At the end of the week before last the amount of physical gold held by the SPDR Gold Trust (GLD), the largest gold bullion ETF, fell to its lowest level since September-2008. What does this tell us?

In many TSI commentaries over the years and in a couple of posts at the TSI blog over the past year I’ve explained that changes in GLD’s bullion inventory are not directly related to the gold price. Neither a large rise nor a large fall in the gold price would necessarily require a change in GLD’s inventory, the reason being that as a fund that holds nothing other than gold bullion the net asset value (NAV) of a GLD share will naturally move by the same percentage amount as the gold price.

However, there is an indirect relationship between the gold price and GLD’s bullion inventory. At least, there has been such a relationship in the past. I am referring to the long-term correlation between the gold price and the GLD inventory that stems from changes in sentiment.

As traders in GLD shares become more optimistic about gold’s prospects they sometimes buy aggressively enough to push the market price of GLD above its NAV, which prompts an arbitrage trade by Authorised Participants (APs) involving the issuing of new GLD shares and the addition of physical gold to GLD’s inventory. And as traders in GLD shares become more pessimistic about gold’s prospects they sometimes sell aggressively enough to push the market price of GLD below its net asset value (NAV), prompting an arbitrage trade by APs involving the redemption of GLD shares and the removal of physical gold from GLD’s inventory.

That is, changes in GLD’s market price relative to its NAV create opportunities for arbitrage trades that adjust the supply of GLD shares and the amount of physical bullion held by the fund, thus ensuring that the market price never deviates far from the NAV. This modus operandi is common to all ETFs.

Since traders in GLD shares tend to become more optimistic in reaction to a rising price and less optimistic in reaction to a falling price, the most aggressive buying of GLD shares will tend to occur after the gold price has been trending higher for a while and the most aggressive selling of GLD shares will tend to occur after the gold price has been trending lower for a while. This explains why the following chart shows that the long-term correlation between the gold price and the GLD inventory is strongly positive and why the major downward trend in GLD’s inventory began well after the 2011 peak in the gold price.

The upshot is that the price trend is the cause and the GLD inventory is the effect.

In conclusion, here are three implications of the above:

1) Anyone who claims that the gold price has trended lower over the past few years due to the selling of gold from GLD’s inventory is getting cause and effect mixed up.

2) Anyone who claims that gold is being removed from GLD’s inventory to satisfy demand in Asia (or elsewhere) is either clueless about how ETFs work or is telling untruths to promote an agenda.

3) The early-August decline in GLD’s bullion inventory to a new multi-year low was consistent with the price action. It was evidence that GLD traders were getting increasingly bearish in reaction to lower prices. They loved it at $1600-$1900 and they hated it below $1100.

Almost every major price move in the financial markets looks predictable after it happens. This is called “hindsight bias”, which is defined thusly at Wikipedia:

“Hindsight bias, also known as the knew-it-all-along effect or creeping determinism, is the inclination, after an event has occurred, to see the event as having been predictable, despite there having been little or no objective basis for predicting it.”

Almost everyone suffers from hindsight bias to some degree. Of special relevance to me, many newsletter writers and other commentators on the financial markets are afflicted by it. After the event they are quick to explain how a big price move was totally predictable, but often forget to explain why they didn’t predict it ahead of time or perhaps even predicted the opposite of what happened.

It’s important to recognise hindsight bias when it occurs in the market-related opinions/analyses/ramblings you read and when it occurs in yourself. And with regard to the latter it is important not to beat yourself up or wallow in regret when the future turns out to be different from what you expected. Regardless of how predictable an outcome appears to have been with the benefit of hindsight, you can be sure that prior to it happening there were other realistic possibilities. It’s just that these other possibilities shrank to nothingness when they didn’t happen.

The best way to deal with the fact that nothing is certain without the benefit of hindsight is to simply accept the possibility that the future will not pan-out as you expect and position yourself accordingly. In particular, don’t bet so heavily on a specific outcome that you will be financially devastated if something different happens.

Here are ten basic gold-market realities that are either unknown or ignored by many gold ‘experts’.

1. Supply always equals demand, with the price changing to maintain the equivalence. In this respect the gold market is no different from any other market that clears, but it’s incredible how often comments like “demand is increasing relative to supply” appear in gold-related articles.

2. The supply of gold is the total aboveground gold inventory, which is currently somewhere in the 150K-200K tonne range. Mining’s contribution is to increase the aboveground inventory by about 1.5% each year. An implication is that there should never be a shortage of gold.

