i. Social conventions whereby people act in a coordinated manner, yet share a similar distaste for the ensuing outcome. We label them consensual oppressive norms. Think of the informal norm prohibiting the remarriage of widows in India, dowry payments (India), or funeral rituals (China) that cause parents to have a preference for sons and result in the scourge of ‘missing women’, or honor crimes and revenge killings in tribal societies.

ii. Rules prescribing a behavior about which different groups of the population have antagonistic preferences: they involve conflicts of interests and we therefore call them conflicting oppressive norms. Examples are norms of purity in the Hindu caste system, kidnap marriages,⁴ or exclusive inheritance practices that bar daughters from inheriting from their parents in patriarchal societies.

22.2.1 The Formal Law as a Focal Point

This is a version of the rival-claimants game that Basu (1983, 1997) has used to argue that market interactions cannot be sustained unless their participants observe a specific type of social norm. Myerson (2004) adapts the game to highlight how a range of institutions, including property rights law, a legislative assembly, a national constitution, as well as traditional leaders, can improve efficiency of interactions within a social group by reducing the incidence of conflict.

If C_W, C_I > 0, it is the best-response for each party to forfeit the property when the other party lays claim to it and vice versa. Therefore, the game has two Nash equilibria: (Claim, Forfeit) and (Forfeit, Claim). The actual outcome can then depend on a focal point, in the sense defined by Schelling (1960), which dictates the expectations of both the widow and the in-laws regarding the behavior of the other. The history of previous instances of widowhood or the customary practice seems, in such situations, to be a plausible candidate for the focal point. We would thus reason that widows forfeit their husband’s property and the husband’s relatives take hold of it because doing so is the best-response for each given their expectation about what the other party will do and, furthermore, that these expectations are shaped by customary practice (e.g. Akerlof, 1984; Kuran, 1988, 1995; Greif, 2006, pp. 3–53; Basu, 1997, pp. 312–314, 2000; Bicchieri, 1997, pp. 25–28). The outcome can then be described as an institution in the sense of Aoki (2001, p. 10): ‘a self-sustaining system of shared beliefs about a salient way in which the game is repeatedly played’. It also evokes the idea of a (stable) social convention found in Young (1996) and Bicchieri (2006, pp. 34–42).¹⁰

In this scenario, the formal law can strengthen the property rights of widows if it succeeds in altering the focal point (i.e. if people mutually believe in the representation that the law stands for). To put it more precisely, there needs to be a decree that, in the absence of a will, widows should hereafter inherit the property of their deceased husbands and in-laws have no claim over it, and this decree must successfully change the expectations about the actions of each party. Only when this second condition is satisfied may we obtain an equilibrium where the widow indeed inherits the property of her deceased husband. However, the key question (i.e. under what conditions the formal law can cause a change in the focal point) is outside the remit of the theory.

The same framework highlights an alternative mechanism through which the formal law can influence the custom. If a legal reform succeeds in changing the value of the parameter C_W such that we obtain C_W > 0, then the modified game will have a unique Nash equilibrium in which the widow will ‘Claim’ the property and the in-laws will ‘Forfeit’. Thus, the legal reform will succeed in strengthening the property rights of widows. In concrete terms, if a legal reform is to influence the customary practice, it must improve the widow’s prospects in a conflict sufficiently such that she would choose to fight for her deceased husband’s property even if she expected that her in-laws would do the same. This can happen only if, in the event of a conflict, the formal law is able to provide a ruling in favor of the widow as well as enforce its ruling, at least with some positive probability.

22.2.2 The Formal Law as a Threat Point

In the previous section, we explored the possibility that the formal law influences the custom by affecting people’s expectations about how others will behave when faced with a strategic choice. An equally legitimate approach, with considerable grounding in empirical studies, is the idea that the custom corresponds to an outcome of a bargaining game, for which the formal law serves as the threat point. Specifically, in the case of the widow and her in-laws, it is reasonable to suppose that the division of the deceased man’s property is the subject of informal negotiations and either party can initiate formal legal proceedings if they choose to.

Banerjee et al. (2002) propose this mechanism as a possible way through which an agricultural reform law passed in West Bengal, India in 1977 strengthened the property rights of tenant farmers and led to an evolution of the terms of sharecropping contracts in their favor. In a different context, Chiappori et al. (2002) show that the supply of labor by different household members is influenced, among other factors, by the extent to which divorce laws favor the husband or the wife; thus, the household division of labor, which is, arguably, shaped by both custom and bargaining among household members, is also influenced by formal laws dealing with an entirely different aspect of marriage.

Dixit (2007) highlights how bargaining theory can serve as a tractable framework for analyzing informal negotiations between trading partners in the shadow of the formal law. The situation under study in this chapter is broadly similar to the aforementioned and, if a particular customary practice can be represented using a framework of bargaining, then results from bargaining theory can be applied readily to obtain insights about the influence of the formal law.

Let us denote by α the share of the deceased man’s property which accrues to the widow after bargaining between the two parties. Let V_W (.) and V_I (.) be the indirect utility functions of the widow and the in-laws from a given level of wealth. Suppose that the formal law prescribes that the widow should inherit a share f of the deceased man’s property, with the remainder going to the in-laws, and that the total cost of accessing the formal court, including legal fees and the delay in obtaining a decision, can be represented by a disutility of C for both parties. Then, the Nash bargaining solution to the problem can be written as:

(22.1)

where A, as before, indicates the size of the deceased man’s property. Although both parties are likely to have other forms of wealth, we omit these factors in Eq. (22.1) as they do not affect the analysis that follows. Figure. 22.1 shows the Utility Possibility Frontier of the widow and her in-laws from different ways in which the property of the deceased man may be divided. The point x depicts the threat point that, we have argued, can correspond to the outcome of formal legal proceedings. If the cost of accessing the formal court and the subsequent legal battle is significant for both parties, utility levels corresponding to the formal verdict would lie within the Utility Possibility Frontier. Point y depicts the Nash bargaining solution corresponding to threat point x.

