The Production Cost Of Gold May Surprise You

Apr.23.13 | About: SPDR Gold (GLD)

Lost amidst all the noise and drama in the gold market is a simple, fundamental truth:

Gold is simply a thing that must be produced and then sold.

In this sense, it is no different from any other object in the world, whether it's a house, a barrel of crude oil, or an iPad. For our purposes today, the question of why a person wants it is irrelevant. All that matters is that someone does and that someone else has the capacity to make the thing in demand.

You have a curve that represents the supply of the thing and a curve that represents the demand of the thing. This is the model that explains what the price should be.

Today we're going to look at cost. How much does it cost to make an ounce of gold and how much did it used to cost?

There's obviously a much more in-depth, robust way to conduct a study like this. But for guys like you and me who are simply looking for a hint to the big picture, today's data and methodology will suffice. I simply selected the biggest gold miner in the world, Barrick (NYSE:ABX), and combed through all of the audited financial reports that I could find.

The funniest thing about this whole exercise was watching how the design and aesthetic of these annual reports change over the years.

My personal favorite was this one from 2008:

(click to enlarge)Click to enlarge

It serves as a reminder that, ultimately, these financial reports are really just another marketing document.

It also explicitly lets you know how both Barrick and their customers view gold. Gold is anything but a safe haven -- the annualized standard deviation of gold is over 21%. That's more volatile than the stock market. And Gold features a history where it trended lower in price for two decades. Not a very safe haven, gold.

The part of me that cares about things like this hopes that this is just marketing and that Barrick doesn't truly believe that gold is a safe haven. If so, it would be a spectacular example of cognitive dissonance and behavioral bias.

Anyway, while the flavor and message of these reports change from year to year, the good news is that there is one data point that stays constant throughout:

Total cash costs per ounce.

I noticed that at various times and places in these reports they have utilized several different "alternative" measures of the production cost of gold. In the latest report, there's a brand new one called "All in sustaining cash costs per ounce."

The important thing to understand today is that none of these numbers are ironclad. In reality, it's difficult to define precisely how much it costs to produce one ounce of gold. The way that Barrick does it is by taking their total cost of sales, adjusting for depreciation, and then backing out a bunch of other little things like other metal sales and sales costs applicable to

The result is "Total cash cost of sales." That number then simply gets divided by the total amount of gold produced and -- voila! -- total cash costs per ounce.

What this means is that this number has the potential to change a lot depending on how the whole company's cost structure changes. And just because the cash cost to produce one ounce this year was $584, doesn't mean it can't be lower next year. Barrick can do all sorts of things to reduce their cost structure, and thereby lower the cost per ounce.

As it happens, this is quite common. While the price of gold was dropping, dropping, dropping in the late 90's. Barrick was getting more and more efficient, lowering their cost per ounce year after year. They even boasted about this fact in one of those old reports. This should serve as a reminder that the psychology of a deflating gold market is completely different than an inflating market like today. Today their report is loaded up with supplemental data and alternative calculations that seem to imply that the cost to produce an ounce is a lot more expensive than you'd think based on the raw financials.

In any case, the cost to mine an ounce of gold doesn't necessarily have to rise year after year ad infinitum until all the gold is gone. Any business can get more efficient. This is something to think about when you start to have a rational conversation about how low gold prices can go.

Now we're ready to take a look.


(click to enlarge)Click to enlarge

Pretty crazy, huh?

Some notes:

1. The spot price is an average price of the previous year.

2. "Cash Costs" is the same as total cash cost per ounce and also represents an average through the course of the entire year.

3. "Gross Cash Margin" is simply one related to the other. (#2 divided by #1). I couldn't think of a more creative name for it. Sorry.

Here are a few points that jumped out to me.

There has always existed a reasonably consistent "spread" between the total cash cost per ounce and the price of gold in the market.

I'm not in the gold business, but if I was, I'd guess that a lot of what it costs me to produce is a function of the price in the market. If prices are falling, I'm working extra hard to make my business as efficient as possible (and lowering that cost/ounce). If prices are rising, I am aggressively expanding and operating the mines that are the most expensive to run at high capacity.

As the price of gold was rising steadily through the aughts, the cash costs were increasing at a somewhat similar rate. In 2003, when gold was trading at $367, cash costs were $189 (~50%). By 2009, when gold was averaging $981/oz, guess what, cash costs had increased to $466/oz.

