You are here

Facts, Evidence and Logical Inference by Frank A. J. Veneroso

Section: Essays

Facts, Evidence and Logical Inference
A Presentation On Gold Supply/Demand, Gold Derivatives and Gold Loans

By Frank A. J. Veneroso

INTRODUCTION Currently head of Veneroso Associates, formerly partner of the hedge fund Omega Advisors where he was responsible for global investment policy formulation. Through his own firm, Mr. Veneroso has been an investment and economic adviser in investment strategy to institutions and governments around the world in the areas of money and banking, financial instability and crisis, privatization, and development and globalization of securities markets. His clients have included the World Bank, the International Finance Corporation, The Organization of American States.

He has advised the Governments of Bahrain, Brazil, Chile, Ecuador, Korea, Mexico, Peru, Portugal, Thailand, Venezuela and the United Arab Emeritus.

Frank is a graduate from Harvard and has authored many articles on the subjects of international finance.

Well, James Turk gave you interesting detective work that shows the possible hand of Government intervening in the gold market---sexy stuff. I am going to talk about the dry stuff---which is statistics on gold supply and demand. I sort of apologize for this but I guess it is an important part of the whole case. I am going to try and focus just on facts and evidence and simple logical inferences from them---rather than allegations, footprints, paper trails and the like. You might want to know why I have come all the way down here to participate in this conference. I find it extremely annoying that there is a hell of a lot of obvious evidence out there that something is happening in the gold market---that there are very large supplies coming into the market---larger than the consensus would claim---and no one is willing to discuss it.

I have had interviews with the press. After the interviews, its has always turned out that the articles were killed. I have requested debates with Goldfields Mineral Services and they have refused to show up. I have asked the World Gold Council to fund pertinent research studies and they have not responded. I never get a response that counters the evidence that I can bring forward. I simply get extreme silence. Only GATA has looked at this evidence and taken it to the public, and so, as a result, I feel it is incumbent on me to present it once again in their forum because I think that it represents evidence of very large undisclosed official supplies in the market that is systematically ignored. If there are any producers here who have influence on organizations like the Gold Council---if you find this persuasive---you should go to them and say, "Hey, listen, this guy has real evidence. He may not be right but it poses serious questions. It should be addressed. Why isn't it being addressed?"

Now, when we went to Congress in May of last year---exactly a year ago today---the subject was a gold derivative banking crisis. I will be talking about gold banking and gold derivatives, but I think you first have to understand what they are. We just had an article in the FT about the gold conspiracy being debunked or disproved by Antony Neuburger - a consultant to the World Gold Council. I have spoken to Mr. Neuburger. I've told him what I will tell you. He systematically has ignored it. Implicit in the work of a lot of these people is a flawed understanding of gold lending and gold derivatives. So I think it is best if I explain to you the process of gold lending and gold banking and the origin of gold derivatives first so that what I then discuss later makes more sense.

It is a simple, simple idea. Central banks have bars of gold in a vault. It's their own vault, it's the Bank of England's vault, it's the New York Fed's vault. It costs them money for insurance - it costs them money for storage--- and gold doesn't pay any interest. They earn interest on their bills of sovereigns, like US Treasury Bills. They would like to have a return as well on their barren gold, so they take the bars out of the vault and they lend them to a bullion bank. Now the bullion bank owes the central bank gold---physical gold---and pays interest on this loan of perhaps 1%. What do these bullion bankers do with this gold? Does it sit in their vault and cost them storage and insurance? No, they are not going to pay 1% for a gold loan from a central bank and then have a negative spread of 2% because of additional insurance and storage costs on their physical gold.

