Nathan Rothschild, an 18th century British nobleman who made a fortune buying in the panic that followed the Battle of Waterloo against Napoleon, is most famous for his quip that, "The time to buy is when there's blood in the streets." A minor detail lost in the popular mythology is that, actually, Rothschild bought when he knew ahead of anyone else, that the British had prevailed at Waterloo. What he should have quipped, is that "The time to buy is when you know something that no one else does."
This subtle distinction is critical to successful contrarian investing.
I am no Rothschild, but I could not help but be attracted to the opportunity in Randgold Resources (GOLD), the gold miner whose share price plunged 30% in the days following a coup d'etat in Mali on March 22. I chose Randgold because it is a pure gold miner and holds about 80% of its reserves in Mali, so it is reasonably easy to isolate the drivers of the stock, and I'm intrigued by the possibility of understanding some things about it that the rest of the market does not.
As I write this update, the coup leaders have just signed an agreement to hand power back to civilian authorities, and the threat of international sanctions has been removed, but the story is not over and if the stock is not up 20% on Monday, I think it is still a buy.
Given the recent volatility of the stock and the popularity of gold-related themes in general, it is surprising how little has been published on this event in either Seeking Alpha or other sources. What I gain from reading the tiny amount that has surfaced can be summarized as follows:
1. No one wants to bother with trying to understand the politics of Mali. It's an extremely impoverished country just south of the Sahara and completely landlocked. It's most famous town, Timbouctou, is famous for being synonymous with an elusive, hard-to-reach destination. To a contrarian, it doesn't get much sweeter than this.
2. There is no agreement on how a resource company should be valued. Also sweet.
3. Few investors seem to have noticed, let alone understood why gold stocks under-performed the metal during a bull-market during a multi-year bull market for gold and gold miner earnings, so there was an unresolved mystery even before the chaos was unleashed.
Until recently, Mali was one of the more stable and democratic countries in all of Africa, but it has become a beautiful mess as an unintended consequence of the Libyan revolt, which accidentally led to a massive change in the power structure in Mali amongst the military, the democratically elected leadership, the tribal rebels in the north, extremist Islamic factions, and neighboring countries. Ironically, it is the neighboring countries, in my view, who actually pose the greatest threat to the mining operations.
Randgold initially began its precipitous decline following news of a coup d'etat on March 22. Had this been the only real problem, the stock would have already recovered because the coup itself was a non-event for the miners. The coup leaders are nothing more than disgruntled employees with guns. They have no political ambitions, no philosophical beliefs, no economic policy, and no desire or intent to disrupt the normal life of the people of Mali. They had a disagreement with the civilian leadership over military strategy. Within days of securing power, they met with mining companies to ensure that all the needs of the mining companies were being met and miners universally reported being satisfied with this meeting.
Whether you are a coup leader or a democratically elected president, the realities of Mali's economy are the same. Mali is an extremely poor country, with more than half the population living on income of less than US$1.25/day. Gold is the country's leading export and the only export that is not susceptible to drought or disease. No government can afford to treat the gold mining industry poorly and expect to survive.
Then, just as abruptly as it started, the coup ended on April 6 with the military rulers agreeing to step down and return power to a civilian government in exchange for amnesty and an end to economic sanctions. I guess people fear the worst when they hear the words "coup" and "Africa" in the same sentence, but this is why you should always do a little research on the personalities involved. Coup leader Amadou Sanogo, on the day before the coup, was an English teacher for the Army. He was trained by the US Marine Corps at the Quantico, Virginia base. He soon began calling himself the president of the National committee for the Return of Democracy and Restoration of the State. This is not the stuff of another Afghanistan or Somalia in the making.
Within 10-days of Sanogo's take-over, allegedly to correct the government's mismanagement of the rebellion in the north, the rebels had over-taken control of the regional capitals of Kidal, Gao, and Timbuktu leading Reuters to dub the coup a spectacular "own goal." Ethnically Berber, nomadic people living in West Africa's Sahel and Sahara regions have staged multiple uprisings in Mali and Niger for greater autonomy since 1916, but after the failure of the previous rebellion, Tuareg fighters volunteered as legionnaires in the Libyan Army. At the end of 2011, following the defeat of the Libyan regime, several Tuareg from the Libyan Army returned to the Azawad regions of northern Mali. Armed with significantly more lethal weaponry and tactical skills than ever before, they formed the National Movement for the Liberation of Azawad in October of 2011. The NMLA launched a series of offensives from January 1 and by February were reported to be in control of several parts of northern Mali. By late March, NMLA was in control of the entire region that they had set claim to.
