Cogs and Monsters: What Economics Is, and What It Should Be, Diane Coyle. Princeton University Press, 2021, 257 pages.

The ambitions of Diane Coyle’s Cogs and Monsters are written clearly into its subtitle ‘what economics is, and what it should be’. Coyle argues that standard criticisms of economics are outdated, but that the discipline does need to adapt in response to a collection of more contemporary issues – a mismatch between research and policy; new social, economic, and technological conditions that blunt key elements of the economic policy toolkit. I expect academic philosophers and economists will find the latter argument (what economics should be) more engaging than the former (what it is). The length of the book and the intended (general) audience means that Coyle’s dismissal of previous criticisms of economics – that it is too mathematical, too focused on selfish individuals, etc. – is unlikely to shift existing viewpoints one way or the other. Where Coyle’s perspective is more insightful is on the future of economics. She offers a lucid discussion of the problems that existing forms of economic policy analysis face. In short, shifts in technology mean that a whole host of common tools and frameworks (from metrics such as GDP, to competition policy centred on consumer prices, to maximizing social utility via preference satisfaction) no longer apply. This argument is persuasive. And Coyle’s suggestion that germs of the changes required to update economic policy analysis already exist is interesting. But, to this reader at least, her conclusion that all we need to do is put these germs together into a new single paradigm seems unwarranted. It is not obvious that a new unified economic policy perspective is what is necessary. Moreover, the idea that we should aim for a new unified perspective seems at odds with arguments Coyle makes elsewhere in the book for policy analysis that does not claim to have one answer for all circumstances.

1. What economics is

The first three chapters of Cogs and Monsters follow a similar pattern of argument – ‘economics is not bad for all the reasons commonly cited but does need to improve in several more incremental ways’. Coyle argues, for example, that even though economics played some role in creating the conditions for the Global Financial Crisis, the discipline is not really to blame (that lies at the door of bad policy and political philosophy). Moreover, contemporary economics differs markedly from its caricature as free-market ideology concerned more with mathematical models of selfish individuals than reality. Coyle tells us that although economists do have a penchant for market solutions, it is empirically grounded. Most economists are sober pragmatists and empiricists – as evidenced by the developments of behavioural economics, institutional economics, and experimental economics. The issue, Coyle argues, is that economists have not done a good enough job of communicating what they actually do. Echoing the public understanding of science pleas from the 80s and 90s, she argues that economists need to engage more with public debate. Although Coyle is probably correct that there is a mismatch between what economists do (mainly some form of applied micro) and what people think they do (often macro forecasting), it is not clear that a failure to bridge this mismatch is the main reason for economics’ bad reputation. The idea that if only economists would communicate more effectively the public would better understand that they are empirically minded pragmatists may be partly correct. But it seems equally plausible that it is the ways in which economists have previously engaged in public debate and policy advice that have damaged the discipline’s standing – see, for examples, the perception that there has been a revolving door between US government economic policy advice and high finance, or the values assigned to human lives in early climate cost benefit models (assigning much lower values to lives in developing countries).

Coyle’s most interesting insights come when she reflects upon the area she is most familiar with: policy economics (the economic tools and analysis used within government). Coyle suggests that policy economics has not moved with economic research, and it is here that the changes are really needed. She argues that there is a core paradox in policy economics that practitioners have insufficiently acknowledged: economic analysis aims to take the position of the objective outsider but by putting economics into practice, policy economists step into their models. Because of this inherent reflexivity, policy economists need to think more about the political economy of economic policymaking. What Coyle means by this is that policy economists should pay more attention to how policies are implemented and to potential behavioural responses. This, I think, is broadly correct. But Coyle could have taken the political economy of economic policy further. Given that she goes on to argue (in Chapter 3) that economics cannot be completely value free, some discussion of how different value judgements should be balanced in the economic policy process would have been interesting. In the Introduction, Coyle nods to the significance of diversity. Could policy advice be more diverse? Is the role of economists in policy too large? Coyle offers an unequivocal ‘no’ to the latter question. But she doesn’t give a particularly strong argument for this answer. She simply lists a collection of important innovations due to economic advice – the idea of opportunity cost, cost benefit analysis, competition policy, among others. This doesn’t do much to compare the costs and benefits of the counter-factual scenario in which economists and sociologists have had more equal footing in policy in the past 70 years. Such an involved comparison is too much to ask for in Coyle’s short book, but it is also too complex to simply assume in favour of economics by pointing to a few innovations.

Coyle argues that one obstacle to improving policy economics is the idea that ethical issues should be separated out from proper positive economic analysis. This idea is not suitable for economic policy because policymakers frequently need to compare the outcomes of measures that have winners and losers (ruling out relying on the Pareto criterion). Coyle argues that cost–benefit analyses do not solve the problem because they assume the sum of costs and benefits to individuals is the same as the social costs and benefits, something she argues is not the case. Coyle argues that a more policy relevant version of economics must accept a partial move away from positive economics. She is not the first to say something like this. But she adds an interesting fresh reason to this point. Coyle claims that a central reason that welfare economics needs to be revitalised is the way that technology is changing the economy. She argues that networked technologies have increased the instances of social spill-over – network effects, economies of scale, externalities, agglomeration effects. This leads to increasing divergences between private and collective interests and makes general equilibrium theory and Pareto-based welfare economics less and less relevant to concrete policy questions.

