The American Banker article yesterday on our new fintech paper has sparked a good deal of discussion, some aimed at my dark view of giant platform-company monopoly risk.  The paper readily acknowledges the smarter, faster, and even better advantages of artificial intelligence over legacy systems.  But, for all the merits of the smarter, faster, better arguments, there’s also a dark side to giant fintech platforms due to opacity, models so complex that even company CEOs are stumped, and enormous, concentrated power over the critical infrastructure of the global economy.  Commodore Vanderbilt, Jeff Bezos, and Mark Zuckerberg have a lot in common, as a quick look at the lessons America learned the hard way about monopoly in the 19th century makes all too clear.

As a thoroughly readable, but authoritative biography of Cornelius Vanderbilt tells it, the early railways were an amazing hodge-podge of different standards.  Tracks were laid with whatever gauge an investor – read speculator – liked, with railway carriages then built to ride each type of track and only that type of track and hook up to its type of carriage and only that type of carriage.  The result was railways that were far faster and better than horse-drawn transport, but only until you got to the end of one short line and needed to move on to another.  A trip from New York to Washington thus took almost as long, once connections were taken into account. Steam ships – not railways – remained the dominant transport mode until about the Civil War.

The man who liked to be styled a commodore was already making his own railways (e.g., from New York to Boston) not just smarter and faster, but also better by virtue of what we now call seamless connectivity and inter-operability.  However, Vanderbilt did not scruple at what we now call rent-seeking – that is, as he came to control transport, he also controlled the pricing of what went on it.  Very quickly, farmers and manufacturers were forced to pay his price, no matter its cost.

With the money he quickly amassed from rail lines and steam ships, Vanderbilt proceeded to gobble up other railways.  From this, he created an efficient, increasingly monopolistic rail network across the East that, as the country expanded, led to other rail lines (e.g., in Chicago) owned by others we would now call oligarchs.  Many of them started their railways after Vanderbilt and emulated him in engineering and rent-seeking, making them quickly Vanderbilt’s match in influence and wealth.  Vanderbilt was thus thwarted in his ambitions to control all of America’s railroads, but so were the other railway owners when they hit the lines he controlled.

Solution?  Something we now call collusion.  After a period of economic warfare in which Vanderbilt tried to take out rival oligarchs and they did their best to outdo him, a truce was called by J.P. Morgan.  You will not be surprised to learn that Morgan did not act out of pity for small-time farmers and manufacturers.  He was instead interested in creating a monopoly coast-to-coast transport system offering favorable pricing only to his own ventures.  This he got, firing up even more protests.

Repeated financial crises sparked sometimes by speculative railway finance and even more often by speculating railway oligarchs at the end of the 19th century powered up the progressive movement.  But by then the railroad empires were – another 21st-century word – scalable.  That is, the antitrust solution – e.g., breaking up Standard Oil into regional component parts – would have recreated the slow, inefficient, broken-up railways that policy-makers remembered all too well.  Instead, railways were subject to a new invention – the Interstate Commerce Commission – and to top-down federal price regulation.

Which brings me back to fintech.  Our new paper shows that an antitrust solution will not work for the giant platform companies except in areas in which opacity is a risk and scalability is not an efficient benefit of vertical integration.  Like railways over a century ago, cloud computing and many other AI functions are extraordinarily efficient due largely to scale.  Breaking giant platform companies back into small bits will thus undermine the many benefits genuinely brought alongside increasingly overt rent-seeking by our new-age oligarchs.  Instead, we need to reconstruct the successful solution to 19th-century railway monopolies to make it fit for purpose for gigantic, unregulated, and infrastructure-essential financial products and services.

Was 19th century utility regulation an unmitigated success?  Of course not.  Legislation in the 1930s and considerable changes since sought to mold Progressive-Era cures to new market-integrity ailments.  However, the lesson then is more than instructive now:  unmitigated market power may be smarter, faster, and even better than what went before.  But, with strength comes power and power used for personal profit puts the public good at grave risk.

How much do we want one company to know about what we do when and with whom where?  How safe is a financial system predicated on just two or three companies controlling critical infrastructure?   We learned about systemic risk the hard way in 2008.  Are we ready for redux? Is it as efficient – let alone fair – for a company to pick financial products on which our economic equality depends as it is for Alexa to micro-target a song? 

Policy-makers have their qualms about resilience, opacity, and oligopoly risk, but few are willing to advance solutions beyond still more monitoring.  I fear that, by the time all of them agree that there’s a bit of a problem, we’ll be riding the same railroad at rigged prices to a destination someone else has chosen for us.