What does it Mean to Make a ‘Joint’ Decision? Unpacking Intra-household Decision Making in Agriculture: Implications for Policy and Practice

Abstract

Strategies to empower women in development contexts frequently address their authority to take decisions within their household, including decisions that are taken jointly by couples. Assessing empowerment in joint decision-making has traditionally followed a dichotomous approach: decisions are either joint or not, with the former associated with women’s empowerment. This paper contributes to a more nuanced understanding of the empowerment effects of joint decision-making, based on case study data from Uganda. We present survey data revealing significant gender differences in perception of decision-making over the adoption of agricultural practices and consumption expenses. Women reported joint decision-making more often than men, who presented themselves more as sole decision makers. We supplement the survey data with an in-depth study in Lodi village, where we reconstruct meanings attached to joint decision-making using focus group discussions, a decision-making game and participant observation. Reported joint decision-making included a range of practices from no conversation among partners to conversations where female spouse’s ideas are considered but the man has the final say. The findings suggest that local interpretations of joint decision-making, in at least this case of a dominantly patriarchal context, can limit its potential for assessing women’s empowerment.

1. Introduction

Women’s authority in intra-household decision-making in rural areas has been at the centre of considerable research for development. On the one hand, assessing women’s role in domestic production and consumption decisions in agricultural households has been central to efforts to determine their level of empowerment (Ibrahim & Alkire, 2007; Kabeer, 1999; Murembe, 2015). In these efforts, the degree to which women have sole or joint decision-making authority in different areas is considered to indicate their level of empowerment (Alkire et al., 2013). This stream of research takes women’s empowerment as a desirable goal in itself to which increased decision-making authority contributes. Apart from its intrinsic value, increased decision-making power of rural women has also been found to positively correlate with important development outcomes such as reproductive, maternal, neonatal and child health; increased expenditure on household health and education; and household nutrition (Leigh, Reynolds, & Gugerty, 2017).

Within this context, increasing joint decision-making in intra-household decisions has been identified as a way of transforming the power-relations between men and women (Hillenbrand, Karim, Mohanraj, & Wu., 2015), thereby contributing to both women’s empowerment and improved development outcomes (Ambler, Doss, Kieran, & Passarelli, 2017; Leigh et al., 2017). This understanding has influenced development efforts aimed at improving rural livelihoods (Satterly, 2016; USAID, 2012). Targeting couples’ decision-making, rather than women or men separately, is, for example, a key pillar of the ‘Gender Action Learning System’ (GALS) methodology, which has been used in development programs in several countries (IFAD, 2014a; Mayou & Novib, 2014).

Intra-household decision-making patterns among spouses in agricultural and domestic spheres have traditionally been assessed through surveys based on a dichotomous understanding of joint spousal decision-making: a couple was considered to have taken a decision either jointly or not jointly and a ‘joint decision’ was assumed to reflect more empowerment than a decision reported as being made solely by the male. This concept rests on the assumption that women would participate as equals in the ‘joint decision’. More recent efforts, such as the Women’s Empowerment in Agriculture Index (WEAI), have transcended this dichotomy and introduced scales that allow respondents to differentiate the extent to which spouses participate in decisions (that is ‘not at all’, ‘small extent’, ‘medium extent’, ‘to a high extent’), although this approach still partly leaves open what the different values on the scale actually mean in practice (Alkire et al., 2013).

Traditionally, surveys about decision-making in rural households were administered to one spouse, normally the head of the household. But what transpires when these questions are asked to both spouses within the same household? Recent studies have found that spouses often report differing perceptions on how specific household decisions are taken (Ambler et al., 2017; Deere & Twyman, 2012; Ghuman, Lee, & Smith, 2006; Leigh et al., 2017). These disagreements suggest the presence of gendered differences in the perception of decision-making processes. They also point to the need to consider that ‘joint decision-making’ may involve complex intra-household negotiation processes and that the concept can be interpreted differently depending on the local and social context. Unpacking these differences in spouses’ experiences and interpretations of joint decision-making processes has been identified as a fundamental need for a better understanding of the relationship between participation in decision-making processes and women’s empowerment (Seymour & Peterman, 2018).

This paper aims to add to the literature on intra-household decision-making and how we measure empowerment. We use data from a case study in the rural District of Nwoya in Northern Uganda to develop a more nuanced understanding of ‘joint decision-making’ between spouses. Uganda presents an important case: in both academic and development spheres, women in Uganda have been portrayed as having low levels of decision-making authority within their households and policy and development programs currently aim to increase their decision-making power (IFAD, 2014a; Nwoya, 2015; The Republic of Uganda, 2013), which between rural household spouses typically includes production and consumption decisions. In the case study region, where climate change is negatively affecting the livelihood of smallholder farmers (Mwongera et al., 2014) and where the adoption of climate resilient agricultural practices has become central to major development investments (IFAD, 2014b), a focus on decisions to adopt climate-smart agricultural (CSA) practices (that is practices that provide mitigation, adaptation and increased productivity for farmers, FAO, 2013), along with selected consumption decisions, is suitable to assess the degree of joint decision-making across a range of intra-household decisions. With this focus, our first research question is: Are there gendered differences in men’s and women’s reporting of intra-household decision-making? The second research question aims to unpack the understanding of joint decision-making across a wider range of agricultural and consumption decisions: What are the gender-differentiated meanings attached to taking joint agricultural and consumption decisions? While the first question is addressed with a survey, the second question is answered through an in-depth study based on focus group discussions, a decision-making game and participant observation in the village of Lodi, Nwoya District.

2. Background

The desire to increase joint decision-making power in households has been recognised by policy and development programs around the globe (USAID, 2017; WFP, 2016). In Ugandan policies, joint decision-making as a way to empower women has been introduced fairly recently. The National Country Vision 2040, which provides long-term development paths and strategies, acknowledges the lower participation and influence of women in household decision-making and calls for men and women to be treated as ‘equal partners in development right from the household to the Country level’ (The Republic of Uganda, 2013, pp. 96–97). Similarly, the National Climate-Smart Agriculture Programme 2015–2025 emphasises the cultural limitations that restrict women’s decision-making power in their households (MAAIF, & MWE, 2015, p. 18). Furthermore, the national Climate Change Policy encourages that ‘both men and women participate meaningfully in planning, testing and rolling out adaptation and mitigation activities’ (The Republic of Uganda, 2015, p. 14).

