COVID-19 and the TMT industry: What is transient and what is permanent?

See the sub-headline above: Esquire UK says that “the way that we live and work will be altered irrevocably” by the current pandemic. My own research suggests that is unlikely to be the case: even though the global influenza epidemic of 1918-20 had at least 50 million deaths (many more than COVID-19 is expected to have) there appear to have been few or no significant permanent changes in consumer behaviour as a result. Trends that were already in place before the pandemic were disrupted briefly, then resumed their previous trendlines within a year or two.

Last week, Nielsen looked back at a 2017 hurricane to show that when people are trapped inside they watched 40-50% more minutes of TV per day than when they were not housebound, and rapidly went back to normal after the crisis was over, see chart below. But that event was localised and of short duration, so are not really comparable to a global pandemic expected to last months or years. Can we find a better example?

Precedent and some notes:

Yes. The global influenza pandemic of 1918-1920. And there is good US data available. There is one very important difference between the two outbreaks: the current crisis is a combination of pandemic related disease/illness/deaths and a severe economic recession/stock market decline due to the lockdown measures taking place, while the 1918-20 epidemic coincided with a world war and its aftermath: there was no massive economic contraction, see charts below. US GNP actually rose through the pandemic, although it dropped in 1921 and 1922. Meanwhile the stock market was basically flat over the course of the pandemic. (It is also worth remembering that the 1918-20 epidemic came in three waves in the US, with the second wave (winter 1918/19) being the worst.)

Before looking at the actual data, my prediction would have been as follows:

“Human behaviour has a lot of inertia. Most times we change something in the short term due to a war or pandemic or whatever, but we go back to what we were doing surprisingly quickly. But there were likely some big shifts or inflection points a century ago. I would have guessed that adoption of the telephone or electricity would permanently move higher, since post-pandemic people would be interested in technologies that made communicating or living at home safer or more comfortably. On the other hand, given the fear of contagious disease, people would be less likely to go to a movie theatre for many years after.”

Consumer technology adoption:

In February, global sales of smartphones fell 40% year over year, while laptop, desktop and computer peripheral sales are up as millions are now working from home. There were no computers or smartphones back in 1918 of course, but there were things that were “new technologies and services”: telephones, cars, electricity and stoves.

As you can see from the chart below, the percentage of US households who had these four devices/services continued to rise through the pandemic period, which I have shaded. This is in sharp contrast to adoption during the Great Depression, where they slowed or even declined. You can see a similar effect in the 1981-82 recession. (These are not new observations on my part: we’ve known that weak economies have this effect on technology adoption for decades.)

Using a magnifying glass, the growth rate (the steepness of the line) for automobile adoption drops very slightly from its previous trend. To a smaller extent, so do the lines for electricity and (to an even smaller extent) telephones. And for cars, electricity and phones, the growth in adoption returns to its previous trend line quite quickly after the pandemic was over[1]. But, despite my initial expectations, did NOT shift higher.

My first conclusion is the current slowdown in smartphone sales is primarily due to the current economic issues, with only a small effect from the pandemic itself. The uptick in PC sales is due to work from home demand, and will revert to the previous trend once those are over. Once the pandemic is over (and the recession) I would expect annual sales of most consumer devices to return to their previous trends.

Pandemics and advertising:

Although social media usage is up 40-60% during the COVID-19 epidemic, that traffic is not translating into ad revenues: several companies have already pre-announced shortfalls (although not the exact magnitude yet) and one analyst estimates that the annual impact will be a decline of about 20-30% in ad dollars for the digital ad companies, and an 11% decline for the overall US ad market, to $212 billion, or approximately 1% of US GDP. Another estimate suggests a 12% decline for linear/traditional media, with digital actually up 4%, and strong rebound in 2021.

What happened to ads a century ago during the influenza pandemic? To be clear: those were the very earliest days of measuring advertising in the US. We have a good data series starting in 1919, and I have had to stitch together data from a different source for the war years, which is not ideal, but better than nothing. As you can see in the chart below, the impact of the pandemic on ad spending appears to be minor, while the Great Depression was much more significant. Part of the growth in ad spend in absolute dollar terms in 1918-1920 would have been due to post war recovery, but even so: the pandemic itself did not cause a material drop in ad spending in either dollar terms or as a percentage of GDP.

My second conclusion is the current slowdown in ad spending is primarily due to the economic issues, with only a small effect from the pandemic itself. And that once the pandemic is over (and the recession) I would expect things to return to their previous trend of long term decline as a percentage of GDP, although still showing some annual growth in dollar terms. [2]

Ups and downs – communications and media:

So far in March 2020 we have seen some big changes in use of media and communications services. Streaming video of all kinds (from Netflix to Twitch to HBO Now) is up 20-70% compared to last month, traditional broadcast TV viewing was up 8% in the US (news was up 40-130%) and 17% in the UK (daytime TV audience was up 29%)…but streaming music is down about 8% and podcast listening is down 10%. Children’s books and games are up 40% in the last week. Internet use is up 50-70%, and congested networks are seeing speeds drop on both fixed and mobile broadband networks in the last week or two.

