A revealed preference study on intra-household resource allocation

How to cite

Arslan, Cansın; Gregg, Daniel. 2019. "A revealed preference study on intra-household resource allocation." AEA RCT Registry. AEARCTR-0004877. https://doi.org/10.1257/rct.4877-1.1

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

Cansın Arslan* , Daniel Gregg

* Primary investigator

Registry ID
AEARCTR-0004877
DOI
https://doi.org/10.1257/rct.4877-1.1
First registered
Last updated
Status
Completed
Trial dates
2019-10-22 – 2020-06-30
Countries
Uganda (Kapchorwa District, Eastern Uganda)
JEL codes
Q1, J1
Registry record
http://www.socialscienceregistry.org/trials/4877

Abstract

Unitary and collective household models assume efficient allocation of resources within the household. A body of literature indicates, however, that the preferences of the household may be pareto-inefficient. Inefficiencies are potentially associated with substantial losses to individual and household welfare in particular when public goods are underprovided or when household members have different consumption preferences. Using data from 400 households, we test the inefficiency hypothesis using a novel revealed preference approach. We elicit the willingness of women and men to pay to receive a small and instant cash transfer in private instead of in front of their partners. Using follow-up survey data collected a week after the experiment, we analyze the spending patterns of those who receive money in private vs. in front of their partners.

Intervention

We propose to employ a mechanism widely used in the experimental economics literature – the Becker, De-Groot, Marschak (Becker et al. 1964) mechanism, or the BDM. The BDM approach provides an explicit measure of the personal cost of transfers of private income to the public pool. The information provided by the BDM approach utilises a direct mechanism which is aligned with real-world decisions regarding entering income into the public account of the household or keeping it private. In this case the willingness to pay (WTP) for keeping endowments private reflects both the difference in preferences for expenditures combined with a lack of agreement on how income should be spent within the household (i.e. allocated to public versus private expenditures).

Endowments used for these components will be worth approximately 1 to 3day’s labour wages, or 15,000 and 30,000 Ushs. The person obtaining the initial endowment (the ‘participant’ in the BDM game) will be randomly selected using stratification across genders (male vs. female) ensuring balance in the sample of male-senders and female-senders (200 of each) for the BDM mechanism treatments.

Experimental design

We randomly assign either male or female household head of each household in our sample to participate in the BDM treatment (400 households). 200 male household heads in the BDM treatment represent a control group for 200 female household heads in the BDM treatment.

Endowments used for these components will be worth approximately 1.5 or 3-day’s labour wages (15,000 or 30,000 Ushs). We will randomly assign either 15,000 or 30,000 Ushs to BDM treatment participants. 200 participants will receive 15,000 and 200 participants will receive 30,000 Ushs endowment.

Design details

Randomization method
Randomization of treatment assignments is done by a computer
Randomization unit
individuals
Clusters
NA
Observations
400 households
Arms
200 households control 200 households treatment
Primary outcomes
Efficiency losses in the household, expenditure patterns (public vs private)

Cited by

No citation data available for this registration yet.

Keywords