A An Illustrated Typology

Any social order is built upon a mix of formal and informal institutions, yet the proportion of either type can vary considerably from country to country for complex reasons that have to do with the political system, administrative traditions (centralised versus decentralised), and cultural influences, in particular. Along the modernisation path of a given country, formal institutions are predicted to grow in importance relative to informal ones. This is not only or necessarily because the former may displace the latter as the result of their being more effective in providing the same service – as witnessed by the replacement of informal insurance or mutual help mechanisms by formal social security systems in advanced Western European countries – but also because they fulfil new functions for which informal arrangements are ill-suited. It is therefore useful to get a better understanding of the way these two types of institutions or rules interact with each other. To help us in this effort, we use a typology proposed by two social scientists, Helme and Levitsky (Reference Helme and Levitsky2004), and we discuss it in a way that extends the presentation made in Baland et al. (Reference Baland, Bourguignon, Platteau, Verdier, Baland, Bourguignon, Platteau and Verdier2020), a paper to which we directly contributed. Here, we make a special effort to illustrate the different cases highlighted in this typology, which is constructed along two dimensions and provides a static picture (the situation obtained at a given point of time).

The first dimension (represented on the vertical axis of Table 1.1) is the degree to which the outcomes of formal and informal institutions’ rules converge or diverge, depending upon whether the latter produce substantively similar or different results from those expected from a strict and exclusive adherence to the former. Divergence points to a substantial discrepancy or contradiction across these outcomes, whereas convergence is obtained when the outcomes are not substantively different. The second dimension (on the horizontal axis) indicates the effectiveness of the formal institutions, understood as the extent to which formal rules and procedures are enforced and complied with in practice. A high level of effectiveness thus means that individuals’ choices are actually constrained or enabled, and there is a high probability of official sanctions in the case of a violation of the rules. Conversely, people expect a low probability of enforcement (and hence a low expected cost of violation) when institutional effectiveness is small.

According to Table 1.1, four different types of institutional interaction patterns can arise: complementarity, accommodation, competition, and substitution. Let us consider them in turn.

B Complementarity

A combination of effective formal rules and convergent outcomes produces complementarity between informal and formal institutions. In such cases, informal institutions may cover contingencies that are not properly allowed for by formal rules, or they may facilitate the pursuit of individual goals within the formal institutional framework, or else they may serve as pillars that support the functioning of formal institutions. In the latter instance, their role is to create or strengthen incentives to comply with formal rules that might otherwise exist merely on paper.

Fafchamps (Reference Fafchamps, Baland, Bourguignon, Platteau and Verdier2020) puts much stress on complementarity when he writes that formal institutions are best regarded as enabling informal ones to perform better. His focus is on the enforcement of market transactions, and his central point is that, since interpersonal relationships are not eliminated by formal institutions (and contracts in particular), the role of good formal institutions is ‘to reinforce the forms of social interactions that lead to a more efficient, more inclusive outcome, and to discourage those interactions that reduce efficiency and ostracise certain groups and individuals’ (p. 376). Thus, formal institutions promote markets less by enforcing contracts directly than by seeking to reinforce informal contract enforcement mechanisms, especially those based on reputation. More precisely, they contribute to more active and more efficient markets by: (i) providing uniform measurement rules and quality standards; (ii) minimising conflicts stemming from fraudulent information or any form of misrepresentation by one party to a transaction; (iii) regulating fraud, bankruptcy, and the conditions under which a relational contract (e.g., an employment contract, a rental or a land lease contract) may end; (iv) curbing violent forms of informal contract enforcement (the reliance on thugs and criminal organisations, in particular); and (v) offering a strictly organised process for the adjudication of contractual disputes which agents may optionally use (pp. 380–1). This view of mutually supporting formal and informal institutions, long held by social scientists (Polanyi, Reference Polanyi1957; Granovetter, Reference Granovetter1985; Ensminger, Reference Ensminger1992), is anchored in the idea that economic exchanges take place between individuals who are necessarily embedded within a social context that does not disappear with the introduction of formal institutions.

C Accommodation

When effective formal institutions are combined with divergent outcomes, we have a situation of accommodation between informal and formal institutions. This situation arises because informal institutions create incentives that prompt people to behave in ways that alter the substantive effects of formal rules, yet without directly violating them. In other words, the effect of informal institutions is to contradict the spirit, but not the letter, of the formal rules. Accommodation occurs when a contradiction emerges between outcomes generated by the formal rules and prescriptions emanating from customary or informal rules. What impedes an outright change or open violation of the formal rules is their very effectiveness. Conflicting dimensions are present within the existing formal institutional arrangement, and what accommodation does is to somehow reconcile them through the implementation and interpretation of formal rules by actors that are subject to informal prescriptions.

