PI makes every political faction hate it. Winning the administration deficit not only zeroes solvency, but also the net benefit.

This brings us to what we take to be the real disagreement between Atkinson and ourselves: the politics of PI. In our 2007 article we stipulated that in the trilemma we can avoid only two of three horns—that is, reverting to soft workfare or by contrast, fully embracing an unconditional BI—by accepting a third, the need for intrusive and costly ‘iron-clad administration’. In his second response to our trilemma, Atkinson claims that faced with this trilemma, he would accept the need for more administration, something which he contends is desirable more generally. Reiterating a major theme outlined earlier in Inequality, he insists ‘the state should invest in better social administration, recognising that this requires a higher labour input, and emphasising the quality of service rather than merely measuring cost efficiency.’ 11 Furthermore, he argues that since PI does not involve means-testing, its introduction would free up considerable administrative resources that could be reassigned elsewhere in the system. In our view, it is unlikely that policy would take this turn in the short run, given current governments’ focus on budget restraint and cost-containment following the 2008 Great Recession. If political feasibility is one of the criteria for accepting a particular choice, then embracing iron-clad administration is, for the time being, a non-starter.2

However, our disagreement with Atkinson is grounded in a deeper concern. We share Atkinson’s personal preference for investing in administrative capability to accommodate PI implementation challenges. But it is in fact precisely that—a personal preference. What Atkinson fails to appreciate when responding to our trilemma in this manner are the administrative challenges to instituting PI —and the hard choice of how to accommodate them at the level of implementation. All of these are bound up with the coalitional politics of trading-off competing policy goals in a political environment where deep disagreement persists about both the goals and the instruments of income support. The point is not whether Atkinson—or we—have good reason to opt for one or other way out of the trilemma; rather, the issue is that in the political climate in which PI is presented as a solution to the challenges of instituting an unconditional BI, the trilemma drives a wedge between different political factions, each of which supports PI for different reasons.

Atkinson’s reluctance to appreciate fully the political nature of the trilemma is surprising in light of his own assertion in his original article that PI was a necessary concession to the more radical BI proposal for distinctive political reasons. He writes, ‘I believe it will be difficult to secure support for a citizen’s income while it remains unconditional on labour market or other activity.’ 12 However, what emerges from our analysis is that PI itself faces considerable difficulties in attracting a robust and stable coalition of support. The idea of a PI may be supported by a broad faction for a wide array of reasons, but when faced with the administrative challenges these diverge both in the extent to which they see administration as a problem—including its cost and intrusion in claimants’ lives—and the direction of how to resolve it.

Those already enthused by the BI ideal will naturally veer to abolishing conditionality.13 By contrast, those who insist on the participation requirement are more likely to converge on the need for a more restrictive policy.14 Those willing to embrace investing in administrative capabilities, such as Atkinson, opt for a third avenue but while doing so part company from their erstwhile allies. For Philippe Van Parijs and Yannick Vanderborght the cost of participation income itself offers an interesting trajectory for pushing towards basic income in the short or middleterm. They suggest ‘the weakening of the participation condition ... or indeed its complete removal could be sold ... as being in the material interest of all parties: less burdensome for the beneficiaries and the administration in charge of checking that the conditions are fulfilled, and less costly to the taxpayers.’ 15 But once we allow taxpayer cost to become part of the argument, there will equally be those who want to move from PI to a more restrictive workfarist option, which could be seen as more easily administrable, and hence could offer some cost savings by embracing categories that are more easily monitored. It remains to be seen, therefore, whether a shift from PI to BI can indeed be sold as being in the interest of all parties. Favouring one horn over the two others, as Atkinson proposes, only works if the coalition calculus is sufficiently skewed so as not to break down when faced with inevitable challenges of implementation. The option of investing in administrative capabilities for the sake of implementing a more inclusive social protection system may appeal to Atkinson—and, again, to us—but we should not assume without further argument that it offers a stable political outcome. This brings us back to a point we made elsewhere: administrative instability breeds political instability—and that is the reason why PI in our view is a non-starter.16