A Contribution to the Defense of Liquid Democracy

TL;DR

This paper defends liquid democracy by critiquing prior analysis, showing that previous concerns about delegation structures and computational complexity are not practically relevant, thus supporting its viability.

Contribution

It provides a high-level rebuttal to Caragiannis and Micha's critique, clarifying misconceptions and demonstrating the robustness of liquid democracy against their concerns.

Findings

Caragiannis and Micha's critique is largely irrelevant in real-world contexts.

Incorrect assumptions in the $oldsymbol{ extalpha}$-delegation model undermine their critique.

Optimal delegation structures are not necessary for correct outcomes in liquid democracy.

Abstract

Liquid democracy is a hybrid direct-representative decision making process that provides each voter with the option of either voting directly or to delegate their vote to another voter, i.e., to a representative of their choice. One of the proposed advantages of liquid democracy is that, in general, it is assumed that voters will delegate their vote to others that are better informed, which leads to more informed and better decisions. Considering an audience from various knowledge domains, we provide an accessible high-level analysis of a prominent critique of liquid democracy by Caragiannis and Micha. Caragiannis and Micha's critique contains three central topics: 1. Analysis using their $\alpha$-delegation model, which does not assume delegation to the more informed; 2. Novel delegation network structures where it is advantageous to delegate to the less informed rather than the more informed; and 3. Due to NP hardness, the implied impracticability of a social network obtaining an optimal delegation structure. We show that in the real world, Caragiannis and Micha's critique of liquid democracy has little or no relevance. Respectively, our critique is based on: 1. The identification of incorrect $\alpha$-delegation model assumptions; 2. A lack of novel delegation structures and their effect in a real-world implementation of liquid democracy, which would be guaranteed with constraints that sensibly distribute voting power; and 3. The irrelevance of an optimal delegation structure if the correct result is guaranteed regardless. We conclude that Caragiannis and Micha's critique has no significant negative relevance to the proposition of liquid democracy.