In the month since I wrote that “gold’s recent rally, though undeniably impressive, is running out of steam,” gold has dropped some $60 per ounce.
The reason that I suggested in mid-April that gold was turning down was that the gold market timers I monitor had become too enthusiastic; from a contrarian point of view that was bearish. I should have added another reason as well: Gold’s seasonal tendencies had become unfavorable.
I’m referring to bullion’s six-months on, six-months off seasonal pattern that is loosely analogous to the “Sell in May and Go Away” pattern in the stock market. Gold’s unfavorable six-month period, however, begins two months earlier than the stock market’s -- lasting from March through August, instead of May Day through Halloween.
The accompanying chart summarizes the historical data. It focuses on bullion’s average March-through April and September-through-February returns since 1973, when gold began trading freely in the U.S. The difference shown in this chart is statistically significant.
Note carefully that these results are based on multidecade averages, and not every year adheres to the overall pattern. This year, for example, gold exploded from mid-March until mid-April, notwithstanding the unfavorable calendar, helped along by a favorable sentiment environment. This serves as a cautionary reminder that seasonality is not the only factor influencing gold’s short-term term direction.
It may also illustrate that, when contrarian analysis and seasonal tendencies disagree, sentiment is the stronger factor to which we should pay attention. Regardless, as I mention below, sentiment now points in the same direction as seasonal patterns: A lower gold price over the next couple of months.
Why would gold exhibit this seasonal pattern?
As you can imagine, there are several theories. One was first put forth in an academic study in 2012 by Dirk Baur, a finance professor at the University of Western Australia. He argued that the effect was likely caused by a combination of three factors: “hedging demand by investors in anticipation of the ‘Halloween effect’ in the stock market, wedding season gold jewelry demand in India and negative investor sentiment due to shorter daylight time.”
Where does gold sentiment stand now?
When I wrote about gold market sentiment a month ago, the average recommended gold exposure level among several dozen gold timers I monitor was 52%, which was higher than 79% of all daily readings since 2000. Today, even though gold has declined over the intervening four weeks, this average exposure level is nearly as high -- 50%, which is higher than 76% of all daily readings 2000. (This average exposure level is what is measured by the Hulbert Gold Newsletter Sentiment Index, or HGNSI).
Not only is the current HGNSI level uncomfortably high in absolute terms, it’s also high in relative terms. That’s because the normal pattern is for the HGNSI to rise and fall more or less in lockstep with gold’s price. It therefore indicates a significant amount of stubbornly held bullishness that the market timers are sticking to their exposure levels despite gold’s weakness.
The bottom line: Gold is now battling not just one, but two, stiff headwinds: Unfavorable sentiment, and unfavorable seasonal tendencies.
While this doesn’t guarantee that gold’s price will continue declining over the near term, the historical odds suggest that gold will trade at lower prices at some point over the next month or two.