DARPA vs the Wheely Suitcase
Or, The Quantity vs Quality Theories of Innovation
by Jonathan Haskel and Stian Westlake
It sometimes seems that there are two ideal-typical books about innovation policy: ‘Apollo, therefore socialism’ and ‘wheely suitcases, therefore libertarianism’.
One school of thought is based on the underlying economic reality that innovation investments have spillovers. The problem with innovation having spillovers, as Kenneth Arrow pointed out half a century ago, is that if you rely on self-interested businesses to look after society’s innovation investment, you’ll get massive underinvestment and slow economic growth. Instead, the argument goes, governments need to subsidise, fund, or directly undertake a large amount of innovation investment. The classic example is the Internet, which exists because of decades of generous investment of US taxpayers’ money by organisations like DARPA — they weren’t necessarily trying to transform the economy, but by making investments that no-one else would, they did so anyway. The DARPA/spillovers case was powerfully made in Mariana Mazzucato’s The Entrepreneurial State, and is widely accepted by mainstream economists.
The other school of thought is that innovation depends on brilliant entrepreneurs trying out lots of things, and occasionally finding just the right combinations of ideas that proves to be very valuable. The classic example is the Rollaboard, the first wheely suitcase: after the fact, it seems an obvious invention to many; it combined two already existing inventions (the suitcase and the wheel), and it was carried out by an entrepreneurial amateur. The kind of books that make this point include Edmund Phelps’ Mass Flourishing and Matt Ridley’s How Innovation Works. In Ridley’s description of the invention of the wheely suitcase, what matters was not the amount of investment in R&D that went into it, but the power of the market forces that called it into being. In Ridley’s words, “When the world is ready, the idea will be already out there, waiting to be employed: in America, at least. Nothing like this happened in Communist Russia or Mao’s China.” Writers like Ridley and Phelps tend to be hostile to government intervention, arguing that it often leads to wasteful boondoggles (Quaero, the EU’s attempt to build its own visual search company, is a classic example), saps the vigour of innovators by loading them down with the rules that tend to be associated with government involvement, or weakens the market forces that call forth innovation.
Our reaction has always been ‘it’s a bit of both’. Entrepreneurs are important, but you can’t write spillovers out of the equation. A libertarian economy would underinvest massively in innovation, and would suffer as a result. But recently we’ve been trying to understand the Wheely Suitcase position a bit more, and we think it reveals an interesting dichotomy or dilemma.
It strikes us that a few ideas are implicit in the Wheely Suitcase idea. Firstly, the idea that innovation investments are not all alike. The value of any given dollar of R&D or design or software varies wildly. Secondly, that innovation is combinatorial: as in the wheely suitcase example, a successful innovation usually involves combining a set of investments in just the right way (innovation investments exhibit potentially large synergies). These two properties are related: finding the magic combination can make your innovation orders of magnitude more valuable.
Professor Paul Lewis of King’s points out that these characteristics are central to how Austrian economists think about capital in general — they talk about ‘recombinance’ and ‘heterogeneity’ in the capital stock. Without passing judgment on Austrian Capital Theory in general, it seems intuitive to us that innovation investments (and intangible capital generally) exhibit more recombinance and heterogeneity than physical capital like a machine tool, and office building or a heavy goods vehicle. So in an economy with more intangible investment in general, recombinance and heterogeneity would be more important phenomena.
There is also an assumption that the systems used to allocate public funding or subsidy can sometimes disincentivise high-value innovation. Whether or not this is empirically true, you can see intuitively and anecdotally how it might work. Researchers frequently talk about how the bureaucracy involved in research funding (which is put in place by democratic governments seeking to be responsible stewards of public money) deters really breakthrough research and encourages gaming. The idea that higher taxes (to pay for the R&D subsidies) or activist public bodies (to coordinate it) distort markets and weaken incentives for entrepreneurs is a common talking point, and while not everyone agrees how important a phenomenon it is, it’s almost certainly sometimes true.
These concepts helped us understand the Wheely Suitcase case for more entrepreneurship and the concern over public funding. If you believe that the value of innovation investments has a wide range and has a large skew to the right, you want lots of attempts to try out different combinations in the hope that you finally hit on the right one. You also want to avoid any kind of policy that creates perverse incentive that puts off the innovators who find the winning combinations, even if it increases the total amount of R&D undertaken. (For an example of a system that might increase the amount of R&D done but discourage some of the best R&D projects, consider Peter Higgs’ comments on research culture.) You could call this the Quality Theory of Innovation. Fewer investment dollars, but, to quote Lenin, better fewer.
If, on the other hand, you believe that the value of innovation investments has a narrower range, and closer to a normal distribution, then what really matters is the total amount of investment, even if to get it you have to have a set of rules that might deter some of the highest-value innovations. This is the Quantity Theory of Innovation: quantity has a quality all of its own, as Stalin is supposed to have said.
To our eyes, the model you prefer depends on what your mental model of how innovation investments pay off. Imagine some generous soul offers you a large number of financial securities for free. Then she makes a follow-up deal. She’ll double the number of securities she has given you, but in return you have to give her the most valuable 5 per cent in a year’s time. Whether you accept the top-up deal depends on what you think the payoff profile of the shares looks like. If they’re high-quality corporate bonds, say, with little variability of return, you’d accept the deal: getting 97.5% more is definitely worth the loss of the top performers. But if they’re lottery tickets, the deal is terrible: losing even a few of the top performers means the portfolio is close to worthless. Taking the deal is the Quantity Theory, turning it down is the Quality Theory.
(Incidentally, we think the same dichotomy might apply to other types of intangible investment — for example to post-18 education, and the debates over apprenticeships, the UK’s Augar Review, and the neverending question of whether too many young people go to university. But for the sake of brevity, we’ll leave that story for any time.)
In fact, neither the pure Quantity Theory nor the pure Quality Theory exist commonly in the wild. The Entrepreneurial State also includes a theory of ‘innovation quality’, in the form of ‘mission-oriented investment’ (roughly speaking: there are a number of societally important ‘missions’, and innovation investment oriented towards them is disproportionately valuable, so government should be empower to do this). Equally, the blog posts of UK government adviser Dominic Cummings are steeped in the Quality Theory (for example, the idea of heterogeneity of innovation investment — the case for a UK ARPA, which is central to his vision of post-Brexit Britain is predicated on the idea that the right funding system combined with smart people can lead to massively valuable innovations) — but along side that he has pushed for big increases in public R&D funding, the definitive Quantity Theory policy.
It’s probably true that most people subscribe to some mixture of the Quantity Theory and the Quality Theory — but it seems to us that it is useful dichotomy for anyone thinking about innovation policy, or for that matter investment in general.