They are intermediaries---they are in the business of making money on financial intermediation. So they take the physical gold and they sell it spot and get cash for it. They put that cash on deposit or purchase a Treasury Bill. Now they have a financial asset---not a real asset---on the asset side of their balance sheet that pays them interest---6% against that 1% interest cost on the gold loan to the central bank. What happened to that physical gold? Well, that physical gold was Central Bank bars and it went to a refinery and that refinery refined it, upgraded it, and poured it into different kinds of bars like kilo bars that go to jewelry factories who then make jewelry out of it. That jewelry gets sold to individuals. That's where those physical bars have wound up---adorning the people of the world.

Now, this bullion banker is net short gold when he conducts this operation. Remember he borrowed gold and now he has a dollar financial asset. He is making a 5% return on the spread, but he now has a gold price risk. As a banker he is not normally in the business of putting on speculative positions like this. He is an intermediary, so what does he do? For the most part what he does is he hedges his gold price risk. He goes long the forward market to offset his physical short. Now if he goes long in the forward market someone else must go short, because every such contract in the forward market has two sides---a long and a short. In doing this he allows private market participants to go short the forward market. Who are those private participants who go short the forward market? They are producers hedging future production, they are jewelers who are hedging their inventory, and they are speculators who want to go short the gold market because they believe the price will go down and they earn a forward premium or 'contango' which happens to be, in this case, roughly equal (though not quite) to the difference between the rate of interest on the dollar asset held by the bullion bank and the rate of interest paid on the gold loans by the bullion bank.

So, basically, in doing this operation the bullion banker has a hedged position on the gold price and he takes a small margin---like a half of one percent---from this intermediation. In doing so, he allows private market participants to go short gold. That's why we elide the two phrases---going short in the gold market and gold borrowing. The ultimate borrowers in the gold lending operation are these shorts in the gold futures and forward markets. Now that we understand what this mechanism is all about, I am going to talk about the commodity case for gold.

The objective here is to try and put together a lot of information that we have on commodity supply and demand. The conclusion is that when we put all this information together we find out that the consensus estimates of supply and demand that we are all familiar with from Goldfields Mineral Services understate demand and understate supply and it appears that the understated supply is largely gold lending that each year substantially exceeds the consensus estimates of gold lending. Now it's a real simple supply/demand balance we are going to talk about. We are going to talk about it in terms of tonnes of gold.

Here is the consensus data and what we see in this consensus data is that demand ten years ago was about 3000 tonnes of gold and more recently it has been about 4000 tonnes of gold. Demand must be equal to supply in any given time period so therefore supply is the same. Where does the supply come from? Well, most of it comes from mine production, some of it comes from scrap, and some of it comes from the central banks.

The central banks have taken gold bars out of their vault and given them to the bullion dealers. The bullion dealers have been selling them spot and they have then been made into jewelry.

In any given year about half of the movement of these gold bars are gold sales and about half are gold loans. This is the consensus understanding of the gold market in terms of supply and demand flows.

Okay, lets go forward...

This chart simply shows the percentage of the supply that is coming from central bank selling and lending. This supply depresses the gold price---if it weren't there the price would be somewhat higher

Now we have a different simple supply/demand framework. In our framework demand is consistently higher and supply is consistently higher and the additional supply is all coming from central bank sales and central bank loans. Now we have a conservative set of numbers based on a lot of data and inferences as well as a more aggressive one. Here you will see, in our conservative assessment, a lot more gold is coming out of the central banks, thereby depressing the gold price---more than the consensus supply/demand statistics would suggest.

Now I will also give you a more aggressive alternative.

I should also say that these are annual flows of central bank bars.

Over time, such annual flows accumulate. Gold lending began about a decade and a half ago. Year after year more and more gold has been lent, which means more and more gold has left the central bank vaults, so when the IMF tells you there are 33,000 tons of gold in them thar' vaults---well that really isn't so. Part of it has been lent. What you see here is that the consensus statistics indicate that roughly 3000 to 5000 tonnes of that 33,000 tonnes has been lent. Now that's a pretty small percentage and if you thought that was the case you would believe that the central banks could keep dumping gold into the market for quite a long time.

