Thread by @edenhofer_jacob

This (👇) illustrates just how popular citizens' assemblies (CAs) have become in the past two decades, with some arguing that the ills of representative democracy can be addressed by deploying CAs more extensively. Here is why I tend to be sceptical about these arguments. 1/n

https://twitter.com/Josh_Self_/status/1759491365203738961

By way of explanation, let me distinguish between 4 distinct justifications.
1⃣ CAs will produce more congruent policies and correct responsiveness biases (in favour of the rich).
Representative democracies, the argument goes, have been shown to ... 2/n
docs.google.com/document/d/1vw…

favour the rich, producing policies that are relatively far from the median voter's preferences (see👇for the difference between congruence and responsiveness). Because CAs are - at least socio-economically - representative of the population and everyone's voice has the ... 3/n

same weight, the resulting policies will be more representative. Indeed, the policy recommendations by CAs might allow us to get a better sense of the true "responsiveness/congruence" gaps because they might be more likely to reflect the population's enlightened ... 4/n

preferences about complex policy issues, compared to "throw-away" answers to survey questions.
To me, this is the strongest reason for grappling with CA-centred reform proposals. I'd object by saying that the fundamental problem with CAs' policy proposals is the lack ... 5/n

of accountability and therefore incentive-compatibility.
▶️CAs do not bear the costs of their policies, and have little incentive to ignore the constraints policymakers face.
▶️This means policymakers have strong incentives not to implement their ...
6/n

https://x.com/edenhofer_jacob/status/1759508035808952745?s=20

costly and more meaningful recommendations.
2⃣ CAs will reduce the influence of money and lobbying.
This is related. I bring it up here in response to sth. @landemore said on Capital'isnt (@zingales + @bethanymac12). There might be sth. to this ...

7/n chicagobooth.edu/review/capital…

argument, but, given their limited / non-existent influence, there has thus far been little reason for interest groups to lobby CAs. If they CAs were granted more powers, we'd expect this to change. Only then could we evaluate @landemore's argument.
8/n

3⃣CAs increase the legitimacy of the policymaking process and thus the acceptance of contentious decisions.
The causal chain is sth. like: CAs → consultation → greater trust / understanding of process → higher legitimacy 9/n

The Irish CAs seem to suggest that this can sometimes work. Yet, a recent paper by @DamienBol et al. cautions against pushing these qualitatively derived conclusions too far: citizens are most in favour of CAs when they believe many agree with ... 10/n

https://x.com/DamienBol/status/1536295918290583553?s=20

their pre-existing attitude on an issue.
Second caveat: Even if the above causal chain holds, the number of people that would have to be involved in CAs (even accounting for diffusion across social networks) and the temporal persistence of the legitimacy-boosting effect ... 11/n

strike me as implausibly big / long, for this effect to make much of a dent in the aggregate perceived legitimacy of the policymaking process. Ultimately, this is, of course, an empirical question.
4⃣ CAs will help us address the problem of a passive citizenry / make ... 12/n

better citizens.
This is incredibly important, but the question is: are CAs the best tool to address the problem?
The work by Fishkin and co-authors is encouraging in that regard. Note, though, that, if this is the objective, then CAs are ...
13/namazon.co.uk/Democracy-When…

first and foremost about civic education, rather than policy influence. This has to be communicated clearly then.
The more serious risks with CAs is that they overlook the fact that some citizens are passive because they are busy eking out a living or because the ... 14/n

opportunity cost of being informed and engaged - i.e. being a good citizen - are prohibitively high, as a result of, say, lower education. This can easily give rise to a system that tilts the playing field in the favour of those fortunate enough to have the time as well ... 15/n

intellectual resources to be informed.
Where does that leave us? My overall conclusion is that CAs can work well for issues with low / none distributional consequences, especially when preferences diverge on these issues primarily for moral or ...
16/n

https://x.com/edenhofer_jacob/status/1759507165310562727?s=20

epistemic reasons. Abortion and gay marriage in Ireland illustrate this. For issues, where the distributional stakes are high, CAs run into incentive-compatibility problems (see 1⃣) and can, if at all, only serve as legitimacy-boosting and citizenship-fostering devices. 17/END

@threadreaderapp unroll