Bill Miller, CFA
Chairman, Chief Investment Officer, and Portfolio Manager, Legg Mason Capital Management

Michael Mauboussin
Senior Vice President, Chief Investment Strategist, Legg Mason Capital Management

"Thought Leader Forum Wrap-Up"

This is a wrap-up. There are many different takeaways. I'd like to share four specific ideas that I hope everybody drew from the last day or so.

The first idea is the role of strategy in investing. The average turnover for a mutual fund is something close to 100%. That turnover ratio is much higher for many hedge funds that are active in the markets. So if you're looking at a 12-month or 6-month time horizon, you typically get two or three data points-a new product introduction or earnings report. Also, quantitative funds, which have become more prominent over recent years often use statistical arbitrage techniques. Strategy is not particularly important for those kinds of investors, and understandably so. In contrast, if you're a long-term-oriented investor, strategy is really crucial to what you do. What we try to do is buy businesses and understand what the company and industry will look like in the future. A typical time horizon for us would be 3 to 5 years. Strategy is absolutely crucial to thinking that out properly, and-to put it more bluntly-thinking about which companies will win and which will lose in the marketplace.

While the disruptive innovation framework is fantastic and is one that we do use, it's one of a number of different strategy frameworks we put to work. We talk a lot about profit pools, five forces and value chains and so on. It's an important part of a broader toolbox that we use to try to understand problems as robustly as we can. So, strategy is really crucial for a long-term-oriented investor but not necessarily for all investors.

The second big idea from my perspective is that innovation has patterns. The disruptive innovation framework can help us recognize those patterns. One of them that we heard about is whether an innovation is sustaining or disruptive. Am I trying to cram a product into the blue space where I'm unlikely to succeed? Am I conceiving the offerings as jobs or am I looking at categories that may or may not give me proper context? Disintegration: is the industry more likely to be vertically integrated or horizontally integrated as modularization takes hold. These are all very robust ways of thinking about real business problems. They don't give us any certainty of being correct, but they increase the probability that we can properly anticipate change.

The third idea has to do with something that Clay didn't talk about but has written about and has been deeply influential to me. It's circumstance versus attribute-based theory. Theory is something that tries to link cause and effect. Many business people don't like the word "theory" because they associate it with "theoretical," or "airy," or "not the real world." But we all operate with theories in our heads, whether we acknowledge it or not. In some of his papers, Clay has laid out some robust thoughts about how theory proceeds. A key insight from that work is that as theory improves, as it becomes more robust, it goes from being more attribute-based to being more circumstance-based. This is a big deal in the investment world.

For example, twenty-five years ago a bunch of researchers found that low P/E stocks deliver higher returns than the current model predicted. That led to a simple attribute-based theory: if you buy low P/E stocks, you'll outperform the market. As you know, if you ask anyone in the market whether a low P/E is good or bad, they have to answer, "it depends." It depends on the circumstances surrounding it. The same goes for a high P/E. So, looking at the world from a simple, attribute-based framework can lead you down many empty paths. The other example from today is the notion of whether a company should outsource or not. Thinking through the notions of modularization-how you go from proprietary to modularized markets and what that means for wealth transfers-is a very robust, circumstance-based theory that can help you gain some important insights.

All the data we have is backwards-looking. What we're trying to do is anticipate the future. If you're doing that with yesterday's data and an attribute-based perspective, that approach can very often lead you to the wrong area.

The fourth and final point I would make is something we should never lose sight of as long-term investors. We're all ultimately trying to generate excess returns for our portfolio. That's what it's about. There are two parts to that equation. We mostly talked about one side of that equation, not both. The first side of the equation is about understanding fundamentals. How will a company perform based on its financials: return on capital, growth prospects, the competitive landscape and so on. As Steven Crist said last night, that concerns how fast the horse can run. Then he said that a properly-trained handicapper can do a decent job of predicting the race and rank of the horses. Those tools are out there and we can understand that stuff. But what gets you paid is not understanding the fundamentals, but the relationship between fundamentals and expectations. One of the greatest errors in the investment business is the failure to properly understand this distinction. When things are going well and fundamentals are good, everyone wants to buy. That's the environment we see today. When things are going poorly, everyone wants to sell. One of the disciplines we iterate over and over is this distinguishing between fundamentals and expectations.

