Do technology stocks have consistently higher returns?

<<< This blog post is also available as a YouTube video. >>>

From Daniel Chancellor  and  Gerd Kommer  

In recent years, investors have achieved higher returns from technology stocks than from “old economy” sectors such as transportation, energy, engineering, trading, finance or the overall stock market. A number to illustrate: Over the five years from October 2015 to September 2020, the global information technology stock sector produced a return of 23.1% p.a. compared to 9.4% p.a. for the general world stock market (MSCI ACWI IMI Information Technology Index and MSCI ACWI IMI Index, nominal returns each in euros). Technology stocks also weathered the Corona crash in the first half of 2020 better than the rest of the market.

If a sub-segment of the stock market - e.g. If a country, an industry or an investment strategy – delivers a high excess return relative to the overall market over several years, some private investors conclude that this excess return is “normal” and will continue in the future.

In this blog post we will show that the idea of ​​the technology sector's systematically (statistically sufficiently reliable) higher returns is an old misconception that pops up again every few years, like a zombie that doesn't want to die. The zombie last roamed the private investor community in the late 1990s. It temporarily disappeared when the dot-com bubble began to burst at the beginning of 2000. Now he's back. 

The fact that investors could systematically take advantage of the supposedly higher returns from technology stocks was not true then, is not true now and will not be true in the long-term future either. 

Before we get to some numbers that illustrate our thesis, let's first answer the question of why so many private investors fall for the technology stock fallacy.

The most important among several causes is the so-called Recency bias – the all-too-human error of thinking that data from the recent past are fundamentally more important for the future than older data.  [1] Since the high-tech sector has outperformed almost all other sectors in recent years, many investors believe that this is now “the new normal”. 

The false conclusions promoted by the recency bias are further reinforced by investment pornography spread by the financial media, i.e. publications that present scientifically unproven conclusions as something that is indubitable or at least very likely - in each case mixed with the unspoken or stated implication that private investors can reap attractive additional returns from them. A financial pornographic example of this is the article headline "Technology stocks remain the only choice - they shone during the crisis and they continue to do so. Investors can bet on that" (financial portal GodmodeTrader, June 15, 2020).

It hardly needs to be emphasized that the financial industry jumps on every trend in its product marketing that allows gullible investors to take money out of their pockets. Of the thousands of stock ETFs sold in Germany, around a third replicate a sector or theme index. These ETFs have, on average, higher ongoing management fees than more diversified non-industry ETFs. The same applies to actively managed equity funds and to the over one million certificates sold in Germany.

In investor thinking, the “built-in” recency bias, the financial pornographic headlines in the media, the fashion-driven marketing of the financial industry and investment pseudo-knowledge are interwoven into a subjectively rounded, coherent story for the investor: Sectors with strong technical progress and a higher proportion of young “disruptive” companies must inevitably be more profitable for investors than old economy sectors with only slow technical progress and little growth. That is “the logic of digitalization”.

As we will see, this simple story will only be questioned when, a few years later, the level of disappointment in the portfolio statement has exceeded a significant intensity. At this point in time, however, there will already be a new “investment opportunity of the decade” that, in our perception, appears to be just as logical and inevitable as the one with which the investor has just suffered a belly landing. Now the game can begin again. 

But now to some numbers that make it clear that the technology sector does not produce more attractive returns in the long term than the overall market and many ancient low-tech industries.

We use U.S. stock market data for this statistical exercise because it has the highest data quality and granularity for periods extending far back in time, and because it is publicly available to everyone on American economics professor Ken French's website. US data in this case is likely to be representative of the rest of the world as our focus is on the relative comparison of industries.

The following table contrasts the return of the High Tech sector  [2] with other selected industries over the past 50 years. The underlying industry classification breaks down the entire U.S. stock market into ten major sectors. The High Tech sector also includes four of the five FAANG stocks (Facebook, Apple, Netflix and Google, but not Amazon). We benchmark this sector with the overall market as well as with 49 other, more granular sectors, which together form the overall market.

Table: Comparison of returns of the US High Tech (HT) equity sector with the overall market and other selected sectors for the period 07-1970 to 06-2020 (50 years) - inflation-adjusted (real) returns in USD, before costs and taxes

► [A] The decade from 07/2000 to 06/2010. For this analysis, the 50-year period was divided into five non-overlapping decades. ► [B] Microchip manufacturing is one of four industries in the 49-industry segmentation that make up the “High Tech” sector in the 10 classification. ► Telecommunications is not part of the high-tech sector here. However, including TK would not significantly change the numerical results in the table. The “notorious fall” of the Telekom share in Germany is probably atypical for the share returns of the telecommunications sector worldwide. ► Overall market = CRSP 1-10 index. ► Data source: Ken French Data Library.

What are the main conclusions that can be drawn from the numbers in the table?

(a) Yes, over the last five (or even seven or ten) years, technology stocks have beaten the general stock market in the US (and worldwide).

(b) Over the entire 50-year period, the high technology sector produced a return that was almost exactly the same as the general stock market. However, the HT sector had a noticeably higher risk than the overall market, namely one and a half times as much volatility (yield fluctuation) and a maximum cumulative loss (maximum drawdown) of minus 81% in September 2002 compared to minus 55% for the overall market in September 1974 (both values ​​adjusted for inflation). “Risk-adjusted”, high tech was therefore worse than the overall market over 50 years. The best shareholder return among the 49 industries was in weapons manufacturing, and the second highest was in tobacco.

