Price Anomalies in the Gold Market
Random Walk, the name of a well-known theoretical market model of random price movements, definitely does not exist in the gold market. But then how does one explain gold's statistically proven price anomalies?
A look at Chart 1 is a good way to approach this issue. It shows gold's 24-hour intra-day price movement from 12:00 A.M. through 12:00 P.M., New York time on three consecutive days during the year 2004. The first day is coloured blue, the second day red and the third day black. You can clearly see that on all three occasions the price of gold fell on each day at around the same time, namely in the morning following the open on New York's COMEX exchange; coincidence or intent?
To find out let's examine a much higher number of days on an intra-day basis. Since using a chart to view intra-day data over a long period of time can be a little unmanageable and to eliminate any subjective conclusions, we will use an average. Chart 2 shows the average intra-day price movement of all days for which intra-day historical data exists, that is from August 1998 to September 2005. The y-axis shows the price, the x-axis the time. The average is reached by taking the mean results of every minute of the day over the approximately 1800 days and connecting them together. This so-called seasonal intra-day chart therefore shows at a glance the average intra-day price movement of gold at any point of the day based on the last seven years. Again we see that prices tend to fall during trading hours on the New York futures market. Additionally prices tend to form local lows between the London AM and PM price fixing as well as at the close of the American markets. A strong drop is noticeable around the PM fixing whereby most of the price drop has already taken place after the London fixings i.e. after the cash market loses influence in favour of the futures market.
Now let's expand the observed time period. Since historical intra-day data is available only until August 1998 we will use the price of the AM London fixing and the closing price on the New York markets as a reference to delve further into the past. This data is available as far back as 1991. So now we are dealing with the price of gold at two definable points of the day. We can see in Chart 2 that the technically decisive intra-day price moves, namely downwards around the New York open and especially at the London afternoon fixing, and upwards during the remaining part of the day i.e. during Asian and European trading hours, coincide only approximately with these two price points. For the more recent past a reference price a few minutes after the London afternoon fixing would be better suited to our purposes than the closing price, but since these are not available before 1998 the closing price must be used. For the period prior to 2001 (since gold started rising) this approximation is close enough. Chart 3 shows the intra-day course schematically using data from 1991.
This shows that the difference between the New York close and the London AM fixing is an approximation for intra-day price movement during New York trading hours (>NY=), mathematically expressed as NY Close - AM Fix. Accordingly, price development during the remainder of the trading day (>Overseas=) can be expressed as the London morning fixing minus the previous day's close in New York (AM Fix - Previous Day's Close NY). Subtracting one term from the other results in the relative price movement over this daily time period: ((NY Close - AM Fix) - (AM Fix - Day's Close NY)). Chart 4 shows this relative price movement smoothed with a 125 day average. If the line is below the zero level it means prices in New York were down compared to the remaining trading hours, the opposite case if the line is above zero. The lower half of the chart shows Gold's actual price movement.
Clearly visible here is that the gold price tends much more negatively during New York trading hours as during the remaining hours of the day. The beginning of these price anomalies (irregularities) can be pinpointed to the day. They started on August 5, 1993. This is of significance because it allows the accompanying circumstances leading to the irregularities to be more clearly defined. September 1999 is also notable as gold jumped following the Washington Agreement (WAG). More recently the line is above the zero level. Since then prices tend to rise toward the New York close. However in the current year (2005) the anomaly of a sharp drop toward the afternoon fixing continues. The basic premise of a typical sharp drop during the Comex session hasn't changed. Only the test procedure, chosen because of data availability is too inexact.