Those would be my four thoughts. Now I'll turn it over to Bill.

Bill: Just to touch on that last point. I was at a Santa Fe Institute meeting in DC a few weeks ago and heard a talk on creative genius by Dean Simonton-not just ordinary creativity, but geniuses like Darwin and Einstein. The question is whether there are attributes or an attribute-circumstance model that leads to this kind of genius. He drew a distinction at the outset between expertise and genius. Expertise requires a narrowing: you study more and get more focused. That way you get a skill set that allows you to solve some kind of task. He said that the creative geniuses are very different than that. They typically have expertise but they also cast very wide nets looking for patterns where other people don't see patterns. It's like the entrepreneur who sees things in his or her mind that are different from what other people are seeing. That's part of why we do these conferences and part of the great benefit of the books we read outside of our field. I noticed that Doug Erwin is here. Doug is one of the leading experts on extinction and one of the great scientists of his generation. Doug was talking with Clay at one of the breaks and said that there are a lot of biological analogs between disruptive innovation and what you see in mass extinction events. Doug has also done some amazing work on innovation and evolution. Doug, do you want to make a remark about the analogs between business and biological evolution?

Doug: I won't say anything about business, about which I know nothing, but I was struck by the similarity of patterns that we see in ancient life. When Clay talked about sustaining innovations, it reminded me of a topic that we've known about for twenty years or so in the fossil record. It's the principle of "incumbency". It's very hard to displace existing organisms. Yet, mass extinctions and other biodiversity crises displace existing organisms rather effectively. One of the things I do is look at what happens afterwards. It's a process called niche construction and is analogous to building businesses or new products, I suppose. When Clay was talking about the pattern of moving out into the green space in business, that's similar to the process of creating a new niche and new eco-space in a biological sense. We're trying to understand how that happens in biology. There are some interesting similarities between what you're seeing in business and what we're looking at in evolution.

Bill: Thanks, Doug. Clay also mentioned the ability to envision things being different outside of our own models. I remember talking to Meg Whitman [CEO, eBay] just after they bought PayPal and they paid a billion-and-a-half in stock. That was about fifteen percent of eBay's total market cap. I said to her, "do you really believe that this company is worth 15 percent of the entire future of eBay?" She said, "Yes," and as it turned out, a billion-and-a-half dollars was a bargain for PayPal. I was also struck by John's comment that nobody inside of eBay really questions Skype. They think it's doing great. As outsiders, we can't tell that kind of stuff the way they can. Certainly if you look at some of the patterns, they're consistent with Skype becoming something far more than the market is currently valuing.

Clay made a comment about the S-curve and plotting it on a log scale. One of the things you tend to see in complex systems are power-law distributions. In capital markets people mistakenly talk about fat tails of the distributions as though that adequately captures the nature of the distributions. We've observed that market shares in Internet businesses follow a power-law distribution. That means there is a fixed proportional relationship between the market share of a company and its rank as number one, two, three and so on. There's a lot of discussion about why that's the case yet no one has a robust theory about it. But empirically it fits the facts and as long as these relationships are stable, this provides a real interesting way to think about the possibilities.

The last thing has to do with eBay's mission statement about making inefficient markets efficient. We can learn a lot about companies by how they describe themselves. In part, that description is a shorthand version of a broader theoretical approach of how they see their business. If eBay's mission is to "make inefficient markets efficient," and Amazon's is to "enable customers to find and buy anything they want online," and Google's is to "organize the world's information," these mission statements can tell you a lot about each company's respective strategy. It would be interesting to ask John whether he thinks systematically about where the big, inefficient markets are. Clearly telecom is one, hence their purchase of Skype. But there are a couple of other monster markets out there like health care. Are there ways to disrupt that? We heard about one of those ways today. Education is another huge, inefficient market globally.

Michael: With that we'll open it up to any questions.

Question: From an investment standpoint, when I think about a stock tripling or doubling in a relatively short period of time, would you view that as a mistake in the sense that analysts didn't foresee it? Corporations tend to move so slowly, so when I see large companies double or triple their stock price, how do you view that?

