Gold and Gold Stocks - More Divergences Emerge

by: Acting Man

Initial Reaction to the Payrolls Report

The October payrolls report looked strong on the surface, but that wasn't the whole story. It seems the headline number was essentially the only thing good about it (for a detailed write-up see Mish's summary). While 204,000 jobs were added according to the establishment survey, 735,000 were lost according to the household survey. The latter number was influenced by the pseudo-shutdown, but the BLS says it should have been even higher, as many of the furloughed bureaucrats apparently realized they were merely on a paid vacation and therefore didn't refer to themselves as jobless in the survey. The main reason why the unemployment rate rose only by one tenth of a percentage point was that the labor force participation rate collapsed to levels last seen 24 years ago. In spite of already being at an extremely low level, the participation rate has gone into something of a free-fall lately.

(Click to enlarge)

Labor force participation rate since 1977 – a full round-trip.

The chart annotation needs to be amended actually: such a steep decline in the participation rate hasn't been seen during previous recessions either (never mind recoveries). It continues a trend though that has begun at the end of the technology mania in 2000. Incidentally this is one more reason to suspect that the year 2000 market peak must be regarded as the "orthodox" top, in addition to the fact that the market peaked against gold at the time and reached record high valuations. The slump in the participation rate is the major reason for the decline in the unemployment rate since 2009. People who are deemed to no longer be looking actively for work are not considered unemployed. However, if they are not unemployed, then what are they? There is a certain degree of demographic drag reflected in the participation rate, but it is a good bet that most of those no longer counted as unemployed didn't choose leisure over work voluntarily.

The market focus on the payrolls report makes little sense, especially in view of how often the numbers get revised before they are considered definitive. On the other hand, the numbers are held to be an important input in the Fed's deliberations over the "appropriate" degree of monetary pumping. The gold market tends to react to the headline numbers rather than the innards of the jobs report, at least initially. However, as noted previously, the initial knee-jerk reaction is probably less important than what happens in the week after the release (e.g. the strong June payrolls report was initially sold, but a rally started in the following week).

(Click to enlarge)

December gold contract, daily. With the sell-off on Friday, gold has returned to an area of technical support (blue lines = support, red lines = resistance). The sell-off after the June jobs report was followed by a sizable rally, so the activity in the week after the report is probably more telling than the initial knee-jerk reaction.

Is there any good reason to look for the possibility that that action that follows on the heels of the June payrolls report may be repeated? Only one, but it is a fairly good reason, even though the daily chart of gold as such is not exactly inspiring much confidence.

The next chart shows a close-up of the activity on Friday with the HUI index superimposed. This was the third occasion since the October FOMC decision that has seen the gold stocks diverge significantly from the metal, and this time the divergence was quite notable, as the gold stock indexes managed to close in positive territory on a day when gold lost more than $20.

Whether or not the buyers will be proved correct remains to be seen, but it is clear that there is lately interest in the sector that wasn't in evidence previously. Consider in this context that the opposite happened around the time of the gold peak in 2011: the gold stock indexes at the time frequently diverged negatively from gold.

(Click to enlarge)

December gold 5-minute chart – the action in the HUI (bar chart) on Friday is superimposed. While gold went sideways near the day's lows, the index opened near its lows and then proceeded to close quite strongly.

A caveat to all this is that the rest of the stock market was very strong on Friday as well, and that may have lent some support to the sector. Gold stocks often don't really care much about what the broader market is doing, but that is not a hard-and-fast rule.

Gold Stocks, Technical Conditions

On a daily chart of the HUI, Friday's action doesn't look particularly impressive, but it has to be seen in context. The context is provided by the HUI-gold ratio, which has risen to a minor interim high.

Note that the HUI needs to overcome the once again declining 50 day moving average, which has stopped all bounce attempts over the past week. Incidentally, there has been no change in the GDM bullish percent index (it remains stuck at the 30 level, which means that 30% of gold stocks are currently on a point & figure buy signal). We will be posting that chart again once there is a notable change.

(Click to enlarge)

HUI, daily. On Friday a small reversal candle was put in, even though gold declined sharply.

(Click to enlarge)

HUI-gold ratio – a bullish divergence from gold's short term trend.