Rebel held territories and Randgold Resources Mine Locations
Source: Company reports, mapsofworld.com
They had displaced more than 100,000 people without meriting any attention from Western media or having any effect on the share price of gold miners in the region. They unilaterally announced a cease fire when they had accomplished their mission and promised that they had no further territorial ambitions, which is important, because there are no gold mines in the territory they claim. The Tuaregs do not care about the gold. Gold miners did not appear to be overly concerned about the Tuareg's either, as none of them ever altered any of their operating plans during all of this time or thought it pertinent to mention any of this in their presentations to investors.
The latest analysis indicates that despite the lack of any international recognition, the Tuaregs cannot be removed from their positions. The world could get used to this. It had no bearing on gold prices before and really shouldn't.
The Tuareg movement is far from monolithic. It involves numerous factions, among which include the Islamic group calling itself Ansar Dine, which, unlike the MNLA, does not seek independence but rather the imposition of sharia law across all of Mali. However, all reports on the ground suggest that Ansar Dine has no military capacity to carry out its ambitions. Rather, they seem to ride in on the coattails of the MNLA and fill in the void left by MNLA's relative lack of administrative infrastructure. It is difficult to know how accurate these reports are, but clearly they would lose Tuareg support beyond the territory of Azawad (please see map), and since 90% of Mali's population claims to be Muslim, predominantly Suni, but women participate in the economy and do not wear veils, so it seems unlikely that AD could gain any degree of popular support in the more heavily populated South.
Ironically, the gravest threat to the mining operations turns out to be the well-meaning neighbors who have refused to recognize either the MNLA clams of independence or the coup leaders claim of authority. Throughout all of the chaos of these events, mining operations continued almost uninterrupted, but the international community's threat of closing borders and imposing trade sanctions would have surely stopped all transport of inventories and export of product. Such threats had almost no hope of having any influence on the rebels because they are motivated by ideology and they have no cross-border commerce, and even if they did, they probably wouldn't even know it. Sanctions would not have hurt anyone except the citizens that they would have been designed to help, and it is fortunate these have been averted. Still, in the process of writing this report, I worked out the valuation effect if the sanctions had interrupted trade for as long as 6 months. The impact is trivial, even in the case of the only real threat to operations that was ever plausible.
Randgold Resources was founded in 1995 by Mark Bristow, a geologist with a PhD from Natal University, South Africa, with one mine and 100 mn ounces of gold production in its first year. He has since built the company to include 4 mines and more than 700mn ounces in production per year. The company engages in discovery, development, and delivery of gold and only gold.
Major discoveries to date include the 7.5 million ounce Morila deposit in southern Mali, the 7 million ounce Yalea deposit and the 5 million ounce Gounkoto deposit, both in western Mali, the 4 million ounce Tongon deposit in the Côte d'Ivoire and the 3 million ounce Massawa deposit in eastern Senegal.
Randgold Resources financed and built the Morila mine which since October 2000 has produced approximately 5.8 million ounces of gold and distributed more than US$1.6 billion to stakeholders. It also financed and built the Loulo operation which started as two open pit mines in November 2005. Since then, an underground mine has been developed at the Yalea deposit and construction of a second underground operation is underway at the Gara deposit. The company's new Tongon mine poured its first gold on 8 November 2010.
Mining started in January 2011 at Gounkoto, another deposit on the Loulo permit in Mali, and it is scheduled to start supplying ore to the nearby Loulo plant in the second half of 2011. Randgold is fast tracking Kibali in the Democratic Republic of Congo, where construction of the mine began in 2011. In 2009 the company acquired a 45% interest in the Kibali project, which now stands at 10 million ounces of reserves and is one of the largest undeveloped gold deposits in Africa.
More than a few poor contributing souls to Seeking Alpha have been lambasted for the crudeness of their approach to valuing mining companies, but all valuation methods involve simplifying assumptions. If your assumptions are valid and consistent, than all methods should lead to the roughly same conclusion. A stock can't be cheap by one method and expensive by another. The stock is either cheap or it's not. The method is not what matters.
A mining company is worth all the future cash flows that it is able to generate from extracting the assets it controls and will control in the future, discounted by a rate that compensates for all of the risks associated with that process and the value of the time needed to complete the extraction. You should be able to get to this value by either valuing its assets or its cash flows. Doing both helps to verify the underlying assumptions.
Also important, There is no "ME" in "VALUATION." As a skilled analyst, I take great pride in my ability to get any price I want from a valuation process, but my assumptions do not matter. The point is to figure out what the market is assuming, and more importantly, to calculate what the market might assume next and how that might change the price.
Discounted Cash Flow
Actually, I never understand why anyone uses a discounted cash flow. For one thing, over very long-periods of time that you need to construct a valid discount model, earnings and free cash flows should be the same, and for another, the cash flows do not belong to us as passive investors so any difference in value that stems from the difference in timing actually renders the cash flow model wrong for the purposes of the passive investor. According to Mary Buffett, Warren Buffett uses an earnings discount model because it's easier to forecast, and whether this is true or not, it is what works for me.
