Richard Feynman on Knowing vs. Understanding

Richard Feynman was one of the most brilliant thinkers of the 20th century. His specialty was physics, but he had an amazing ability to clarify a problem, cut through unnecessary detail, and explain complex concepts. His techniques for problem-solving are valuable to anyone, in any field.

 

This is a short clip of Feynman explaining how we should think about competing theories, and why understanding the philosophical basis for a problem is more important than having a black box model that can spit out an answer. The parallels to investing couldn’t be clearer.

To anyone interested in learning more about Feynman, I strongly suggest the autobiographical book Surely You’re Joking, Mister Feynman as well as Genius by James Gleik.

Bad Incentives and Portfolio Managers as Salesmen

This is part one in a series about how active management could be better. This is my perspective alone. Of course, the dynamics I describe do not apply to all managers. In a prior piece, I explained how active managers could theoretically outperform their benchmarks; now we arrive at the question: why don’t they?

Part One: Bad Incentives and Portfolio Managers as Salesmen

Incentives are powerful. They unconsciously motivate everything we do. Even with the best intentions, the wrong incentives will keep pushing us to do the wrong thing. “I think I’ve been in the top 5% of my age cohort all my life in understanding the power of incentives, and all my life I’ve underestimated it,” says Charlie Munger.

Across vast swaths of the money management business, the incentives are broken, causing a vicious cycle that erodes returns. I will describe how this dynamic plays out.

There is a giant pool of portfolio managers (PMs) trying to build a track record. In such a large field of PMs, some will have periods of strong performance. Some perform well due to talent, others due to luck.

For a PM who shows good short-term performance, the incentive is to sell based on that performance. It’s hard to resist because the PM gets most of his income from a percentage fee on assets. The more the PM wants that new vacation house, the less he cares about his strategy’s fit for prospective clients. The prime directive is to grow AUM and ask questions later. He might employ a sales team whose bonus is tied to short-term inflows. Because the sales team has no long-term incentive, they will be eager to please. With these incentives, every prospective client seems like a perfect fit for the strategy.

Second prize is a set of steak knives.
Due to the laws of probability, the strategy does not outperform forever. Now that the PM has built a sizable AUM, his incentives shift to retaining assets. When performance lags, there is enormous pressure from clients to defend every quarterly return. The PM may say “some periods of underperformance should be expected in this strategy,” but that’s not how he pitched it when times were good. Clients are upset. The PM becomes myopic, short-term focused. There is pressure to hug the benchmark. It becomes harder and harder to stick to the strategy that gave him success in the first place.

It doesn’t help that many large clients want active managers to hug the benchmark. Many prefer closet indexers. Many consultants rank PMs using “optimization formulas” that heavily penalize managers for deviating from their benchmark. In my opinion, this is an area where the industry has lost its mind. If you don’t want deviation from the benchmark, don’t hire active managers—buy an index fund. If you just want less deviation from the benchmark, allocate less to active management!

Most equity managers don’t believe they can outperform every single year, much less every month. Not everyone is Renaissance Technologies. Most PMs have long-term investment strategies; they expect to have good years and bad years.

But misaligned incentives create a bad dynamic: when performance is good, the incentive is to sell fast, grow AUM, and ask questions later. This creates a poorly-fitted client base of impatient investors with unrealistic expectations. When performance lags, PMs feel pressure to become defensive or hug the benchmark, forgetting what brought them success in the first place.

What can be done about the problem of incentives?

Underlying almost every problem with active management is an uncomfortable fact: there is still more of it than we need. Too many PMs try to grow AUM without making sure clients are a good fit. Too many clients hire active managers without understanding the strategy. They develop unrealistic expectations and don’t stick with the strategy long-term. For many investors, indexing is a better approach. Encouraging more investors to use index funds will improve active management.

The other problem affecting PMs and their clients is a culture of impatience and desire to get rich. For most of us, active management shouldn’t be about getting rich quickly; it should be about getting rich eventually. Active managers should charge lower base fees (fees based on a percentage of AUM). When possible, they should use long-term performance fees and invest more of their own personal savings in their strategies. Aligning these incentives places the PM’s primary focus on long-term returns, not month-to-month AUM.