3. Although supply always equals demand, the price of gold moves due to sellers being more motivated than buyers or the other way around. Moreover, the change in price is the only reliable indicator of whether the demand side (the buyers) or the supply side (the sellers) have the greater urgency. An implication is that if the price declines over a period then we know, with 100% certainty, that during this period sellers were more motivated (had greater urgency) than buyers.

4. No useful information about past or future price movements can be obtained by counting-up the amount of gold bought/sold in different parts of the gold market or different parts of the world. An implication is that the supply/demand analyses put out by GFMS and used by the World Gold Council are generally useless in terms of explaining past price moves and assessing future price prospects.

5. Demand for physical gold cannot be satisfied by “paper gold”.

6. Prices in the physical and paper (futures) markets are linked by arbitrage trading. For example, if speculative selling in the futures market drives the futures price down relative to the physical (or cash) price by a sufficient amount then arbitrage traders will profit by selling the physical and buying the futures, and if speculative buying in the futures market drives the futures price up relative to the physical (or cash) price by a sufficient amount then arbitrage traders will profit by selling the futures and buying the physical.

7. The change in the spread between the cash price and the futures price is the only reliable indicator of whether a price change was driven by the cash/physical market or the paper/futures market.

8. In a world where US$ interest rates are much lower than usual, the difference between the price of gold in the cash market and the price of gold for future delivery will usually be much smaller than usual. In particular, when the T-Bill yield is close to zero, as is the case today, there will typically be very little difference between the spot price of gold and the price for delivery in a few months. An implication is that in the current financial environment the occasional drift by gold into “backwardation” (the futures price lower than the spot price) will not be anywhere near as significant as it would be under more normal interest-rate conditions.

9. Major trends in the US$ gold price are determined by changes in the general level of confidence in the Fed and the US economy. An implication is that major price trends have nothing to do with changes in jewellery demand, mine supply, scrap supply, central bank buying/selling, and the amounts of gold being imported by India and China.

10. The amount of gold in COMEX warehouses and the inventories of gold ETFs follow the major price trend, meaning that changes in these high-profile inventories are effects, not causes, of changes in the gold price.

Commentators on the financial markets often make statements like “it’s a bull market” and “the trend is up” as if these were indisputable facts, but such statements are always opinions.

A statement of fact could reasonably be phrased along the lines of “the market was in an upward trend between date X and date Y”, because if a sequence of rising lows and rising highs occurred between two dates then the trend was, by definition, up during that period. However, it is impossible to know the direction of a market’s current price trend with absolute certainty, let alone the direction of its future price trend. The reason is that even if a market has just made a new high/low there will be some chance that this will turn out to be the ultimate high/low.

For example, it’s a fact that gold was in a bear market in US$ terms from its peak in September of 2011 through to 24th July 2015 (when it hit a 4-year low of $1072), but it is a matter of opinion as to whether gold is now in a bear market. The bear market could obviously still be in progress, but there is also a possibility that it ended on 24th July 2015. At the time of writing, nobody knows for sure.

Some market participants and commentators will draw a line on a chart and then make a statement such as “I will consider the trend to be up (or down) unless the market proves otherwise by moving below (or above) my line”. Fine, but there’s a big difference between claiming to know the direction of the price trend and working under the assumption that the trend is in a particular direction unless/until proven otherwise by some predetermined event. The valley of shattered financial dreams is littered with traders who were determined to stay ‘long’ or ‘short’ because they thought they KNEW the direction of the price trend.

The impossibility of knowing whether a bull/bear market or an up/down trend is going to continue, or even whether the market is currently in bull or bear mode, makes risk management essential. Someone who knew the future would never have to bother with risk management; they could, instead, risk everything on a particular outcome because for them it wouldn’t be a risk at all. But ordinary mortals always face a degree of uncertainty when making investment decisions and, as a result, always need to face the reality that these decisions could prove to be wrong. Be wary, then, of advisors who claim that there is only one possible direction for the future price of an investment.

But while unwillingness to acknowledge the possibility of being wrong is a defect in the approach of some investors, other investors suffer from the opposite problem in that they have a hard time maintaining a bullish or bearish view unless that view is continually being validated by the price action. That is, they are incapable of remaining confident in any opinion that doesn’t happen to conform to the current opinion of the manic-depressive mob. As a result they routinely get ‘sucked in’ following large price rises and ‘blown out’ following large price declines, as opposed to taking advantage of the mob’s proclivity to be wrong.

Therefore, as investors the challenge we all face is to strike a balance between staying the course in rough weather and preparing ourselves for the possibility that there could be unseen rocks up ahead.