HUI and SPX Correlation
Adam Hamilton April 22, 2005 3044 Words
With the flagship HUI unhedged gold-stock index spending the better part of the last five weeks plunging from 225 to 180, popular fear is mounting in this realm and many bearish theories are flourishing.
Since it is ultimately the bull markets in gold and silver that drive the future profits and hence stock prices of the companies that mine these precious metals, I remain very bullish on the HUI. Gold and silver, while consolidating just like the HUI, remain above their 200-day moving averages and primary bull-market support lines. Their global supply/demand fundamentals remain very bullish as well.
If gold and silver continue to rise in the years ahead, then the profits to be made in mining them will also rise. The greater the profits of the precious-metals miners multiply, the lower their valuations will become. Eventually, when HUI valuations get low enough, even conventional value investors will pour in to reap the low P/Es and hefty dividend yields. As go gold and silver, ultimately so goes the HUI.
With the probabilities apparently in favor of another great HUI upleg erupting sooner or later here, I have been trying to combat the usual crop of bearish theories that grow like weeds during any consolidation. Last week I discussed gold futures and the week before that gold-stock sentiment and technicals. My goal with these essays is to expose bearish myths so prudent contrarians are not misled by today’s irrationally pessimistic sentiment.
This week I would like to take a look at yet another fear factor weighing heavily on contrarian hearts. It involves the correlation between the HUI and the general US stock markets.
As has become more evident in the past six weeks or so, the general stock markets are deteriorating. Based on their extremely high valuations and almost total lack of any sense of fear, the ongoing general market weakness is no surprise at all. Valuation mean reversions tend to run for 17 years or so and we are only about 5 years into this one so far. Eventually the piper is always paid for speculative excess.
While this Long Valuation Wave winter is inevitable and will fully run its course, contrarians have no problem planning around it and thriving in the midst of it. Great commodities bulls tend to soar when stock valuations are mean reverting, so capital can be diverted from bearish general stocks into various commodities-related investment and speculation vehicles.
This is all well and good, but what if commodities-producing stocks are sucked into the black hole of deteriorating general markets? What if stock-market fear spreads to commodities stocks like a virulent plague? Could precious-metals stocks become collateral damage in a general market selloff?
As all serious students of the markets know, the answer to a “could” question is always yes. Anything can happen at any time. There is never certainty that a particular course of action will or will not come to pass. As such, we must deal in probabilities. Ultimately the markets are nothing but the greatest probabilities game on the planet. The right question to ask is not “could it happen”, but instead “is it likely to happen”.
In order to investigate this alarming thesis, I examined the HUI bull market to date relative to the S&P 500 (SPX), the best major proxy for the US stock markets as a whole. The behavior of the HUI relative to the S&P 500 over the past four to five years should give us an idea of how closely related the HUI is to the general markets. While past correlation or lack thereof is no guarantee of uniform behavior extrapolating into the future, it does give us a basic understanding of the probabilities involved.
The tools used in this analysis are simple. The HUI has had 5 major uplegs and 5 major corrections/consolidations bull to date. Using daily data we ran correlation analyses on the HUI versus the S&P 500 in all 10 of these major intermediate moves in the HUI. The results are quite interesting and generally do not support the bearish thesis that stock-market downlegs are dangerous for PM stocks.
In these charts the vertical blue lines divide up the 10 major HUI moves. During each of these 10 HUI uplegs or corrections, the blue number notes the absolute percentage gain or loss in the HUI while the red number highlights the S&P 500’s performance over the same slice of time. The yellow numbers reveal the individual correlation coefficients for each major HUI move.
Interestingly both major HUI uplegs and corrections can happen during both major S&P 500 bear-market rallies and downlegs. The general stock markets don’t appear to have too big of influence at all on the HUI bull to date. It is ultimately the prices of gold and silver that drive the HUI, not prosperity or carnage in totally unrelated general stocks.
There is a lot going on this chart, so it is probably best to start with the big strategic picture and then zoom in to the tactical details. The HUI bottomed in November 2000 and has been powering higher in a secular bull market ever since. This HUI bull, like all bulls, consists of a series of powerful uplegs (numbered 1-5) punctuated by necessary and healthy bull-market corrections to bleed off short-term excesses of euphoria.
The S&P 500, meanwhile, carved a more complex macro pattern. After topping in March 2000 after a spectacular bull market born way back in 1982, this flagship index fell sharply over the subsequent three years. This secular bear period was marked by vicious downlegs punctuated by periodic bear-market rallies to keep fear from ballooning to astronomical levels.
Then stocks carved a triple bottom near 800 in July 2002, October 2002, and March 2003. In March 2003 the spectacular war rally erupted over the exuberance ignited when Washington’s invasion of Iraq went from a nagging uncertainty to a fait accompli. Since then the stock markets have been in a cyclical bull market, although momentum is rapidly fading with the S&P 500 still trading today where it was in early 2004.
