Spouses’ Relative Influence in Household Decision-Making: Does decision domain make a difference?

This study does not consider a randomized intervention. This is a lab-in-the-field study where the intervention is a decision-making experiment with spouses to measure relative influence in different decision domains.

The main question of this study is whether spouses’ relative influence in household decision-making differs between decision domains; in this study an investment decision and a household contribution decision. The investment decision concern an investment in a lottery and the contribution decision concern a voluntary contribution to a local social fund for poor households in the village to enjoy next years TET holiday. Decisions are monetary incentivized. Furthermore, this study investigate what socioeconomic factors that determine relative influence in household decision-making. Specifically, if the importance of socioeconomic factors differ between decision domains.

All included couples in the study will first make two individual experimental decisions and then make the same two experimental decisions jointly.

In the individual investment decision subjects receive a personal endowment of 115 000 VND and decide what amount ‘e’ he/she wants to invest in a lottery. If the subject wins the lottery, he/she receives twice the amount invested. If the subject loses the lottery, the investment he/she made is lost and the subject only receives the (115 000-e) VND he/she decided not to invest. The investment decision is first made individually and then jointly. In the joint decision the spouses must agree to make the same investment in the lottery.

In the individual contribution decision the subjects are informed that the household receives an endowment of 115 000 VND and the subjects decide the amount "e" of the household endowment to voluntary contribute to a social fund for the poor households in the village. The amount contributed to the social fund will be multiplied by 1.2 in the experiment. The poor households in the village will receive a cash donation from the social fund ahead of next year’s TET celebration. Prior to making the decision participants are informed about the total number of households in his/her village that are included in the study and hence will make the same contribution decision. Subjects are also informed about the approximate number of households in the village that are classified as poor and hence will be the beneficiaries of the social fund. The contribution decision is first made individually where each spouse makes an individual decision about what amount of the household endowment he/she want to contribute on behalf of the household. The amount not contributed to the social fund is split equally between the spouses. In the following joint decision, spouses must agree on the amount of the household endowment of 115 000 VND that they want to contribute.

The order of the investment and contribution decision in the experiment is randomized at the household level. This is an attempt to alleviate concerns that my experimental measures of domain specific relative influence in household-decision making will be biased due to order effect and turns taking in the experimental decisions.

2021-05-06

2021-05-25

The key outcomes are the following:

a) Relative influence in investment decision
b) Relative influence in contribution decision
c) Difference between domain specific relative influence in household decision–making

The key approach to measure the primary outcomes are the following:

a) Relative influence in investment decision: Calculated based on spouse’s individual experimental decision and the spouse’s experimental joint decision. A relative influence parameter is calculated for each household using the formula; (Joint decision-Husband’s individual decision)/(Wife’s individual decision-Husband’s individual decision).

Joint decision refers to the amount "e" of a personal (one for each spouse) endowment of 115 000 VND that the spouse’s jointly decide to invest. That is, husband and wife must agree to invest the same amount "e" of their personal endowment of 115 000 VND. The husband’s individual decision refers to the amount "e" of his personal endowment of 115 000 VND that the husband individually decides to invest. The wife’s individual decision refers to the amount "e" of her personal endowment of 115 000 VND that the wife individually decides to invest.

b) Relative influence in contribution decision: Calculated based on spouse’s individual experimental decision and the spouse’s experimental joint decision. A relative influence parameter is calculated for each household using the formula; (Joint decision-Husband’s individual decision)/(Wife’s individual decision-Husband’s individual decision)

Joint decision refers to the amount "e" of a household endowment of 115 000 VND that the spouse’s jointly decide to voluntary contribute. That is, husband and wife must agree on an amount "e" of the household endowment of 115 000 VND to voluntary contribute. The husband’s individual decision refers to the amount "e" of the household endowment of 115 000 VND that the husband individually decides to contribute. The wife’s individual decision refers to the amount "e" of a household endowment 115 000 VND that the wife individually decides to contribute.

With joint decisions in between or equal to any of the two individual decisions the two relative influence parameters provide measures of relative influence, where the wife’s relative influence is increasing with parameter value. This means that the wife’s relative influence in the joint decision is increasing in the relative influence parameter. However, existing experimental research conducting similar experiments to mine show that a non-trivial proportion of couples make joint decisions out of the range set by the individual’s choices. The strategy to deal with such choice shift will be to set the relative influence parameter to 0 or 1 depending on which spouse’s individual decision is closer to the joint decision. In a robustness section, all relative influence measures out of range will be excluded.

