Most analysts today view gold as a commodity. As such, they try to make sense of the gold price the way one would approach, say copper, or nickel, by looking at changes in the physical supply and demand for the metal. This failed during the 90s, when analysts called for a higher gold price, and the gold price declined. Then, since the turn of the century, they have been calling for lower gold prices, yet the price of gold is rising.

Have you heard that the gold price just has to go up, because total worldwide mine-production is only 2,400 tonnes per annum (average from 1990 to 2002) while fabrication demand is more than 3,300 tonnes? Surely this, almost 1,000 tonnes per annum deficit, has to drive the gold price higher.

Of course, what is left out of the equation are things like official sector sales, scrap sales, gold loans, forward sales, hedging, etc. Gold Fields Mineral Services (GFMS) adds up the total supply of physical gold and compares that to the total demand for physical gold, including all the items listed above, and some others. The difference between total supply and total demand is deemed to be net investment demand, which can be either negative (disinvestment) or positive (investment).

Whether GFMS' derived net investment demand is absolutely accurate is not all that important; we can assume that the overall trend of investment flows is correct because it is influenced by changes in the supply and demand for gold from categories that we can measure.

So if the assumption that gold can be analyzed as a commodity is correct, there should be a correlation between net investment demand and the gold price. Does net investment demand have any effect on the gold price? Could the decline in the gold price between 1990 and 2000 have been a result of a decline in net investment demand?

A quick glance at the volatile nature of net investment demand (Figure 1) could lead to the conclusion that there is no correlation between it and the gold price. I also plotted the net investment demand trend line from 1990 to 2000, just in case the correlation was hidden in the volatility. While it's too early to know for sure, it seems as if there has been an uptrend in net investment demand since 2000, which is why the trend line truncates in that year.

Looking at the trend line, it appears that while net investment demand was trending down, the gold price was declining. This corroborates the view that gold could, indeed, be analyzed as a commodity.

Or, could we say that as long as the gold price was declining, net investment demand was trending down? These are not the same. Is investment demand driven by the gold price, or is the gold price a function of investment demand?

Looking at the chart above may lead the casual observer to conclude that net investment demand is in fact the cause of changes in the gold price. After all, the gold price bottomed, and started its bull market rally only after the downtrend in net investment demand was reversed in 2000. Clearly net investment demand picked up prior to the inception of the gold price rally. But it didn't.

Net investment demand is in the order of a few hundred tonnes of gold per annum whereas annual physical gold trading is in the order of thousands of tonnes. Last year, for example, gold trading on the London Bullion Market Association (LBMA) alone was in excess of 6,600 tonnes, while the net investment demand for gold was only 137 tonnes, or 2% of the trading volume. From 1997 to 2002, worldwide net investment demand represented on average only 1.96% of the gold traded on the LBMA. Considering that the LBMA is not the only place in the world where physical gold changes hands, we can conclude that net investment demand represents less than 2%, and probably less than 1%, of physical gold trading.

Even though price is set at the margin, I don't believe the magnitude of net investment demand is sufficient to influence the actual gold price in any significant manner.

Furthermore, the gold market is not confined to the United States, even though the gold price is ubiquitously quoted in US dollars. The price of gold is determined by many factors not indigenous to the United States, such as economic prosperity around the world, concern for currency devaluations including, but not limited to, the US dollar, international political instability, social insecurity, war etc.

Given that net investment demand represents the aggregate of global economic trends, it makes no sense to try and correlate it to the gold price in US dollars alone, without considering the price of gold in the rest of the world, which by definition must incorporate all currency exchange rates.

An analyst in Lusaka, for example, may find value in analyzing the gold price strictly in terms of Zambian Kwacha, but in reality his analysis will reflect mainly the Kwacha exchange rate, with a minor contribution from the actual gold market. The same applies to an analyst in New York, working strictly from a US dollar perspective.

To solve this problem I created a global gold price using a GDP-weighted index of 35 currencies, representing in excess of 75% of the world's economy. Since 1998 I have been looking at the gold price from a whole-world perspective, and it's been very interesting. For one, it allowed me to be very, very bullish on gold as early as 1998.

It's obvious (Figure 2) that there is absolutely no correlation between net investment demand and the gold price. It is also quite obvious that the increase in the gold price preceded the upturn in net investment demand by several years. Net investment demand is therefore not a good leading indicator for the gold price. If anything, it's a reaction to the gold price.