Learning from the Death of the Afternoon Paper

In this issue:

• Learning from the Death of the Afternoon Paper

• Navigating Bubbles

• Monopoly and Monoculture

• US/China Relations: Unchanged

• How Vaccines Get Made

• The Container Shortage

The news business's painful transition from print to digital has been astoundingly well-documented. People in media are very diligent about reporting on media, at least when there are forces outside of their control. It's a bit like trend pieces, where anything three separate people in Brooklyn are doing is now an unstoppable global phenomenon. Twentysomethings in Greenpoint are to culture news what mice are to health news.

So Facebook's overestimate of video view time ($, WSJ) (and a precisely offsetting underestimate of ads' conversion rate relative to view time) got reams of coverage, and was even blamed for the evaporation of businesses like Mic and Mashable. (Left unasked: how did the accuracy of Facebook's buggy view-time metric compare to display ad views, Nielsen ratings, or whatever ad view count is imputed from a print media outlet's circulation stats?)[1]

The media shifts happening today—text to video and audio, ad-supported to subscription, newsroom to solo owner/operators—are not the first time the industry has been completely changed. TV and radio hurt the print news business, and cable TV and AM talk radio hit them even harder. Legacy media companies have been dealing with competition since Gutenberg.

The period from roughly the 50s through the mid-80s looks like the golden age of newspapers. It was certainly a period where valuable companies like News Corp, Capital Cities, Gannett, the New York Times Company, the Washington Post Company, and Knight-Ridder profited handsomely from owning papers. Warren Buffett famously minted money both by owning newspaper stocks and through direct newspaper ownership (he bought the Buffalo News for $33m in 1977; a decade later, its pretax profits were $39m).

But that long bull market in newspaper assets was actually a painful transition. The single biggest category in the newspaper business at the start of that period had almost completely died out by the end. In 1950, 60% of daily newspaper subscriptions were for afternoon papers, not morning papers. Evening papers were larger until the mid-80s, and by the early 2000s, they were only 13% as large as morning papers.

The separation of morning and afternoon papers was a convenient way for competing newspapers to split the market. In 1910, more than half of the cities in the US with daily papers had two. By 1971, it was 2%. Since then, the list has dwindled to a handful (since that article came out, at least two of the papers profiled have shut down).

Evening papers faced two problems:

1. There was obvious competition from newer media. TV stations geared their early-evening programming to the same kinds of news content papers had, and video is easier to passively consume than text.

2. Generally, blue collar jobs have earlier—and stricter—start and stop times than white-collar jobs, so the death of the evening paper partly reflects the decline in manufacturing employment over the same period.

But fortunes were made while the industry's biggest player was in terminal decline, for a simple reason: one-paper towns were a great monopoly business. They were, in fact, such a good business that after Buffett bought the Buffalo News, he was sued by a competing paper for launching a cheap Sunday edition. From that lawsuit, quoting Buffett:

In the case of monopoly papers... The price hovers right around three times gross revenues... A competitive paper is an entirely different animal.

The News went for 80% of gross revenues.

In general with regional monopoly businesses, the number that can be supported by the market is a floating-point number, while the number of businesses meeting that demand is a positive integer. Running one of two newspapers in a town that can support 2.5 newspapers was great; running one of them when the town can support 1.5 was terrible. But funding one of two newspapers in a 1.5-newspaper town until the other one shuts down was a wonderful business decision.

And, as it turns out, it wasn't just good for the cold-blooded capitalists who were buying at 0.8x revenues and selling for 3x a higher revenue number. It was good for employees, too; newspaper employment rose at a healthy pace until the late 80s, when it peaked and then plummeted.

This puts the decline of the news business in a broader perspective. It's been a tough business for about a generation, although the NYT's recent performance has shown that the subscription model can work, at least as long as it can be be amanuensis for the Tweeter-in-chief. But it's also a business facing a tough comp; rarely in business history has an entire category of company stumbled into a monopoly position because by outlasting its direct competitors against its indirect competitors. The newspaper business had a randomly good run, produced some great writing, exercised immense cultural influence, and put some people on the Forbes 400. Generally, the newspapers that survived were better than the ones that failed, and they reinvested their monopoly profits in better coverage—the world record for the longest newspaper in history was set by the NYT in 1987, quite late in that cycle.

But it's good to be suspicious about narratives, particularly from people who manufacture them for a living. And the people who joined the newspaper industry when times were good assumed that this was the natural state of the market. It turned out to be a fortunate golden age driven by factors outside of their control. And, for the same reason, it ended.

[1] The Wikipedia article on the "Pivot to Video" shows many reports of pivots and subsequent layoffs, but they seem to mostly antedate Facebook's bugs. Since the bug was revealed in September 2016, it might have been responsible for the failure of Mashable's video shift in May of 2016, but not Mic's decision to start focusing on video a year later.