As an illustration, consider the following example, collected from fieldwork in West Africa by one of the authors. In Mali, judges in the formal court system often deal with inheritance cases involving rural people in ways that rest on a compromise between the formal law and custom. Thus, when one son and one daughter disagree about their entitled share of the wealth of their deceased father, the judge may choose to persuade the defendant (the son who asks for the enforcement of the custom according to which he should inherit the entire wealth of the father) and the claimant (the daughter who asks for the enforcement of the statutory law that has established the principle of gender equality in matters of inheritance) to accept a verdict based on Islamic law (the daughter receives one-half of the brother’s share). In this manner, the authority of Islam is invoked (the claimant and the defendant are both Muslims) with a view to avoiding a confrontational approach that is likely to disrupt family relations. This is a typical instance in which an actor, the modern judge, plays upon several legal registers to find a solution that is acceptable to the different parties involved in a case. The formal law is not ignored, since it serves as a reference point with respect to which the compromise with other types of laws is devised. Incidentally, the example shows that the stark opposition often drawn between French colonial countries, which rely on written codes, and British colonial countries, which rely on the common law, may be easily overdone.Footnote ⁴

D Competition

The next possibility combines ineffective formal rules and divergent outcomes, giving rise to competition between informal and formal institutions. Here, informal institutions structure incentives in ways that are incompatible with the formal rules: to follow one rule, actors must blatantly violate another. Examples of such situations occur, for instance, when particularistic informal institutions, such as clientelistic relations and clan-based nepotism, arise in various contexts of weak formal political or economic institutions (Hoff and Sen, Reference Hoff, Sen and Barrett2005, Reference Hoff, Sen, Bowles, Durlauf and Hoff2006; Bardhan and Mookherjee, Reference Bardhan, Mookherjee, Baland, Bourguignon, Platteau and Verdier2020). Not surprisingly, competing informal institutions are often found in postcolonial contexts in which formal institutions that were imposed on indigenous rules and authority structures dominate, and have been retained by the new independent states. In postcolonial Ghana, for instance, civil servants were officially instructed to follow the rules of the public bureaucracy, but most believed they would incur a significant social cost (such as a loss of standing in the community) if they ignored kinship obligations that made it a duty for them to provide jobs and other favours to their families and villages (Price, Reference Price1975). The same attitude of favouritism may be found in non-tribal societies: for example, in India where loyalty is due to the jati (subcaste) rather than the kin group (Kakar, Reference Kakar1978).Footnote ⁵ In the worst cases, the main goal of the officials consists of extorting revenue in order to distribute gifts to families and patronage networks. The temptation of such biased behaviour is increased when official posts are on sale, generally within a limited circle of people and groups. The winners are then forced to exact kickbacks to recoup their investment, such as is observed in parts of Pakistan and in Afghanistan (prior to the seizure of power by the Taliban in August 2021).

In some countries – African countries in particular – the problem of kin-based favouritism plagues not only the public sector but also, potentially, the private world of business firms (Kennedy, Reference Kennedy1988). A series of studies of indigenous and non-indigenous (immigrant) firms located in certain African countries, and based on detailed interviews with a sample of trading and manufacturing firms, reveals interesting evidence (see Fafchamps (Reference Fafchamps2004) for an extended synthesis, and Platteau (Reference Platteau, Akyeampong, Bates, Nunn and 337Robinson2014: 177–9) for a summary). Not only is trade with relatives and friends extremely rare in Africa but, when it happens, it harms firm performance. This suggests that, as much as possible, businesspeople want to avoid involving their kin networks in their business, for fear of the costs that are likely to follow. Involvement of relatives is ‘the surest way to go out of business, while selling on credit to relatives and neighbours amounts to ‘signing the death warrant of the firm’ (Fafchamps, Reference Fafchamps2004: 173). Entrepreneurs thus complain that it is difficult to keep business with relatives within the confines of an economic transaction. For example, it is hard to collect payments from relatives, whether in relation to a loan or the delivery of a good. More generally, payment problems are frequent because friendship and family ties get in the way of exerting pressures on clients. In addition, firms buying from family and friends encounter more late delivery problems.

A second relevant finding is the absence of evidence of systematic mechanisms whereby information about trustworthiness of clients is shared among African manufacturing firms, other than direct recommendation by common acquaintances (Fafchamps, Reference Fafchamps2004: 173, 256–7, 295). Among agricultural traders, too, trust-based relationships constitute the dominant contract enforcement mechanism, implying that trust is established primarily through repeated interactions, with little role for referral by other traders. Information on bad clients does not circulate widely, which severely limits collective punishments for opportunistic breaches of contract (such as non-payment). African-managed firms face more cases of non-payment than other firms, and they also complain more frequently about deficient quality (Fafchamps, Reference Fafchamps2004: 92, 109, 117, 135). Their transaction costs are consequently higher. By contrast, within stranger communities, information circulates rather freely, and client referral is a common practice. For all these reasons, non-indigenous firms operating in Africa are at an advantage, as illustrated by the fact that in Kenya, for example, it is only within the Indian community that first-time customers are able to obtain trade credit from the date of their first purchase.Footnote ⁶