Finally, active managers need to focus on setting the right expectations and building trust. Some PMs have been model citizens in this regard. Seth Klarman’s Baupost Group has gone so far as returning money to investors when it doesn’t see enough good investment opportunities. If clients understand an investment strategy and trust it for the long term, the manager won’t need to do backflips every time quarterly performance lags a bit.

Changing these incentives would certainly benefit clients, whose fortunes are on the line. For many PMs, this could mean less money. But the system would be fair—the best active managers could thrive just by implementing their strategy. If they skillfully invest a sizable chunk of their own money, they will earn great returns for themselves. With great returns, they can attract a well-fitted client base and build a solid foundation of trust. It may take a bit longer, but everyone will be better off in the end.

Pulling Teeth: The FANG Year and Financial Memes

After reading Michael Batnick’s excellent posts about the distribution of stock market returns (read The Skew and The Other Side), I was inspired to look back at the market of 2015.

Perhaps you’ve heard about 2015, the fabled year when the four “FANG” stocks (Facebook, Amazon, Netflix, and Google) accounted for all of the gains (!) of the S&P 500. Many articles and blog posts were written about this phenomenon in 2015, and the “FANG year” became something of a meme in financial discourse. The S&P 500 returned only +1.4%, but the four FANG stocks returned +34%, +118%, +134%, and +46%, respectively. They accounted for the all of the gains (!) of the S&P. Crazy, right?

Will Deener of the Dallas Morning News wrote, “One of the hallmarks of a healthy bull market is that a broad swath of stocks and sectors move higher in tandem. Currently, that is not the case, with both the S&P 500 and Nasdaq being propped up by only four stocks.”

Only four stocks!

Many people surmised that investors who missed out on the FANG stocks must have had a dismal year. If you picked a random stock from the S&P 500, you would almost certainly not pick a FANG stock, so you must have had a terrible year, right?

Let’s look at the numbers. If you had thrown darts at a board to pick one random stock out of the S&P 500 and held it for the entire year, what were your chances of picking one that outperformed the index?

47.19%.

Yes, 47% of stocks in the S&P 500 outperformed the index that year. So slightly less than half. But still, almost half. You would have had a 47% chance to outperform the S&P 500 just by picking any stock at random and holding it for the year.

What were your chances of selecting a stock that had a positive return?

50.37%. You would have had better-than-even odds of a positive return by picking any stock at random that year.

Let’s look at the distribution of returns in the S&P 500 in a histogram:


The above chart shows the number of stocks that had a 2015 total return at various minimum ranges. The best performing stock (Netflix) returned +134.38%, and the worst (Chesapeake Energy) returned -76.76%. How did the rest do?

  • 50% of stocks returned at least +0.13%.
  • 40% of stocks returned at least +3.83%.
  • 30% of stocks returned at least +10.09%.

So in the infamous “FANG year,” when supposedly all of the gains (!) in the S&P 500 went to just four stocks, 30% of stocks in the index actually returned over +10%.

So what’s going on here? Why was this meme so misleading?

The S&P 500 is market cap-weighted (companies are weighted by their market value). A stock’s weight times its return equals its contribution. Bigger companies have a larger weight, so they can have a larger impact on the index’s return. In 2015, some of the best-performing stocks also happened to have some of the heaviest weights in the index. So it’s true that if you owned the S&P 500 index except for those four stocks, your +1.4% gain would be gone. These four stocks had a combined contribution of almost +2%.

In any given year, the index includes positive contributors and negative contributors. If you took all the stocks with gains in 2015, their combined contribution was +8.9%. This is the gross contribution, or what most people would consider all of the gains. If you took all of the stocks with losses, their combined contribution was -7.4%. This is all of the losses.

When you put the gains and losses together, it was a fairly flat year for the market. The net of all gains in the S&P 500 was just +1.4%. But wait, commentators said, the contribution of the four FANG stocks was more than +1.4%. So voilà, just those four stocks accounted for all of the gains (!).

Naturally, when gains and losses net close to zero it becomes easy to pick out a small group of stocks whose contributions add up to the net market return. What’s more, the FANG stocks weren’t even the top four contributors that year. Amazon was number one, but Microsoft was number two and G.E. was number four. Saying only the FANG stocks mattered in 2015 is like saying the only reason a football team won 28-21 is one particular touchdown. Every point counts.

