money supply. Now, ZeroHedge has posted the story.

ZH is a hardworking group of bloggers led by "Tyler Durden". They do break some stories and analytical views. I read them daily, but they do throw a lot of stuff up against the wall, especially when it gets to economics. The problem is they are not economists, they are traders, so when it comes to pure economics, the stuff they throw can sometimes, well, miss the wall completely. The Turk piece is an example of such.

James Turk contends that the money supply is being undereported. To reach this marvelous conclusion, he uses some keen insights coupled with non sequiturs and odd definitions that seem to float about in the universe without any known connection to the universe, Big Bang connection or otherwise.

Here's a Turk insight:

As the financial crisis has unfolded over the last two years, the Federal Reserve has been responding in a variety of unprecedented ways.

This is true. He goes on:

The quantity of dollars in circulation is being underreported (sic)by relying upon the traditional and now outdated definitions used to calculate M1 and M2.

This he does not prove.

He states:

Cash currency of course is simply printed, but every note issued is recorded on the Federal Reserve’s balance sheet. Basically, the Fed ‘monetizes’ an asset by turning it into currency.

If, for example, a bank sells a $1 million T-bill to the Fed, the Fed ‘pays’ for it with $1 million of newly printed cash currency. The Fed records the T-bill as an asset and the cash currency it issued as its liability. These Federal Reserve Notes are the “currency” component in the definition of M1 above.

The creation of deposit currency is similar. When a bank makes a loan or purchases a security, it records the loan or security as its asset and creates deposit currency as its liability. Simple bookkeeping entries increase the bank’s assets and liabilities by the same amount.

New deposit currency is created because the bank deposits the amount of the loan in the borrower’s checking account, or similarly, credits the account of the seller of the security it is purchasing. These dollars are now available on ‘demand’ of the borrower or the seller of the security.

This description is a little shaky, but it is not where I want to focus, so I will just let it ride.

He then throws out this whopper:

For historical reasons unimportant to the point of this analysis, the Federal Reserve in the past has only created cash currency.
I have no idea where he gets this from. Does he think when the Fed buys Treasury securities, the Fed just backs a truck up to the seller and dumps pallets full of newly printed bills? Turk is just way, way off here. (But, we can charitably say he is still in the universe at this point.)

He then goes on:

However, the unprecedented changes it has engineered over the past two years have resulted in a vast amount of deposit currency being created by the Fed. Instead of purchasing paper from the banking system solely with cash currency – its traditional form of payment to ‘monetize’ assets by turning them into currency – the Federal Reserve since the start of the financial crisis has increasingly relied upon deposit currency to purchase paper.

Regardless how the Federal Reserve pays for the paper it purchases – cash currency or deposit currency – it is creating dollar currency and perforce expanding the money supply.

There is a difference in how much money is ultimately created depending upon if, what Turk calls "cash currency", or "deposit currency," is in issuance, but again this covers a direction I am not interested in discussing, just know there is a difference that Turk doesn't seem to get, namely that banks don't lend against money in circulation that is out in the economy, but they sure do loan against reserves deposited at the Fed, what I think Turk means by "deposit currency".

Now, let's get into the expanded universe section of Turk's "analysis". At one point he writes:

The total amount of deposit currency in checking accounts is the “demand and other checkable deposits” component in the definition of M1 above.

After saying that "deposit currency" is found in "other checkable deposits" in M1, he writes:

More to the point, none of this deposit currency is captured in the traditional definition of the Ms.

Ah, M1 would be one of the M's. But let's also look at his statement that "other checkable deposits" are somehow "deposit currency." According to the Fed's statements, other checkable deposits are "Other checkable deposits at commercial banks(4) and thrift institutions (5)." The footnotes state that:

4. NOW and ATS balances at domestically chartered commercial banks, U.S. branches and agencies of foreign banks, and Edge Act corporations.

5. NOW and ATS balances at thrift institutions, credit union share draft balances, and demand deposits at thrift institutions.

An ATS accounts and NOW are what is being reported here (An ATS is a kind of NOW account for corporate, as opposed to individual accounts). But, I see no connection between any definition I can think of that Turk might be using for "electronic currency" that matches with "other checkable deposits". Also keep in mind that Turk writes that the Fed has in the past used only what he calls "cash currency", but that since the crisis they have started to use "electronic currency":

For historical reasons unimportant to the point of this analysis, the Federal Reserve in the past has only created cash currency
He then produces a chart of what he calls "electronic currency", but it is a reference to reserves at the Federal Reserve. So is all this Turk talk about "electronic currency" just another name for bank funds on deposit with the Fed, i.e. reserve funds? I think so but, you really can't tell for sure, since Turk really is all over the place. It is my best guess, though, as to what he most often means by "electronic currency". Notice, though, that he does not differentiate between required reserves and excess reserves.

He then produces a chart showing that M1 money supply is $1,716 billion (he references the Fed data for this number). He then shows in this chart that when you add "electronic currency" to M1 the money supply is $2,918 billion which is an increase of $1,206 billion (Which I think is the bank reserves he is referencing).

With all due respect, this is the most confusing, bizarre analysis I have ever seen. And I have read Keynes, so I have seen confusion.

It really seems to be that Turk, as I say above, is really discussing reserves on deposit with the Fed, but a large portion of this is excess reserves, which means it is sitting at the Fed and not circulating or impacting the economy. A topic I have discussed many times using conventional terminology.

There is no undereporting of the money supply--at least any based on Turk's analysis.

Class dismissed.


  1. I found your site some time ago while browsing ZH; probably the best info I ever got from there. They are a hard working bunch and I still read it, but all articles seem to eventually circle back to the joys of goldbugs fondling their physical silver and gold.

    I've learned to appreciate reasoned economic analysis much than tin foil hat-based paranoia.

  2. Wenzel - if you're going to get all "professorial" on us I must remind you of the dangers of the "ivory" tower syndrome. Some real world subjects are so complex they compel simplifying assumptions in an attempt at analysis. But when the assumptions begin to define the subjects, the analysis loses its insight.

    I agree Turks terminology gets a little fuzzy and you are probably correct that he has stumbled upon the increase in reserves that you more accurately discussed. However, I am interested in the notion (which I think Turk is struggling with) that the money supply is difficult to define and impossible to measure.

  3. It doesn't matter. They can print all they want. Banks have few worthy people to lend to and the public does not want to borrow. The cash at the FED will be used to pay off bank losses on CRE and other defaults coming up. In essence, a wealth transfer. The losses being transferred to government, meaning the public, which will be felt later down the road through currency debasement and taxes.

  4. Anonymous - here's why I think it matters. the FED's M's (M1, M2, etc.) are not the only things that people can use for money. Some identify this with a shadow econony or black market. But while the FED is waving this bright shiny credit creating machine in our face, the market is making its own adjustments which are not precisely measurable. So, is the interest rate too high or too low? Does inflation exist or does deflation reign? If the money supply is unmeasurable and there are no fixed relationships in economics (all things are variable) there is no way to know the answer to these questions. And this makes it very difficult to make investment decisions.