Aggregate measures of relative influence may mask within household heterogeneity. A third key outcome is the within household difference in relative influence in the two decision domains:

c) Difference between domain specific relative influence in household decision–making: Calculated as Relative influence in contribution decision parameter-Relative influence in investment decision parameter.


The study includes analysis of the following main research questions and hypothesis:

1) Do husbands have a stronger relative influence in household decision-making independent of decision domain (investment and contribution decision)?
Hypothesis 3: Husbands have a stronger influence than wives in household decision-making irrespective of decision domain.

2) To what extent are relative influence in household decision-making between decisions domains (investment and contribution decision) correlated?
Hypothesis 2: Spouses' relative influence in different decision domains are positively correlated.

3) Does spouses' relative influence in household decision-making differ between decision domains (investment and contribution decision)?
Hypothesis 1: Spouses' relative influence in household decision-making differ between decision domains.

4) What background factors determine relative influence in household decision-making and do these factors differ between decision domains (investment and contribution decision)? Of particular interested is the association between survey measures of spouses’ self-reported responsibility for the household’s social relations.
Hypothesis 4: Wives relative influence in household decision-making increase in her responsibility for the household’s social relations.

Key explanatory variable:
a) Responsibility for the household’s social relations

Other key variables collected:
b) Knowledge difference voluntary contribution decision
c) Beliefs about spouse's individual decisions
d) Social distance to beneficiaries of the social fund
e) Most risk averse spouse

Key explanatory variable collected and constructed:

The main explanatory variable of interest in this study is spouses’ responsibility for the household’s social relations. This measure is constructed from a set of 4 questions in an individual survey where the subjects answers questions about which household member is responsible for the social relationships with relatives, friends and neighbors with respect to i) attending weddings, funerals and death ceremonies ii) keeping contact and socializing iii) economic transactions and iv) discussing important matters. These questions are answered by both husband and wife individually and include the possible answers “Only you” “Mostly you” “You and spouse together” Mostly your spouse” “Only your spouse” “Other person”. If subject answer “Other” the subject get a follow up questions where he/she is asked to think only in terms of oneself and the spouse without the option of answering “Other”. Both husbands’ and wives raw answers will be recoded on a categorical scale from 1 to 5;
1=Only husband
2=Mostly Husband
3=Husband and wife together
4=Mostly wife
5=Only wife

Based on the set of 4 questions in the individual survey I construct 3 different measures of spouses’ responsibility for the household’s social relations: The first measure put equal weight on the husband’s and the wife’s individual answers, the second measure is based exclusively on the wife’s answers and the third measure is based exclusively on the husbands answers.

I consider the measure applying equal weight to both spouses’ individual answers the main measure since this is assumed to give the most fair description of actual responsibility for the household’s social relations. However, for exploratory purpose I consider the measures based on the wife’s individual answers and the husband’s individual answers respectively since what matters for relative influence in joint decision making may be individual perceptions of the responsibility for the household’ social relations.

Other key variables of interest collected and constructed:

Knowledge difference voluntary contribution: The study includes an individual knowledge quiz including three monetary incentivized questions with respect to the own household and other household previous voluntary contributions to a village social fund for the poor to enjoy the TET holiday. Based on the number of correct quiz answers I will construct a variable measuring the knowledge difference between the spouses equal to Wife’s number of correct guesses-Husbands number of correct guesses.

Beliefs about spouse's individual decisions: The study includes a part where spouses make monetary incentivized guesses about his/her spouse’s individual experimental decisions. I construct two sets of two dummy variables. In the first set the dummy variables each of them are equal to 1 if the spouse made the correct guess about the amount his/her spouse invested and contributed respectively in the individual experimental decisions. In the second set of dummy variables each of them are equal to 1 if the spouse's individual investment and contribution decisions respectively is within the same range as his/her guess about his/her spouse’s decision.

Social distance to beneficiaries of the social fund: The study includes a question of whether the respondent know anyone that has received TET gifts from a social fund raised by voluntary contributions from the village. Both the husband and the wife answer this question and I construct a dummy variable equal to 1 if the wife's answer is YES and the husband's answer is NO.

Most risk averse spouse: I construct a dummy variable equal to 1 if the wife invest a smaller amount in the lottery in the investment decision than her husband.