Michael: I think we would say that it's unusual for fundamental values to change dramatically in so short a period of time. This is similar to a topic that we had for the forum several years ago. Why are markets not efficient all the time, reflecting the fundamentals? We know that they reflect the fundamentals most of the time, but occasionally, for many psychological reasons, people tend to get uniformly optimistic or uniformly pessimistic about a particular industry or company. This creates a gap between fundamentals and expectations. As the fundamentals continue to grow, the expectations can send the price into an overshoot situation. Part of the answer is a little bit of what Steve Crist talked about last night-the psychological reasons and behavioral reasons that values and prices deviate. Sometimes it's also the S-curve, like Clay talked about. There are companies that grow at an exponential rate when they're starting out.

Bill: Clearly, when that degree of change happens, it's a question of a change in expectations. So, why do they happen so rapidly? I think that part of the answer has to do with probabilities and that the cone of uncertainty widens as time goes on. If I say to you, "what are the chances that IBM will be bankrupt tomorrow morning," the answer is, "virtually zero." How about next year? Five years? One hundred years? Clearly, you'd attach a higher probability to the one-hundred-year time span than the five-year time span. The data comes in every day, and that can change our view very quickly about the various paths that a company can be on. If a company beats projections four quarters in a row, people might say, "this company can grow faster and longer than I thought." That will generate a much higher value. Alternatively, these things are iterative. If oil was $10 and now it's $80, there's a combination of observing the price change, trying to figure out what explains the price change, and the longer the phenomenon persists, the greater your confidence that your explanation is consistent with what you're seeing. That allows you to extrapolate more, and that creates iterative effects.

Question: I'd be interested in your thoughts on the following. Enron was named the most innovative company at least five years in a row by Fortune magazine and during its heyday was considered by many observers of business and finance as a very innovative company. Yet, ultimately, the company collapsed pretty spectacularly. Innovation actually did take place in the company. There was almost a focus on innovation at the expense of thinking about some of the more practical realities of running the business. I'd be curious about your comments on Enron as a case study of an innovative company that didn't work out.

Clay: I personally have stopped using the unmodified word "innovation" in my vocabulary because it's a very broad word that doesn't lead me to any accurate predictions at all. I do use the phrases "disruptive innovation" and "sustaining innovation" because modifying it in those ways gives me more predictive power. One of our former faculty members, Phil Rosenzweig, recently wrote a great book called The Halo Effect. It's quite new. The LMCM team is inviting Phil in to discuss the book on October 9th. He points out a broad tendency amongst business observers to say, "this company was successful, let's look at what they did and then assume that if everyone else in the world did those things, they would be successful, too." The very fact that you were successful gives you a halo that shines favorable light on all the things you did.

I love Jim Collins. He's written some very interesting books. But Phil is just devastatingly critical of the design of the research that Jim Collins used in the book Good to Great. I'm not saying this to be critical of Jim, but to be critical of the sorts of awards that would be given to an Enron that had a halo around its head.

Collins' research design started with companies whose stock prices had been pretty stagnant. Then he found eleven whose stock prices went up sharply. Then he found a comparable company in each of those industries whose stock price went flat. The former went "good to great" while the latter went "good to good." Then he looked for the characteristics that distinguished the good-to-great companies from the other ones. His book then encouraged everyone else to mimic what the good-to-great companies did in order to be successful.

One of the characteristics of the good-to-great companies was they were "hedgehogs." They burrow down and they know what they're good at. If you look at those eleven companies, they all were hedgehogs. Some of them that didn't go good-to-great were actually a little less focused than the other ones. That raises the question, "should every company always be a hedgehog?" In fact, there were roughly 200,000 other companies out there that didn't go from good to great, and a lot of them were hedgehogs, too. The research design observed a correlation within a small sample, but can it be trusted as a law of nature?

We all ought to be concerned when we read that a company was successful and everyone ought to do it their way. We should say, "no, these guys were successful, but what was it about the situation they were in that caused them to be successful." Then, is there any suggestion that they might be moving into another situation where the formula that worked before won't work anymore. I think that's a much more robust way of thinking about things.