A change in the formal law that gives the widow a greater share of her deceased husband’s property would improve her utility from the threat point and lower that of her in-laws. In a large number of cases, this would also improve her welfare from bargaining at her in-laws’ expense, but such an outcome is not inevitable. In particular, following the reasoning in Lundberg and Pollak (1993), if both the widow and the in-laws derive utility from, and provide funds for, certain expenditures (e.g. expenditures on the deceased man’s children) the change in the formal law may leave the outcome of bargaining unchanged. The reason is that, in a non-cooperative equilibrium, the party that loses out from the change in the law would simply cut back on ‘public-good’ expenditures, and the other party would make up for it. A decline in the cost of accessing the formal legal system would improve both parties’ utility from the threat point and therefore its effect on the actual division of the property decided upon is ambiguous. However, Ngo and Wahhaj (2012) show, for constant elasticity of substitution utility functions, that such a shift in the threat point would favor the party that is initially in a stronger bargaining position. Thus, if the widow is initially in a weaker bargaining position, making the formal courts more efficient can, in fact, weaken her property rights further.

We provide a simple illustration of this result. Let us assume, for ease of exposition, that A = 1 and V_W(.) = V_I(.) = V(.), with V′ > 0 and V′′ > 0. It is possible to show that the first-order condition to (22.1) is both necessary and sufficient for determining the value of α^*.11 Using the first-order condition to (22.1), we obtain:

(22.2)

Differentiating throughout (22.2) with respect to C, we can show that:

(22.3)

where G(α,f,C) is negative in the solution to the bargaining problem.¹² Therefore, for α^* > 1/2 (so that V′(α^*) > V′(1 – α^*)), we have that ∂α^*/∂C > 0. This means that if the widow were initially getting less than half the share of her husband’s property, then a decrease in the cost of accessing the formal court (a decrease in C) would, in fact, cause her share to decline.

The intuition behind this result is as follows. As the cost of accessing the formal court decreases, so does the surplus from bargaining. Consequently, the difference in their utility levels from the threat point becomes more important relative to the surplus, and hence more important for their respective bargaining powers. Thus, we obtain the paradoxical result that if the cost of accessing the formal courts are lowered for both parties, such a change will favor the party that is initially in a stronger bargaining position.

22.2.3 The Customary Authority as a Strategic Player

In Sections 22.2.1 and 22.2.2, we considered models where the customary practice was represented by the strategic choices made, or an agreement reached, by widows and in-laws. An alternative approach would be to represent the custom as being set by an individual, or group of individuals, in order to maximize his/her/their own objective function. In practice, the customary authority may consist of a council or assembly composed of influential members of the community (e.g. elders, lineage heads), such as the shalish in Bangladesh or the bagarusi (courts of elders) among the Haya of Tanzania (Chirayath et al., 2005; Davis, 2009).

The customary authority is likely to have his own preference about what the customary practice ought to be, but may be willing to adapt his position for strategic reasons that we describe below.¹³ Individuals whose rights are not well-protected by the customary practice may consider appealing to a formal court of law if the latter promises better protection of their rights. An appeal, however, would ostensibly weaken the authority and prestige of the customary authority. To avoid such an outcome, there are two ways in which use of the formal court by community members may be discouraged: (i) by the threat of social exclusion and (ii) by adapting the custom so that those who are favored within the formal legal system are content with the customary ruling regarding any conflict. Social exclusion can be costly for the community, as it will effectively lose one of its members (at least temporarily); however, it would have to be exercised to maintain a credible threat for the future.

Aldashev et al. (2012b) develop a simple model along these lines to highlight how the customary practice may be influenced by the cost of accessing formal courts, the uncertainty in the outcome of formal legal proceedings, as well as the formal law itself. We adapt the basic framework introduced in Section 22.2.2 to illustrate the main features of this model.

Suppose the division of property between the widows and the in-laws is decided by a customary authority with the following utility function:

(22.4)

where α_C denotes his preferred custom, A is the size of the property, E_W (E_I) is a variable that takes a value of 1 if the widow (in-law) has been excluded from the community and 0 otherwise, and S_W (S_I) is the social cost of excluding the widow (in-laws). We modify the utility functions of the widow and in-laws as:

(22.5)

where F_W (F_I) is a variable that takes a value of 1 if the widow (in-law) has appealed to the formal court and 0 otherwise, C_W (C_I) is the cost of accessing the formal court for the widow (in-laws), and P represents the penalty of social exclusion for the person being excluded. In both Eqs. (22.4) and (22.5), α represents the final decision regarding the division of the deceased husband’s property, whether this decision is made by the customary authority or the formal court.

The sequence of events within the game is as follows.

There are two decisions that, arguably, should depend on incentives faced by individuals, but are triggered automatically in the model. The first is the social exclusion imposed on individuals who appeal to the formal court. Although social exclusion is costly for the community, Aldashev et al. (2012b, p. 8) argue that community members would be willing to carry out such punishment ‘to maintain the prestige of the customary authority, on which they depend not only for the resolution of conflicts but also for a variety of other functions, such as representation of the community in negotiations with outside agents’. The second decision indicated above is the ruling provided by the formal court. After presenting the main results within this framework, we shall consider the implications of modeling the formal judge as a strategic player.

We define two threshold values for the custom, α_I and α_W, at which the in-laws and the widow, respectively, are indifferent between accepting the verdict of the customary authority and making an appeal to the formal court. For ease of notation, we assume hereafter that A = 1. Then, using (22.5), we obtain:

(22.6)

Neither party will appeal to the formal court if the local authority declares a custom in the interval [α_W, α_I] containing f.¹⁵ Therefore, if α_C lies in this interval, it is optimal for the local authority to declare a division of property corresponding to his preferred custom (i.e. α_D = α_C) and this will be accepted by both parties.

If α_W > α_C, the widow will not accept the local authority’s preferred custom (i.e. she will appeal to the formal court if he chooses α_D = α_C). In this situation, the local authority has two alternatives: he can choose α_D = α_W to ensure that the widow does not appeal his verdict or α_D = α_C and have his authority challenged. From (22.4), we see that the local authority will deviate from his preferred custom and choose α_D = α_W if α_W > α_C + S_W and α_D = α_C otherwise. Similarly, if α_I < α_C, the in-laws will not accept the local authority’s preferred custom. In this situation, the local authority will choose α_D = α_I if α_I > α_C – S_I and α_D = α_C otherwise. Figure 22.2 summarizes the three possible cases when the formal law is more favorable to the widow than the customary authority’s preferred custom (f > α_C). For f > α_C, we obtain three additional cases corresponding to Cases A, B and C in Fig. 22.2.

