Feminist Knowledge | Land Issues / Economics

Women, Land Tenure and Inheritance in Ghana: Equal Access or Discrimination? (Book Reviews)

by Dzodzi Tsikata

Two studies in the 1990's on women's land rights in Ghana raise very important questions about customary land tenure and whether or not it is disadvantageous to women and therefore whether or not it is in need of reform. The two studies are Women's Access to Agricultural Land in the Household: A Case Study of three selected Districts in Ghana by Prof. George Benneh, Dr R. Kasim Kasanga and Ms Doris Amoyaw (1995) and Women in Agriculture in Ghana by Beatrice Akua Duncan (1997).

Empirical studies such as these are very crucial for a clearer understanding of gender relations in Ghanaian society. For women's rights advocates in particular, they are important because of the numerous charges of over-generalisation they face in their work.
Indeed, Benneh et al. are critical of the lack of empirical basis of two studies they consulted in the course of their work. These are an National Council on Women and Development (NCWD) study which identifies access to land and access to credit as major problems facing small scale women farmers and a book by Prof Dolphyne which argues that the root causes to women's problems with land and credit, are marriage and inheritance systems among others. "Though some of these assertions might be true, they appear to have been over-generalized, making no room as most have done so far for regional, district and even village level variations in tenurial and inheritance systems" (Benneh et al, p. 1). Duncan's study is very useful in this regard because it cannot be dismissed as not having a basis in empirical study; particularly so in view of its different conclusions from those of the Benneh study.

It is a welcome development when well known and respected experts on land and agricultural issues, two of them men, turn their attention to a subject related to women's rights (in this case, women's access to land), particularly when they engage with the writings of female and feminist scholars. Such a study can be seen as a contribution to the process of de-ghettoisation of gender and women's rights studies which is usually seen as a sideshow to serious mainstream academic pursuits and the preserve of women academics. However, those who rejoice too soon will find that there is hard work along the way before this joining of issues can work in support of women's rights.

Similarities in Findings; Differences in Conclusions

Both studies involved surveys of communities in Northern and Southern Ghana. Benneh et al focus on selected comunities in the Wa and Nadowli Districts in the Upper West Region and the Kumasi Metropolitan District in the Ashanti Region, while Duncan studied five rural communities each in the Northern, Volta, Brong Ahafo and Ashanti Regions. The Benneh study is essentially a report of the findings of the community survey. The Duncan report on the other hand, in addition to the community surveys, employs interviews with institutions such as the Ministry of Agriculture and discusses customary laws and local legislation implicated in women's land rights. These differences in method are responsible for the different emphases of the studies. However, other factors have also contributed to the other differences in the studies, one of them being that while there are some points of convergence in their findings, their main conclusions are widely divergent. For example, both studies find that many women farmers exercise user rights on lands belonging to their husbands (Duncan's study showed that this was more widespread in the Volta Region and Northern Region (55% and 30% of respondents) than in Ashanti and Brong Ahafo (24% and 12%). Benneh et al. also found that "an overwhelming majority of the respondents, men and women alike, gained access to farmlands by either occupying family land, or were allocated land by family heads" (p. 97). Family land occupation accounted for 52% in Kumasi, 83% in Wa and 85% in Nadowli (pp. 122, 143, 162). In Kumasi, family head allocation 8%, gifts 7%, share cropping- 6% and others 15% accounted for much of the rest. In Wa and Nadowli, allocation by the tendamba and family heads respectively, were important sources of land (13% in Wa and 8% in Nadowli). Duncan on the other hand found many cases of share cropping relations in Southern Ghana (58% in Brong Ahafo and 24% in the Volta Region in her sample) and many cases of pleading (45%) in Northern Ghana. Pleading here refers to asking a community member for land which is given on no terms and reverts to the owner whenever he or she needs the land back. Benneh et al. found their respondents satisfied with their land tenure arrangements (99% in Wa and Nadowli and 88% in Ashanti), while Ms Duncan found them dissatisfied.

Ms Duncan thus concluded that her findings "render debatable the general notion among Ghanaians that women have uninhibited access to land. Contrary to such a perception, there is overwhelming evidence to support the fact that by contrast to men, women do not generally own the lands they farm and tend to have smaller lands than men". Benneh et al. on the other hand reached the conclusion that " in spite of the differences in geographical location and land ownership patterns, therefore, women's access to land in the three study areas are reasonably assured and secure" (p. 97). For the Wa area, they conclude that, "Cushioned by substantial supplies of undeveloped agricultural land, it is difficult to reconcile the idea of discrimination against women in respect of landholding, distribution and benefits derived therefrom" (p. 68). The Nadowli study areas were found to be in a similar situation. For Kumasi, they find that "no insecurity or discrimination was discernible or apparent for any group of people, women and men alike" (p. 31).

Thus while Duncan advocates land reform to streamline and unify customary laws on property settlement upon divorce similar to the intestate succession law, Benneh et al. think it unnecessary. "There is no need or any apparent justification for any tenurial reforms or high-handed government intervention (e.g. nationalisation or vesting orders)" (Benneh, p. 106). Like other studies before them, they argue that customary land tenure provides a strong basis for development and therefore, it needs no reform (See Gyasi, 1994).

Similar Concepts; Different Meanings

One of the main reasons for the differences between Benneh et al. and Duncan, are their use of concepts and the meanings they apply to them. One of these is Discrimination. While both studies apply it as important tool for making findings, only Duncan attempts to define it. She identifies three main definitional sources - broad, intermediate and narrow and employs all of them to examine her subject. The broad source is a dictionary definition "an act or system of treating different groups of people in different ways, especially unfairly; the intermediate is found in Article 17 of the Constitution which defines it in terms of giving different treatment to different people attributable only or mainly to their respective descriptions by race, place of origin, political opinions, colour, gender, occupation, religion or creed, whereby persons of one description are subjected to disabilities or restrictions to which persons of other description are not made subject or are granted privileges or advantages which are not granted to persons of another description. The narrow source is the Convention on the Elimination of All Forms of Discrimination whose definition is quite similar to that of the constitution's human rights provisions (Duncan, pp. xxx-xxxii). The definition of concepts is useful for both researchers and readers of such studies - firstly to test hypotheses; make findings and also to understand clearly the bases of these findings. While it is not clear what definition of discrimination has informed Benneh et al's findings, it can be deduced that they are applying a standard lower than that demanded by Duncan's definition.

An important conceptual tool which feminists have contributed to the discussion of land rights is the distinction between ownership (legal title), access (ability to use) and control (the right to use and determine use). Both studies do not employ these concepts strictly and tend to use the term access very loosely. For example, in her discussion of women's rights, the tension between the theory and practice of customary law in relation to women's land rights, Duncan writes in a chapter entitled "Women's Access to the Means of Production": "in practice however, women have not been generally known to have equal access to stool or lineage land which is known as the usufructuary or indeterminable interest" (p. 74). However, Duncan distinguishes between various rights implicated in the distinctions between ownership, control and access - the allodial title, the usufruct and user rights arising from a husband's usufruct and discusses the implications of these differences for women's land rights both in theory and practice. Benneh et al. collapses the different rights under land into ownership and access. They dismiss ownership as not important because it is mere trusteeship, one that very few people enjoy it and discuss all other rights in terms of access.

In so doing, Benneh et al. ignore the importance of ideology in land relations, as a result the theory and practice of trusteeship are very far apart in many cases. While the absence of land sales in the study areas in Northern Ghana might justify this faith in the trusteeship ideology of land relations, the situation in Kumasi points to the difficulties of this understanding of ownership. Their own study in Kumasi show that chiefs are entitled to most of the income from land sales while individuals who lose their usufructory rights in this new land market are compensated with a plot of land. "Besides payments to the Asantehene, proceeds from the land are divided into three at Apeadu: 1/3 to the chief for his personal use; 1/3 to the stool and 1/3 to the Town Development Committee, which is lodged in an official account and controlled by the Committee through its treasurer. Money for the stool is further divided into three portions, one third of which goes to the chief for protocol. The other two thirds go to the various sub-chiefs for their upkeep and buying and maintenance of their regalia" (Benneh, p. 32). It is again pertinent to note that the community's enjoyment of this land is through development projects funded by the 1/3 going to the development committee.

