Land ownership-the path towards rural women empowerment : A case from Southern Ethiopia

Ethiopia has revised its radical land reform of 1975 via land proclamations. Although the status quo of state ownership of land has been maintained, there is progress in terms of gender. This study is conducted to examine the impact of the land tenure reforms on women empowerment in land management decisions at household level using survey data collected from 394 wives and female head of a family in 2007/08. The results showed that land related decision making is still dominated by men although it has shown a modest change in favour of women ‘with’ the land tenure reform. Women perceived that they have a better autonomy over fixed assets and resource use ‘with’ the reform. However, the improvement in land related decision making and asset control was not accompanied in income control, decision making on self earned money and political participation. In the latter cases a reversal tendency was emerging. The women empowerment index computed using principal component analysis for the ‘with’ and ‘without’ land tenure reform situation indicates that the overall women empowerment status shows improvement after the land tenure reform. However, around 31% of women respondents’ empowerment index has shown decline. This suggests that the affirmative action taken in the land proclamations in favour of women may not be welcomed across the board. A single intervention is proved inadequate to remove shackle of women’s disempowerment that arises from the pervasive traditional and cultural norms. This might call for interventions in strengthening women’s organizations, extensive education and awareness that ensures women’s empowerment in all aspects of life.


INTRODUCTION
In spite of an increasing awareness on women's role in development, appreciation of the economic and social values of women's role in countries development falls far short of what is due.Overall women have made little gains in economic well being.As a result the feminisation of poverty has become prominent in recent years (Ferrell, 1999;Holden et al., 2011;FAO, 2013).This calls for a concerted effort to enhance women's empowerment.There are several vital reasons for evaluating, promoting and monitoring the level of women's empowerment with respect to decision making, access to and control over resources.Enhancing women's status and their empowerment can play a decisive role in the success of many development programmes and bring about positive social changes.Women's empowerment is necessary for ensuring not just uplifting of individual woman in isolation but to ensure the well-being of the entire household which has a multiplier effect on the society as whole (UNFPA, 2005;Peterman, 2011).From this perspective, an individual woman's empowerment could trigger the empowerment process at the household, community and societal levels and thus challenge prevailing patriarchal gender norms.The other valid reason for women's empowerment is the discrimination against women is a *Email: dagted@gmail.com.violation of their basic human rights as it hampers women's abilities to realize their full potential.Millions of rural women around the world have very limited land tenure rights including rights held by an individual or group to land, rights to own, control and use the land 1 .For women, control of land rights has always been difficult, but it has been made worse by the ongoing shift from food crops to commercial crops and population explosion which gradually undermine women's role in agriculture in most traditional society.
Like many other traditional societies, Ethiopian society is highly gender stratified.The low status of women in Ethiopia has been a matter of concern for many years, and the Ethiopian government has implemented various policies and programs to improve their situations.The Federal 2005 land proclamation stipulated that land certification shall be issued jointly for husband and wife and land can be transferred to family members without gender discrimination.The study regions viz., Oromiya and SNNPR land proclamation (ORS, 2007;SNNPR, 2007) are explicitly acknowledge women co-ownership of land and issued joint land certificate.The proclamations are also stated that women can inherit land as men do and entitled equal right to share land holding upon divorce.However, right to land, whether privately or jointly with husband or family, is not always a matter of legal rights.Security of tenure and autonomy on land can be attained by ownership and access, but it also depends on control of the land and the right to use it.And this is often determined by tradition, legal provision and market forces.Shivji suggests that women's land issues are less about access and more about control and ownership because the evidence shows clearly that women are the real producers in most less developed countries (Shivji, 1996).Hence examining the effectiveness of the current land policy to curb the old aged cultural norms and tradtion which undermine women land right is need of the hour so as to draw timely feedback to overcome the gender related hardle which hampers welbeing of women.

Problem setting
Land policy in Ethiopia has undergone some changes in the last four decades.The radical land reform of 1975 has vested ownership on the state and provides only use right on the free hold to the needy.The reform had far reaching consequences.First, it abolished the exploitative land lord-tenant relationship which particularly widespread in the southern part of the country in pre 1975 period.Secondly, via the reform individual households had granted user rights to land albeit the rights could not be transferred by sales, lease or mortgage.Although the reform accredited in what followed the egalitarian agrarian structure in Ethiopia, it fell short of the expectation in terms of promoting tenure security, agricultural productivity and women's land right.A rarely found study by Teklu (2005) cited in Berhanu and Fayera, (2005) stated that the assumption of land reforms such as the 1975 of Ethiopia that took a household as a homogenous unit and distribute land in the name of head of the household proved wrong.Such gender neutral land reform has maintained the status quo of power imbalance between men and women both at household and community level.
Land policies have been extensively studied in Ethiopia both by nationals, expatriates and donor agencies.Many studies uncover the link between the country's problems of agricultural growth; persistent food insecurity; pervasive rural poverty; tenure insecurity; low level of investment; land fragmentation etc with the existing land policy.However, the gender impacts of the land policies is rarely a subject of inquiry and this calls for examining land policy through gender lens.In addition as of late 1990s there have been some changes which range from decentralization of land administration to regional state level to the introduction of joint land certificate for husband and wife.There is also a progress in terms of inheritance and land share upon divorce.The latest land proclamations both at federal and regional level address issues of equal inheritance right for sons and daughters, equal share upon divorce and specify terms and conditions of land rental market.