We are now in a position to analyze the impact of a change in the formal law regarding the division of a deceased man’s property on the custom. Suppose that, initially, the formal law corresponds to the custom (i.e. f = α_C). Then α_C is in the interval [α_W, α_I]. In this case, it is evident that the legal reform that leads to a small increase in f will have no impact on the custom. As f increases, the interval [α_W, α_I] shifts to the right till we obtain α_W = α_C (i.e. f = α_C + C_W + P). At this point, further increases in f will cause the custom to evolve in the direction of the formal law. This ‘magnetic effect’ of the formal law will persist till we obtain α_W = α_C + S_W (i.e. f = α_C + C_W + P + S_W). Any further increase in f will cause the customary authority to revert back to his preferred custom. A similar reasoning applies if we begin to decrease f starting from f = α_C.

Thus, within this model, the impact of the formal law on the custom is non-monotonic. Neither mild reforms (f close to α_C) nor radical reforms (f much larger or smaller than α_C) can induce the customary authority to deviate from his preferred custom. By contrast, moderate reforms will ‘pull’ the custom in the direction of the formal law. Specifically, the formal law has an impact on the custom in the intervals [α_C + C_W + P, α_C + C_W + P + S_W] and [α_C − C_I − P − S_I, α_C − C_I − P].

In some respects, the conclusions we reach using the model presented in this section are similar to what we obtain using the bargaining model. In the latter case, the formal law can influence the custom by affecting the threat point of the bargaining game. In the former case, as we have just seen, the formal law can change the costs and benefits of maintaining a particular custom for the local elite. However, we obtain the additional insight that, if the customary authority finds it very costly to deviate from its preferred custom, radical reforms may well be counter-productive.

Thus far, we have assumed that the judge in a formal court of law would provide a ruling in the dispute in accordance with the letter of the law. However, in reality, the formal judge may have strategic reasons to deviate from the prescribed law just as the customary authority can deviate from the existing custom. One possible reason why the formal judge may not follow the letter of the law is that he has his own views about how the property of the deceased man should be divided between his widow and his relatives.

Aldashev et al. (2012b) assume that the formal judge receives utility from following the letter of the law (called the ‘benefit of law abidingness’) and a disutility for deviating from his own preferred judgment. Then, for a given law, some judges will follow the letter of the law, while others may ignore it and provide judgments according to their own sense of justice or attachment to tradition.

A change in f may cause some formal judges to accept it; however, others, with a different view of the matter, may come to rebel against it. However, as Aldashev et al. (2012b) show within this framework, when the formal law is reformed to make it more favorable to one party over another, the one that is favored in the letter of the law may actually do worse in terms of actual outcomes in the formal court. We briefly explain the intuition behind the result.

Let us assume that a legal reform favoring the widow is carried out. The formal judges will respond to this change in three ways. Among those who were previously following the letter of the law, some will continue to do so. Others, however, may find the new law too radical and begin to provide rulings according to their own sense of fairness. A third group, which found the old law too biased against widows and therefore provided them a greater share of the property than that prescribed by the law, would find the new law more acceptable: however, in following the prescriptions of the new law, they would be providing judgments less favorable to widows than before.

Thus, the first group of judges mentioned above becomes more ‘progressive’ in response to the legal reform, but the second and third groups become more ‘conservative’. Therefore, it is conceivable that, on average, widows receive verdicts less favorable than before when they take their disputes to the formal court. Such a result is more likely to be obtained when the formal law is considered too radical by the majority of judges in the formal court. In this instance, a radical legal reform may again prove to be counter-productive.

22.2.4 The Influence of the Formal Law When the Custom is Dynamic

In the models considered above, the custom has no scope for evolving except through exogenous events that alter the focal point or the relative payoffs from conflict. It is important to recognize, however, that the custom can also change as a result of internal forces, specifically as a result of actions by individuals who live within it. How would the presence of a formal legal system affect the internal dynamics of a custom? Bowles (2004) has shown how institutional change driven by accidents, idiosyncratic behavior, or collective action can be modeled using stochastic evolutionary game theory, pioneered by Foster and Young (1990). We will show, using such a framework, how legal reforms can successfully accelerate changes in the custom, even if there is no visible impact on customary practices in the immediate aftermath of the reform.

We continue to make use of the normal-form game introduced in Table 22.1 and Section 22.1 to analyze the rights of a widow in the property of her deceased husband. However, we extend and modify the framework as follows.

We assume that in each period, there is a new group of widows and in-laws, arranged in pairs, who face the situation depicted in Table 22.1. A fraction 1 – ω of widows and a fraction 1 – ω of in-laws are assigned strategies according to the distribution of strategies used in the preceding period. We can think of these fractions as simply copying the strategies they have seen or heard about being used by older relatives or neighbors. The remaining fraction ω of widows and in-laws opt for the strategy that yields a higher expected payoff given their expectations about the strategy to be used by the opposing party. These expectations are based on the distribution of strategies used in the preceding period. The players are thus myopic but, as Bowles (2004) argues, ‘this updating process … may realistically reflect individuals’ cognitive capacities’ (p. 408). However, with some small probability ɛ, the players (in the fraction ω) choose a non-best response strategy for reasons not explicitly modeled. In the context of a widow’s property rights, this may include whim, error in judgment, or temporary shocks that affect one’s opportunity costs.

If C_W, C_I > 0 and ɛ is small, as Bowles (2004) shows, the system will be ‘drawn’ to either of the Nash equilibria – (Claim, Forfeit) or (Forfeit, Claim) – of the original game. That is, in the long run, most in-laws will claim the property and most widows will forfeit or vice versa. The system thus exhibits persistence and history-dependence in the sense that the customary practice in the past is a strong predictor of customary practice in the future. However, given the small probability that the widows and in-laws actually opt for suboptimal strategies, it is possible for one equilibrium to be ‘dislodged’ in favor of another.

