Social Choice for Social Good: Proposals for Democratic Innovation from Computer Science

<p>Driven by shortcomings of current democratic systems, practitioners and political scientists are exploring <em>democratic innovations</em>, i.e., institutions for decision-making that more directly involve constituents. In this thesis, we support this exploration using tools from computer science, via three approaches: we <em>design</em> practical algorithms for use in democratic innovations, we mathematically <em>analyze</em> the fairness properties of proposed decision-making processes, and we <em>identify </em>extensions of such processes that satisfy desirable properties. Our work mixes techniques from computational social choice, algorithms, optimization, probabilistic modeling, and empirical analysis.</p> <p>In Part I, we apply the frst two approaches to <em>citizens’ assemblies</em>, which are randomly selected panels of constituents who deliberate on a policy issue. We analyze existing algorithms for the random selection of these assemblies, and we design new algorithms for this task that are provably fair and now widely used in practice. In addition, we design algorithms for partitioning assembly members into deliberation groups, which allow more members to interact than before.</p> <p>Part II identifes extensions to <em>liquid democracy</em> and <em>legislative apportionment</em>. First, we demonstrate that a variant of liquid democracy, in which agents are asked for two potential delegates rather than a single delegate, reduces the concentration of power observed in classic liquid democracy. Second, we extend legislative elections over parties to approval ballots, and give apportionment methods for this setting that satisfy strong proportionality axioms. Finally, we extend a proposal for the randomized apportionment of legislative seats over states to satisfy additional monotonicity axioms.</p> <p>In Part III of this thesis, we engage with a specifc policy topic, <em>refugee resettlement</em>. We design algorithms for allocating resettled refugees to localities in a country, which improves these refugees’ chances of fnding employment over the status quo and is now being used by a major US resettlement agency</p>