Bubble or Boom? Rob Arnott and Cathie Wood Share Perspectives on Equity
A debate on Tesla, disruption, and reversion to the mean.
Editor's note: This article first appeared in the Q4 2021 issue of Morningstar magazine. Click here to subscribe.
The opening keynote of this year’s Morningstar Investment Conference was a widely anticipated discussion with Rob Arnott and Cathie Wood.
Arnott, founder and chairman of the board of Research Affiliates, and Wood, founder and CEO of ARK Investment Management, have at least one thing in common: They both aim to beat a market-cap index over the long term. But they go about that in very different ways. Arnott is a diversified, contrarian, value-oriented manager, while Wood is a concentrated, disruptive technology-focused, growth-oriented investor.
Their strong opinions on electric vehicles, the process of disruption, and market valuations made for a lively debate. This transcript of our conversation has been edited for length and clarity.
Daniel Needham: Over the last decade or so, we’ve seen a big spread between value and growth stocks. Rob, is there going to be mean reversion, or is this time different?
Rob Arnott: We always hear the observation that it’s different this time. It’s always true. Things are always different this time. They’re just not necessarily different enough to matter.
We’ve observed over the course of decades that the spread in valuation between growth and value stocks widens and narrows, widens and narrows. And one of the marvelous things is that when it widens, that’s a good predictor for subsequent performance of value relative to growth. The same holds true when it narrows. It was abnormally narrow in 2007, and that’s when we had the quant crash. Everyone was piling into value, and value cratered.
So yes, I do think they will mean-revert. The most powerful evidence is that over the last 13 years, value—defined in the Fama-French fashion as book/price—has underperformed growth peak to trough—the trough was last August—by 58%. But value got cheaper by 70%. If you have a stock that’s down 58%, and its P/E ratio or price/book is down 70%, that means it’s cheap. If I’m not willing to buy now, I never will.
Needham: Cathie, you’ve said that almost half of the S&P 500 is threatened by disruption and risks being a value trap. Does technological disruption make concepts like mean reversion and contrarian investing strategies riskier, less relevant?
Cathie Wood: We are focused on the five major innovation platforms that are evolving at the same time today: DNA sequencing, robotics (especially collaborative or adaptive robotics), energy storage, artificial intelligence, and blockchain technology. The seeds for these five platforms were planted during the 20 years that ended in the tech and telecom bubble. Too much capital chased too few opportunities, and it all ended badly. But that does not change the fact that the seeds were planted. They have been gestating for 20, 25 years, and now we’re on the cusp of transformations to every sector globally, the likes of which we have not seen since the early 1900s. The major technological platforms back then were telephone, electricity, and the automobile. Life as we knew it back then certainly changed and never reverted to the mean. We did not want to go backwards. I think the same will be true this time as well.
Needham: One industry you’ve got a relatively strong view on is electric vehicles and specifically Tesla TSLA, which some may consider to be at a high price relative to fundamentals.
Wood: Our research is centered around Wright’s law, a relative of Moore’s law. Moore’s law is a function of time; Wright’s law is a function of units. Theodore Wright made this observation during the early days of airplane manufacturing. For every cumulative doubling in the number of units produced—so, one to two, two to four, four to eight—the costs associated with these new technologies declines at a consistent percentage rate.
In the case of the battery pack systems within electric vehicles, that cost decline is 28% for every cumulative doubling. Last year, we globally produced and sold roughly 2.2 million electric vehicles. Based on that cost decline in battery pack systems, the largest cost component of electric vehicles, we believe that the average electric vehicle price will drop below that of the average gas-powered price in the next year or so and will continue to decline. In 2025, the average electric vehicle, like a Toyota Camry electric vehicle, will cost $18,000, while the Toyota Camry will still be roughly $25,000. Electric vehicles are better-performing vehicles. They have fewer parts, so much less servicing is required, and there is no gasoline trip every week. I have not had to go to the gas station since September of 2018. The total cost of ownership today of electric vehicles, when you incorporate everything, including insurance, is already lower.
We believe that the number of electric vehicles sold will scale from 2.2 million last year to 40 million in 2025, which is almost half of total car sales globally that we expect. That’s a twentyfold increase. This is exponential growth simply based on this notion that these cars are going to become more affordable than gas-powered vehicles.
Needham: Why will Tesla be the one that benefits from that? Why won’t the more traditional autos or the many other electric vehicle manufacturers capture that trend?
