Tremors in the Gold Bubble

This fall I commented more than once (1, 2) on how frothy the gold craze has gotten, and explained why I think the metal equivalent of a tulip bulb in 1636 is in a bubble that could result in a dramatic crash—but due to a lack of necessary resources (software and manpower), I haven't been able to provide the quantitative analysis to back up my observations. Fortunately a team from the Research Institute of Mining Geomechanics and Mine Surveying (VNIMI) and the Russian Academy of Sciences (PAH) has taken up the problem, and published their working results this month.

The paper, entitled “Log-Periodic Oscillation Analysis and Possible Burst in the ‘Gold Bubble' in April–June 2011”, by Sergey Tsirel, et al., describes how the team applied Didier Sornette‘s approach to predicting crisis points in financial markets to the price of gold for various time periods, focusing on the accelerating climb that started in 2003. Moreover, they compared gold to the price pattern preceding the 2008 crash in oil (which I identified as a bubble nearing its end-stage about two weeks before the crash occurred) and concluded that towards the end of the bubble

…the temporal distance between the critical moment and the time of the [latest] observation first increases. However, approximately two months before the crash of the oil price, the position of the critical point stabilizes, and the interval between the critical point and the actual bubble burst turns out to be about 1–1.5 months.…

We do not see a clear stabilization of the forecasted critical point date [in gold], but [there has been a] discontinuation of strong oscillation and a gradual approach to the (approximately) finite value during 2010.…Various methods of [extrapolation] indicate that…the most probable time of the gold bubble burst is May 2011.

My own chart comparing recent bubbles illustrates the pattern, but doesn't predict when it will end:

To further reinforce their prediction, Tsirel's group superimposed the log-periodic price oscillations in gold on a hyperbolic trend found in socio-economic processes. This method places the critical time point in the first week of July 2011. Putting together all three analyses, the authors predict that the gold bubble is at greatest risk of collapsing in May or June of 2011, and that the price per ounce will top out in the $1500–$1600 range.

Why shorting gold now is a bad idea

Despite the accuracy of many (though by no means all) of Sornette's analyses and the additional checks that Tsirel, et al. performed against their primary analysis, a crash in gold is far from inevitable. Before all you gold bugs out there lash back with citations of bubble predictions that were wrong, let me repeat a few disclaimers from Sornette's writings and the Tsirel team's paper:

• A critical point merely predicts when a market will be most vulnerable to a crash, not that one will necessarily happen.
• If traders begin to believe bubble predictions, prices may level off and enter an orderly consolidation rather than a crash.
• Government intvervention—by the Federal Reserve or China's central bank, for example—can drastically alter how a bubble ends.

Regarding the overall economic impact should gold actually crash, Tsirel, et al. write that major players' huge losses could set off a crisis that propagates through financial markets and ultimately inflicts serious damage on the world economy. “On the other hand,

investments in gold, which caused the very “gold rush”, also lead to the diversion of funds from stock market investments and to the reduction in the production of goods and services. If at the time of the collapse some promising areas of investment appear in the developed and/or developing countries, [investment capital] can move to those markets, which…will contribute to the production of new goods and services and accelerate the way out of the crisis.

Not being prone to selling short myself, I'm rooting for a leveling-off rather than a crash. Nonetheless, I definitely would not want to be a buyer of gold right now.