The World Gold Council Demand Trends Data Survey

Executive Summary

The World Gold Council estimates its survey encompasses 80% to 85% of all global demands for physical gold as estimated by Gold Fields Mineral Services. The Council surveys only jewelry, bar, and official coin demands. It does not survey demands for electronics, dental, decorative, industrial and medallions.

We pose the question, does the 80% to 85% global coverage of the World Gold Council (WGC) demand survey refer to just jewelry, bar and official coin demands or does it refer to all gold uses? The WGC assumes the latter. However, if that were so, the WGC 1999 survey would cover roughly 96% of global demands for jewelry, bar, and coin.

Gold Fields Mineral Services has provided estimates of jewelry consumption in 1998 for only seven of the many, many countries the WGC does not survey. Jewelry consumption in these seven countries alone, according to GFMS, is equal to more than 8% of their estimate for total jewelry fabrication demand worldwide. These GFMS estimates make it clear that the WGC survey encompasses much less than 96% of global jewelry demand.

If one compares the aggregate incomes of the countries the WGC surveys versus the aggregate incomes of those countries it does not survey, it would seem that the WGC survey may encompass less than 80% of the global markets it surveys---jewelry, bar, and official coin.

We "gross up" the WGC survey data from an assumed 83.2% coverage of the jewelry, bar and official coin markets to a full 100% coverage. We then add GFMS' estimates of demand for those gold markets or uses the WGC does not survey. We also make small additional adjustments for WGC under-estimates of Chinese demands, Gold Fields underestimates of dental and other demands, and positive inventory demands at the fabrication and wholesale levels. This procedure argues that GFMS global gold demand estimates were 500 to 600 tonnes too low in the 1994-1996 period. It indicates that GFMS global gold demand estimates were 800 to 900 tonnes too low in 1999.

The above analysis suggests that the WGC demand data, with possibly 80% or less coverage of the global markets it covers, might need to be grossed up by more than 25%---not just 20%---to arrive at an estimate for global gold demands for jewelry, bar, and official coin. This procedure points to total global demands and a gold market deficit that was 1000 tonnes or more higher in 1999 than the GFMS estimate.

If the GFMS demand data is correct, one may need to assume only a small level (200 tonnes or so) of undisclosed official selling in 1999. If we (and the WGC, though it does not realize this) are correct about global demand, there was absolutely huge undisclosed official selling in 1999.

The World Gold Council Demand Trends Data Survey
Does It Point To A Large Undisclosed Official Seller?

Introduction

World Gold Council demand surveys have indicated a large rebound in end user demand for gold worldwide as well as a sharp decline in liquidations of gold by parties in financial distress in emerging Asia (which are classified as scrap by Gold Fields Mineral Services). At the same time, the Washington Accord in September 1999 and the subsequent gold price explosion has changed risk perceptions by market participants. Trend following funds have covered short positions and have gone long. Mining companies have started to reduce their aggregate hedge position. Bullion bankers using value at risk and other trading disciplines have probably started to reduce net short positions. Improvement in supply/demand fundamentals and a move by market participants who were formerly short sellers of gold to cover short positions should have had a dramatic positive impact on the gold price. Instead, the gold price explosion of September/October was almost completely reversed in late 1999 and the recent rally has met with stiff resistance and has effectively been reversed. The failure of the gold price to rally despite improved supply/demand fundamentals and short covering by former short sellers points to a large undisclosed official seller.

The strength of the above argument depends upon one's assumptions about the underlying gold market supply/demand structure. If one assumes the GFMS framework, there has been virtually no recent growth in global gold demand and the gold market deficit was modest in Q4 1999. We have turned to the World Gold Council's estimates of gold demand for 1999 to argue that gold demand has rebounded strongly from the Asian crisis lows of 1998 and that there was a large gold market deficit in the fourth quarter of last year. Therefore, our case for an implied large undisclosed official seller in that quarter depends upon our interpretation of the World Gold Council's recent demand surveys. In this report we will defend our interpretation of the WGC demand estimates and our contention that there has been a very large undisclosed official seller since September of last year.