Wood: Tesla has created four major barriers to entry. One is battery costs: It built its cars on cylindrical batteries. Most other auto manufacturers base their cars on lithium-ion pouch batteries, which cost roughly 15%, 20% more than the cylindrical batteries that Tesla uses. Tesla is riding down the cost curve of the consumer electronics industry. Most other auto manufacturers are not.
The second barrier to entry is the artificial intelligence chip that Tesla designed. Tesla is taking a leaf from Apple’s AAPL book. Apple created the concept of a smartphone. It believed that we would have a computer in our pocket. Nokia NOK, Motorola MSI, and Ericsson ERIC did not believe that. And you know where they are today. Apple is dominating the profits in the smartphone market.
The third barrier to entry is the number of real-world miles driven that Tesla has collected. It has more than a million robots out there collecting data and sending it back every day. My car is one of them. Therefore, it is able to discern corner cases and design its full self-driving system to incorporate these corner cases in a way that other auto manufacturers cannot.
The fourth barrier to entry is that Tesla is still the only car doing over-the-air software updates to improve performance and prevent breakdowns. Since September 2018, I have not had to take my car in for any reason except a nail in the tire. Well, software is not going to solve that problem.
Needham: Rob, you’ve got some opinions on electric vehicles and Tesla.
Arnott: I certainly do. We wrote a paper earlier this year called “The Big Market Delusion,” which looked at up-and-coming industries that are disruptive. The market’s pricing every company in these industries at lofty multiples, as if they are all going to succeed. Yet they’re competing against one another, so there will be winners and losers. Look at the Internet bubble, or the supercomputer bubble in the early ’80s.
Disruptors get disrupted. Palm was spun off from 3Com back in the year 2000 and had an initial value that was more than 3Com was valued at before the spin-off. Palm was disrupted; BlackBerry BB came along with a better product. BlackBerry was disrupted; Apple came along with a better product.
Disruptors are massively important to the economy and to economic growth. But you have to look at (a) how disruptive are they, (b) how much of a premium are you paying for that disruption, and (c) are they vulnerable to being disrupted themselves?
When we looked at the electric vehicle market, we found that as of the end of 2020, out of eight specialists in electric cars worldwide, Tesla was the second cheapest. At the time it was about 25 times sales. The highest multiple was well over 10,000 times sales, because it was a company that had aggregate sales from the previous year that was measured in three digits. I don’t mean three digits of production. I mean three digits in dollars.
Peeling the Onions
Needham: So, we can say that you think valuations are stretched within the EV space?
Arnott: Yeah. In the Internet bubble, there were countless disruptors, and they were radically reshaping the way we communicate, the way we transact, the way we interact with our clients. And thank goodness for that Internet revolution, because when COVID came along, all of those innovations turned out to be massively important to all of us.
But the market got ahead of what was likely to happen. Cisco CSCO was briefly the largest market-cap stock on the planet. Since that time, they’ve had double-digit annual growth for 21 years, but their share price is lower than it was at the peak in 2000. The market was expecting stupendous growth. It got [only] impressive growth.
That’s where I’d be inclined to push back. Cathie, not to take anything away from you, there have been 108 mutual funds and ETFs over the last 30 years that have gone up over 100%. Five of those are yours. Wow! But 70% of them were down the following year, 80% of them were down the following three years, by an average of 53%. My question: What’s the sell discipline that can protect those gains? What’s the sell discipline that can rotate you into undiscovered disruptors?
Wood: First of all, just to support Tesla’s position out there, there may be a lot of electric vehicle manufacturers, but they are tiny. That is a function of the four barriers to entry. In the early days of its battery, auto manufacturers and auto analysts laughed at Tesla. It’s building its car on top of cell phone batteries that are blowing up in airplanes? Ha ha ha. Now there is a Bolt recall because its batteries are catching on fire. Tesla nailed that one down, and the traditional auto manufacturers are having problems.
In terms of the differences between our styles and our strategies, we are looking forward, not backward, and many of the questions you’ve just asked are backward-looking. We are looking at exponential growth opportunities that have evolved as these innovation platforms have started to mature and move into prime time.
These seeds, as I mentioned earlier, were planted in the 20 years that ended in 2000. I think the exponential growth rate that we’re going to see is a function of how long they have gestated. When I throw out a number like 88% or 89% compound annual growth in units that sounds preposterous. But it is tracking. Wright’s law gives us a very nice guide as to how the costs should decline. If you pass that down in prices, the uptake will be there. This exponential growth trajectory is going to be cemented.
I think most of the world is used to looking at the linear growth opportunities. That’s the kind of growth that has dominated ever since telephone, electricity, and the internal combustion engine evolved. We got a glimpse of exponential growth with the Internet, but the costs were too high. The technologies weren’t ready. Now the costs are low enough, the technologies are ready, and not only are we seeing these five innovation platforms evolve at the same time, we are beginning to see them converge. And we’re beginning to see S-curves feeding S-curves feeding S-curves.