From the standpoint of ten years with several thousand intra-day movements it could be expected that the line would be both above and below the zero level to the same degree, which would mean the gold price sometimes rises in New York and sometimes during the remainder of the trading day and vice versa. However that isn't so. Let's examine all possible pairs of movement. On days with evenly dispersed price movement the line falls within the expected range. There are about an equal number of days on which prices rise both in New York and during the remaining hours (717) as well as the reverse case (falling 699 times in both parts of the day). If the opposite price movement is observed results are completely different: There were 948 days on which the price fell in New York while rising during the remaining part of the day. In the reverse case it was only 543 times (days when prices were unchanged were not considered). Before May 2001 the ratio was substantially more than 2 to 1. This massive price irregularity is far from what one would call a Random Walk scenario. Considering the high number of days sampled in this test and the number of times the anomaly is measured it can be viewed as statistically very significant. What could be the reason for this price behaviour?
For a long time there have been rumours in the market of official intervention against gold's price. An organisation was even set up (Gata) to collect evidence on the suspected interventions. Data indicates at most isolated instances, but put together they deliver a rather sound view. At the same time the observed intra-day price movements present an undeniable price anomaly with no satisfactory market explanation to support it. These irregularities however can be explained through just this type of intervention, because interventions are not random but intentional. And it stands to reason that the futures market with its power of leverage would be the preferred instrument. Thus, the price irregularities as statistical proof reinforce the total picture based on individual observations of systematic price suppression.
Watergate scandal reporter, Bob Woodward, said in his book Maestro about Alan Greenspan, that during the stock-market crash of 1987 the futures market was also used for price intervention purposes. This action was obviously the model for holding down the gold price. This also applies to secret actions which apparently tend to promise better success. A third parallel can be observed in that private banks execute the work in detail while official institutions make public guarantees and coordinate actions.
The always needed physical gold is provided by the central banks not only through sales of the metal but also through the loaning of it. In this instance, the central bank loans gold to a private bank which then sells it in the market. These agreements can be continually prolonged as the central banks can act as if they don't need the metal returned. The extent of the loans is never made public and the loaned gold is put on the central bank's books as if it were available. The public never learns how much gold is still being stored in the vaults and how much physical gold as been released to the market through loans. This type of creative book keeping can also lead to errors such as physical supplies being double-booked (for instance when the loaned gold ends up on the books of another central bank).
But what warrants the high degree of concern by the central banks for a metal that is hardly still used as money? The answer can be found in the notes of a FED meeting in July 1993, the meeting just prior to the beginning of gold's irregular price movement. Then FED governor Angell expressed concern about the rising price of gold saying it could create inflation fears. Additionally he spoke of the connection between long-term interest rates and the gold price as well as gold's relation to the U.S. Dollar. In summary you could conclude the physical currency, gold, was in competition with paper money. But why does keeping the price of gold down help? For one it helps keep interest rates low: when stock markets are weak, but gold is made unattractive by holding its price down money tends to flow into the stock market. Keeping the price of gold down also helps keep expectations of inflation in check which also leads to a lower rate of inflation if savers i.e. companies and private households react accordingly. A stronger dollar is another desired effect as gold is the biggest competitor to the dollar as world reserve currency. The August 1993 FED meeting was also where America's strong-dollar policy was first officially introduced.
One of the unintentional effects of the gold price interventions of course was that they also contributed to the asset market bubble as excess money and low interest rates caused asset price inflation. Rounding out the accompanying circumstances was also the fact that FED Chief Angell also noticed that central bank vaults held several times the yearly production of gold - a prerequisite for pushing the gold price down through intervention. A similar situation is not present in other world markets as they do not have access to material on loan. But worth noting is that Angell himself explicitly formulated the wish for a suppressed gold price as well as the possibility of its achievement, saying "We can hold the price of gold very easily."
It was understandable, if short-sighted, considerations that led to the systematic suppression of gold prices. This cannot be done forever if physical supplies are limited. There are some indications that it is becoming increasingly difficult to hold gold's price down. Gold price suppression has been the most influential factor of price development on the gold market over the past 12 years. But on the other hand, it could also be nearing its end. The gold price is lower then it would be without intervention and therefore has more upward potential. Additionally market balance has been skewed through a lack of new mine openings. Aside from that, gold price suppression could in the end prove to be a boomerang in the event of a financial crisis by weakening trust in the central banks.