The Level of Demand and the Gold Market Deficit

The World Gold Council does a quarterly survey of gold demand in those countries where it has an active presence. It surveys demands for jewelry, bar and official coin. It no longer surveys dental demands. It has never surveyed gold demand for electronic, decorative and other industrial uses as well as medallions. The Council estimates that its surveys comprise 80% to 85% of all global demand for physical gold as estimated by Gold Fields Mineral Services.

In the Gold Book Annual 1998 we assumed that the World Gold Council's (WGC) coverage (roughly 80% at that time) applied to only those markets or uses that it surveyed. Applied to the current juncture, we would interpret this to mean that the WGC estimate of global gold demand of 3278 tonnes of gold in 1999 encompassed 80% to 85% of global demands for gold in jewelry, bar and official coin only and not for dental, electronics, decorative and industrial uses and for medallions. Employing this assumption, we have "grossed" up the most recent WGC demand estimates by roughly one fifth (from 83.2% to 100%) to obtain an estimate of global gold demand for just jewelry, bar and official coin.

In the Gold Book Annual we added Gold Fields Mineral Services' estimates of demands for gold in dental, electronics, decorative, industrial and medallions to arrive at a complete global gold demand estimate. Our analysis indicated that further adjustments had to be made for underestimates of Chinese demand by the WGC, for underestimates of dental, electronics, and other demands by Gold Fields, and for inventory demands by fabricators and wholesalers.

Of greatest importance is the question, "Have we been correct in assuming that the World Gold Council survey comprises 80% to 85% of global gold demand for only jewelry, bar and official coin uses?" Our position calls for a justification since the World Gold Council has argued that its demand survey encompasses 80% to 85% of all global markets for gold including dental, electronics, decorative, industrial, and medallions.

Why did Veneroso Associates assume in the Gold Book Annual that the World Gold Council survey encompassed (at the time) roughly 80% of only those markets the Council surveys and not all markets for gold use? We came to this interpretation by looking at the incomes of the countries that the WGC surveys and comparing their aggregate incomes to the aggregate incomes of those countries it does not survey. We compared these incomes along regional lines, as different regions exhibit different "intensities of gold use". Such comparisons argued conclusively that the WGC survey covered 80% of global gold demand only for those markets it surveyed---not all world markets for gold.

In April of 1999 Gold Fields Mineral Services provided some estimates of jewelry consumption in a handful of countries the World Gold Council does not cover in its survey work. These GFMS estimates of gold end use greatly support the interpretation of the Gold Book Annual regarding the extent of the coverage of WGC surveys.

Let us explain in more detail the above thesis. In January of this year, GFMS announced a preliminary estimate of global demand for gold in fabrication and bar hoarding of 3875 tonnes. We can exclude their estimate of Western investment demand of 203 tonnes, as that largely referred to short covering by funds and has no counterpart in the WGC survey. To make the GFMS data more comparable to the WGC data, we should subtract from their estimate of total fabrication and bar hoarding demands their estimate of 464 tonnes of dental, electronics, etc. demands which the WGC does not survey. This provides a GFMS estimate of global fabrication and bar hoarding demands for those uses for gold that the WGC does survey of 3411 tonnes in 1998. The World Gold Council's estimate of comparable demands "by the trade" is 3278. The WGC estimate is equal to 96% of the Gold Fields' global subtotal for such demands. The difference between the two estimates of 133 tonnes is equal to less than 4% of GFMS' estimate of such demands for the whole world and is equal or less than such demands in many individual countries.

