Thread by @saadgulzar

Just accepted @The_JOP

"Information, Candidate Selection, and the Quality of Representation: Evidence from Nepal" with @ZuhadHai & @bnpaul20

RCT: party leaders update candidate selection when polling info is presented to them, ⇧ vote share!

dropbox.com/s/a4fxhyq6cwm4…

THREAD

Selecting candidates for elections is a key job of party leaders. Yet we know little about this process, perhaps because party operations are often black boxed to researchers.

Working with a pol party in Nepal we study how cand selection can be improved for parties *and* voters!

Two unique features of our setup: 1. we observe the full set of feasible candidates that usual election data do not record as party elite preferences are typically unknown 2. we observe party and voter rankings over this set

First, we show that in our case party top leader pref on candidates matches voters' tops person in only a third of cases -- selection is as good as random chance.

We posit two reasons for this discrepancy: lack of info among leaders, and/or strategic divergence

Second, we poll voters to retrieve their ranking of prospective candidates and randomize the provision of this information to party elites. Two results emerge: 1. leaders update their candidate selection, displaying their top cand with voters'; and 2. party vote share improves!

Studying candidate selection in an actual election with real stakes allows us to analyze impacts on vote shares.
Important because we can then interpret results as suggestive of party leader candidate selection diverging from citizen preferences because of a lack of information!

Results show that a closer correspondence between party elite behavior &public opinion has the potential of improving the electoral performance of the party and aligning who is elected to office with voters’ preferences–factors that together improve the quality of representation!

See the paper here: dropbox.com/s/a4fxhyq6cwm4…

and send us ideas for follow-up work with parties!