Now we have a conservative set of gold lending numbers and we have a more aggressive set of such numbers. Our range of estimates implies that somewhere between 10,000 and 16,000 tonnes of the official sector gold position has left those vaults by way of the lending process.

And you can see that these are very substantial percentages. In the more aggressive case, almost half of the central bank's gold has left those vaults and it is now flowing out at a pretty rapid rate. So not only have we depleted the central bank physical gold stock more than most people have thought, but what metal is left in the vaults is diminishing at a more rapid rate than people think.

Now why do we think this? I am going to give you six pieces of evidence and lines of reasoning---more or less independent of one another---that lead us to this conclusion.

I started out on this crazy voyage with a statement that was made by a man from the Bank of England---Mr. Terry Smeeton---who was in charge of the gold operations of the Bank of England. On something like November 21st or 22nd of 1995 at the 5th Annual Banking Conference in the city of London, he addressed the issue of gold lending. He gave some statistics. He basically said that gold lending had roughly doubled over the last year and a half. Precisely, what he said was that gold loans more than doubled and gold swaps increased by more than 50%.

But he didn't give us any absolute numbers. However, he made a similar speech in Australia in March of 1994 and I went back and I checked what he said then. There he said that gold loans were 1500 tonnes based on a recent survey the Bank of England did, but he didn't make any reference to swaps. I called him up and asked him what the Bank thought the total swaps were in early 1994---a year and half earlier. He said to me he didn't remember exactly but he thought they were about 400 tonnes. What that meant is that the Bank surveys indicated that roughly 1900 tonnes of official gold had been lent around the beginning of 1994 and 18 months later---around the middle of 1995---the number had roughly doubled to 3700 tonnes, perhaps more. Now that was interesting because 3700 tonnes was a substantially larger figure than the consensus estimate of all those lendings, which at the time was about 2200 tonnes.

I thought this was intriguing and I did some analysis. I went to Mr. Smeeton from time to time under fairly casual circumstances and I asked him to give me an interpretation of his data. What he told me was that the Bank of England had done a survey of the fourteen principal market makers in the City of London and they had reported this data. I said to him, "Well, did that include the Swiss banks for example?" and he said, "No, absolutely not---only the fourteen principal market makers in the City of London." So I went to these fourteen bankers and I asked them "When the Bank of England came and asked you about this what did you tell them?"

I found out something very interesting. Some of these characters said to me, "I didn't report anything. I don't keep a big book in London---most of my book is outside of London. He doesn't know my loan position." Some of them said, "Oh, we complied. We gave them our global book, not only what was in London, but everywhere." From these conversations, I came up with the impression that, for these fourteen bullion bankers, this number was a partial total---it wasn't a complete total because many had not disclosed their books outside of London.

Then a bullion banker friend of mine said, "Frank, that's only the half of it---those fourteen (14) principal market makers in the City of London." He said, "Take down this list of all the guys who take gold deposits from central banks." And I took down the list and there were thirty-seven (37) of them. So I took the other twenty-three (23) who were not surveyed, I put them in a list, and I put that list next to the list of the 14 that were surveyed. I went to ten bullion bankers and I asked, "Of these two groups, which are the most important?" Nine out of ten of those bullion bankers said to me that the 23 who were not surveyed were as important or more important in terms of their aggregate position as the 14 that were surveyed. I sat down and I said to myself, this is very interesting.

The Bank of England survey showed that only 14 bullion bankers had lent 3700 tonnes and that total was partial. If I grossed up the 14 to their total and I threw in an equivalent total for the 23 that were not surveyed, I came up with some gigantic numbers. Perhaps 9000 or 10,000 tonnes of gold had been lent, based on this Bank of England survey. This was all by inference mind you, but none the less, it was very striking. The total estimated borrowed gold in the official Gold Fields Mineral Services statistics was only on the order of about 2200 tonnes at the time. There was a giant discrepancy. It was so giant that I decided to be conservative, and, for no good reason, I chopped the 9000 tonnes down to 6000 tonnes because that 6000 tonne figure was already so far removed from the official numbers. In any case, this Bank of England survey implied big, big errors in the consensus supply/demand balances and a hell of a lot more gold lending than anyone thought. This is written up in chapter 2 of our Gold Book and I will put it up on GATA's website for you so that you can see more precisely how this all started.