METHODOLOGY
The study was carried out in two regional states of Ethiopia viz.Southern Nation Nationalities and People Region (SNNPR) and Oromia Regional State (ORS).Three districts, Shashemene and Arisi Negel from the ORS and Meskan from the SNNPR state were selected for the study.A three-stage stratified random sampling technique was employed in selecting the sample.In the first stage districts were selected purposefully as they are located close to National City Addis Ababa where investment boom is significant and pressure on land is high as a result of urbanization and population explosion.In the second stage Peasant Associations (PAs) were clustered into peri-urban and rural based on their distance from the district centers.Five peri-urban and six rural PAs were selected from the three districts.In the third stage a sampling frame was prepared.A total of 394 (three hundred ninety four) households which comprise 5% of the total population were randomly selected.A structured questionnaire was prepared for wives and women headed households and administered in 2007/08 through personal interviewing method.Eight variables namely decision on land related investment, decision on production activities, decision on resource use, income control, autonomy to use self earned money, asset control, village assembly participation and membership in women association were used to compute women empowerment index in order to examine impacts of the reform in 'with' and 'without' situations.
The study used PCA in order to produce linear combinations of the covariates which are uncorrelated with each other.This will avoid the multicollinearity problem that arises because of including independent variables that are related to each other in the regression model.Moreover, the linear combinations chosen have maximal variance; that means, it identifies independent variables combination that explain gender empowerment variation in the study households.A good regression design chooses values of the covariates which are spread out.The following procedure was adopted in converting raw data into normalized form.To start with the best and worst collected values in a variable are identified.The best and the worst values will depend upon the nature of a particular variable.In case of a positive variable, the highest value will be treated as the best value and the lowest will be considered as the worst value.Similarly, if the variable under consideration is negative in nature, then the lowest value will be considered as the best value and the highest will be the worst value.Once the best and worst values are identified, the following formula is used to obtain normalize values: where NV represents normalized value of the variable xi in an indicator j, the best, worst and observed values are as defined earlier.Normalized value always lies between zero and one.Using the normalized value women empowerment index was computed using Principal Component Analysis (PCA).PCA is a multivariate statistical technique used to compute index and to reduce the number of variables in a data set into a smaller number of 'dimensions'.In mathematical terms, from an initial set of n correlated variables, PCA creates uncorrelated indices or components, where each component is a linear weighted combination of the initial variables.Let us consider the variables X1, X2,...,Xp.A principal component analysis of this set of variables can generate p new variables, known as the principal components, PC1, PC2,...,PCp.The principal components can be expressed as follows: where amn represents the weight for the m th principal component and the n th variable.The weights for each principal component are given by the eigenvectors of the correlation matrix, or if the original data were standardized, the co-variance matrix.The variance (λ) for each principal component is given by the eigenvalue of the corresponding eigenvector.The components are ordered so that the first component (PC1) explains the largest possible amount of variation in the original data, subject to the constraint that the sum of the squared weights (a 2 11 + a 2 12 +…+ a 2 1n) is equal to one.Formally estimation of first principal component involves solving maximization problem with the help of Lagrange multipliers method as a standard procedure for maximizing a function of several variables subject to one or more constraints.Formula for the variance of first principal component is: Applying the Lagrange multiplier method, we have: In order to have a solution for a1 other than the null vector, then λ must be chosen so that the characteristics equation i.e. |Σ-λI| = 0. Thus a non -zero solution exists if and only if λ is an eigenvalue of Σ.But Σ will generally have p eigenvalues, which all must be nonnegative as Σ is positive semidefinite.These eigenvalues are denoted as λ1, λ2,..., λp and an assumption that they are distinct can be freely made, so that λ1>λ2>...> λp> 0. Differentiating with respect to a1 gives: Hence a1 is an eigenvector of Σ corresponding to the eigenvector λ1.To prove λ1 is the eigenvalue of PC1 premultiply of equation by a T 1 will give us The second principal component, PC2 = a T 2 X is obtained similarly but with one extension.In addition to the scaling constraint that a a T 2 2 = 1 there is another constraint that PC2 should be uncorrelated with PC1; continuing this argument, the j th principal component has to be associated with the j th largest eigenvalue.In case when some of the eigenvalues of Σ are equal there is no unique way of choosing the corresponding eigenvectors, but as long as the eigenvectors associated with multiple roots are chosen to be orthogonal, then the argument carries through.As the sum of the eigenvalues equals the number of variables in the initial data set, the proportion of the total variation in the original data set accounted by each principal component is given by λi/n where n is the number of variables.The number of principal components extracted can also be defined by the user, and a common method used in the literature is to select components where the associated eigenvalue is greater than one (Seema and Lilani, 2006).Finally, women empowerment index for each observation is computed using the formula, where WEI is women empowerment index, Xi is the i th variables in

