What is a ‘policy hack’?

Cross-posted from The Mandarin.

I

Since I used the term ‘policy hack’ in my presentation “What economic reform thinking might look like if we’d bothered to do it”, I’ve had a number of exchanges with Martin Wolf, my discussant that evening, about what I mean. Here’s how I defined the term when I first used it in the essay I ran up to support the presentation:

I use the term ‘hack’ here to mean “a tip, trick, or efficient method for doing or managing something” (Dictionary.com). Though the term is sometimes taken to imply inelegant effectiveness, the policy ‘hacks’ covered here are often simple, but, because they typically work by some clear distinction being made (for instance between funding and provision or property rights and externalities) they are often also elegant.

The point is best made by example. Speaking of the wave of economic reform from the 1970s on I suggested that, while its political impetus was declining economic performance in the wake of the 1970s oil shocks, its intellectual underpinnings went back to a cluster of ideas originating from the 1950s on. Moreover, I distinguished between ‘ideas’ and ‘hacks’. Friedman’s ‘idea’ was unbundling delivery from funding leading to the ‘hacks’ of vouchers and income-contingent loans, for instance. Coase’s ‘idea’ was to think about externalities as an artefact of the definition and assignment of property rights, the corresponding ‘hacks’ being such things as pollution permits and spectrum auctions.

By contrast, I pointed to George Stigler’s research into utility regulation in the 1950s which documented the results of ‘regulatory capture’. This provides us with an ‘idea’ (price regulation isn’t all it’s cracked up to be.) It also leads us to ask if we could change things to improve this – change regulatory governance or whatever. But where, in the cases above, the hacks arise as ‘aha’ moments from the analysis (even if such aha moments turn out to be a dead end or require lots more development), Stigler’s critique, his ‘idea’ tells us something’s wrong but doesn’t lead directly to any ‘aha’ moment as to how to fix it.¹

II

My point in all this was to put the ideas leading to policy hacks high in the hierarchy of contributions economists can make to their species. The things I’ve done in economics that have been most useful, conform to my definition of ‘hack’. In each of these cases, some new way of looking at things produces an array of policy tricks or hacks. And there’s preferably some elegance to the way the ‘hack’ flows from some idea.²

Robert Shiller’s ideas for new markets in risk are all hacks, all stemming from the ‘idea’ that the way the financial products that do exist are defined is some relatively arbitrary product of history. That means numerous potentially very useful markets don’t exist. Or, since all the suggestions are ‘why don’t we create a market in that’, perhaps one could say they’re a single hack in a litany of guises.

Peter Martin’s recent list of a Magnificent Seven new policy ideas forces me to further distinctions. Are they all hacks? Well if I make them all hacks, then my new term ‘hack’ doesn’t say anything distinctive. It seems to me that most of the proposals could be described as hacks, though not surprisingly, some of the best proposals are not original. Another, proposing that we tax super to fund aged care is a hack in the sense that it’s coupled with a reasonable idea (taxes on super earnings are both progressive and fall more heavily on those closer to needing aged care because older super beneficiaries have larger portfolios), but I don’t think the match between the idea and the hack is particularly elegant or pleasing.

I think one of the best policy ideas which is better targeting of welfare isn’t really a hack. It’s the solution of an optimisation problem that the researchers gave themselves – with some fancy new modelling they’ve managed to do. Good on them. Still, I could find ways to disagree with this conclusion. One could argue that the ‘idea’ from which the hack arises is that of ‘optimising’ welfare payments to minimise poverty.

III

Is a policy hack just what philosopher Karl Popper called “piecemeal reform”? Well yes. But no.

I’m happy for ‘hacks’ to be regarded as a subset of piecemeal reform. How could they be anything else? But Popper’s injunction to limit oneself to piecemeal reform doesn’t seem to rule out much. His wider mission was seeking to develop criteria of demarcation by which one could identify the One True Path – towards good science, good social science and so, good policy. In each case he saw himself as advocating modest, open, empirically informed, humble and corrigible approaches against hubristic, closed, rationalistic, self-justifying and incorrigible folly and megalomania (of both the intellectual and political kind).

As part of a swingeing attack on what he called ‘historicism’ which represented the totalitarian urge (most particularly Marxism) Popper juxtaposed ‘piecemeal’ and ‘utopian’ social engineering. The latter involves a comprehensive reordering of life in pursuit of utopian ends. But ruling this out as a policy objective doesn’t tell us much. If it rules out any actual policymaking in history, it rules out the chaos of some revolutions particularly the French one with its resetting of the calendar, (but probably not the American one) and the wilder more homicidal or suicidal cults of history. But nothing much else.³

His point seems to be the pretty unarguable one that we’re terribly ignorant and the social world massively complex, so it’s important that the change we make is corrigible – and corrected for apparent mistakes as we go.⁴ Still, he also concedes that there’s an inevitable “piecemeal haphazardness” to ‘utopian reform’ to address endless unintended consequences (p. 68). Popper’s enemy really seems to be a kind of ‘spirit’ of what he calls “activism” not unlike Adam Smith’s objection to the ‘Man of System’ and Burke’s concern about the utopianism of the French Revolution.⁵

