Positive Economic Commentary
January 4, 2000
Is Greenspan "Spiking" the Punch?
Former Fed Chairman William McChesney Martin is credited with the comment that the Fed's job is to take away the [money and credit] punchbowl before the [economic] party gets out of hand. The recent surge in bank credit and the broad M3 money supply suggests that, if anything, Fed Chairman Greenspan has been spiking the monetary punch rather than taking away the bowl. (I wish I had come up with this "spiking" analogy, but I did not. If I could remember where I read it recently I would give proper citation credit to the originator.)
The chart below shows the explosion in bank credit (loans, exclusive of interbank loans, and investments of commercial banks located in the US) and the M3 money supply (currency/travelers checks, liabilities of depository institutions excluding bond/debenture liabilities and money market mutual fund liabilities). In the 8 weeks ended December 22, 1999, bank credit growth surged to an annualized rate of 23.4 -- eclipsing the 23.0% growth in the 8 weeks ended November 2, 1998. M3 growth has soared to an annualized rate of 21.8 in the 8 weeks ended December 20. This surpasses the prior peak M3 growth of 14.8% set back in the 8 weeks ended November 30, 1998.
The surge in bank credit and M3 growth back in the fourth quarter of 1998 was due, in part, to the "seizing up" of the money and capital markets in the aftermath of the Russian government's default on its debt and the insolvency of the Long-Term Capital Management hedge fund. Because it became inordinately expensive for corporations to borrow in the money and capital markets, many of them tapped their bank credit lines for funding. This increase in bank loans -- assets on the books of banks -- was matched by an increase in banks' liabilities, which are included in the M3 money supply.
But what is responsible for the current accelerated growth in bank credit and the M3 money supply? The money and capital markets did not seize up in the fourth quarter of 1999. The yield spread between Baa corporate bonds and 10-year Treasury securities appears to have fallen below 200 basis points in December (based on Fed data through December 29), its narrowest this year. A narrowing Baa-Treasury 10-year yield spread is inconsistent with a seizing up of the corporate bond market. In the two months ended November 1999, growth in commercial paper net issuance was the strongest since the two months ended March 1998. Again, this is just the opposite of an indication of money markets seizing up.
Some analysts are ascribing the current surge in bank credit and M3 money supply growth to Y2K effects. One of these Y2K effects is the increased demand for currency by nonbank entities in the event that ATM machines would not work or credit card purchases could not be transacted as the calendar flipped to 2000. Banks' demand for currency, or vault cash, also has increased so that they could honor customers' desired currency withdrawals. Indeed, both the nonbank public's and banks' demands for currency have shot up in the past two months. In the 8 weeks ended December 20, the annualized growth in currency held by the nonbank public was 21.3%. This is the fastest growth in the currency component of the money supply over an 8-week interval since the 22.6% growth back in January 28, 1991. Did the M3 growth rate also soar into the low 20s in the 8 weeks ended January 28, 1991? No. The best it could do was reach 7.7%. What about bank credit growth? Did it jump into the 20s? Hardly. In the 8 weeks ended January 30, 1991, the annualized growth in bank credit was just barely above zero at 0.2%. In the past 15 years, the correlation between 2-month growth in currency and 2-month growth in the M3 money supply is minus 0.17. In contrast, the correlation between 2-month M3 growth and 2-month bank credit growth is positive 0.51. The historical record, then, does not support the argument that strong currency growth implies strong M3 growth. The two are negatively correlated, not positively correlated as the Y2K currency explanation for current strong M3 growth implies.
As mentioned above, the nonbank public is not the only entity that has increased its demand for currency as Y2K approached. Depository institutions did too. In the 8 weeks ended December 29, depository institutions' holdings of currency increased at an annualized rate of 561.2. This is the fastest growth in vault cash since January 1975, the inception of the data series. Could this recent surge in vault cash account for the recent surge in M3 money growth? Again, the historical record suggests not. The correlation between the 2-month growth rate in vault cash and the 2-month growth rate in M3 is minus 0.1 for the past 15 years, or essentially no correlation.