All this FANG talk gave many people the false impression that only four stocks had any gains at all that year. But there were plenty of other stocks with huge returns that year. NVIDIA returned +67%. Hormel foods returned +54%. Starbucks returned +48%. Kroger returned +32%.

Nevertheless, the “FANG year” meme had spread across financial media like World War Z. ‪It’s a fun idea, but not all that meaningful.

Netflix, Precision, and Stories

We think in terms of stories and analogies. We were born to do it. In SapiensYuval Noah Harari argues that our penchant for telling and believing in stories is the “mysterious glue” that allowed humans to cooperate and dominate the world.

While stories may encourage cooperation, they can sometimes blind us to reason. Daniel Kahneman illustrates how stories can breed fallacies in Thinking: Fast and Slow.

He describes an experiment involving a fictional woman named Linda. In the experiment, Kahneman and his partner Amos Tversky described Linda as a young, single, outspoken, brilliant philosophy major “concerned with issues of discrimination and social justice, [who] also participated in antinuclear demonstrations.”

They then asked people which was more probable:

(1) “Linda is a bank teller.”

(2) “Linda is a bank teller and active in the feminist movement.”

They found that “85% to 90% of undergraduates at several major universities chose the second option, contrary to logic.”

We can be so willing to take a narrative, form a mental picture, and categorize things that we set aside reason.

The same thing can happen when discussing businesses, and few companies are as well discussed as Netflix. In a recent appearance, CNBC guest David Trainer argues that Netflix’s valuation is “disconnected from fundamentals.” He claims that that investors “seem to be unbelievably gullible these days,” and that Netflix is a “story stock.” Trainer mentions profit margins and multiples to suggest that Netflix is hugely overvalued, and explains that investors are falling victim to a false narrative.

But which is the false narrative? Although Trainer derides Netflix as a “story stock,” isn’t he constructing his own story here? He seems to apply a standard valuation framework, compares Netflix to other content creators like Disney, and posits that it’s “a tough business to be in”. He asks, “can you name any businesses in the history of the world that have consistently [sic] created profitable new content?” The framing of that question sounds suspiciously close to a story.

The problem with this way of thinking is that some businesses have predictable economics and some have highly uncertain economics.

In an electric utility, for example, we can be somewhat confident in forecasting future cash flows. We might look at population trends, trends in energy consumption, and the utility’s cost per watt. The industry is highly regulated, so utilities have limited ability to increase margins. Given those constraints, if we try to place a value on XYZ Utility, maybe we’ll say it’s worth $1 billion, or maybe it’s worth $1.2 billion, but it’s definitely not worth $10 billion. With this type of business, we could use multiples or industry comparables and feel pretty good about our estimate of value. It’s a fairly predictable story.

With a business like Netflix, the future is incredibly uncertain. We do not know what its programming content will cost in ten years or how much pricing power it will have. We don’t know what its innovations will be. Trying to use a quantitative approach like discounted cash flow or even comparables to value this type of business is fraught with danger.

So why might Trainer be wrong? Why might the market think Netflix is worth $60 billion?

Consider what the business is, who runs it, and what it could be in ten years. The company is building a massive stream of recurring revenue with sticky subscribers. It has an innovative team. We can see the trends in cable vs. streaming, and we can imagine a future in which Netflix could raise prices without losing many subscribers. A business with almost no need for physical capital has the potential to send additional revenue directly to the bottom line. Netflix might not live up to lofty expectations, but there is something real here. Maybe investors aren’t just being “gullible.”

Sometimes, using a valuation formula or comparables can be dangerous. To quote Seth Klarman, “Any attempt to value businesses with precision will yield values that are precisely inaccurate.” We should be skeptical of any stock that has a sky-high valuation, but we should also be skeptical of valuing a business by formula or by analogy. Valuing a business like Netflix is like looking through fog. If you’re looking through fog for land on the horizon, don’t use a magnifying glass.

Disclosure: I do not currently own shares in Netflix.

Stock Picking and the Zero-Sum Game

Most stock fund managers fail to outperform their benchmarks. This is a widely discussed and repeatedly corroborated phenomenon. There are studies showing abysmal rates of outperformance among active U.S. equity managers in recent years. Warren Buffett made a famous bet against hedge funds that looks better every day. The flow of funds from active to passive (index) investing is one of the most popular discussion topics in modern investing.

All this raises the question: is it even possible for actively managed funds, as a group, to outperform passive funds?