It is often argued that estimates of world gold demand by the WGC seem high relative to estimates by GFMS because the WGC estimates demands by retailers (the "trade") whereas GFMS estimates demands by fabricators. To be sure, inventory change at the level of fabricators and wholesalers causes a disparity from time to time between these two measures of demand. However, such inventory demands should be on balance positive; therefore, they should tend to make the GFMS estimates higher, not lower, relative to those of the Council. (See chapters one and four of the Gold Book Annual 1998). For the time being, let us provisionally assume the inventory issue is not material. We can then pose the question, "Does the World Gold Council survey capture 96% of all global demands for gold for jewelry, bar and official coin or does it capture only 80% to 85% of such demands?" We can definitely conclude that the latter is the case. To explain our unqualified answer to this question, let us briefly consider the various regions of the world in which the WGC conducts its surveys.

Europe: The World Gold Council surveys Germany, France, Italy and the UK only. It does not survey Greece, Portugal, Spain, Austria, Switzerland, Belgium, Netherlands, Denmark, Sweden, Norway, Finland, Ireland and the smaller European principalities. Gold demand for jewelry, bar and official coin was estimated at 274 tonnes in 1999 for the four countries the WGC surveyed. Gold Fields estimates jewelry consumption alone for Greece, Portugal, and Spain was 107 tonnes in 1998. Therefore, jewelry consumption alone in just three countries accounts for almost the entire difference between the WGC and GFMS estimates of demand for those gold markets---jewelry, bar and official coin---which the WGC surveys. The combined income of all the other European countries which the WGC does not survey---Austria, Switzerland, Belgium, Netherlands, Denmark, Sweden, Norway, Finland and Ireland---equals or exceeds the incomes of every country in Europe except Germany. This would suggest that these countries taken together may consume almost 60 tonnes or more of gold in jewelry, bar, and official coin. These considerations suggest that demands for gold for jewelry, bar and official coin in all those countries in Europe that the WGC survey does not cover may have approached 170 tonnes in 1999. This one region alone accounts for more than the entire difference between the WGC and GFMS estimates of demand for gold for jewelry, bar, and official coin.

The Near East: The World Gold Council does not survey demand in Iran, Iraq, Israel, Lebanon, Jordan and Syria. Gold Fields Mineral Services estimates that jewelry demands in Iran and Syria alone were 72 tonnes in 1998. The odds are that total demands in these six countries for jewelry, bar, and official coin might have been well in excess of 100 tonnes in 1999.

Latin America: The World Gold Council makes estimates for gold demand in only two Latin American countries: Brazil and Mexico. These two countries had an average gold demand of 63 tonnes in 1999. All the remaining counties of Latin America---Argentina, Uruguay, Paraguay, Chile, Peru, Bolivia, Ecuador, Columbia, Venezuela, Guyana, Suriname and French Guyana---are not covered. Neither are the small countries of Central America or the island countries of the Caribbean. Gold Fields estimates that jewelry demand in Colombia alone was 20 tonnes in 1998. Columbia's national income is less than 5% of all of Latin America and less than 10% of that of Brazil and Mexico combined. The combined incomes of the countries that the WGC neglects are almost equal to the combined incomes of the two that are surveyed, which, according to the WGC, consume 126 tonnes of gold. Conceivably the demands for gold in jewelry, bar and official coin of these counties not surveyed could be on the order of 100 tonnes.

The Indian Subcontinent: WGC does not survey Sri Lanka, Bangladesh, Nepal, Bhutan, Mauritius, and Afghanistan. Their combined incomes approach that of Pakistan, which consumed 122 tonnes of gold in 1999.

Far East Asia: The WGC does not survey Myanmar, Laos, Cambodia, Philippines, North Korea, Mongolia and the small Pacific island nations. Based on income and ethnicity these countries combined may generate gold demand equal to one medium-sized country in the region, which might be somewhat less than 100 tonnes.

The Former East Bloc: This region is not surveyed by WGC. Gold Fields estimates that jewelry demand in the former Soviet Union was 43 tonnes in 1998. Income levels in East Europe and the former USSR would point to much higher levels of demand for jewelry, bar and official coin---possibly on the order of 100 tonnes overall.