Now look, gold lending began in earnest in the early 1980s. By 1995 it was a process that had been going on for more than ten years. Now, what if there was 6000 tonnes of gold loans---not 2000 tonnes of gold loans as implied by the consensus supply/demand statistics. That means that there had been 4000 tonnes more lending, most of it over the last ten-year period. Gold lending was a small activity during the 1980s. It was a much bigger activity during the 1990s, so obviously it was a business that was occurring on an increasing scale. If the discrepancy was 4000 tonnes over ten to fifteen years, 300 to 400 tonnes a year---well, then it was probably 200 tonnes a year in the 1980s and it was probably nearer 600 tonnes a year by 1995. That meant supply and demand was underestimated by something like 600 tonnes a year. Now, the Bank of England survey results suggested a yet higher rate of lending over the past two years alone, but we thought it was more conservative to spread the final total out more smoothly over more years. In any case, I looked at this data and I said to myself, "How can this be? Is there any corroborating evidence? Low and behold we found corroborating evidence---so now lets go further.

Now let us look to the World Gold Council. They do a demand survey. Their survey uses a lot more people than does Gold Fields Mineral Services--- all of the people they work with in the countries where they have a presence in the promotion of gold consumption. In the case of the United States that means 28,000 point of sale players.

The World Gold Council survey is only partial. It only surveys three uses---jewelry, official coin, and investment bar. It only encompasses the countries where the Council has a substantial presence. The Council has estimated that its geographical coverage of jewelry, official coin, and investment bar demands encompasses roughly 80% to 85% of those demands. We have done studies to verify this, and those studies suggest its coverage is perhaps 80% of the total, possibly less. From the Gold Fields Mineral Services data we have estimates on the gold demands the Council does not survey---like electronics demands.

We have played around with this data, trying to project a total for global gold demand from the Council's partial statistical coverage. We have produced one such exercise in our Gold Book. We have provided another in our contribution to GATA's gold derivative banking crisis document presented to the US Congress last May. We have yet another in a draft research proposal to the Gold Council. The upshot of our research is that projecting the Council's gold demand survey data to a global total suggests that total global gold demand exceeded Gold Field's estimates by perhaps 500 or 600 tonnes a year in the mid 1990s and probably by much more in recent years.

We also compared the growth rate in demand for gold in the 1990s in the World Gold Council and Gold Fields Mineral Services data series. The growth rate in the World Gold Council data series is persistently higher that it is in the Gold Fields series. This suggests that the Gold Fields data may have been falling ever further behind a real and more positive demand trend that is reflected in the Gold Council data. The upshot of all this is simple and quite satisfying: we found a completely independent and alternative series on gold demand that pointed to a short fall in Gold Fields estimates of global gold demand. That demand short fall corresponded in magnitude to the short fall in Gold Fields estimates of official supplies suggested by our extrapolations from Bank of England survey work.

The World Gold Council data, then, was quite corroborative, quite significant.

Since then we have done some more analysis and we have found four more bodies of evidence that point to the same conclusion: consensus (Gold Fields) estimates of global gold demand and supply are significantly understated.

The first of these four is data on the gold derivatives of commercial banks reported to the BIS in conjunction with BIS capital adequacy requirements. In the US there is a monthly report by the US Controller of the Currency on the balance sheet of the US banking system which includes US commercial bank derivatives. I had seen the gold derivatives data included in this report over the years, but I did not know what to make of it. I thought it described a lot of derivatives that were duplicative. I didn't think it was significant.