There is also a lack of consistency in the argument about women traditional office holders and land rights. At one level, they make much of the fact that there are queen mothers who control and sell land in Asante. "The community survey and findings conclusively confirm that land custodianship in the Kumasi peri-urban study area is not an exclusive preserve of men... In theory, all lands under the Kumasi paramountcy are vested in the Golden Stool... In practice, however, stools and sub-stools occupied by Queenmothers and Chiefs administer the lands on behalf of the Asantehene for the benefit of the whole community. The evidence from Emena and Boadi, with stools being occupied by Queenmothers - debunk the general charge of discrimination against women in landholding in the study areas" (Benneh et al, pp. 35-36). We are not told how many villages there are in peri-urban Kumasi and therefore what proportion are ruled by queen-mothers. Faced with the absence of women among the members of the land-holding institution in the Wa area (Tendamba), they argue that this is not significant. "Admittedly, no tendana was a woman. Some might view this position as discrimination against women. However, if this were so, then the discrimination could equally be applicable to almost every man and the chiefs, with the exeption of a few tendamba in each village. In custom and tradition, the tendamba are only titular land owners, holding the land in trust for the whole community" (Benneh, p. 67)

Another problem is that Benneh et al. also ignore the significance of another important interest in land, which though inferior to the allodial title (the highest title in land held by chiefs, and tendambas etc.), guarantees the rights of the indeterminable beneficial user - the usufruct. Thus while both studies found that "in principle, all stool subjects and lineage members regardless of sex have inherent rights of access to stool and family land" (Duncan, p. 74), Duncan distinguishes the principle from the practice. In practice, women's access to land is stymied by a number of factors, an important one being that land clearance. This is men's work in the division of labour, and establishes the usufructory right which is not the same as the ability to use another person's usufructory right in land because you are that person's wife, what both Benneh and Duncan describe as access to land. While Benneh et al. also confirm that in the Wa and Nadowli districts until the mid 1980s, women were not supposed to hold a hoe and bend down like men to farm, were not to make yam mounds or establish new farms and that it was men and their sons who established farms, they give no significance to this in their discussion of land rights.

Differences in Treatment of Customary Law Practices Affecting Women's Land Rights

In addition to the land clearing issue, there are other factors identified by Duncan as hampering women's ability to acquire usufructory rights in land. This is another source of difference between her study and the Benneh one, in that they give different weights to customary practices which affect women's rights in land. Benneh et al. do not consider those which are indirect such as the rights and obligations in marriage and ownership of financial resources to be a problem of land tenure. Duncan views them as important to the issue for women's rights in land. As she notes, women are not generally known to have equal access to stool or lineage land because of early marriage, marital obligations, the lack of financial resources, the emergence of permanent crops and the growing phenomenon of land sales. In our view, Duncan's interpretation is more correct because there is a chicken and egg situation - are women poor because of their inability to access large pieces of good quality land, or its is their poverty which makes them unable to invest in larger tracts of land? Do women not have land because they lack the credit to buy land, or they do not have credit because they do not own the land they farm on. The list is endless. What is important, is that these indirect sources of disadvantage are implicated in a land tenure system which has requirements women cannot fulfil because of those disadvantages. For example, the requirement that you need to have cleared the land to hold the usufruct is discriminatory.

A related issue is the different ways in which the two studies view land rights arising from marriage. Both studies found that particularly in the patrilineal areas of inheritance, in Northern Ghana and in the Volta Regions, many women farmed on land given to them by their husbands or worked together with their husbands on the same piece of land. While Duncan point out the limits of this security generated by marriage, Benneh et al. feel that this makes women's access to land secure. And yet, in their study, in both Nadowli and Wa, Benneh et al. show a keen awareness of the implications of marriage and divorce for women's land rights when they write:

In these traditional communities, the practice of women farming with their husbands as a household economic unit is still largely the case. Apart from widows, divorcees and single women, scarcely would any married women (outside small vegetables, rice or groundnut fields), establish their own farms outside the household farm unit. Most men felt they would probe their wives to know the real motive for establishing such farms, when their efforts could best be expended on their family farms...Access to family lands through the husband or to vacant communal land through the tendamba (free of charge) guarantees women's access to land. A woman could feel dissatisfied with her husband or his family. But if the matter cannot be resolved by the family head, elders or chief, her recourse usually is to go off home to her father's house. Any settlement then is best discussed under relations between lineages and clans. If such a woman divorces eventually, she would have access to land through her father. Her rights, by custom, are unfortunately subordinated to that of the sons. Given the current abundance of land, there appears to be no apparent discrimination against women's access to agricultural land. However, the potential for discrimination against divorcees and single women in the fact of land scarcity and differential access to good quality land cannot be ruled out... the economic trees belong to the man. However, the picking, storage and marketing is done by the wife or women generally. The proceeds are usually shared to mutual benefit. If a woman could go out of her way to acquire her own farm, however, it is asserted that all economic trees found there would belong to her" (Benneh et al., p. 81).

