November 28, 2017

The Theory of the (Tech) Firm



Yesterday, I had the pleasure of speaking on a panel at the American Enterprise Institute, which asked the question, “Should Washington break up Big Tech?” Antitrust, unsurprisingly, was at the fore of everyone’s mind. But the title of the panel suggests that there’s a fundamental problem associated with “Big Tech”—a problem that antitrust may (or may not) be able to solve. So what is the problem?

In general, the current critiques du jour of American tech firms seem to boil down into one of these categories:

  1. Tech firms are too big, too, powerful, and too few;
  2. Tech firms are inhibiting competition; and
  3. Tech firms are destroying the institutions of liberal democracy.

Unfortunately, none of these criticisms holds up to scrutiny.

To start, concerns over the size of tech firms are entirely overblown. It’s true that the level of industry concentration (the presence of fewer, more dominant firms) across many sectors of the American economy has increased in recent decades, but whether this is problematic and why it is happening remain unclear.

One theory, proposed by James Bessen in a recent paper, suggests that much of the increase in firm size can be attributed to greater effectiveness at implementing proprietary IT systems. Unlike generic computing, such systems are far slower to diffuse to rival firms. As a result, these disparities in effective implementation give rise to heterogeneous firm productivity and access to skilled labor (especially if many of those skills are learned on the job with firm-specific proprietary technology). This, he argues, widens the productivity gap, leading to more efficient and consumer welfare-enhancing outcomes emanating from larger, more concentrated industries.

Part of the concern over firm size also goes to the problem of rent-seeking and the concentration of political power. Large corporations can have an outsized influence on high-level political actors and use that power to maintain their dominance in an industry, often by lobbying for the erection of regulatory entry barriers. However, there is no indication that the dominant tech firms have done anything to diminish the competitiveness of the Internet landscape through such means. In fact, quite to the contrary, many of these same companies regularly lobby against legislation that would undermine the openness and contestability of the market in which they operate (the concerted effort to oppose sweeping changes to Section 230 of the Communications Decency Act is just the most recent example of such efforts). Of course, it is certainly true that such opposition invariably benefits the large dominant firms; but it also works against attempts to close-off the Internet to new entrants.

Ultimately, we should be far more concerned about the behavior of firms, not their size or structure. To that end, so long as a market remains contestable, entry barriers are minimal, and the firms enhance consumer welfare, the size and quantity of companies in a given industry don’t matter much.

This leads to my second point: the idea that competition in the technology sector—and the online service provider segment of that market—is stifled by the presence of dominant firms is entirely unfounded. Determining the level of competition in the tech sector can be difficult, in part because services are generally provided at a price of zero in exchange for consumer’s data. So what’s a good means of assessing competitiveness in the absence of clear prices that might allude to symptoms of excessive market power?

I’d argue a good starting point would be examining expenditures on R&D. Investment in R&D is essentially a firm placing a big bet today in the hopes of very uncertain future returns—the type of bet a firm is unlikely to make unless it is concerned about retaining, or expanding, its market position. In its 2017 Global Innovation 1000 study, PricewaterhouseCoopers makes it pretty clear which industry dominates R&D spending: the American technology sector.

The R&D spending of the five biggest tech firms (Amazon, Alphabet, Microsoft, Apple, and Facebook) ranges from $1.6 billion to $16 billion, with R&D intensity (the ratio of R&D expenditure to total revenue) ranging from 2.16 percent to 27.5 percent (Amazon is the pack leader in absolute spending, while Facebook leads in R&D intensity). In both absolute spending and intensity, American tech firms spend far more on R&D than any other industry. By contrast, the top R&D spending by utilities, which are commonly cited as the quintessential examples of non-competitive monopolies, capped out at $650 million, with an associated R&D intensity of 0.75 percent. It’s also worth pointing out, that of the utility firms that managed to eek their way into the PwC Global Innovation 1000, not a single American utility firm made the cut (the $650 million number was for Électricité de France S.A.).

Another good metric would be looking at barriers to entry relative to other industries. While entry costs to effectively competing with Google or Facebook may include things like access to advanced data analytics and AI, large quantities of consumption pattern data, and scalable network effects, none of those are barriers are any more anti-competitive than the initial capital requirements for building a new factory or investing in inventory stock for a new bodega. What the Internet offers, which other industries simply cannot match, is non-capital intensive access to the market, with bare minimum requirements amounting to little more than a computer and Internet access. In that sense, the Internet, as a broad market, is not marred by the types of stifling barriers to entry that characterize other economic sectors.

If R&D spending is high, and barriers to entry are (and remain) low, it’s fair to conclude that the market is competitive, and doesn’t unnecessarily burden new entrants. The key here is that the contestability of the online service market remains high. So long as the market constantly threatens to unseat incumbent firms, concerns of durable and everlasting monopolies in the tech sector are unjustified.

Third, criticisms over the role that social media played in undermining American democracy in the 2016 election are wildly overstated, and completely ignore more fundamental socio-cultural (and political) problems. Here it’s worth reiterating something that people seem to forget, or miss entirely. The Internet is not divorced from culture and society; it is a reflection of all those things we commonly associate with the “real world.” As I discussed in an old blog post:

The Internet … mirrors society and, by extension, each of us: our preferences, associations, and world views. In a sense, our online tools are just means by which we transpose our lives into the digital realm, and those tools, wonderful though they may be, are just that: tools that we can use to shape our online bubble. Google may be great at allowing us to access information we previously did not know, or refutes our long-held beliefs. But there’s a chasm of a difference between being confronted with evidence suggesting our perspectives are ill-informed, and internalizing that information and readjusting our beliefs accordingly. Simply having access to more information doesn’t necessarily make us better informed.

Criticisms of Facebook in the wake of the Russian election interference are pretty easy to come by. But charging the company with contributing to the degradation of American political institutions by providing a platform from which people can express themselves is a wildly incongruent perspective. It’s odd to make the claim that our democracy is undermined by a service that is premised on vibrant and robust expression and interconnectivity, when those are the very values that characterize the open and inclusive society America has always billed itself as. The tech sector may make for a tempting target of scorn in the current political climate, but leaving the blame for dwindling institutional trust at “Big Tech’s” doorstep is a woefully unsatisfying answer to the question of, “how did America get to now?”

Last summer, Jonathan Rauch’s appropriately-titled Atlantic article, How American Politics Went Insane, argued that part of the reason for declining trust is in part a symptom of the breakdown in party machines and political networks—the American people abandoned the establishment, not the other way around. “The biggest obstacle,” he argues, “is the general public’s reflexive, unreasoning hostility to politicians and the process of politics. Neurotic hatred of the political class is the country’s last universally acceptable form of bigotry. Because that problem is mental, not mechanical, it really is hard to remedy.” I think this is broadly correct. The point, however, is that claims of “Big Tech” undermining American democracy are overly-simplistic and incomplete answers to larger and far more complex issues.

I discussed these and a lot of other issues with my fellow panelists at AEI. If you’re interested, feel free to check out the entire event below.