Bill Nygren (Trades, Portfolio), a renowned value investor, is manager of the Oakmark Fund, Oakmark Select Fund and Oakmark Global Select Fund, as well as chief investment officer for U.S. equities at Harris Associates, a Chicago-based firm with approximately $118 billion in assets. His Oakmark colleague, Win Murray, co-manages the Oakmark Select Fund with Nygren and serves as director of U.S. equity research at Harris Associates.
Together, the two gurus answered questions about investing that GuruFocus readers asked recently. Read their responses in the interview below.
GuruFocus question: How do you control your emotions when dealing with a position at a loss, where you know that it is still undervalued?
Bill Nygren (Trades, Portfolio): First, you say “where you know that it is still undervalued.” I think it’s important to state that we never have enough certainty to say we know a stock is undervalued. We may believe it is undervalued, but it is extremely important to keep an open mind to new information that could prove the thesis incorrect. As a value investor, discipline and patience are prerequisites for success. However, taking them too far and becoming stubborn can be a fatal flaw. A great way to prevent becoming stubborn is to always search for non-confirming information.
I think the best way to keep emotions under control is to remain focused on business fundamentals rather than stock price. When we buy a stock, we establish a roadmap for how we expect the business fundamentals to progress. If the fundamentals are meeting our expectations but the stock has declined, we often use that as an opportunity to add to our position. On the other hand, if the fundamentals aren’t aligning with our expectations, we will usually consider our thesis broken and move on, regardless of how cheap the stock may look relative to our original expectations.
Question: How do you avoid value traps? How do you identify if you’ve invested in one, and how do you get out of it?
Win Murray: A value trap is a stock that appears cheap on traditional metrics, but in reality is fairly priced (or even overpriced). Existing holders don’t want to sell, telling themselves it “looks too cheap,” even though there are far better investment opportunities available.
As value investors, we look for companies selling at a discount to our estimate of intrinsic value. Many times these companies are out of favor in the market, and remain so for an extended period of time after our initial purchase. To restate your question: How can we tell which of these out-of-favor securities are outstanding ideas to which more capital should be deployed, and which are the ones that should be sold because our estimate of intrinsic value is incorrect?
Put simply, it comes down to whether the company’s fundamentals are tracking to our original expectations or not. When an Oakmark analyst recommends a stock, he or she lays out a thesis as to why the company is undervalued, what the revenues, margins, cash flows and capital structure are expected to be in a few years, and sets an estimate of the stock’s intrinsic value per share. Over time, this estimated share price should grow with the discount rate (minus dividends) as the company meets our fundamental expectations. If this is occurring and yet the stock price is underperforming, we are happy to increase our investment.
Sometimes, however, the company’s fundamentals disappoint, and our estimate of intrinsic value falls short of the original expectations. Many times when this happens, the company’s stock price drops to a greater extent than our value has changed, and it’s quite tempting to declare that the stock is cheaper than ever and, thus, is an even better idea than it was initially. This, however, is precisely the set-up that we’ve found potentially leads to a value trap.
Fortunately, the analysts at Oakmark are generalists (see Bill Nygren (Trades, Portfolio)’s recent commentary on why this is important here), which allows us to utilize a number of different strategies in the face of a possible value trap. We can transfer coverage of the stock to a different analyst for a fresh look, we can have an analyst write a “Devil’s Advocate” memo on the name (explaining to the entire investment department why it should be sold), we could have the current analyst review the stock at our weekly investment meeting, or we could just exit the position entirely. By staying vigilant to possible value traps and utilizing one or more of these techniques, we try to get out of these situations before they can cost more of our capital.
Question: What do you think about the current valuation of bank stocks? Which valuation metrics do you use? How do you adjust valuation for quality of management? Should banks be valued like commodity cyclicals?
Nygren: Banks account for more of the assets in Oakmark and Oakmark Select than any other single industry, so it follows that we believe banks are very attractively priced today. Before I get into valuation metrics, a word on management quality: Because of the high financial leverage at banks, rather than adjusting price targets for management quality, if we don’t like the management and culture of a bank, owning it is simply a non-starter. Banks are one of few industries where traditional metrics like price-to-book (P/B) or price-to-earnings (P/E) do a pretty good job of capturing value. We think a well-run bank should be able to earn 13-15% on its tangible equity and sell at 13-15x earnings. That equates to a price-to-book target of roughly two times.
Most banks today are priced at a small premium to book value and at less than ten times earnings. Most pay an above-market dividend and are repurchasing enough stock to produce double-digit earnings per share growth without requiring much net income growth. We don’t think banks have been given the credit they deserve for how much their risk profiles have been reduced compared to 10 to 20 years ago. Most have ratios of tangible equity to assets that are twice as high as they were going into the 2008 recession. Having twice the capital means, all else equal, they are now half as risky. But all else isn’t equal. Banks have also had much stricter underwriting standards following the real estate collapse, so their assets are now less risky. I believe many investors are treating banks as if the losses they suffered in 2008 are what should be expected in any recession. We view 2008 as a once-a-generation occurrence as opposed to a normal cyclic low.
