Causes and Consequences of Hidden Income

How to cite

Zhang, Sally. 2022. "Causes and Consequences of Hidden Income." AEA RCT Registry. AEARCTR-0009348. https://doi.org/10.1257/rct.9348-2.0

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

Sally Zhang*

* Primary investigator

Registry ID
AEARCTR-0009348
DOI
https://doi.org/10.1257/rct.9348-2.0
First registered
Last updated
Status
Completed
Trial dates
2022-05-10 – 2022-07-01
Countries
Kenya (Busia County)
Registry record
http://www.socialscienceregistry.org/trials/9348

Abstract

Why do couples hide income and what are the possible consequences? I hypothesize that couples hide income from each other because they have misaligned preferences, and an individual is more likely to hide from their spouse if they don’t have a say over how to allocate household resources. I recruit married couples from Kenya to a lab experiment, where they allocate lab endowment between a private pot and a household pot. I study how exogenously unobservable income affect individual’s choices to share income and to consume.

In addition, I measure spousal hidden income using a household survey. I ask each individual to report their own income, savings, transfers and expenditures, as well as their spouse's income, savings, transfers and expenditures. I hypothesize that individuals underestimate spouses income, and they may also underreport their own income when asked in front of spouse.

Intervention

Experiment:
1. Observability of income: experimental income is observable to spouse or unobservable to spouse (within-subject)
2. Share of household pot: the individual receives 25%, 50% or 75% of the household pot, with the spouse receiving the rest (between-subject)

Survey design:
1. Self report vs. other report: each respondent is asked to report their own income/expenditures/savings/transfers and estimate those of spouse (within-subject)
2. Privacy of survey condition: interviewed privately or in front ot spouse (within-subject)

Experimental design

The participant receives an endowment and decides to allocate their money between two pots. Contributions to the private pot will not be observed by the spouse. Contributions to the household pot will be multiplied by 2x and divided between the participant and their spouse. Contributions to the household pot will be revealed to both spouses. A participant that maximizes household income should allocate all experimental endowment to the household pot.

The participant is told that they will be able to purchase some vouchers in the lab. They can either purchase these goods with allocation from the private pot, in which case, their consumption will not be seen by the spouse. They can also purchase any of these goods together with their spouse after their investment proceeds from the household pot is distributed.

The game will be played for six rounds. One of the rounds for the participant and one of the rounds for each spouse will be randomly implemented.
1. Three rounds of fixed endowment (100 Ksh., 150 Ksh., 250 Ksh.): endowment is observable to spouse.
2. Two rounds of lottery endowment: spouse will know that participant drew from a lottery but not the realization. The lottery pays out 0 Ksh. with 10% probability, and 100 Ksh. / 150 Ksh. / 250 Ksh. each with 30% probability.
3. One round of choice: participant chooses between 150 Ksh. fixed endowment (observable to spouse), or lottery endowment (realization unobservable to spouse). The choice is unobservable to the spouse.

Participant couples will be cross randomized in division of the household pot. Husbands and wives will receive 25%/75%, 50%/50%, or 75%/25% respectively of the household pot.

To summarize, a participant who maximizes household income allocates all endowment to the household pot. There are three reasons why a participant may allocate endowment to the private pot: (1) control of resources, as the participant retains full control over resource in the private pot, (2) hiding income, as the resource in private pot is unobservable, and (3) hiding consumption, as the consumption choices from private pot is unobservable. I study whether participants allocate more to the private pot when income is unobservable. In addition, I study whether they choose different types of gifts when income is unobservable and whether they behave differently depending on the share of the household pot they will receive.

Using the household survey, I study whether couples have incomplete knowledge of spousal income, and whether misestimation of spousal income is due to hidden income. The couples will first be surveyed separately, where they are asked about their own income, expenditures, savings, transfers, as well as their knowledge of their spouse's income, expenditures, savings and transfers. At the end of the study, the couples will be asked a few questions together.

Design details

Randomization method
Randomization done in SurveyCTO
Randomization unit
Couple level randomization for between-subject treatment
Clusters
510
Observations
1020
Arms
170 husbands get 75% of household pot, wives get 25% of household pot 170 wives get 75% of household pot, husbands get 25% of household pot 170 husbands get 50% of household pot, wives get 50% of household pot
Primary outcomes
Experiment: allocation to private pot, experimental earnings, consumption of in-lab gifts Household survey: self-reported income, expenditures, savings and transfers; other-reported income, expenditures, savings and transfers

Cited by

No citation data available for this registration yet.

Keywords