The Anglo Saxon Countries: Canada, Australia and New Zealand are not surveyed by the WGC. GFMS' estimates Canadian jewelry demand at 24 tonnes in 1998. Comparison with US and UK income levels would point to total demands for jewelry, bar and official coin in these three countries of almost 50 tonnes.

Africa: This region is the most intriguing. Only Egypt is surveyed by the WGC. GFMS provides no consumption estimates for the countries of Africa other than Egypt. The World Bank estimates that the income of Egypt is only roughly one tenth of that of all of Africa. The cultures of many African countries suggest a generally high gold intensity of use. Egypt consumes 125 tonnes of jewelry, bar and official coin. Surely the rest of Africa's demands are not nine time those of Egypt, but they could be two or three times that of Egypt's. Two times Egypt's demand would be 250 tonnes.

Conclusion

From such country by country considerations it is clear that the countries surveyed by the World Gold Council do not comprise all but 4% of global demand. Gold Fields Mineral Services' estimates of 1998 jewelry demand alone in only seven countries---Greece, Portugal, Spain, the former USSR, Iran, Columbia, and Canada---total 268 tonnes, equal to more than 8% of GFMS' estimates for such demands globally. We have looked at eight different sets of countries the WGC survey does not encompass. The average likely level of demand of each of these sets may approximate the 133 tonne difference between the WGC demand total and the GFMS total for these same demands. Our assumption that the GFMS survey encompasses 83.2% of global demand for the gold uses that it surveys assumes that all the other countries of the world have such demands totaling 650 tonnes. The above country comparisons indicate that assumption is reasonable. If anything it would suggest a higher, not a lower, demand total for the combined countries not surveyed by the WGC.

In a recent Gold Watch we argued that, if the WGC demand surveys encompass 83.2% of total global demands for jewelry, bar, and official coin, the WGC demand estimate for the fourth quarter of 1999 of 806 tonnes would project total global demands of almost 1100 tonnes at a minimum and possibly 1200 tonnes. Mine and scrap supply in that quarter was slightly more than 800 tonnes, suggesting a Q4 1999 deficit of almost 300 to 400 tonnes. The above analysis suggests that this recent estimate of the fourth quarter deficit may have been too low.

We believe that prior short sellers of gold---funds, producers and bullion banks---reduced their shorts in the fourth quarter of last year by at least several hundred tonnes on a combined basis and probably more. A quarterly market deficit of 300-400 tonnes plus short covering of a comparable or greater magnitude had to be offset by an equal volume of selling. With private market participants covering short positions, selling on such a scale could only have emanated from the official sector. Yet only roughly 100 tonnes of official selling has been announced for the fourth quarter. Therefore, we argued that there must have been undisclosed official selling of many hundreds of tonnes and possibly even 1000 tonnes or more in that quarter. The analysis presented in this report suggests that, if anything, our earlier estimates of the fourth quarter 1999 gold market deficits and their implied undisclosed official selling may be too low.

Let us consider the implications of the WGC demand survey for global gold demand on an annual basis. If we gross up the WGC 1999 total assuming a market coverage of 83.2%, total demands for gold in jewelry, bar and official coin by the trade were 3934 in 1999. To this we must add 464 tonnes of gold absorbed by fabrication of electronics, dental, decorative, and industrial products plus medallions for a larger total demand figure of 4398 tonnes. We would add an additional 400 tonnes for a WGC underestimate of Chinese demand, an underestimate of dental, electronic and other demands by GFMS, and positive inventory demands at the fabrication and wholesaler level not captured by the WGC survey (See Chapter one, Gold Book Annual 1998). This generates a new global demand total of 4798 tonnes. If the suggested demands for the eight sets of countries not surveyed by the WGC are in the ballpark, total global demand for gold in 1999 was well in excess of 5000 tonnes. It follows that the GFMS estimate of global gold demand and the gold market deficit could be too low by 1200 tonnes or more.

Veneroso Associates: All portions of this work copyright 2000, all rights reserved.