Rank Incentives within Firms

How to cite

Ashraf, Anik. 2015. "Rank Incentives within Firms." AEA RCT Registry. AEARCTR-0000981. https://doi.org/10.1257/rct.981-1.0

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

Anik Ashraf*

* Primary investigator

Registry ID
AEARCTR-0000981
DOI
https://doi.org/10.1257/rct.981-1.0
First registered
Last updated
Status
On going
Trial dates
2015-10-01 – 2017-05-25
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/981

Abstract

This study aims at understanding whether provision of information to workers on their relative performance can increase their productivity, and if it does, what are the mechanisms that drive such change. The information is delivered in the form of ranks computed from their performances in the previous month. This study is in collaboration with a Bangladeshi sweater factory, where the workers are paid based on a piece-rate scheme. I introduce rank information to treatment group workers in two different forms - private and public. The outcomes of key interest are productivity, absenteeism, quality of outputs, cooperation, and communication across workers.

Experimental design

This study is conducted with workers from the manual knitting section of a sweater factory in Bangladesh, where workers knit sweater parts for monthly compensation based on piece rates. There are 380 such workers divided into 15 administrative groups called 'blocks'. The workers will be randomly assigned into one of three different experimental groups through a public lottery. These groups consist of one control group and two treatment groups. Workers in the treatment groups receive information about their relative ranks among their peers (within their respective treatment groups), while the specific kind of information they receive varies across the two groups. The ranks are computed from the average time a worker takes to complete jobs of each style assigned to him. In the first treatment group each worker gets to know only about his own rank, and no one knows anyone else's rank - the information provided is thus 'private' in nature. In the second treatment group, all the workers get to know each other's ranks - the information provided is thus 'public' in nature. The treatment will be provided through individual letters delivered on a monthly basis, while the ranks will be computed based on their performances in the previous month. To negate any effect caused by introduction of letters and not the information per se, the control group will also receive letters but with minimal summary statistics of performance within their group, something that does not introduce rank incentives in the control group and will also be included in the treatment group letters.

As already mentioned, the randomization will be done at worker level. But before workers are drawn into the experimental groups, each of the 15 blocks will be randomly selected into one of two categories. The first category of blocks will have relatively more 'public' treatment workers than 'private'. The second category of blocks will have the opposite - relatively more 'private' treatment workers than 'public'.

Design details

Randomization method
Public Lottery
Randomization unit
In this factory, workers work in administrative groups called 'blocks'. At first, we randomly assign each of the blocks into one of two different categories - (i) 'block' with relatively bigger share of 'private' treatment, and (ii) 'block' with relatively bigger share of 'public' treatment. In the second step, we randomly assign workers of each block into one of one control and two treatment groups through a public lottery.
Clusters
380 workers
Observations
380 workers
Arms
128 workers in Control arm, 126 workers in 'Private' Treatment arm, 126 workers in 'Public' Treatment arm.
Primary outcomes
The outcomes of key interest are productivity, absenteeism, quality of outputs, cooperation, and communication across workers.

Cited by

No citation data available for this registration yet.

Keywords