So we are running correlation analysis of a secular HUI bull market versus a secular S&P 500 bear that gave way to a cyclical bull. If you are not familiar with these terms, secular generally means a primary trend running for three years or more. Cyclical means a primary trend running for more than one year but less than three years. Naturally the HUI bull should have higher correlations with the recent bullish stock-market years than the preceding wickedly bearish ones.
Since 2000, the whole enchilada, the HUI and S&P 500 have run a weak negative correlation of -0.44. This means there is a tendency, although not particularly strong, for gold stocks to be up over time when the general stock markets are down. This big picture view alone definitely refutes the bearish notion gaining currency today suggesting that HUI stocks are going to get sucked down along with the general markets.
As far as individual major-HUI-move correlations, they are all over the map. As you can see above they ran from -0.72, a fairly moderate negative correlation, to 0.88, a strong positive correlation. Interestingly if you average the individual major-HUI-move correlations, all the yellow numbers above, the result is an amazing 0.05. This means that, on average, major HUI uplegs and corrections have no correlation whatsoever with the general stock markets.
A correlation of 1.00 means that the HUI would always move in lockstep with the general stocks, rising when they rise and falling when they fall. A -1.00 reading indicates that the HUI would always move in perfect opposition with the general stocks, rising when they fall and falling when they rise. But a reading of 0.00, of course, indicates no correlation at all. Regardless of general-stock action, the HUI is equally likely to be up or down or flat. Averaging all the major move correlations showed this to indeed be the case.
Another thing to consider when pondering correlations is the r-square value. When a correlation coefficient is multiplied by itself, the r-square is the result. It explains how predictable one set of data is likely to be given the changes in the other set. A rudimentary understanding of this additional step to correlation analysis is necessary to comprehend just how weak these HUI/SPX correlation numbers really are.
If the HUI and SPX were perfectly correlated, they would have a correlation of 1.00 and an r-square of 100%. But as the actual correlation drops, the r-square falls exponentially. In market analysis I consider a 0.90 correlation to be very strong, yielding an r-square of 81%. This suggests that 81% of the price variability in one stock, index, or commodity can be explained by the variation in the other. Such high levels can be quite useful for timing trading.
But if the correlation merely falls to 0.80, the r-square plunges to 64%, a moderate relationship. 0.70 yields 49%, 0.60 yields 36%, and 0.50 yields a paltry 25%. So once correlations fall under 0.80 or so, there is probably no readily definable tradable interrelationship. At 0.70 or less, or 49% r-square, you may as well just flip a coin to decide if price A is likely to predict price B at any given moment in time or not. Below 0.70 in the markets, with scarce capital on the line, is seldom worth getting excited about.
Using this 0.70 correlation threshold, 8 of the 10 major HUI moves above had no level of useful correlation with the S&P 500. Neither did the entire dataset from 2000 nor the average of all 10 major moves. Of the 2 exceptions to this rule that did show individual major-move correlations above 0.70, together they illustrate that the HUI can do whatever it wants regardless of general stock market pressure.
Back in early 2002, the HUI rallied a massive 145% in its single greatest upleg to date. During this very short period of time, just a couple quarters, the S&P 500 fell an ugly 9% in the initial third or so of its steepest bear-to-date downleg. This peculiar period of time had the strongest negative correlation of all, -0.72. Yet still the r-square was only 52%, there were only even odds that the HUI action would be predictable by the SPX action. Even when stocks started freefalling, the HUI remained strong.
A year later in 2003, the HUI rallied again, up 125% in its second largest upleg to date. But during this time the general markets soared by 23% over the resolution of the Iraq invasion uncertainty. This resulted in the strongest correlation witnessed so far, a high 0.88 between the HUI and SPX. This yields a high r-square of 77%, indicating that 77% of the daily moves in the HUI were predictable by the general markets. This parallel rally ran about 3 quarters in duration.
Thus, out of about 18 quarters in the HUI’s bull to date, there have been just 3 where the gold stocks and general stocks have had a high positive correlation. 2 more of those quarters witnessed a fairly moderate negative correlation. This leaves 13 of 18 quarters where the correlation between the HUI and S&P 500 was so low that it had no trading significance. General-market action, bullish or bearish, is generally not a useful predictor of HUI fortunes over any given intermediate-term time horizon.
And while we are considering the two greatest HUI uplegs in this bull to date, 2 and 4 above, it is fascinating to note that they occurred in very different general-market seasons. The massive HUI upleg 2 erupted when the SPX was topping and then gradually starting to fall into a parabolic vertical tailspin into its worst downleg. But upleg 4 blasted higher when stocks were rising in their biggest rally since 2000. The HUI is demonstrably capable of strong uplegs in both bull and bear general-market conditions!
Before we delve into our final chart, there is one more point of interest here. It deals directly with today’s theory that any steep stock-market selloff is likely to hammer the HUI stocks inflicting heavy collateral damage on contrarians. The bull-to-date record really doesn’t support this assertion to any material degree. Sharp general-market selloffs can happen without hammering the HUI and the HUI can also plunge just fine on its own accord without the stock markets encouraging it.