RESULTS AND DISCUSSION
A comparative analysis of women's empowerment was carried out for 'with and without' land tenure reform situation2 .It has been noted from the figure in Table 1 that without land tenure reform only 37% of women participated in land investment decision whereas with the reform 47% of women participated in the same.About 22% of women respondents participated in production decision 'with' the reform while only 14 per cent did so 'without' the land reform.The finding suggests that most investment and production decisions are still dominated and made by husbands.Although, with the land tenure reform positive changes in favour of women are emerging; the achievements so far are far from being adequate.Examination of women autonomy over economic resources reveals that women's participation in resource use showed an improvement with the reform; however their authority on income control has declined by around 6%.In addition while 57% of women reported that they decided on self-earned money without the land reform, the percentage of women who reported the same declined to 54 %with the land reform.Women respondents perceived that their overall asset control capability (such as land, house etc.) has improved with the reform.The level of women's participation in community matter showed a reversal tendency.While close to 39 %of respondents participated all the time in village assembly without the reform, only 29 %did so with the reform.The level of membership in women's association 3 also showed a decline 'with' the reform.
To start with women empowerment index construction, orthogonal rotated solution was chosen to obtain uncorrelated components using varimax rotation method.The rotated component matrix of PCA led to the selection 3 One of the major reasons for decline in membership in women association is there is no benefit associated with being a member as it is used to be during the socialist regime.After the wake of privatization in Ethiopia all commodities supplied through service cooperatives in a subsidy to members were abolished with the adoption of structural adjustment progarmme since 1991.Since the association do not assume a new and dynamic role to motivate women membership, most women are quitting their membership.
of three components explaining women's empowerment for the 'without' land reform situation (Table 2).The approach commonly used for extraction is ''root greater than one'' criterion.Originally suggested by Kaiser (1958) cited in Dillon and Goldstein (1984), this criterion retains those components whose eigenvalues are greater than one.For the case of 'without' land reform situation three PCs were retained.The first principal component accounted for 38 %of the variations in the variables used to explain women's empowerment.The second and the third principal components independently explain 18 and 14 %of the variation which are not explained by the first principal component.The three components jointly explain about 70 %of the total variation explained by the variables incorporated in the analysis.
A perusal of figures in Table 3 for the 'without' land reform situation reveals that the extracted principal components have reflected different aspects of women's empowerment.The first principal component explained land related decision making (production and investment decision) and assets control (income from farming and fixed assets).Principal component two and  three describe autonomy to use resources and political participation respectively.The first principal component makes up the largest proportion of the total variability (38%) in the set of variables used.The second component accounts for the next largest amount of variability (18%) not accounted by the first component and so on for the higher order components.The empowerment components can be easily interpreted by analyzing the signs and size of the indicators in relation to the new component variable.Likewise for the 'with' land reform situation the orthogonal rotated solution was chosen to obtain uncorrelated components using varimax rotation method (Table 4).Three components with eigenvalues of greater than one were retained.The three components namely component 1 (38%), component 2 (19%) and component 3 (16%) jointly explain 73 % of the total variation in the original data set.The principal component extracted for 'with' land reform situations has also reduced the original eight variables into three women empowerment indicators (Table 5).The first component describes involvement on land related decision making (production and investment) and assets control (income from farming and fixed assets).The second principal component can be interpreted as indicator for women's autonomy over resources including self earned income.The third principal component can be interpreted as indicator of women's political participation.Women's empowerment index was computed using the extracted three principal components which accounted for 70% of the variation in the case of 'without' land reform and 73% of the variation for 'with' land reform situation in the original data set.In addition scree plots criteria was employed for comparison purpose (Figures 1  and 2).The advice of the scree plot would be to pick 2 components because the elbow occurs at component 2 (the slope broke at component 2).This would suggest that the first two components account for a disproportionately large amount of the combined variance.Hence in this case only component one and two were used for the construction of women empowerment index.In both cases (eigenvalues of greater than one criterion and a scree plots), the WEI was ranged from 0 to 1. Zero value implies the wife has hardly engaged in any sort of decision making in land matters, no autonomy over resource use and self earned income and no authority in the control of resources while the value of one implies a complete empowerment on the same indicators.The disaggregation of women empowerment index by headship status shows women headed households were more empowered than the wife of the main sample household.Under 'without' land reform situation the mean WEI for the former was 0.53 while it was only 0.387 for the latter.This result indicates that married women are enjoying less empowerment status as compared with women headed counter parts with respect of the empowerment indicators considered in this study.The situation for 'with' reform situation also confirmed the same conclusion (Table 6).However, the change in women empowerment index from 'without' land reform situation to 'with' land reform situation was 6.6% for women head of a households while the change was 12.6% for wife of the main household head.This implies that the land reform has empowered married women more than women headed household.
District wise inspection of women empowerment index for 'without' and 'with' land reform shows comparable results (Table 7).The mean women empowerment index 'without' land reform was highest in Meskan district (0.421) and lowest with mean of 0.388 in Shashemene.The 'with' land reform women empowerment index followed exactly the same trend and the mean value was highest in Meskan (0.477) and lowest in Shashemene (0.454).The result obtained using scree plot criterion was not different from the engine value greater than one.The mean value of women empowerment index for the whole study area was found to be 0.406 'without' land reform and it has shown an improvement to 0.462 'with' the reform.The finding shows that the land reform has improved women level of empowerment in all districts.The joint land title issued 'with' the land reform has changed the land rights of the women in study regions from subordinate position (secondary right) which is obtained through marriage or being a daughter to having equal holding rights with men.However, comparison of individual women empowerment index 'without' and 'with' land reform shows that the change in status quo women being the property of men to property owner is not welcomed across the board.And some men feel less tenure secure with their wife co-ownership of land and the reform in some cases has triggered land dispute and divorce and this was more so in polygamous households (Tefera and Holden, 2008).
Examination of the figures in Table 8 indicates that the average WEI score 'without' land reform was 0.406 while the average WEI 'with' land reform the score was 0.462.The mean difference between these scores is 0.056.The two-sided null hypothesis that the mean difference is equal to zero was rejected at less than 1% level.The one-sided tests are given on each side of two-sided test in the bottom centre of the output.The result indicates that women are better empowered in land related decision making and in their autonomy with respect to

