Cross-Community Audit Methodology for Bias Detection
RCTReview
Observational transaction data cannot isolate whether price differences stem from true credit risk differences, bargaining skill, or pure identity discrimination. Unbiased detection of ethnic market distortion requires controlled audit experiments that systematically rotate buyer identities across sellers.
Picture this
To test if store owners treat people differently based on heritage, researchers send out actor buyers from different ethnic backgrounds who follow the exact same script and purchase the exact same items. By comparing what happens to each buyer at the exact same stores, researchers can catch discrimination red-handed.
What the evidence says
Standardized multi-community audit visits successfully isolated seller pricing differences (-12% for Marwari sellers vs. +5% for Andhra sellers) and buyer-seller match effects independently of product quality or negotiation script variations.
- Who was studied
- 107 pen and stationery shops in Chennai, India, visited by local entrepreneur auditors representing Tamil, Marwari, and Andhra groups.
- How
- Randomized audit matching where each shop was visited at least once by auditors from each of the three communities using standardized negotiation scripts.
What to do
Deploy matched-pair audit buyer teams adhering to identical negotiation protocols to measure discrimination across supply chain vendors.
From the source
"These auditors were randomly matched to the wholesalers in such a manner that each shop would be visited at least once by auditors from each of the three communities."
Importance_of_Ethnic_Networks_in_Business_Transactions_in_India.pdf