If predictions are to be trusted, we should all be dead.
It has come to our attention that the end of the world occurred a few weeks ago. Or at least it should have, according to an ancient Mayan calendar. Although the precise manner in which we were to perish is unclear, it seems that The End was to be brought down upon us by some combination of large solar maximum, an interaction between Earth and a black hole at the centre of the galaxy, or Earth's collision with a planet called "Nibiru." Apparently, December 21, 2012 marked the end date of a 5,125-year cycle and also the end of the universe.
Outlandish as this idea may appear, it was but one among a vast number of predictions made every single day. Most of the predictions we make, of course, do not fall into the cataclysmic, end-of-the-world variety, but are more humble in nature as Nate Silver explains in the excellent The Signal and the Noise. Every time we go outside for a walk, decide whether to go on a vacation, or choose a route in the car to beat traffic, we are making a forecast about how the future will proceed. Predictions are indispensable to our lives. Yet they often go astray.
History is littered with examples of predictions made by experts that turned out to be categorically wrong. These anecdotes are supported empirically by a number of studies that highlight how poorly most experts forecast. Experts seldom have any forecasting advantage over chimps randomly throwing darts at a board. As it turns out, humans are not especially gifted at making them, particularly the more complexity becomes a factor. And yet our societies persist in making forecasts in exactly the kind of complex scenarios we know to be difficult.1
What drives us to this obsession? To paraphrase Nate Silver, the answer seems to be a deeply engrained pattern recognition. Ten thousand years ago, our Paleolithic hominid ancestors spent their days as huntergatherers. In the tall grass of Sub-Saharan Africa, the gentle rustle of leaves in the bush might not have merely been the wind, but a lion from which to flee; the crack of thunder in the sky, a sign of a potential downpour. We had no claws, no ability to fly, no ability to camouflage. All we had were our wits. In these circumstances, recognizing patterns became an extremely useful trait.
Genetically, not much has changed for us since our Stone Age days. We are wired to detect patterns and to spot opportunities and threats in the landscape. The problem is that these Stone Age advantages have become Information Age disadvantages. We see patterns in situations where none exist. And there are few places where this is more obvious than in the stock market.
Stock markets are exactly the kind of complex systems that make for bad predicting. Complex and adaptive, stock markets are based on numerous interactions between the economy, businesses and human psychology. They are dynamic and nonlinear systems, where the behaviors of the system at one point in time influences its behavior in the future. In sum, they are among the worst systems in which to attempt to predict. Even the job of the weatherman is easier.
This tension points to one of the central challenges in portfolio management and the main theme for this year's annual letter: prediction amid uncertainty. Or, more specifically, how can we manage a portfolio if we cannot predict with certainty the events that will unfold?
It is a theme that is particularly appropriate given the macroeconomic uncertainty that dominated market discourse over the last year. Debates over whether Chinese growth would falter, or whether Europe would emerge from its sovereign debt woes, or the U.S. would go over the fiscal cliff, exhausted media airwaves and investor attention in 2012. Prediction was often a knee-jerk reaction to the unclear direction of the global economy; an attempt to achieve respite from the fear that clings in our chests when there is a lack of certainty. Indeed, we spent the better part of our written communication to clients this year addressing our concerns about relying on predictions.
But where we focused earlier this year on the faults of prediction, we now focus on what can be done instead. If accurate and systematic prediction of events is impossible, how can we adequately manage uncertainty? What processes do we employ to do so? And how does resiliency actually manifest within our clients' portfolios? While we certainly cannot claim to possess the answer to managing risk amid uncertainty, we do hope that what follows helps clarify our approach, both in how we manage risk and how we tame our own inner Mayan.
Imagine that tonight you turn on the TV to your local news channel. Just as you're about to hear the final score for your favourite sports team, the regular programming is interrupted by some breaking news. The news anchor informs viewers that tomorrow there will be a major and deadly disaster to befall your city – and then, suddenly, the TV and all your electronics go dead before you can hear the rest. You have no clue what the disaster will be or when it will strike. How do you prepare?
How you choose to prepare will likely be determined by what you believe is the most probable disaster to hit. If you live in a city prone to earthquakes, like San Francisco, you may hide your family under a sturdy table. If you live in a city prone to tornados, you may head to the basement. But if you have been reading articles on the recent resurgence of the Bubonic plague, you might assume that a deadly infection has spread and stock up on face masks and food. Or, if you've stayed up too late during the holidays watching Zombie shows, you might believe that the Zombie apocalypse has come and go and grab your baseball bat.