IV

Nevertheless it seems to me that Popper’s idea of utopianism does provide a way into something important that my idea of a hack seeks to get at. A hack is a specific policy arising from a specific insight and it’s in pursuit of clear and easily identifiable benefits. A lot of economic reform was not pursued in this way. It was utopian in this sense outlined by Popper (note however that Popper still admits it as acceptable piecemeal reform):

The politician who adopts this ¹ method may or may not have a blueprint of society before his mind, he may or may not hope that mankind will one day realize an ideal state, and achieve happiness and perfection on earth. But he will be aware that perfection, if at all attainable, is far distant and that every generation of men, and therefore also the living, have a claim.⁶

A lot of the policymaking of the reform period was precisely this kind of reform. Reformers might not have described themselves as in pursuit of perfection on earth, but they had a comprehensive vision of transformation in which liberalisation was somehow to be pursued if not quite for its own sake (it was for the sake of higher GDP) – then as an article of faith that such action would achieve that objective.⁷

Hacks will generally be much more specific in their practical motivation. Generalising from the ones I can think of, they will mostly also involve impacts which are localised to the systems in which they arise with impacts outside their immediate sphere being second order. Some, like public-private digital partnerships, would radiate impacts more widely, but this is primarily because they are more ‘connected’. And other hacks might sound attractive but involve daunting complexities which could count against them. Thus, my proposals for competitive neutrality would have straightforward impacts in most areas in which they were tried, including superannuation, but in central banking they change the nature of the institution to which they’re applied with consequences which are potentially much more wide-ranging and difficult to comprehend in advance.

Footnotes

1. I wonder whether things might be improved if a regulatory body were governed by some board chosen by random selection from a cadre with domain expertise but without conflicts of interest. But not only do I have no great confidence that this would make things much better, there’s no direct sense in which it follows from Stigler’s analysis. In that sense, there’s no ‘hack’ there.↩
2. • “Central Banking for All” and “Commonwealth Super for All” are policy hacks with the conceptual innovation behind both being competitive neutrality as a competitive sword in addition to competitive neutrality as a shield for business.
   • Public-private digital partnerships. This is a policy hack arising from the characterisation of Google and Facebook as ‘public goods by choice’. The entrepreneurs building them realised that the value of the services they’d deliver as free public goods would dwarf their value as private goods delivered for a fee behind a paywall. Because the cost of delivering the service was so low compared with the value it created they could become as rich as Croesus by giving it away and monetising through ads. This suggests the existence of other potential digital public goods that cost more to deliver than can be recovered without charging a fee. Public-private digital partnerships can accordingly unlock such possibilities and many exciting ones can be envisaged with additional benefits from the state ensuring that the public interest predominates in system design and pricing.
   • Convening the emergence of standards: Market leaders of various kinds have an interest in releasing their own data demonstrating their relative prowess. But there’s little incentive to do so because, there being no standard against which to report, they would gain nothing from it. Therefore convening groups to define such standards could unleash a torrent of useful data which could improve information flows in important markets. “Windows on Workplaces” is one hack to which these ideas led though there would be many more.
   • Digital access regimes: Given its market power in the sector and its being the result of network externalities, not product superiority, we should impose an access regime on Microsoft Office products whereby its file formats should be fully documented publicly to permit anyone to write to them without the bugs that occur now and other producers of word processing software should be able to copy the look and feel of the Microsoft’s products. This neatly interdicts the source of Microsoft’s market power – network externalities. Similar targeted access regimes could be imposed on dominant platforms cleaving off the benefits of network externalities from the services the platforms provide, though I can’t see it solving all problems – for instance the advertising monopoly on Google search.
   • Rights to alternative regulatory compliance. Those who are regulated would have the right to demand some statement of objectives from regulators and if they could meet those objectives in some other auditable way, they would have a right to do so.For instance:↩
3. Indeed Popper pretty much concedes this and the difficulty in actually drawing the distinction he has made between piecemeal and utopian engineering, retreating back into the ‘motivations’ of the engineers.↩
4. In fact, as one proceeds through Popper’s quite scrupulous consideration of exceptions, things become quite murky:”It may be questioned, perhaps, whether the piecemeal and the holistic approaches here described are fundamentally different, considering that we have put no limits to the scope of a piecemeal approach. … ²onstitutional reform, for example, falls well within its scope; nor shall I exclude the possibility that a series of piecemeal reforms might be inspired by one general tendency, for example, a tendency towards a greater equalization of incomes.”In this way, piecemeal methods may lead to changes in what is usually called the ‘class structure of society’. Is there any difference, it may be asked, between these more ambitious kinds of piecemeal engineering and the holistic or Utopian approach? And this question may become even more pertinent if we consider that, when trying to assess the likely consequences of some proposed reform, the piecemeal technologist must do his best to estimate the effects of any measure upon the ‘whole’ of society.” (Popper, K. R. 1957. The Poverty of Historicism, The Beacon Press, p. 68)↩
5. Thus the difference between Utopian and piecemeal engineering turns out, in practice, to be a difference not so much in scale and scope as in caution and in preparedness for unavoidable surprises. (Popper, Historicism, p. 69).↩
6. Popper, The Open Society and its Enemies, p. 158.↩
7. I’d say the same about the framing of the PC report into competition policy in human services (implicit in the terms of reference it received from the Government) in which the aim was always to contract services out and increase competition, rather than to achieve specific goals in human services or efficiency.↩

1. piecemeal[↩]
2. C[↩]