Starting way back in 2001, the two wickedly ugly V-bounces in the S&P 500 that year didn’t seem to affect the HUI to any significant degree. The HUI just continued climbing higher when it was in an upleg and grinding sideways when it was consolidating, oblivious to the general-market fires burning around it.
In 2002, the HUI held near bull-to-date highs for all but the very last weeks of the SPX’s devastating freefall. But this damage wasn’t particularly bad, as the HUI was due for a correction at the time anyway and merely fell to a higher interim low at its crucial 200-day moving average support line. The HUI showed some moderate weakness near the second V-bounce of late 2002 and the final one of early 2003, but the actual correlation numbers are so low as to be useless.
Meanwhile in spring 2004, the HUI entered a vicious freefall of its own, slicing through its key 200dma like a bullet through cardboard. The wailing and gnashing of teeth in precious-metals land a year ago was extraordinary as folks fled for the hills in terror. Note, however, that the general markets barely moved at all while the bottom temporarily fell out of the HUI.
So bull to date we have had about a half-dozen fear-driven major V-bounces in the general stock markets but perhaps only one, and that is debatable, managed to taint HUI sentiment in the summer of 2002. Meanwhile the HUI was perfectly capable of plunging by itself regardless of stock-market action last spring. This supports my assertions that it is not general stocks that drive the HUI, but the prices of gold and silver.
Our final chart zooms in on just the recent cyclical bull years in the general stock markets. While we should expect higher correlations between a secular-bull HUI and a cyclical-bull SPX than during the SPX’s secular bear, the results are still unimpressive. Regardless of popular fears, in general the HUI is just not correlated with the stock markets to any significant degree.
Now over this entire time frame the HUI/SPX correlation ran 0.81, seemingly high. Yet, with both the HUI and S&P 500 in bull-market mode, we should really expect strategic correlations to be decent. But once again if we delve into the tactical realm where trades are actually made, the individual major-move correlations, with one exception, remain very weak.
The HUI upleg 4, running at 0.88, was the only high-correlation major move during the parallel HUI and SPX bulls of recent years. It alone is probably responsible for most of the HUI/SPX correlation since 2003. If we actually average the correlations shown above for all five major HUI moves, the result is a paltry 0.28. Taken to the r-square level, this means, on average, that only 8% of the HUI’s daily action during this period could be explainable by the S&P 500.
After this one spectacular upleg, the relationship broke down considerably. In early 2004 the HUI was grinding lower while the S&P 500 was still reaching a new interim high. By spring last year the HUI was plummeting like a meteor while the general markets continued grinding blissfully sideways in their low-volatility no-fear consolidation. After this sharp HUI correction though, it entered its fifth major upleg, almost all of which happened while the general markets continued grinding lower until the US presidential elections.
Since then the SPX has hit three subsequent cyclical-bull-to-date highs in late 2004 and early 2005 while the HUI entered a very pronounced downtrend. Even over periods of time when both the HUI and stock markets are rising in bull markets, they are just not that correlated. It isn’t the general markets that drive the HUI, but the underlying bull markets in gold and silver.
Regardless of popular perceptions today and bearish fears spawned by the HUI weakness, the reality of the hard data is that the HUI and general stock markets are really not correlated at all. There may be rare exceptions when a correlation does arise, but in the vast lion’s share of the time the HUI seems to move completely independently of the stock markets.
In light of this analysis, I would assign a low probability to the HUI getting sucked into a major general-stock selloff. Sure, it could happen, just like the NASDAQ could jump to 4000 tomorrow if enough capital bid on it, but it probably won’t. And investing and speculating is all about betting on high-probability outcomes, not fretting about low-probability threats.
And if a stock-market selloff does spark a HUI selloff, then the only ones contrarians will have to blame is ourselves. Over the short-term it is psychology and sentiment that drives markets, and if HUI players allow themselves to be spooked by a general-market selloff they may start selling anyway regardless of what history shows. But if we as a contrarian herd heed history and ignore a stock selloff, odds are the HUI will weather it beautifully.
At Zeal we have been ignoring the irrational pessimism surrounding the HUI lately and have been layering in elite gold and silver stocks in the past months, both majors and juniors. It is these very times when investors are scared that have been the greatest moments to buy historically. Please subscribe to our acclaimed Zeal Intelligence monthly newsletter today to see what stocks we are buying, when we are buying them, and why.
The bottom line is the hard correlations between the HUI and general stock markets in this bull to date are very low, mathematically uncorrelated for the most part. The HUI can rise or fall anytime the gold and silver prices drive it to, regardless if the general stock markets happen to be rising, falling, or grinding sideways.
Personally I am not concerned about a stock selloff igniting a dire HUI plunge. Sure, it could happen, but it probably won’t.
Adam Hamilton, CPA April 22, 2005 Subscribe at www.zealllc.com/subscribe.htm