While these policies acknowledge the lower decision-making power of Ugandan women, they do not directly promote joint decision-making at the household level. However, the National Agriculture Sector Strategic Plan (ASSP) 2015/16–2019/20 for Uganda foresees training activities in joint decision-making in all community-based development programmes (Ministry of Agriculture, Animal Industry and Fisheries (MAAIF), 2016, p. 63):

“MAAIF […]will mainstream gender analysis and gender-based budgeting in all community-based development programmes. The activities will include: 1) Facilitating training in joint decision making and planning of the household agroenterprise(s); promoting gender equity in ownership, access and control over production resources such as: land, livestock, agricultural equipment, labour and capital/credit (…)”

Sub-national development plans also emphasise the promotion of joint decision-making processes between spouses. For example, the five-year development plan (2015/16–2019/20) of Nwoya prioritises ‘couple counselling for informed couple decision-making’ (Nwoya, 2015, p. 199). However, these policies leave the concept under-defined and ambiguous. They neither elaborate what type of joint decision-making the trainings and counselling would aim at nor do they specify how much voice and decision-making authority each member of the household should have.

The ambiguous terminology surrounding joint decision-making between spouses reappears in current iNGO and NGO programs in Uganda. An example is the Gender Action Learning System (GALS) methodology, which aims to ‘give women as well as men more control over their lives and catalyse and support a sustainable movement for gender justice’ (Mayoux & Novib, 2014, p. 8). The GALS methodology has been widely adopted in Uganda and is currently used by multiple NGOs and development organisations (for example, Caritas Uganda, Oxfam, IFAD) operating in several districts, including in Northern Uganda. Some of the activities that the GALS methodology proposes, such as the ‘Gender Balance Tree’, have a clear focus on achieving a more equal distribution of ownership and decision-making among couples (Oxfam Novib, 2014). However, the GALS manual does not clearly define ‘joint decision’, leaving the concept open to a variety of interpretations. This ambiguity is transferred to development projects that use the GALS methodology or variations of it. For example, we found no explanation what would be constitutive of a ‘joint decision’, or what a joint decision would entail, in IFAD projects using the GALS methodology in Uganda such as the District Livelihoods Support Programme (DLSP) and the Project for the Restoration of Livelihoods in the Northern Region (PRELNOR) (IFAD, 2013, 2014b, 2016).

Overall, while the promotion of joint decision-making has permeated policy and development practice in Uganda, the relative voice and decision-making authority of the spouses entailed in the concept is rarely an explicit consideration. The ambiguous use of the term extends to the expected contribution of joint decision-making to women’s empowerment, which would vary significantly depending on the quality of women’s participation in a ‘joint’ decision.

3. Examining ‘joint decision-making’ in research for development

The extent to which women participate in intra-household decision-making processes, either individually or with their spouses, is often used as a metric for women’s bargaining power (Leigh et al., 2017) and women’s empowerment (Alkire et al., 2013; Seymour & Peterman, 2018). ‘Empowerment’, following Kabeer (1999), refers to the expansion of people’s ability to make life choices in cases where this ability was formerly denied to them. ‘Bargaining power’ is often used as one of the components to assess women’s empowerment and refers to the ability of one person to exert influence over another during a negotiation process (Alsop & Heinsohn, 2005; Doss, 1996; Ibrahim & Alkire, 2007; Kabeer, 1999; Malhotra, Schuler, & Boender, 2002; Narayan-Parker, 2005). Decision-making authority is commonly used as an indicator of women’s bargaining power (Leigh et al., 2017).

Participation in decision-making is usually measured by examining ‘who’ makes a specific decision. As Doss (2013) explains, survey questions surrounding who in the household makes decisions are believed to capture, at least partially, women’s bargaining power and therefore how involved and influential women are in decision-making. Assessing women’s participation in intra-household decision-making about agricultural production and domestic consumption can be achieved through examining the degree to which they make autonomous or joint decisions, with the literature being unclear which contributes more to women’s empowerment. For example, the WEAI aggregates joint and sole decision-making without further differentiation (Alkire et al., 2013, p. 73), while other studies prioritise sole over joint decision-making (see, for example, de Brauw, Gilligan, Hoddinott, & Roy, 2014).

The way in which ‘joint’ decision-making is measured in nationwide representative datasets is not homogeneous and varies across survey instruments. For example, in the USAID’s woman’s model questionnaire (Phase 7) of Demographic Household Surveys (DHS), ‘joint’ is either both included directly in the question and as an option for the enumerator to code¹ or it is not included in the question and just remains an option for the enumerator to code². In contrast, in the Living Standards Measurement Study Integrated Surveys on Agriculture (LSMS-ISA) – a household survey implemented in eight African countries – questions about production–related decision-making deploy the phrasing “Who are the decision makers concerning … ?”, inviting the respondent to name up to two household members (see, for example, Uganda Bureau of Statistics, 2013).

Whether a survey included ‘joint’ as an explicit category (for example, DHS) or invited respondents to freely name decision makers in the household (for example LSMS-ISA), the subsequent analysis typically categorises responses into decisions take alone by the women, jointly with her spouse, jointly with other members of the household, or woman’s non-participation (see, for example, Leigh et al., 2017; Twyman, Useche, & Deere, 2015). Only few studies – most notably the WEAI and pro-WEAI indexes (Alkire et al., 2013) – analyse the kind and degree of women’s and men’s participation in reported joint decisions. This is problematic since, even though the term ‘joint’ denotes a certain degree of equality of input between the spouses, earlier studies found that this is not always the case (Hindin, 2004). Therefore, it remains unclear to what extent ‘joint’ decisions as measured in representative surveys reflect meaningful participation by women (Seymour & Peterman, 2018).