Finally, foreign firms hesitate to enter into business relations with indigenous firms, which they generally deem to be unreliable. In particular, they find fault with African managers for continuously trying to renegotiate delivery and payment terms ex post (Fafchamps, Reference Fafchamps2004: 110). Clearly, the multilateral reputation mechanism which, according to Avner Greif (Reference Greif1994), characterises so-called collectivist cultures is conspicuously absent in sub-Saharan Africa. What we find, instead, is the bilateral reputation mechanism that is typical of ‘individualistic cultures’. This is a rather paradoxical conclusion, yet it is perfectly congruent with the idea that in this region kinship/ethnic ties and their associated obligations are more an impediment to private capital accumulation than a social capital that can reduce transaction costs (Kennedy, Reference Kennedy1988). Note that inefficiencies arising from kin-based relations are not only caused by non-contract performance; they are also caused by powerful redistributive pressures exerted on economically successful kin people. In fact, the two problems are often related, since if African businesspeople do not estrange themselves from the realm of the family, they will typically be compelled to sell on credit to their relatives and friends, and the risk will be high that the loans will never be returned. The problem is that borrowers do not feel morally obliged to repay debts incurred from a prosperous relative. Refusing to return a loan is an accepted way to preserve a rough egalitarianism among kin people, or to maintain the prior state of status inequality within the kin group (when a clan leader claims a higher economic position). Insofar as the credit has been granted under this informal pressure, the gift involved results not from a spontaneous but from a forced transaction (Platteau, Reference Mauro2014: 170).

E Substitution

The last type of interaction combines ineffective formal institutions and converging outcomes, and it involves substitution between the two types of institutions. In short, informal institutions achieve what formal institutions were designed, but failed, to achieve. A well-known example is the persistence of traditional and informal forms of intra-community income sharing and mutual help in the face of absent or highly imperfect (formal) insurance markets, due to a lack of verifiability and/or asymmetric information between contracting parties (Scott, Reference Scott1976; Platteau, Reference Platteau, Ahmad, Drèze, Hills and Sen1991, Reference Platteau1997; Dercon, Reference Dercon2005; Udry, Reference Udry1994). A second illustration concerns the problem of public order and public goods provision. A most glaring reflection of weak state capacity is its failure to provide key public goods, including physical security and protection for its citizens. When this happens, there is a tendency for informal groups, networks, or organisations to fill the gap. There is thus abundant evidence of such groups emerging or extending their role to provide emergency help to people hit by a natural disaster (a flood, a forest fire, an epidemic), to build and maintain rural roads or water control infrastructure, or to supply basic education and health services. When it comes to public order, what springs to mind is the situation of failed states in which, by definition, the state is unable to fulfil its minimal function of guaranteeing law and order to all its people. In such contexts, law and order is generally established on fragmented portions of the national territory, at the level of regions or sub-regions, tribal and ethnic entities, religious communities, and so on. It is implemented by warlords, clan militia, sectarian movements based on religion or a millenarian ideology, or some forms of ‘village republic’, military groups backed and directed by a foreign government, or criminal organisations.

There is a serious risk that the substitution of informal law and order agencies for state power entail what Chabal and Daloz (Reference Chabal and Daloz1999) have called a ‘re-traditionalisation of society’. But reality may be more complex and point to various forms of amalgamation of the state system with traditional agencies (Bayart, Reference Bayart1993; Reno, Reference Reno1995; Jones, Reference Jones2009). In Liberia, for example, we learn that ‘tribalism is not an ancient form of organisation which pre-dates the nation-state, but an essentially modern concept which is inherently connected to that of the nation-state … It is a political resource which enables individuals and factions to pursue their interests in a national state’ (Ellis, Reference Ellis1999: 198). Therefore, ‘the apparent uniformity of the formal system of government which has its centre in Monrovia was in reality a patchwork structure in which local communities and their leaders reached a variety of accommodations with the centre, in the process of which old institutions of government, such as chieftaincies and religious sodalities, acquired new characteristics’ (p. 207). Interactions between informal and formal institutions may thus evolve dynamically depending on the institutional path followed by society: the two types of institutions influence each other in a dynamic two-way process.

A Smooth Adaptation of Institutions: Price-Induced Institutional Change

One approach to institutional change is to view institutions as more or less smoothly adapting to changing circumstances or a changing environment. This adaptation can be conceptualised either as the outcome of the changing behaviour of rational agents or as an evolutionary process driven by agents with only limited rationality. The former view, which we discuss in the present subsection, is well reflected in the so-called induced institutional innovation theory advanced by Hayami and Ruttan (Reference Hayami and Ruttan1985) and Hayami and Kikuchi (Reference Hayami and Kikuchi1981). The basic idea is that changes in the environment are translated into the price realm, and agents respond to changes in relative prices by modifying not only the technology they use but also the institutions that regulate their lives.Footnote ⁷ The explanation is thus based on an important simplification: institutions are treated analogously to ordinary goods and services, and market forces are conceived as exerting their influence upon them in the same way as they do for goods and production factors. In short, the market is the driver of institutional change.

As Hayami later admitted (Reference Hayami1997), this is a quite naïve model, especially because institutional change typically requires collective action, and it is simplistic to assume that it is automatically organised whenever the aggregate social benefit resulting from such a change exceeds the social cost. In his words, ‘if such a naïve mechanism of induced institutional innovation always operated, all the economies would have grown smoothly, and no great income gap would ever have emerged between developed and developing economies’ (p. 20). Still, the naïve model could well be valid in broad terms of progress in human history, which seems to suggest that, with enough time, people eventually find solutions to overcome the incentive problems that stand in the way of collective action.

It is therefore no coincidence that the work of Boserup (Reference Boserup1965, Reference Boserup1981) has been a major source of inspiration for development economists seeking to explain institutional change as a pseudo-market mechanism, that is, as the outcome of changes in relative factor scarcities. In her detailed account of the evolution of agricultural systems and patterns of land use under pre-industrial conditions in both temperate and tropical regions, she persuasively argued that food has been increasingly produced with the help of labour-intensive technologies. The dynamic changes involved – which entailed the reshuffling of land rights, the redefinition of gender roles, and other social and institutional changes – had the effect of increasing land productivity while simultaneously maintaining labour productivity and standards of living in the presence of growing population pressure.

