Thursday, September 22, 2011

Valuing Gold From Inflation and Money Supply Growth Rates


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 ("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 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.


Thursday, September 15, 2011

Global Debt Watch: $95T and Counting

TotalMktDebt BIS Dec 89 Dec 10

As of Dec 2010, worldwide marketable/tradable debt outstanding neared some USD$95Trillion, according to the Bank of International Settlements (BIS). The historical data above indicates that debt markets have more than doubled from Dec 2002 to Dec 2010, with the largest increases stemming from domestic issuances, at first (2002-2007) from mortgage and asset-backed security issuances, and then more recently (since 2008), from sovereign government issuances. The United States maintains the largest debt market, at some $32.5T, ~35% of the worldwide total market [1].

Gauging and tracking marketable/tradable debt is key to understanding global capital market stability. Though total debt levels can also contain "nonmarketable" debt, such as nonmarketable sovereign government debt, it is the marketable debt that has the greater systemic influence across debt, equity and derivatives markets, since market participants price and trade that debt; however, the influence of nonmarketable debt levels should not be understated. I have started to maintain a "Global Debt Watch" page HERE, with the intent of providing historical trends and data analysis at regular intervals from a variety of international sources.

The fantastic growth in the debt markets has been assisted by three primary factors: (a) the reduced borrowing costs made possible by central bank monetary easing policies worldwide; (b) the aggressive use of short-term funding markets, such as the repurchase agreement (repo) and commercial paper (CP) markets, to borrow cash short and buy longer-dated, higher-yielding debt; (c) government policies that promote debt issuance, and government-sponsored entities (GSEs) that "back" such issuances. Government sanctioned credit rating agencies have also been a factor in the growth of debt markets.

The growth and decline of the repo and CP markets in the last decade coincide with that of the growth and decline of mortgage and asset-backed securities (MBS/ABS) and collateralized debt obligations (CDOs), structured pools of MBS/ABS. The total repo market size between the two largest markets, U.S. and Europe, stood at approximately $12.6T Dec 2010, after falling from a 2008 high of $17.5T [2]. Unlike the CP markets, the repo markets are not reported in the BIS debt data above; repo markets are (usually) very fluid, with the majority of transactions composed of overnight or very short-term maturities. The relative opacity of repo markets, plus their vulnerability to liquidity issues due to collateral quality, counterparty risk and capital cushions, make tracking these markets imperative to gauging stability and "systemic risk." Conceivably, the more transparency in the repo markets, the better able the system would be to handling (greater) market liquidity dislocations, such as that experienced from the credit crisis of 2007/8. However, such transparency does not solve the problem of debt accumulation sponsored by central bank monetary and sovereign government fiscal policies.

[1] The Securities Industry and Financial Markets Assoc. (SIFMA) estimates U.S. debt markets at $35.5T, ~37% of the total worldwide market. SIFMA includes offshore centers and CDOs issued in USD.
[2] I assembled these estimates from two sources, the NY Federal Reserve (Fed) and the International Capital Market Association (ICMA). The NY Fed data only reports primary dealer repos, from a survey of Fed primary dealers, and does not count private OTC repos handled by bank holding companies. That may likely increase the U.S. total by some 30%, according to BIS.


Global Derivatives Watch: Netting, Collateral and Capital Cushions Matter

DerivMkt BIS Dec 87 Dec 10

In an attempt to demystify the derivatives markets, which have endured an abundancy of malign from both somewhat credible and uninformed critics, I have started to maintain a "Global Derivatives Watch" page HERE, with the intent of providing historical trends of these complex markets along with data analysis at regular intervals. Not only do I trade these markets (namely the options, futures and forex markets), but I see enormous value in derivatives to (a) hedge risks with a defined loss and (b) place bets with a defined loss, when structured properly and controlled with defendable margins and prudent capital controls.

The first fallacy that pervaded the derivatives markets, only to lead to the road to ruin for certain participants in these markets, was that there was ever such a construct as "risk free." Let me assure you that there is nothing of the sort. The second fallacy was that due diligence was not necessary on the derivatives counterparties with which a participant executes a trade, particularly of the sort that includes high leverage or risky collateral, and also particularly on custom over-the-counter (OTC) contracts that don't trade through a clearinghouse or on an exchange. Most equity, futures and forex options and forwards trade on exchanges and have benefitted from the transparency and liquidity. Nearly half of all interest rate swaps (IRS), the largest derivatives market by almost an order of magnitude over all the others in terms of gross market exposure, are cleared through a central clearinghouse, where netting and collateral control can be verified. To be sure, there are many specialized, one-off OTC derivatives that will still be executed between parties directly or through a dealer, but that doesn't minimize the importance in netting, monitoring of underlying or backing collateral quality and value, and capital cushions should a trade go south. (The poster child for failure to execute these tenants is AIG, which should have never been bailed out for its incompetence and risk mismanagement.)

As I like to say, derivatives don't bankrupt people, people bankrupt people. (Sound familiar?)

As the chart above shows, the derivatives markets have grown exponentially over the last two decades, signifying the demand for the ability to hedge and define risks. Much of this demand is due to heightened risks in markets that result from uncertainties in macroeconomic factors, such as interest rate risk, foreign exchange rate risk, systemic default/credit risk, and futures curve (commodity price) risk. Though the credit default swaps (CDS) markets have taken the brunt of the criticisms from an array of punditry, and have spawned one too many sensational bestsellers filled with overblown rhetoric, these instruments do serve a purpose in hedging both macro and microeconomic risks, particularly specific to securities, indexes of securities, companies, institutions and governments.

Just so I don't come off as a rep for one of the industry or lobby groups for derivatives, I will provide my own criticisms of the derivatives markets thus: not everything is known about these markets to gauge systemic stability in the event of gross risk shocks (dislocations) that in turn might bring about market liquidity issues (i.e., selling panic) and lead to serious market degradation and inadvertent losses. Of particular interest is the study of interest rate swaps when rates correct markedly higher or experience extreme volatility, as a result of market panic and inviting more market panic. Understanding the vulnerabilities in markets (not just derivatives, but also equity and debt/bond markets) is valuable and the role of market transparency, disclosure and visibility critical to the task.