The above quotation raises a number of issues. First of all, the fact that the majority of wives farm on land belonging to their husbands implies that their rights in the land and the farm is inferior to those of the husband no matter how much labour and capital they expend on it. In the case of enjoyment of economic trees on land, what is "mutual benefit" in Benneh et al.'s terms?

There is evidence that many married women do enjoy user rights on land belonging to their spouses. As Duncan argues, "there is overwhelming evidence to show that the user rights of women on their husbands farm initially found tremendous security during the subsistence of the marriage" (Duncan, p. 77-78). Two factors demand that caution be exercised in interpreting what these user rights really mean. One of them is the nature of the marriage contract and the other is the nature of customary divorce. Benneh et al., Duncan and other studies on marriage in Ghana note the independence and autonomy of the two parties to a marriage. Benneh et al. stress this and employ statistics from the survey to show this. In such a situation of separate incomes, expenditure and property, the property rights of one party do not guarantee the enjoyment of the other during the subsistence of the marriage, much less in situations of divorce or the death of the property owning spouse.

In situations of marital conflict or divorce, the precariousness of a wife's situation becomes apparent for all to see. She goes to her father's home - not to her own and any settlement is seen as a family affair (Benneh, p. 81). What are her guarantees that her contribution to Benneh et al.'s "household economic unit" will be recognised and secured? As Duncan notes, "the security of tenure of women who farmed with their husbands however, ended with the demise of their marriage either upon divorce or death" (Duncan, p. 79). Duncan found that in the case of divorce among the Ewe in the Volta region a wife could expect drinks and some monetary compensation to her family. In the two Akan speaking areas studies, respondents mentioned monetary compensation, drinks, cloth, and in some cases a portion of the husband's farm given directly to the woman. In the Northern region, Duncan found that women received nothing on divorce and also lost custody of children, except if the husband decided to give them one. In matrilineal areas, children belonged to their mother while in the patrilineal areas they went to their father. According to Duncan, the divorce settlement appears to depend on who was judged to be the cause of the marriage break-down. When the divorce is effected, each party can submit a claim for expenses incurred on the other's behalf during the marriage and such expenses could include on the part of the husband the bride price, debts, valuable gifts and money advanced for trading (p. 113). These rules imply that women could not automatically claim a portion of a farm in all the areas or custody of children's in patrilineal areas, thus limiting the claim that marriage provides security of tenure.

Should a woman's husband die before her, Benneh et al. have no problems with a situation where it is her children who inherit and she benefits through them. They note an evolving patrilineal system where a brother or a son would inherit and be expected to look after everyone, to one in which children and wives inherit, but if the children were young, a brother would take over as caretaker until the children were old (Benneh, p. 62). The fact that a wife was not mentioned as a possible caretaker could imply that either her rights to that property are in name only or that she is also considered to be a minor. According to Benneh et al., women's leaders and representatives declared themselves to be satisfied with the inheritance system, both because they were more concerned about their children's security than their own and also that they expected to be looked after by the children. In the case of Muslim women, they were entitled to a portion of their deceased husband's estate (Benneh et al., p. 62). The assumption that children's interests are co-terminus with their mother's is at best optimistic. And what if a woman has no children of her own and is not a Muslim?

The Meaning of Matrilineal Inheritance

Benneh et al. are very positive about what matrilineal inheritance implies for women's rights in land. "Under the matrilineal system of inheritance, women have an edge over men in respect of access to, and the control of family agricultural land" (p. 63). This assertion is debatable and it is not clear how they came to that conclusion about women's edge over men under matrilineal inheritance. There seems to be a confusion of the matrilineal inheritance system with inheritance by women. In any system of inheritance, the lineage is central. What distinguishes the matrilineal from the patrilineal is who is considered a member of the lineage or the inheritance group. In the matrilineal areas, this group is traced from a female ancestor through females. However, the group includes both men and women. Thus in the case of a man, his lineage members are his uterine brothers and sisters, his sisters' children and other male and female relations related to him through his female ancestor, while in the case of a woman, her lineage group consists of her brothers and sisters, her own children and so on. One person from the lineage is appointed successor, someone who is usually of the same gender as the deceased, and that person controls the property and makes judgements about what should accrue to other beneficiaries. Benneh et al. confirm the above when they find that while a substantial number of respondents in the matrilineal area of peri-urban Kumasi mentioned their children as inheritors of their farmlands and houses, the majority of respondents could not specify who would inherit them. Follow up interviews had found that respondents were expecting customary successors in line with the matrilineal system of inheritance. Thus Benneh et al. conclude that the matrilineal system "had come to stay in the Kumasi peri-urban study area" (p. 100).

