We now understand one of the main reasons that many leading firms have been winning relative to others, resulting in higher markups, profits, and wage inequality:

The biggest companies in every field are pulling away from their peers faster than ever, sucking up the lion’s share of revenue, profits and productivity gains. Economists have proposed many possible explanations: top managers flocking to top firms, automation creating an imbalance in productivity, merger-and-acquisition mania, lack of antitrust regulation and more. But new data suggests that … IT spending that goes into hiring developers and creating software owned and used exclusively by a firm is the key competitive advantage. It’s different from our standard understanding of R&D in that this software is used solely by the company, and isn’t part of products developed for its customers.

Today’s big winners went all in. …Tech companies such as Google, Facebook, Amazon and Apple—as well as other giants including General Motors and Nissan in the automotive sector, and Pfizer and Roche in pharmaceuticals—built their own software and even their own hardware, inventing and perfecting their own processes instead of aligning their business model with some outside developer’s idea of it. … “IT intensity,” is relevant not just in the U.S. but across 25 other countries as well. …

When new technologies were developed in the past, they would diffuse to other firms fast enough so that productivity rose across entire industries. … But imagine instead of power looms, someone is trying to copy and reproduce Google’s cloud infrastructure itself. … Things have just gotten too complicated. The technologies we rely on now are massive and inextricably linked to the engineers, workers, systems and business models built around them. … While in the past it might have been possible to license, steal or copy someone else’s technology, these days that technology can’t be separated from the systems of which it’s a part. … Walmart built an elaborate logistics system around bar code scanners, which allowed it to beat out smaller retail rivals. Notably, it never sold this technology to any competitors. (more)

A policy paper goes into more detail. First, why is the IT of some firms so much better?

Proprietary IT thus provides a specific mechanism that can help explain the reallocation to more productive firms, rising industry concentration, also growing productivity dispersion between firms within industries, and growing profit margins. … There is a significant literature that identifies IT-related differences in productivity arising from complementary skills, managerial practices, and business models that are themselves unevenly distributed. Skills and managerial knowledge needed to use major new technologies have often been unevenly distributed initially because much must be learned through experience, which tends to differ substantially from firm to firm.

Yes, skills vary, but there are also just big random factors in the success of large IT systems, even for similar skills. What can we do about all this?

While there may be other reasons to question antitrust policies, the general rise in industry concentration does not appear to raise troubling issues for antitrust enforcement at this point by itself. …

Both IP law and antitrust law pay heed to … balancing innovation incentives against the need for disclosure and competition, balancing concerns about market power against considerations of efficiency. … This balance has been lost with regard to information technology … the policy challenge is to offset this trend. … This problem might require some lessening of innovation incentives. … The challenge both today and in the future for both IP and antitrust policy is to facilitate the diffusion of new technical knowledge and right now the trend seems to be in the wrong direction. …

To the extent that rising use of employee noncompete agreements limits the ability of technical employees to take their skills to new firms, diffusion is slowed. Similarly, for extensions of trade secrecy law to cover knowhow or the presumption of inevitable disclosure. Patents are required to disclose the technical information needed to “enable” the invention, but perhaps these requirements are ineffective, especially in IT fields. And if patents are not licensed, they become a barrier to diffusion. Perhaps some forms of compulsory licensing might overcome this problem. Moreover, machine learning technologies portend even greater difficulties encouraging diffusion in the future because use of these technologies requires not only skilled employees, but also access to critical large datasets.

It seems that making good backroom software, to use internally, has become something of a natural monopoly. Creating such IT has large fixed costs and big random factors. So an obvious question is whether we can usefully regulate this natural monopoly. And one standard approach to regulating monopolies is to force them to sell to everyone at regulated prices. Which in this context we call “compulsory licensing”; firms could be forced to lease their backroom IT to other firms in the same industry at regulated prices.

Note that while compulsory licensing of patents is rare in the US, it is common worldwide, and it one of the reasons that US drug firms get proportionally less of their revenue from outside the US; other nations force them to license their patents at particular low prices. So worldwide there is a lot of precedent for compulsory licensing.

The article above claimed that backroom IT is:

inextricably linked to the engineers, workers, systems and business models built around them. … While in the past it might have been possible to license, steal or copy someone else’s technology, these days that technology can’t be separated from the systems of which it’s a part.

I’m not yet convinced of this, and so I want to hear independent IT folks weigh in on this key question. I can see that different IT subsystems could be mixed up with each other, but I’m less convinced that the total set of backroom IT of a firm depends that much on its particular products and services. Maybe other firms in an industry would have to take the entire backroom IT bundle of the leading firm, rather than being able to pick and choose among subsystems. But when the leading IT bundle is so much better, I could see this option being attractive to the other firms.

The leading firm might incur some costs in making its IT package modular enough to separate it from its particular products and services. But such modularity is a good design discipline, and a compulsory licensing regime could compensate firms for such costs.

Note that I’m not saying that it is obvious that this is a good solution. I’m just saying that this is a standard obvious policy response to consider, so someone should be looking into it. At the moment I’m not seeing other good options, aside from just accepting the increased IT-induced firm inequality and its many consequences.

Added 12:30: Okay, so far the pretty consistent answer I’ve heard is that it is very hard to take software written for internal use and make it available for outside use. Even if you insist outsiders do things your way.

So assuming we are stuck with industry leaders winning big compared to others due to better IT, one worry for the future is what happens when leaders of different industries start to coordinate their IT with each other. Like phone firms are now coordinating with car firms. Such firms might merge to encourage their synergies. They we might have single firms as big winning leaders in larger economic sectors.