This article updates an earlier study, Anomalous Selling in COMEX Gold 1985 to November 2000, which I published in late 2000. It validated Dr. Harry Clawar's earlier findings of consistent, predictable COMEX selling against a rising overseas market component. Dr. Clawar has since published updated conclusions, which reveal that in four out of five trading days over a one-year period the COMEX closed lower than the London AM fix. My earlier study utilized the concept of preemptive selling, a condition that occurs when the dominant forces on the COMEX sell gold three times lower than the same day's decline on London market.

Dr. Clawar's counter trend COMEX selling tests the hypothesis that the COMEX gold market can be predicted to close below the next morning's London AM fix. Predictable market behavior such as this represents abnormal and biased operations. Figure #1 below shows the updated series from 1985 through May 2001. Note that for over a year counter trend selling has stayed above the mean, indicating a sustained selling effort by the dominant forces on the COMEX. This selling continues to be predictable and runs counter to the activity in the overseas component of the gold markets.

Figure #1a below presents an expanded resolution timeline. Note the drop in counter trend activity just prior to the Washington Agreement (WA) in September 1999. During that period gold exhibited horizontal pricing. Perhaps the implied intervention was no longer necessary. Then, late in September 1999, the rally in gold prices triggered by the WA called forth a far more aggressive intervention, resulting in a large spike in counter trend activity, which has persisted at high levels almost to the present.

By testing a relative condition between the COMEX and the London Bullion Market, preemptive selling is more sensitive at detecting trading patterns and their degree than a simple yes/no test. Figure #2 below is an updated and annotated chart from Anomalous Selling in COMEX Gold 1985 to November 2000. The series has been divided into two falling gold price epochs in order to highlight one of that study's several control mechanisms. The temporal resolution units are months. In a freely traded market, selling events in one price declining epoch should be similar to selling events in other price declining epochs. The original conclusion of clustered, anomalous selling behavior stands.

Figure #2 shows that extreme selling events have continued into the present year. The March 2001 event touches the three standard deviation ceiling. These three standard deviation selling events would be expected to happen only once in 83 years of COMEX trading (1,000 periods). In Epoch 1, there are no selling events which exceed two standard deviations. Epoch 2 contains seven such events. Event B, a four standard deviation selling event, would be expected to happen at random only once in 833 years of COMEX activity (10,000 periods). The implications of such an event are ominous. These extreme selling events started during the Clinton Administration. Perhaps not coincidentally, Event A occurred right around the time that the Federal Reserve, after a lapse of 64 years, assumed the two American seats on the board of the Bank of International Settlements. This aggressive selling effort nevertheless only affected the price of gold in a minimal way. Event B was necessary to break the resilience of the gold market, which thereafter fell into a major multi-year decline.

Figure #2a above gives an expanded time line from April 1998 to today, including annotations relating the over two standard deviation preemptive selling events to related market events. As yet there are no obvious market events to associate with the March 2001 spike. However, at least until very recently, counter trend selling has remained remarkably high around this March preemptive selling event. Perhaps more effort has been required to suppress the price of gold than in prior periods.

These updated figures show a continuation of the abnormal COMEX gold trading pattern that began in 1994, including another in a series of high standard deviation selling events. These extreme selling episodes did not happen prior to 1994. The updated figures also validate for the first part of 2001 Dr. Clawar's observation that the COMEX gold market follows a predictable pattern of selling against a rising overseas price component. Such a market trading pattern cannot be regarded as random. The recent slight reduction in counter trend selling below the mean should be watched. It could indicate the beginning of a significant change in trend.

G. Michael Bolser
Valrico, Florida
June 23, 2001

The author wishes to thank the Gold Anti-Trust Action Committee, Bill Murphy, Reg Howe and Don Lindley for their kind assistance.