(c) If you divide the 50 years into two halves, high tech performed noticeably worse than the overall market in the first half and better in the second half. But even in these second 25 years, HT only finished in the top mid-range of returns of 49 industries and was almost overtaken by the “old” shipbuilding industry on the 25-year “middle route”.

(d) If you look at the worst decade for HT among the five non-overlapping 10-year periods in the 50 years under consideration, HT is again not on the winning side: 9.6% average loss per year from July 2000 to June 2010 versus only 3.1% loss per year for the overall market. HT stock performance this decade appears even more dismal relative to the high 10-year return of the “stone age” coal mining sector. It should be mentioned here that this weak decade for HT shares was not that long ago.

Interim conclusion: If you include a longer period of time in a data analysis, instead of just the last few years selectively, then a “slightly different” overall picture emerges than what technology stock fans like to convey.

It is sometimes claimed that there has been some kind of structural break in the stock market recently, which is why data longer than e.g. B. ten years ago, are no longer relevant with regard to technology stocks. This argument, which has never really been proven, seems comically naive. This is clear from the fact that the buzzwords, concepts and technologies used today to explain the high returns of the high-tech sector in recent times - Internet, digitalization, network effects, artificial intelligence, scaling and so on - all existed 20 or 30 years ago. Konrad Zuse built the first computer in 1936. 

The question “what’s really new?” indirectly leads us to an explanation of why HT shares have performed averagely historically and why this will probably continue to be the case in the future. However, seeing and understanding this real explanation requires two things: 

• When it comes to empirical data, one must make an effort to look beyond the selective short-term data from recent years and past the story of individual stocks that the financial industry and media typically disseminate.
• When it comes to factual logic, you have to stick to old lady science and not to simple-sounding stories and gut-feeling economics. 

From a scientific perspective, some important factual facts on this topic look like this: 

Sectors have not been identified as systematic drivers of return and risk for stocks in empirical financial market research over the past 60 years. The only “systematic drivers” of return and risk are the so-called factor premiums (see our blog post on this Factor investing – the basics from May 2019). Understanding these is of course less exciting than the colorful stories about the latest industry trends, sensational product innovations in consumer electronics and famous technology celebrities such as Steve Jobs or Elon Musk (see our blog post on this The best stock in the world from June 2020).

Of course, the industry a stock belongs to has an impact on its return. This is trivial. However, mere influence does not help investors if this influence is not systematic in nature, i.e. cannot be “exploited” in the future. Of course, companies in the information technology sector have better growth prospects than most other sectors. That too is trivial. 

However, these better growth prospects are already priced in at any given time because the entire market knows them and simply assumes them for the future. You will therefore no longer have a systematically exploitable return advantage for the shares in question moving forward. One indicator of this pricing is the high valuations of technology stocks today. For example, the price-to-earnings ratio of the five FAANG stocks at the end of September 2020 was approximately 61 (simple average). That's around three times higher than the P/E ratio of the overall market. This essentially means that the market expects business figures for the five companies in the future (in total) to be three times as good as for the market as a whole, so that the stock return will only be the same as that of the market as a whole.  [3] 

However, to produce higher shareholder returns than the overall market in the future - as FAANG investors expect - the five companies' earnings must grow faster than just three times the market. Three times as much is already priced in. 

In our opinion, if this extreme growth does not occur in the future, no one should be surprised because it has already happened in the past. A multi-year under-return in the technology sector is not unusual, as our table shows.

Since we all love stories, let's end with the story of the “yield race” between Walmart and Intel. Walmart is a boring supermarket retailer, albeit the largest in the world by sales. In supermarket retail, margins are notoriously razor thin and technical progress is rather slow. Intel is the world's largest microchip maker, a global technology leader that owns 40,000 patents and has high margins.

Intel's publicly available stock price data goes back a good 40 years to February 1980; for Walmart, the available data series is longer. In the 40.6 years to September 2020, the old economy stock Walmart generated an impressive return of 20.7% p.a. versus “only” 14.6% p.a. for Intel (nominal in US dollars including dividends, before costs and taxes). Walmart has also been ahead over the last five years and the last 20 years.

Conclusion

The opening question of this blog post was “Do technology stocks have consistently higher returns?” As we have now seen, historically this has not been the case. However, in order to recognize this banality, one must not Data snooping i.e. selectively choosing only short periods of time and ignoring other evidence that does not support one's own thesis.

Regardless of whether you break down the world stock market into ten, 49 or 100 sectors, there are always periods of five or more years in which individual sectors significantly outperform the overall market and individual sectors for which the opposite is the case. The more you unravel, the greater the differences between the industries become apparent. It had nothing to do with high tech versus low tech.

Can these inevitable phases of outperformance of individual sectors be reliably identified in advance? No. Even the few real investment superstars that still exist today - most notably Warren Buffett - don't claim that; neither does science (Bessembinder 2020). 

Anyone who, from the perspective of this science, is looking for a sensible return-risk combination for their assets in the future is ill-advised to go this route Sector picking to try. It is smarter to have all sectors in your portfolio.

Endnotes

[1] In our blog post Ten big investor mistakes In September 2020, we addressed the recency bias and eight other common investor mistakes.

[2] “High Tech” is the term in the Ken French database for what is typically called technology stocks/technology sector in industry jargon.

[3] The American stock sector Information Technology had a P/E ratio of around 33 at the end of August 2020 (MSCI USA Information Technology Standard).

literature

Bessembinder, Hendrik (2020): “Extreme Stock Market Performers, Part II: Do Technology Stocks Dominate?”; July 22, 2020; Internet reference: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3657609