Tuesday, September 23, 2008

Currency is a Lagging Indicator

Various measures of money supply have been popular over the past four decades, but only two of them gained prominence in the stock market for any period of time, M1 and M2.

M2 - Milton Friedman's Favorite

The money measure designated M2 consists of currency in circulation, demand deposits at commercial banks and savings deposits at commercial banks. Over reasonably long time frames M2 did go up and down along with the economy and inflation. The problem with M2 is that the lead time was insufficient to use as an economic forecasting tool. Therefore, the markets never truly focussed on the M2 number when it was released.

The reason M2 has little forecasting value is that savings deposits are held at the discretion of the saver. If stocks were flying, people might cash in their savings and move into the market. The Fed could not force an increase in savings deposits, though if inflation rose for a sustained period of time, savings deposits would naturally increase.

M1 - The Market's Choice

In the 1980's the weekly release of the M1 number, which consisted of only currency in circulation and demand deposits at banks, was a big deal. If the number released on Thursday afternoon was larger or smaller than expected, the bond markets would react strongly, as would stocks when they opened for trading on Friday mornings.

And M1 did have some forecasting value. A sustained increase in M1 usually forced the Fed to react by either tightening further, or switching away from a previous period of easy money. If the increase in M1 went on too long, the economy usually rebounded, and if it persisted longer yet, inflation would start marching upward.

The Monetary Base - The Logician's Choice

The monetary base consisted of three factors, currency in circulation, total bank reserves, and an adjustment factor supplied by the Fed of St. Louis which accounted for the impact of previous changes in reserve requirements. The monetary base also had a good track record for anticipating future Fed moves, economic activity and the rate of inflation.

Effective Reserves - The Best of All

By the year 1976 I was convinced that the Monetary Base was the better measure of the three discussed above. Then one night it dawned on me that Currency (the cash in our pockets, in the cash registers, in the mattresses, etc.) should not be considered a leading indicator of anything. This is because we generally don't anticipate higher prices by carrying more currency. Instead, we react to past price increases by carrying more currency because we start running out of cash sooner than we used to. Why? Because prices have already risen! Currency, then, is a lagging indicator of inflation, and since inflation lags economic activity by a year or even two to three years, clearly currency did not belong in a forecasting measure.

Bear in mind that currency is a major component of both of the better monetary forecasting tools, the Monetary Base and M1.

So, I took the Monetary Base Series and subtracted currency out of it, leaving just the total reserves plus the adjustment factor, the combination of which I called Effective Reserves.

Effective Reserves - A Once-Great Forecasting Tool

When I plotted Effective Reserves versus three measures, the stock market, GDP (GNP at that time) and the inflation rate (in the form of the GNP Deflator) the strongest relationship I obtained was the one where Effective Reserves forecast the direction of GNP almost exactly six months in the future. For example, if the economy was doing fine, and then Effective Reserves slowed to zero or negative growth for an extend period, the reported economic statistics would start to flash recession almost exactly six months after the initial slowing of Effective Reserves. Similarly, if the economy was in recession, those same statistics would start to flash recovery almost exactly six months after Effective Reserves had bottomed out and started growing again.

Things Change - Unfortunately

Now, when I came up with all this I was the manager of the Bond Portfolio of the Tennessee Consolidated Retirement System so I had access to what most of the economists on Wall Street were saying about money supply measures, the economy and inflation. As I wrote in the Monograph, only one Wall Street economist appeared to be homing in on the concept and that was a couple of years after I initially formulated it. Then the rules changed drastically and NOW accounts were introduced. This, for reasons explained in the Monograph, changed the game completely and Effective Reserves could no longer logically be a useful forecasting tool, nor could M1, for that matter, as people were dumping demand accounts (which paid no interest) in favor of the new NOW accounts.

Demand Deposits - A Useful Present-Day Measure

Over time reserve requirements have been reduced so low (a mistake, in my view) that I no longer trust them as a forecasting measure. However, if M1 is adjusted by removing Currency, you get just Demand Deposits which still exist because businesses carry non-interest bearing deposits for cash settlement purposes. While I am certain that the old Effective Reserves measure was an excellent forecasting tool, there seems to be some value in Demand Deposits now that they've settled down again. (It took years for the shift from Demand Accounts to NOW Accounts to work their way through the system.) Simply put, if Demand Deposits peak and then start falling, we stand a reasonable chance of entering a recession about six months later. Similarly, if Demand Deposits have been growing strongly, no recession is indicated. However, this statement has to be severely tempered because the banks every now and again introduce a new mechanism that allows companies to utilize their cash balances more efficiently, e.g., sweep accounts.

The monograph explains in more detail the lag times involved, and I really suggest that one refer to it if these posts have piqued curiosity. Here's the link again: The Monograph on Monetary Policy

In the next post I'll discuss the relationship between Effective Reserves and the GNP Deflator that I observed in the 1950-1980 data.

Next Post: Forecasting the Inflation Rate

No comments: