Printer Friendly

WOMEN EMPOWERMENT AND INTRA-HOUSEHOLD DIETARY DIVERSITY IN NIGERIA.

Introduction

Empowering women is one of the key objectives of development policy in Africa. The United Nations Sustainable Development Goals (SDG5) emphasized on the need to achieve gender equality and empower all women and girls. Achieving these goals has a broader economic development impact. For instance, Alkire et al. (2013) traces female empowerment to improving household productivity. Likewise, female empowerment (1) in the agricultural sector is seen as essential for achieving food security and reduce hunger, as well as enhancing the efficiency of policy interventions (FAO, 2011; World Bank, 2011). More so, gender equality can be achieved through women empowerment for poverty reduction, hunger eradication, fertility decisions, technical efficiency, and improvement in food security (Phan, 2013; Malapit et al., 2015; Sharaunga et al., 2015; Seymour, 2017). These outcomes from empowering women are hinged on the fact that women play important roles in household and care services, which affects a larger proportion of individuals within the society.

Like other developing countries, in Nigeria, one important role of women within households is their ability to ensure and enhance an efficient dietary intake among household members. Most women in Nigeria are largely involved in food production, (2) distribution and consumption. This implies that within households, women are primary decision makers in relation to food (Diego and Quentin, 2010; Efobi, 2016). Meanwhile, the growing demand for food has drastically increased, especially because of the growth of Nigeria's population from about 159 million in 2010 to 182.2 million in 2015. Nigeria is in food-deficit and depends on imports of grains, livestock products and even fish (IFAD, 2012). More so, the extent of nutrition (3) diversity shrinks as food cost becomes expensive, while consumption remains on the increase.

To tackle this situation, there is a renewed interest in furthering women's empowerment status through interventions and development programs. (4) Despite these actions, the need to develop scientific indicators for measuring empowerment and to examine the relationship between these measures and food related outcomes provides relevant policy options that can be used to monitor the interventions to empower women. This paper, therefore, investigates the linkage between empowerment and the nutritional diversity of households using the 2012-2013 World Bank's General Household Survey for Nigeria. The components of the empowerment variable were used to identify how specific domains and indicators are associated with nutrition diversity. Particular attention was directed at understanding how the estimated linkage differs across gender in the household and empowerment index that is similar to construction of a survey-based approach empowerment index of Alkire et al. (2013) was used. The Alkire et al. index assesses empowerment based on five domains such as agricultural production, access to and control over productive resources, control over the use of income, leadership in the community and time allocation.

The modification of Alkire et al. (2013) measure of empowerment considers domains that pragmatically reflect empowerment indices within Nigeria, such as, access to and control over productive resources, leadership in the community, education, information and connection, and insurance. The inclusion of index such as, completion of a post-secondary education, ability to use and own technology that enhances information and connection, and the ability to own an insurance/savings scheme for measuring empowerment is motivated by three reasons. First, the consideration of this study goes beyond households in the agricultural sector and therefore, the need to develop an index that considers other measures of empowerment outcome that though relates to agriculture, but can fit beyond the agricultural sector. Second, some of the development programs that are initiated in Nigeria and directed at empowering women have considered outcomes such as ensuring educational attainment and human capacity development, as well as owning and using connection and information gadgets to enhance social connectivity (see Department for International Development, 2005; Aderinoye et al., 2007). Third, considering insurance as an important outcome for empowerment is receiving considerable attention since it reflects the individual's ability to invest in the future and escape the poverty trap--or out-of-pocket expenditure that weighs much on the household's income (Grown et al., 2006; World Health Organization, 2012; International Fund for Agricultural Development--IFAD, 2012; African Development Bank, 2015). This is particularly linked to countries with high-income inequality like Nigeria.

Studies have shown that this study is the first attempt to examine empowerment and dietary diversity for all household members (and females) in Nigeria, especially using a modified Alkire et al. (2013) empowerment index. This is important therefore, considering the ongoing policy interventions that are targeted at empowering different gender classifications in Nigeria. More so, the approach to this study adds to the scanty literature on measuring empowerment and advances the Alkire et al. (2013) methodology by including other components that directly define the extent of empowerment within the Nigerian context. This study uses an instrumental variable, alongside the local polynomial and the ordinary least square regression, to arrive at the following results: the empowerment index of all household members and across gender is significantly and positively associated with dietary diversities. The dietary intake of households that are female-biased (5) tends to respond positively and significantly to a change to the extent of empowerment. These results explain the different needs of individuals and household composition across gender. Thus, gender should be seriously taken into consideration when considering empowerment alongside improved dietary diversity as this paper suggests.

The sections of this paper are distributed as follows: the next section considers a background to the study that discusses women empowerment and stylized facts in Nigeria. The third section considers conceptual definitions and review of related literature on the relationship between empowerment and dietary diversity. The fourth section focuses on methodology-data, empirical specification, and variables. The results and discussion are presented in the fifth section, while the sixth section concludes with suggestions for further study.

Stylized Facts: Women in Nigeria

Apart from the important role of women in food production and quality of consumption, the need to study the impact of empowerment across gender and on dietary intake is motivated by the following stylized facts. First, challenges that confront women in patriarchal societies, like Nigeria, (6) are enormous. Beyond economic disadvantages, women face higher inequality in school enrolment than their male counterparts. Table 1 presents the primary, secondary and tertiary enrolment status across the gender of household members. There is a consistent gender dimension to the pattern of education enrolment across the years and at the different levels of education. The Table shows that the gender gap for education widens when considering tertiary education. The unequal enrolment of women in tertiary education is often associated with poverty (see Department for International Development, 2005) and cultural stereotypes, as most cultures in Nigeria do not support/encourage girl-child education. At the level of enrollment, decisions to further education almost entirely rest on the individual (and household's support) since the average age for this level of enrollment is 18 years. Individuals of this age are considered as adults and are therefore responsible for the outcome of their livelihood. Therefore the ability to advance in their education to the university (tertiary) level, despite the odds that confront the individual, can be traced to some level of empowerment as well as household effort that intrinsically motivates the individual for this level of achievement.

Secondly, the unequal representation of women and men in school enrolment in Nigeria may have some health effect on the household and children. According to UNICEF (2011), the relationship between women's level of education and children's malnutrition rate is positive. Smith et al. (2003) and Von Grebmer et al. (2009) also observe that in South Asia, the low status of women in education (among others) contributes to chronic child malnutrition and food insecurity. Luchuo et al. (2013) associated the educational status of mothers in Sub-Saharan African countries to nutrition, sanitation and common disease prevention strategies that logically reduce malnutrition-related mortality and morbidity. Yet Asongu and De Moor (2015) and Asongu and Nwachukwu (2016a, b) note that the bulk of inclusiveness in development agenda must include women empowerment.

Figure 1 (Panel a) below, shows a trend of the common diseases across gender of individuals in Nigeria and as reported by the state ministries of health. Female individuals are more prone to diseases than male individuals. More so, females die from these diseases at a higher rate than their male counterparts, and this has been over the years (see Figure 1, Panel b). Among the causes of these deaths include the inability of some women to pay for health care cost and poor nutrition, compared to male individuals (see Olatinwo, 2017). Therefore, when considering health and wellbeing, male individuals in Nigeria outperform the females; these raises concern as to the intensity of the challenges that confront women.

Thirdly, women lack access to economic resources and basic wage earning opportunities compared to men. Figure 2 displays the percentage of men to women who own a house, land and even received wages in the past year. From the chart on Figure 2, men are about 7 times more likely to own a house than women. More so, the percentage of men who own land alone is about 6 times higher than that of women. In a similar trend, fewer women receive wages in the past year compared to men: men are twice more likely to receive wages compared to women. Evidently, this shows that fewer women have access to economic resources and job paying opportunities than men, which contributes substantially to the poverty rate across gender. (7)

Considering these trends and assertions across gender within the household, this study tends to find out the extent to which empowering women will improve dietary intake. Therefore this study provides a background to advocate for possible policy actions that should be directed at women, especially by providing empirical evidence on how empowerment (across gender) can improve intra-household dietary intake.

Conceptual Definitions and Review of Literature

Dietary diversity is important because of its positive health impact like the reduction of the vulnerability of individuals to certain health disorders (Drescher et al., 2007). The dietary diversity index has remained the common approach to the measurement of dietary intake within the household. The index is developed by counting the frequency and number of consumed food items and food groups by household individuals within a given period (see Thorne-Lyman et al., 2010; Taruvinga et al., 2013).

Women play an overwhelming role in ensuring dietary intake that supplies essential food vitamins and minerals within the household (see FAO, 2016). FAO notes: "... despite their role as the backbone of food production and provision for family consumption in developing countries, women remain limited in their access to critical resources and services..." (FAO, 2016: 3). There is the need to ensure that women empowerment is given paramount attention in policy decisions especially in developing countries. Some of the instruments for empowerment that are identified by FAO (2016) include education, training and extension services, access to decision-making responsibility and credit facilities. Apart from the limitation of women's access to resources, considering empowerment of women matters, there are considerable data suggesting that household members act in diversity of manner when making decisions over household matters (including food consumption). For an instance, women within households may not have the same preferences as their male counterparts (Alderman et al., 1995; Haddad et al., 1997). Therefore, the choices of dietary composition will vary across gender and will be further enhanced for females, especially with empowerment (see e.g. Ibnouf, 2009; Sraboni et al., 2014).

Empowerment therefore entails access to both productive and non-productive resources with the motive of improving the value of individual's orientation towards making decisions that affect both the individual and other related entities. Kabeer (2001) have considered empowerment to be the expansion of an individual's ability to make strategic life choices, especially in contexts where such ability had been earlier denied. Bertelsen and Holland (2006) also describe empowerment as the capacity to make effective choices and then transform those choices into desired actions and outcomes. These two definitions consider empowerment as making quality choices, especially pertaining to the betterment of the individual's life. However, measuring empowerment has remained difficult in scientific studies. Some measures identified by Alkire et al. (2013) that are popularly used, especially at the aggregate level include: the ratios of girls to boys in primary, secondary and tertiary education; the share of women in wage employment in the non-agricultural sector; and the proportion of seats held by women in national parliament. These approaches have two main weaknesses. First, they do not capture heterogeneities that exist across individuals and, secondly they do not directly measure the empowerment that different individuals experience per time.

To circumvent the challenge of measuring empowerment, Alkire et al. (2013) developed an empowerment index that considers women in agriculture. The index is survey-based and it is designed to measure the empowerment, agency and inclusion of women in agriculture based on data collected by interviewing men and women within the same households. This index is classified into five domains: women's input in productive decision and their autonomy in production; women's decision over the ownership, purchase and sale of productive assets as well as access to and decision over credit; control over the use of income; women's involvement in groups and ability to speak in the public; and women's decisions over time for leisure and work. This measure of empowerment has gained credence in some studies like Sraboni et al. (2014) and Malapit et al. (2015).

Despite the breakthrough of Alkire et al. (2013) standardized measure of empowerment, the authors acknowledge that empowerment is inherently context-specific and it is shaped by socioeconomic, cultural and political conditions of the specific country or region that is being studied. Thus using a single measure of empowerment for a broad comparison across countries may not be efficient in understanding its impact (see Malhotra and Schuler, 2005). In this current study, some components like women being able to decide on advancing their educational level, owning an/a insurance/savings scheme, and owning and using technology that can enhance information and connections were included in deriving the empowerment index. The approach adopted for this study to the modification of Alkire et al. (2013) is similar to studies like Mutua et al. (2014) who modified the Alkire et al. index to study social and economic empowerment in Kenya.

Studies (e.g. UNCTAD, 2014; Huh, 2016) have shown that indices such as completion of post-secondary school education, using and owning information and connectivity gadgets, and having an insurance/savings scheme are direct measures of the level of empowerment in some developing countries. For instance, patriarchal practices in developing countries like Nigeria, where families invest more in educating the male child (over female), makes considering a post-secondary education an act of quality decision by the female-child and her family, ceteris paribus. Taking up insurance and savings schemes is also considered as an important measure of empowerment in countries where there are incidences of inequality in access to finance (see Buvinic and Furst-Nichols, 2013). Waller (2014) also used savings outcome to measure the extent to which Malawian women are empowered, in terms of self-reliance and the need to improve their social and economic status.

In addition, having access to and using technological gadgets like mobile phones and other connection and information devices is another important indicator of empowerment in this study. Information and communication devices have spread rapidly over the last decade in Africa (Asongu, 2013). (8) Despite this increase, it is noted that not every person owns (or is able to use) a communication/information gadget. Since these gadgets are personal devices, that if owned and not shared, provide the owner with a degree of independence and autonomy, then it becomes necessary to include this variable as an indicator of empowerment. In essence, women owning mobile phones could arise as a result of their level of empowerment. See Buvinic and Furst-Nichols (2013) to further buttress information on how technology adoption and effective use of such technology can be a direct outcome of empowerment.

As a summary, this study assumes that empowering women will result in enhanced household dietary intake. Hence the testable hypothesis that underpins this study include: (i) there is a significant relationship between empowerment and dietary intake for female household members than their male counterpart; (ii) the impact between dietary intake and empowerment is higher for households that are female biased compared to those households that are male biased.

Research Methodology

Data

The data for this study was sourced from the World Bank's Living Standards Measurement Study (LSMS)-Integrated Household Survey, conducted in collaboration with Nigeria's National Bureau of Statistics. The LSMS dataset (9) is a household type of microdata that contains variables that reflect household conditions across the different states of Nigeria. It also includes information on women's contribution to decision making process on land assets, buildings and income usage, data on savings and insurance schemes that are engaged by women (and other household members), as well as their access and usage of information and communication gadgets. Information such as household income, consumption distribution, gender (including that of the household head), count of household asset and infrastructure, are also contained in the dataset.

The latest LSMS_ISA wave (i.e. 2012/2013) household data was used for our analysis. The 2012/2013 LSMS_ISA data consist of 5,000 households and contain other additional data on agricultural activities, other household income activities, and household expenditure and consumption. Specifically, the second wave of the LSMS_ISA data was conducted in two visits (post-planting visit in September-November 2012 and post-harvest visit in February-April 2013). The post-harvest data was used for this analysis because it adjusts for households that have changed location after the post-planting visit.

Empirical Specification

The following equation was estimated to test the relationship of interest:

D.Di = [[beta].sub.i] + [[alpha].sub.i]empowerment + [[theta].sub.i]x + [[epsilon].sub.i]

Where D.Di is a vector of dietary diversity of the individual, [beta], [alpha] and [theta] are unknown parameters that are to be estimated, x is a vector of the household-level characteristics, and [epsilon] is the error term.

The household dietary intake was measured using the count of food groups based on the 7-day recall household food consumption data in the LSMS_ISA survey. The food groups used for this study are based on 11 food categories as identified in the Food and Agricultural Organization (FAO) guidelines for measuring dietary diversity. These categories include cereals, tubers, legumes, meat, egg, vegetables, fruits, oil, sweets, milk and fish. This measure is increasingly been used in computing dietary diversity (see Taruvinga et al., 2013). More so, Sraboni et al. (2014) admonish the use of our measure, other than the calorie availability indicators that would have been a suitable alternative measure. These authors criticized calorie intake measure because it does not reflect the quality of foods available to the household.

The empowerment index, which is the key independent variable for this study, is computed using the individual level data of household's male and female. This study presents the different domains of the empowerment index with their indicators in Table 2. Each of the domains has equal weights, as are the indicators within the domains. Individuals are defined as empowered when the index tilts towards a higher value closer to 1 and are disempowered in the situation when the index tilts away from 1 and towards 0. The different domains of empowerment are defined as follows:

(i.) Education: this domain concerns the individual's decision to pursue a post-secondary education.

(ii.) Resources: this domain concerns ownership, access, and decision-making over productive economic resources in the household such as land, wage income and building.

(iii.) Insurance: this domain considers the individual's ability to make decisions over owning a savings and/or an insurance scheme for future and contingent events.

(iv.) Group activities: this domain contains the individual's involvement in economic or social groups.

(v.) Information/connection: this domain is focused on the individual's ability to own and use information and technology equipment that can enhance information and connection with the immediate and distant environments.

The other control variables in the equation include average age of the household, the share of household members that are educated, household size, the share of females in the household, income per capita of the household, the share of adults in the household and household expenditure on electricity per capita. The summary statistics of the all the variable used are presented in Table 3. Likewise, the statistics for the share of the different domain of empowerment in the overall index is presented in Figure 3.

Estimation Strategy

The relationship between empowerment and dietary intake across households was first tested using the non-parametric regression based on the local polynomial regression approach. This type of regression fits the relationships between the variables of interest, such that separately fitted relationships are obtained at different values of the independent variable, so as to accurately predict the regression lines. This technique has its unique advantage, which includes: first and unlike the parametric linear regression technique, it allows for a relaxation of the linearity assumption and can predict estimators and inference procedures that are less dependent on functional form assumptions (Yatchew, 1998; Frolich, 2006). Second, different forms of relationships can easily be explored between variables of interest, which makes it useful for exploratory data analysis and for practical and policy-relevant analysis. Third, the non-parametric regressions permit, in many cases, an estimation of variables despite the endogeneity status (Frohlich, 2008).

The parametric regression approach (in the form of the Ordinary Least Square (OLS) regression and the Instrumental Variable approach were also used in this study. The OLS is included to provide a baseline analysis of the estimated relationship. However, some household characteristics are affected by the same factors that influence dietary intake. This therefore makes the empowerment variable to be prone to endogeneity issues. This study applies the standard instrumental variable technique to correct for possible endogeneity bias. The following instruments at the household level were adopted: (i) the difference in ages between the primary male and female decision makers and (ii) the type of building in which the household currently resides. The first instrument is computed based on the age difference between male and female household members that are adults and who are capable of making decisions within the household. The motivation for this instrument is that it reflects the relative bargaining strength within the household (see Quisumbing and Hallman, 2005). The type of building the household resides in is indicative index of households' level of physical capital.

For evident reasons, the parametric estimations allow for the option of controlling for other important variables that affect intra-household dietary intake apart from the main explanatory variable (empowerment) for this study. Assuming these variables are not controlled for, the regression analysis in this study tends to be confronted with omitted variable bias and the estimated relationships will not be rightly predicted. Another important advantage of the parametric estimation technique is the capacity to control for possible endogeneity issues that is vivid in our predictive relationships. The likely omitted variable bias is handled with both the OLS and the Instrumental Variable techniques, while the endogeneity issues are taken care of using the instrumental variable estimation techniques.

Results and Discussion

Descriptive and Summary Statistics

The pie chart in Figure 3 shows the size of the different domains of empowerment for our sampled male and female household members. There is no much difference across the domains of empowerment and across the gender of the household members. The connection and information domain had the highest contribution to empowerment. This is followed by control over the resource of the households, which contribute about 24 percent of the total empowerment. The education and group activities domain are the third and fourth important contributors to empowerment, while insurance scheme contributes only a marginal fraction of the pie. The different types of household members are better empowered in terms of information and connection. However, women are mostly disempowered in terms of access to insurance schemes.
Figure 3 Contribution of Each of the Five Domains to Gender Empowerment
in Nigeria

Female                            Male

Insurance,                1.06%   Insurance,                0.77%
Resources,               24.33%   Resources,               25.19%
Education,               10.70%   Education,               10.25%
Information/Connection,  [VALUE]  Information/Connection,  [VALUE]
Group Activities,        [VALUE]  Group Activities,        [VALUE]

Note: Table made from pie chart.


Figure 4 presents the descriptive statistics of the aggregate empowerment variable and dietary intake. On the average the different household types have similar empowerment index but male household members have a marginal higher bar than the female. The slight difference in the empowerment scores across gender of the household members (0.233 for female and 0.238 for male) connotes that on the average and generally, the empowerment scores are low. Household members sample used for this study have only 23 percent positive scores of the 10 indicators of empowerment. Focusing on the average dietary intake displayed in Figure 4.2, female household members' dietary intake is slightly higher than males. Despite that there is no much difference (in terms of evident contrast) across gender of the household members, however, the bar for female members exceed that of the male.

The mean and standard deviation for the entire variables in the econometric model are presented in Table 3. Some important highlights from the Table include: on the average, many of the households do not own an asset. This is not different when considering the households across gender. On the average, the mean age of households is about 17 years, showing that the households mostly consist of younger individuals. About 1 person in the households have post-secondary school education, while 33 percent of the households have individuals that are 18 years and above. The average household size is 4 individuals. This statistics is similar across gender of the household. The income per capita of the household is higher for male than the female counterparts. Overall, the income per capita for the entire household is only 2604.59 local currency, which is equivalent to 16.5 US$, using the exchange rate as at the year of the survey. The empowerment index and the dietary intake are not different from the graphs presented in Figure 4.

Estimation Result

Non-Parametric Regression

The estimation result begins with considering the non-parametric regression lines of dietary intake and empowerment across the gender of the household member. The local polynomial regression is used to estimate kernel-weighted regression lines of dietary intake and empowerment. This type of analysis presents a first-hand display of the regression lines between our variables of interest. The graph presented in Figure 5 reports that household members with low empowerment index have a lower dietary intake. The importance of individual dietary intake constantly increases with rising empowerment index. This evidence is consistent with the predictions that the proportion of individual dietary intake increases with the level of empowerment. FAO (2011) and World Bank (2011) likewise confirm this result that empowerment essentially increases food security and reduces household hunger--for which dietary intake is an important component.

The kernel-weighted local polynomial regression for the variables--dietary intake and empowerment--across the household location (i.e. rural vs. urban) was also presented in Figure 6. This study finds evidence of gender differences across different locations of the individuals. For the individuals in the urban area, it is evident that a unit increase in empowerment has a higher impact on dietary intake for female household member compared to their male counterparts. Evidently, the gap between male and female household member widens as the empowerment index increases. Thus, females are more likely to improve their dietary intake with an increase in empowerment, compared to their male counterparts. A similar trend is seen for the regression line between empowerment and dietary intake for households in the rural area, except for the fact that the gap between male and female shrinks. As the empowerment index increases, the dietary intake for female household member still maintains a marginal gap compared to their male counterparts. One important point to note is that; women improve their diet better than their male counterparts at similar levels of empowerment. This connotes that a higher impact can be achieved on improving the intra-household dietary intake if women are better empowered. More so, the household location does not really matter since a similar impact is achieved for dietary diversity with empowerment.

Parametric Regression

The standard OLS regression (with the household fixed effect) and the instrumental variable estimations are presented in Table 4. This study presents the parsimonious econometric estimates in Table 4, which contains the effect of the empowerment variable on household dietary intake for the entire households, the male and female household member. The OLS results across the three models suggest that generally, the empowerment variable is positively and significantly associated with dietary intake. The coefficient was larger for female household members than male counterparts. In essence, women dietary diversity will be better enhanced with empowerment compared to male individuals.

The instrumental variable (IV) estimations and its diagnostics are presented at bottom of Table 4. The endogeneity test for the IV estimations across the columns of Table 4 implies that the endogenous variables are relevant and are, in fact, endogenous. This supports the need for the IV estimation to handle the endogeneity problem. The over-identification and under-identification test confirm that the instruments are valid and the models are efficiently identified.

The signs and significant values of the coefficient show a similar pattern as in the OLS estimations. From the IV estimation results, it is evident that empowerment has a positive effect on dietary diversity. The magnitude of the coefficient is higher for female household member. In addition, the coefficients in the IV estimation are higher than those in the OLS regression, meaning that neglecting endogeneity of the empowerment measures may underestimate the impact of increasing empowerment on our outcome. In fact, the coefficient for female household members is about nine times higher than that of male. Our finding agrees with studies in Bangladesh that have shown a positive impact of empowerment on dietary intake, with higher impact for females (Kumar and Quisumbing, 2010; Malapit et al., 2015).

A broader model that controls for other explanatory variables is computed and presented in Table 5. The estimations in Table 5 control for other variables that explain the dietary intake across households, but which were not included in the parsimonious estimations in Table 4. From the Table, the sign and significant values of the empowerment variable remained consistent throughout the models and even across the gender of the household members. More so, the coefficient for the empowerment variable for female household members show a higher effect of empowerment on dietary intake compared to the effect observed for male counterparts. Evidently, an increase in female empowerment will result in a significant improvement of about 39 percent of dietary intake. Males will only experience a 14 percent increase in dietary intake with an improved empowerment index.

Moving on to the individual indicators in Table 5, we find out that the share of females in households matters, but not consistent across different households. For male household member, the female share has a positive and significant impact on dietary intake. This is also similar to the size of the household. Household size and share of female do not play a significant role in the extent of dietary intake for female household members. Other variables like income, education and the cost of electricity have a consistent positive and significant role in defining the extent of dietary intake across the different household types. The signs and levels of significance of these variables are consistent with the results of studies like Sraboni et al. (2014) and Malapit et al. (2015). Thus, better human capital development, income and the amount per capita that are spent on electricity significantly increase household dietary intake. The share of adult and average age of the household shows both negative and positive effect; however, their levels of significance were inconsistent across the different columns of Table 5.

The relationship between dietary intake and empowerment is also examined at the household level. Two household groups were considered based on the gender of the household head and the proportion of female in the households. (10) Households with equal representation of female and male individuals are not included in this analysis for clarity of discussion.

From the report in Table 6, the empowerment variable remained consistently positive for households that have larger percentage of female. These households tend to have a positive and significant improvement in dietary intake with an increase in empowerment. Likewise, for households with female heads the empowerment variable remained positive and significant. An improvement in female empowerment will result in a further increase of about 36.5 and 43.2 percent of household dietary intake for households with more women and those with female heads. For the other categories of households (those with a smaller share of female members and male headed households), it is evident that the coefficient of the empowerment variable was not significant, despite that it was positive. This result confirms the earlier finding that when related to food consumption and intake, empowerment interventions that benefit a large proportion of women will be more impactful on the household than when excluding more women. The signs and significant values of the control variables were not discussed since they are not of interest for this study.

This paper also uses an alternative measure to capture the empowerment variable. This new measure consists of the gender parity gap in which the difference between male and female empowerment index are generated. The gender parity gap is equal to zero if the women empowerment score is equal or exceeds the value for males. The results of the regression analysis (both the OLS and IV estimation) are presented in Table 7. This Table indicates that not much difference is observed for the main explanatory variable--empowerment. The coefficient of this variable maintained a positive value across the different estimation technique. One important difference between these results and those reported in Table 5 is that in this new estimation, the empowerment variable lost its significance (based on the OLS estimation). Since this study do not base the inference on this estimation technique and the signs of the variable remained unchanged, it does not raise much concern. A more important point to note from this new estimation in Table 7 is that female empowerment still maintained a higher impact on household dietary diversity, unlike male empowerment.

Conclusion

This paper investigates the linkage between empowerment and the nutritional intake of households using the 2012-2013 World Bank's General Household Survey for Nigerian households. This study also used the components of the empowerment variable to identify how specific domains and indicators are associated with nutritional intake in Nigeria. Particular interest was directed at the differences in the estimated impact across the gender of the household. An empowerment index that is similar in construction to the survey-based approach of Alkire et al. (2013) was used in this study. The testable hypothesis include (i) there is a significant relationship between empowerment and dietary intake for female household members than their male counterpart; (ii) the impact between dietary intake and empowerment is higher for households that have more female and that also have a female head compared to those households that are male dominant

The results from the regression analysis include: First, and as generally expected, average household empowerment is an important determinant of dietary intake. In addition, consistent with gender differences in dietary intake and across different levels of empowerment, the effect of empowering female household members differs considerably from those of male genders. Secondly, this study finds that the empowerment effect on dietary diversity differs across the extent to which households are female dominated. For households with higher proportion of female and as well as households with female head, shows a significant and higher dietary diversity impact, compared to those that are less biased towards female. Thirdly, more empowerment will result to an improved dietary intake, especially when empowerment is directed at females and households that are female dominant.

This result is not in any way suggesting that male empowerment is less important. However, the results suggest that any action taken towards improving female empowerment (compared to male counterparts) will result in a significant and higher impact on the dietary intake of the household. There are two important explanations for the result we have observed. The first is that with empowerment, women will have the capacity to make better choices on the quality and mix of diets that the household will be consuming. This is unlike males, whose main responsibilities in the household are not directly tied to diet or decision over the choice and mix of diets. Second, women play a significant role in household food consumption and production. This makes the female gender to be better acquainted with vast choices of meals that can be advantageous for the household consumption. Therefore when considering the household health through the diversity of diet, it is important to pay much attention and initiate programs and interventions that can empower women.

An important limitation of our study is that this study measured empowerment from a self-reported survey, where the opinion of the household members (affirmatively or otherwise)--in terms of different indices to measure empowerment--was relied upon to determine the extent of their empowerment. Though this approach is widely used in literature, it is important to consider experimental kind of data, where actual observation of empowerment interventions is measured against the average household dietary intake. More so, measure of empowerment used for this study was not in full compliance with the Alkire et al. (2013) measure of empowerment. Two reasons motivate our approach: this study intended to create an empowerment index that reflects the socio-cultural condition of our sample; and due to data constraint, most of the indicators in Alkire et al. (2013) measure of empowerment were not available in the survey data used in our study. Hence, we are not able to adequately compare our result with that of Alkire et al. (2013). Therefore, future studies that will be making use of survey data (apart from the LSMS data) can consider including the empowerment indicators of Alkire et al. (2013) in their survey instruments. This will aid estimations that can be compared with other studies that have used similar empowerment measurement.

Acknowledgements

The first version of this paper was presented at the conference Changing Food Systems in Africa: Agro-ecology, Food Sovereignty and their Roles in Nutrition and Health, organized by Alliance for Food Sovereignty in Africa (AFSA), the Ecological Organic Agriculture Initiative for Africa (EOA-I), AfrONet and IFOAM--Organics International, Ethiopia, 24-26 November, 2016. The second version was presented at the 1st international conference The Use of Tanzania National Panel Survey and LSMS Data for Research, Policy and Development, jointly organized by the World Bank and Tanzania National Bureau of Statistics, Dar es Salaam, Tanzania, February 7-8, 2017. The authors are indebted for the useful comments and suggestions from the participants and resource persons at these important conferences. In addition, the comments from the anonymous reviewers and the journal editor are gratefully acknowledged.

NOTES

(1.) Here empowering women entails having access to productive resources with the motive of improving the value of individual's orientation towards making decisions that affect both the individual and other related entities.

(2.) Nigerian women contribute 70 percent of the total agricultural workforce, 50 percent of animal husbandry related activities and 60 percent of food processing activities (National Coalition on Affirmative Action, 2009).

(3.) Nutrition and dietary diversity will be used interchangeably. They mean the same thing for this study.

(4.) Such as National Gender Strategic Framework, Information, Communication and Value Re-orientation, Capacity Building and Skill Development, and some other Women Empowerment Programs that have been initiated by Federal Ministries that support women and agricultural development in Nigeria.

(5.) These are households that have more female members or where a female is the household head. Further discussions on this category of household are discussed in future sections.

(6.) Nigeria is a lower-middle-income country in West Africa with a per capita income of about 1280 US$ and a poverty rate of about 62.6 percent. In addition, the common feature across the major tribal groups in Nigeria is its patriarchal values. Most of the patriarchal practices demean and clearly deprive female gender of similar opportunity and resources that are accessible to male gender.

(7.) Poverty rate for women is about 70 percent as of 2013.

(8.) For example, the number of mobile phone and Internet users per 100 persons has risen from 12.0 and 2.0 in 2005 to 75.7 and 22.4 in 2015 (World Bank, 2016). Africa also has the fastest ICT growth rate, compared to other regions of the world (see Asongu and Nwachukwu, 2017).

(9.) The dataset, structured questionnaires, manuals and codebook are available online at World Bank Webpage (http://go.worldbank.org).

(10.) The group of households was divided across the sample according to the share of females that are living within the household (i.e. 1 if the proportion of female in the household is higher than 50 percent of the entire household member and 0 otherwise).

REFERENCES

Alsop, R., M. Bertelsen, and J. Holland (2006). Empowerment in Practice from Analysis to Implementation. Washington, DC: World Bank.

Aderinoye, R. A., K. O. Ojokheta, and A. A. Olojede (2007). "Integrating Mobile Learning into Nomadic Education Programmes in Nigeria: Issues and Perspectives," The International Review of Research in Open and Distributed Learning 8(2). http://www.irrodl.org/index.php/irrodl/article/view/347/919

African Development Bank (2015). Empowering African Women: An Agenda for Action. Addis Ababa: African Development Bank Group. http://www.afdb.org/fileadmin/uploads/afdb/Documents/Publications/African_Gender_Equality_Index_2015-EN.pdf

Alderman, H., P. A. Chiappori, L. Haddad, J. Hoddinott, and R. Kanbur (1995). "Unitary versus Collective Models of the Household: Is It Time to Shift the Burden of Proof?," The World Bank Research Observer 10(1): 1-19.

Alkire, S., R. Meinzen-Dick, A. Peterman, A. Quisumbing, G. Seymour, and A. Vaz (2013). "The Women's Empowerment in Agriculture Index," World Development 52: 71-91.

Asongu, S. A. (2013). "How Has Mobile Phone Penetration Stimulated Financial Development in Africa," Journal of African Business 14(1): 7-18.

Asongu, S. A., and J. C. Nwachukwu (2016a). "The Role of Governance in Mobile Phones for Inclusive Human Development in Sub-Saharan Africa," Technovation 55/56(September/October): 1-13.

Asongu, S. A., and J. C. Nwachukwu (2016b). "The Mobile Phone in the Diffusion of Knowledge for Institutional Quality in Sub-Saharan Africa," World Development 86(October): 133-147.

Asongu, S. A., and J. C. Nwachukwu (2017). "Mobile Phones in the Diffusion of Knowledge and Persistence in Inclusive Human Development in Sub-Saharan Africa," Information Development 33(3): 289-302.

Asongu, S. A., and L. De Moor (2015). "Recent Advances in Finance for Inclusive Development: A Survey," African Governance and Development Institute WP 005. file:///C:/Users/Udechukwu/Downloads/SSRN-id2575667.pdf

Buvinic, M., and R. Furst-Nichols (2013). Measuring Women's Economic Empowerment: Companion to A Roadmap for Promoting Women's Economic Empowerment. United Nations Foundation and Exxon Mobil Foundation. http://www.womeneconroadmap.org/sites/default/files/Measuring%20Womens%20Econ%20Emp_FINAL_06_09_15.pdf

Department for International Development (2005). Girls' Education: Towards a Better Future for All. London: DFID.

Diego, A., and W. Quentin (2010). "Income Generation and Intra-Household Decision Making: A Gender Analysis for Nigeria," in J. S. Arbache, A. Kolev, and E. Filipiak (eds.), Gender Disparities in Africa's Labour Market. Washington, DC: The World Bank: 381-406.

Drescher, L. S., S. Thiele, and G. B. Mensink (2007). "A New Index to Measure Healthy Food Diversity Better Reflects a Healthy Diet Than Traditional Measures," The Journal of Nutrition 137(3): 647-651.

Efobi, U. R. (2016). Who is Benefiting from Access to the Seme Border of Nigeria? A Gender Study. Geneva: UNCTAD VI. http://vi.unctad.org/tag/docs/tagmr/vimrtagnigeria16.pdf

Food and Agricultural Organisation (2016). Women and Sustainable Food Security. Rome: FAO. http://www.fao.org/docrep/x0171e/x0171e02.htm#TopOfPage

Frolich, M. (2006). "Non-parametric Regression for Binary Dependent Variables," Econometrics Journal 9(3): 511-540.

Grown, C., G. R. Gupta, and A. Kes (2006). "Taking Action to Empower Women: UN Millennium Project Report on Education and Gender Equality," Global Urban Development 2(1). http://www.globalurban.org/GUDMag06Vol2Iss1/Grown,%20Gupta,%20&%20Kes.htm

Haddad, L., J. Hoddinott, and H. Alderman (1997). Intra-household Resource Allocation in Developing Countries: Models, Methods, and Policy. Baltimore, MD: Johns Hopkins University Press for the International Food Policy Research Institute.

Huh, Y. (2016). "Gender Empowerment and Educational Attainment of US Immigrants and Their Home-Country Counterparts," Feminist Economics 23(2): 120-145.

Ibnouf, F. O. (2009). "The Role of Women in Providing and Improving Household Food Security in Sudan: Implications for Reducing Hunger and Malnutrition," Journal of International Women's Studies 10(4): 144-167.

IFPRI (2012). Food Security Portal--Nigeria. http://www.foodsecurityportal.org/nigeria/resources

International Fund for Agricultural Development (2012). Transforming Agricultural Development and Production in Africa: Closing Gender Gaps and Empowering Rural Women in Policy and Practice. https://www.ifad.org/documents/10180/3006b7d0-f241-4b27-8972-9ea655181fd2

Kabeer, N. (2001). "Reflections on the Measurement of Women's Empowerment--Theory and Practice," in A. Sisask (ed.), Discussing Women's Empowerment--Theory and Practice. Stockholm: Novum Grafiska AB, 17-54.

Luchuo, E. B., K. A. Paschal, G. Ngia, P. K. Njem, S. Yelena, B. Nsah, et al. (2013). "Malnutrition in Sub--Saharan Africa: Burden, Causes, and Prospects," The Pan African Medical Journal 15: 120.

Malapit, H. J., E. Sraboni, A. R. Quisumbing, and A. U. Ahmed (2015). "Gender Empowerment Gaps in Agriculture and Children's Well-being in Bangladesh," IFPRI DP 1470. Washington, D.C.: International Food Policy Research Institute. http://ebrary.ifpri.org/cdm/ref/collection/p15738coll2/id/129730

Malhotra, A., and S. R. Schuler (2005). "Women's Empowerment as a Variable in International Development," in D. Narayan (ed.), Measuring Empowerment: Cross-Disciplinary Perspectives. Washington, DC: World Bank, 219-246.

Mutua, E., E. Waithanji, L. Karir, and E. Mukewa (2014). "Measuring Women's Social and Economic Empowerment," International Livestock Research Institute RB 36, December.

National Coalition on Affirmative Action (2009). National Gender Policy. http://www.aacoalition.org/national_policy_women.htm.

Nigerian Bureau of Statistics (2014). Statistical Report on Women and Men in Nigeria. Abuja: Nigerian Bureau of Statistics.

Olatinwo, T. (2017). Making Healthcare Affordable: Women's Savings and Loans Clubs to the Rescue. Washington: USAID. http://www.mcsprogram.org/legacy-alma-ata-declaration-integrating-maternal-newborn-child-health-services-primary-care-copy/

Phan, L. (2013). "Women's Empowerment and Fertility Changes," International Journal of Sociology of the Family 39(1/2): 49-75.

Seymour, G. (2017). Women's Empowerment in Agriculture: Implications for Technical Efficiency in Rural Bangladesh. Agricultural Economics 48(4): 513-522.

Sharaunga, S., M. Mudhara, and A. Bogale (2015). "The Impact of 'Women's Empowerment in Agriculture' on Household Vulnerability to Food Insecurity in the KwaZulu-Natal Province," Forum for Development Studies 42(2): 195-223.

Smith, L. C., U. Ramakrishnan, A. Ndiaye, L. Haddad, and R. Martorell (2003). The Importance of Women's Status for Child Nutrition in Developing Countries. Washington, DC: IFPRI.

Sraboni, E., H. J. Malapit, A. R. Quisumbing, and A. U. Ahmed (2014). "Women's Empowerment in Agriculture: What Role for Food Security in Bangladesh?," World Development 61: 11-52.

Taruvinga, A., V. Muchenje, and A. Mushunje (2013). "Determinants of Rural Household Dietary Diversity: The Case of Amatole and Nyandeni Districts, South Africa," International Journal of Development and Sustainability 2(4): 2233-2247.

Thorne-Lyman, A. L., N. Valpiani, K. Sun, R. D. Semba, C. L. Klotz, K. Kraemer, et al. (2010). "Household Dietary Diversity and Food Expenditures are Closely Linked in Rural Bangladesh, Increasing the Risk of Malnutrition Due to the Financial Crisis," Journal of Nutrition 140(1): 182S-188S.

UNCTAD (2014). Measuring ICT and Gender: An Assessment. Geneva: UNCTAD.

UNICEF (2011). Situation Analysis of Children and Women in Nigeria. https://www.unicef.org/nigeria/SITAN_UNICEF_Nigeria_2011_FINAL_2012_Sept.pdf

UN-Women (2016). Progress Towards Meeting the MDGs for Women and Girls. http://www.unwomen.org/en/news/in-focus/mdg-momentum

Von Grebmer, K., B. Nestorova, A. Quisumbing, R. Fertziger, H. Fritschel, P. Lorch, et al. (2009). Global Hunger Index, The Challenge of Hunger: Focus on Financial Crisis and Gender Inequality. Bonn, Washington, DC, and Dublin: Deutsche Welthungerhilfe/IFPRI/Concern Worldwide.

Waller, M. (2014). Empowering Women through Savings Groups: A Study from the Wellness and Agriculture for Life Advancement (WALA) Program. Baltimore, MD: Catholic Relief Services. http://www.crs.org/sites/default/files/tools-research/empowering-women-through-savings-groups.pdf

World Bank (2016). World Development Indicators. Washington, DC: World Bank. http://databank.worldbank.org/data/reports.aspx?source=world-development-indicators

World Bank (2011). World Development Report 2012: Gender Equality and Development. Washington, DC: World Bank.

World Health Organization (2012). "The Impact of Health Insurance in Africa and Asia: A Systematic Review," Bulletin of the World Health Organization. http://www.who.int/bulletin/volumes/90/9/12-102301/en/

Yatchew, A. (1998). "Non-parametric Regression Techniques in Economics," Journal of Economic Literature 36(2): 669-721.

BELMONDO TANANKEM VOUFO

tanankemvoufo@yahoo.fr Ministry of Economy, Planning and Regional Development, Cameroon

EFOBI UCHENNA

uche.efobi@covenantuniversity.edu.ng Covenant University, Ogun State, Nigeria (corresponding author)

SCHOLASTICA NGOZI ATATA

ngatata@gmail.com Federal University of Agriculture Abeokuta, Ogun State, Nigeria

Received 20 February 2017 * Received in revised form 17 June 2017

Accepted 19 July 2017 * Available online 15 August 2017

doi:10.22381/JRGS7220173

How to cite: Tanankem Voufo, Belmondo, Efobi Uchenna, and Scholastica Ngozi Atata (2017). "Women Empowerment and Intra-Household Dietary Diversity in Nigeria," Journal of Research in Gender Studies 7(2): 39-66.
Table 1 School Enrolment Rate in Nigeria (2010-2013)

      Primary enroll.  Secondary enroll.  Tertiary enroll.
      Female  Male     Female  Male       Female  Male

2010  45.715  54.285   45.341  54.659     41.106  58.894
2011  45.888  54.112   45.088  54.912     42.728  57.272
2012  47.770  52.230   44.278  55.722     42.555  57.445
2013  47.921  52.079   47.359  52.641     43.906  56.094

Source: Nigerian Bureau of Statistics (2014)

Table 2 Empowerment Index: Domains, Indicators and Weights

Domain                  Indicator                       Weight

Education               Education                       1/5
Resources               Right to sale of land owned     1/20
                        Own income                      1/20
                        Decides on the income           1/20
                        Own building                    1/20
Savings/Insurance       Own savings/insurance           1/5
Group activities        Part of a finance group         1/10
                        Part of a social group          1/10
Information/connection  Access to information           1/10
                        Access to the connection        1/10

Source: Authors

Table 3 Descriptive Statistics of Entire Variables

Variable        Total              Male Hh            Female Hh
                Mean     S.dev     Mean     S.dev     Mean     S.dev

Age (Years)       16.78     20.59    16.27     19.61    17.38    21.60
Educ. per cap      0.13      0.32     0.13      0.33     0.12     0.32
Share adult        0.33      0.44     0.32      0.44     0.33     0.44
Share Female       0.51      0.47     0.09      0.19     0.97     0.12
  Hh Size          4.16     13.02     3.53      3.81     3.84     3.20
Income per cap  2604.59  31728.93  3254.45  44179.22  2014.73  9881.47
Elect. per cap    16.44     49.23    15.14     29.19    17.75    63.14
Dietary Intake     0.13      0.34     0.13      0.34     0.12     0.34
Empowerment        0.23      0.14     0.24      0.14     0.23     0.13

