How much is gold worth? Is gold in a bubble? Is it undervalued, even at $1900/oz? Is the current selloff providing a buying opportunity or a reason to sell? How can we tell?
One way to value gold is to look at long-term compounded annual growth rates (CAGRs) in inflation and money supply metrics, and use those rates to calculate the future value (FV) of gold from some traded price level in the past. Since I have not seen this type of analysis anywhere, it seemed germane and useful, as gold demand in terms of traded price level responds to both inflation and money supply growth, among other factors. Not surprisingly, inflation and money supply metrics grow at different rates over different time periods, due to ebbs and flows of economic activity and Federal Reserve (Fed) monetary policy.
The table above lists CAGRs of two inflation metrics, the consumer price index (CPI) levels reported by the Bureau of Labor Statistics (BLS) and the pre-1983 BLS CPI reported by ShadowStats.com ("SGS CPI") [1]. Also shown are CAGRs for five broad money supply metrics, M2, M3 (and M3b, which is M3 calculated by other sources [2] after the Fed discontinued its M3 series in early 2006), M3c [2], MZM, and TMS [3].
The results: Gold's Aug 2011 FV based on inflation rates range from $209-$5840, and based on broad money supply growth rates range from $539-$7141. Taking a closer look, the lows of these ranges stem from a relative market low price level at the start of the compounded period, namely 1970, when gold was artificially kept at a $35 peg (the market trading price level was slightly higher, $36), and 2000, which represented a decades market low for numerous commodities such as oil and gold, due to an exceptionally stronger dollar from a short-covering rally. The consistently low FVs based on the BLS CPI CAGR, $209-$1673, can probably be thrown out, given the specious definition for the metric after 1983 [1]. That leaves a FV range of $539-$7141. The averages over all of the FVs from this set span from $1709-$2757.
As one might expect, the wider ranges for gold FVs come from the money supply metrics, especially M3c, which contains the significant effects from recent Fed monetary stimuli programs, 2007-present. M3c ~ $17.32Trillion as of Aug 2011 [2], well exceeding nominal GDP ~$14.5T. Perhaps most telling is the FV range from long-term gold price appreciation since 1925: $3174-$7141. The lower bracket here is provided by the FV from the SGS CPI inflation rate, $945.
A few notes on the growth rates. The highest rates come from SGS CPI growth from 2000-2011 and 1995-2011 (yet we were told by the gov't and the Fed that inflation was exceedingly low over these periods!), M3c from 1995-2011 (thanks to "special" Fed monetary stimulus programs), and TMS from 2000-2011. Growth in the latter, TMS, is interesting, as TMS differs from the other metrics through its inclusion of Fed special memorandum data in its monetary aggregates database that many argue should be construed as money substitutes, while it excludes less liquid substitutes that have shown downside volatility from financial instability.
This analysis provides a realization that if gold buyers specify demand through price based on historical inflation and money supply growth, gold FV price levels can easily exceed where it is currently trading, and we have a concrete series of FV price levels to measure from.
Gold buyers should keep in mind that as we have gold sell off sharply on days such as today, Sept 22, 2011, that the selloff is driven largely by margin calls as all price levels fall in a volatile trading session in the broader commodities futures and equities markets, i.e., there is sharp deleveraging leading to forced selling. If gold's long-term trend is motivated by inflation and money supply growth rates that we have outlined here, that provides a floor for the traded gold price, as the expectation is that the Fed will continue its monetary stimulus measures as the dollar strengthens and the commodity and equity markets wane.
Author's Note, 9/23/2011: As the metals market continues sharp declines on a second day of volatile trading, gold closes at ~$1643, some 13.3% below the average high close of $1895 on 9/6/2011. Gold would have to go down to ~$1335 from this average high close to equal the peak-to-trough decline experienced in 2008 (~29.5%: $1011 on 3/17/2008 down to $712.50 on 10/24/2008), a recent period of comparative selling. On a longer scale, gold selloffs exceeding the 2008 decline occurred between 1980-1982 (~65.1%: $850 on 1/21/1980 down to $297 on 6/18/1982) and 1983-1985 (~44.2%: $509 on 2/15/1993 down to $284 on 2/25/1985). A 65% decline from $1900 is $663, below the lowest low of the FVs based on the SGS CPI CAGR shown in the above table. The 2008 and 1980-82 declines resulted in sharp reversals, while the 1983-1985 decline didn't yield any significant price appreciation until gold hit a multi-decade low in July 1999 ($252) after meandering around a average price of $355 (sigma ~ $52) between 1985-2004, before taking off in 2005.
[1] The SGS CPI mimics the inflation metric reported by the BLS before 1983. Starting in 1983, the BLS CPI was changed to de-emphasize food and energy inflation and allow for hedonic adjustments and substitutions, among other changes. The SGS CPI tracks the original definition, which maintains a measure of the cost of living needed to maintain a constant standard of living. HERE is a great article discussing just how misleading the current BLS CPI is.
[2] M3b is estimated by NowandFutures.com HERE. M3c is essentially M3b on steroids: it includes the money supply growth estimated from the Fed's various special liquidity and monetary easing programs since 2007, namely TAF, PDCF & TSLF, QE,... .
[3] TMS is "true money supply," which is M2 minus small time deposits and retail/institutional money funds, plus demand and time/savings deposits due to foreign banks/institutions, U.S. gov't demand/note balances at banks and depositories, and U.S. gov't general account balances at the Fed. TMS seeks to track standard money plus money substitutes ("perfectly secure and immediately convertible, par value claims to standard money"). Up-to-date TMS is tracked HERE. Whether TMS is better and more accurate than M2 or M3 or MZM at tracking broad money supply will be discussed in a later article.
No comments:
Post a Comment