Elsewhere

Navigating Bubbles

Verdad Research usually produces fine quantitative work, but this time they've done something more qualitative: assessing when the 90s started to look like a bubble (and then, back to the usual format, quantifying what happened next). As it turns out, the 90s bubble was obvious—even before the Netscape IPO:

"I think we're approaching a blow-off phase of the U.S. stock market," Ray Dalio told Pension & Investments in 1995. "Price acceleration on the upside is preceding a significant correction—20% beginning over the next 18 months." Peter Lynch echoed Dalio’s concerns in an article in Worth Magazine in 1995, warning that "not enough investors are worried."

Something similar happened ahead of 2008: searches for "housing bubble" peaked in the summer of 2005, and skeptical research on the housing market started coming out a few years before that. While the long-term results from betting on a bubble are generally poor, the realized return in dollars from betting against one is also not great; plenty of people make their bet at one level of assets under management and then get their vindication when they're managing a fraction of what they had before. Clearly, current valuations are stretched, but that's also how things looked to some smart people in 1995, when the S&P had hit the breathtaking level of 500.

Monopoly and Monoculture

When AWS stopped providing service to Parler, one argument went like this: AWS is hardly a monopoly; they only have 32% market share! Surely Parler can go somewhere else. Technically, that argument was true, but it ignores the dynamics of how companies choose who to work with, and what effects it has: AWS couldn't just say "Parler is slightly outside the bounds of what we tolerate. It was basically a coin-toss, but we decided not to work with them." That's guaranteed to produce blowback. Instead, AWS carefully documented all the most egregious Parler content they could find, and used that to argue that Parler was far beyond the bounds of acceptability, and that they had no choice but to stop serving them.

This does not exactly make anyone at Azure enthusiastic about scooping up some market share. (Parler, for what it's worth, did find hosting in Russia. Edit: DDoS protection, not hosting.)

The same story played out in alternative accommodations. Airbnb chose not to offer reservations in Washington DC during the inauguration. At first, VRBO continued to do so. But they've also changed their minds.

Terms-of-use violations are often on a continuum. Before Parler's recent round of adverse PR, it was getting attention because its lax moderation invited adult content. That, too, might have been against terms-of-service somewhere. But once the switch has been flipped, it's flipped definitively, and any nonconformist platform needs to very carefully articulate what it allows and why. The default outcome is that once a major company declines to provide service to a controversial user, its biggest direct competitors will follow suit.

US/China Relations: Unchanged

Politicians like to highlight how they differ from one another, to give voters a reason to show up (or mail in, as the case may be). But it's worth noting—as I argued a few months ago—that an adversarial stance with China is likely to continue. Recently, there's been a flurry of comments, from Trump officials who've repudiated Trump and from incoming members of the Biden team, to this effect. H. R. McMaster wants continuity. Anthony Blinken wants a tough line on China and more friendliness with Taiwan. Janet Yellen wants a full array of tools ($, Nikkei) to curb China's behavior. Interestingly, there's been a gradual convergence between the human-rights narrative, the abusive trade story, and the great power conflict theory. The category that affects all of these is technology supply chains.

How Vaccines Get Made

There have been many writeups on specific parts of the vaccine supply chain, from stories about vial shortages to pieces on the complexity of the cold chain needed to deliver vaccines. This piece, via Marginal Revolution, is the most thorough yet, tracing vaccine production from ingredient lists and vats of bacteria to shots in arms.

And in other Covid news, data from Israel indicates that vaccine recipients are unlikely to infect others, at least after the second dose. Unfortunately, there's going to be a very fraught period over the next year where vaccinated people can return to normal, but there's no visible indicator that they're vaccinated. There is not a good way to distinguish, on sight, a frontline healthcare worker who got an early vaccine from someone who's ignoring the pandemic entirely. This is a tension that won't be resolved until most of the population is vaccinated, although I do expect t-shirt entrepreneurs to clean up with "I'm not irresponsible, just vaccinated" gear in the meantime.

The Container Shortage

About a year ago, the first macro concerns about Covid-19 started trickling out. The Wuhan lockdown was a year ago on Saturday, and at that time the most common macro worry about the epidemic was its effect on imports from China. This, as it turns out, still matters. Prices for shipping goods from China to Europe have quadrupled in the last two months ($, FT), hitting record highs. This is an early symptom of a shift that will continue for years: from the early 90s through about 2010, China had, in effect, an infinite supply of labor. Any increase in global demand could be met by increasing manufacturing in China. Now, both for political reasons and for demographic ones (China's working-age population is in decline, and will likely continue to do so for the next generation or more), demand spikes run into real constraints. Today, those constraints are driven by limited shipping capacity—easy to ignore for consumers of high value-to-weight products like electronics, but unmissable for low value-to-weight goods like... weights—but in the future, these constraints will hit other goods and services, too. The long deflation in manufactured goods was the result of several trends that coincided at the same time, but it couldn't last forever. This won't necessarily lead to higher overall inflation, but it's likely to cause more volatility in price levels.