Friday, September 12, 2014

Gold Sentiment Plunges to Summer 2013 Levels

by Pater Tenebrarum

Newsletter Writers Turn Very Bearish

This is a little addendum to our recent gold update. Shortly after we had posted it, Mark Hulbert published an article at Marketwatch regarding the recent moves in the Gold Newsletter Writer Sentiment Index (HGNSI). Note here that this sentiment measure must be seen in the context of market action. As we have pointed out previously, there have been a number of very significant ‘misses’ of this indicator, especially in early 2003 and early 2004, when following its message would have been a grave mistake.

However, there is also the fact to consider that today’s gold-focused newsletter writers are probably not the same bunch that was active 10 years ago. Many of those who were active in 2003 had survived a 20 year long bear market, so their collective judgment was at times actually quite good.

To Mr. Hulbert’s credit, we must concede that his indicator has worked better in recent years, and his interpretations of it in the course of this year have largely been on the mark. That’s actually a good thing, as the indicator is currently showing an extreme in negative sentiment.

Here is the chart:

1-HGNSI

Hulbert Gold Newsletter Writer Sentiment Index – currently it stands at – 40.6% – click to enlarge.

What the percentage means is that at the moment, gold timers are recommending that their clients allocate 40.6% of their gold-related assets to shorting gold. Although the measure fell to even lower levels in the summer of 2013 and also in 1998/1999 if memory serves, this is still a historically fairly extreme level.

Mr. Hulbert notes:

“Consider the average recommended gold market exposure level among a subset of short-term gold market timers tracked by the Hulbert Financial Digest (as measured by the Hulbert Gold Newsletter Sentiment Index, or HGNSI). This average currently stands at minus 40.6%, which means that the typical gold timer is recommending that clients allocate nearly half their gold-oriented portfolios to going short the market.

That’s a particularly aggressive bet that gold will keep declining, and — at least according to contrarian analysis — these timers are unlikely to be right. As recently as last week, the HGNSI had not fallen below minus 21.9%. That was less than the lows to which this sentiment index fell last December (minus 36.7%) and in the summer of 2013 (minus 56.7%). And that, in turn, led me to conclude that contrarians were not yet ready to bet on even a short-term rally.

[…]

The usual qualifications apply, of course. Sentiment is not the only thing that moves the markets. And even when contrarian analysis is right, it doesn’t necessarily have pinpoint accuracy. But, because sentiment analysis has been on the correct side of this gold market in recent months, it’s definitely noteworthy that it’s now more optimistic.”

(emphasis added)

Now for the context: since gold is mired in another short term downtrend, it is not surprising that bearishness is high. However, gold also remains above the lows made in 2013, and gold stocks have built multiple divergences with the gold price, in the process beginning to hold up somewhat better on a relative basis. So the context is at least to some extent a mixed bag and not entirely negative.

Note by the way that yesterday’s decline in cap-weighted gold stock indexes like HUI and XAU was distorted by Anglogold (AU) announcing that it will hive off its non-South African assets into a new company. Normally such an announcement would be seen as bullish for the stock, the problem is though that in order to get approval from the SA Reserve Bank for the deal, AU had to find a way to extinguish a lot of debt. In order to to that, it plans to do a very large share issue. That was not what the market wanted to hear, and the stock was clobbered by more than 15.5%.

Given that AU is one of the world’s largest gold mining firms, it has a pretty large weighting in the indexes and has been responsible for about 50% of yesterday’s decline. Needless to say, the timing of the announcement is a real head-scratcher. We have no idea why AU’s management decided to make it just as the stock was hitting multi-month lows. However, we have often seen such announcements being made at especially inopportune times. Apparently the managers of many gold mining companies don’t consider that it can actually cost shareholders money when such announcements are mistimed (since greater dilution will now be required than would otherwise be the case).

2-AU

AU crashes on news of a capital raise that has become necessary to receive SARB approval for its planned demerger – click to enlarge.

Other Sentiment and Positioning Data

Below is a quick overview of other sentiment and positioning data. Commitments of traders in gold have seen hedgers covering and speculators selling, and as of last week Tuesday, the net speculative long position (small and large speculators combined) stood at 103 K contracts. This is not very much compared to recent years, and it is highly likely that it will have become noticeably smaller in the meantime.

Note though that there is no telling where this indicator will go. If a new bear market leg lies ahead (i.e., a multi-week downtrend), then this position could conceivably be reversed altogether (from net long to net short). So it currently looks mildly encouraging compared to the levels of recent years, but it there is always the caveat that it can worsen further in the event of a more protracted bear market. Luckily for bulls, it coincides with the above mentioned extreme in the HGNSI and several other sentiment extremes – so at the very least the downside should be somewhat limited.

3-Gold-CoT

Commitments of traders in gold – the hedger net position in blue, the small speculator net position in red. The latter group is by now probably either flat or slightly net short (we are guessing, as this snapshot is more than a week old. Unfortunately the release of CoT data is not very timely) – click to enlarge.

Lastly, a look at the discounts to NAV of closed end bullion funds, as well as Rydex precious metals assets and cumulative cash flows. Normally we consider it better when these discounts contract, but when they hit what are historically extreme levels, then they serve as contrary indicators as well. Again, similar to what Mr. Hulbert notes w.r.t. the HGNSI, these are by no means very accurate timing indicators. They can only tell us whether there is froth or fear in the market, and right now there seems to be quite a bit of fear.

We’ve adapted a chart that was recently posted by Erin Swenlin (formerly of Decisionpoint, now with Stockcharts) and have left the trendlines she drew on the gold price chart in. It should be noted that the triangle formed over the past year or so with its two lower highs does not look particularly encouraging, so once again, it is not a big surprise that bearish sentiment is quite pronounced at the moment. While the long term chart picture is currently discouraging, it wouldn’t take much to turn it completely around. On the other hand, if the 2013 lows were to break, that would obviously be very bad news.

The engagement of Rydex traders in the sector meanwhile is reminiscent of 1999/2000. Note though that it is not enough that sentiment is bad – that by itself will not turn the market around.

4-Gold sentiment data

Discounts to NAV of CEF and GTU (two closed end bullion funds) plus Rydex precious metals assets and cumulative cash net cash flows – click to enlarge.

Conclusion:

Bulls should actually hope that these sentiment trends bottom out and reverse, but that a certain amount of skepticism remains if and when they do. However, in the meantime the relatively extreme levels of the data suggest that the potential additional downside should actually be limited in the short term. It then remains to be seen how things evolve in the next bounce.

Addendum:

Note that this article was written prior to Thursday’s market open, so the charts reflect Wednesday’s close. Thursday’s market action does however not substantially alter the situation described above, except insofar that it can be inferred that several of the data points may have deteriorated slightly further.

See the original article >>

No comments:

Post a Comment

Follow Us