TMS or M3? A Reply to Paul van Eeden

By: Steve Saville | Tue, Jun 10, 2008
Print Email

Below is an extract from a commentary originally posted at on 8th June, 2008.

A few weeks ago Paul van Eeden (PVE) posted an extremely bearish outlook on bonds that he justified, in large part, by the rapid expansion of M3 money supply. We responded that while we are long-term bearish on bonds (we expect bond yields to move much higher over the coming 5 years), we thought that PVE's premise was wrong. Our reasoning: M3 is a poor indicator of monetary inflation, whereas a vastly superior monetary aggregate, namely the True Money Supply (TMS) developed by Murray Rothbard and Joseph Salerno, reveals a relatively slow rate of monetary inflation.

PVE has since responded to our response, and in doing so has raised some good points that we'll address in today's report. Before we do, though, it's worth noting that we never mean any disrespect when we critique another analyst's work. In fact, we usually won't take the time to read, let alone comment on, articles by analysts we don't respect. Furthermore, we have absolutely no problem being taken to task by others who consider our analysis to be off the mark, as long as the 'taking to task' is done in an objective manner. And we certainly aren't averse to changing our opinion when the evidence suggests that a change is appropriate. For example, we considered M3 to be a reasonable indicator of monetary inflation until early this year, at which time the large divergences between different monetary aggregates prompted us to delve much more deeply into what should, and should not, be counted in the money supply. The research we did during the first few months of this year convinced us that a) flaws in the compositions of M3 and MZM were causing these broad measures of money supply to generate 'major league' false signals, and b) TMS, while perhaps not ideal, was a much better measure of money supply.

Turning our attention to the above-linked PVE article, the first thing that deserves to be clarified is that TMS is NOT suggesting deflation. There cannot be deflation as long as the money supply is expanding and TMS has increased by 4.8% over the past 12 months. Also, TMS's current 12-month rate of change is the highest since April of 2005, so TMS is increasing AND its rate of change appears to be in the early stages of a new upward trend.

To make sure there's no doubt as to where we stand on this issue we reiterate that we are strongly of the view -- a view consistent with the performance of TMS -- that the US will experience inflation at varying rates for many years to come. The current difference between M3 and TMS is not the difference between inflation and deflation; it's the difference between extremely high monetary inflation and moderate monetary inflation.

Next, PVE acknowledges that TMS is a better measure of MONEY supply than M3 when he states: "TMS was constructed with the intention of defining a true monetary aggregate that could tell us something about the true nature and extent of Americaís money supply. M3, on the other hand, includes components of the banking system, such as money market mutual funds, that are clearly not money and which make M3, strictly speaking, not a true "monetary" aggregate." So, despite the to and fro it seems that we are in agreement regarding TMS's superiority as a measure of total US money supply.

After acknowledging that TMS is the better measure of money supply, PVE goes on to explain why he will continue to use M3. In a nutshell: he will continue to use it because it is a better predictor of events within the real economy and the markets.

We disagree that M3 has been the better predictor and will soon explain why, but in any case this is a separate issue. To say that TMS is the better measure of money supply and that M3 is the better predictor/indicator of the economy and the financial markets is to say that changes in the supply of money aren't the only things that affect the economy and the markets. This, of course, is true.

We'll now deal with the contention that M3 has been a more useful economic/market indicator.

A lot of PVE's analysis relies on long-term comparisons between money-supply growth rates and changes in money purchasing power, with changes in money purchasing power represented by year-over-year changes in the Consumer Price Index (CPI) calculated by John Williams at It is quite likely that this version of the CPI comes closer to reflecting reality than the CPI reported by the government, but even an honest attempt to come up with a single number that represents the economy-wide average price level will necessarily fail because the whole concept of an economy-wide average price is nonsensical. We've explained why in several prior reports/articles, including the one posted HERE. Unfortunately, not everything worth measuring can actually be measured.

