From Hartlepool to Habermas: resolving the Progessives’ Prisoner Dilemma
I don’t know if the coming Hartlepool by-election will be lost to the Tories because the previous Brexit party vote goes Tory, or even if one of the ‘insurgent’ parties will come through the middle and win dramatically. The former is more likely, of course.
But I know that the usual calls and counter-calls have been made, with Labour saying a vote for anyone other than Labour is effectively a vote for the Tories, ‘insurgent’ parties saying that Labour has taken votes for granted too long and that time is over, and those who are desperate for some kind of anti-Tory coalition and willing to countenance compromise with other ‘tribes’ looking desperately for a way to mediate. I wish the latter well, but fear it will be in vain.
So how do the coalition builders do better? How do we stop the 2023 or 2024 election becoming a whole lot of Hartelpools, made even worse than 2019 by the potential entry into the fray of splinter parties claiming social democratic credentials but selling these on a sub national basis?
I suggest a good place to start is by looking at the problem in the abstract.
In that abstract, all we really have is a variation on the classic Prisoner Dilemma problem. Simplifying for now to a two-party problem — let’s say Labour and the Lib Dems  — our parties’ dilemma goes something like this:
1) If Labour stands back and lets the LibDems have a clear run, then an interest common to both — not being punished by a new Tory MP — stands a better chance of being met.
2) If the LibDems stand back, then the same applies but vice versa.
3) But both LibDems and Labour are reluctant to take the first step because they don’t trust the other side to reciprocate in other places. Worse, they fear that standing back and allowing a clear run will be badged as weakness, or as “they knew they couldn’t win”.
4) So neither party does anything, because at least that ay the odds are not stacked against them next time around.
In other words, rational choice can beget bad outcomes.
But we also know, from thousands of experiment in the wake of the original setting of the Prisoner Dilemma, that it can be different; in cases where there is trust that the other party will do the right thing, the dilemma is resolved and a common interest met.
So how do we instill trust where there is none?
One way, appealing to John Rawls’ and Amatya Sen’s notion of contractual behaviourism , is for a third party (let’s say Compass) to propose a covenant to future trust.
It works like this. A Labour party candidate, in order to secure a LibDem withdrawal, signs the following statement:
I, the undersigned, agree to abide by the following terms.
1) In the event that the Lib Dem candidate withdraws, and I go on to win the election, I will engage honestly in a process to establish a Progressive Citizen’s Assembly, drawn fairly and equitably from the membership and support network of progressive parties in the constituency*, which is tasked with selecting a candidate for the election after the one for which I am now the selected progressive candidate.
2) In the event that the Labour party is able to form a government at any point before the election after the one for which I am now the selected progressive candidate, I will promote and support legislation that introduces proportional representation to the national electoral system, recognizing that if such a system is introduced the commitments I make in this covenant then become null and void.
3) In the event that:
a) the Labour party is not able to form a government at any point before the election after the one for which I am now the selected progressive candidate, and that no shift to proportional representation is made at national level, and
b) in the event that the Progressive Citizen’s Assembly (established under term 1) finds that the Labour party is not their progressive party of choice to context the next election on behalf of an anti-Right coalition;
then I will abide by the decision of the Progressive Citizen’s Assembly.
Of course, the same would apply vice versa, and in adapted form in a multi-party situation
There are of course issues that fall out of such a covenant, some of which I will address below, but first we should firm up the Rawlsian principle at work here.
The idea of a covenant to future trust is that, by contracting the immediate beneficiary of an altruistic act to a possible altruistic act of their own in the future, the current behaviours of all parties to the agreement are shifted towards cooperation, because — unlike the prisoner dilemma where neither party has evidence of what the other will do — actions of the other party become more predictable. Of course their remains the possibility of reneging on the covenant (in this case the Labour party candidate deciding to stand again), but the formalization of the contractual commitment means that to do will be risky, given the potential for concerted opprobrium for previous partners to the deal and a wider local public. 
There is, of course, still a leap of faith to be made by both parties. The Lib Dems, making the first move, need to believe that the contract will bind Labour, but Labour also need to believe that the subsequent Progressive Citizens’ Assembly will not become kind of stitch up, or at the very least there will not be an immediate expectation of candidate ‘swapsies’ (though of course that might be the decision the Assemblu takes).
And in this respect, Rawls’ contractualist model for resolving the prisoner dilemma can only take us so far, until Habermas takes over.
For Habermas, as for the Labour and Lib Dem candidates in our example, a leap of faith is needed if a new constitutional norm is to come into being. In the terms Habermas uses , this faith constitutes an initial validation of a new ideal form of communication — a validation which may ‘overshoot’ the facticity of that communication, but in so doing triggers a virtuous circle of improvement . What is important for our purposes though is that such validation actually relies on the abstract procedurality of the change being proposed.
The same can apply here. The leap of faith towards an anti-Tory coalition can occur because the Labour candidate, benefiting from the initial covenant process, is in a position to defer concerns about how a subsequent Progressive Citizen’s Assembly may be drawn up and then conduct its work. Then, the first step is taken, and with that the prisoner dilemma is resolved for the other party.
It’s all in the choreography of the first step.