The 1918-20 pandemic saw a similarly mixed bag. All data for the charts below is from the invaluable chapter on communications from the US Census report Historical statistics of the United States, Colonial Times to 1957.

·        As mentioned above, telephone penetration continued to grow during the pandemic, although growth was a little slower than in the periods before and after, see chart below. The Great Depression has a significant impact on phone adoption, but from 1907-1929 the blue line is remarkably straight.

·        In terms of telephone usage, the average number of daily local calls grew relatively consistently at about five percent annually, went negative in 1919 and 1920, and then resumed its previous growth rate (but did not accelerate) until the Depression, when it fell sharply.

·        The telegraph data series only begins in 1917, so it is hard to see a clear historical trend, but it would make sense that they would be high during war-time and the pandemic, and drop off a bit in 1921 when both were over, and then returned to relatively steady growth out to 1929.

·        It’s odd to think of postage stamps as a good way to measure communications technology, but in 1918-20 mail would have been an important part of how Americans communicated. The data below shows an uptick in demand in 1919, a decline the following year, and by 1921 annual growth resumed its previous growth trend, broken only by the Great Depression.

Film and live theatre were both strongly affected by the influenza epidemic. Although not all theatres were forced to close, many were, usually for 6-7 weeks. The closure of so many venues brought film production nationwide to a virtual standstill. I can’t find any numbers for admissions to either films or live theatre before and after the epidemic…but according to one article “It was like you flipped a switch. Businesses and theatres opened up again.”

My third conclusion is the current changes in things like streaming video and music, podcasts and traditional TV are temporary, and primarily due to working from home, lockdowns and less commuting. Once the pandemic (and recession) is over I would expect things to return to their previous trend.

Conclusion:

To my surprise, I found no significant permanent effect or inflection points (up, or down) on a variety of devices, services, and media due to the 1918-20 pandemic. Given that COVID-19 appears to be a less severe pandemic, I expect even less of an effect to occur this time. Could I be wrong, and why?

“It’s different this time.” Always possible, but in my experience when people say it’s different this time…they are almost always wrong. Remember 9/11, and the predictions that it would would change the airline industry forever? It certainly depressed Americans’ propensity to fly for 2-3 years, but after that it went back to growing annually at about the same rate as before, until the 2009 financial crisis. Then it slowed again for 2-3 years, before resuming growth at about the same rate again.

A bigger wild card could be the various shelter in place and lockdown initiatives.

As we try to flatten the curve, we are simultaneously conducting a real-time experiment across millions of companies and hundreds of millions of employees about just what massive work from home (WFH) looks like. If enough companies and/or employees like it, we could see a significant and permanent growth in the number of people who work from home compared to the past: 5.2% of Americans worked entirely from home as of 2017.

WFH (and the associated changes in commuting patterns) does seem to materially affect media consumption, internet use (both mobile and fixed), device sales, and also cloud services: Microsoft announced this weekend that “we have seen a 775 percent increase of our cloud services in regions that have enforced social distancing or shelter in place orders.”

Will it happen? I have heard anecdotes both ways: some say working from home makes them appreciate the office more than ever, others are posting cartoons and memes.

There was no equivalent of WFH in 1918. But there is a possible proxy: fleeing the city.

In the summer of 1665, there was a mass exodus from London during their Great Plague. In 2020, many Parisians went to the French provinces just before the lockdown took effect. And “in 1918, the virus was regarded as an urban problem.” I expected to see a change in the percentage of Americans who lived in cities following the pandemic. As you can see in the chart below, although there may have been some temporary population shift to the country as a result of the pandemic (the data is measured only once per decade), it was temporary: the line from 1910 through 1920 and to 1930 is completely straight, and is only broken by the Great Depression.

Over and over, severe and prolonged economic depressions seem to cause significant changes in the tech, media and communications industries…pandemics don’t. It’s clear that Q1 and Q2 are going to be economically weak…but if we can snap back in the back half of 2020, I expect the TMT industry of 2021 to look much as it did in 2019.

[1] The data I am relying on is annual. The 1918-20 pandemic was seasonal, so it makes sense to me that there could have been a more pronounced negative effect on adoption of these devices/services of short duration, followed by a bounce back when the influenza was less active: individual months or quarters could be bad, but the year as a whole could still show growth.

[2] As a note, ad spending as a percentage of GDP has been declining for years now; it is believed the growth in digital advertising is behind that trend.