But then something strange happened in 1999---there was a big shift in the locus of these derivatives. About that time there was also an explosion in the gold price, after the Washington Accord. People then began to focus on this derivative data. Reg Howe did a lot of research on these derivatives and he found similar derivative data on the German banks and on the Swiss banks. We took this data and we tried to interpret it. Now the World Gold Council, Jessica Cross, Antony Neuburger and the bullion banks--- they've all said this is meaningless data---it's transaction data---it's data not about stocks or positions but about transaction flows and you can't draw any conclusions from it about the amount of gold lending outstanding. In fact, Reg Howe has shown that, in the BIS' own accounts of what they are compiling and reporting, they make it very clear that this is in fact position data, not transaction data.

Now we believe we have figured out what this data means. Let us go back to the process of bullion banking. The central bank deposits its gold with the bullion banker. The bullion banker sells the physical gold into the spot market, and then goes long forward to hedge his physical short. That forward is a derivative. If all that was done by a bullion banker was to go long forward against the gold deposited with him by the central banks, there would be a one to one correlation between his gold deposits and his gold derivatives. But in fact that is not what happens. What happens is that sometimes he hedges with options rather than forwards. In the latter case the hedge is the delta on those options. I won't go into the details---basically these options will have a nominal or face value that is several times the value of the gold deposit operation being hedged. Now if we mixed together some option hedges and some forward hedges, a bullion bank's gold derivatives would be a small whole number multiple of the deposits (or at least those deposits that were hedged). I want to say that what James Turk was talking about earlier today---that some bullion bankers will borrow gold, sell it spot, and take positions in currencies like the dollar without a hedge---these operations, these gold carry trades, carry no associated derivative. Gold derivatives are only spawned if bankers choose to hedge their physical gold shorts. But based on the way I described the operation you will see that it is likely that the derivatives would be perhaps two or three times a bullion banker's deposits, assuming that they hedge most of the physical gold shorts generated out of their gold deposits. Now it is likely that there is some double counting in this data due to duplicative positions that would make this ratio somewhat higher. On the other hand, some bullion bankers presumably have unhedged physical shorts and, in this case, there would be deposits without an associated derivative that would tend to make this ratio somewhat lower.

Now we have a survey of our own of gold deposit taking from central banks, not deposit taking from other bullion banks, but just from central banks. The survey encompasses only a partial sub-set of the gold dealers. Of importance is that some of those gold dealers are also among the ones who report their gold derivatives. We took a ratio of what we thought were their deposits and what were their derivatives disclosed to the BIS, and that ratio came out almost exactly to what you would think given my description of bullion banking operations. It turned out that, for this small sub-set (and it was small and it could be unrepresentative), the face value of gold derivatives was maybe three times our estimates of gold deposits. That meant that, from the derivative data that had fallen into the public domain, we could infer a number on total gold loans from official lenders outstanding based on our analysis of the data on the gold derivatives. Once again, our sample is partial--- it is not total---but we believe it's pretty representative. We estimated from BIS data that the total amount of the gross gold derivatives of the bullion bankers, all 37 of them, has been somewhat more than 40,000 tonnes. That would suggest something like 10 to 16 thousand tonnes of gold have departed from the official sector as a result of official gold lending. This is an inference from a small sample, but it's an interesting corroborative piece of evidence.

Okay, now, let us look to yet another piece of evidence for more support for our gold loan estimates. In the Gold Book Annual we analyze the gold market like you would analyze any commodity market using microeconomics. What is analysis of commodity markets all about? It is all about elasticities with respect to price of the market's supply and demand variables---variables like jewelry demand, or mine supply, or scrap supply. It is also about changes in these variables over time. What is important about such changes over time with commodities is that, for a constant real gold price, for most commodities, demand grows less rapidly than global income. In economics we say that the income elasticity of that commodity is less than unity. In commodity jargon we say that this commodity has a declining intensity of use. Now, almost all commodities have an income elasticity of less than unity; in other words, they almost all have a declining intensity of use over the long run, at least in modern economies. BUT NOT GOLD. Excluding the monetary use of gold and focusing only on jewelry, on electronics, and the like, if you look at 200 years of data until 1997 what you find is that gold has an income elasticity in excess of unity. That is, demand rises more rapidly than global income over periods in which the gold price is constant in real terms.

Now, in commodity analysis you also have to look at the supply side over time. What we find out also about commodities is that, because of technological change and exploring new lands, for a constant given real gold price, mine supply increases. In the jargon we say that, because of technological change and the exploitation of new lands, the supply schedule for the commodity shifts outward at a certain rate. Now for most commodities, it shifts outward fairly rapidly. A combination of an income elasticity that is less than unity and a rapidly outwardly shifting supply curve implies that most commodities have, on trend, a declining real price. For copper, silver and the like, what has happened to their real inflation adjusted price over 100 years? It has fallen in real terms by 70%. But gold has not---gold has kept a constant real price.

This is an amazing thing because half of demand 100 years ago was monetary demand and today there is no monetary demand. In fact, monetary demand is negative because the central banks are dishoarding their monetary stock. So, if we just looked at the commodity dynamics of gold we would see that the gold price would tend to rise in real terms. Why is that? It is because demand tends to rise more rapidly at a constant real price than global income and mine supply tends to rise less rapidly. And, in order to make supply equal demand every year, the gold price has to rise in real terms to ration down price elastic demand and encourage more supply and thereby clear the market. Two hundred years of data from Eugene Sherman suggests this and 25 years of Gold Fields Mineral Services data, which is somewhat better data, suggest the same between 1971 and 1996. Now here is what is interesting. If you look at the Gold Fields data since 1996, what you find out is that, despite a big decline in the real gold price (remember gold demand is elastic with respect to the gold price), and despite perhaps 3% per annum increases in global income, demand has only increased by 10%.

Now if you apply the income elasticities that we have estimated from 200 years of data and the income and price elasticities that we have estimated from 25 years of data to the last four years---1996 to 2000---demand should have risen by something like 40% - 45% over those four years. Income went up, and the real gold price went down by a lot. The World Gold Council's demand series shows that demand went up by 20%---not 45 %---but the Gold Fields data contends that demand only went up by 10%. To assume it went up by only 10% implies that gold's income elasticity and gold's price elasticity have totally changed relative to history. We don't think that's plausible. We think at a minimum that the Gold Council's data is more reasonable; it allows for a certain amount of reduction in demand versus historical trends, perhaps because gold has gone somewhat out of fashion. But the "official" Gold Fields data is almost unbelievable. Now remember, the Gold Council's data shows an increase in demand much less than history would suggest; yet, it implies much higher levels of demand and much higher levels of supply.

Let us present yet another piece of corroborating evidence. These under reported official gold flows we are talking about are coming out of the depositories of the central banks. Now, in keeping with their unwillingness to be transparent, the central banks don't like to tell us what physical gold is in these depositories. However, we have data on what is in two of these depositories---the BIS and New York Fed. The physical gold in these two vaults at the beginning of the 1990's accounted for about one third of the officially held gold. Now, if you take a look at that 1/3 window on the total, what you find out is that we have lost almost four thousand tonnes of gold. The amount that has left those depositories that comprise only a third of all the gold in official depositories is almost equal to all the gold (5000 tonnes plus) that has supposedly left the official sector vaults in this decade through both selling and lending. If we prorate this drawdown from one third of the official depositories---that is, if we assume basically that there was the same kind of drawdown out of the other depositories (the country vaults, the Bank of England depository, etc.)---we come up with a draw down or liquidation that is consistent with OUR numbers on total gold lending and gold sales and not the official statistics. I should add, however, that this inference supports our more conservative estimates of outstanding gold loans (10,000 tonnes) and not our more aggressive estimates.

Now, in addition to the above three corroborative bodies of evidence we did a little bit of field research---we have had other people make inquiries with bullion bankers. (We went to other parties to make the inquires, since we feared that, as analysts, these dealers would be less forthcoming with us.) Some of these bankers had left bullion banking, some had been fired and felt disaffected and inclined to speak, some are still employed. In any case, they were willing to talk. We have gotten, albeit crude, estimates of gold borrowings from the official sector from about 1/3 to 1/4 of all the bullion banks. We went to bullion dealers and we asked, "Are these guys major bullion bankers, medium bullion bankers, or small scale bullion bankers?" We classified them accordingly and from that we have extrapolated a total amount of gold lending from our sample. That exercise has pointed to exactly the same conclusion as all of our other evidence and inference---i.e. something like 10,000 to 15,000 tonnes of borrowed gold.

So I have now given you 6 completely independent pieces of evidence that a hell of a lot more gold has left those official vaults than the consensus would contend. This implies that the flow, the draw down rate, the liquidation rate from official gold stocks is substantially higher than what the consensus contends.

Now let us put these two suppositions together. Remember our pie chart--- a big chunk of the official gold reserves reported to the IMF has already gone. Remember our supply/demand balances---this outflow on an annual basis is substantially larger. Now, let us project forward. First of all, as the years pass, global income will rise. At a constant real gold price we could then expect demand would rise somewhat. At the same time the current very low gold price is close to the total cost of production and we are having less exploration. We have more or less exhausted the pipeline of projects that we created through higher exploration expenditures years ago. Also, cash flow strapped miners are high grading the eyes out of their mines. Mines deplete in any case by about 7% a year. Mines are depleting more rapidly now because miners are high grading. In essence, we are not coming up with new projects to replace what is being depleted, and depletion is occurring at a rapid rate. Overall we can expect mine output will fall over time. Therefore, we can assume some growth in future demand, we can assume some decline in supply, so that the deficit in the gold market---that rate of flow of gold out of the central bank vaults---should increase in order for the gold price to remain at its current level in real terms.

Now, we will make two sets of assumptions. First let us take the current rate of drawdown and project it forward. Second, let us also assume some growth in that rate of drawdown. Let us then take our estimates of what is left in them thar' vaults and figure out how long this process can go on.

First, we take our conservative numbers--- our lower rather than our higher estimates of gold lending. Here we project how long this process can go on if we assume no growth in demand and no decline in supply and conclude it will take a decade to empty the vaults. In this alternative projection we have assumed some growth in demand and some decline in supply. It will take about 7 years to empty the vaults.

If we use our more aggressive numbers, we have less in those vaults and it is flowing out at a faster rate; consequently, it takes less than seven years to empty the vaults.

So whatever is happening in the gold market--- whatever is keeping the gold price down---if our numbers are correct, it can't go on that much longer, because we know not every central bank will lend or sell all it's gold. In fact, if our analysis is correct, the official sector knows what is coming. If the official sector is rational, it knows what will happen to the gold price when this large flow that is depressing the price abates and ultimately ends---the price will go up by a lot. Therefore, some rational central banks will not sell and lend down to the last ounce. Instead they will start to buy. So regardless of what has been happening in the gold market, if our data is correct, then, within a couple of years, whatever the official sector is doing, it will terminate and the gold price will rise.

What are the implications of all this dry statistical analysis for the claims of GATA? To our mind, it is very simple. There is much evidence that the consensus data on supply and demand is wrong and that the supply coming from the central banks is higher than the consensus estimates. In our opinion, the fact that the central banks do not acknowledge this but simply keep affirming the consensus data---despite abundant evidence to the contrary---represents considerable support for the allegations of GATA that there may be something deliberate and intentionally clandestine about the large flows of official gold that have been depressing the gold price.