Table 4 Regression Results--Empowerment and Dietary Intake

                                      Total
                                      OLS           IV

                                       0.074 (***)   0.326 (***)
Empowerment                           (0.000)       (0.000)
                                       0.108 (***)   0.201 (***)
Constant                              (0.000)       (0.000)
R-squared                              0.009         0.100
F(1, 4813)                             7.140        42.340
Prob.                                  0.000         0.000
Under ID test p, Ho: underidentified                 0.000
Weak ID test stat                                    5.530
(Cragg-Donald Wald F statistic)                     65.810
Endogeneity test p, Ho: exogenous                   (0.000)

                                      Male Hh
                                      OLS           IV

                                       0.063 (***)   0.322 (***)
Empowerment                           (0.000)       (0.000)
                                       0.110 (***)   0.200 (***)
Constant                              (0.000)       (0.000)
R-squared                              0.006         0.100
F(1, 4813)                            76.330        43.900
Prob.                                  0.000         0.000
Under ID test p, Ho: underidentified                 0.000
Weak ID test stat                                    5.530
(Cragg-Donald Wald F statistic)                     65.120
Endogeneity test p, Ho: exogenous                   (0.000)

                                      Female Hh
                                      OLS           IV

                                       0.082 (***)   3.081 (***)
Empowerment                           (0.000)       (0.000)
                                       0.106 (***)  -0.595 (***)
Constant                              (0.000)       (0.000)
R-squared                              0.001         0.022
F(1, 4813)                            26.870        11.340
Prob.                                  0.000         0.001
Under ID test p, Ho: underidentified                 0.000
Weak ID test stat                                    6.600
(Cragg-Donald Wald F statistic)                     24.760
Endogeneity test p, Ho: exogenous                   (0.000)

Note: The superscripts (*), (**) and (***) imply 10, 5 and 1 percent
levels of significance. The values in parenthesis are the probability
values. The instruments used for the IV estimations are the type of
building of the household and the difference in ages between the
primary male and female decision-makers s. We used only individuals
that are of the age 18 years and above to compute this variable.

Table 5 Regression Results--Empowerment and Dietary Intake

                                 Total
                                 OLS           IV

                                  0.026 (***)   0.563 (***)
Empowerment                      (0.003)       (0.000)
                                  0.003         0.004 (**)
Share Female                     (0.117)       (0.032)
                                  0.185         0.058
Hh Size                          (0.225)       (0.747)
                                  0.008 (***)   0.007 (***)
Income per cap                   (0.000)       (0.000)
                                  0.003         0.102 (***)
Educ. per cap                    (0.426)       (0.000)
                                 -0.011 (***)  -0.025 (***)
Share adult                      (0.002)       (0.000)
                                  0.019 (***)  -0.029
Age                              (0.010)       (0.723)
                                  0.063 (***)   0.050 (***)
Elect. per cap                   (0.000)       (0.000)
                                 -0.806        -0.289
Constant                         (0.256)       (0.730)
R-squared                         0.032         0.120
F-Stat                           48.670        88.320
Prob.                             0.000         0.000
Under ID test p,                 --             0.000
[H.sub.0]: under-identified
Weak ID test stat                --             7.250
(Cragg-Donald Wald F statistic)
                                 --            25.500
Endogeneity test p,              --             0.000
[H.sub.0]: exogenous

                                 Male Hh
                                 OLS           IV

                                  0.023 (*)     0.140 (***)
Empowerment                      (0.086)       (0.000)
                                  0.002         0.013 (***)
Share Female                     (0.677)       (0.021)
                                  0.215         0.004 (*)
Hh Size                          (0.258)       (0.066)
                                  0.008 (***)   0.006 (***)
Income per cap                   (0.000)       (0.000)
                                  0.003         0.236 (***)
Educ. per cap                    (0.495)       (0.000)
                                 -0.011 (**)   -0.036 (***)
Share adult                      (0.026)       (0.000)
                                  0.019 (*)     0.011 (***)
Age                              (0.097)       (0.000)
                                  0.057 (*)     0.024 (**)
Elect. per cap                   (0.000)       (0.055)
                                 -0.941        -0.594 (**)
Constant                         (0.285)       (0.018)
R-squared                         0.031         0.120
F-Stat                           22.900        48.160
Prob.                             0.000         0.000
Under ID test p,                  --            0.000
[H.sub.0]: under-identified
Weak ID test stat                 --            8.750
(Cragg-Donald Wald F statistic)
                                  --           24.930
Endogeneity test p,               --            0.000
[H.sub.0]: exogenous

                                 Female Hh
                                 OLS           IV

                                  0.077         0.391 (***)
Empowerment                      (0.517)       (0.000)
                                 -0.002        -0.008
Share Female                     (0.675)       (0.153)
                                  0.134        -0.005
Hh Size                          (0.489)       (0.180)
                                  0.008 (***)   0.007 (***)
Income per cap                   (0.000)       (0.000)
                                  0.009 (**)    0.074 (***)
Educ. per cap                    (0.047)       (0.002)
                                 -0.005        -0.020 (***)
Share adult                      (0.269)       (0.000)
                                  0.012         0.109
Age                              (0.151)       (0.174)
                                  0.065 (***)   0.056 (***)
Elect. per cap                   (0.000)       (0.000)
                                 -0.560         0.502
Constant                         (0.533)       (0.164)
R-squared                         0.033         0.128
F-Stat                           28.630        47.530
Prob.                             0.000         0.000
Under ID test p,                  --            0.000
[H.sub.0]: under-identified
Weak ID test stat                 --            5.530
(Cragg-Donald Wald F statistic)
                                  --           23.790
Endogeneity test p,               --            0.000
[H.sub.0]: exogenous

Note: The superscripts (*), (**) and (***) imply 10, 5 and 1 percent
levels of significance. The values in parenthesis are the probability
values. The instruments used for the IV estimation are the type of
building of the household and the difference in ages between the
primary male and female decision-makers. We used only individuals that
are of 18 years and above to compute this variable.

Table 6 Regression Results--Empowerment and Dietary Intake across
Different Household Types

                                  >50% Female
                                  OLS           IV

                                   0.034 (***)   0.365 (***)
Empowerment                       (0.005)       (0.001)
                                   0.011         0.053 (***)
Share Female                      (0.557)        0.013 (0.440)
                                   0.118         0.021
Hh Size                           (0.686)       (0.509)
                                   0.007 (***)   0.007 (***)
Income per cap                    (0.000)       (0.000)
                                   0.044         0.068 (***)
Educ. per cap                     (0.414)       (0.006)
                                  -0.007        -0.021 (***)
Share adult                       (0.140)       (0.000)
                                   0.017 (*)     0.016 (**)
Age                               (0.070)       (0.036)
                                   0.067 (***)   0.051 (***)
Elect. per cap                    (0.000)       (0.000)
                                  -0.422         0.237
Constant                          (0.718)       (0.472)
R-squared                          0.033         0.131
F-Stat                            29.51         47.73
Prob.                              0.000         0.000
Under ID test p,                   --            0.000
[H.sub.0]: under-identified
Weak ID test stat (Cragg-Donald    --            8.750
Wald F statistic)
                                   --            8.190
Endogeneity test p,                --            0.004
[H.sub.0]: exogenous

                                  <50% Female
                                  OLS           IV

                                   0.014         0.135
Empowerment                       (0.262)       (0.158)
                                   0.061         0.009
Share Female                      (0.008)       (0.000)
                                   0.180         0.002
Hh Size                           (0.342)       (0.494)
                                   0.008 (***)   0.008 (***)
Income per cap                    (0.000)       (0.000)
                                   0.009 (*)     0.156
Educ. per cap                     (0.076)       (0.431)
                                  -0.014 (***)  -0.017 (***)
Share adult                       (0.006)       (0.000)
                                   0.003 (**)    0.002 (***)
Age                               (0.021)       (0.133)
                                   0.007 (***)   0.064
Elect. per cap                    (0.000)       (0.000)
                                  -0.773        -0.119 (**)
Constant                          (0.379)       (0.612)
R-squared                          0.031         0.136
F-Stat                            23.090        43.990
Prob.                              0.000         0.000
Under ID test p,                   --            0.000
[H.sub.0]: under-identified
Weak ID test stat (Cragg-Donald    --            7.250
Wald F statistic)
                                   --            1.999
Endogeneity test p,                --            0.159
[H.sub.0]: exogenous

