The state doesn’t experience consequences for risk – which means markets are always superior and the state risks crowding out superior innovation

Baumol (2002) argues that the market economy is in effect an innovation machine. By having no choice but to arms race each other, firms are forced to find new niches, improve existing products, and find new markets. That is, innovation becomes the primary weapon in each individual firm’s battle for survival. If the right knowledge is out there, and if the framework conditions are right, an entrepreneur will find it and turn it into something valuable. If that person is successful, more firms will follow suit and further refine the innovation, and of course introduce competing technologies.

A new wave of thinking about growth has come to incorporate entrepreneurship more explicitly. Acs et al. (2009) remark that part of “the endogenous” in endogenous growth theory is the creation of technological opportunities, driven by conscious investments in knowledge. In their model, entrepreneurs exploit knowledge that leaks out from incumbent firms and unto the rest of the economy. An entrepreneur is something like a dynamic conduit for finding, probing, and ultimately spreading knowledge. Acs et al. (2009) show that greater regulation, administrative burdens, and market intervention will reduce the knowledge available to us.

To be sure, this role of private enterprise does not preclude a substantial role for the state; it is not a point of ideology. But the knowledge-carrying function of entrepreneurship implies that intervening with the process comes at a risk. It is certainly possible that by steering firms into more productive, or more socially desirable activities, the state may improve upon things. This much seems to have been known by almost every ruler since the invention of enterprise (see for instance, the exposé in Landes et al., 2010). But this almost self-evidently correct claim—the could—does not in and of itself lead to the conclusion that the state should be actively involved in these things, because we are not living in a first-best world. This would require the state to have some method for picking the right paths among many possible ones.

A key thing to understand is that private sector entrepreneurship has not been invented; it has evolved (Winter & Nelson, 1982). In a cocktail of local and regional institutions, available human capital, and input from markets, small and large firms have butted heads. What remains is, with simplification, either deeply battle-tested, currently challenging the status quo, or funded by investors who are too patient. Actors in social systems can indeed copy behaviors of others and entrepreneurship is no exception (Andersson & Larsson, 2016), so it is by no means impossible for top-down entrepreneurs to do well by learning from the past and the present in similar ways. In fact, the best-case scenario—when one entrepreneurial state figures something out that really works, other states copy it—can work under the right circumstances. But key for the state is to know two things: first, it must challenge the status quo in the right way. Even when the state appears to have been highly successful, it is possible that the results are disastrous if it has crowded out better solutions to the problem at hand. Second, it needs to be able to tell when it has turned into an overly patient investor that should cut its losses.

There is a key difference between the state as innovator and the free-market system the way Baumol describes it: it is not clear how the state can bear any costs of being wrong (in part because it is unlikely to know when it is wrong, as noted above). The arms race that drove firms to innovate to stay ahead of the competition is not there for the state, except for in a few exceptional cases, such as literal arms races. The one real cost that the state does bear for being wrong is possible embarrassment a long time into the future, if at that point it is obvious that it backed the wrong horse. That is, the state does not carry much risk, mostly none at all, and as such it cannot act entrepreneurially the way we normally define the term (see e.g. Shane & Venkataraman, 2000, p. 222). It simply is not its area of expertise, just as running the police force is no area for private businesses. That the incentives are skewed the wrong way is not evidence that the state cannot mimic entrepreneurial behavior, or for that matter, that competing private enterprises could not run a working police force. But these are extraordinary claims, and we should require extraordinarily good arguments from those who want to go down those paths.