While women benefit from matrilineal inheritance, they do so as lineage members, but not as wives, and also not as children if the parent involved is a man. These are important details in any reckoning of the benefits of the matrilineal system for women. Even in cases where women are entitled to lineage land, marriage residence and marital responsibilities prevent many of them from taking advantage of their rights. In any case, the lineage heads who control much of family land are male as are the majority of stool occupants.

One related issue of inheritance arises from Benneh's findings that "most lands have now being reduced to family and stool lands" and "vast permanent family landholdings now constitute the bulk of agricultural land, apart from the substantial undeveloped land" (p.85) in Ashanti and the Wa and Nadowli districts respectively. Such lands revert to the family on the death of the family member who farms them. As Benneh et al. point out, such land cannot be dealt with under the Intestate Succession Law. It is also the case that wives and children cannot inherit such property, no matter what their contribution to developing it is.

Thus while the farm may have been made by the nuclear family's labour, the ownership of the land on which it was made is what determines its inheritance. Once the land is family land, it reverts to the family on the death of the member. The distinction between a farm and the land on which it is made, is important in determining security of tenure. Thus it cannot be the case that in Nadowli, "if any man or woman is able and willing to farm, the full enjoyment of the fruits of one's labour are absolutely guaranteed by society and families" (Benneh et al., p. 85).

While Benneh et al. are right to point out that security of tenure is the issue and not legal ownership, they are wrong in assuming that access gives security of tenure. In addition to the example of family land discussed above, Duncan found that share cropping, another source of land for women, was also problematic. In her study, many share croppers complained about the arbitrariness of landlords in changing the terms of the tenancy at will, a situation made easy for them by the verbal nature of many of these arrangements (p. 77). Other studies on these share cropping arrangements have found them disadvantageous to tenants, in particular, the increasingly favoured practice of turning over half of the crop to the landlord as opposed to 1/3 or 1/4 (Gyasi, 400-401). The banks, which could be a source a credit for women farmers will not do business with those who have no written interests in their land and that includes user rights in husband owned land, share cropping and other such arrangements which Benneh et al. refer to as access.

Different Reading of Statistics

Statistical findings tend to present the old conundrum of whether something is half empty or half full. Secondly, it may not be enough to simply report the views of the majority of respondents for any particular issue. The minority figure and even the minority's reasoning might provide very important qualifications to findings and could be an indication of new and growing trends, indicating the source of future problems which need attention now.

For example, it was found that while bush fallowing remains the system of farming in Nadowli, only 27% of women as oppposed to 42% of men practised fallow periods for longer than 6 years (p. 88). Again, 10% of women farming with male support in Nadowli and 7% in Wa said that their husbands took care of their incomes. In Nadowli, 14% of women farming without male support mentioned joint family control. Surely, these figures have some implications for the finding that women and men have control of their own incomes (pp. 147 and 166). On inheritance of farm land in Kumasi, 36% of men mentioned their children, 17% mentioned their brothers while 6% mentioned spouses, 22% did not know and 8% mentioned others. In the case of women, an average of 44% mentioned children, 5% mentioned spouses while 33% mentioned others and 8% did not know. In the inheritance of houses in Kumasi, 12% of men as opposed to 25% of women mentioned children, while 9% of men and 3% of women mentioned spouses. The majority of respondents 67% men and 68% of women mentioned others. These figures suggest that more women would leave property to their children than men and very few men and women would leave property to their spouses. In Wa, no men or women mentioned spouses while more women mentioned children (91%; 69%). 19% of men would leave property to brothers and sons while 11% of women would do the same. In Nadowli the figures for children are similar to Wa. 27% of men will leave property to their brothers and sons as opposed to 14% of women. This category of brothers and sons is an important minority finding in Wa and Nadowli and points to discrimination in inheritance practice (148,167).