It is here where the uniqueness of Coyle’s perspective starts to come out. The first half of the book (Chapters 1–3, roughly focusing on ‘what is economics’) collects a nice set of reasons to think that economics is partly performative, that it cannot always be value free, and that there is not going to be one perfectly objective way to devise policies from experimental evidence. But those engaged in philosophical topics around economics are unlikely to learn much new about performativity, evidence, or the role of value judgements in economics. However, to knock Cogs and Monsters because it does not offer fresh insight into thorny philosophical questions would be to miss the point. What is interesting about the book is how Coyle sees these issues in relation to forthcoming problems that economics must navigate. This comes to the fore more in the second half of the book.

2. What it should be

Coyle argues that the increasing significance of the digital economy requires a profound rethinking of economics’ standard toolkit. She argues that GDP is outdated as a measure of economic success, particularly when we consider the value being created in sharing platforms and data aggregation, which does not register in GDP statistics. She notes that in the digital economy power is concentrated among a few very large firms to a greater degree than ever before (challenging standard ideas about competition), and suggests this is due to a number of features inherent to how digital platforms work – everyone being on one platform benefits all, etc. She also argues that shaping the preferences of individuals is a key part of digital business to a degree not seen before. The prevalence of advertising and data on individual choices is now so great that it makes even less sense to think of individual preferences as given than it did. When this is combined with network effects, the policy goal of maximizing utility via preference satisfaction seems at best very difficult to satisfy.

This leads Coyle to one of the most striking passages in the book. She argues that although everyone has always known that the models of general equilibrium are not completely accurate, policy economists have treated them as close enough. This has justified the dual ideas that the goal of policymaking should be maximizing preference satisfaction and that market solutions are the most effective means to achieve that goal – specific externalities or market failures can be identified and rectified through clever policy. Coyle’s key claim is that this is just not correct anymore: ‘the degree of interdependence and increasing returns in the modern economy means that this is incorrect and an unhelpful framework for designing economic policies’ (178). A rethink is required.

Coyle weighs up the idea that this rethink should entail a return to some sort of socialist calculation. She quickly dismisses that idea but notes that the issues that block socialist calculation – a lack of perfect data on everything that happens in the economy; increasing returns to scale and network effects that make optimization problems impossible to solve analytically – also apply to mainstream market economics. Coyle gives examples from competition and industrial policies to make her point. Existing competition policy can’t take into account envelopment (companies growing so large they can easily take over neighbouring markets) and data barriers to entry. Moreover, the concentration of most major tech companies in two rival states (the USA and China) means geo-politics is starting to play a more explicit role in the degree of anticompetitive behaviour policy makers are willing to stomach. Where determining what counts as healthy competition may once have been adequately thought of as a technocratic question, it is now much more political. Coyle draws a similar conclusion about industrial policy – technocratic cost–benefit analysis has not worked for ‘left behind’ areas, and strategic planning and more political forms of decision-making may be the way of the future.

This leads Coyle to argue that we need a new form of policy economics for the modern mixed economy. This new policy economics should, given the non-convexities (conditions that block optimization problems from having single analytical solutions) involved, be sensitive to changing circumstances and contexts. Coyle argues that nothing is, yet, on offer. The climate of ideas has moved away from the market mainstream but has not settled elsewhere. The fundamental problem, as Coyle sees it, is that there is no settled framework for assessing social welfare in the context of the non-convexities now common. Coyle doesn’t solve this problem but does give us a list of assumptions she thinks the next paradigm will have to start with: ‘increasing returns to scale, information asymmetries, pervasive network effects and externalities, principal-agent relations, interactions between decision makers’ (209).

Coyle also offers hope by noting that these assumptions exist in one form or another in some of economics’ specialist subfields – endogenous growth theory, narrative economics, etc. Coyle’s overall conclusion is, thus, partly optimistic: economics has all the tools required to make sense of our changing world, we just need a new way of synthesizing these ideas together to create the next policy paradigm. She ends with an interesting point, however, that slightly pulls away from this prognosis: because non-convex optimizations are intractable, there are always going to be multiple answers to any policy question. There is, in other words, not one way to run the kind of economy we now have. Context and adapting with circumstances will be the order of the day. This is something I broadly buy (and it is something philosophers such as Nancy Cartwright have said in different words for some time). Yet it seems to jar with the search for a new unifying economic policy framework. Maybe what we need instead is constantly evolving ad hoc knowledge and theories about specific domains, countries, regions, or aspects of the economy. Maybe the idea of overarching frameworks and one discipline dominating policy debates itself is what needs to shift – maybe those disciplines that offer better training in how to draw out specific context-specific causal mechanisms should play a greater role. Coyle will surely disagree, but her description of how the digital world has changed the economy offers the ammunition to those who think otherwise.