Let me give you some of the drama here. I already gave you the 28% cost decline for battery pack systems in electric vehicles. If you look at DNA sequencing costs, for every cumulative doubling in the number of whole human genomes sequenced, those costs drop at a 40% rate. The genomics revolution is alive and is exploding. In the industrial robotics space, collaborative robot costs are dropping for every cumulative doubling at a 58% rate. In the world of artificial intelligence, AI training costs are dropping 68% per year.
I don’t think many analysts and portfolio managers are as focused on this as we are. This is all we do. The more we peel these onions, the more astonished we are at the growth rates that are evolving, mostly because of the convergence.
An example of the convergence among three of our platforms is autonomous taxi networks. This is the next big leg of valuation for Tesla, we believe. Autonomous taxi networks are the convergence of robots, energy storage, and artificial intelligence. We’ve got cost declines in these three areas. You get the convergence of that, you have an explosion. We think no one’s even close to Tesla. The closest might be in China, but we think what’s going on in China right now is going to turn off a lot of innovation.
Needham: You’ve talked about the cost side, but demand’s always been hard to predict. How do you think about the range of possible outcomes and the possibility that you’re wrong when you’re valuing these businesses?
Wood: We listen to every earnings call. We’re talking to managements, just like other teams. And we get market share data and unit data all the time. We are constantly evaluating whether this is either happening more slowly or more quickly than we expected.
On the electric vehicle side, the one risk that has evolved is China, because China is going to be a big source of demand for electric vehicles. China just said it’s going to close down all its coal-powered plants, so it seems to us it’s moving more aggressively to the electric future—the electrification of transportation.
How do we value these exponential growth opportunities? Our investment time horizon is five years. Usually, by the way, demand does come through. Better, cheaper, faster, more creative, more productive typically works. The demand is there if the costs are right. And we project out the cash flows, based on Wright’s law and our cost trajectories, to arrive at an EBITDA in year five.
For many of the categories that we invest in, except perhaps genomics, we will simply put a FAANG-type multiple, a mature innovation company multiple, on these stocks of 18 times EBITDA. We do not believe our companies will be mature in year five. We are making a conservative assumption. Our minimum hurdle rate of return is 15% at a compound annual rate. Right now, if you were to look at our flagship strategy, our compound annual rate of return expectation is 30%.
Possible or Plausible?
Needham: Rob, Warren Buffett said, “Beware of past performance ‘proofs’ in finance. If history books were the key to riches, the Forbes 400 list would be full of librarians.” Your approach looks at historical fundamentals and relies on some of those relationships to hold. How do you think about disruptive elements?
Arnott: Mean reversion is the most powerful factor at work in the capital markets. It shows up on earnings growth. When you have very rapid earnings growth, it tends to mean-revert down. When you have tanking earnings, it tends to mean-revert up. (Not in all cases; there are value traps.) A lot of our work is based on long-horizon mean reversion and a recognition that there are things you can take from the past and things you can’t. You can take from the past measures of mean reversion. You can take from the past measures of how fast aggregate earnings grow relative to the macroeconomy.
It’s not widely known, yet it’s a simple truism, that earnings per share for the broad market cannot rise as fast as GDP growth on a long-term basis. Why can’t they? Because GDP growth consists of growth in existing companies and creation of new companies, the kinds that Cathie’s looking for. The creation of the new companies is an enormous engine for economic growth.
But I wouldn’t suggest that the pace of innovation today is radically different from the past. That’s an area where I’m sure that Cathie would disagree. If you go back to the start of the Industrial Revolution, there’s been one revolution after another after another after another. One of the most interesting things is that the key beneficiaries of those innovations are usually not the shareholders, but the customers.
When you’re looking at a whole spectrum of disruptive companies, there will be some that turn out to be spectacular. Go back to the first tech bubble. How many of the 10 largest tech stocks by market cap in the year 2000 outperformed the market over the next 10 years? Zero. Not one. How many outperformed over the next 20 years? One, Microsoft MSFT. What about Amazon.com AMZN and Apple? They weren’t anywhere near the top 10 back then. Apple was perceived to be poised on the brink of ruin.
Our definition of a bubble is a very simple one: If you’re using a discounted cash flow model or some other valuation model, you’d have to use implausible assumptions to justify today’s price. We plugged in 50% growth for 10 years for Tesla, assumed profitability matching the best of any automaker of any single year of the last 10 years, and we came up with a net present value of 430 bucks. I view 50% growth as implausible. Cathie does not. So, I view Tesla as a bubble. Cathie does not.