I think it is interesting that one of the pushbacks we get about our large bank investment is that there is risk of political overreach that could so regulate banking that it becomes similar to electric utilities. Well, electric utilities are now priced in the market at about 20 times earnings and over two times book value. We don’t wish for more regulation, but clearly the “risk” of becoming priced similarly to electric utilities is not to be feared.
Last, another reason I think the large banks are far superior to commodity cyclicals is the competitive advantage they get from technology spending on initiatives like cyber security and mobile banking. Per dollar of deposits, those costs are much higher for small banks. Data suggests that banks we own, such as Bank of America (BAC, Financial), Citigroup (C, Financial), Ally Financial (ALLY, Financial), Capital One (COF, Financial) and Wells Fargo (WFC, Financial), are developing a significant cost advantage relative to small- and mid-sized banks. The large banks have been prohibited from growing via acquisition since the financial crisis, but they have grown market share just the same by using their cost advantage to win the battle for new customers.
Question: With the exception of Tesla (TSLA, Financial), the automakers all seem to sell at unjustifiably low P/E and P/CF ratios. Can you discuss your rationale for your holdings and explain what you think other investors are missing?
Nygren: An investment in Tesla requires more imagination than we have as to how different the auto industry may look in the future. Tesla needs to be a very different company than it is today to justify its current valuation.
We think there are a couple of valid reasons that auto manufacturers have sold at such low multiples. First, the industry was cyclical, with little change in peak sales from one cycle to the next. The companies reported high income at peaks and high losses in troughs. To account for the bad years, the stocks had to sell at very low multiples in good years.
Second, most of the reported earnings for auto manufacturers were not available to distribute to the shareholders. Capital expenditures were higher than depreciation, off balance sheet liabilities required constant funding, and for some of the companies, any remaining cash was used to rebuild captive finance companies, which were viewed as necessary to support sales during downturns.
Today, however, the view through the windshield appears quite different than what investors see in the rearview mirror. The large auto manufacturers have become global businesses that are benefiting from an expanding global middle class that wants to own cars. That has not only made the industry less cyclical, but has added an exciting growth element with peak sales higher in each successive cycle. Further, most of the companies are now projecting much better cash conversion such that most of reported earnings will be available for dividends and share repurchase. For those reasons, we would expect today’s very large gap in P/E multiples between autos and industrials to meaningfully shrink.
Companies like General Motors (GM, Financial) and Fiat Chrysler (FCAU, Financial) now earn almost all of their income from selling SUVs and pickups, which continue to take share from passenger cars. With their stocks selling at low single-digit P/Es and now also single-digit multiples of free cash flow, it isn’t farfetched to think they could be takeover candidates.
Question: How do you value businesses? Are asset-based or earnings-based valuations more useful?
Murray: A business is worth the sum of its future cash flows, discounted to today. Mathematically, that’s the only truly accurate way to value a company. At Oakmark, we estimate a business’s pre-tax, pre-interest cash flows (assuming a normal margin), apply an estimated tax rate, capital required for growth, and a projected capital structure, then estimate an intermediate growth rate. These factors, combined with an appropriate discount rate, tell us what multiple of cash flows the business likely deserves.
You mention two other valuation metrics: earnings based and asset based. Both have utility as shortcuts to a discounted cash flow model, although cash flows ultimately determine value. If the company’s earnings are approximately equal to their after-tax cash flows, then an earnings-based valuation model will likely be an accurate tool for determining the business’s value. This is not uncommon in a lot of mature industries that aren’t too research and development or amortization heavy and, therefore, is pretty widely used by investors.
Asset-based valuations are especially useful in very cyclical industries during time periods when current cash flows are far lower (or higher) than a company’s “normal” long-term cash flows. Take, for example, a deep-water drilling company. The company owns dozens of deep-water drillships, each worth hundreds of millions of dollars. When oil was more than $100 a barrel, these ships were contracted at exceptionally high rates, producing amazing cash returns on the original asset prices. However, when the oil price collapsed, many of these ships were left idle, producing no cash flows at all (and, in fact, requiring maintenance cash flows to keep shipshape for the future). A reasonable valuation methodology here would be to determine the replacement value of these assets if someone were to try to build the same fleet (adjusted for deprecation) and use this as a basis for valuing the company long term. You have to be careful when doing this, though, because ultimately all that matters are the cash flows. If, for example, circumstances in the industry have changed such that it would be impossible to get a good long-term return on a newly built drillship, then an asset value model would overstate the company’s value.
Nygren: Value investors usually look at much slower-growth companies than Alphabet and are therefore accustomed to getting more of their return through the free cash flow yield than from organic business growth. In the case of Alphabet, it has achieved annual organic growth of 20% or so while still producing a 3% free cash yield, which, if there was nothing more to the story, would still be pretty amazing.
More importantly, though, we believe that looking simply at an Alphabet P/E or price-to-free cash flow (P/FCF) ratio ignores or values many of its assets at negative numbers. Let’s start with cash. The company is expected to exit 2019 with nearly $150 billion of net cash, or $212 per share. That cash earns about 1% after tax. Cash would have to sell at a P/E of 100 to fully reflect its value. We prefer to subtract the cash from the stock price and subtract the interest income from earnings per share to compute the P/E for the business.
