Who in the household has decision making power over various things (kids going to school, health seeking behavior of individual members) either alone or jointly with someone else in the household makes up a set of questions that often find their way into surveys (e.g. a version is included in most Demographic and Health Surveys). An interesting new paper by Amber Peterman and coauthors takes a hard look at these questions and what they might, or might not, be telling us.
Let’s start with some theory and motivation. We might care about this as some indicator of the bargaining process as it relates to an outcome such as the efficiency of allocation within households. However, this is still a un- (or at least less) developed area of theory and so, using it for this purpose is still quite a stretch.
What this question may more readily provide is a direct indicator of agency, an end in itself. When an individual has the power to exert control over a decision that affects his/her welfare (and I am including the welfare of her/his children in that), then she/he is better off. So, when this question comes up on a survey, the fact that the respondent has either sole or joint decision making power in some domain of the household is a marker of agency or, more broadly, empowerment.
Peterman and co. unpack this question in four interesting ways. First, they give us some indication of how these indicators look in three countries: Ecuador, Uganda and Yemen, using respondents selected for evaluations of cash and food transfer programs. Second, they use a framing experiment to see if how the question is set up matters for the response that we get. Third, they go through some correlates to see what seems to be correlated with various uses of this measure. Finally, they look to see if the transfer programs they are evaluating (in other papers) move the needle.
Before looking at some numbers, it’s worth going over the questions that Peterman and co. are looking at (not least because they are some of the folks who have thought about these questions more than the average). The key questions they start with are: ““Who in your household usually has the final say in the following decision?” and is asked across the following domains: (1) woman’s own work for pay; (2) children’s education; (3) children’s health; (4) woman’s own health; (5) small daily food purchases; (6) large purchases of items like furniture, cattle, TV, or other assets; (7) bulk food purchases; (8) use of family planning; and (9) opening of bank accounts or borrowing money.” Women are given the following choices: they alone decide, that their spouse decides, that they decide together, someone else in the household decides or that the woman decides with someone else in the household. Then Peterman and co. go on to ask if there’s been a disagreement about these decisions in the recent past and, if so, who the final decider was (and also they hypothetical what if there was a disagreement). Finally, they throw in a more aspirational type of question by asking (in 2 of the countries): “In an ideal situation, who in your household would make the decision.”
As indicated above, it’s not clear whether joint or sole is “better” so Peterman and co. construct counts and indices of these answers which look at sole decisions and sole plus joint decisions. Looking at just the sole decision making, the domain which has the highest level of this across countries is related to the woman’s own health (74 percent in Ecuador, 66 in Uganda, and 47 in Yemen). The lowest is large asset purchases (34 percent in Ecuador, 37 in Yemen, 19 in Uganda). When they add the domains up, women have sole decision making in 4.5 of 9 domains in Ecuador, 2.54 of 6 in Yemen and 2.49 of 6 in Uganda (they didn’t ask about all domains in all countries). Adding joint to sole decision making gives a significant boost in Ecuador (7.5/9) and in Uganda (4/6) but less so in Yemen (3/6). For the final say after disagreement, Ecuador comes in at 6 of 9 domains where the woman is the decider, and Yemen 2.4 (it wasn’t asked in Uganda).
They then use these counts (including the ideal decision maker as well) to construct indices. And what is striking is how low the correlations between these indices are. For example, the correlation between sole decision making and the sole plus joint decision making is only 0.31. This implies that whichever way we choose to interpret these answers has significant implications for how households will rank.
The next question is: does framing matter? In Uganda, Peterman and co. randomly asked the questions in three ways: with a positive framing (many women in Uganda are able to exert control….), a negative framing, and no framing. What surprised me is that (in Uganda at least) this framing seems to have no effect on responses – there’s no significant difference across any of the measures. Maybe this fact, in itself, is a measure of empowerment?
Peterman and co. then go on to look at how these decision making measures correlate with other variables that are sometimes used to proxy for women’s empowerment or household welfare. Not much shows up as significant; the only factor that matters across countries is that women’s age is correlated with more decision making power.
Finally, Peterman and co. turn to the transfer programs. One of the neat things here is that all of their transfer programs had some (randomized) variation in what folks (the women in Ecaudor and Uganda, men in Yemen) were given – mostly varying between cash or in-kind transfers. For Ecuador, the transfer program doesn’t significantly shift decision making power. In Yemen, food transfers, relative to cash, lead to an increase in women’s ideal decision making. So this is shifting aspirations, if not actual decision making. In Uganda, cash transfers (relative to both food and the control group) leads to an increase of about 0.5 decisions that women decide either alone or with their spouse (there is no significant increase in sole decision making). So these women are joining their husband in more decisions.
As Peterman and co. say, this paper raises more questions than it answers. These measures are clearly picking up something, but more work needs to be done to understand what it is and how it relates to other measures of agency or empowerment, as well as to how households work. For example, some of the coauthors on this paper have another paper looking at the effects of the Ecuador transfer program on intimate partner violence and find that the program significantly reduced it. But the decision making variables don’t seem to be moving in the same way. On another dimension, the broader findings (plus a nice example in the paper) also suggest a lot of within- (as well as across) country heterogeneity in how couples do or do not work together in different domains that is worth understanding. It would also be nice to also have the partners’ responses and to see whether they have the same view, and then what we could learn from the cases where they do versus those where they don’t. And, finally, once we’ve done that, it might be a good idea to understand the role that others in the household might play in these decisions. As in my own relationship, the journey to better understand household decision making (and how it matters), continues.
Let’s start with some theory and motivation. We might care about this as some indicator of the bargaining process as it relates to an outcome such as the efficiency of allocation within households. However, this is still a un- (or at least less) developed area of theory and so, using it for this purpose is still quite a stretch.
What this question may more readily provide is a direct indicator of agency, an end in itself. When an individual has the power to exert control over a decision that affects his/her welfare (and I am including the welfare of her/his children in that), then she/he is better off. So, when this question comes up on a survey, the fact that the respondent has either sole or joint decision making power in some domain of the household is a marker of agency or, more broadly, empowerment.
Peterman and co. unpack this question in four interesting ways. First, they give us some indication of how these indicators look in three countries: Ecuador, Uganda and Yemen, using respondents selected for evaluations of cash and food transfer programs. Second, they use a framing experiment to see if how the question is set up matters for the response that we get. Third, they go through some correlates to see what seems to be correlated with various uses of this measure. Finally, they look to see if the transfer programs they are evaluating (in other papers) move the needle.
Before looking at some numbers, it’s worth going over the questions that Peterman and co. are looking at (not least because they are some of the folks who have thought about these questions more than the average). The key questions they start with are: ““Who in your household usually has the final say in the following decision?” and is asked across the following domains: (1) woman’s own work for pay; (2) children’s education; (3) children’s health; (4) woman’s own health; (5) small daily food purchases; (6) large purchases of items like furniture, cattle, TV, or other assets; (7) bulk food purchases; (8) use of family planning; and (9) opening of bank accounts or borrowing money.” Women are given the following choices: they alone decide, that their spouse decides, that they decide together, someone else in the household decides or that the woman decides with someone else in the household. Then Peterman and co. go on to ask if there’s been a disagreement about these decisions in the recent past and, if so, who the final decider was (and also they hypothetical what if there was a disagreement). Finally, they throw in a more aspirational type of question by asking (in 2 of the countries): “In an ideal situation, who in your household would make the decision.”
As indicated above, it’s not clear whether joint or sole is “better” so Peterman and co. construct counts and indices of these answers which look at sole decisions and sole plus joint decisions. Looking at just the sole decision making, the domain which has the highest level of this across countries is related to the woman’s own health (74 percent in Ecuador, 66 in Uganda, and 47 in Yemen). The lowest is large asset purchases (34 percent in Ecuador, 37 in Yemen, 19 in Uganda). When they add the domains up, women have sole decision making in 4.5 of 9 domains in Ecuador, 2.54 of 6 in Yemen and 2.49 of 6 in Uganda (they didn’t ask about all domains in all countries). Adding joint to sole decision making gives a significant boost in Ecuador (7.5/9) and in Uganda (4/6) but less so in Yemen (3/6). For the final say after disagreement, Ecuador comes in at 6 of 9 domains where the woman is the decider, and Yemen 2.4 (it wasn’t asked in Uganda).
They then use these counts (including the ideal decision maker as well) to construct indices. And what is striking is how low the correlations between these indices are. For example, the correlation between sole decision making and the sole plus joint decision making is only 0.31. This implies that whichever way we choose to interpret these answers has significant implications for how households will rank.
The next question is: does framing matter? In Uganda, Peterman and co. randomly asked the questions in three ways: with a positive framing (many women in Uganda are able to exert control….), a negative framing, and no framing. What surprised me is that (in Uganda at least) this framing seems to have no effect on responses – there’s no significant difference across any of the measures. Maybe this fact, in itself, is a measure of empowerment?
Peterman and co. then go on to look at how these decision making measures correlate with other variables that are sometimes used to proxy for women’s empowerment or household welfare. Not much shows up as significant; the only factor that matters across countries is that women’s age is correlated with more decision making power.
Finally, Peterman and co. turn to the transfer programs. One of the neat things here is that all of their transfer programs had some (randomized) variation in what folks (the women in Ecaudor and Uganda, men in Yemen) were given – mostly varying between cash or in-kind transfers. For Ecuador, the transfer program doesn’t significantly shift decision making power. In Yemen, food transfers, relative to cash, lead to an increase in women’s ideal decision making. So this is shifting aspirations, if not actual decision making. In Uganda, cash transfers (relative to both food and the control group) leads to an increase of about 0.5 decisions that women decide either alone or with their spouse (there is no significant increase in sole decision making). So these women are joining their husband in more decisions.
As Peterman and co. say, this paper raises more questions than it answers. These measures are clearly picking up something, but more work needs to be done to understand what it is and how it relates to other measures of agency or empowerment, as well as to how households work. For example, some of the coauthors on this paper have another paper looking at the effects of the Ecuador transfer program on intimate partner violence and find that the program significantly reduced it. But the decision making variables don’t seem to be moving in the same way. On another dimension, the broader findings (plus a nice example in the paper) also suggest a lot of within- (as well as across) country heterogeneity in how couples do or do not work together in different domains that is worth understanding. It would also be nice to also have the partners’ responses and to see whether they have the same view, and then what we could learn from the cases where they do versus those where they don’t. And, finally, once we’ve done that, it might be a good idea to understand the role that others in the household might play in these decisions. As in my own relationship, the journey to better understand household decision making (and how it matters), continues.
Join the Conversation