Control variables collected about the household and individual characteristics:
Income, education, age, years of marriage, marriage arrangement, number of children, relative of husband live in household and relative of the wife live in household

The study will be conducted with couples in two separate villages in the Mekong River Delta, Vietnam. From each village 108 households are randomly selected from the Vietnam Census 2019 data. Enumerators collect the data in selected households in teams of two enumerators, one female enumerator and one male enumerator. The study is conducted in respondents’ home and consist of different parts, the order of the different parts is fixed for all households and include the following:

01) Pre-study questions: Check experience of lotteries, experience of voluntary contributions to the village social fund for the poor and social distance to such social fund beneficiaries in the village as well as comprehension of probabilities. (spouses answer separately)
1) Incentivized Knowledge quiz: Measure knowledge about previous voluntary contribution to social funds for poor households to enjoy the TET holiday in the village. (spouses answer separately)
2) Individual survey: Demographic questions, questions about household members responsibility for the households social relations, questions about household-decision making across decision domains. (spouses answer separately)
3) Experiment Individual decisions: Order of investment and contribution decision is randomized at the household level. Detailed information below. (spouses answer separately)
4) Incentivized Guesses about spouse’s individual decisions. Respondent make guesses about the spouse’s individual decisions in previous experimental part. (spouses answer separately)
5) Experiment Joint decisions: Same order of experimental decision as in the individual experimental decisions. Detailed information below. (spouses answer together)
6) Household survey: Demographic questions. (spouses answer together)
7) Determine payment relevant choices. (Both spouses present)

Experiment design individual decisions (Part 3) and Joint decisions (Part 5):
All subjects make one individual decision and one joint decision with respect to two different decision domain; an investment decision, which concern an investment in a lottery, and a voluntary contribution decision to a local social fund for poor households in the village to enjoy next years TET holiday.

For the investment decision in Part 3 (Experiment Individual decision) subjects receive a personal endowment of 115 000 VND and decide what amount "e" he/she want to invest in a lottery. If the subject wins the lottery, he/she receive twice the amount invested. If the subject loses the lottery, the investment he/she made is lost and the subject only receive the (115 000-e) VND he/she decided not to invest. The investment decision is first made individually and then jointly in Part 5. In the joint decision the spouses must agree to make the same investment in the lottery.

For the contribution decision in Part 3 (Experiment Individual decision) the subjects are informed that the household receives an endowment of 115 000 VND and the subjects decide the amount "e" of the household endowment to voluntary contribute to a social fund to the poor households in the village. The amount contributed to the social fund will be multiplied by 1.2 in the experiment. The poor households in the village will receive a cash donation from the social fund ahead of next year’s TET celebration. Prior to making the decision participants are informed about the approximate number of households in his/her village that are included in the study and hence will make the same contribution decision. Subjects are also informed about the number of households in the village that are classified as poor and hence will be the beneficiaries of the social fund. The contribution decision is first made individually where each spouse makes an individual decision about the amount of the household endowment he/she want to contribute on behalf of the household. The amount not contributed to the social fund is split equally between the spouses. In the following joint decision in Part 5, spouses must agree what amount of the household endowment of 115 000 VND that they want to contribute.

The order for which the investment and contribution decision is presented in the experiment is randomized at the household level in an attempt to alleviate concerns that my experimental measures of domain specific relative influence will be biased due to order effect and turns taking the experimental decisions.


After participants has finalized the last part of the survey, enumerators randomly draw which decisions in the experimental parts are randomly selected for payment using bags with balls in different colors. For the investment decision, subjects will get payment either from their individual decisions or from their joint decision. For the contribution decision, subject will get payed either from the husband’s individual decision, the wife’s individual decision or from the joint decision.

To inform the decision of the study, survey material and experimental material pre-testing with students and small field pilots with coupes was conducted in Dec 2020- Jan 2021 (1 pre-testing with students & 2 small field pilots) as well as April 2021 (1 pre-testing with students & one small field pilot).

A computer randomly select the sample of household from the pool of married households in the Vietnamese Census 2019 data. Adding to that, a replacement list of households are randomly sampled. Whenever a household in the original sample of included households decline participation, are absent during the time period of the data collection or have divorced this household will be replaced by a replacement household in the list of replacement households. Replacement is stratified by age since pre-testing of the experiment in the field (conducted in January 2021) suggested that previous replacement procedure induced over-representation of households with older spouses.

Household

No

Not relevant

216 households (implies 432 participating subjects). All subjects make both the investment and contribution decision.

216 households (implies 432 participating subjects). All subjects make both the investment and contribution decision.