Investing in technological change

The beginning of the end?

A few years ago, a well-known investor uncovered an interesting conundrum within markets, known as the ‘top dog’ problem.^[1] He found that investing in the largest company within the US stock market on a rolling basis yielded terrible returns. Others^[2] have since verified this, creating an index investing in only the top stock on a rolling one-year basis and comparing its total return to the wider S&P 500. The index starts at 100 in 1949 and closes on 4 at the end of 2023! What worked for six or so decades after 1950 was suddenly spectacularly wrong in the last 10 or so years - concentration at the top (the share of the overall US stock market capitalisation commanded by the top 10 companies) has doubled. Corporate titans to rival the previous peak in concentration in the late 1960s trounced all contenders via a range of means. As usual there are many potential explanations,^[3] most only obvious after the fact.

The degree of concentration in the US stock market is certainly rare, as are the valuations and levels of profitability, reflected / highlighted in the unique power (and potentially vulnerability) conferred by huge intangible asset accumulation, certainly an artifact of the moment we are in. As usual, the sensible investor will diversify their portfolio beyond the recent batch of winners. They will keep a foot in the more unfashionable corners of the world’s capital markets where valuations, profitability and expectations are much less of an impediment to prospective returns. It is easily possible that it will be some of these more unloved sectors where the gains from the continuing rapid advances in the technological frontier will show up most vividly.

All the same, this week’s earnings reports from the titans suggest that they are not done yet. The sector’s profits and cash flows continue to surge, as does its capital spending. For the latter, there are understandable concerns about what this implies for future profitability and the ‘asset light’ investment case. However, for now, these businesses remain formidable for the most part. Below we explore some of the things for investors to wrestle with in positioning for the road ahead.

Profiting from technological revolutions

The relationship between innovation and investment returns is often misunderstood. Some will simply look at certain types of companies as the beneficiaries, or perhaps the instigators of, innovation. The technology sector, and some of the sub-sectors within, tend to be very popular. Thematic baskets of stocks purporting to contain directly related exposure to Artificial Intelligence (AI), quantum computing or whatever else grabs the imagination, tend to sell well. Investments organised around themes can help cut through in what tends to be a cluttered and noisy investments shop window – a very helpful trait in the context of time-poor investors trying to grapple with a world often moated by inaccessible jargon.

However, there are dangers here too. A close study of the innovation of the past and its key beneficiaries speaks of the need to cast your investment net wider than these thematic baskets often allow (if held in isolation). Similarly, it is tempting to focus excessively on the ‘sexier’ breakthroughs, such as generative AI or quantum computing, at the expense of the more prosaic. Few were drooling with excitement as the standardised shipping container arrived on the scene, or the Excel spreadsheet. However, these were some of the most transformative innovations in terms of productivity of the last century.⁴ In the case of the former, it is important to remember that it was not the shipping industry that saw the benefit of multiplying productivity, but end consumers in emerging markets among others scattered around the world. The best investment advice would have been to diversify.

How inventions interact with each other- and us - has always been unpredictable. In 1867, the invention of the dynamo allowed for the practical generation of electricity. This, in turn, allowed a separate power source to be attached to each machine in the factories of the time, which in a relatively short space of time allowed Henry Ford to mechanise production with a moving assembly line. On came standardised parts and quickly followed mass produced, affordable cars. The Model T and its successors transformed American (and later European) society. It allowed people to move around more quickly and cheaply than ever before. It provided high paying work to many immigrants who could not easily converse in English. It helped to create the suburb, the shopping centre, the domestic tourism industry, the motel (even the sex lives of Americans were transformed).^[5]

Some thoughts on how that future is currently valued

As discussed in the article on bubbles, they tend to be hard to spot. More often than not, what appears at the time to be irrational exuberance is often revealed later to be a step change in growth prospects. Nonetheless, some measures of valuations, such as Nobel Prize winner Robert Shiller’s famous cyclically adjusted price to earnings ratio (CAPE), sit at historically concerning levels.

CAPE is a valuation measure derived by dividing the price of an index by the average of 10 years of corporate earnings adjusted for inflation. The idea originated from Benjamin Graham and David Dodd’s investing bible ‘Security Analysis’ published in 1934 and was later popularised by Robert Shiller and John Campbell in a 1988 paper. The sensible point made was that one-year forecast or realised earnings could be volatile and could therefore provide an extremely unsteady valuation platform from which to evaluate equity investments. CAPE’s many disciples point to the fact that over the last 100 years, a high CAPE value has, on average, been followed by lower than average 10-year annualised excess returns with the reverse also being true. Case closed?

As always, it is not that simple.

First, smoothing the denominator does make it less volatile, but it also means that almost all of the statistical leg work – the explanatory power - comes from changes in price – the numerator. Meanwhile the bit of information that doesn’t come from price is still capable of jumping about as the effects of major write downs and recessions enter and exit the average.

Second, we are looking at this dataset with perfect hindsight, assuming that the investors of the past knew more than they possibly could have. A more honest ‘out-of-sample’ approach would involve sorting starting valuations by using only past data available to investors at the time.

Perhaps the biggest problem with using CAPE as a tactical indicator is where its practical application would have left you as an investor. The rising trend in the CAPE seen over the last century would have left its tactical adherents persistently underinvested in stocks – always waiting for the market to come back to the mean that they had in mind, but all too rarely finding those moments. Of course, all that may simply indicate that stocks have further to fall when that day of reckoning does finally arrive. However, for those arguing for a reversion to the long-term mean, that wait has so far been extremely costly in performance terms.

Those who would argue that we are on the cusp of a reversion to the CAPE’s 100 year moving average, so, in part, justifying this long-term underexposure, may underestimate many things, from changing sector composition to increasingly unrecognisable accounting standards. However, they may also underestimate the effect of dramatically reduced transaction costs. The price of acquiring and maintaining a diversified investment portfolio has plunged over the last century in particular. This has raised the net returns from stocks and facilitated greater diversification, so lowering the associated risk in equity investment. These are among the factors that suggest that at least some of this higher trend CAPE is justified.

The broader point though is that while the CAPE has reasonable (but not jaw dropping) predictive power with regards to long-term equity returns, its record over the shorter term is weak. This is not limited to CAPE in fact, but most mainstream valuation metrics. Those wisely focused on that medium term can take reassurance from the fact that the level of the dividend (and buyback) yield for developed stocks, allied to a sensible assessment of the prospective growth in that shareholder compensation, still points plausibly to mid to high single digit annualised percentage returns for the medium term, even without further valuation expansion.

The post war history of stocks contains many patterns and trends. Deciding which are relevant to the moment we occupy and which are wasting our time (and money) is one of the most challenging parts of investing. The fact that it’s not just different this time, but every single time, further complicates the task. There are always plenty of overconfident voices telling us to frame the moment in one particular way or another. The more conviction on display, the less you should trust them as Yeats suggested.^[6]

¹ https://ioandc.com/rob-arnott-sell-the-top-dogs/
² https://www.themebfabershow.com/episodes/3g6O04KWI6D
³ https://www.imf.org/en/Publications/fandd/issues/2024/09/america-must-rediscover-its-dynamism-michael-peters
⁴ Robot Macroeconomics: What can theory and several centuries of economic history teach us? – Bank Underground
⁵ Lipsey, R.G; Carlaw, K.I & Bekar, C.T (2005) – Economic Transformations – General Purpose Technologies and Long-Term Economic Growth – Oxford University Press pp.4
⁶ W.B. Yeats – The Second Coming – “The best lack all conviction, while the worst are full of passionate intensity.”