If we are initially in the state where all in-laws ‘Claim’ and all widows ‘Forfeit’ then, using Table 22.1, widows will continue to find the strategy ‘Forfeit’ optimal unless and until a fraction of in-laws greater than –C_W/(A – C_W) (recall that C_W is negative) choose to ‘Forfeit’ for idiosyncratic (exogenous) reasons. Hereafter, more and more of widows will choose to ‘Claim’, making it less and less attractive for in-laws to do so; this gives rise to the possibility that eventually most widows will ‘Claim’ and most in-laws will ‘Forfeit’. The same reasoning applies if a fraction of widows A/(A – C_I) chooses to ‘Claim’ for idiosyncratic reasons. Therefore, the fractions –C_W/(A – C_W) and A/(A – C_I) are, as argued by Bowles (2004), measures of the persistence of the custom (Forfeit, Claim). The greater are these thresholds, the less likely is it that the custom will evolve as a result of internal dynamics to a situation where widows have stronger property rights.

What role can the formal law play in this framework? We reasoned in the previous section that if the formal law has some influence over the outcome of a conflict (i.e. at least some widows seek recourse in the formal court when both parties ‘Claim’ the property and the formal court has some ability to enforce its ruling), a legal reform can, arguably, be represented by a change in the parameters C_W and C_I. If the initial customary practice is that widows ‘Forfeit’ and in-laws ‘Claim’, a legal reform that strengthens the position of widows (an increase in the value of C_W and/or a decrease in the value of C_I) will weaken its persistence, as represented by the fractions –C_W/(A – C_W) and A/(A – C_I). We argued in the previous section that the legal reform can cause a switch in the equilibrium if it can make it worthwhile for widows to ‘Claim’ their deceased husbands’ property even when they expect their in-laws to do the same (i.e. if it succeeds in making C_W positive). In a framework where we allow the custom to evolve according to its own internal dynamics, we see that even a small increase in C_W or a small decrease in C_I can, by weakening the persistence of the current institution, increase the probability that there is, eventually, a change in the custom.

22.2.5 Subjective Game Models

Aoki (2001) proposes a game-theoretic framework for studying institutional change in which, it is assumed, each agent has a ‘limited, subjective perception of the structure of game that he or she plays’. Such a notion may be pertinent for an analysis of the impact of formal law on customary practices as individuals in traditional societies may be unaware of the strategic choices that are available to them through the formal legal system, especially if there have been recent changes to the formal law or if the formal institution itself is a recent phenomenon. For example, following the introduction of a new Family Code in Morocco in 2003, covering all matters relating to marriage, parentage, children’s rights, and inheritance, a 2008 survey in three regions of the country found that only between 24% and 38% of respondents were able to explain one or more of its clauses (Chaara, 2011).

In this section, we consider potential insights about the role of formal law in shaping customary practices as they can be gleaned from the framework proposed by Aoki. In modeling institutions as specific equilibria of a game, Aoki (2001) identifies four key factors in determining what type of institution will arise:

However, Aoki argues that ‘individual agents cannot have a complete knowledge of the technologically determined rules of the game, nor can they make perfect inferences about other agents’ strategic choices’. In particular, Aoki argues that an agent’s choice of action may be drawn, not from A, but from an ‘activated subset of strategic choices’ A^*, representing the strategic alternatives that he or she is aware of at a point in time. An institution, Σ^*, is represented by the set of beliefs from E that are common to all agents. Given Σ^*, each agent has a ‘subjective consequence function’ CO^* that he or she uses to infer the consequences and payoffs from his or her choice of strategy. Finally, given A^*, Σ^*, and CO^*, each agent has a perceived ‘best-response choice rule’ S^* that may differ from S.

Within this framework, institutional change may occur through Aoki (2007, pp. 19–20):

… changes in the activated sets of individual choices due to the accumulation and development of skills, learning, innovation-induced new action possibilities, and so on. Or, they may be changes in technological and environmental conditions that result in different physical consequences for the same action choices. They may be new laws or fiats which are enacted as consequences of the game in the political domain, but appear as exogenous changes in the parameters of consequence functions in other domains.

In the context of a widow’s right of inheritance in her deceased husband’s property, there may be a number of possible actions available to her when the property is being contested, including appeal to the customary authorities in her husband’s community, request for support in the dispute from her own paternal relatives, use of the formal legal system, or some combination of these. However, if the widow has not previously had experience of the formal law, which is a plausible assumption in societies where the division of a deceased man’s property is normally determined by custom, she may not consciously consider making use of the formal legal system. In this case, her ‘activated subset of strategic choices’ need not correspond to her strategy set.

If the widow is ignorant of the option of going to the formal court, it may be possible to sustain certain strategy profiles as equilibrium outcomes even though they do not qualify as Nash equilibria of the objective game. To illustrate this point, we shall consider a modified version of the rival-claimants game as follows.

In Table 22.2, we assume that C_W, C_I > 0; α + β = f + g = 1; α, β, f, g > 0 and 0 > B > A. Compared to the game depicted in Table 22.1, we have added two new strategies for each player, ‘Share’ and ‘Formal Court’. In specifying the payoffs, we have assumed that if both parties ‘Claim’, or if one party proposes to ‘Share’ while the other party chooses to ‘Claim’ the entire property, a conflict ensues that results in the payoffs C_W and C_I. The other strategy pairs consisting of combinations of ‘Claim’, ‘Share’, and ‘Forfeit’ do not lead to conflict: each party, in effect, receives what it asks for. The payoffs associated with any configuration of actions involving an appeal to the formal court by at least one player are fB and gB, because we assume that its ruling overrides any ruling made by the customary authority.

If we ignore the strategy ‘Formal Court’ for each player, the resulting game has three pure-strategy Nash equilibria: (‘Forfeit’, ‘Claim’), (‘Share’, ‘Share’), and (‘Claim’, ‘Forfeit’). If there is no functioning formal legal system, or if both players are ignorant about it, these three equilibria represent the possible customs that can arise in the community.

Following the reasoning offered by Aoki (2001), the strategy ‘Formal Court’ could become part of either player’s ‘activated subset of strategic choices’ through decentralized experiments of agents trying new strategies from the given sets of action choices. In our context, this could mean that a widow through her own initiative finds that appealing to the formal court is a viable option. Alternatively, she may become aware of the formal law because of information campaigns launched by the government or non-governmental organizations (NGOs).