If 25% of male farmers in Nadowli as opposed to 1% of women own land larger than 6 acres, could it not point to problems of access? If 28% of women who have male support in their economic activities as opposed to 2% of men and 0% of women without male support have no idea what acreage they farm, what does this imply? (Benneh et al, p. 162). All the respondents allocated land by the family head were women in the Nadowli sample; the majority of those allocated land by the Tendamba in the Wa sample were women without support and all the share croppers in the Ashanti sample were women. Again, what could those findings suggest about women's access to family land? A closer reading of the statistics most helpfully provided by Benneh et al. in their appendixes raises all kinds of interesting questions.

Again, there are some tantalising findings which could not be explored because of the construction of the survey questions. Thus we do not know why the vast majority of respondents in both Wa and Nadowli were satisfied with their land tenure system and only a small minority were unhappy or why 10% of women and 2% of men in Kumasi were unhappy with both the inheritance system and existing land tenure system. Also, why in Wa, it was found that 20% of women hired labour as opposed to 10% of men, or why in Nadowli, 85% of women and 59% of men reported that they lived in family houses, while 41% of men and 15% of women lived in their own houses, remain a mystery. In that connection, the Duncan study was even less helpful in her presentation of the statistical findings. It would have strengthened the study if the statistics not being used immediately in the body of her report had been put in an appendix as was done in the Benneh study.

Regional and Local Differences are Important

One of the strengths of the Benneh study is their insistence that regional and local differences be considered in any intervention on land relations. Ironically, it is their study which comes to the conclusion that in spite of the important differences in land tenure in their study areas of Southern and Northern Ghana, the land situation in those two areas remains similarly satisfactory. The two areas could not be more different - abundant land in one area and land scarcity in the other; the high development of a land market in one and its absence in the other; the conversion of former agricultural land to urban housing and other uses in one and the absence of this phenomenon in another, etc. In one area, women have a longer history of farming and are even in the majority as farmers while in another area, it was only in the mid 80s that women begun to farm in their own right; one area is matrilineal and the other is patrilineal, etc.

Duncan on the other hand, found some significant regional differences in her study in the areas of farm decision making and control of resources. For example, she found that in Ashanti and Brong Ahafo, women appeared to have a greater degree of control over what crops to grow and the proceeds from the sale of crops, in contrast to the Volta and Northern Regions where the findings suggest that there was a low degree of control over the choice of crops and control over incomes. Duncan attributes this to the fact that women in these areas are greatly dependent on their husbands for land (p. 78).

Gender Bias Versus Women's Rights Advocacy

Beyond the conceptual and methodological issues, another difference represented by the two studies is the gender bias apparent in the Benneh study and the explicit interest in women's rights represented by the Duncan study. Duncan, a lawyer and member of FIDA, an organisation of women lawyers engaged in work around women's rights, makes the important point that women's work particularly when it is unpaid is seen as an extension of their domestic duties in the home. This is confirmed in the way Benneh et al. look at women's work on land owned by their husbands. Describing women's contribution to agriculture in the Wa and Nadowli Districts, Benneh et al. fall for this stereotype of women as complementing male work rather than making a substantial contribution. "Men and their sons established farms, did the clearing, weeding, mound making, some harvesting, and threshing of cereals. Women and girls complemented (on the same farm by sowing/planting, transplanting, harvesting, transporting, processing, storing and marketing" (Benneh, p. 63). Surely, it is problematic to see all these many tasks as complementary.

Given that the Benneh study focuses on women's rights to land, it is surprising that when the rules of patrilineal inheritance are discussed in the Wa case, only the situation of a male deceased is addressed. Are there no rules of inheritance for when women die, or are they not considered important? There are indications of this kind of bias throughout the book. Another example of bias, are the sweeping conclusions about land tenure and women's rights that the Benneh study comes to in spite of the clear cases of discrimination identified in their study. One such conclusion reads as follows:

"Clearly then, women's cultural, social and eonomic security are generally guaranteed in all three study districts. No significant insecurity or discrimination was apparent in Wa and Nadowli districts. The significant level of insecurity recorded amongst women in the Kumasi peri-urban area stem largely from acute poverty and the inability to make ends meet, rather than inhibitions emanating from the land tenure systems" (p. 101), they write. Again, they conclude even earlier, "Women's access to agricultural land in the household in Wa peri-urban study villages are reasonably assured and absolutely secure under current tenurial and inheritance systems. So are the rights of men (p. 76)."