There are three key differences between a mining company and a normal industrial company. The first is that the assets are nearly indestructible. This comes in handy during wars and famines, for example. I would not have been so interested in a producer of generic drugs in a war-torn country, with its highly precise equipment that could easily be destroyed by misplacing a tiny dust particle. On the other hand, the future market value of the end product, which drives the value, is more likely to change in unpredictable ways. Finally, the assets deplete, so instead of having to calculate a terminal growth rate and value for which there is not much theory and even less evidence, the mining company's terminal value conveniently falls to zero.
At the current rate of output, Randgold Resources is running down the value of its existing reserves at a rate of 2% per annum. This presents an interesting forecasting problem, in that, based on this information, the mines will go on producing for another 50 years and then suddenly become worthless. Here we go with the assumptions, because it is absurd to even think that you know what the marginal profitability of mining gold will be over the next 50 years. What we know is that the stock will have a price and that the price, whether anyone likes it or not, will imbed numerous assumptions about the next 50 years, and that by going through this process we can figure out what kind of assumptions are reasonable, and what kind are not. The most important reason for creating such a model is not to come up with some "fair value" which the stock will almost assuredly never trade at, but rather to give us the ability to calculate very precisely what the effect on value would be of any real, near-term alterations from the general path, say, a war. This alone gives you Rothschild-like knowledge ahead of the pack.
In constructing a base case, I like to start with consensus estimates for near term earnings. According to Bloomberg, the consensus expects the company to earn US$686mn in 2012, US$ in 2013 and US$6 in 2014. (I used the ADRs in this model.) Because the price of gold, and the margins earned by gold miners over the past several years have been so strong, it seems prudent to me to assume that over the following five years, profits decline at a 5% rate per year. This is a pretty radical decline because we expect production to increase throughout this period, but it is not unprecedented so I think it is a prudent assumption.
Thereafter, I assume that gold prices keep pace with extraction costs and that production increases at a rate that fades at a constant rate, starting at 6% and fading to zero. I make a simplifying assumption that reserves increase by an amount equal to the amount needed to maintain the useful life of the operation at 50 years. At the end of the forecast period, the mines are worth zero. These are wild assumptions. I don't have any reason to believe any of them except that I think they are prudent. Reserves have recently increased at a much more radical rate, but we know that there is a limit to how much Gold is left in the ground. For all we know, in 50 years'-time, Randgold might transform itself into a life-insurance company, but that would not be a prudent assumption and the market would never pay for that kind of assumption.
I use the 10-Yr US treasuries as the risk-free rate. "Risk-Free" is a misnomer. Treasuries are far from risk free, but what we are really looking for is a bench-mark. I don't think 1-Yr treasuries are a reasonable benchmark for a long-term investment.
Now we arrive at a problem that is unique to gold mining operations. The stock's beta is 0.6, but this is because it has a low correlation to the market, not because it has a low level of risk. As I think I've been fair and honest about, the assumptions we have to make to get this done are all crazy and the ease with which any or all of them could prove wrong are at least as high as that of the average stock in the market. The beta of such an operation has to be more than 1, but I do not know of any really sound theoretical means for determining the right beta any more precisely than this.
If we start with "something more than 1," we will be on the right track. Fortunately, I am only interested in what the stock market will pay for my stock when I try to sell it, rather than what it should theoretically pay, and I have a kind of elegant solution that is pure voodoo, but it works for me. I back out the beta needed to get to the price that has prevailed in periods with similar characteristics to what I consider to be normal. For Randgold, that beta turns out to be about 1.2. To get to the current price, we would need a beta of 1.5, but since we know that the recent price is based on events and conditions that pretty much define abnormal, the beta of 1.2x is a more prudent assumption. I told you it was voodoo, but let's see what we are able to do with this tool now that we have built it.
If the assumptions we have now built in to the model are accurate, the fair value of the stock is $104 implying a 28% upside from the current price.
Now we can ask, "What if operations are halted for 6 months because the company is unable to access supplies?" By halving the current consensus projections for 2012 earnings, the fair value drops only $3, to $101. Whether or not we have the right base case, we can very soundly determine that the market has panicked outrageously over an event that in the worst conceivable case is trivial. The key to driving this conclusion is the fact that we are evaluating a mining operation where a downward revision would only reflect the time value of the delays in extraction and have no effect on the underlying assets. A six month stoppage of any other business would have much greater repercussions because of the potential for lost reputation, damaged equipment, and lost market share.