Of course, real life is a lot more complicated than this deliberately abstract schema, in which ‘the candidate’ has been portrayed as a single, rational entity working through the best way to maximize gains and avert losses.
Any fully developed project working towards an anti-Tory coalition which then creates the governmental space for the introduction of proportional representation will need to take account of the local party structures and, in particular, the interests and powers of national party hierarchies. Space does not permit a full exploration, so suffice to say that the prisoners’ dilemma also works at that level too, as we saw in the Autumn of 2019, when the parties were unable to work out a way to remove the pre-election Johnson government from power, even though the parliamentary numbers stacked up for a temporary deal followed by an election in which, with trust triggered, the mechanisms proposed here might have been applied. 
But even to discuss the procedures for local covenants might have a triggering effect on the national parties of a type that nothing else can, not least because of the painful experience of 2019, but also because there would be a glimpse of what lies beyond the current impasse, where proportional representation cannot be introduced until there is a government on power that benefits from proportional representation.
If nothing else, serious discussion of covenants in the abstract would make a change from the frankly tedious, often confused, moralizing about how evil parties are if they ignore the realities of First Past the Post and stand candidates where they cannot win ; at least it might move the debate on to consideration of what might actually be done by the parties to resolve that problem, even if that is not as proposed here, and that would be a step forward in trust-building.
I say this because there is a danger in pursing coalition politics and proportional representation by the route I suggest.
The danger, especially if it takes a number of election to get to proportional representation, is that progressive parties become tied into a deal with other parties, however those other parties act, and however, un-progressive they might become in their (non) electoral appeal to voters. In a world where a radical politics is going to be needed to help avert climate catastrophe, for example, there is a danger that a cosy covenant could lead to a cosy league of centrists, where there is no promotion for exciting new teams, and no relegation for defence-minded outfits who are long past their best.
Still, at least in the short term, the risk of living under a long term Tory government intent on making the UK a kleptocratic state may be greater than the risk of a conservative-minded, but not corrupt-by-design, anti-Tory coalition. Failure to bring about a coalition government intent on reintroducing norms of democratic governance may mean, soon rather than later, there is no longer a democracy to govern.
[1[In my mind’s eye, however hard I try written in the abstract, I think of Southport in December 2019, lost to the Tories despite my door-knocking, because Labour and Lib Dems both claimed to be best placed to beat them, each of them using different arguments to justify their position that were each valid in their own terns. This kind of two-party dilemma is still likely to be more common in 2024 than the Hartlepool multi-party one, especially given that two-party covenants of the type advocated here will pull in some votes from the minor parties that do stand.
 Here, I have drawn particularly on Amartya Sen’s (2009) interpretation of Rawls contractualism, and in particular this passage, which for me links Rawls’ notion of the veil of ignorance to Habernas’ notion of procedurality in the abstract as the key to improvements in the ethical discourse:
First, even though the idea of contract is used by Rawls to determine the nature of just social institutions and corresponding behavioural demands, Rawls’s analysis relies not so much on strong-armed enforcement of the agreement (as in many contractarian theories), but on people’s willingness to follow how they have, as it were, ‘agreed’ to behave. This way of seeing the issue has tended to distance Rawls from the need for punitive enforcement, which can be entirely avoided, at least in theory. Behavioural norms, then, take a post-contract reconstructed form …….The demonstration of mutual advantage as a prelude to the contract in the original position yields the contract, and that in turn — at least the imagination of it (since it is a purely hypothetical contract) — shapes the behaviour of human beings in societies with just institutions set up through the principles embedded in the contract.
Second, another feature that takes Rawlsian analysis well beyond the usual arguments for decent behaviour for the sake of mutual advantage is Rawls’s way of ensuring that in the original position no one can argue or bargain from the knowledge of his or her actual position in society, but has to do so from behind the veil of ignorance. This moves the exercise from the pursuit of actual advantage for oneself to the promotion of advantage for the community as a whole, without knowing what one’s own personal advantage would be in that overall picture. There is surely impartiality enough in this respect in the Rawlsian story, and yet the tie with advantage-seeking justification of cooperation, in this case in an impartial form (thanks to the veil of ignorance), is not transcended through this extension. (p204)
Sen, A (2009) The Idea of Justice (pdf)
 Technically, such considerations feed the mathematical complexities of the iterated prisoner dilemma.
 For an earlier take on why it is not useful to moralize about parties’ unwillingness to accept to ‘realities’ of the First Past the Post system, and put forward candidates even though it allows the Tories in by the backdoor, see my essay ‘The Integrity of the Greens’. The argument there draws on Bernard Williams’ insistence on the overriding ethical importance of the ‘life project’, even where a focus on same leads to negative social consequences. I still think there is validity to that argument, alongside the one I make here.
 I have covered the concept of validation overshoot, by way an example of going to the Coop to get beer, in chapter 3 of a longer essay.
 At a more detailed level, there is also the problem of how to cater for those candidates who, under national arrangements, are asked to “do the decent thing”, given that for many becoming an MP is a life’s ambition not readily given away for a greater good. That is in part a problem of selection processes which promote this kind of person, but this is less likely to be resolved in the short term than the anti-Tory coalition problem, so contingencies will be needed.