Surveys assessing reported decision-making patterns of both spouses (see, for example, Ambler et al., 2017; Leigh et al., 2017; Seymour & Peterman, 2018; Twyman et al., 2015) add an additional layer of complexity. They found that spouses often disagree how a specific decision was taken, with levels of spousal disagreement varying across different decision issues and study regions. Recent intra-household decision-making studies conducted in Ecuador, five Asian countries, Bangladesh and Tanzania found, for example, rates of disagreement ranging from 34 to 57 per cent, 25 to 50 per cent, 38 to 55 per cent and 29 to 40 per cent, respectively (Ambler et al., 2017; Deere & Twyman, 2012; Ghuman et al., 2006; Leigh et al., 2017). Women are more likely to report joint decision-making than men, who tend to report sole decision-making (Allendorf, 2007; Ambler et al., 2017; Deere & Twyman, 2012), however, again with variation across types of decisions and study regions. For example, Bradshaw (2013) found that women in Nicaragua largely considered contraception as a ‘female’ decision, while men favoured ‘joint’ for the same type of decision. In contrast, Leigh et al. (2017) in rural Tanzania and Ghuman et al. (2006) in five South Asian countries found that husbands tended to report more authority for their wives than wives reported for themselves.

While some of these asymmetries between spouses’ answers are thought to result from random measurement error, the literature seems to agree that this alone does not account for all the differences. For example, intentional asymmetries were found in Nicaragua, where reported joint decision-making by women was influenced by the presence of their spouses during the survey, indicating that at least some responses might have been given as a result of inhibition (Bradshaw, 2013). Other authors hypothesise that unintentional asymmetries might happen, for example, when decision-making questions are posed generally, without including reference to a specific timeframe, since spouses might respond in reference to different specific instances (Deere & Twyman, 2012; Seymour & Peterman, 2018), or when spouses have different threshold or might attach different meaning to the questions and decision-making processes (Ghuman et al., 2006).

That spouses may disagree on decision-making authority makes it difficult to assess women’s involvement in household and agricultural decisions and by extension their bargaining power or empowerment level. In light of these asymmetries in spouses’ responses, contextualising how decision-making processes take place and capturing the associated meanings that female and male respondents attach to different types of decision-making processes is key. In this sense, complementing quantitative surveys with in-depth qualitative research could provide valuable insights into these differentiated meanings and decision-making negotiation processes (Ghuman et al., 2006; Richardson, 2018; Seymour & Peterman, 2018). In particular, applying an interpretive approach (Wagenaar, 2011; Yanow, 2000) can help unpack the different meanings attached to joint decision-making processes from different angles of inquiry. This would facilitate an examination of decision-making authority that goes beyond considering ‘who’ makes the decisions to exploring ‘how’ decisions are understood to be made. Understanding people’s social construction of meaning and interpretation surrounding joint decision-making will in this way prove useful in considering the potential for employing this concept as a measure for bargaining power, and consequently women’s empowerment, in development settings.

4. Research methodology

This study was conducted in Nwoya, a district located in the Acholi sub-region in the Northern Region of Uganda. Acholi society is patriarchal, with men generally considered to be the heads of the households (Omona & Aduo, 2013). Rain-fed smallholder farming systems are predominant in the district, with groundnut, beans, and rice as the main cultivated crops (Mwongera et al., 2014). Climate change has increasingly affected rural livelihoods in Nwoya and climate-resilient approaches such as CSA have become central to major development programs in the area (IFAD, 2014b; Mwongera et al., 2014). The study first uses a survey to quantitatively examine differences in spousal perception in decision-making patterns, and then a qualitative in-depth study in the village of Lodi to delve into the gender-differentiated meanings that are associated to decision-making processes.

4.1. Survey

To answer the first research question (Are there gendered differences in men’s and women’s reporting of intra-household decision-making?), we use quantitative data from a survey conducted in December 2014 in the District of Nwoya. Through a probability proportionate to size sampling approach (Skinner, 2016), we sampled a total of 585 households from the four different sub-counties of Nwoya (Figure 1). We administered the survey separately, and in parallel, to the principal male and female decision-maker in each household. Previously trained enumerators conducted the interviews face-to-face and in Acholi, a dialect of the Luo language. Questionnaires took on average 180 min to complete. The survey had a focus on intra-household decision-making processes, adoption of CSA practices and agricultural production (Mwungu et al., 2017).

Figure 1. Map of Nwoya District, portraying the surveyed sub-counties (Alero, Purongo, Anaka, Koch Goma) and location of the in-depth study site (Lodi).

From the sample, in 99 households only one principal decision maker was interviewed (either because they were single-headed households or for other reasons), and in 12 the two principal decision makers interviewed were not spouses. This study only considers the households where two spouses were interviewed (n = 474). Our analysis focuses on men and women’s reported intra-household decisions in consumption choices (that is who decides on budget for food cost, expenses for the children, purchases of major household items) and in the three most frequently adopted CSA practices in the area (that is intercropping, fallow and seed selection). For each of the decisions, the principal male and female decision makers in the households were asked: ‘Who decided on … ?’ Interviewees could mention up to three members in their households. The analysis of the survey data was conducted using R version 3.3.0 (R Core Team, 2016) and RStudio version 1.0.136 (RStudio Team, 2016).

4.2. In-depth case study in Lodi village

The second research question (What are the gender-differentiated meanings attached to taking joint agricultural and domestic decisions?) is answered through an in-depth case study in Lodi, located in the Purongo sub-county (Figure 1). Lodi was selected because among the villages included in the survey, it had the highest share of households practicing a range of CSA practices (that is intercropping, fallow, seed selection, channels, covers cropping and live fences). The in-depth study, however, considered decisions beyond the adoption of CSA, covering a wide range of agricultural and domestic decisions. The in-depth study used three main research methods: focus group discussions, a decision-making game and participant observation.

(a) Focus Group Discussions (FGDs) were organised with 16 couples from Lodi, and one additional female that attended the FGD without her spouse. Discussions took place in Acholi dialect and were moderated by five previously trained native research assistants. We purposively selected the couples from the survey participants in order to capture a diversity of decision-making patterns related to the adoption of CSA. Discussions took place concurrently and separately in two groups, one with the male (n = 16) and the other with the female participants (n = 17), to provide spaces in which women and men could express their ideas among peers and would allow for a separate dialogue on gender issues. The FGDs evaluated decision-making perceptions of 24 agricultural and domestic activities. In the two groups, each participant was given three colour cards: green (for decisions made by the man), pink (for decisions made by the woman), and yellow (for decisions made together by the couple). Each of the 24 activities was read out loud and we asked participants to raise a card, depending on with who they instinctively associated the decision-making process. To reduce mutual influence, we asked participants to raise cards at the same time. For each activity, after the number of cards of each colour was noted, follow-up questions were asked to encourage a group discussion.

(b) Decision-making game: The game involved two events with 16 different couples. The first event took place on 22 April 2017, as couples arrived to the FGDs described above. In order to study the couples’ interaction in a concrete decision-making process, we offered them to choose between two varieties of maize and two varieties of beans, after having been briefed about the characteristics of each variety. Seeds were offered to participants as a sign of appreciation for having participated in the FGDs. The second event with the couples took place on 2 and 3 May 2017, during which we asked the male and female spouse individually whether they had felt part of the decision when choosing the seeds. Separating the first and second part of the exercise, while potentially introducing memory bias, was seen as necessary to limit the effect of FGD on participants’ responses.

(c) Participant observation: The lead author and a female Acholi translator shadowed three different couples from Lodi for a period of seven days each. This approach facilitated access to everyday practices of decision-making (both domestic and agricultural activities) and, to an extent, allowed us to grasp the intra-household power dynamics that are involved in these processes. We purposively selected the couples for stating different views and levels of spousal accord in intra-household decision making in the adoption of CSA in the survey used for the quantitative analysis.

5. Results

5.1. Survey results: joint decision-making patterns

Table 1 provides an overview of summary statistics that describe the sample used in the analysis. Demographic profiles of households, disaggregated by spouses’ reported decision-making patterns, were generated through canonical discriminant analysis (Johnson & Wichern, 2002) using the candisc package of R (RStudio Team, 2016) and are presented as Supplemental Material.

Table 1. Summary statistics

Variables	N	Mean	SD
Age male spouse	474	39.37	12.24
Age female spouse	474	33.77	11.64
Education of male spouse	474	7.92	1.77
Education of female spouse	474	8.24	1.18
Total family size	474	6.92	2.68
Proportion of adult men in household (ages 16+)	474	0.25	0.13
Proportion of adult women in household (ages 16+)	474	0.23	0.10
Proportion of young male in households (ages 0–15)	474	0.26	0.18
Proportion of young female in households (ages 0–15)	474	0.26	0.17
Total cultivable land held by household (acres)	473	10.62	13.16
Housing % Dried bricks (wall)	474	97.6	n/a
% Mud & dung (floor)	474	87.34	n/a
% Thatch (roof)	474	97.47	n/a

Source: Own calculations based on author’s survey.

Pearson’s χ² Test returned a test statistic of 357.57 and a p-value of 2.2e-16 for adoption of CSA practices and a test statistic of 23.91 and p-value of 2.6e-05 for consumption decisions, indicating a strong association between gender and perceptions about decision-making. From the analysis of standardised residuals, for both CSA adoption and consumption decisions, men are disproportionately more likely than women to report sole male decision-making while women are disproportionally more likely to report either joint decision-making or sole female decision-making (Table 2).

Table 2. Counts and standardised residuals χ² test for independence of sex and decision-making on adoption of agricultural practices and consumption decisions

		Joint couple		Sole male Spouse		Sole female Spouse		Others
		Count	standardised Residuals χ2	Count	standardised Residuals χ2	Count	standardised Residuals χ2	Count	standardised Residuals χ2
Intercropping	Female (n = 252)	189	2.06	14	−6.03	49	3.92		n/a
	Male (n = 222)	117	−2.20	101	6.42	4	−4.18		n/a
Fallow	Female (n = 198)	146	1.95	25	−4.39	27	3.07		n/a
	Male (n = 170)	85	−2.10	84	4.74	1	−3.32		n/a
Seed Selection	Female (n = 178)	93	1.47	6	−6.86	79	6.25		n/a
	Male (n = 181)	68	−1.46	112	6.81	1	−6.19		n/a
All three CSA adoption	Female	428	3.30	45	−10.01	155	7.72		n/a
	Male	270	−3.45	297	10.48	6	−8.08		n/a
Children expenses	Female (n = 469)	324	0.76	96	−2.81	44	3.52	5	0.23
	Male (n = 467)	296	−0.76	159	2.82	8	−3.52	4	−0.23
Food expenses	Female (n = 474)	164	1.58	38	0.31	270	−1.11	2	−1.00
	Male (n = 471)	125	−1.59	34	−0.31	306	1.12	6	1.01
Major household purchases	Female (n = 259)	152	1.03	87	−1.45	12	−0.25	8	1.59
	Male (n = 304)	152	−0.95	134	1.34	16	0.23	2	−1.46
All consumption decisions	Female	640	1.78	221	−2.96	326	0.19	15	0.47
	Male	573	−1.75	327	2.91	330	−0.18	12	−0.46

Source: Own calculations based on author’s survey.

Comparing differences in the perception of decision-making between spouses within the same household confirmed this pattern (Table 3). For this analysis, only couples in which both members agreed that a specific decision was taken or a specific CSA practice was adopted were considered. Overall, asymmetrical spousal perception (spousal disagreement) of decision-making on CSA adoption was more common (59.3% of couples) than spousal symmetrical perception (40.7%). Regarding the consumption decisions, asymmetrical and symmetrical spousal perceptions constituted 48.8 and 51.2 per cent of the cases, respectively (Table 3).

Table 3. Decision-making perceptions on adoption of agricultural practices and consumption decisions between spouses

	Spousal agreement on DM in CSA adoption				Spousal disagreement on DM in CSA adoption
CSA practice	Husband alone	Wife alone	Joint decision couple	M: Husband alone F: Joint Couple	M: Husband alone F: Wife alone	M: Joint couple F: Husband alone	M: Joint couple F: Wife alone	M: Wife alone F: Joint Couple	M: Wife Alone F: Husband Alone	Others forms of disagreement	Total
Intercropping (n = 160)	2,50%	1,30%	40,60%	35,60%	9,40%	3,80%	6,90%	0,00%	0,00%	0,00%	100,00%
Fallow(n = 99)	8,10%	0,00%	38,40%	33,30%	9,10%	4,00%	7,10%	0,00%	0,00%	0,00%	100,00%
Seed Selection (n = 88)	3,40%	0,00%	23,90%	30,70%	29,50%	0,00%	12,50%	0,00%	0,00%	0,00%	100,00%
Mean	4,33%	0,60%	35,74%	33,70%	14,41%	2,89%	8,38%	0,00%	0,00%	0,00%	100%
	Spousal agreement on DM in consumption decisions				Spousal disagreement on DM in consumption decisions
Consumptiondecisions	Husband alone	Wife alone	Joint Decision couple	M: Husband Alone F: Joint Couple	M: Husband alone F: Wife alone	M: Joint Couple F: Husband alone	M: Joint Couple F: Wife alone	M: Wife Alone F: Joint Couple	M: Wife Alone F: Husband Alone	Others formsof disagreement	Total
Children expenses(n = 465)	9.25%	0.00%	45.16%	21.29%	3.01%	11.18%	6.45%	1.72%	0,00%	1,94%	100,00%
Major household purchases (n = 201)	14.93%	0.00%	33.33%	27.36%	3.48%	14.93%	1.00%	2.99%	0,50%	1,49%	100,00%
Food expenses (n = 471)	0.85%	38.22%	10.19%	2.34%	4.03%	2.55%	13.80%	22.08%	4,25%	1,70%	100,00%
Mean	6.77%	15.83%	28.58%	14.51%	3.52%	8.27%	8.53%	10.38%	1,85%	1,76%	100%

Source: Own calculations, based on authors’ survey. (M = male spouse respondent; F = female spouse respondent).

In the case of the adoption of intercropping and fallow, the most common decision-making pattern was spousal agreement on joint decision-making. This was followed by spousal disagreement with male spouses reporting they solely made the decision and female spouses reporting the adoption decision was jointly made. In decisions about the adoption of seed selection, the latter constituted the most common decision-making pattern. Agreement on joint decision-making was the third most frequent answer for seed selection (Table 3). With regard to decisions on children expenses and major household purchases, the most common decision-making pattern was spousal agreement on joint decision-making, followed by spousal disagreement where male spouse reported own sole decision-making and female spouse reported joint decision-making. Perceived decision-making for food expenses differed considerably from the other types of decisions investigated in the study. The most common decision-making pattern for decisions regarding food expenses was spousal agreement of sole female spouse decision (38.2%), followed by spousal disagreement where the male spouse reported sole decision-making by female spouse and the female spouse reported joint decision-making (22.1%).

We have reason to assume that the differences in the response patterns found across practices partly reflect gendered roles and responsibilities. For example, compared to other practices, women might be relatively more inclined to perceive the adoption of seed selection as a decision solely taken by themselves if they are normally in charge of preserving or selecting seeds in the households. Similarly, the fact that food preparation is a domain predominantly feminine in Acholi culture (Omona & Aduo, 2013) could explain the distinct pattern that decisions on food expenses presented in comparison to the other types of decisions investigated. Such differences in gendered attribution underline the importance of assessing a wide range of agricultural and household decisions in efforts trying to assess women’s bargaining power and women’s empowerment.

5.2. In-depth study: examining gendered perceptions of joint decision-making

5.2.1. ‘Joint decision-making’ as the most frequently reported decision-making pattern by men and women

During the FGDs joint decision-making was the most frequent response to most of the intra-household decisions discussed in both groups. Comparing women’s and men’s perception of joint decision-making (Table 4), the decisions on ‘when to harvest’, ‘to hire labour’ and ‘to start a new agricultural practice’ were perceived more frequently as joint by men than by women. Women explained that since they were more often in the field, they knew when harvest was ready and thus starting to harvest was a decision they could take individually. In contrast, hiring labour was perceived as a decision taken individually by their male spouses: ‘it is men who normally keep our money and savings, so he sees what we have and decides’. Similarly for the adoption of new practices, women explained that men were normally more mobile and had opportunities to identify new technologies so they were the ones that decided to implement them.

Table 4. Perception of joint decision-making from the focus group discussion card exercise

	Husband alone		Wife alone		Joint couple
Type of Decision Male group: n = 17; Female group: n = 16	Male Group	Female Group	Male Group	Female Group	Male Group	Female Group
To buy a cow	1	4	0	0	15	13
To sell a cow	4	4	0	0	12	13
To use own bulls for ox-ploughing	16	17	0	0	0	0
What and where to plant	10	3	1	1	5	13
When to clear the land	1	4	0	3	15	10
How will the land will be cleared	1	4	0	0	15	13
When to start planting	4	0	0	2	12	15
When to weed*	2	0	12	14	1	3
When to harvest*	4	0	0	13	11	4
To hire labour	3	10	0	2	13	5
To leave land fallow	10	2	1	6	5	9
How much, where and when to sell crops	4	0	0	1	12	16
How to spend income from crops or livestock	6	1	0	0	10	16
To start a new agricultural practice*	0	9	0	0	15	8
To start growing a new crop	4	1	0	0	12	16
To get a loan	2	0	0	1	14	16
To sell land	11	14	0	0	5	3
Who will inherit the land	4	0	0	0	12	17
Usage of money from savings group	2	0	0	0	14	17
To use harvest income for personal use	4	0	0	0	12	17
To use harvest income for major expenditures	1	0	0	0	15	17
To use harvest income for school fees and children necessities	2	0	0	1	14	16
What to cook	0	0	13	17	3	0
Average	25.7%	18.7%	7.6%	15.6%	66.8%	68.3%

Source: Focus Group Discussions * indicates n = 15 in male group, one participant left temporarily the group to attend phone calls.

In contrast, decisions regarding ‘what and where to plant’ and ‘to sell land’ were more frequently perceived as joint by women than by men. Elaborating why the decision of what to plant was perceived as joint, one woman explained: ‘men always dictate, when you come up with your idea, they don’t normally accept it if it does not side with theirs’. This perception of a joint decision implied that a female spouse was able to raise her opinions, although the male spouse was the ultimate decision-maker. For the same type of decision, what to plant, the men explained that since it was their land, they decided: ‘the woman does not know my land, it’s me to plan and she should just plant where I show her’. In this case, we see that while men consider that it is their decision because they tell their wives where to plant, many women perceived this as a joint decision as there is an interaction between spouses in the same physical location.

Finally, Table 4 also reveals that women were perceived mostly as decision makers conjointly with their spouses (66.8%) but rarely individually (7.6%), as observed by their partners. This confirms the results obtained through the survey (Table 3). Male spouses perceived women having sole decision-making power only for responsibilities of which women are traditionally in charge, namely cooking and weeding. This contrasts with women’s perceptions of their own decision-making authority. Women saw themselves more often (15.6%) as having sole decision-making authority in particular for some crop production and land management decisions (for example, clearing the land, leaving land fallow, start planting, hiring labour). These gender differences in perceptions were particularly strong with regard to the decision over when to harvest, with 13 out of 16 women associating it as their own decision and none of the men perceiving women as sole decision makers.

5.2.2. Perceived joint decision-making might not involve a conversation between spouses

During the decision-making game, the majority of couples established some type of interactions in deciding which variety of maize and beans to take (Table 5). In most of the cases, the male spouse, who was the final decision maker, largely dominated the discussion regarding the variety of maize to choose. In the case of beans, it was largely the female spouse who would suggest the variety to take while the male spouse, as final decision maker, would approve her choice. We suggest that these differences can be explained by gendered patterns of division of labour in Nwoya: beans are largely cultivated as a crop for home consumption and thus fall into women’s domain, while maize is seen as a cash crop and therefore falls under the men’s domain. However, for both crops, the final decision remained largely with the male spouse.

Table 5. Interactions between couples for seed selection (direct observations)

Couple interactions during seed selection (n = 16)		Maize	Beans
Did not discuss		4	3
Discussed	Man dominated the discussion	8	3
	Woman suggested the variety, husband approves (final decision-maker)	2	10
	Equal say during discussions, man preferred variety chosen	1	0
	Instant agreement (both partners wanted the same variety)	1	0

Source: Own calculations, based on experimental game.

In 4 out of 16 maize cases and 3 out of 16 beans cases, no conversation was established between the couple when selecting the variety of seed (Table 5). However, when the male and female spouses were separately asked ten days later whether they had felt part of the decision, all gave an affirmative answer with the exception of one female spouse. That means that with one exception, spouses still reported that they felt part of the decision-making process even when no actual conversation among the couple had been observed. It remains an open question whether this reflects a social desirability effect or a genuine sense of involvement.

5.2.3. Reported joint decision-making processes are largely dominated by male spouses

5.2.3.1. We decide everything together

Over the three weeks of participant observation, joint decision-making processes were frequently referred to when general discussions about farm and household decisions came up in day-to-day conversations. Both spouses would frequently use variations of the statement, ‘Whatever we do, there must be agreement between me and my spouse’, when asked generally how they took farming and domestic decisions at home. In households 1 and 3, the male spouse repeatedly emphasised the importance of listening to their wives’ ideas for the wellbeing of their households, even at the risk, they explained, of other men in the community thinking that their wives ‘overpower’ them. Finding this reasoning in two households could suggest that decision-making authority is customarily associated with male domination in the case study region or that social norms dictate that men should publicly appear to dominate decisions even if in the intimacy of the household decision-making processes are more equal.

The apparent prevalence of joint decision-making was frequently met with conflicting statements. While generally both spouses would report joint decision-making, this would be frequently contradicted by the male spouse’s accounts of his authoritative decision-making power in the household, when not asked directly about how decisions were taken. For example, the spouses of Household 2 would almost exclusively assert joint decision-making processes when asked directly. However, observations of their actions found a different story. In one episode, while we were in a specific plot inquiring about the future plans for next seasons’ crops, the female spouse asserted that they had not yet made plans for the crops to be planted there. However, the male spouse immediately followed up on her statement by saying, ‘I will be planting cabbages, eggplants and okra here’. This was not an isolated incidence. When the male spouse was not directly queried, he would almost always use the first person singular pronoun (‘An’) rather than first person plural (‘Wan’) for explaining decisions and activities made in the household and farmland. However, when directly queried he would normally attest that decisions in the house were jointly made.

We encountered similar patterns in the other two households. In Household 1, the male spouse regularly made statements that directly opposed an egalitarian decision-making couple, for example: ‘A wife has no right to decide how many children to have, but her acceptance is a blessing to me. I decide on major things for the house, she has to obey me if she wants to stay in my house.’ Similar statements were obtained from the male spouse in Household 3, who had largely reported joint decision-making processes in all the agricultural and domestic activities between the spouses:

“I decide on when to sell the crops and livestock. Women have no rights on any livestock. I also decide when my wife can go visit her family. Normally I let her go once per year. She cannot decide that, if she does, she has broken my rules and I cannot accept that.”

The constant contradictions in perceived decision-making encountered, in which an apparent dominance of reported joint decision-making was frequently undermined by male spouses’ authoritative statements, made us suspect that either reporting joint decision-making within couples could at times constitute the ‘socially correct answer’ that people by default gave in the area to outsiders, or that joint decision-making could be perceived in a different manner by the local Acholi people compared to researchers or development workers. Additionally, the fact that responses and interactions often occurred when spouses were together might have affected the type of response given by female and male spouses in each situation. That joint decision-making processes could be perceived as the ‘socially desirable’ behaviour transpired also during the field observations. In one instance, we witnessed the female spouse of Household 2 telling her (male) neighbour, who was working alone in the field, that if we saw that he was never working peacefully in the field with his wife, we would not include their household in our ‘program’.

What does joint decision making mean?

“Wabedo wani dako na piny ki awace tama ki en oyee tam ma mega”

The Acholi statement above, ‘I sat her down, told her my ideas and she accepted’, was frequently used by male spouses to describe how a process of taking a joint decision would normally take place. Correspondingly, all female spouses regularly reported to feel part of a decision when the husband informed them about what would be done.

In Table 6, we present a selection of quotes that capture different perceptions of joint decision-making. Overall, they show that ‘joint decision-making’ typically involved an actual conversation between the spouses. A perceived joint decision would normally take place when the husband initiates a conversation and informs the female spouse about his decision, a process through which she would feel part of the decision at hand, even if she had not shared her opinion (for example, Table 6 – Household 3, female spouse). We also encountered several instances when the male spouse would listen to the woman’s ideas and sometimes take them into consideration in a spousal discussion even though the opinion of the male spouse ultimately prevailed (for example, Household 3, male spouse). Less frequently, a woman would bring a proposition to her spouse and wait for his approval in order to proceed (for example, Household 2, female spouse). The spousal dialogue of Household 1 (Table 6) depicts a process of negotiation during which the female spouse needs to be convinced why her partner’s ideas are better while the pressure on the woman to accept his ideas is a sign of respect for him as the head of the household. Far from being an isolated case, this limited level of decision-making power of the female spouses was rather frequent. Overall, in the majority of the decisions that we witnessed and inquired about, the male spouses initiated the deliberations with a suggestion and had the final say. These observations display a much higher power of men in joint decision-making processes compared to their female spouses.