The price of gold increased 167% while costs increased 145%.

Did the price of gold rise because it got more expensive to make? Or did Barrick (and presumably other miners) just spend a whole lot more as prices were rising?

Certainly a bit of both, but if I had to guess, I'd say it's a whole lot more due to the latter. All I know is that when my business is booming and demand for my product is rising, I'm spending as much as my CFO allows me to. Did an ounce of gold really become that much more expensive to mine?

Things have gotten pretty crazy lately.

CONSPIRACY ALERT! As the price of gold was bubbling in 2011, the "gross cash margin" was just over 70% for Barrick. This represents the biggest margin that I could find, by a considerable amount. Perhaps that's why 2012 was the year where they introduced this new "all in cash sustaining costs" concept.

That alternative metric was $945 in 2012 which, COINCIDENTALLY, would represent a "gross margin" of almost exactly 50% of the all-time high price of gold. In case you were curious, the average "gross cash margin" prior to 2009 was 48%.

(I mention this just in case you thought there wasn't still creativity in the world of accounting.)

Also, as the cost to produce one ounce stayed relatively constant in the post crisis years -- which makes total sense because a lot of businesses were managing themselves very conservatively during those years -- the price of gold was taking off. 2011-2012, the peak of the bubble, was the window that featured the biggest 1-year change in the history of their production costs. The cost to produce one ounce of gold jumped over 25% between 2011 and 2012. Why?

After four straight years of almost perfectly consistent costs, why a 25% jump?

If you think about it, 2012 was in a lot of ways the "all clear" year. This was the year that we all started to feel that things were getting back to normal. Or at the very least, we started feeling confident that it wasn't all going to fall apart again. Maybe that's when Barrick finally felt better about aggressively spending more money and ramping up internal costs?

After lifting their hedges back in 2009, perhaps some general operational conservatism might also explain why costs stayed constant in those years before finally jumping once the bubble got really crazy.

If you think that gold magically got 25% more expensive to pull out of the ground 2012, maybe think again.

What does this mean for the future gold?

I'm already on public record about gold being in a bubble and that the risk/reward at these levels is tremendously unfavorable. Over the long, long run, gold should, by definition as a dollar-denominated asset, appreciate at a rate roughly similar to the inflation rate. That's it. Any trend that goes too far above that baseline is bound to correct towards the historical mean.

Gold's actual rate of appreciation has vastly exceeded the inflation rate, of course. Even if you go all the way back to the early 70's and even if you use alternative, more aggressive ways of measuring "inflation," it's very difficult to justify any price higher than around $400-700.

Is it coincidence that that band also resembles the cash cost of producing an ounce of gold in recent years? You be the judge.

Full disclosure: I do own a bit of the gold ETF (NYSEARCA:GLD) in two of my Alpine Advisor portfolios (Black Diamond Global Growth, and Green Valley Conservative), but it's in there by force of mandate since these are diversified portfolios. Gold is different and doesn't correlate and that makes it a tremendously valuable tool for any diversified portfolio. But gold has been at its lowest permissible allocation threshold since 2011. One should always employ a tactical asset allocation approach with gold.

Regarding the Aggressive portfolio, which takes concentrated directional bets, this is what I wrote back in the February issue:

Best case, gold seems to be locked in a channel, and worst case, it could be on the brink of completely falling apart as the descending pennant breaks.

I eliminated it from that portfolio entirely after it had been a major holding since I started these portfolios. I really should have clipped it out a year earlier. But with bubbles, I suppose the "better late than never" maxim still applies. Also, investors who waste time hypothesizing about alternative histories are probably better off focusing their energy elsewhere.

If you're still heavily long gold, ask yourself why. I hope it's not the money printing or the conspiracies or because you're betting on the apocalypse. Check out Ritholtz's 12 Laws of Goldbuggery and make sure you're not committing any of those sins.

Projecting the price of gold from here is a fool's errand. It's an impossible exercise that's bound to make you look stupid. But now that we've got some data on what it costs to make this "stuff" we have a better idea about some of the fundamentals that ultimately have to support the market price.

If Barrick simply cannot mine gold for a profit when prices are $400/oz, they aren't going to mine much gold. Eventually supply and demand will clear and prices will move back to a level where it makes sense for them to produce more gold.

Any way you slice it, we're still far above that level.

Disclosure: I am long GLD. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.

Additional disclosure: For more disclosure, see