When index funds gained popularity in the ‘80s and ‘90s, many professional investors dismissed them, believing that any decent business school graduate could outperform the market. Now, some commentators dismiss outperformance as an impossible challenge, arguing that market “efficiency” has wiped out the possibility of achieving excess returns.

Amid this discussion, a well-worn maxim is the idea that stock picking—and therefore active management—is a zero-sum game. The idea holds that market participants, as a whole, cannot outperform the market, and therefore active participants, as a whole, cannot outperform passive participants. The zero-sum game concept is often touted as an argument for passive management or for the futility of active management.

By rephrasing an assertion from William Sharpe, we can define the zero-sum game concept of markets in the following way:

The Zero-Sum Game Rule of Markets: Before costs, for any distinct set of securities, the collective return of all active participants must equal the collective return of all passive participants.

This is simple enough logic. For a distinct set of securities, any excess return achieved by one active participant comes at the expense of another.

This logic can be misused in common discourse by suggesting that actively managed funds as a group cannot, by some immutable law of finance, outperform the market. Although there is ample evidence that most actively managed stock funds have performed poorly in recent years, the zero-sum game concept does not suffice as an explanation. It is theoretically possible for active funds, as a group, to outperform passive funds, as we will see below.

Actively-Held Stocks vs. Passively-Held Stocks: a True Zero-Sum Game 

Let’s imagine the universe of equity shares split into two buckets: passive and active. The passive bucket represents all shares held by those who can only hold stocks at the market weight. The active bucket represents all shares owned by those who can hold and trade shares at any other weight. The active participants determine the market weights of all equity shares, while the passive participants must accept the market weights. Combined, the two buckets represent all publically traded equity shares. For simplicity, we will exclude fees and assume that the passive investors can invest perfectly with zero tracking error. The two buckets look like this:

Given these assumptions, the passive bucket and active bucket will always have the same return. The returns of individual active participants may differ widely, but as a group, their stocks cannot outperform the stocks held by passive participants. This is a true zero-sum game.

But, as William Sharpe also points out, there are a number of ways that active equity funds could outperform the market.

Active Funds vs. Passively-Held Stocks 

Remember that the zero-sum game rule holds true for a distinct set of securities (i.e., the set of all stocks). But active equity funds are not bound by this set; they can choose to substitute some amount of cash, bonds, or other securities outside of the set held by the passive funds, as illustrated below:

In theory, one advantage held by active equity managers is that they have a broader investable universe than index funds. And because they can reach beyond the set of securities held by the market, they could collectively outperform it. In theory, they could hold more cash in falling markets and less cash during rising markets. They could short equity futures in falling markets or sell covered calls in flat markets. This is the crucial distinction: while the returns of active stocks must equal the returns of passive stocks, the returns of active funds do not necessarily equal the returns of passive funds.

Consider the following example in a hypothetical stock market. Imagine there is one passive fund, one active fund, and other active participants who trade stocks at a variety of weights. For the first half of the year, the stock market returns -5.00%. For the second half, the market returns +10.53%. To exclude the effect of stock-picking, imagine the active fund invests all of its stocks at market weights and only adjusts the size of its cash position.

For the full year, the return of the passive fund equals the market return. But imagine that the active fund held 15% in cash for the first half of the year and 1% in cash for the second half. Assuming cash returns 0%, the results are summarized below:

The active fund outperformed the passive fund simply by modifying its cash position. Because active funds can reach outside of stocks into a broader investable universe, the zero-sum game rule does not prevent them from outperforming the market.

Active Funds vs. Other Active Participants 

Of course, there is a second (and perhaps more obvious) source of potential outperformance. Excess returns for active funds could simply come at the expense of other active investors. Anyone else who trades stocks at non-market weights including individual investors, corporations, and employee stock-plan participants (the supposed “dumb” money) could underperform, effectively transferring excess returns to professional fund managers (“smart” money).

Conclusion

This exploration is simply a reminder that actively managed equity funds could, in theory, outperform the market, and that the zero-sum game rule does not prevent them from doing so. None of this is intended as an argument for, or against, active management.

The interesting thing is that most active funds don’t outperform the market. The reasons for this deserve further exploration. Structural causes of underperformance, along with thoughts on how to tackle them, will be the subject of future pieces.