                                  Female Head
                                  OLS           IV

                                   0.034 (*)     0.432 (***)
Empowerment                       (0.059)       (0.000)
                                  -0.003         0.003
Share Female                      (0.912)       (0.701)
                                   0.200        -0.005
Hh Size                           (0.797)       (0.966)
                                   0.007 (***)   0.008 (***)
Income per cap                    (0.000)       (0.000)
                                   0.005 (**)    0.082 (***)
Educ. per cap                     (0.458)       (0.002)
                                  -0.007        -0.024 (***)
Share adult                       (0.916)       (0.005)
                                  -0.003         0.003
Age                               (0.765)       (0.714)
                                   0.008 (***)   0.007 (***)
Elect. per cap                    (0.000)       (0.000)
                                   0.998         0.056
Constant                          (0.782)       (0.964)
R-squared                          0.031         0.125
F-Stat                            15.220        22.52
Prob.                              0.000         0.000
Under ID test p,                   --            0.000
[H.sub.0]: under-identified
Weak ID test stat (Cragg-Donald    --            5.530
Wald F statistic)
                                   --           10.990
Endogeneity test p,                --            0.000
[H.sub.0]: exogenous

                                  Male Head
                                  OLS           IV

                                   0.023        -0.030
Empowerment                       (0.349)       (0.710)
                                   0.003
Share Female                      (0.703)       (0.703)
                                   0.224         0.003
Hh Size                           (0.142)       (0.300)
                                   0.008 (***)   0.090 (***)
Income per cap                    (0.000)       (0.000)
                                   0.003         0.007
Educ. per cap                     (0.734)       (0.653)
                                   0.014 (*)    -0.012
Share adult                       (0.060)       (0.112)
                                   0.016         0.035
Age                               (0.905)       (0.978)
                                   0.099 (***)   0.095 (***)
Elect. per cap                    (0.000)       (0.000)
                                  -0.978        -0.175
Constant                          (0.167)       (0.461)
R-squared                          0.036         0.134
F-Stat                            12.83         20.04
Prob.                             (0.000)       (0.000)
Under ID test p,                   --            0.000
[H.sub.0]: under-identified
Weak ID test stat (Cragg-Donald    --            7.250
Wald F statistic)
                                   --           25.18
Endogeneity test p,                --            0.000
[H.sub.0]: exogenous

Note: The superscripts (*), (**) and (*) imply 10, 5 and 1 percent
levels of significance. The values in parenthesis are the probability
values.

Table 7 Regression Results--Empowerment and Dietary Intake

                                    Total
                                    OLS           IV

                                     0.011         0.391 (*)
Empowerment                         (0.371)       (0.000)
                                    -0.002        -0.008
Share Female                        (0.673)       (0.153)
                                     0.121        -0.005
Hh Size                             (0.536)       (0.180)
                                     0.009 (*)     0.007 (*)
Income per cap                      (0.000)       (0.000)
                                     0.009        -0.074 (*)
Educ. per cap                       (0.052)       (0.002)
                                    -0.005 (*)    -0.020 (*)
Share adult                         (0.255)       (0.000)
                                     0.015 (***)   0.001
Age                                 (0.100)       (0.174)
                                     0.007 (*)     0.006 (*)
Elect. per cap                      (0.000)       (0.000)
                                    -0.493         0.501
Constant                            (0.586)       (0.164)
R-squared                            0.032         0.128
F-Stat                              29.820        47.530
Prob.                                0.000         0.000
Under ID test p, [H.sub.0]:          --            0.000
under-identified
Weak ID test stat                    --            5.530
(Cragg-Donald Wald F statistic)
                                     --           39.75
Endogeneity test p,                  --            0.000
[H.sub.0]: exogenous
Nos. of Households                  1128          1128

                                   Male Hh
                                   OLS          IV

                                    0.016        0.099 (*)
Empowerment                        (0.169)      (0.000)
                                    0.005        0.017 (*)
Share Female                       (0.314)      (0.006)
                                    0.190        0.001 (*)
Hh Size                            (0.293)      (0.000)
                                    0.008 (*)    0.009 (*)
Income per cap                     (0.000)      (0.000)
                                    0.006        0.363 (*)
Educ. per cap                      (0.190)      (0.000)
                                   -0.015 (*)   -0.072 (*)
Share adult                        (0.002)      (0.000)
                                    0.002 (**)   0.001 (*)
Age                                (0.012)      (0.000)
                                    0.006 (*)    0.002 (***)
Elect. per cap                     (0.000)      (0.229)
                                   -0.815        0.919 (*)
Constant                           (0.330)      (0.002)
R-squared                           0.030        0.107
F-Stat                             22.29        42.77
Prob.                               0.000        0.000
Under ID test p, [H.sub.0]:         --           0.000
under-identified
Weak ID test stat                   --           7.250
(Cragg-Donald Wald F statistic)
                                    --          36.27
Endogeneity test p,                 --           0.000
[H.sub.0]: exogenous
Nos. of Households                 1109         1109

                                   Female Hh
                                   OLS          IV

                                    0.077        0.391 (*)
Empowerment                        (0.517)      (0.000)
                                   -0.002       -0.008
Share Female                       (0.675)      (0.153)
                                    0.134       -0.004
Hh Size                            (0.489)      (0.180)
                                    0.008 (*)    0.007 (*)
Income per cap                     (0.000)      (0.000)
                                    0.009 (**)  -0.074 (*)
Educ. per cap                      (0.047)      (0.002)
                                   -0.005       -0.020 (*)
Share adult                        (0.269)      (0.000)
                                    0.012        0.001
Age                                (0.151)      (0.174)
                                    0.065 (*)    0.006 (*)
Elect. per cap                     (0.000)      (0.000)
                                   -0.560        0.501
Constant                           (0.533)      (0.164)
R-squared                           0.033        0.361
F-Stat                             28.630       47.53
Prob.                               0.000        0.000
Under ID test p, [H.sub.0]:         --           0.000
under-identified
Weak ID test stat                   --           6.66
(Cragg-Donald Wald F statistic)
                                    --          23.79
Endogeneity test p,                 --           0.000
[H.sub.0]: exogenous
Nos. of Households                 1128         1128

Note: The superscripts (*), (**) and (***) imply 10, 5 and 1 percent
levels of significance. The values in parenthesis are the probability
values. The instruments used for the IV estimation are the type of
building of the household and the difference in ages between the
primary male and female decision-makers. We used only individuals that
are of the age 18 years and above to compute this variable.
COPYRIGHT 2017 Addleton Academic Publishers
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2017 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Voufo, Belmondo Tanankem; Uchenna, Efobi; Atata, Scholastica Ngozi
Publication:Journal of Research in Gender Studies
Article Type:Report
Geographic Code:6NIGR
Date:Jul 1, 2017
Words:10400
Previous Article:HOW WELL CAN THE GENDER PAY GAP IN MISSISSIPPI BE EXPLAINED?
Next Article:THE ROLE OF GENDER AND MOTHERHOOD IDEOLOGIES IN PERPETUATING WORKPLACE INEQUALITY.
Topics:

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters