“Value investors have bored momentum investors for decades by trotting out the axiom that the four most dangerous words are, ‘This time is different.’ For 2017, I would like, however, to add to this warning: Conversely, it can be very dangerous indeed to assume that things are never different.” – Jeremy Grantham, Co-Founder, GMO
There is much concern over the level of the stock market. Price to earnings (PE) ratios generally, and the Shiller CAPE PE specifically, stand at levels not seen with the exception of the roaring 20’s and dot-com bubble. To the extent that investors believe the future will look like the past, these PE ratios indicate that the market is overvalued.
But of course, change is the only constant, and the future will be different than the past. What matters to investors is how quickly and in what ways that change will occur. As Warren Buffett wrote in 1981, the greater the rate of change, the more we need to examine the assumptions underlying our historical valuation measures:
“While investors and managers must place their feet in the future, their memories and nervous systems often remain plugged into the past. It is much easier for investors to utilize historic p/e ratios or for managers to utilize historic business valuation yardsticks than it is for either group to rethink their premises daily. When change is slow, constant rethinking is actually undesirable; it achieves little and slows response time. But when change is great, yesterday’s assumptions can be retained only at great cost. And the pace of economic change has become breathtaking.”
The pace of economic change that Buffett found breathtaking in the early eighties has only accelerated. One metric that bears this out is the average dominant period of companies listed in the S&P 500:
Among, if not the principal force driving the increased pace of disruption has been the advancements in information technology. As the venture capitalist Marc Andreessen put it in the title of a now famous 2011 article, “Software Is Eating The World”. And indeed it is. According to Bloomberg News, the technology sector recently reached 25% of the S&P 500, a milestone not seen since the tech bubble.
As software becomes an increasing proportion of American business, it follows that the unique economics of software businesses will impact the character of the U.S. economy and stock market.
One likely and, to my knowledge, little commented upon outcome of software’s increased share of the economic activity is that stock market earnings will be better protected from inflation.
Software, Marginal Costs & Inflation
It is under-appreciated, even in 2018, that stocks fare poorly in inflationary environments. Yet as far back as the 1970’s, Warren Buffett was writing about the types of business investors should prefer to own under inflationary conditions:
“Such favored business must have two characteristics: (1) an ability to increase prices rather easily (even when product demand is flat and capacity is not fully utilized) without fear of significant loss of either market share or unit volume, and (2) an ability to accommodate large dollar volume increases in business (often produced more by inflation than by real growth) with only minor additional investment of capital.” – Warren Buffett, How Inflation Swindles The Equity Investor
Software-based businesses satisfy Buffett’s second criteria; they can increase the volume of their business with only minor additions of capital. To understand why this is special, and how it relates to inflation, it’s useful to start by describing why most traditional businesses do not satisfy Buffett’s criteria.
Historically, most businesses were manufacturing businesses. I.e. they made stuff. And thus had both fixed and variable costs. Fixed costs describe any costs associated with running the business and equipment that does not result from additional units of product. Marginal costs are those that can be directly associated with the cost of making an additional unit of product.
Using General Motors (GM) as an example, it’s plants and the equipment in them represent the fixed costs required to run the business. While the supplies and labor that are required to make each individual car represent the marginal costs. Both types of costs represent a consumption of capital, but only one type of cost relates directly to the volume of business the company can do.
Software is special because it is the first product in the history of the world that costs nothing to replicate and, with the advent of the internet, nothing to distribute. This property creates businesses that have almost exclusively fixed costs and zero marginal costs.
The fixed costs of software businesses tend to be very large. However, once these costs have been spent, it costs nothing to serve an additional unit of product. For example, it cost billions of dollars to build Google’s search engine and advertising platform. But now that it’s created, it costs virtually nothing for Google to provide an additional billion search results and advertisements associated with them. General Motors, by contrast, has to incur significant capital outlays if it wants to increase the number of cars it manufactures in a given year.
This difference in marginal costs becomes material to investors because in an inflationary environment two things happen: 1) the cost of capital increases and 2) input prices increase.
When the cost of capital increases, investors would rather be taking cash from their investments and reinvesting it elsewhere. Unfortunately, for most businesses, inflation also raises costs. That means that for companies like GM, the price of their marginal costs (parts & labor) increases, and so instead of taking money out, the company actually requires more capital just to achieve the same level of return.
For a sense of how changes in the cost of capital would feel in day-to-day life, imagine being forced to own a car that becomes less fuel efficient (less miles per gallon), as the price of gas increases – a double whammy.
A software company like Google on the other hand, by virtue of it’s near-zero marginal costs, could maintain or even increase its level of output without requiring any additional high-cost capital. Its owners thus enjoy the ability to take capital out of the business when it’s most scarce and, to the extent possible, even increase earnings without the need for additional capital. It’s this difference that makes software’s low marginal cost economics so attractive to investors during inflationary environments.
With respect to market valuations, this dynamic is relevant to investors because of software’s increased share of the U.S. economy. To the extent that software drives more of American businesses towards a low marginal cost model, the risk that returns on capital will decline in an inflationary environment should diminish. That, in turn, should allow investors to justify paying a somewhat higher multiple of earnings relative to history.
Software, The Internet & Profit Margins
Software’s impact on the character of American business extends further than its ability to maintain its earnings power during inflation. Evidence indicates that software is underwriting a trend towards monopolistic corporations earning supernormal profit margins. Numerous articles and academic papers note the increasingly “Winner Take All” nature of stock market generally, and the super-normal profitability of firms in the technology sector specifically.
My favorite explanation of the mechanisms underlying this trend comes from the brilliant Ben Thompson, writer of the blog Stratechery. His piece, Aggregation Theory explains that software and the internet is shifting the corporate value chain from one where profits were created at physical distribution points, to one where digital distribution creates profits for firms that leverage technology to create superior user experiences. He writes:
“First, the Internet has made distribution (of digital goods) free, neutralizing the advantage that pre-Internet distributors leveraged to integrate with suppliers. Secondly, the Internet has made transaction costs zero, making it viable for a distributor to integrate forward with end users/consumers at scale….
This has fundamentally changed the plane of competition: no longer do distributors compete based upon exclusive supplier relationships, with consumers/users an afterthought. Instead, suppliers can be aggregated at scale leaving consumers/users as a first order priority. By extension, this means that the most important factor determining success is the user experience: the best distributors/aggregators/market-makers win by providing the best experience, which earns them the most consumers/users, which attracts the most suppliers, which enhances the user experience in a virtuous cycle.”
Google, Facebook and Amazon are the clearest examples of how this dynamic has played out. But in my opinion, Apple is the most compelling case for how technology reaches beyond the realm of pure-software to impact the profitability of more traditional businesses.
I single out Apple because software monopolies such as Google and Facebook so clearly benefit from classic network effects, that it obscures the fact that what’s driving monopolistic firms more broadly is software’s ability to differentiate traditional physical goods via a superior user-experience. (Does Apple benefit from some network effects? Yes, but I think it would be hard to argue that they are the sole basis for its dominance.)
“What makes this model so effective — and so profitable — is that Apple has differentiated its otherwise commoditizable hardware with software. Software is a completely new type of good in that it is both infinitely differentiable yet infinitely copyable; this means that any piece of software is both completely unique yet has unlimited supply, leading to a theoretical price of $0. However, by combining the differentiable qualities of software with hardware that requires real assets and commodities to manufacture, Apple is able to charge an incredible premium for its products.”
Again, Apple’s success at differentiating hardware via software is worth recognizing because it shows that the trend towards super-profitable firms is not limited to “internet companies” like Facebook and Google. And as such, it may be an indication that the trend towards more concentrated and profitable firms has an underlying mechanism to extend itself even further.
The trend towards higher profits and the sustainability of that trend is highly important to stock market investors because higher future profits allow investors to pay higher multiples of current earnings and expect similar returns. Thus, to the extent that information technology is driving a trend towards more profitable firms, what look like high P/E ratios by historical standards may not be as expensive as they appear.
Profit Margins, Pricing Power & Inflation
Thompsons framework is also useful for evaluating what the trend towards more highly-profitable firms may mean for inflation risk. Recall that Buffett’s dictum for investing during inflation called for buying businesses with two characteristics:
1) The ability to raise prices even if demand has not increased.
2) An ability to increase the volume of business without adding incremental capital.
We explored how software based, zero-marginal-cost business models fulfilled his second criteria. The question remains, do the monopolistic profits of software-enabled Aggregators satisfy his second?
My first intuition was, yes. Monopolies, after-all, earn supernormal profits because they have a competitive advantage which allows them charge higher prices. However, upon further reflection, there are two reasons to think that tech-centric monopolies may not have the ability or willingness to raise prices in the ways monopolies have historically:
- Historically monopolies commanded pricing power by controlling supply. Today’s monopolies do not control supply.
- Software monopolies rely on aggregating users to leverage positive feedback loops which incentivizes them to prioritize the volume users/activity on their platforms, possibly at the expense of profitability.
To the first point, note that historically monopolies achieved pricing power by controlling the supply of a good or service. That meant the consumer’s choice was binary; you either bought the product from the monopolist or didn’t get it at all. Need oil? No problem, you just have to buy from Standard Oil, and they set the price.
Today’s software monopolies, by contrast, are monopolies of distribution, not supply. Google may have the best advertising platform but it is not the only place to advertise. Google maintains a near-monopoly of ad-market share by delivering the best experience for both its advertisers and search users but there exists a price at which that advantage is not enough and its customers (in this case advertisers) would begin to switch to a competitor. Thus relative to historical monopolies of supply, today’s tech monopolies are disadvantaged in terms of their ability to raise prices and maintain the same volume of business.
The second limiting factor speaks to a feature of the aggregator business model. As Ben Thompson pointed out in Aggregation Theory, competition has shifted towards differentiation by delivering superior user experiences. How each firm achieves this varies. But in many cases their dominance is based on some combination of network effects and/or positive feedback loops which feed off of the activity of consumers and suppliers. Because firms understand how important these dynamics have been to their dominance, they are likely incentivized to sacrifice profit in order to maintain their position.
Put another way, sacrificing users in order to maintain profitability may pose a real or at least perceived existential threat to these companies business models. It’s worth remembering that these firms entire institutional experience has reinforced the idea that user growth and engagement should be prioritized over short-term profitability. As such, management may sacrifice profitability to ensure that the feedback loops which enabled their dominant position continue to operate. Their not unjustified fear being that by sacrificing users/engagement for profits they might inadvertently create the conditions under which a competitor is able to successfully leverage their new-found market-share and challenge the incumbent firms’ dominance.
With these caveats in mind, it’s important to recognize that these businesses in aggregate do likely possess the market-power to raise prices by some degree. Further, to the extent that they can both raise prices and run a low marginal cost business model, they embody the beau ideal of an inflation protected business, according to Buffett.
Valuing Inflation Protected Stocks
We’ve explored how software’s increasing share of American business is diving a trend towards firms with low marginal cost business models and greater pricing power. Both of these trends imply that relative to history, corporate earnings will be better protected from inflation. Because investors are concerned about real rates of return, investors should be willing to pay a higher price to earnings ratios than they have historically. But how much more?
It’s impossible to say with certainty but the bond market offers clues. As of this writing, the difference between 30-year nominal and inflation protected bonds is about 2%. Using the implied inflation estimate from the bond market, we can say that the upper limit for inflation protection in the market today is worth a 2% reduction in yield.
To translate that into stock market terms, consider the following hypothetical; Imagine the U.S. Government is offering a security that pays the earnings from the S&P 500, adjusted for inflation. How much would a rational investor be willing to pay for it?
As in the bond market, investors would be willing to pay more for a security that protects them from inflation but only up to their estimate of future inflation rates. As stated previously, the bond market’s best guess of inflation over the next 30 years is about 2%. Thus, reducing the current yield of the S&P 500 by 2% should give us an approximation of what investors would be willing to pay for a security that offered the earnings power of the stock market with the addition of inflation protection.
Adjusting the current equity market price to earnings multiple would reduce the yield on the S&P 500 from 3.91% to 1.91% and take the PE from 25 to 52. The price of the hypothetical Inflation Protected S&P 500 would be $5,607 versus ~$2,700 for the nominal S&P 500. Again, this hypothetical illustrates the extreme upper limit on the value of inflation protection in today’s market.
For a more accurate estimate of the value of inflation protection we need to moderate it’s expected value by the following two factors:
- The proportion of economic activity software-based business will represent in the future relative to the past.
- The extent to which such businesses through a combination of pricing power and/or low-marginal cost will insulate investors from inflation
For example, using the historical average Shiller PE of 16 and a historical inflation rate of 3%, and under the assumption that software-based businesses will represent an incremental 20% share of U.S. stock market earnings and 30% of those earnings were effectively insulated from the risk of inflation, the incremental PE increase would be 0.50. E.g. a 16 PE becomes a 16.5 PE.
|Historical Inflation Rate||% Of Market Tech||% of Tech Inflation Resistant||Historical Avg PE||Adjusted Fair Value PE||PE Difference|
Looking at a range of assumptions, as I’ve provided in the below table, can provide a better sense of the potential magnitude of the inflation protection impact:
|Historical Inflation Rate||% Of Market Tech||% of Tech Inflation Resistant||Historical Avg PE||Adjusted Fair Value PE||PE Difference|
As you can see, the range is considerable. In terms of an adjustment to PE ratios, the low end of the assumptions call for almost no adjustment, and at the high-end as much as an additional 2.69 points of PE. E.g. a fair value PE of 16 would increase to 18.69.
Reasonable people will disagree about the appropriate value for these assumptions and the impact of tech and its relationship to inflation generally. Recognizing that, and the limitations of this simple framework, my hope is that it is at least a useful starting point for appreciating that even this esoteric dynamic could have a non-negligible impact on investors estimates of fair-value.
The intention of this article was not to suggest that the relationship between technology or inflation should be a principal determinant of stock market valuation. To the contrary, in my opinion, the impact of technology’s relationship with inflation on the overall valuation of the stock market is likely dwarfed in comparison to other factors such as, profit margins, tax and interest rates.
Instead, my goal has been to illustrate the potential of a relationship between two infrequently associated aspects of the economy, and thereby underscore Buffett’s wisdom on the danger of anchoring ourselves to historical metrics in an increasingly dynamic world.