Michael: I have just a pure financial statement analysis commentary. When you model a company, the return on incremental invested capital equals the maximum supportable growth rate of the company excluding external financing. The ROIIC represents the fastest a company can grow without tapping external capital. Based on Enron's public financial statements, their ROIIC was in the high single digits, roughly equal to their cost of capital, yet they were growing at greater than 20%. Mathematically, you knew that they had access to a lot of outside capital. They did that through straight debt, equity and the off-balance-sheet stuff. But sooner or later, one of two things has to happen. Either returns have to go up-which often happens as the business matures-or the growth has to come down. It turns out I made a presentation to the board and senior management in 1999 and the whole theme of the talk was this precise issue. They recognized at the time that they needed to get their returns up. My own take on it is that a number of their core operations were very innovative businesses and they ran into a pretty old-fashioned liquidity problem. The trading operation lost access to capital and folded up pretty quickly. They also had a lot of start-up and ancillary businesses in which there was some monkey business going on where the economic profiles were questionable.

Bill: Just a slight addition to that. Enron had investment-grade ratings. We did very good research on Enron, played off all the hard assets against the debt and figured the non-balance sheet stuff had zero equity value. It came down to the trading operation. As long as it had access to capital, it was fine. When the rating agencies pulled their investment-grade ratings, that was what ended Enron. You might say that it should have died, but there's a similar situation over in England with Northern Rock, a company that did everything right except having access to funding.

The Halo Effect is a very illuminating book. Its core message is that when we look at successful or unsuccessful companies and extract the attributes that appear to explain what we're observing, they may actually have no relation to what we're observing. A weakness of all that kind of attribute-based literature is that it makes a post-act abstraction but it doesn't make any predictions. They should be able to look at other companies out there that are currently good and predict that they will become great at some point. Clay actually is doing this-taking real money and making predictions based on his model. So far, so good!

Question: I know you have been invested in Level 3 for a long time. Jim Crowe [CEO, Level 3] has been talking about disruptive innovations for ten years in telecommunications and how the market was going to disaggregate and go to modular systems. It seems like that company is now diluting itself with stock while regional bells have built nationwide networks by combining. I would love to hear your dissection of Level 3 in relation to what we've been talking about all day.

Bill: Level 3 is evolving in line with Jim's vision when he put the company together. It's been through the telecom boom and bust. The core of it is the horizontalization and modularization of the business, as verticalization doesn't make any long-term economic sense. Specializing in their layer will require high capital costs up front. The marginal cost of moving stuff through is minimal. As Internet traffic grows and the use of bandwidth accelerates with the advent of services like YouTube, eventually prices start to rise and they make vast amounts of money. It's a long time coming, but it's still on track.

Answer: The only thing I would add is that you've seen a lot of this horizontalization in telecom but you're also seeing a push to reverticalization by AT&T and Verizon, based on nothing more than a strong management preference. For them, verticalization doesn't necessarily carry any economic benefits but it's just how they do things. They're willing to trade off ROI to do that. That's interesting in thinking through the psychology of what's happening. That adds time to the whole process or increases the dynamic nature of it. About 60% of Internet traffic right now is video, typically YouTube. A lot of the economic models there are still unfinished. It's hard to believe that we would end up with no economic model. But at the same time, until there is a model, it's hard to assess where the charges down through the chain will end up. Theory is incredibly valuable in a dynamic situation where preferences are changing or it's unclear how things will manifest.

Bill: Lots of companies that are at Level 3's stage of development have multiple evolutionary pathways to follow. Being dogmatic about how you think things are going to develop is much less effective than considering the various pathways a company could follow.

Question: We heard a lot about the younger generation at Red Hat and how they worked. From a Legg Mason standpoint of hiring new, young investment talent, have you observed differences in how you have to reach the new generation?

Answer: There is one process that our company is based on and that we try to adhere to, no matter what role one is playing. That process is not dependent on your age or gender or other preferences. While people certainly work in different ways-the younger people in research are much more likely to be online-the fundamental output is standard.

Bill: We're looking for cognitive diversity among people that we hire in order to get diverse perspectives and views. But we have a single, unifying mission to try to help clients earn excess returns. I was up at Fidelity a week or so ago talking to a group of 200 analysts. Someone asked me whether we would hire someone at Legg Mason who had their own way of doing things that was different from our model-would we hire someone who was maybe just technically-oriented towards managing mutual funds? That's not the style and culture that we have. We have a distinctive way of approaching what we do. But if we did bring in someone who used a much different model, then we'd have to reorganize research. Fidelity has centralized research and decentralized portfolio managers, each of whom has his or her own way of doing things. The only thing that holds them together is if they outperformed, they're good, and if they underperformed, they're bad. With that approach, you'd need modularized research that would be oriented towards the style of the individual.

