Expert knowledge is contingent and useful forecasting constrained by the interplay of complex systems

I recognize but am dubious of the use of government power to prevent the emergence of monopoly markets.  I am fine with regulations but am unconvinced of the ability of the government to either understand or effectively forecast, plan and manage markets, especially rapidly evolving markets.

This skepticism about anti-monopoly actions is closely related to my skepticism of "experts".  There are certainly experts with narrow and deep domains of knowledge but almost all complex problems are a function of the interaction of multiple loosely coupled evolving non-linear chaotic systems.  Single domain expertise is insufficient and multi-domain expertise domains rare or non-existent.  

As illustrated in this reporting from T-Mobile Proves That Mergers Can Benefit Consumers by Thomas W. Hazlett.  The subheading is That should give pause to today’s overzealous antitrust enforcers.  

The government has become increasingly suspicious of major mergers over the past decade, under both political parties. The Justice Department under Donald Trump sued to prevent AT&T from buying Time Warner. The Federal Trade Commission under President Biden is continuing a case the Trump administration initiated against Meta, parent of Facebook, to force the firm to cough up Instagram and WhatsApp, which it swallowed during the Obama years. In January JetBlue Airways’ plans to merge with Spirit Airlines and Amazon’s plans to acquire iRobot were deterred under regulatory pressure.

In April 2020, however, T-Mobile and Sprint managed to sneak past regulators, merging to reduce the number of major U.S. mobile networks from four to three.

[snip]

T-Mobile’s takeover of Sprint was controversial among analysts. “If this merger is not anticompetitive,” Eleanor Fox, a trade regulation and antitrust law professor at New York University, told reporters in 2020, “it is hard to know what is.” Yale economist and antitrust scholar Fiona Scott Morton delivered her verdict on the deal in a co-authored 2021 article: “The era of aggressive price competition in wireless is over.” The authors predicted that the wireless industry, whittled down to a big three, would “nestle into a cozy triopoly.”

Respectable and credentialed experts from the most prestigious institutions made forecasts which could be validated.  

And they were wrong.

The prediction proved wrong. Average monthly mobile subscription fees dropped sharply. In the three years before the merger, according to government price data, mobile charges declined in real terms by about 8%. In the three years following the merger, the real price decline has been nearly 12%.

These trends were even more impressive given dramatically improving network performance. Before the merger, the top four U.S. carriers delivered data download speeds averaging about 26 megabits per second, nearly all via 3G or 4G. By early 2023, with 5G deployments spreading, Verizon and AT&T data flowed 24% to 39% faster, while T-Mobile was more than three times as fast as before. T-Mobile’s high-speed coverage had also expanded; half of its connections were via 5G by January 2023, against just 10% to 20% for its rivals.

T-Mobile’s aggressive deployment of the Sprint spectrum rights it had purchased paid dividends in enhanced services for customers. It was also a boon for shareholders. As T-Mobile pulled market share from its rivals, its stock soared. Between 2018 and 2023, T-Mobile shares outperformed the S&P 500 Index by 50%.

Further evidence that the merger of T-Mobile and Sprint was pro-competitive was seen with Verizon and AT&T share prices. From 2018 to 2023, Verizon and AT&T stock prices declined sharply, losing more than a third of their real value. The postmerger marketplace was a great victory for T-Mobile but a blow for its rivals. The cozy-cartel thesis collapsed.

It might have turned out otherwise.  The future is hard to predict.

Which is pretty much the point.  Experts ought to be adjudged and assessed based on the utility of their knowledge as reflected in their capacity to accurately make forecasts.  If they are not able to accurately forecast, it is not necessarily an indictment of their narrow domain of knowledge.  It merely suggests that the larger set of systems are so complex that the narrow domain is insufficient.

The real lesson is not so much that experts are uselss.  The real lesson is that all knowledge is contingent and that forecasts of complex systems are almost certainly going to be wrong and that passionate conviction or even shared consensus has little to do with actual measured reality.