Because PVE's comparisons of different monetary measures and the CPI are based on the false premise that it is possible to determine a single number that consistently/accurately reflects changes in the purchasing power of money, there's no need for us to specifically address the conclusions he draws from these comparisons.

This could be a case of beauty being in the eye of the beholder, but when we look at long-term charts showing year-over-year changes in TMS and M3 we arrive at the conclusion that TMS has generally been a better predictor than M3 of various economic and market-related trends. With reference to the following charts showing the year-over-year percentage changes of TMS and M3, here's why:

1. During 1978-1979 the TMS growth rate plunged from +8% to -5%. This suggested that the commodity boom was close to an end, which proved to be the case. M3's growth rate pulled back during this period, but remained fairly high and therefore failed to signal an end to the 1970s commodity boom.

2. TMS's growth rate exploded upward between November of 1981 and July of 1983, peaking at an extraordinary rate of around 50%. This suggested that a new inflation-fueled boom was beginning, which again proved to be on the mark (the stock market boom began during the second half of 1982). At the same time, there was nothing in M3's performance to indicate that monetary conditions had gone from 'tight' to 'easy'.

By the way, to answer a question posed by PVE near the end of his article: the explosive growth in TMS during the early 1980s was driven to a large extent by the shifting of money from time deposits (not included in TMS) to savings deposits (included in TMS), which was, in turn, related to the realisation that interest rates had peaked.

3. M3's growth rate fell steadily during the first three years of the 1990s, prompting some analysts to warn of deflation and to become very bearish on the stock market. At the same time, TMS's powerful growth suggested that the monetary backdrop was becoming easier and that the monetary stage was being set for a strong stock market. TMS proved to be correct.

4. TMS's downturn during 1993-1994 'predicted' the equity bear market of 1994.

5. There was strong growth in both TMS and M3 during 1998-2003, leading to the inflation-fueled booms in commodities, real estate, mortgage lending, debt securitisation, and the emerging markets.

6. The sharp decline in TMS's growth rate during 2004-2006 suggested that the most vulnerable of the above-mentioned inflation-fueled booms would burst. TMS's performance is therefore consistent with what actually happened to the real estate and credit markets. On the other hand, M3's acceleration suggested that the real estate and mortgage-lending markets were not going to run into major difficulties. Again, TMS transmitted the more correct message.

7. As PVE notes, TMS is much more volatile than M3. This, however, is not a weakness of TMS because in this respect TMS is simply reflecting reality. The sad truth is that under the current monetary system the supply of money not only increases way too fast over the long-term, it does so in a very haphazard and volatile way.

8. The relationship between interest rates and money-supply growth is complex because major changes in long-term interest rates are not directly determined by monetary inflation; they are determined by inflation expectations. For example, there was plenty of monetary inflation during the 1980s and 1990s, but during these two decades the main beneficiary of the inflation was the stock market and rising stock prices are never perceived to be indicative of an inflation problem.

M3 Chart Source:

As noted above, the downturn in TMS's growth rate during 2004-2006 and its relatively minor rebound over the past 18 months pointed to problems for the housing industry and over-extended credit providers. These problems are likely to continue. There's also a significant risk that the reduced rate of money-supply growth will be the catalyst for a sizeable downturn in commodity prices over the coming months, although the commodities market, being the international variety, is influenced by changes in GLOBAL monetary inflation. Unfortunately, at this stage we don't have a global equivalent of TMS.

We aren't offering a free trial subscription at this time, but free samples of our work (excerpts from our regular commentaries) can be viewed at:



Steve Saville

Author: Steve Saville

Steve Saville
Hong Kong

Steve Saville

Regular financial market forecasts and analyses are provided at our web site:

Also, samples of our work (excerpts from our regular commentaries) and additional thoughts on the financial markets (and other stuff) are provided at our blog:

Copyright © 2003-2016 Steve Saville

All Images, XHTML Renderings, and Source Code Copyright ©