This is surprising given their own accounts. For example, on inheritance in Wa, they note that "even though things are changing in favour of children in the past, one person, the deceased's brother or his son, inherited and took care of the rest of the family" (p. 78). Again, they note in relation to Nadowli that "the potential for discrimination against divorcees and single women in the face of land scarcity and differential access to good quality land cannot be ruled out" and cite two incidents of two divorcees being thrown out of their rooms in their family houses (pp. 81-82). How can a system of inheritance which in Benneh et al.'s own words practices widow inheritance, relegate girls to the background with the rationale that they are other men's wives in other villages and place all emphasis on boys as the carriers of the family tree (Benneh et al., p. 79) have land tenure relations which can be considered fair to women?

In another example of bias, they argue that "an important finding is that the agricultural problems were not gender biased - women as well as men suffered almost equally from the same problems". It is not clear how this "almost equally" was determined. Is it simply about the existence of the problem or its severity or degree. Benneh et al. do not seem to be sure when they make another such statement "….rural poverty appears to be gender neutral. However, the burden seems to be heavies on women without male support" (p. 91).

The Opinions of the Oppressed are Central but are not the Whole Story

Benneh et al. clad their findings in the legitimacy of the opinions of the respondents as does Duncan. Thus Benneh et al. find that "Both men and women opined that there was no discrimination against women in respect of farming rights" (Benneh, p. 64) while Duncan notes "Most women were of the view that generally speaking, the system of land acquisition was acceptable in that any serious minded woman would not be denied access to a piece of land through various methods. They did however express a preference to have been the actual owners rather than subordinate owners, because full ownership according to them would give them the full authority to grow economic or tree crops such as cocoa and increase their chance of benefiting from credit facilities" (Duncan p. 77).

Researchers are trained to respect the opinions of their respondents, and so they should. However, while the opinions of the respondents are sacrosanct, it is important that to remember that they can be programmed by biased questions, the position/situation of the respondents, the conditions under which respondents are interviewed, the weight of culture and socialisation, the existence of successful coping mechanisms and other factors. It is not enough to say that respondents found nothing wrong with the land tenure system. The question is whether researchers should have their own standards and how therefore they should measure responses.

This is even more crucial in the light of some of the findings of the Benneh study. For example, they find that the majority of people had never heard of the pieces of legislation passed by central government to address some of the problems of land tenure, and yet those same people had said that those laws were unneccessary. How is it possible for them to determine the usefulness or otherwise of laws they know nothing about. Duncan on the other hand found that many women did not know about laws or could not avail themselves to such laws because they lacked confidence in the judicial system as too distant (both physically and psychologically) and as too expensive. Both studies correctly point out the failures of governments in their attempts to disseminate information and give education to support the implementation of laws affecting the rights of many people. However, Benneh et al. come to the controversial conclusion that the laws are not useful because they were imposed from Accra, Ghana's capital without knowledge of local conditions. In fact, Benneh et al. go as far as to state that the focus on land tenure may be a diversion in that more serious problems facing women in agriculture will be missed. Thus they state in their conclusions that "imposed national laws ought to be based on findings from empirical research towards correcting the inadequacies of functioning local laws. There is the need for a convergence between evolutionary local law and nationally imposed general law. But adhoc general law on property rights ought to be discouraged" (p. 106).

Flawed Studies make Flawed Recommendations

In spite of the seeming sureness of their findings, Benneh et al. call for further research as does Duncan. More research is needed, particularly in view of the controversy generated by these studies. However, more research will not solve the problem of the fairness or otherwise of land tenure systems in Ghana if it proceeds on flawed premises as in the case of the Benneh study.

While Benneh et al. make the important point that the imposition of laws without due cognisance of local conditions is a problem since among other things the laws cannot be properly enforced, the situation where people who do not know these laws claim not to need them should be distinguished from the judgement that such laws are not needed. While not disregarding the resilience and dynamism of customary law rules, it is important not to romanticise them. A number of studies have made the point about the deliberate construction of customary law as we know it today. Moreover, it cannot be the case that to seek some uniformity, certainty and equity in the rules of land tenure as they are being applied now will be bad for the development of Ghana. It is the process which should be at issue and not the goal. It is not the case that all customary law problems are solved by evolution. In cases of glaring human rights abuses such as female circumcision, cruel widowhood rites and the lynching of witches, can we afford to wait until customary law practices sort themselves out.