Monday, June 10, 2013

The Money Problem

Last month, I wrote “Why Monetary Stimulus Is Broken”, in which I argued that (1) the money supply is not responsive to growth in the Fed’s balance sheet because banks are hoarding cash instead of lending; and (2) that NGDP is not responding to the (modest) growth in the money supply because of a high liquidity preference at households and businesses. (I did not and will not today go into the whole expectations problem, which is also very important.)

Of the two transmission vectors, the one that seems most broken is that between the monetary base (the Fed’s balance sheet) and the money supply (M2). The monetary base has grown exponentially since Lehman, from roughly $1T to $3T, while M2 has only grown from roughly $8T to $10T. The Fed would appear to be pushing on a string, and monetary velocity continues to decline.

Banks are leaving the Fed’s money on deposit at the Fed instead of lending it. (A bank deposit at the Fed is not money, it’s a bank asset; money is a bank liability which is created when a bank makes you a $1 million loan and deposits $1 million into your checking account, which is your claim on the bank. Now you have $1 million, and the money supply has risen by $1 million.)

There are clear linkages between bank loan growth and money growth, and between overall credit growth and nominal growth. Let’s take a look at these two variables: bank loan growth, and overall credit growth.

Bank Credit Growth (FRB H.8)

Bank credit (now $10T) was growing around 10% before the Crash, contracted  during the crash, rebounded to 6% after the Crash, but then slowed down to the current anemic 3.5%. This is not suggestive of robust monetary or economic growth. Total bank loans and leases, a slightly smaller slice ($7.3T), shows a similar pattern: 12% growth pre-Crash, a sharp contraction, recovery to 5% then declining to 3% at present. Bank credit growth is not robust.

Total Credit Growth (FRB Z.1)

Next we’ll look at the Fed’s Flow of Funds database to observe total credit growth. Looking at the broadest possible credit aggregate (TCMDO) which is now $57T, we see precrash growth of 10%, a mild contraction, and a current, rather weak, 3% growth rate. Now let’s look at the four components: households, business, government, and finance.

This is not a pretty picture, and would seem to be at the root of the problem. HH credit ($13T) was growing at a feverish 12% precrash, then it began to contract and has continued to do so ever since. It is still contracting, which goes a long way to explain what’s wrong with the recovery.

The Business picture is much brighter than the HH sector. Business credit ($9T) was growing at an overheated 13% precrash, contracted during the crash, but has since recovered to a very healthy 9% growth rate. Business credit does not appear to be broken.

Credit to governments has been the mirror image of the private sector. As private sector credit contracted during the crash, government credit took up the slack. Regional government credit ($3T) grew rapidly during the crash, but has since stopped. Federal credit ($12T) grew very rapidly during the dark days after Lehman (35%) and is still growing at a 10% rate.

Most people exclude the financial sector from an analysis of domestic credit growth because it is not an important component of the economy, it is a derivative of the real economy at best, and it is rife with double-counting. Nonetheless, I think that it worth a look, because it has its own (unhappy) story to tell. The pattern of financial sector credit growth bears a striking  resemblance to the HH sector. Once again, there was a feverish precrash growth rate of 13% followed by a very sharp contraction which has not yet ended. Banks and the Street are still licking their wounds.

The net of all this is that bank loan growth (which is a monetary policy transmission vector) is very weak, and is in fact slowing.  Overall credit growth, which plays an important role in determining nominal growth, is also very weak. While business credit has fully recovered from the Lehman shock, household credit continues to shrink and thus retard the recovery.

The Fed’s policy of relying on the banking system to stimulate growth isn’t working. This is reminiscent of Japan which has had the same problem for two decades. Buying bonds from banks creates bank reserves, not money. And it would appear that however great the banks’ reserves get, they are not creating inflation.

It may be that when rates are at the zero bound and the banking system is broken, the appropriate policy instrument may not be to buy bonds from banks, since buying them doesn’t seem to affect the price level. Bernanke was certainly correct that the Fed could create inflation by dropping money on citizens from helicopters, but that would be a rather blunt instrument. It seems to me that the Fed needs to buy something besides Treasury and agency bonds, and from someone besides banks. The obvious alternative to Treasuries would be foreign government bonds, or gold. Since the former would constitute a “currency war”, that would seem to leave gold. Gold is not a bank asset, it's a "civilian" asset. Buying gold from nonbanks creates money and inflation.

I have no doubt that if the Fed were to announce that it will buy gold until it has achieved 2% inflation and 6.5% unemployment, it would get there. It would disrupt the gold market (and enrich some of the wrong people) but that is a small price to pay. No foreign government could object to the Fed buying gold; it’s been doing it for 100 years.

That was an idle thought-experiment, in the sense that buying gold has not even been discussed at the FOMC or anywhere else. But I am quite confident that it would work better than the current policy of buying Treasury bonds from banks. And it is not unprecedented.


comet52 said...

How would the Fed buying gold improve the employment picture, exactly?

Christopher T. Mahoney said...

Employment growth is a function of real growth, which is a function of nominal growth. We will never get to 4% real growth with 3.5% nominal growth. The best way to get to, say, 6% nominal growth is to either target nominal growth or a derivative, such as employment, which the Fed has done. The optimal policy instrument for creating nominal growth is something that creates inflation, which bond-buying clearly isn't doing. It could be anything at all, such as apartment buildings, but gold is a conventional policy instrument that is guaranteed to create inflation. I see no better alternative.

comet52 said...

I can see massive Fed gold buying inflating the gold price, but everything else? Gold rose 6-fold from 2001 to 2011 and everything else didn't do that.

Christopher T. Mahoney said...

I don't know what the price elasticity of gold is, but there is a lot of gold out there. I think that the Fed can create inflation without an excessive impact on the real price of gold. The Fed can smash any attempt at a corner with occasional sales.