Binswanger and co-authors have followed up on the idea that changes in endowments propelled by population growth are an important source of institutional change. They thus proposed a theory in which agrarian institutions and their evolution are largely explained as a function of population density (Binswanger and Rosenzweig, Reference Binswanger and Rosenzweig1986; Binswanger and McIntire, Reference Binswanger and McIntire1987; Binswanger et al., Reference Binswanger, McIntire, Udry and Bardhan1989). More precisely, relative scarcities of key production factors, jointly with material characteristics of agricultural activities and the pervasive information problems associated with them, are seen as playing a major role in explaining changes in contract forms, the intensification of agriculture, and the emergence of small family farms in areas of high population density. The conditions in which plantations exist can also be derived. The idea that under the influence of population growth and market integration informal land tenure rights are gradually transformed into forms closer and closer to freehold rights has been studied in detail by Platteau (Reference Platteau2000: Chap. 4, Reference Platteau and Dercon2004). More recently, Guirkinger and Platteau (Reference Guirkinger and Platteau2015, Reference Guirkinger and Platteau2017, Reference Guirkinger, Platteau, Baland, Bourguignon, Platteau and Verdier2020; Guirkinger et al., Reference Guirkinger, Goetghebuer and Platteau2015) have further developed this line of explanation and support it with first-hand empirical evidence about the nuclearisation of farm households in West Africa.

Interestingly, fascinating applications of the approach of induced institutional change have been made to important episodes of historical development in the Western world, Japan, and Russia. Since it is beyond the scope of the present book to review them, we are content with making a passing mention of a few particularly salient studies. To begin with, Smith (Reference Smith1959) has offered us a detailed and original account of the rather gradual transformation of agrarian contracts and the demise of serfdom in Tokugawa Japan. These profound changes are traced back to a major expansion of economic opportunities sparked by the development of rural (silk-producing) industries. This process appears to be in striking contrast to events in Russia, where serfs were emancipated (in 1861) as a result of a top-down, state-directed reform abolishing serfdom. Things may have been more complex, though, since under the initiative of enlightened landlords, gradual reforms were introduced in some estates even before 1861 (Markevich and Zuravskaya, Reference Markevich and Zuravskaya2018). Another illustration is provided by the work of Voigtländer and Voth (Reference Voigtländer and Voth2013), who traced the origin of late marriage in Europe to the Black Plague period. The underlying mechanism lies in a change of the opportunity cost of women’s involvement in husbandry production, rather than in grain production, following the abrupt decline of the population caused by this plague. The same line of argument, based on the relative importance of grain and husbandry in the prevailing agricultural system and the specific characteristics of the husbandry technology, has been used by the latter authors to account for the differential evolution of the marriage pattern between northwestern Europe, on the one hand, and Mediterranean and Eastern Europe and even China, on the other hand.

At this final stage of our discussion of the first approach, two remarks are in order. First, in many post-Hayami studies the analytical framework actually departs from the original and crude idea of price-adjusting behaviours by agents endowed with so-called parametric rationality (agents take prices as a given which they cannot influence). The preferred theoretical approach has been the principal–agent model, in which a principal wants a task to be performed by an agent whose actions or characteristics s/he cannot directly observe. Since the agent is then incited to opportunistically exploit the resulting information gap, the principal needs to design a contract or a scheme that will induce the agent to behave in such a way as to satisfy the principal’s interests. In this type of model, individuals are assumed to possess strategic rationality, meaning that they are able to anticipate how others will respond to their own decisions. Instead of adjusting to relative prices, they make constrained optimal decisions in which the constraint is often set by relative factor scarcities (in land, labour, or capital) or by the importance of external opportunities (as reflected in the agent’s reservation utility).Footnote ⁸

Second, it is not coincidental that in the above-cited examples, institutional change is largely the outcome of individual decisions taken in a decentralised context. In such instances the induced institutional innovation model does not appear to be too naïve, although it is based on a comparative-static reasoning rather than on a genuinely dynamic argument. When a change occurs in some parameters of the social, technical, and economic environment, the optimal institution, rule, or contract is modified and, being somehow able to recognise this, rational individuals bring about the new arrangement. If that does not happen, the persistence of the inefficient institution is typically attributed to undue meddling of the government or another authority. In contrast to the above approach, the one to which we now turn is truly dynamic in the sense that it depicts the path that leads from one institution to another.

B Smooth (but Slow) Adaptation of Institutions: The Evolutionary View

An alternative view of how institutional change can come about when the environment changes – say because of population growth, the emergence of new economic opportunities, or an external threat or challenge – is anchored in evolutionary theory. According to this view, the emergence, diffusion, and demise of rules or institutions are the outcome of an organic process of Darwinian natural selection, which epitomises the competitive pressures of the market and the invisible hand. Institutions or rules are thus seen as evolving unconsciously and gradually as a result of the pursuit of individual interests as agents repeatedly face the same types of social problems or situations. In the simplest version, inefficient institutions are expected to have a low evolutionary fitness, and therefore they tend to be displaced in the long run by more efficient institutions (see, e.g., Schotter, Reference Schotter1981; Axelrod, Reference Axelrod1984; Sugden, Reference Sugden1986, Reference Sugden1989).