How ever you ultimately prepare, the manner of your preparation would depend in large part on the assumptions you make in that moment. Clearly, some preparations would be appropriate under a certain disaster and totally inappropriate under another. What good is a baseball bat against a hurricane? Being right about which specific disaster will unfold would be less important in this scenario than being prepared for the variety of disasters that could occur. It is better to be capable of responding to a number of outcomes well than to one outcome perfectly.
The concept in investing that best reflects this kind of preparedness is what we refer to as resiliency. At its core, resiliency in investing is about having the ability to bounce back from whatever scenario unfolds. Where traditional risk management focuses on the prediction of risks in order to protect against them, a focus on resiliency aims to protect capital regardless of the events that unfold. Instead of just finding the bat, resiliency might suggest that you stock up on canned goods and face masks, and would advocate for buying a house with a solid foundation in the first place.
Resiliency is, in other words, an approach to managing uncertainty. Why is this important to clarify? Because
there is an important distinction between uncertainty and risk, and there are implications for portfolios depending on which one you choose to manage against. While both words describe the probability of an event occurring or not, one can be measured (risk) and the other cannot (uncertainty). And in our experience, the investing world is much more dominated by things that cannot be properly measured.
As an example, imagine that you are rolling a die and you make a bet that a 6 will come up. Although the outcome is unknown, you know that you have a 1-in-6 chance that a 6 will be rolled. This is risk. Now imagine that you make the same bet, but this time you don't know how many sides are on the die. What will happen is still unknown but now you can't even measure it. This is uncertainty. Poker and blackjack are games of risk. The outcome of war is uncertainty. How you choose to act in a given situation depends greatly on which one you think you are dealing with.
Let us return now to the term, "resilient." The thought of the word might conjure up images of impenetrable fortresses and great, big rocks. Things that look and feel sturdy and are fashioned to be immovable. These images might incite feelings of safety, but they do not necessarily reflect the kind of qualities that create resilience within a complex, adaptive system like an investment portfolio. Rather, the kind of resilience that we refer to when we use the term has its etymological roots in ecology. Instead, it is closer to ecological resilience. In ecology, resilience is the ability to cope with unexpected disturbances in the environment by resisting damage and recovering quickly. When a fire is caused by a random lightning strike, and the forest recovers and thrives, that is ecological resilience. Often, when an unexpected disturbance occurs in a resilient ecology, the system itself grows stronger. Like in ecology, an investment portfolio is resilient when it can limit its downside from unpredictable negative events and continue to compound wealth.
What are some indications that a portfolio is resilient? In our experience, resilient portfolios share four traits: they are adaptive, they are diversified, they demonstrate convexity, and they leave room for a margin of error. They are adaptive because they respond to evolving conditions in a flexible manner and continue to compound wealth; they are diversified in that they are balanced in their exposure to risks; they demonstrate convexity because they lose less in negative events than they win in positive ones; and they leave room for a margin of error, which helps guard against human bias and the inevitable occurrence of wrongly placed bets. While these concepts are simple, by no means are they easy to implement. As usual, the devil is in the details. How do you build resilient portfolios in today's environment?
GUERRILLA RISK MANAGEMENT (BOTTOM-UP RESILIENCE)
In 1950, the French were frustrated. They had been fighting the Indochina war in the thick jungles of Vietnam for four years and they just couldn't seem to win. The problem for the French was not that they were outgunned; in fact, they had a sixteen-to-one artillery advantage. The problem was that were competing against a headless enemy. The Viêt Minh was not fighting the big, decisive battles that the French were used to. They were using Guerrilla warfare tactics. They fought in small, independent groups of between 5 and 10 members that could respond to local threats organically. On one occasion, the French even won a large battle and thought the war would be over, only to be surprised later on that their enemy was still fighting.
