James Bessen’s The New Goliaths landed on this reader’s desk for review just as the present AI fanfare started to gain momentum. While the book offers little in the way of concrete predictions as to the development of competition and innovation in AI, its prescience should not be underestimated. Indeed, this eminently readable account of the economic and broader societal impact of complex, proprietary IT systems – described in shorthand as ‘software’ – represents essential reading for lawyers and policymakers interested not only in narrowly framed ‘digital markets’, but any market in which firms harness the power of new technology to better understand and capture consumers. Despite focusing on the US economy, much of the economic analysis is in concert with findings from other parts of the world by the OECD, which, incidentally, recently tabled an engaging discussion of the book that can be accessed here.
This review shall focus on two important insights. The first insight concerns how the development of software has fundamentally altered the ways in which firms navigate the trade-off between efficiency and heterogeneity in their product and service offerings. Whereas in the ‘old economy’ a degree of homogenisation was essential to unlocking efficiencies, in recent decades advancements in technology have reduced the marginal cost of customisation to the extent that firms increasingly compete on complexity. Moreover, firms have evolved “new types of organization and institutions” as part of the technological gearshift. This combination of investment in complex, proprietary software and the evolution of organisational structures is exemplified by Walmart, which pioneered the decentralisation of decision-making in order to more efficiently and accurately meet the diverse needs of its consumers. The result is a so-called ‘superstar economy’, in which a small number of leading firms across different sectors have become entrenched thanks to their software-driven complexity advantage. In other words, the gap between top tier and second tier is too wide to leapfrog. Ultimately, Bessen draws a convincing causal link between firms’ investments in software and industry concentration and contests the “disruption myth” embodied in the writings of Joseph Schumpeter and Clayton Christensen. Turning to AI, one of the pivotal questions yet to be answered is whether new technologies will prove to be a disruptive or a compounding force.
The second insight is that the prevailing preoccupation with castigating ‘Big Tech’ may, in certain situations, be rather misguided. As Bessen points out, for example, the cloud industry was essentially created by Amazon’s decision to open-up its highly advanced, proprietary IT infrastructure to the public. This transformation from ‘closed’ to ‘open’ – referred to as ‘unbundling’ – deserves positive attention and should be held-up as a kind of (informal) blueprint for the widespread diffusion of technological innovations. The lesson here is one of discrimination. Designing policies that incentivise technologically advanced firms, regardless of the sector, to license-out critical parts of their IT infrastructures deserves equal, if not greater, attention than many of the headline-grabbing ‘populist’ antitrust directives currently being pursued across a number of jurisdictions. The crucial question to ask is whether firms are “primarily engaged in winning a greater share of the economic pie” or “expand[ing] the size of the pie to the greater benefit of society as a whole”. The ‘superstar economy’ is charactised by the former, and the risks of inhibited diffusion range from diminishing aggregate productivity growth to growing income inequality. Given the general purpose nature of AI technology, its potential to increase long-term productivity is considerable. The key will be to sustain what appears to be vibrant early ‘open’ entry, in particular in the development of foundation models, and to ensure that the economic fruits of AI are equitably distributed.
Although Bessen’s articulation of problems is novel and compelling, the ‘unbundling’ solutions feel rather too general. Proposals for inter alia mandatory licensing, open standards and data access, and for reform of laws that inhibit skilled worker mobility – all of which would contribute to the transformation of firms from ‘closed’ to (more) ‘open’ – are not new. And the design and operationalisation of interventions under any one of these approaches demands a far more comprehensive economic and legal analysis than a book of this nature could undertake. That being said, one readily actionable take-away stands out. Complex, proprietary software makes the task of regulating systemic and existential risks ever more challenging, and the book adroitly joins the dots between the subprime mortgage collapse that precipitated the Great Recession of 2008, the Volkswagen ‘Dieselgate’ emissions scandal, and the Boeing 737 MAX catastrophes to elucidate the perils associated with the widening disconnect between regulators and the regulated. Invoking Louis Brandeis’s observation that “sunlight is said to be the best of disinfectants”, Bessen concludes that “exposure to code and data to many eyeballs can go some way toward leveling the balance of power between corporations and regulators”. This reader would observe that AI may prove to be one of the most powerful instruments in the regulatory toolkit, driving the efficiency with which complex information can be analysed and helping to realise the ambitions of initiatives such as the Computational Antitrust project.
The New Goliaths is a thought-provoking, and highly approachable, book that makes an important contribution to the ongoing discussion concerning the role and impact of rapidly advancing digital technologies in increasingly concentrated markets across sectors and across continents. It is also a work from which many questions and lessons may be drawn to inform the current debate concerning the regulation of AI. Bessen is widely regarded as something of an academic Goliath for good reason.
Ben Evans is a Postgraduate Researcher at the School of Law and Centre for Competition Policy, University of East Anglia.
About the book
- James Bessen
- Published June 2022
- £20, Hardcover
- ISBN 9780300255041