Looked at in this way, institutions emerge not as a result of rational, purposeful design by any individual or organisation of individuals, but as the result of spontaneous evolution. This means, for example, that people learn from experience that following a given constraint or custom can actually serve their own individual interests (Aoki, Reference Aoki2001: 40). Possessing a limited (or bounded) rationality, they follow trial-and-error behaviours: what works well for an individual is more likely to be used again, whereas what turns out poorly is more likely to be discarded. They look around them, gather information, and ground their decisions on the basis of fragmentary information. Because they have only an incomplete idea of the way the world in which they operate works, they do not fully understand the strategic implications of their choices, and may not be especially forward-looking (Young, Reference Young1998: 5–6). Thanks to imitation, trials and errors, and takeovers, however, effective strategies are more likely than ineffective ones to be retained (Axelrod, Reference Axelrod1997: 47–8).

The key mechanism driving change in the evolutionary setup is the so-called replicator mechanism. Whereas in biology the inheritance of parental genes is the replication mechanism (called natural selection) that ensures the survival of the fittest over successive generations (the more effective individuals are more likely to survive and reproduce), in economics replication is often assumed to be effected through the types of behaviour just discussed: imitation and takeover of unsuccessful agents or firms by successful ones, trial-and-error experimentation, and learning.Footnote ⁹ In such ‘reinforcement’ mechanisms, it is one’s own past payoffs that matter, not those of other agents. The principle is that ‘the probability of taking an action in the present increases with the payoff that resulted from taking that action in the past’ (Young, Reference Young1998: 28). It should be noted that, in line with the assumption of limited rationality, the diffusion of more effective strategies does not require that the agents fully understand the strategic implications of their choices (i.e., they are not assumed to have a perfect ability to reason inductively about a feedback mechanism between their own choices and the choices of the other players). Rather than explaining how they would rationally pick actions in a given situation, evolutionary theorising is concerned with understanding how behaviour evolves or persists over time (Rasmusen, Reference Rasmusen1989: 121).

Interestingly, many pioneers of the development economics discipline implicitly held a sort of (co-)evolutionary view of institutional and cultural change. Thus, Arthur Lewis (Reference Lewis1955) thought that religious beliefs, for example, may evolve and be reinterpreted depending on the economic environment confronting societies. In other words, traditional values and attitudes, whenever they are hostile to economic advancement, will eventually adapt themselves to new economic opportunities (p. 106). And Alfred Hirschman (Reference Hirschman1958) pointed out that traditional images of change will remain a critical bottleneck for constructive action for economic development until experience modifies them in the appropriate direction (see also Bauer and Yamey, Reference Bauer and Yamey1957; Meier and Baldwin, Reference Meier and Baldwin1957; and the discussion in Platteau, Reference Platteau, Platteau and Peccoud2011).

Unlike what immediate intuition might suggest, and defeating the excessive hopes placed in them by the economists who first used them, evolutionary models do not unambiguously point to efficient adaptation of human institutions: there is absolutely no certainty that optimal rules or institutions will emerge from evolutionary processes. Contrary to appearances, the evolutionary theory of institutional change is therefore in clear opposition to the induced institutional innovation approach, and the former is better able than the latter to explain diverging institutional trajectories across countries. And since diverging paths are often encountered, the interest of the evolutionary approach ought not to be underestimated. This is particularly evident in the case of cultural norms and habits, modes of social interaction, and political regimes.

It would be wrong to think that the competitive selection of group-level institutions rather than individual strategies might lead to more positive conclusions in the form of more efficient arrangements. In fact, it raises even more severe problems than those confronting invisible hand arguments applied to individual traits. Several factors account for this (partial) failure, and they deserve our attention (see Bowles, Reference Bowles2004: 90–1, Chap. 13).

First, the repertoire of institutions and behaviours among which selection operates may be highly restricted: being absent from the available repertoire, many institutions remain unknown or untried.Footnote ¹⁰ Moreover, ‘the creation of novel institutions is akin to the emergence of new species: it requires the confluence of a large number of improbable variations in the status quo’ (Bowles, Reference Bowles2004: 91).Footnote ¹¹ Binmore (Reference Binmore1992) makes essentially the same point when he stresses that in many evolutionary models attention has been artificially restricted to a few strategies, often arbitrarily chosen. No clue is given as to why particular strategies are there while innumerable other conceivable strategies are ignored (p. 434). If such an approach allows the evolutionary modeller to derive efficient institutions, it does so without really explaining their emergence, since the appearance of the beneficial strategies that lead to them is itself unaccounted for. For example, if followed by everyone, a strategy consisting of respecting a claim to property made by an individual who first occupied a piece of land can lead to the establishment of the institution of private property rights. Yet, the question remains as to how such a strategy did emerge. Also, what happens when several individuals came to occupy a land at the same time, or believe and claim that they did, is an unresolved issue.