Conclusion
The land tenure reforms in Ethiopia improve women's empowerment particularly in land management viz., land related investment (tree planting, conservation, agricultural input use), production decision and autonomy in resource use.Hence, the legal provisions given in the 2005 land proclamation of Ethiopia assured women coownership of land which opened a window of opportunity for women's empowerment that has been witnessed in terms of participation in household level decision making and confidence to claim equal share during divorce.However, the present finding also shows that for empowerment of women to happen a multidimensional intervention is necessary.This is because in traditional society where social norms and cultural practices are against women complete empowerment cannot be ensured with the improvement of legal provision alone.This might call for intervention in extensive education and awareness raising and strengthening women organizations so as to ensure women's empowerment in all aspects of life.The respective regional governments should provide due attention to design appropriate interventions in this regard.On the basis of the study finding the following policy recommendations are drawn.
1.For rural women's empowerment the legal rule establishing co-ownership of land upon marriage appears important under the prevailing patriarchal family structure in the study area.The finding of the study concurs with the importance of legal provision given for women coownership of land and its practical implementation should be followed up for further improvement to be made.2. The finding shows that women's participation in land related investment, production decision and autonomy in resource use improved after the land proclamation which ensured their co-ownership of land.This improvement should be accompanied by measure which improves economic benefit sharing and resource control.3. Rural women's empowerment demands multidimensional measures such as economic, social and political empowerment.This call for an integrated policy framework such as supporting women through gender friendly extension programme, promoting non-formal education and improving women's participation in leadership and political position is a critical pathway to ensure rural women empowerment.

Table 1 .
Description of variables used in women's empowerment index construction.Lij is factor loading value of the i th variable on j th indicator; λj is the engenvalue of the j th factor.

Table 2 .
Total variance explained for 'without' land reform period.

Table 3 .
Rotated component matrix for 'without' land reform period.

Table 4 .
Total variance explained for 'with' land reform period.

Table 5 .
Rotated component matrix for 'with' land reform period.
Extraction Method: Principal Component Analysis.Rotation Method: Varimax with Kaiser Normalization.

Table 6 .
Women's empowerment index by headship status.

Table 7 .
Women's empowerment index by district.

Table 8 .
Paired t-test for 'with' and 'without' land tenure reform situation.