Table 6. Selection of quotes on what constitutes joint decision-making

Household 1	Household 2		Household 3
Couple	Male Spouse (MS)	Female Spouse (FS)	Male Spouse (MS)	Female Spouse (FS)
What crops to plant - (Wife) In case we disagree on what to plant in the field, we sit and discuss together - (Husband) I will explain the reasons why I want to plant these crops in this field. After she has understood why it’s a better option then she will agree. - (Wife) But only when I have understood. - (Husband) When I want to do something (take a decision) I explain to her and as the wife who respects her husband she will accept as I’m the head of the household	Budgeting for money ‘I tell my wife to sit down so that we plan on how to use the money. We together plan on how much to spend on school fees, savings, farming and other necessities.’	What crops to plant ‘He normally decides what to plant. But last season it was my idea to plant beans on this field, I told him my idea and he allowed it.’	What crops to plant ‘Most of the ideas are given by me but she can also sometimes give her ideas and I can accept to use her ideas sometimes. So, it’s a joint decision when she gives an idea about it even if it is used or not. But at the end, as the household head, I make the final decision.’	Budgeting for money ‘ We plan together for the money after crop sales. I feel we make the decision together when he informs me about it even if he uses the money the way he wants it. If he does not inform me, then it means he has made the decision alone. If he at least asks me, then I would feel part of the decision even if he makes the final decision.’

Source: Author’s own participant observation.

5.3. What do women understand by joint decision-making?

Examining what women in Lodi considered as a joint decision is fundamental for understanding whether and to what extent promoting joint decision-making in policies and NGO programs could indeed lead to women’s empowerment in this local context. During women’s interventions in the workshop, the decision-making game and throughout the three weeks of participant observation, we were able to identify an array of situations in which women felt that a decision was jointly made. We grouped these into four distinct categories (Table 7). We realised that a perceived joint decision could occur with (section A) or without (section B) an actual conversation between spouses. When a conversation did not take place, the female spouse could feel part of the decision by the fact that the husband had informed her, either before (Table 7, row A.2) or after the decision had already taken place (row A.1). When a conversation between spouses took place, women felt part of the decision, whether they voiced their opinion (row B.2) or not (row B.1).

Table 7. Typologies of perceived joint decision-making by women

Types of joint decision		Illustrative quote
A. No conversation between the couple	A.1: Male spouse ‘informs’ the woman about the decision after the fact	‘Even after my husband has already done something and tells me (after the fact), I feel part of the decision because he has informed me’
	A.2: Male spouse ‘informs’ the woman about the decision before the fact	‘A man will always inform me even if the final decision is made by him, sometimes I keep quiet and when I keep quiet it means that I have accepted his proposal and participated in the decision’
B. Conversation between the couple	B.1: Male spouse ‘informs’ the woman about the decision before the fact. Female spouse’s ideas are neither given nor considered	‘The decision on how to spend the income from crops and livestock is made jointly. In most cases he informs me, and by the fact that he has informed me I feel that I have participated in the decision’
	B.2: Female spouses’ ideas are considered, but man has the final say	‘Anyone can come up with the idea to discuss, we discuss it but it is the man who has the final say. He can either accept or refuse your idea’

Source: Authors’ observations during experimental game, focus group discussions and participant observation.

In most of the perceived ‘joint decision’ stories told by women, female spouses had no say in the discussion (A.1, A.2, B.1). The perception of participating in the decision came through being present in the same physical space with their husband in which he communicated to her, either before or after the fact, about the decision. It was the fact that the man shared the decision with his female spouse that was perceived by her as having participated in the decision. In other instances, the female spouse’s views and opinions were raised during the spousal discussions (B.2) but the male spouse ultimately took the decision. In none of the cases, women had an equal say with their male counterparts.

6. Discussion

In this study, we used a mixed-methods approach to assess intra-household decision-making patterns in rural Uganda and to examine the interpretations of ‘joint decision-making’ applied to a variety of household and agricultural decisions in a local context. Similar to other studies (Ambler et al., 2017; Deere & Twyman, 2012; Ghuman et al., 2006; Leigh et al., 2017), the survey found that a large percentage of couples reported differing perceptions on consumption decisions (48.8% of households) and decisions on adoption of CSA practices (59.3% of households). Except for decisions on food expenses, men tended to report lower levels of women’s participation in decision-making (either individually or joint) compared to women’s reports, a pattern that has been also found in other regions (Alwang, Larochelle, & Barrera, 2017; Ambler et al., 2017; Twyman et al., 2015). The qualitative findings from this study appear to support the previously posed hypothesis that these differences may derive from discrepant understands of what constitutes being part of a decision (see, for example, Ambler et al., 2017; Ghuman et al., 2006). Male spouses, for example, might have reported sole decision-making in situations where they had the final say in the decision even though their spouse might have had some degree of involvement in the decision. Women in turn, might have considered this same situation as a decision that was taken jointly.

The in-depth study allowed us to unpack the concept of joint decision-making and the different understandings of what it means to take part in a decision (see Table 7). The common denominator of these variations of joint decision-making was the unquestioned role of the male spouse as the final decision maker. This emphasises the need to complement questions regarding ‘who’ took a specific decision with questions regarding ‘how’ the decision was taken. The fact that we found no instances in which alleged joint decision-making actually meant a decision in which both spouses had equal say and negotiated the final decision as equals, must raise serious doubts as to whether a focus on spousal joint decision-making – without a prior in-depth knowledge of the context area – can constitute an adequate strategy for promoting women’s empowerment.

In this sense, our findings raise questions about the reliability of data collected solely by surveys (especially those collecting data from only one spouse and without an in-depth investigation of the local context), as there might be gendered sources of response bias that amplify or reduce any underlying differences in perceptions. For example, in this study, if the research had stopped at the survey level, we could have arrived at the conclusion that women were active participants in decision-making processes and thus had a significant level of agency. Similarly, couples in which both spouses agreed that the decision had been jointly made could have been classified as a ‘cooperative spousal relation’. With this, however, we do not want to disregard promising survey-based work investigating relationships between different types of perceived decision-making with autonomous motivation (Seymour & Peterman, 2018); the work resulting from implementing Women’s Empowerment in Agriculture Index (WEAI), which includes more nuanced questions on decision-making and qualitative modules in some of its variations; and the positive correlations found between spousal agreement of decision-making processes and improved development outcomes (Ambler et al., 2017). Rather, we want to emphasise the potential that combining quantitative with qualitative methods, particularly an interpretive approach, has in advancing our understanding of decision-making perceptions, especially in those circumstances where spouses hold differing understandings.

Finally, the fact that joint decision-making processes are being emphasised in policy and development programs in Uganda, including the district of Nwoya, without an assessment of what type of spousal collaboration is sought and what degree of women’s empowerment this entails is problematic. The semantic flexibility resulting from this ambiguity in the conceptualisation of decision-making leaves room for the beneficiaries of development projects to interpret the concept in ways that fit their lived realities. The ambiguous connotation of ‘joint decision-making’ thus enables multiple interpretations to co-exist, including those that imply minimal change to a patriarchal status quo.

7. Conclusions

This paper has unpacked the multiple meanings that are attached to the concept of ‘joint decision-making’ as a strategy for empowering women within spousal households. Contrary to the implicit understanding of national and sub-national development programs and policy discourses on women’s empowerment in Uganda, the perceptions of joint decision-making of women in Lodi did not amount to decisions in which both spouses had an equal say in the discussion and in the final decision.

While no generalisations beyond the community of Lodi can be made, the results obtained through this case study provided interesting considerations as to why certain assumptions in policy and academic literature might be problematic in terms of measurement of women’s empowerment in specific locations. (1) Conceptually, this research has highlighted the relevance of unpacking the different understandings of the concept of joint decision-making and the need to reject the assumption that spouses always have an equal say in these processes, at least in the local context of Lodi. If bargaining power is conceptualised as the ability of one person to exert influence over another in a negotiation process (Doss, 1996; Kabeer, 1999), and the decision-making power of women is used as an indicator, the consequence should be a consideration on what types of perceived joint decision-making actually reflect women’s bargaining power. Clearly, when women report joint decision-making when the actual processes did not involve a conversation with their spouse, such answers do not constitute a suitable indicator of bargaining power and women’s empowerment since in these cases the ability of women to ‘exert influence’ on the decision at hand is extremely limited if not null. (2) Politically, if joint decision-making processes are used in policy and development programs with the assumption that both spouses collaborate equally in decisions, there is a risk to create a ‘dialogue of the deaf’. In this sense, policymakers, development practitioners and local villagers could make reference to the same concept but with profound conceptual differences and implications for the degree in which a spousal joint decision reflects women’s empowerment. Understanding these differing interpretations is necessary to make the introduction of these development notions truly gender-transformative, as interpretations that do not improve women’s empowerment within their households undermine any transformative ambition.

(3) Methodologically, our research has emphasised the added value of supplementing survey-based studies with qualitative in-depth examinations into the processes of decision-making and their differential understanding by spouses. Research on intra-household decision-making could benefit from including both ‘Who’ and ‘How’ lines of inquiry, and from unpacking interpretation of each spouse’s participation in a joint decision. We argue that survey questions that only inquire which members in the household were part of a decision but fail to capture the specific contribution of each member can easily lead to flawed conclusions. Future research using household surveys would in this sense benefit from questions that also target the process through which a decision was taken. Examples of such questions could be: Who had the final say? Did you participate in a conversation about this decision? Were you informed before the decision was taken? Were you informed after? Could you influence this decision if you wanted to? The latter, already in use in the WEAI, acknowledges, for example, preferences for decision-making, rather than assuming that increased decision-making is always desirable, and in turn aims at promoting equal decision-making authority in those decisions where women would want to contribute. Furthermore, qualitative studies could conduct text analysis to examine patterns of discourse, utterances or words that could shed additional light into when decision-making is referred to in different situations.

While this research has raised questions about the reliability of the term ‘joint’ according to which spouse is interviewed and in which country or context, and how the term is currently being interpreted and analysed, we consider that reported joint decision-making is still more desirable than women not participating in decision. In this sense, even if reported joint decision-making, especially by women in the case of Lodi, might not entail an equal say in the discussion it does denote a certain level of perceived agency. Furthermore, we join Seymour and Peterman (2018) in asserting that researchers should put at the forefront of their efforts examinations towards women’s preferences in decision-making, and acknowledge that joint decision-making might not necessarily always be preferred. Notwithstanding this, our research has shown how adding an interpretive inquiry to quantitative studies investigating decision-making authority of women can help to unpack the different meanings and degrees of joint decision-making that exist in a specific context. An ensuing step would be to assess the extent to which each of these degrees of spousal participation in decisions relates to women’s empowerment. This, coupled with more clearly stated goals and strategies by which development programs and policies aim at changing intra-household decision-making patterns, is a prerequisite for more transformative practice towards gender equality.

Acknowledgements

This work was implemented as part of the CGIAR Research Program on Climate Change, Agriculture and Food Security (CCAFS), which is carried out with support from CGIAR Fund Donors and through bilateral funding agreements. For details please visit https://ccafs.cgiar.org/donors. The views expressed in this document cannot be taken to reflect the official opinions of these organisations. We thank Simon Riley for his assistance in developing the R scripts used for part of the analysis in this study. We thank Penninah Aryemo for her invaluable support in the field and all the farmers that collaborated in this study.

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplementary Material

Supplementary Materials are available for this article which can be accessed via the online version of this journal available at https://doi.org/10.1080/00220388.2019.1650169.

Notes

1. For example ‘Who usually makes decisions about [health care]: you, your (husband/partner), you and your (husband/partner) jointly, or someone else?’ with enumerator choices: 1: respondent; 2: wife/partner; respondent; 3: wife/partner jointly; 6: other.

2. For example ‘Who usually makes decisions about major household purchases’?