Making the Most of Limited Government Capacity

How to cite

Kapon, Samuel; Carpio, Lucia Del; Chassang, Sylvain. 2019. "Making the Most of Limited Government Capacity." AEA RCT Registry. AEARCTR-0004802. https://doi.org/10.1257/rct.4802-2.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Samuel Kapon* , Lucia Del Carpio , Sylvain Chassang

* Primary investigator

Registry ID
AEARCTR-0004802
DOI
https://doi.org/10.1257/rct.4802-2.0
First registered
Last updated
Status
Completed
Trial dates
2019-10-07 – 2021-10-12
Countries
United States of America (New York)
Registry record
http://www.socialscienceregistry.org/trials/4802

Abstract

We study how mechanisms that seek to make an optimal use of limited government capacity perform in realistic settings, in which participants need not play according to equilibrium. Our motivating example is that of a tax authority with only limited capacity to perform audits ahead of collection.

The experiment will offer players the opportunity to settle a debt that they owe to a collection authority. The treatments of the experiment will vary the method of collection, settlement offers and information provided to participants about the settlement decisions of others.

Intervention

We vary the method of collection, the information that players receive about others, and discounts for settling early.

Experimental design

-Sessions run in NYU CESS, with each subject playing all treatments randomly ordered.
-Sessions run on Amazon Mechanical Turk with groups of 10 workers, each only playing one session.

Design details

Randomization method
Randomization done by computer.
Randomization unit
Session level - each group will be randomly assigned one treatment.
Clusters
12 sessions in NYU CESS (each session corresponding to three treatments) ~40 Sessions on MTurk (each session corresponding to 2, 3 or 4 treatments)
Observations
~300 individuals in NYU CESS ~1250 individuals on Amazon Mechanical Turk
Arms
Each player in the NYU session plays all 3 treatments, randomly ordered. On MTurk, each player plays just 1 treatment
Primary outcomes
The outcomes we are interested in are total amount collected, total number of settlers and the speed of settlement.

Cited by

No citation data available for this registration yet.

Keywords