Next, Alphabet’s “other bets” include the autonomous driving subsidiary Waymo, artificial intelligence, cloud computing, Google Play (Android) and many other venture stage investments. In total, “other bets” is projected to reduce earnings per share by about $5 next year. So to get a more accurate picture of value, we add “other bets” losses back to search earnings and also subtract from the stock price the estimated asset values for each investment. Effectively, this step is equivalent to what the GAAP presentation would be if Alphabet made these investments through a venture capital firm as opposed to making them in-house.
Finally, YouTube is intentionally under monetized (lower subscription fee and/or lower ad volume relative to other video content) to maximize growth in hours viewed. Were hours of viewing on YouTube valued similarly to current stock market values of broadcast or cable networks, YouTube would be worth hundreds of dollars per share. If you subtract the value of all non-search assets from Alphabet’s share price and divide the price you are paying for Google by its search income, you get a P/E ratio that is lower than the S&P 500 P/E.
Though on the surface Alphabet doesn’t look like a value stock, when you value the assets separately and sum them up, it is easy to get to numbers much higher than the current stock price.
Question: Given how beaten down brick-and-mortar retailers are, don't you think there are some incredible opportunities in that sector (despite the secular risks)?
Murray: Many value investors use some form of “mean reversion” in their approach to setting price targets, assuming that over the long run, returns will revert to their long-term averages. But in the case of brick-and-mortar retailers, although some look quite inexpensive when applying historic P/E multiples to current earnings, we believe that this frequently far overstates the value of these businesses given the secular challenges many of them face. In addition, a number of them have many years of significant lease obligations, which aren’t properly accounted for in their stated enterprise values. These obligations impede the companies’ ability to restructure their store bases and are a call on future cash flows.
All of that isn’t to say that these companies are worthless, but we must be careful to model the cash flows using the correct (sometimes negative) future growth rates. We’ve owned a number of retailers in the past and continue to believe that many of these businesses have very durable business models (Tiffany, Home Depot, CarMax and Carters come to mind), but we are always quite price-sensitive when making our initial investments.
Question: Netflix (NFLX, Financial) is a large position in both Oakmark and Oakmark Select. What makes you think it’s undervalued given its high P/E and negative cash flow? Isn’t the growth rate in earnings per share that Netflix requires simply to justify the current price almost impossible to achieve?
Nygren: We believe that generally accepted accounting principles (GAAP) do a poor job of measuring business value and change in value for Netflix. Note, this isn’t specific to Netflix, but rather applies to many of the growing number of asset-lite businesses where investments for growth are expensed rather than capitalized.
Using typical GAAP metrics, like earnings per share and book value, Netflix appears ridiculously valued: It sells at about 100x expected earnings, and about 30x book value. But the company reminds us a lot of the cable companies Oakmark owned in the early 1990s. Those businesses had negative book value, negative earnings and negative cash flow, yet their stocks performed well and many were acquired by larger cable companies or private equity. The acquisition prices were consistently around $1,000 per subscriber. If you adjusted the GAAP book value to reflect that $1,000 per subscriber value and adjusted GAAP earnings by adding $1,000 for each net new customer added, the stocks looked cheap.
One of Netflix’s primary competitors is AT&T’s HBO NOW. When AT&T acquired HBO in the Time Warner acquisition, it appeared to pay about $1,000 per subscriber. At $16 per month and a 40% operating margin, that implies about 13x pretax earnings, which seems to be in the right ballpark.
So how does Netflix stock look relative to a $1,000 per sub value benchmark? It is expected to end 2019 with 168 million subscribers, which would be worth $168 billion. Subtracting $10 billion of debt would leave a value of $158 billion, or about $360 per share. Defining earnings in terms of subscriber additions gives 29 million new subs in 2019 for a $29 billion addition to value while adding $3 billion of debt, so a net increase in value of $26 billion. At $325, Netflix has an equity capitalization of $142 billion, meaning the stock is priced at less than 6 times the value added in 2019.
If Netflix subscribers are worth as much as HBO subscribers, it is a very cheap stock. Most Netflix subscribers say their Netflix subscription, despite costing less than $12 per month, is more important to them than HBO, Spotify, Sirius XM, etc., all of which are charging $15 to $20 per month. We believe Netflix is currently intentionally underpricing its product by at least $4 per month because it can create more value through subscriber growth than it could by maximizing current earnings. Because GAAP requires immediate expensing of customer acquisition cost rather than amortizing it over the expected life of the customer, the company reports trivial income relative to the growth in its value. Further, growing the number of subscribers makes it much more difficult for any competitor to catch Netflix. Because it has more subscribers, the company can pay more for programming yet pay less per subscriber than any competitor.
We believe there are numerous examples in the market today, and in the Oakmark Fund, where GAAP makes a stock look overvalued because investment spending, such as R&D or customer acquisition costs, is unduly depressing current earnings.
Stay tuned for the second part of this interview next week.