If either party chooses the strategy ‘Formal Court’, it is assumed that the property will be divided according to the prescription of the formal law, which is represented by the constants f and g. However, there is some cost for accessing the formal law, captured in the assumption that B > A.

If the community is initially characterized by a custom that widows ‘Forfeit’ and in-laws ‘Claim’ the entire property, a widow should deviate to the strategy ‘Formal Court’ when she ‘discovers’ that it is a viable option, since fB > 0. We can argue that, as more and more widows become aware of this strategy, the relations of a deceased man can no longer reasonably expect that the widow will forfeit the property. The original custom will be dislodged. This reasoning can be formalized using the framework from stochastic evolutionary theory adopted by Bowles (2004) and discussed in the previous section.

What would happen in the long run after both players have become aware of the strategy ‘Formal Court’? Depending on the prescriptions of the formal law and the cost of accessing the formal court, one possibility is that the courts will be used for settling such disputes hereafter. However, if βA > gB and αA > fB, note that the strategy profile (‘Share’, ‘Share’) can also be a (evolutionarily) stable equilibrium. Thus, the formal law and increasing knowledge of the law in the society would act as a catalyst for an evolutionary process which eventually leads to a new custom. Beginning with a customary practice in which widows receive very little of their deceased husbands’ property, the following conditions are sufficient for the formal law to play such a role:

The mechanism for institutional change proposed here, based on Aoki’s notion of ‘subjective game models’ is akin to the effects of the formal law on the threat point of a bargaining game, discussed in Section 22.2. However, there is an important difference. In the present instance, the evolutionary process is not sensitive to the precise content of the formal law, in the way that the threat point of a bargaining game would be. The crucial point is that, when a widow is made aware of the formal law, she should find it sufficiently attractive compared to the existing custom to appeal to the court. This will cause a breakdown of the existing custom and provide the conditions for a new, more egalitarian custom to arise.

22.2.6 The Customary Authority as a Forward-Looking Player

In the preceding sections, we have considered various models in which the custom is shaped by strategic choices made by agents involved in a dispute and by others who have the authority to act as arbitrators in such disputes. A common feature in these models is that the agents are short-lived or, more precisely, have a one-period horizon when making their strategic choices.

However, a customary authority whose prestige and wealth stems from the social and economic activity within the community should, arguably, take into consideration the impact of any decision it makes on future membership in the community. Specifically, if the customary law is unfavorable to some section of the community, it is possible that individuals would opt to leave the community even before they are involved in a dispute, to avoid being subject to the customary law at a later date. If the customary authority is forward-looking, he should take this potential outcome into consideration.

Similarly, if the customary authority expects economic opportunities outside of the community to improve in the near future, it may opt to make the custom more progressive, so as to induce individuals to remain within the community. Alternatively, he may decide that a decline in community membership is a fait accompli and, consequently, opt for a conservative custom. To the extent that the formal law affects the quality of life outside of the community, legal reforms would influence community membership in a similar manner to changes in economic opportunities. Therefore, legal reforms as well as expectations regarding future changes in the law can influence the evolutionary path of the custom. However, the preceding discussion suggests that this influence can be either positive or negative, depending on the agents’ preferences, community membership, and outside opportunities.

To determine the conditions under which progressive legal reforms would induce the custom to evolve in the same direction, Aldashev et al. (2012a) develop a dynamic model of the custom and formal law where the customary authority behaves as a forward-looking agent. In the following sections, we provide a brief presentation of the model and discuss lessons we can learn from the exercise.

22.3.1 Theoretical Considerations

In the discussion below, we present three distinct strands of argument that can support the case of moderate legal reforms: the welfare argument, the law enforcers’ behavior argument, and the subjective argument.

22.3.2 Ex Post Mitigation of Ineffective Radical Laws: Examples

That the modern law may be a ‘dead letter’ when its content departs too radically from the prevailing custom is attested by interesting instances where the state has actually stepped back by amending the law to get it closer to the custom. In Gabon and Senegal, for example, the state realized after a while that it was pointless to prohibit polygamy altogether since the practice was deeply rooted in African tradition. Instead, it decided to make polygamy an optional choice to be explicitly mentioned in the marriage contract (Article 177 of the Gabonese Civil Code). Moreover, the law provides that violation of the commitment to either monogamy or polygamy is a legitimate cause for divorce and in Senegal violation of the monogamous option can be invoked to declare the marriage void. However, the code’s prescription has been continuously ignored by husbands who renege on their commitment to monogamy whenever it suits them. In Gabon (although not in Senegal), the reaction of the state has consisted of promulgating a new law (Law 18/89 of 30 December 1989) that allows much more flexibility than the previous ruling: spouses are now authorized to rescind their commitment to monogamy in the course of marriage life (Coulibaly, 2005; Nambo, 2005; Sacco, 2008). For women, the benefit nevertheless remains that, thanks to the law, they may more easily invoke legitimate reasons for divorce than before.

An important domain in which attempts to rule through central legislation have proved problematic from colonial times onwards is the sphere of land relations. State provisions intended to regulate the allocation of rural lands frequently ran counter to local informal rules (which are not necessarily static) and to the widespread belief that such a prerogative belongs to customary authorities and not to the central state.

When the government of Kenya, in the wake of reforms initiated by Lord Swynnerton before independence, introduced the systematic, formal registration of land rights, the ineffectiveness of such a bold step soon became plain. In the words of two researchers (Shipton, 1988; Barrows and Roth, 1989, p. 7):

[Customary law] in fact continues to govern the way in which most people deal with their land, making tenure rights ambiguous. The land law failed to gain popular understanding or acceptance, individuals continued to convey rights to land according to customary law, and a gap developed between the control of rights as reflected in the land register and control of land rights as recognized by most local communities.

Still, the new law was biting enough to spark serious conflicts which forced the Kenyan state to backtrack and reduce the distance between the formal law and the custom. This was done by requiring that any alienation of land by a title holder be subject to prior approval of other lineage members or witnessing by elders. District Land Boards can also hear complaints filed by wives or sons to prevent a man from selling land essential to a family’s subsistence or land promised for sons to inherit (Berry, 1993, pp. 126–127; Haugerud, 1993, pp. 162–175; Mackenzie, 1993; Pinckney and Kimuyu, 1994).