There are conceptual reasons why there is no historical record of a positive correlation between either currency held by the nonbank public or currency held by the banks and the M3 money supply. If the nonbank public wants to hold more currency, how does it get it? By running down its deposit balances. So, if the nonbank public wanted to hold $100 more in currency, it might withdraw $100 from its savings accounts. Would this $100 increase in currency holdings by the nonbank public in and of itself increase the M3 money supply? No. It would merely change the composition of M3. There would now be more currency in M3 and fewer deposits. So, an increased demand for currency by the nonbank public does not in and of itself increase the M3 money supply. If banks held no excess reserves to meet these currency withdrawals, then the Fed would have to inject reserves to keep the fed funds rate from rising above its target in the face of these currency withdrawals. If the Fed increased reserves equal to the currency withdrawals by purchasing government securities from nonbank dealers, then M3 might increase. But the increase would not be some multiple of the increase in currency held by the public.
Will an increased demand for vault cash on the part of banks in and of itself increase the M3 money supply? Again, no. Banks can get more vault cash by running down their reserve deposit accounts at the Fed. This merely changes the composition of bank reserves toward more vault cash and fewer reserve balances held at the Fed. If banks' demand for vault cash is above and beyond the reserves they are required to hold, this will result in banks' excess reserves rising. It could be that the M3 money supply would be boosted by an amount approximately equal to the increased demand for excess reserves by banks in the same way that an increased demand for currency by the nonbank public could increase M3, as explained in the paragraph above. That is, M3 could be increased if the Fed creates these excess reserves by purchasing government securities from nonbank dealers. But because banks want to hold these excess reserves, not lend them out, there would be no multiple expansion of the money supply from this Fed injection of excess reserves.
In the 8 weeks ended December 20, the total of M3 (not adjusted for seasonal variation) is up $244.9 billion, but the currency component of M3 (again, not adjusted for seasonal variation) is up only $23.2 billion and surplus vault cash is up only $12.8 billion. This means that at a minimum, about $209 billion of increased M3 is not being accounted for by increased currency demand on the part of banks and the nonbank public.
The principal factor accounting for M3 growth now is the extension of credit by banks and other depository institutions, either through loans to the nonbank public or the purchase of government securities. This is reflected in the chart above showing the tandem acceleration in M3 growth and bank credit growth in the past 8 weeks. Could this acceleration in bank credit growth be Y2K related? If it were, the strongest growth in bank loans would likely be concentrated in the industrial and commercial category. When businesses borrow from banks to finance their inventories, it is this category in which this borrowing usually shows up. So, if businesses were increasing their inventories as a precaution against Y2K-related supplier delivery interruptions, we might expect to see an acceleration in commercial and industrial borrowing relative to other categories of bank loans.
But that's not the case now. As the chart below shows, growth in both consumer and real estate loans has been outpacing growth in commercial and industrial loans. How is the surge in real estate loans related to Y2K? Were people rushing to their banks for a mortgage at the end of 1999 out of a fear that banks would not be open on January 3? Is the rise in consumer loans related to an increased demand for bottled water and Spam? Maybe, but I doubt it.
But if you really want to see rapid growth in bank lending, take a look at was has happened to loans related to securities transactions. The chart below shows that growth in banks' securities-related loans has skyrocketed in recent weeks. In the 8 weeks ended December 15, 1999, these loans grew at an annualized rate of 1,275.2%! Growth moderated to 743.4% in the 8 weeks ended December 22. Was the increase in these loans Y2K related? Doubtful. More likely stock market margin and IPO.com related.
In sum, the recent explosion in the M3 money supply appears to have more to do with a fundamental underlying strong demand for credit in this country than Y2K issues. If so, then this rapid growth in M3 will not reverse simply with the abatement of Y2K concerns. Rather, it is likely that the Fed is going to have to raise interest rates more to curb the demand for credit. Although Fed Chairman Greenspan might not exactly be spiking the money and credit punch himself, it sure appears as though he is looking the other way as others do. I'm not sure I would want him to chaperone my kid's high school graduation party!
Paul L. Kasriel
Chief Domestic Economist |