Second, the existence and efficiency of an institution often depend on the fraction of the population that is governed by it (a characteristic sometimes called strategic complementarity) and on the set of co-existing institutions.Footnote ¹² Thus, ‘some institutions may be complementary, each enhancing the functioning of the other, while some institutions may reduce the effectiveness of other institutions’ (Bowles, Reference Bowles2004: 90). As a consequence, there may exist multiple stable configurations of institutions, and some of these configurations may be very inefficient and still persist over long periods of time. Because strategic complementarities or other sources of increasing returns – more particularly, setup or fixed costs that must be incurred to create an institution, or learning effects that raise the effectiveness of an institution that acquires growing influence – may thus give rise to multiple (equilibrium) outcomes, institutions may be path-dependent. When institutional evolution is path-dependent, small initial differences may cause distinct societal histories to emerge. Instead of institutional convergence, what is obtained is then the long-term coexistence of distinct evolutionarily stable institutions (Bowles, Reference Bowles2004: 403–4). The evolutionary process follows paths that have different long-run characteristics depending on where they start and on the order in which agents happen to meet, thus leading to different equilibrium configurations (Young, Reference Young1998: 8; see also North, Reference North1990: 92–104).

Third, when the evolutionary framework is enlarged to allow for a coevolution of institutions and preferences (with each exerting an influence on the development of the other), it is quite possible for group-advantageous but individually costly norms and patterns of behaviour to evolve. Thus, between-group competition may have favoured the emergence of groups or nations that have fostered preferences promoting military abilities and war-making capacity rather than individually profitable behaviour. In the process, values centred on glory, honour, valour, and self-sacrifice, as well as the systems promoting them (tribal systems, religions, etc.), take precedence over more selfish individual traits.

Finally, the rates of change induced by selection processes may well be very slow relative to the pace of changes induced by chance events, or exogenous changes in certain key elements in the environment (e.g., knowledge, external influences). If that is the case, and if the outcome of a competitive selection process eventually materialises, it will have become inefficient.

Interestingly, the central message – according to which inefficient (and unequal) institutions can persist over very long periods of time – continues to hold when evolutionary models are made more complex, by bringing multi-level selection into the picture and by introducing players who intentionally pursue conflicting interests through collective action (Bowles, Reference Bowles2004: Chaps. 11–13). We now turn our attention to two remaining approaches which, unlike the above two, are essentially static.

C Institutions as Self-Enforcing Arrangements

This approach, in common with the principal–agent approach, conceives of institutions as self-enforcing (equilibrium) arrangements. However, rather than considering contracts or deals that are entered into in the context of bilateral relationships or multilateral relationships based on pairwise relationships inside a network, it analyses situations in which numerous actors interact simultaneously. Institutions are conceptualised as Nash equilibria because they ensure that beliefs have converged, and actions have been coordinated between individuals. More precisely, all agents choose their own action-choice rules in response to their subjective perceptions (beliefs) of others’ action-choice rules, and it is only when these perceptions are confirmed by observation that an institution or rule is well established (Aoki, Reference Aoki2001: 10–11). Institutions are therefore self-enforcing by virtue of the fact that at equilibrium the actors’ expectations about each other’s behaviour in a particular situation turn out to be consistent with the experience generated by the resulting actions. When this is the case, agents have no incentive to deviate from their own action-choice rule, and institutions represent stable outcomes.

A number of important implications follow from the above characterisation. First, there are multiple possible institutions that can satisfy the requirements of a Nash equilibrium; that is, institutions for which beliefs converge and actions are coordinated. Furthermore, there is absolutely no assurance that they all correspond to socially efficient outcomes. Put in another way, it is only in particular situations, namely in pure coordination problems, that people are indifferent between possible stable solutions: for example, they do not care whether they should stop at the traffic light when the colour is red, green, blue, or orange – the only thing that matters to them is that everybody follows the same rule or convention. Since individual beliefs or expectations play such an important role in many problems of institutional choice, it is highly plausible that people end up coordinating on a rule that is not optimal (in the sense intended by Pareto, that no other rule is available that could improve the situation of one individual without causing a loss of welfare for some other individuals). By definition, if a rule or institution is a Nash equilibrium, people do not deviate from the outcome to which their expectations have led them; cultural inertia appears to be the factor behind this stability (Basu et al., Reference Basu, Jones and Schlicht1987). In the words of Stiglitz (Reference Stiglitz and Bardhan1989), ‘individuals know more about the institutions and conventions with which they have lived in the recent past than they know of others by which they might live’ (p. 26). Therefore, they tend to prefer the status quo to untried solutions, the effects of which they can only anticipate or speculate on.

Owing to the critical role of beliefs, it is even possible to come across paradoxical situations in which individuals choose to support rules or other institutional arrangements which they do not like, or even find repugnant (such as the caste system in India). As shown by Georges Akerlof (Reference Akerlof1976) and Timur Kuran (Reference Kuran1995: Chaps. 6–8), it may thus be rational for an individual to comply with unpleasant rules or to obey a totalitarian regime if there exists an effective network of mutually reinforcing social sanctions against disobedience and a system of converging expectations that sustain the existing arrangement. The underlying intuition is that a bad institution or a harmful rule persists due to mutual suspicion between people, and the immediate implication of the existence of a web of self-reinforcing sanctions is that everyone is both a victim and a supporter of a system in which there need not even exist a power- wielding central authority. For these sorts of effective sanctions to prevail, meta- punishment must be applied; that is, an individual is considered disloyal to a regime or a rule if either s/he does not cooperate or s/he maintains relations with someone who is disloyal (Kuran, Reference Kuran1995: 118–36; Basu, Reference Basu2000: 136–47).