But two things are interesting about bubbles. One, they can go much further and last much longer than any skeptic would expect, so be very careful about short-selling bubbles. The most vivid example of that to me was the Zimbabwe stock market in 2008. Coming into the summer, you could have looked at Zimbabwe and said there’s hyperinflation here. This country’s in trouble. I do not want to own stocks here. In fact, I’m going to short-sell them. But it’s a crazy market. I’ll be conservative. I’ll put 2% in as a short position. Well, Zimbabwe stocks over the next six weeks rose 500-fold, which means in U.S. dollar terms, it rose fiftyfold. You just got wiped out.
The second observation about bubbles is that implausible growth assumptions don’t mean impossible. Amazon in 2000 would have qualified for my definition of a bubble. Amazon was a terrible investment in the 2000s and got it all back with room to spare in the 2010s, when it grew elevenfold.
To justify Tesla’s current price, you’d have to assume roughly fiftyfold growth over the next 10 years. Is that impossible? No, anything is possible. Do you believe it’s plausible? I don’t. As with Amazon in 2000, I could be proved wrong. But as with Amazon in 2000, you might have to wait a while for the market to catch up to the actual growth opportunities if they are as extravagant as Cathie says.
Wood: Tesla is not an auto manufacturer. It’s a robotics company, energy storage company, artificial intelligence company. We have our robotics, energy storage, and artificial intelligence analysts focused on this stock. Or analysts are specialists on technologies and generalists on sectors, because as technology costs come down, the access increases across many sectors. The reason auto analysts missed Tesla—and they all did—is because they’re not the right analysts to be analyzing Tesla.
Amazon is a really good example of what did come out of the tech and telecom bubble. We think Amazon is the prototypical company based on these innovation platforms. We’re going to see years of super growth from these companies, because they are platform-oriented companies, and the companies with the best data and the best AI expertise are going to take the lion’s share of these opportunities. Tesla is certainly in that category when it comes to autonomous networks.
Reflecting on Amazon and the bust generally, that period created such muscle memory in our business that there is real fear when it comes to our strategy. As we’re going through due diligence and client presentations, we get more questions about the risks than we do about the opportunities. I remember all too well in the tech and telecom bubble, no one even thought about the risks. So, we are not in a bubble today. That I know. But I do believe that the market is beginning to understand how profound some of these platform opportunities are and how sustained and rapid the growth rates are going to be.
If You Had to Choose
Needham: Given that you both love large-cap, market-cap-weighted indexes, I’m going to give you a choice. I’m going to give you $1 million. You can take a 10-year Treasury bond that matures in 10 years or you can take a U.S. large-cap index-based ETF. You’ve got to hold for 10 years. Which would you take?
Arnott: I’d find that very difficult, because I expect them both to produce about a 2% annual return for the next 10 years. Can I put it in emerging-markets value stocks?
Needham: You have to choose.
Arnott: Stocks, because there’s more risk of upside surprise than downside relative to our 2% expectation. This is one area where risk is good. The downside risk could take it 10% below that. The upside could take it 15% above that. That’s an asymmetry I like.
Needham: We’ve got Rob going long the U.S. equity market, for the first time in 15 years.
Needham: Cathie, over to you.
Wood: That’s a hard one. I probably would choose the equity, but I think the years ahead are going to be so confusing because of this dynamic of disruptive innovation and creative destruction, which is not isolated to value stocks. There are so-called growth stocks out there that are going to be in harm’s way, especially if the companies have been buying back shares, leveraging up to do so and to pay dividends, and haven’t been spending aggressively enough in the right areas in R&D. We think that there’s going to be a huge deflationary pull out there, which means that these companies are going to be in a bad way. To service their debt, they’ll have to sell their increasingly obsolete products at lower prices.
Needham: So, it sounds like deflation equals a Treasury bond?
Wood: I think the Treasury bond will do well, and a lot of what people consider growth stocks right now will be surprisingly poor-performing.
Arnott: The short answer is we both disliked the question, because you didn’t offer us anything that we liked. Don’t buy the index fund. Don’t buy the Treasury. Buy emerging-markets deep-value stocks, which are likely to mean-revert well enough to roughly double your money each five years of this decade.
Needham: So, I lost control of the panel! Thanks very much, Cathie and Rob. This was a lot of fun.
Daniel Needham, CFA, is president and chief investment officer of Morningstar Investment Management.
Photography by Matthew Gilson.