The Viet Minh's approach was highly unusual and one of the major reasons for ultimate French withdrawal. In traditional warfare, the chain of command flows primarily in a top-down manner from the General to his or her troops. But as the example above shows, there can be a benefit to responding to opportunities and threats in a more decentralized manner. In WWII, killing Hitler meant the end of the war. But who was the equivalent head of the Viêt Minh? The system that the Vietnamese used during the Indochina and Vietnam Wars was remarkably resilient and caught the western world by surprise. As Henry Kissinger noted: "the conventional army loses if he does not win. The guerrilla wins if he does not lose."
In the investing world, the traditional approach to risk management is also mostly centralized. The portfolio manager adjusts portfolio weights according to his or her view on the risks in the world. There might be some differences in how managers ascertain risk and react to it, but it is predominantly centrally controlled. Yet a portfolio is much more likely to be resilient if adaptability, convexity, and room for error can be built into the portfolio on a bottom-up basis. And so the first step in building a resilient portfolio is not with the General but with the troops.
Within our equity portfolios, there are several approaches that we take at the bottom-up level to be resilient: we invest in businesses that can adapt to different environments, we value them on a probabilistic basis, we emphasize balance sheet strength, and we invest alongside managers that are excellent capital allocators. The first approach is relatively straightforward. Companies that are most likely to adapt to changing environments are those with attractive business models and sustainable competitive advantages. Consider the case of one of our portfolio holdings, BHP Billiton, one of the largest miners of iron ore, copper, gold and coal. The Anglo-Australian company has roots all the way back to the 19th century and is among the world's most successful companies.
Mining is a notoriously difficult and cyclical business. Miners produce commodity products that must be sold at the market price and where differentiation is basically impossible. They are subject to huge swings in commodity prices and to the vagaries of global supply and demand, none of which they have any control over. When the market is booming, they make profits deftly, but when the market turns and goes bust, many highcost miners go bankrupt. It is no surprise that long-term weighted average return-on-equity (ROE) figures for the mining industry tend to be much lower than the long-term average ROE of the S&P 500.
Miners now face considerable uncertainty. Since 2000, growth in China has fueled a huge demand and subsequent rise in prices for most mining products, including iron ore. The dollar price per dry metric ton has soared from $12.05 in 2002, to a peak of $187.18 in February 2011, falling back to $120.35 in November 2012. For the last few weeks iron ore prices have been on an upwards path, but in which direction are they headed next? What if the growth rate of demand for steel products – iron ores' main output – falls? Or if demand itself drops? With China shifting its growth model, there is some indication already that the forces that drove the spectacular rise in iron ore over the last ten years may not continue into the future. What does this mean for mining companies and BHP?
A miner might seem like a risky proposition at this point. But investing in BHP is not the same as purchasing a broad mining index. Where high-cost Chinese miners tend to require breakeven iron ore prices of $100- $150 per dry metric ton, BHP has several very large, low-cost assets with breakeven prices closer to $30-40, and should generate returns on capital employed of 8% in the $60 range. Their low cost asset base, scale, and diversity of commodities are significant competitive advantages that provide buffer and flexibility for the company during market busts. This is one of the main reasons the company has weathered different cycles over time and should continue to compound wealth into the future. BHP is likely to survive whether China grows at 6% or at 10%.
BHP also demonstrates why it is so important to value businesses on a probabilistic basis. Most analysts that cover BHP provide a single point estimate for what they think the company is worth. In creating these estimates, the analysts must make a huge array of assumptions, including the price of iron ore, coal, copper and oil, all of which depend on their own highly complex set of assumptions. After making these estimates, the analyst then writes a report and confidently announces the valuation "answer." This is not unlike the Mayans calling December 21st, 2012 the end of the world.
We have found over time that it is better to be approximately right than precisely wrong. There can never be a precise "answer" to the value of a company, as this would depend on having an untainted and complete understanding of exactly how the world will unfold in the future, and this is impossible. However, it is possible to create a forecasted range of prices for what we believe a company could be worth. This is more intellectually honest.
The way that we create this range is through our discounted cash flow models. We build models that forecast a company's cash flows into the future, and then discount these cash flows back to today to arrive at an estimate of what the company is worth. The assumptions we make are based on our understanding of the company's business model, competitive advantages, industry, history, and management team. We test our assumptions through conversations with customers, suppliers and competitors. But the process does not stop there.