Value of Reserves Less cash Costs Per share vs. Share Price
Source: Company reports
Share Price Compared to Earnings
(Click to enlarge)
Source: Company reports, consensus forecasts
It would be far more significant if we dropped the growth rate in the final year of our explicit forecast to 3.5%, from 7%. This would drop the projected fair value to $84, which would eliminate any meaningful upside in the stock. This could happen, but since we've already made adequate allowance for the potential of a decline in the gold price, and, therefore, the operating margins, a decline in assumptions of this magnitude would most likely only come from a re-assessment of the company's ability to increase reserves. By making the model, we know that such a scenario was not in anyone's mind before. It is still a plausible risk, but an unlikely outcome given the company's current pipeline and its track record of developing reserves. Given that it is an unlikely scenario and that the consequences if it does happen are not exceptional, it is a risk that I am willing to take in exchange for the opportunity to take part in the far more likely scenario of a 30% rise. In fact, the risk is balanced by another outside possibility that the rise is much higher even than 30% if momentum caries the stock much higher than its fair value.
Randgold's latest reported reserves totaled 16.4mn tons. The realized price of finished gold during the past quarter was $1,666, with a cash cost of $753. I have not independently verified this cash cost calculation but I'm unaware of any obvious reason to doubt it, so we will just bundle the possibility that they are using liberal definitions of cash cost into the general category of risks. This gives a total gross value of the current reserves of $164 per share, but at the current rate it will take 50 years to extract all of the gold so we must discount this by multiplying it by dividing by 50, then by the discount rate, which, for expedience, we can take from the previous discussion, less the growth rate. This is where the asset valuation model shows itself to be far the more crude because you have to lump all the changes in price and increases in production as well as the potential for new discoveries into one number.
Essentially, there is no difference between what we are modeling, but the asset model is simpler. Instead of multiple phases of growth and explicit earnings forecasts, the asset model uses the cash extraction value and lumps all the growth phases into one number. If we use a rate of 6%, the resulting fair value is $100/share, very similar to our discounted earnings model conclusion of $106. The asset model is extremely sensitive to this growth rate, however. If we use a rate of 5%, and hold the discount rate constant, the value drops to $75. First of all, 5% growth is not what the market was assuming at any time prior to the coup, so something would not be right about it. My position is that the issue is not so much one of the correct growth rate but the correct spread between growth and risk. When you drop the growth rate, you should also drop the discount rate because the less aggressive your forecast is, the less risk there is. Absent the coup, $100 is a reasonable target for this stock.
Does the actual share price ever correspond to anything like this kind of calculation of fair value? It would be pretty surprising if it did, but it is instructive to see how and when market prices differ from theoretical values. Historically, the price of Randgold has not been anywhere near the theoretical value, but the explanation is fairly simple. Historically, the stock has traded at a significant premium to these calculated values because of the anticipated increase in reserves, and the optionality of a potential increase in the gold price.
With the explosive growth in reserves and the price of gold in recent years, however, it is prudent to discount the possibility of further increases more aggressively. The plots of share price to net current asset value and net earnings per share demonstrate that the market, prior to the coup, took exactly this approach, i.e., where the stock had previously traded at a premium to the current value of assets, it recently began to trade at a discount. This was a perfectly reasonable discount, not the completely crazy discount that prevailed at the end of last week.
Therefore, once the dust settles around the coup and the rebellion and investors become more convinced that there will be no further impact on the mining, it is reasonable to speculate that the market price will again reflect these kinds of basic assumptions and re-price the stock some 30 to 40% higher than the price that prevailed at the end of last week. In the worst case scenario, the stock is certainly unlikely to fall very much further so we have the potential for an outsized gain relative to a very undersized risk.
Gold Stocks vs. Gold Bullion
Gold stocks are supposed to outperform the metal in bull markets. That's the rule because gold miners have fixed costs and as a result, their operating profits rise far more in a gold bull market than does the price of the metal. But in the past 10 years, Gold has risen 444% while the stocks rose only 348%. Recently the stocks, have been underperforming the metal at a far more frightening pace, and well before the Mali coup were almost universally trading at rates to the metal not seen before the 2008 liquidity crisis or the depths of the last bear market for gold.
One explanation I have seen that doesn't really work is that the increase in availability of ETFs that own gold has created a shift in investor preferences toward owning the metal. There is no reason for the availability to shift the preference. The economics haven't changed. Gold earnings still rise and fall further than the changes in the metal price. This explanation fails the smell test, primarily because the stocks outperformed the metal both before and immediately after the liquidity crisis exactly as theory suggests they should. The counter-intuitive behavior of the stocks began only in 2011, which does not correspond to the introduction of any gold ETFs. The change more clearly correlates to a divergence between gold and other metals that many of the various gold miners also engage in. As long as one sticks to a pure gold miner such as Randgold, there is no reason to believe that the conventional economics of owning stocks vs. the metal have changed.
Gold Bugs Index / Gold Spot
(Click to enlarge)
Disclosure: I am long GOLD. I have bought puts on GLD to hedge the exposure to volatility in the spot price of gold bullion.