A popular topic lately has been the fact that copper has fallen into bear market territory.  At least, that’s true if you subscribe to the idea that a bear market equates to a 20% fall from a recent high.

Since copper is used in so many industrial applications, then many believe that a dramatic fall in copper prices reflects an economic slowdown, and by extension a likely future decline in stock prices.

We’ve discussed “Dr. Copper” several times in the past, and the correlation between extremes in copper prices and future returns in stock prices is poor.  It’s just not usually a very good predictor.

But let’s go back over the past 25 years for which futures prices are available and look at how the S&P 500 fared after other times copper fell rapidly into a bear market.  By this, we’re looking for a decline of 20% or more within a two-month period, when its highest point was very near a 52-week high.

Out of the 6 instances, only 1 led to an imminent and dramatic fall in stock prices.  The occurrence in November 2007 did ultimately lead to a bear market in stocks, though the S&P rallied fairly strongly for a month or so beforehand.

The other occurrences were not so bearish.  In fact, they were outright bullish for the S&P.  By six months later, the index was up by more than +12% each time.  For whatever reason, perhaps this latest bear market in copper will be more predictive of a future failure in stocks…but its history has not been consistent.

Below are charts of each of the occurrences noted above (click for larger view).

There has been intense interest from traders, regulators and politicians related to the performance in some commodities, particularly silver.  In response, the margin requirements to hold some futures contracts, again particularly silver, have been increased.

Exchanges supposedly set their margin requirements using computer algorithms based on historical and implied volatility, but the fact of the matter is that the ultimate decision is discretionary…and they are not immune to political pressure.

The question now isn’t whether margin requirements on silver are going up – they already have, astronomically – the question is what impact it will have going forward.

What kind of dent in sentiment does such a hike tend to have?

Unfortunately, getting historical margin requirements has proved to be exceptionally difficult.  I have about 10 years’ worth for most commodities, but only 7 for silver.  It’s not for lacking of trying, or expense.

Assimilating all the research, the conclusion boils down to this:

Margin increases tend to greatly lower trading volumes and open interest, and they stabilize price action to where it was before the event that triggered the margin change.  That holds especially true in markets like silver where speculators are a larger influence.

None of the studies suggested that price action itself was consistently impacted by margin changes.  It was only trading volume and open interest that were influenced in statistically significant way.

The chart below shows silver futures, along with the initial margin requirement for speculators, which obviously has gone parabolic.  Aggregate open interest has, true to historical precedent, fallen during that time.

One potential difference this time is that the studies are based on the idea that speculators in silver futures are “jump on the bandwagon” types.  But based on data from the CFTC, we’re not seeing that this time.  Large speculators are not that net long the futures, and didn’t ramp up their exposure as it rose (possibly due to the introduction of exchange-traded funds that are easier and more liquid than futures).

The ETFs may throw a wrench into what the exchanges are trying to do, but historically these margin hikes have led to lower volume and lower volatility.  For those who are suggesting that it necessarily means a decline in silver, they may have a point…but it’s not supported by any empirical research that I can find.  Silver prices will fluctuate based on their own merit, and not based on arbitrary margin requirements.

For those who want more detail, here are the relevant conclusions from the two most topical studies:

“In summary, there is strong evidence that (1) the exchange used margin requirements to affect market activities; (2) margins were usually raised if the trading pattern had been heavy; and (3) except for the open interest in the 1979-1980 period, changes in margin requirements were effective in stabilizing the trading pattern prevalent prior to the change.”

 - Margin Requirements and the Behavior of Silver Futures Prices, Christopher K. MA, G. Wenchi Kao and C.J. Frolich

“There is a clear causal negative margin effect on trading volume and open interest. Trading activity moves away from…the metal with the higher…margins.  [a]10 percent increase in margins is associated with a drop in average trading volume of 1.38 percent, a drop in average open interest of 1.51 percent, and a drop in the growth rate of open interest of 2.96 percent.”

 - Margin Requirements, Price Fluctuations, and Market Participation in Metal Futures, Journal of Money, Credit & Banking, Gikas Hardouvelis and Dongcheol Kim

Hardouvelis and Kim (1995) examine metals futures contracts. They exclude silver contracts during the Hunt Brothers squeeze period arguing that the special circumstances of the period would contaminate their sample.

They found strong evidence that high metals futures margins reduce contract open interest, but no evidence that higher margins attenuate contract price volatility.  Indeed their evidence suggests that any empirical margin-volatility relationship in the data owes to the prudential activities of the futures clearinghouses. Higher volatility leads the clearinghouse to increase margins, but the increase in margin has no measurable effect on contract price volatility subsequently.

On May 4th, we looked at the extremely pessimistic sentiment towards the US Dollar.  Then on May 6th, we discussed the big one-day rally, and its implication when coming off of a 52-week low.

The buck has continued its rally, and has now crossed above its 50-day moving average for the first time since January.  That’s four long months of being below.

Let’s go back to 1973 and look at every other time the Dollar had spent at least four months below its 50-day average, then finally managed to scramble back above it.

In the shorter-term, the Dollar’s performance was OK.  At least it was (barely) better than random.  After that, by three months later, it showed a positive return only 31% of the time.  The percentage of time positive improved by a little after that, but the median return continued to worsen.

The chart below shows a rolling 1-month, 3-month and 6-month correlation between the Dollar and the S&P 500 since the 2009 market bottom.

The correlation between the two has been very negative lately.

The thing about correlations, though, is that they’re not static.  They usually oscillate from one extreme to the other.  We can see that clearly on the chart – other than a stretch during the fall of 2009, when the correlations hit one extreme or the other they tend to start moving the other way.

Right now, the 1-month and 6-month correlations are extremely negative and have just started to reverse.  That does not necessarily mean that the Dollar and S&P will start moving in the same direction at thesame time (a positive correlation), but it does suggest that they relationship between the two will probably become a little less predictable.