Raising the question of institutional change in this new framework amounts to asking how people can possibly extract themselves from the current arrangement. Logically, they can succeed in this only as a result of exogenous shocks that modify some key parameters of the situation produced by their history and culture. As pointed out by Greif and Laitin (Reference Greif and Laitin2004: 633): ‘A self-enforcing institution is one in which each player’s behaviour is a best response. The inescapable conclusion is that a change in self-enforcing institutions must have an exogenous origin’. The shocks can have different sources: they may consist of technological, economic, demographic, environmental, or political changes that end up modifying the payoffs accruing to some or all the agents, of changes that bring new actors onto the scene and/or remove old ones, or of changes that enlarge the repertoire of available actions or alter the expectations of agents regarding others’ actions. In such cases, new equilibrium outcomes become possible while old ones may no longer be accessible.

A well-known illustration is based on the distinction between inclusive and extractive political equilibria, as described by Acemoglu and Robinson (Reference Acemoglu and Robinson2012) and by Sokoloff and Engerman (Reference Sokoloff and Engerman2000). While the latter are based on highly oppressive economic and political institutions, the former are characterised by participatory politics, checks and balances on politicians, and effective property rights for a broad cross-section of society. Once a country has attained either equilibrium, it gets trapped in it. For precisely this reason, an exogenous shock or a chance event is required to move a country from the vicious (extractive) to the virtuous (inclusive) equilibrium. As in the case of the first approach (induced institutional innovation), though, the theory is essentially static and therefore does not provide clues about the pathway leading from one equilibrium arrangement to the other. Since many agents are now involved, the issue of how they succeed in establishing the best institution following the occurrence of a beneficial shock is even more complicated than in the first approach.

To take another example, consider the aforementioned example of collectivist societies described by Greif: those societies were embedded in an institution that was dominated by the mechanism of multilateral reputations and sanctioning. Grounded in continuous relationships between members of a network, this arrangement enabled merchants to build trust, enforce contracts, and expand their activities in Europe and the Mediterranean during early modern times, and it is not coincidental that similar trade networks, or trading diasporas, have been equally successful in Africa as well (Platteau, Reference Platteau2000: 261–2). As trade opportunities expand, however, the limited size of a network may become a constraint on further capital accumulation and growth. What is needed, then, is a shift of the society from the collectivist institutional setup to a more individualistic setup in which political enforcement organisations operate.

Call ‘relation-based governance’ the informal mode of contract enforcement documented by Greif (i.e., contracts involving reputational effects arising from repeated transactions), and ‘rule-based governance’ the system that relies on formal contracts and their enforcement by the courts or the police (Dixit, Reference Dixit2007: 141–2). Li (Reference Li2003) has aptly clarified the analytical conditions under which each system is preferable to the other, and when the latter should replace the former (see also Dixit, Reference Dixit2004). Relation-based governance has small fixed costs, since it can operate on the basis of existing networks of relatives and friends. But its marginal costs are large, and they increase substantially as the scope of trade expands and requires the enlargement of the merchant network beyond close acquaintances who are generally trustworthy. In contrast, rule-based governance entails high fixed costs in the form of laws, regulations, regulatory agencies, and courts. Yet, once these formal institutions are in place, business deals with strangers can be struck at low marginal cost. The conclusion is rather straightforward: the relation-based system is scale-limited, being better at small scales of transactions but inferior to the rule-based system at large scales. But that leaves entirely open the question as to how, starting from an informal relation-based system, a society succeeds in moving to a formal rule-based system when such a shift becomes necessary.

What bears emphasis is that in most cases countries do not switch entirely to purely formal institutions: a lot of economic activity, even in the most advanced countries, continues to be governed by relational and private ordering ‘under the shadow of the law’ (Dixit, Reference Dixit2007: 143–4). To the extent that formal and informal rules are complementary, this feature does not create problems but actually helps promote growth in a rather smooth manner (see Section II, subsection B).

A clear advantage of the last two approaches to institutions and institutional change is that they both lead to the conclusion that there is not one but several pathways to modern economic growth and development. The last approach, to which we now turn, comes to the same conclusion.

D The Political Economy Approach: Why Are Institutions and Policies Inefficient?

The discussion of this last approach proceeds in two steps. To begin with, we examine the many obstacles that can block the decentralised mechanisms whereby losers from socially beneficial institutional change are appropriately compensated. The need for an effective state emerges from this failure of a decentralised functioning of the society. Unfortunately, as argued in the second step of our analysis, the state may also fail for reasons that the political economy approach can help us to understand.

1 The Coase Theorem

Consider the frequent situation in which the aggregate benefits of an institutional change exceed the costs, but some individuals or groups are going to lose from the shift to the new institution. The efficiency principle that underpins the well-known Coase theorem indicates that this should not be an obstacle to change: under certain assumptions, through decentralised bargaining and transfer payments, agents should be able to reach an agreement that ensures that the losers are duly compensated by the winners and therefore agree to depart from the status quo. In these conditions, moreover, the choice of the efficient institution, which is supposed to be unique, does not depend on the a priori distribution of power between the parties involved. The latter will affect only the distribution of the costs and benefits of the change. In short, the issues of efficiency (selecting the efficient institution) and distribution (how the gains from institutional change are shared among all the participants) are separable (see Milgrom and Roberts (Reference Milgrom and Roberts1992: 35–8) for a detailed argument).

