83the fact that efficiency was not necessarily the main value in an economic system, and he often sketched out a nuanced role for the state, as his magazine article A Cautious Case for Socialism (1978) makes abundantly clear. He did not pick the title of the article, but it shows how his thinking, even at the time of working on welfare theorems, was very clearly never that market failure was the ultimate or only motivation for government involvement. Few of the classical economists held that view. Given a certain framework of law and order and certain necessary governmental services, they seem to have conceived that the object of economic activity was best attained by a system of spontaneous cooperation (Robbins, 1978). This meant mostly that we should not intervene in the way we produce or consume, given existing means-ends. What an economist should think of the development of new means-ends is simply not clear from theory alone.

84J. P. Larsson

What can definitely be said is that a dominant innovation policy for a time was promoting R&D with the motivation that market failures caused societies to underinvest (Schot & Steinmueller, 2018). A narrower statement of the market failure theory would propose that “government support of R&D should only be justified on grounds of ‘market failure’” (Nelson, 2011 p. 687). This position certainly does have adherents in the economics profession, and it does limit the scope of policy. Nelson continues by noting that while this position is (Ibid) “not necessarily a high obstacle to the development of active policy in an area, the market-failure orientation starts by asking what the market will not do, rather than what kinds of fruitful roles active policy can play.” He notes that the latter view could probably often be more fruitful when asking what we can do to improve innovation systems. At the center of this issue is the question of what it means to create a market and what it means for markets to fail. This discussion resembles the issue of whether the proverbial falling tree makes a sound or not. Akerlof’s (1970) point in The Market for Lemons was not simply that information asymmetries would make the market function less smoothly. A central argument was that if the information asymmetries are severe enough, there would be no market! But surely, the state is not creating a market by mandating that used car salesmen offer warranties.

Judging by the wide range of innovation policies that exist in all developed nations, as well as those innovation funds for risky ventures and ideas that are now ubiquitous, the fix market failures only paradigm is not too influential at present. Few economists seem to think that the most cited missions—putting a man on the moon or helping plant the seeds for the internet—were in any way bad ideas. It is easy to imagine that part of the reason is that these accomplishments were associated with enormous positive externalities. That is, the investor in them could not keep the gains private. That is, at least to some extent, these successes can be understood through a theory of market failure perfectly consistent with Arrow (1962). There is, in that sense, no real difference between what proponents of the entrepreneurial state or mission-led innovation propose as the old and the new.

It is of course a fact that economists and others subscribe to a market failure way of thinking, but that does not mean that the only reason for government involvement is market failure in the Pigouvian sense in which a tax or a subsidy bridges the difference between private and social costs (and most governments seem to prefer regulation and government ownership, anyway). In many circumstances—I shall argue which below—even most classically schooled economists are probably prepared to go far beyond Pigouvian fixing and into something that we may tentatively refer to as genuine market failure: problems or intriguing opportunities that most can agree on but where means and often ends are obscured.

Historically, with things that have clear outcomes, measurability, and in which buyer–seller incentives are clear enough, economists seem to have been rather supportive of large-scale state innovation efforts, even in the absence of narrowly defined market failures. I would argue that the recurring example of putting a man on the moon is precisely the kind of thing that the classical economists could probably get behind: it is an intriguing feat; it was generally a popular idea among the public; Innovation Without Entrepreneurship: The Pipe Dream of Mission-Oriented. . .

85it was likely to lead to some real insights and technological spillovers; the process would indirectly let us evaluate a large number of theories, and given some time it would be possible to determine whether it was time to pull the plug. Crucially, As Nelson (2011, p. 688) pointed out, the same entity was in fact “the intended and eager user of the technology developed, as well as the funder of the R&D.” Accountability and incentives were in place. The state investing in ARPANET would seem to follow a similar logic. In addition, it was infrastructure, which most consider at least largely a state activity that we generally believe comes with large spillovers if done right. Investment in high-risk infrastructure is of course nothing new and debates on cost-benefit analysis, accessibility, and spillover effects, as well as many other considerations, typically precede its construction.

In fact, the technological advances (the internet, biotech, the IT revolution, etc.) that act as canonical examples of good mission-oriented innovation also come with enormous positive externalities, have completely changed the way we think of information asymmetries, or relate to market failures in other ways. What really constitutes fixing markets that do exist on one hand and researching technologies that can fix markets that do not yet exist on the other can hardly be properly distinguished. So, what really sets the new and old innovation strategies apart? The main difference—other than sheer size and risk of the endeavors—is a sizeable shift in favor of bottom-up, relative to top-down approaches.