Question: If you identify that there is disruptive innovation going on, it seems to me that it's a very hard task to put a price on that and then get a sense of whether the company is overvalued or undervalued.

Michael: Warren Buffett and Charlie Munger repeat this quotation from Ben Graham, "if it's close, we don't play." So, look at the world probabilistically and try to entertain various scenarios. Ideally, if you spot a disruptive innovation, you don't want to pay much for it. Think about possible outcomes, what reasonable probabilities there are, and then look for things that are grossly mispriced on that basis. That's the idea of central tendency of value where we use probabilities, outcomes and other valuation metrics (price to sales, price to book, metrics on a per customer basis). You're looking for situations where there are substantial mispricings.

Bill: Look at Amazon a year or so ago. That's a case where they had declining margins for a couple of years, rising spending, and the stock did very poorly. The company was very clear that that period of investment would come to an end, that they were investing in a wide variety of new services and options to enhance the customer experience. You could see the skepticism about the margins they could obtain. It wasn't really hard to tell what they could obtain because the margins were running about half what they'd done a couple of years ago. All they had to do was stop spending and it would revert to that number. They were doing a lot of things that were potentially disruptive but the market didn't care and wasn't paying for it because the market was focused on the shorter term. Now that some of those things have come to fruition, they've got better margins and accelerating sales growth, and people are all excited about the services they've started, which are going to be huge in a couple of years. Now it's much more difficult. At $90, the expectations are much higher and the intellectual exercise becomes more difficult-the models will give you a vast range. Here's how I look at it. Amazon has a $38Bn market cap right now. If you look at the spaces they're occupying, the sizes of the markets globally, the competitive dynamics, the S-curve, the option value of web services, and so on, you can easily see that in ten years, this company could have a $200Bn market value. Yet the market will go up about 8% per year. The trade-off there is much harder. Do I want to take the risk of selling Amazon at $90 because next quarter it could be $70 or $60 and then lose the position, or do I want to sit tight with it and figure that I get a high probability of beating the market dramatically over a ten-year period? It makes it challenging.

Question: Any good books?

Michael: This year we prepared a list. There's always a lot of good stuff. I will mention two book club books for 2007. One is The Halo Effect, mentioned already. The second is Scott Page's book, The Difference. Scott is a good friend of the firm. He's at the University of Michigan and on the SFI external faculty. The book is a very rigorous treatment of the issue of diversity and the importance of cognitive and value diversity.

A couple of popular books include The Black Swan by Nassim Nicholas Taleb. Super Crunchers recently came out by Ian Ayres. It takes the ideas similar to those in Moneyball and applies them much more broadly-that is, the use of statistical methods. It's a little like the medicine example we talked about earlier today. Where cause and effect are clear and that relationship can be detected through statistical methods, that's very helpful. Steven Pinker has a book called The Stuff of Thought, which I am really enjoying. He's a psychologist at Harvard but the book is an analysis of how language gives us a window into cognition. The words we choose to use reflect how our minds are organized.

There's a book called The Art of Learning by Josh Waitzkin. You may have heard of this guy. He was a chess prodigy featured in the movie, Searching for Bobby Fischer. He burned out on chess and became a world champion Tai Chi master. He achieved excellence in two totally different domains. The book is about how to learn. He said that when someone makes a mistake in chess or Tai Chi, rather than recovering their composure, they get a little flustered, which leads to another little mistake, which gets them even more flustered and they end up spinning out of control. That happens a bit in the investment business, too. People have a stretch of bad performance that gets them agitated and they start making other decisions that are not so good. That leads to a negative snowball.