Imposing radical changes in a top-down manner when personal and civil matters traditionally regulated by customary norms are at stake is likely to give rise to strong resistance or noxious backlash effects. This applies, for example, to the authoritarian way in which the Young Turks and, later, the Kemalists, embraced anticlerical and illiberal ideologies to radically challenge old institutions inherited from the Ottoman empire in Turkey. Mustafa Kemal (Atatürk) and his circle rallied to the most extreme Westernist position of the Young Turk era, holding that European civilization was indivisible and should be adopted in toto. There was thus no attempt to harmonize European civilization with Turkish culture because what they had in mind was ‘a cultural revolution in which not only the ‘high’ Islamic civilization would be exchanged for that of Europe, but also the ‘low’ or popular culture would be transformed’ (Zürcher, 2010, p. 149).

By advocating the closure of madrasas (religious schools), tekkes and zaviyes (Sufi lodges), by banning vows to Muslim saints, by replacing the fez with a new national hat, by prohibiting the veiling of unmarried women, especially during their education, by banning the wearing of the turban and cloak (except for the certified ulama), and by totally transforming the Turkish legal system, they alienated a large part of the population that was still immersed in a traditional Muslim culture (Kuru, 2009, pp. 214–215). The rise to power of the Justice and Development Party (AKP) in 2002 and its rescinding of some extreme secular measures, such as the prohibition of veiling in public places, testifies that legal reforms during the Ataturk era were too radical to be sustainable under a democratic regime. Democracy, indeed, meant that large sections of the Turkish population, until then considered as backward conservative people of lower class or peasant heritage (the so-called ‘Black Turks’), could make their voices heard under the leadership of a pious and economically successful group of Anatolian businessmen (White, 2013, pp. 24–79).

22.3.3 Passing Successful Moderate Laws: Examples

States do not always opt for moderate reforms through backtracking on initially bolder initiatives: they may well understand the need for moderate reforms right from the beginning. In the following, we present two illustrations of a reformist attitude of the lawmaker: in relation to immigrants’ rights of access to land (in sub-Saharan Africa) and in relation to inheritance rights (in India).

22.4.1 The Magnet Effect at Work

The experience that we want to present in this subsection is the so-called PNDC Law 111 whereby the Ghanaian government (the so-called Provisional National Defense Council that succeeded Nkrumah at the helm of Ghana in December 1981) attempted to regulate practices of intestate succession in favor of the wives and children of a deceased man, particularly among people (the Akans) governed by a matrilineal system. In the words of the chairman of the Inheritance Commission (known as the Ollennu Commission by the name of its chairman), the new law is aimed at modifying the matrilineal family system ‘in such a way as to give the wife and children of a deceased person interest in a definite portion of his estate since it is felt that the right to maintenance and training of the children and support for the wife out of the estate, though legally enforceable, is not generally recognized as sufficiently substantial’ (cited from Josiah-Aryeh, 2008, p. 20).

By deciding that the nuclear family should become the focus of succession, PNDC Law 111 ‘appears to be an attack on customary law’ (Josiah-Aryeh, 2008, p. 29). Under the matrilineal system, indeed, the brothers and sisters born of the same mother occupy a more important place than the spouse inside the family sphere. A person is not entitled, therefore, to inherit from his or her spouse or from his or her father: membership of the matri-clan determines all succession rules and it is typically the maternal uncle of a child who is in charge of managing the assets of the deceased father (Woodman, 1985; Quansah, 1987; Gedzi, 2009a, 2009b). Matrilineal societies are thus characterized by a weaker relationship between father and son than patrilineal societies (La Ferrara, 2007).

Upon careful look, PNDC Law 111 is not as radical as it is sometimes purported to be. It thus makes a pivotal distinction between the family assets (the assets inherited from the matri-clan) and the assets personally acquired by a deceased person. While the former continue to be regulated by custom, the latter are to be divided according to a rule that earmarks the highest share of the deceased’s personal assets to the nuclear family (surviving spouse and children) and a smaller one to the extended family. For example, when a man dies and his own parents are dead, the share accruing to the spouse is three-sixteenths, that accruing to the children is nine-sixteenths, and the customary heir receives one-quarter. When there are no children, personal assets are divided equally between the surviving spouse and the customary heir (Josiah-Aryeh, 2008, pp. 21–25). A colonial law, the Marriage Ordinance (1884), was even more favorable to the nuclear family (with one-third of the deceased’s personal assets devolved to the surviving spouse and two-thirds to the children, thus leaving nothing for the matri-clan), yet it applied only in cases of officially registered marriages. Following passionate protestations by traditional chiefs, the law was amended in 1909, awarding a one-third share to the extended family, and a two-thirds share to the spouse and children, but the application was again restricted to officially registered marriages (Woodman, 1985; Awusabo-Asare, 1990; Manuh, 1997).

Although empirical evidence is sparse, it is still sufficient to enable us to somehow figure out the impact made by PNDC Law 111 and a number of findings can be usefully interpreted in the light of the theory of the ‘magnet effect’. First, 22% of a sample of about 250 household heads interviewed around the city of Kumasi declared a preference for following the prescriptions of the law. By contrast, 37% expressed a preference for the customary inheritance practice, while 25% confessed that they make inter vivos gifts either to escape the matrilineal succession custom – gifts are then made to children – or to avoid Law 111’s prescription – gifts then go to members of the matri-clan (Gedzi, 2009a).

Second, most female plaintiffs prefer to take land- and family-related cases to chiefs’ or family courts. The minority of those who prefer the formal court option mention a number of reasons. They think that the court is impartial to women and capable of enforcing its decisions. The court, they believe, has the ability to give necessary protection to those who need it, whereas chiefs’ courts are mainly concerned with maintaining or re-establishing existing or pre-existing relationships between disputants. As a consequence, the informal judges may not assess the case brought to them in a rightful manner. In a consistent manner, women who prefer the customary ways tend to lay emphasis on the ability of indigenous courts to repair and restore damaged relationships through arbitration, mediation and advising. The low take-up rate of litigation relating to Law 111 is often attributable to women’s perception of the high cost of a legal recourse, typically the fear of severe sanctions in the form of separation from their children, ejection from their house, and loss of valuable family and clan relationships (Gedzi, 2009b, pp. 15–17; see also Benneh et al., 1997; Fenrich and Higgins, 2001).