A close-up of the HUI over the past 5 trading days compared with GLD shows that there were two days when it diverged positively, 2 neutral days and one negative divergence (on the Thursday preceding the payrolls report). On three of the five days there was strength late in the day.

(Click to enlarge)

The HUI over the past five trading days (5 minute candles) – it ended the week roughly where it started it, but gold lost considerable ground over the same span.

Sentiment and Positioning

Lastly, a brief update on sentiment and positioning data. On Friday, the daily sentiment index (DSI) published by ended at 12 points, a historically very low level (the record low is 5). The rule of thumb regarding the DSI is that readings below 15 increase the probability of a short term low being close, while readings above the 85 level are a heads-up to expect a short term peak.

The commitments of traders reports have finally caught up again (there was a brief suspension due to the shutdown). The recent changes haven't been overly remarkable, but small speculators have essentially returned to a neutral position, which is actually historically rare. However, we want to point to two divergences in the futures positioning data (net positions) between large and small speculators. A bearish divergence occurred near gold's highs, while in recent months a slightly smaller bullish divergence has become evident:

(Click to enlarge)

Commitments of traders in gold futures. Note the two divergences between big and small speculator net positions in 2011-2012 and in 2013.

Next comes a look at the premiums and discounts of closed end bullion funds, with GTU as the proxy (the data of other closed ends like CEF look very similar). Contrary to GLD, which is of the open-ended variety, they reflect sentiment not by variations in their bullion holdings, but by the extent to which their market value deviates from their net asset value.

When looking at NAV premiums and discounts in GTU, we can clearly differentiate two time periods – one when extreme bullishness pushed GTU's value way above its NAV and the more recent one when it has traded well below its NAV (this means incidentally that it currently offers a way to acquire bullion at less than the spot price).

(Click to enlarge)

Bullish and bearish sentiment on gold as reflected in GTU's premium/discount to NAV. Note also the slight divergence between price and sentiment expressed by the discount that has occurred this year.

The last positioning/sentiment datum we want to look at is the Rydex precious metals fund. At one point in 2011, the fund held approximately 35% of all Rydex sector assets (the biotech sector has reached a similarly exalted position recently). Since then this measure has fallen to less than 5%, which is a very low level historically. In fact, it is only about half the level seen at the 2008 panic low. So bearish sentiment is currently a lot more pronounced than in 2008 on a relative basis.

Total assets held by the Rydex precious metals fund have fallen back to levels last observed in Q4 of 2002, note though that the low in cumulative net cash flows into the fund (not shown on the chart below) was made in early 2012 and thus continues to slightly diverge from the other data points.

Rydex funds only capture a small slice of the market, but we have found over time that this microcosm nevertheless provides valuable information – it is an excellent mirror of the overall sentiment backdrop.

(Click to enlarge)

Rydex precious metals fund, total assets (lowest since Q4 2002) and assets as a percentage of all Rydex sector assets.


One must always keep in mind that for the trend to change, sentiment must actually change as well. Overly bearish sentiment as such is not meaningful – it only tells us that in the event of a trend change, there could be a lot of potential fuel for the new trend. Note also that bigger sentiment extremes and larger lags could be observed in many markets in recent years than previously (i.e., the lag time between sentiment extremes first showing up and actually having an effect).

However, in spite of the renewed weakness in the gold price over the past two weeks, we do have some preliminary evidence that a trend change may be close. As we have often pointed out, given that gold stocks have led the decline since 2011, they should also be expected to lead an eventual advance, something that will be expressed by a rising HUI-gold ratio (the ratio of any of the other gold stock indexes or ETFs to gold will serve as an indicator as well of course).

That said, the HUI's chart looks quite messy (with lots of overlapping waves, which is normally indicative of a corrective formation), and it has as of yet neither broken out the upside nor produced a new downside breakout.

Until then the market action only provides information in the form of proverbial breadcrumbs, but given the time of the year (trend changes in the sector often happen in the October-November time frame), it is nevertheless worth paying close attention.

Charts by: BarCharts, StockCharts, BigCharts, Sentimentrader, Decisionpoint, St. Louis Federal Reserve Research