The Lucifer Effect is another book I'll mention. It's by Philip Zimbardo. Phil is a famous social psychologist from Stanford and conducted one of the most famous social psychology experiments ever, called the Stanford Prison Experiment. They took twenty-four healthy young men from the Palo Alto area and put half of them into a constructed prison as prisoners and half as guards. They became psychotic over a very short period of time and they had to call the whole thing off. This is the first full account that he gives of what happened during that experiment. He gives some general principles as well. One of the most underestimated dimensions of organizations is the role of social context in decision making. Proper social context can be conducive to good decision making while negative social context (people around you stressed out or overly negative) can be very deleterious to decision making. He talks a lot about Abu Ghraib and what happened in 2004.

The last book I'll mention is a bit of a sleeper. This summer, Bill Walsh, the legendary coach of the San Francisco 49ers passed away. I was reading a little about him and found out that he wrote a book which tells you everything that you need to run a great organization. The book is called Finding the Winning Edge. I went to Amazon and the cheapest version was $160. I bought the book. It goes through every aspect of how to run an organization from the meetings to speeches to what he expected from the secretaries. A lot of it was specific to football but many principles apply in a broader context as well.

Bill: I'll mention a few. The first is A Demon of Our Own Design: Markets, Hedge Funds, and the Perils of Financial Innovation by Rick Bookstaber. He will be speaking at our SFI meeting on October 18th. This book came out about six months ago and laid out a template for exactly what's happening now in the credit markets. These things always follow a similar pattern. It's very instructive and written in an engaging style. He has been present at every one of these collapses beginning back in 1987. He was at Salomon Brothers during the Long Term Capital Management collapse. He said that everyone there was brilliant but when the markets went haywire they would just go to Salomon and ask for more capital. Sometimes Salomon was reluctant to but every time they got more capital. If you're doing directional trading, even standard statistical arbitrage, your risk control measures are designed to keep you from being hurt and having your problem compound. If you're doing convergence trades, then the risk procedures are different and you want to throw more money at it.

They went on their own and did terrific at Long Term Capital Management from 1994 to 1998. They kept giving their investors money back and in 1998, they sent a letter out to their investors offering them a once in a lifetime opportunity to make a huge amount of money by putting more in. Not one investor put more money in and that's what doomed them.

There's a new book out called Mistakes Were Made (But Not by Me), which is written by Carol Tavris and Elliot Aronson. It goes through case studies and experiments that show how people will reinvent the past to make themselves right and so they can maintain their own self-image of being right. It's all about how to recognize error and determine when you've made a mistake because there are very powerful psychological forces at work that cause you to not admit error.

One that's a little more out of the way is The Myth of the Rational Voter: Why Democracies Choose Bad Policies by Bryan Caplan. It attacks democracy on theoretical grounds. Democracy is a sort of wisdom of crowds applied to politics. He says in the beginning of the book that the theory of democracy involves a theory of diversity where errors cancel each other out. But just in economics, if there are systematic behavior patterns, they can lead to very bad outcomes. In democracy, there are systematic behavior patterns that do lead to bad outcomes. One of them is free trade. Whether you're a liberal or conservative economist, everyone agrees that free trade is good. However, few average people believe in free trade. They don't support it. They talk about how some country is taking advantage of us. Even if you look at free trade pacts, they're not really so. They're agreements that say, "if you buy more of my exports, I'll buy more of your exports." This leads to suboptimal policies when iterated over a large variety of instances.

Michael: I'll mention one other title for those folks who liked last year's topic on neuroscience. Jason Zweig has a new book out called Your Money and Your Brain. It's probably the best synopsis of the three fields of psychology, neuroscience and economics.

Question: What's something that you had a strong view about last year that you've changed your mind about since?

Bill: I don't have strong views about most anything-they tend to be held rather tenuously. Something that I've clearly been wrong about and baffled by is the oil price, especially in the context of other commodity prices. It's completely baffling to me and everyone in the oil business and the trading business. There's something going on that's different and no one can quite figure it out.

Michael: Again, thanks to all of you for taking time from your busy schedules to join us for the last day or so.

Legg Mason Capital Management ("LMCM") is comprised of (i) Legg Mason Capital Management, Inc., and (ii) LMM LLC.

The comments, opinions and any forward predictions presented about any particular security, the economy or "the market" are based on the analysis of the speaker. These are not necessarily the opinion of, and should not be construed as a recommendation on the part of Legg Mason Capital Management or any of its affiliates.