Mark Hulbert on Gold Sentiment
Mark Hulbert publishes sentiment data on stocks and gold, which aggregate the recommendations of market newsletter writers. The level of gold sentiment index HGNSI indicates the percentage of one's portfolio the respective newsletter writers recommend their readers should allocate to either long or short positions in the market concerned. Keep in mind that these are averages, and that we are talking about people who are in the business of selling newsletters.
They are catering to customers who are often not necessarily interested in good recommendations, but want to see their own biases confirmed. Of course, there are many newsletters that have an excellent track record, published by advisors who are genuinely interested in offering good analysis. All we are saying is: a slight long bias is probably detectable on average.
In his latest column, Mark Hulbert argues that the current average recommendation of the gold timers – namely that people should be net short gold with 10% of their assets – is not yet indicating high enough pessimism. He writes:
“To be sure, the average gold timer is more bearish today than he was this past weekend. But he still is not as pessimistic about gold’s prospects as he was on the occasion of past tradeable bottoms.
My preferred gold sentiment indicator is the Hulbert Gold Newsletter Sentiment Index (HGNSI), which reflects the average recommended gold market exposure level among a subset of short-term gold market timers tracked by the Hulbert Financial Digest. This average currently stands at minus 10%, which means that the average gold timer is now recommending that his clients allocate 10% of their gold-oriented portfolios to going short.
That is a step in the right direction, from a contrarian point of view. When I wrote my column this past weekend, in contrast, the HGNSI stood at 16.7%. But, as you can see from the chart, the HGNSI is still not as low as it was at the bottom of gold’s other corrections over the last couple of years. At the end of last November, for example, prior to a rally that would add nearly $200 to the price of gold, this sentiment average stood at minus 36.7%.
And last June, prior to a gold rally that was even more powerful, the HGNSI got as low as minus 56.7%. To be sure, gold did mount a several-hundred-dollar rally in late 2012 off a sentiment base that wasn’t as bearish as these two instances from last year. Then the HGNSI never got lower than minus 15.7%, only modestly lower than where it stands today.”
Here is the chart that was posted along with the analysis:
The HGNSI over the course of gold's cyclical bear market.
What it All Means
We happen to recall an instance in March of 2003, when gold had just begun to rise again after having gone through a corrective phase. At the time, the HGNSI stood at + 65% (average recommendation: 65% long) and Mark Hulbert warned in no uncertain terms of how dangerously overconfident the newsletter writers were. Over the next 9 months, the gold sector rallied by more than 100% (gold itself didn't do as well, but it rallied as well of course). When the sector finally suffered a fairly noteworthy correction in late 2003/early 2004, the gold timers lost a lot of their previous enthusiasm. Mark Hulbert concluded it was a good time to buy, but it turned out it actually wasn't. A grinding bear market that lasted 18 months ensued (Gold went mostly sideways in a broad range, but gold stocks really took it on the chin).
Now pay attention to the sentences we have highlighted in the excerpt above. It is absolutely true that the two most recent lows were made with much more negative HGNSI readings – in fact, readings that haven't been seen since the late 1990s. It is just as true that in 2012, when gold still looked technically strong, a much lower negative reading was followed by a very big rally. So why is there such a big dispersion in what these readings signify? Are they really any better than a coin flip?
We would suggest that the meaning of these readings simply depends on whether gold is in a bull or a bear market. In bear markets, steeper negative readings are required to give a buy signal. In a bull market, even a high positive reading can mean it's not a bad time to be long.
What we should be interested in at this juncture is this: whether or not gold is in the process of transitioning back into a bull market. If it is, then the HGNSI readings required to indicate short term lows will move to progressively higher levels.
Just consider the last two lows of June 2013 and December 2013. They were at almost exactly the same price level, but the HGNSI was at very different levels. This actually represented a bullish divergence. It showed that the conviction of bears was receding as the same low was approached the second time. The same thing happens in bull markets, only the other way around. The peaks occur with slightly less bullish sentiment than is seen at the immediately preceding interim highs.
In other words, whether the current average "10% net short" recommendation is a contrarian buy signal or not does not depend, as Mr. Hulbert claims, on what the HGNSI readings at the most recent significant lows were. It depends solely on whether gold is about to enter a cyclical bull market or not. If the bear market is still intact, then we may eventually well see much lower readings (along with lower prices). We personally believe the transition phase has begun, but we could of course be proved wrong. However, if we are not wrong about this, then higher HGNSI readings will henceforth be recorded at interim lows.
Chart by: Mark Hulbert / Marketwatch