Due to ‘fear of spiritual reprisals from the family, family and community pressure and the strong moral sense not to wash the family linen in public’, they are reluctant to take the family members of their deceased husband to the formal courts when these members infringe on their inheritance rights (Gedzi, 2009a, p. 27). In short, most disputants prefer to solve their cases in the traditional setting of local informal courts where the procedure is less adversarial than in the formal court. Conversely, the few women who seek formal litigation to defend their rights are strong characters, often with a comparatively high level of education (and outside option), and ready to defy the social pressure and threats emanating from the extended family.²⁰

Third, from another field study (Quisumbing et al., 2001), it appears that a widespread practice followed by Akan household heads – 40% of them when they are males and 48% when they are female – consists of dividing the personal assets of the deceased person into three equal parts: one-third for the surviving spouse, one-third for the children, and one-third for the extended family. This practice, which actually coincides with the prescription of the succession law of 1909, is more favorable for the members of the nuclear family than the custom, yet at the same time more favorable for the extended family than the statutory provisions of PNDC Law 111.

Since the custom has evolved in the direction of the new law, a ‘magnet’ effect seems to be at work. Can it be objected that the whole transformation of the custom is nothing else than an endogenous adaptation to rising land pressure? According to this evolutionary argument, population growth and market integration make investments for land improvement and conservation more necessary, yet these investments require a proper incentive structure which the matrilineal succession system cannot provide. A change in the rule is therefore induced. Two findings nonetheless suggest that the formal law is likely to have played a significant role in the transformation of inheritance practices. On the one hand, the main mechanism behind the endogenous evolution of this system, which consists of the practice of inter vivos gifts to children, has been increasingly resisted by the extended family under conditions of land scarcity. On the other hand, the facts that (i) a significant minority of household heads claim to follow the statutory provisions of the PNDC Intestate Succession Law and (ii) some women appeal to the modern court to settle their case point to a genuine impact of the law.

Law 111 is a moderate law compared to the Marriage Ordinance (1884), which deprived the extended family of any share in the personal wealth of the deceased person. Since the latter Ordinance aroused serious tensions in Ghanaian matrilineal groups, we may think that a radical legal reform is probably less effective and desirable than a more moderate one, which adds new evidence to the examples cited in Section 22.3. It is revealing, in this regard, that under the impact of Law 111, the custom has been driven to evolve toward the prescription of the 1909 law, which for the extended family (whose authority represents the elite in our theory) is less favorable than the custom yet more favorable than Law 111. It is thus as though the 1909 law acted as a focal point at which a new (probably transitory) equilibrium could be established.

That a comparatively small number of women dare seek the assistance of the modern court system to advance their interests is in keeping with the assumption that informal sanctions are severe enough to be taken into consideration whenever a choice of arbitration types is contemplated by a plaintiff. The ‘magnet’ effect nevertheless operates because Law 111 has conferred upon women a new bargaining power that they can use effectively against the upholders of the matrilineal custom.

This is not true everywhere, however. In some remote areas, the custom has not changed because people, and the marginal group (women) in particular, are not, or not well, informed about the new law (Fenrich and Higgins, 2001). A basic assumption on which the ‘magnet’ effect theory rests is therefore violated. Moreover, in Muslim communities, which are governed by a patriarchal social system, opinion leaders including chiefs insist that Islamic principles determine the rules of intestate succession. Any plaintiff who would dare challenge these principles would face very serious sanctions: according to the chief imam of the Volta region, any Muslim woman who wants to enforce her rights under Law 111 would face stronger sanctions from the community than a woman who does not profess the Islamic faith (Gedzi, 2009b, p. 14). In the latter instance, the ‘magnet effect’ is not working because the cost of appealing to the formal court is prohibitively high and the customary authority is highly reluctant to adapt the local rule. Yet, we need to emphasize that in Muslim communities the ‘local rule’ invoked by religious authorities is not the sharia proper, but a more syncretic rule resulting from the blending of Islamic principles with indigenous, pre-Islamic practices. As such, it is therefore susceptible of adapting to changing circumstances (Lapidus, 1988; Bowen, 2003).

22.4.2 Community-Based Change as a Complement to Legal Reform

The theory of strategic interactions between the modern law and the custom does not predict that the ‘magnet’ effect is always in operation. This suggests that other approaches than pure legal reform may be needed when unfavorable conditions prevail. The modern law then needs to be complemented or buttressed by other actions or approaches to be effective. One approach that had precisely that effect is the so-called community-based change of social norms (CBCSN) advocated by the UN Children’s Fund (Mackie, 1996; Mackie and LeJeune, 2009; UNICEF, 2007, 2010; NCTPE, 2003). It apparently met with some success in fighting harmful practices such as FGM in sub-Saharan Africa or footbinding in China.

The CBCSS is a bottom-up approach that calls for the intervention of external facilitators or catalysts typically provided by NGOs. Their role consists of initiating, among a core group of disadvantaged people, an explicit deliberation of local habits and customary rules so that they become aware of the existence of an alternative view of their rights that nevertheless respects their traditional moral values and culture. When the alternative view happens to be embedded in existing statutory laws or in a broader, more universal discourse (e.g. human rights discourse), this aspect is emphasized, thereby facilitating the activation of dormant strategic choices (to use Aoki’s terminology). Thereafter, with the help of extensive media coverage, the catalysts facilitate the diffusion of the new concept of rights among all the disadvantaged people (at local and supra-local levels), encourage consensus-building (rather than an adversarial approach), and help build a momentum susceptible of leading to a declaration that will publicly declare the abandonment of the harmful practice.

