When gold reached $1,000 in March I became concerned that its price had risen too high, too fast and wrote that I would not be surprised to see it fall. I had no fundamental reason for believing gold was over-valued, but the rate at which the gold price had increased made me uncomfortable.

At the time I was still using M3 as a proxy for the money supply in my gold model, and the model suggested gold was fairly valued at around $950 an ounce. Furthermore, the model suggested that if inflation rates remained the same, the gold price should appreciate by approximately 15% per year. So while I felt uneasy about the gold price I had no good reason to be bearish. The revised gold model suggests that gold is in fact over-priced at the moment, and I feel much better now. The gold price could spike as a result of psychological fear, just as it did in 1979 and 1980 -- I cannot predict such events -- and if it does I will be liquidating even more gold and gold related assets. While I would welcome such an event, I would not risk my capital betting that it was imminent.

I don’t believe it’s possible to predict the future and therefore I don’t believe it’s wise to risk capital betting that this, or that, is going to happen. Such bets are not investing but gambling and most people who continuously gamble with their savings end up without savings. That does not mean we cannot figure out intelligent things to do with our capital. There are no guarantees that rational decisions will always lead to investment success but I can assure you that irrational and emotional decisions will, over time, lead to investment failure and the loss of capital. This applies equally well to gold and other assets as it does to stocks.

Subscriber question
The most common response to the gold model has been that it should compare gold to all fiat money and not just to the US dollar. This week I hope to put that notion to rest.

The gold model compares the rate of increase in the supply of US dollars to the rate of increase in the supply of gold to calculate gold’s fair value in terms of US dollars over time. We can do a similar calculation to arrive at a value of gold in terms of other currencies, such as yen, or euros, but it makes no sense to say that we can only arrive at the correct value of gold unless we compare gold to all currencies.

If we wanted to compare the value of the US dollar to the yen, for example, we would compare the inflation rate of the dollar to the inflation rate of the yen. We would not compare the inflation rate of the dollar to the inflation rate of all other currencies because that would tell us nothing about the relative value of dollars to yen. Similarly, if we want to compare gold to US dollars we need only consider gold and US dollars.

But let’s say we did want to look at gold versus all other currencies. We would then include the increase in the supply of all currencies and compare it to the increase in the supply of gold. But that means we would have to start at some point with the supply of all currencies -- we could not start just with the US dollar supply and then add the increase in the supply of all other currencies to the supply of dollars. When we’ve added all currencies together, what would it mean? The sum of all currencies is a meaningless number since it has no units. To convert this number into something useful we would first have to convert it into something with units. So let’s say we wanted to express the result in US dollars. To do that we would have to adjust all the currencies by their respective exchange rates against the US dollar, to arrive at a figure expressed in US dollars that we can now compare to gold. However, the exchange rates between currencies already incorporate their relative inflation rates so the result would again be the same as if we used only the supply of dollars and the supply of gold.

The above assumes that currencies’ respective exchange rates are in equilibrium, which is seldom the case at any one specific point in time, but is usually the case over an extended period of time. Markets can be quite inefficient when you take a snapshot in time but are extremely efficient over time.

Some people want to compare the supply of gold to the total of all currencies so that they could get a sense of what gold would be trading for if each unit of fiat money in the world would be backed by gold. Dividing the sum of all fiat money supply by the total supply of gold only tells us something about the relative value of gold should fractional banking be abolished. The current gold model incorporates the enormous expansion of fiat money to arrive at gold’s fair value accepting the reality that we have a fractional banking system.

If the world adopted a gold standard and simultaneously rejected fractional banking then the gold price would have to be much higher than it is today, but that is not currently the reality. In the meantime we have to accept that gold acts just like any other currency within today’s monetary system.

I agree that if one assumes that the era of fractional banking is imminently coming to an end an argument could be made for a much higher gold price. But there is no evidence of a sufficient revulsion to fractional banking to bring the system down, which means that such a line of reasoning is far more emotionally driven than supported by evidence.

People who rely solely on emotion have a tendency to think that gold is always under-valued and should always be priced higher, regardless of what the price is. I don’t think I need to say more about that -- the fallacy of believing that gold must always be under-valued regardless of its price is self-evident.

My good friend Brent Cook suggested that the gold model may actually be over-valuing gold since historical gold production records understate actual gold production. Many small mining operations don’t sell their gold through official channels and therefore their production is never recorded. It means gold production is higher than the figures I used and therefore gold’s relative value should be lower.

This is very true, but impossible to calculate since that data, by definition, does not exist. Nonetheless, unrecorded gold production is relatively small compared to recorded production and therefore the error so introduced is also small. There are many areas where errors could be introduced in both the gold data and the other monetary data available to us. We should at all times try to identify such potential problem areas so that we can make reasonable adjustments, if necessary. But I do not think the fact that we do not have perfectly accurate data should discourage us from using what we do have to try and make rational investment decisions. There would have been no progress in any branch of science, or any form of technological innovation, if we had not built models and proceeded to develop our understanding of the universe in the absence of absolute knowledge.