Say Chinese demand for base metal products continues at its former rate. In this scenario, the long-term price of iron ore and copper might be $150 per dry metric ton and copper $3.00 per pound. Most miners would continue to meet their cost of capital. But what if China decides it has built enough skyscrapers and infrastructure? If iron ore and copper prices drop to their marginal costs of production, which is consistent with how commodity prices operate over the long run, many miners would not earn enough to meet their costs of capital. How can we reflect this uncertainty into valuation?
One solution is to look at a wide range of outcomes for BHP. In order to get a range of prices, we look at what the company would be worth under thousands of scenarios. This way, we can at least articulate whether a stock is more likely cheap or expensive. In the case of BHP, the stock market appears to be pricing in iron ore and copper prices around $130 and $3.00 respectively, which is significantly above long-term estimates of marginal costs. Since we prefer our models to be conservative, we assume a world in which commodity prices fall to their marginal costs of production, which is believed to be closer to $90 per metric ton of iron ore. We then take long-term prices to around $70 for iron ore and $1.70 for copper. How much could BHP be worth in this scenario?
Even in this world, BHP's stock is probably worth between $39 and $70, equivalent to roughly 7 to 9% annual rate of return if held forever. This is not a spectacular return per se but it is an attractive enough range for a base case. And in the scenario where commodity prices remain at elevated prices forever? The stock is more likely to generate a return closer to 9 to 12%. This would also be reasonable.
How does this process create more resilience? It is a systematic way of incorporating more of a margin of safety into our portfolios. If each individual security we purchase is more likely cheap under a variety of scenarios, we are more likely to have a portfolio of attractively valued stocks. It is also one of the main tools that help our team remove the biases of prediction.
The two other methods we use to build resilience at the stock level are balance sheet strength and management. When it comes to balance sheets, we are cautious with management teams that seem too willing to take on debt or when the company has too much of it in general. As the 2008-2009 financial crisis demonstrated, excess leverage magnifies the potential for losses. This defeats the purpose of trying to build convexity – losing less than we are winning – into the portfolio. High amounts of debt are especially perilous when there is presently a high threat of a banking or liquidity crisis, like we saw in Greece. Our strong preference is to limit balance sheet risk and avoid companies with too much leverage.
Management is another key factor in determining the resilience of a portfolio. Like a General commanding an army, there is only so much information that can be seen from a portfolio manager level. Management teams are on the ground and see opportunities and threats to their businesses that we as investors could hardly be expected to anticipate. Good management teams will steer their businesses prudently through cycles and reduce risk in their business, where bad managers tend to increase risk in the business. This makes the caliber of management of utmost importance.
Over time, we have found that good managers surprise us with how much value they can create and bad managers shock us with just how much value they can destroy. Thankfully, we believe that the vast majority of management teams within our portfolios are good managers, and some are even great. How do we arrive at this conclusion? We evaluate their track records of value creation, the decisions they've made and how successfully they've steered their company throughout their history at its helm.
In sum, building resilience at the security level is a strategy we have found to be quite successful over time. There is value in bottom-up effort. However, there is also an important role that remains for top-down management. The General still has a role.
DON'T GET PUNCHED IN THE FACE!
Muhammad Ali was one of the best heavyweight boxing champions the world had ever seen. Known for his unorthodox fighting style and pre-match trash talk, Ali brought a level of grace and entertainment to boxing in what was then boxing's Golden Era. Unlike most heavyweight boxers, Ali rarely kept his fists near his face to block blows, preferring instead to rely on speed and dexterity to dodge a hit. This unusual style was a bit like dancing and allowed Ali to rarely get hit dead-on.
Boxing is a sport of hitting without getting hit. Heavyweight champs must learn not only to throw a mean punch but also to withstand the huge onslaught of blows they will receive throughout their career. A good boxer knows he will get hit - he just doesn't want to get hit flush in the face. It's much easier to respond to a series of small hits than deal with the consequences of one hard, direct punch. And when does a boxer get hit flush? When they don't see the punch coming.
A similar concept exists in investing. Many negative events come and go and will have limited long-term impact
on the portfolio. Seasoned portfolio managers know they cannot avoid all negative events. It is the big ones that they want to miss. In the world of investing, getting hit flush in the face is equivalent to permanently impairing a large amount of capital.