There are many interesting lessons to learn from the work of Tostan (a Senegalese NGO with headquarters in Dakar and dedicated to community-led development) with village communities in West Africa. Designed to end harmful practices, such as FGM and child marriage, this NGO experimented with the above-described approach on a rather large scale. One lesson that deserves attention is that in Senegal awareness of the law banning FGM (enacted in January 1999 by the government) increased significantly among the poorly educated women of the project areas as a result of Tostan’s intervention. This is despite the remarkable fact that knowledge of the law was initially quite widespread. Thus, 72% of the sample women knew about the law before Tostan started to operate, but this proportion rose to 93% for women who directly participated in the education program organized by Tostan and to 88% for those who knew about the program through neighboring villages, but did not live in a community where it was implemented. In other villages, where only radio programs were accessible, the rise in awareness of the law was significantly less important. As a consequence, the prevalence of FGM receded significantly among the daughters of the first two categories of women, while it barely changed among those of the third category (Diop et al., 2004).

Another noteworthy outcome of Tostan’s action is the following: whereas any (sporadic) attempt to enforce the law was strongly resisted by the populations before this intervention (Ndoye, 2011, Chapter VIII), the situation changed once the harmful effects of FGM had been carefully explained to them. Thus, the proportion of men and women able to cite at least two reasons for not cutting girls rose from less than 10% to about 80% when they had been exposed to Tostan’s education program (Diop et al., 2004) and the underlying social norms had been called into question.²¹ The fact that Tostan’s strategy relied on efforts to break the connection between oppressive social norms and cultural identity needs to be stressed. Members of the local societies had, indeed, to be persuaded that, by giving up certain obnoxious erstwhile practices, they would not betray their tradition considered as the hallmark of their cultural specificity.²² As long as they considered certain social practices as being integral to their tradition and its underlying values, even victimized members (women in this case) were reluctant to question their validity. Within the theoretical framework of the ‘magnet effect’ approach, Tostan’s efforts focused on providing them better information about the true health cost of continuing with these social practices and changing their conception of the social cost – stemming, specifically, from a loss in cultural identity – that would result from abandoning them.²³

Finally, it is evident from the Senegalese experience that the public declaration represents a significant moment in the decision-making process for the community: large-scale abandonment seems, or is expected, to occur following demonstrations of collective commitment (UNICEF, 2010, p. 15). In fact, the public declaration serves the function of proclaiming the new practice that a critical mass of disadvantaged people have started or intend to follow.²⁴ It makes the change of behavior common knowledge, while its official character allows the members of the disadvantaged group to commit. Not only do elite groups, including male elders, take due notice of the ‘change of winds’ and adjust their behavior accordingly (without asking for any compensation), but the government may also become more determined to implement the existing law. Whereas the law banning FGM was only sporadically enforced before the public declaration, this event compelled the government to make a bolder move in 2009. More exactly, the government itself made a public statement to the effect that the law is to be taken seriously: it decided to adopt Tostan’s approach as the centerpiece of their National Action Plan to end FGM by 2015 (Ndoye, 2011, Chapter VIII). The statutory law therefore ends up backing, and being backed by, a declaration of intent professed by the beneficiaries themselves, following a sustained external intervention.

In the setting of the theory of the ‘magnet effect’, the activation of the modern law/court system in response to the public declaration can result from two different mechanisms: (i) women are more ready to assert their rights and, if needed, to have recourse to the modern court in order to demand their legal enforcement, and (ii) following the pressure exerted by the anti-FGM movement, a larger proportion of modern judges is willing to strictly abide by the law. That this second mechanism is also at work is evident from the fact that prior to the declaration, the punishments meted out on law breakers by most judges were much milder than those prescribed by the law.²⁵ This observation, combined with the aforementioned strong resistance to any attempt to enforce the law before the start of Tostan’s action, suggests that in the pre-declaration environment the law was too radical.

It is interesting to note here that the passing of the law in Senegal met with some significant opposition. (i) There was a fatwa pronounced by the leader of a famous Muslim brotherhood (tijania) to which a number of parliament members were affiliated. In his fatwa, the marabout Thierno Mountaga Tall thus cited a Prophet’s hadith from which he inferred that ‘excision is a corollary to circumcision … and both surgeries are intrinsic parts of human nature’ (Ndoye, 2011, p. 134, our translation). (ii) More unexpectedly perhaps, some NGOs lobbied the government to postpone the law on the grounds that it was premature and needed a wide sensitization campaign among the populations concerned before implementation (Ndoye, 2011, p. 151).

In the light of the above, we cannot rule out the possibility that, in certain instances, the enactment of the statutory law is actually induced by a grass-roots movement. For example, in Guinea Bissau, the government passed a law banning FGM in 2011 only after facing pressures from village communities in the north of the country, where individuals had attended public declarations regarding the practice across the border in Senegal.²⁶ In this case, in the areas where the demand for a change in customary practices has arisen from the communities themselves, the statutory law cannot be too radical and it cannot serve as a ‘magnet’ that causes the custom to evolve.

The initiating role of the community (or part of the community) as prime mover of a change in social norms (whether a law exists or not that bans the underlying practices) is not necessarily stimulated by an external organization. A recent event illustrates this possibility. The location is Sourif, a small Palestinian market town not far from Hebron in Cisjordania (the West Bank). There, a young woman called Ayah Baradeya was murdered by her uncle on the grounds that she entertained ‘improper relations’ with a university comrade who wanted to marry her. The discovery of her dead body a year later (May 2011) aroused a wave of angry reactions among the local community. Following massive street demonstrations (first in Sourif, then in Hebron and Ramallah) and protests via the Internet, the Palestinian Authority eventually decided to rescind a law that punished honor crimes very leniently. Moreover, one brother of Ayah, an educated young man who played a key role in the protest movement, persuaded his family to forsake its customary right to impose a collective punishment on the whole murderer’s family (expulsion from the area) in revenge for the crime. It is interesting to observe that earlier initiatives to change the criminal code ended in failure owing to the strong objections of powerful conservative members of the Palestinian Parliament, who themselves belong to big patriarchal families.²⁷

As pointed out above in the Palestinian case, the law may be regressive because it does not depart from the oppressive customs followed in some parts of the society. It may even happen that the passive attitude of the lawmaker is erected as a governing principle, officializing the complete surrender of the state in the matters concerned. Thus, in Bolivia, the new constitution drawn up by Evo Morales’ party and approved in 2009 has legalized traditional justice dispensed by village elders. Such a step may amount to legalizing lynching, stoning, strangulation, or burning with petrol (The Economist, 2012a, p. 39).