Responding to criticism by retail and institutional investors, some major gold mining companies got together under the auspices of the World Gold Council and started working on establishing a set of accounting rules to measure what is being termed the "all-in" cost of mining gold.
Up until very recently, gold miners commonly reported what was termed "cash cost" per ounce. Although not a generally accepted accounting principle (or GAAP for short) this "cash cost" can be defined as:
The costs of mining, milling and concentrating, onsite administration and general expenses, property and production royalties not related to revenues or profits, metal concentrate treatment charges, and freight and marketing costs less the net value of the by-product credits.
In other words, it includes all costs incurred while mining and selling gold with all non-gold byproducts credited against the cost. Typically "cash costs" do not include indirect costs, such as overhead, royalties, capital cost and taxes. It usually also does not include exploration activities. This measure has been criticized, since it gives an incomplete picture of the cost per unit and has been misunderstood by many investors.
Large gold miners such as Goldcorp (NYSE:GG), Barrick Gold (NYSE:ABX) and Newmont (NYSE:NEM) have started to use a different metric called "all-in" sustaining cash costs in their most recent annual reports. Attentive readers of Barrick Gold's 2012 Annual Report will have noticed the following passage:
Our current definition of all-in sustaining cash costs commences with total cash costs and then adds sustaining capital expenditures, corporate general and administrative costs, mine site exploration and evaluation costs and environmental rehabilitation costs. [...] it does not include capital expenditures attributable to projects or mine expansions, exploration and evaluation costs attributable to growth projects, income tax payments, interest costs or dividend payments. [...] In addition, our calculation of all-in sustaining cash costs does not include depreciation expense as it does not reflect the impact of expenditures incurred in prior periods.
This passage already foreshadows the issues that have been criticized by some commentators who think that this definition does not go far enough; mainly for the following reasons:
- The issue of expensing capital costs seems to be a sticking point in the discussions. Mines are expensive to build, and this expense does not enter the cost equation for gold mined at the site once completed.
- Exploration is excluded but it is necessary to maintain a reserve and resource base.
- Money for tax and dividends etc. is part of the cost base which needs to be paid for by revenue generated from the gold that's mined.
A more radical approach has been reiterated by fellow author Hebba Investment here on Seeking Alpha. Hebba proposes to simply add up all cost positions over one calendar year and divide this summarized cost by the ounces mined during this year. This cost is termed as "true cost". Hebba's concept is not exactly new. In 2009, the exact same way of computing cost of precious metal production was proposed in this blog post on Gata.org.
The three listed methods yield vastly different results. Take Barrick's Q1/2013 results as an example. Traditional "cash cost" was computed to $561 per ounce. "All-in" cost is computed to $919 per ounce and Hebba came up with a value of $1215.
Let us deal with Hebba's methodology first, and let us state the main flaw of this method up-front. Using Hebba's approach does not account for cost positions that have an effect beyond the observed year (and therefore skews the result).
Imagine a growth-oriented junior mining company that is operating one mine and decides to build a second mine. Capital expenditure during the year of the mine build will be enormous and will quite probably govern this one year's overall cost summation that Hebba would produce for his calculations. Hebba would then go on and divide this cost summation by the ounces produced in the first mine of our imaginary company. The resulting number has almost certainly absolutely nothing to do with the "true cost" of mining gold.
Consider another scenario. Someone buys a car and drives a moderate amount of miles during the first year. Would it not be foolish to add the cost of the car to the petrol cost, then divide this summation by the driven miles and term the result as the "true cost" per mile? Yet, this is exactly what Hebba is doing in his articles. A large mining company could be compared to someone who owns multiple cars and drives these cars a lot, so the cost of purchase of yet another car is less dominant when computing the "true cost" as proposed by Hebba; however, it still skews the result.
Of course, this is a very simple accounting problem, and most likely Hebba knows the answer on how to correctly consider the new car in the parking lot in the balance sheets. We take the purchase price and divide it by a sensible number of years that we think we will be using the car. Then we expense the car over this time frame and only consider the portion for one year in our cost calculations. If we upgrade the car (say, we want new sub-woofers) we can expense the upgrade, just like we expensed the initial purchase over a sensible period of time. Smaller expenses, like the occasional oil change or servicing the car, typically have no effect beyond the yearly accounting period and are normally simply added to the overall bill.
To compute our unit cost, the cost per mile in this case, for a given accounting period (let's say a calendar year), we would take the purchase price of our car and divide it by the number of years we have chosen as our write-off period. Then we add our ongoing maintenance costs plus the petrol and also interest if we had to finance the car, and divide this summation by the miles driven during the year. Simple. Simple?
Re-interpreting the car example for a mining company would implicate that capital expenditure (capex) necessary to build a mine could (and should) be expended over a conservative life-of-mine estimate. The same goes to say about mine upgrades, mill extensions etc. (which would be comparable to our new sub-woofers in the car example). Operating expenditures such as new wheels for the mining trucks or cyanide for our leaching tanks are comparable to changing the oil or buying petrol in our car analogy.
Preferably, this would be the model we would like to see from mining companies when computing all-in cash costs. We build the mine, make sure it remains operational and mine the gold at a certain "all-in" cost per ounce that includes a yearly cost position resulting from the initial capex and associated financing cost.
However, there are still details that need to be included. How would we deal with write-downs at our mine? This would equate to having an accident in our car analogy. Unless we can spread the cost of the accident over a longer period of time (through financing for example) we would have to front up the cash to repair the car immediately and would therefore include this cost in our end-of-the year unit cost calculations. Same goes for mining gold.
How do we account for exploration and development of new projects for future mines? In our simplistic analogy, this would compare to searching for yet another car to park out the front of our house. We would have to invest time and transport and other costs to source a suitable vehicle. These sourcing costs have nothing to do with the unit cost of driving our other cars. We would therefore pay for the sourcing from our discretionary spending account.
Where does this discretionary money come from? Well, I forgot to say that in our analogy, we don't just drive our cars; we are actually getting paid transporting people with our cars, i.e. we are a taxi company. As we drive people from A to B, we charge them at a fixed rate (which would be called the spot price when we put our mining investor hats back on). Since we manage to charge more per mile than we incur in cost, we end up with a profit. At the end of the year, we calculate this profit and happily (not) pay our taxes for the greater good. We also look at what is left from the profit after tax and decide on a budget. Part of the profit will end up in our exploration budget; err, discretionary spending account. Other parts of the profit will be paid as dividends to shareholders of our taxi company. And so forth.
In other words, we do not see taxes, dividends and exploration cost for new projects as items that should be lumped in with cash costs for gold mines. (Yes, we have our gold mining hats back on now). However, we would interpret near mine exploration to maintain the reserve base of a mine as part of mine maintenance, and therefore include near-mine exploration into our cash cost.
To conclude, here is our wish-list for the World Gold Council when they meet again to decide on the final definition of "all-in" gold cash cost:
- Please expand the concept used by Barrick, Goldcorp, Newmont and others and include capital expenditure and associate financing cost in the definition of "all-in" gold cash cost.
- Please do not include items such as tax, dividends, green field exploration etc. in the "all-in" gold cash cost.