One of the main ways we avoid permanently impairing capital is by emulating the heavyweight boxer. The market continuously provides information on events as they are unfolding. If an investor perceives a large, negative event headed her way, it is prudent to shift slightly to avoid full impact. Even if the event never actually realizes, a slight shift is better than being hit flush. Likewise, an investor will also move slightly towards what she perceives as opportunities. The goal is to eliminate blind spots and leave the portfolio resilient. In the last year, our team identified many opportunities and threats to the portfolio. These included Chinese growth, central bank intervention, European banks, deleveraging, Canadian commodity exposure, and a potential oversupply of heavy oil in North America due to breakthroughs in drilling and completion techniques. While these events were often labeled as risks and consequently developed a negative connotation, none of them are inherently good or bad. They merely represent a set of outcomes that investors may or may not want their portfolios to be exposed to at certain levels. Rather than try to make predictions about how the world will unfold, our team labours to identify opportunities and threats, ascertains our exposures to them, and decides whether we believe this is an acceptable level of exposure.
As an example, consider the potential impact from the development of oil fracking technologies. Over the last ten years, the North-American natural gas market has been transformed by technologies that can cheaply unlock the commodity from shale and other so-called tight formations. Technology has enabled huge amounts of natural gas to be unlocked in plays such as Marcellus and Haynesville, creating a glut of natural gas and driving prices down to $3 per MMCF and lower. Now, it looks like the same technology might create a similar scenario in North American oil markets, so much so the IEA is forecasting the U.S. could overtake Saudi Arabia as the world's largest producer of oil by 2017.
Oil from shale is therefore an outcome that must be managed. But is it positive or negative? That depends on how the earnings in our portfolio would be impacted. A glut of oil would likely be negative for Canadian heavy oil producers but positive for U.S. manufacturers who could benefit from cheaper energy. The choice here is, therefore, how to be exposed. In this particular case, our team has chosen to reduce our exposure to heavy oil producers and modestly increase our exposure to U.S. manufacturers. In a similar vein, owning Chevron creates some potential negative exposure to oil shale technology but positive exposure in the event of war in the Middle East.
As another example, let's consider the potential for rising interest rates. The way we balance this risk in our portfolios is through what may seem like inherent contradictions: some of our companies benefit from rising rates and some benefit from a low rate environment. Among our Canadian companies, we own both Manulife and First Capital. As a life insurer, Manulife's portfolio of investments and ability to meet liabilities benefit from rising rates. As a real-estate company, First Capital benefits from a low rate environment. Because we own them both, we expect to at least partially benefit regardless of the scenario that unfolds for Canadian rates. This kind of flexibility is an important part of building resilience.
Again, the effort being made here does not focus on predictions. The effort here is to balance the potential events to which the portfolio is exposed, as we never want to be caught flat footed. It is best to avoid the kind of events that can knock you out.
A Final NOTE ON THE MAYANS
At one point in Douglas Adams' Hitchhiker's Guide to the Galaxy, the author notes:
"the major difference between a thing that might go wrong
and a thing that cannot possibly go wrong
is that when a thing that cannot possibly go wrong goes wrong
it usually turns out to be impossible to get at or repair."
If only this view was more broadly accepted in society, we suspect that uncertainty would be much better managed.
Prediction, despite its faults, is likely here to stay. And not just in the few areas where we are good at it, but in the many areas where we are not, like investing. More than ten thousand years of evolutionary programming is unlikely to be erased at will. Awareness of this bias, however, is probably a good starting point for building systems that can be resilient over time. This is true both for the investment management industry and for society at large.
Yet awareness alone is not enough. We must accept and even embrace the role of uncertainty in our lives. There is nothing inherently evil about it, even if it has the tendency to rattle us as humans. Doing so is probably one of the best ways for our skills as decision-makers in complex systems to get better over time. As for our own investment process at Mawer, it is an ongoing effort to calm our inner Mayan. Awareness and systematic processes, such as our investment philosophy and valuation models, are some of the main ways we keep ourselves in check. Increasingly, we are moving away from macroeconomic outlooks and focusing more on managing portfolios under uncertainty.
2012 was a year that was dominated by macroeconomic uncertainty. Over the year, we laboured to make our clients' portfolios as resilient as possible, using all of the methods that we mentioned above. We continue to believe that the challenge of investing is won by not losing. While we cannot claim to know the specific path that will unfold in the next twelve months, we continue to build resilient portfolios for our clients and to provide them with peace of mind.