The problem is that, whether explicit or implicit, the assumptions required for effective decentralised bargaining and for absent impact of power asymmetries on the efficiency of institutional choices are very restrictive. We review them because their limitations point to important barriers to institutional change. Let us start by stating the efficiency principle: if people are able to bargain together effectively and can effectively implement and enforce any agreements they reach, they should be able to realise all the gains caused by a shift from an inefficient situation to an alternative that everyone would prefer (Milgrom and Roberts, Reference Milgrom and Roberts1992: 24). Two explicit assumptions are: (i) negotiating costs are nil and (ii) no wealth effect is at play. The former assumption is clearly violated when many people are involved, or they have heterogeneous characteristics. It is true that bargaining can be entrusted to group representatives, but then all sorts of collective action problems arise, including the delicate issue of leadership (see Baland and Platteau, Reference Baland and Platteau1996: Chaps. 5–7, and Reference Baland and Platteau1997, Reference Baland and Platteau1998, for an extensive treatment). Moreover, if one party can more easily solve these problems than the other can (think of the old Smithian problem of a well-organised group of colluding producers facing a large group of consumers), the resulting institutional choice will favour the former and not necessarily coincide with the optimal solution. As for the second assumption, it is realistic only when the amounts of the compensatory transfers are small relative to the agents’ financial resources; that is, when the stakes of institutional change are not too high. If that is not the case, the very payment transfer process is going to affect the wealth of the negotiating parties, who will then modify their transfer demands (for the losers) or their willingness to satisfy them (for the winners).

But there is another bargaining-related assumption that is often ignored: the reasoning is typically based on a two-agent framework where the outcome of bargaining is rather straightforward. As Shapley and Shubik (Reference Shapley and Shubik1969) have shown, however, when there are more than two agents, a bargaining solution may not exist. More precisely, an efficient solution may exist, yet the parties will not be able to reach it through decentralised bargaining. Moreover, whether a bargaining solution is attainable or not may ultimately depend on the initial assignment of rights to the parties; that is, on the initial distribution of bargaining powers. In this instance, the separability between efficiency and distributive issues is clearly broken (Baland and Platteau, Reference Baland and Platteau1996: 51–2).

To these major difficulties a number of other snags must be added. Three of them deserve particular attention (see Platteau, Reference Platteau, Brousseau and Glachant2008: 447–9). First, people may not behave rationally, implying that they may seem to act against their interests. This is the case not only when they stick to a particular cultural prescription (a social norm may even prohibit compensatory cash transfers, at least when they involve certain categories of people), but also when they are imbued with a strong sense of justice that takes precedence over cold calculus. As an example of the latter possibility, common people may oppose awarding compensation to an erstwhile elite (say, big landowners) who are going to lose from a socially beneficial land reform. Their opposition may be justified by the excessively large benefits that this elite drew from the existing arrangement in the past. Ideology then trumps rational bargaining based on present and future gains.

Second, negotiations are complicated if winners and losers have a different assessment of the costs and benefits of institutional change. Situations that immediately come to mind here are those in which poor people, whose preferences exhibit high discount rates, live in the same institutional setup as richer people, who have more future-oriented time preferences. Asymmetric evaluation of institutional change may also originate in particular worldviews or ideologies rooted in past experiences. An apt illustration is provided by the resistance of domestic producers against an opening of the national economy to external trade on account of an excessively pessimistic evaluation of the costs that such a change would entail for them. Such kind of resistance is expected to be especially strong in countries with a protectionist tradition. Of course, it may just be the case that genuine uncertainty prevails about the benefits and costs of institutional change. It is then highly likely that among the several bodies of knowledge or expertise available, the parties at stake will invoke the one that serves their particular interests best. For example, in fishing conflicts between small-scale fishermen and industrial vessel owners, the former tend to attribute their declining catches to the encroachment of the latter in coastal waters, while the latter tend to blame excessive fishing effort on the part of the former, say as a result of adoption of more effective technologies. Each party seeks the support of marine biologists who produce evidence in their favour, and while small-scale fishermen demand the enforcement of an exclusive fishing zone for themselves, industrial fishermen and owners strongly resist this demand.Footnote ¹³

A third source of failure of the Coase theorem is the double commitment problem, which makes the outcome of decentralised bargaining hard to implement and enforce. Potential losers may typically not trust the potential winners’ promise to compensate them once the institutional change has occurred. Realising the transfers before the advent of the change is of no help: the inverse problem would then arise, since the potential winners may now fear that the incumbents will opportunistically collect the payment and then oppose the agreed-upon change.

For all of these reasons, efficiency and distributive issues are generally not separable, and institutional choices tend to reflect the distribution of power in society, rather than efficiency considerations. To increase efficiency, the intervention of the state is needed. However, as soon as one drops the naïve assumption of a benevolent state or social planner, that is, the assumption of a central agency that maximises the aggregate welfare of the population, the problem of power resurfaces. Precisely this problem is at the heart of the political economy approach: its main novelty consists of positing the state as a full-fledged actor pursuing its own (selfish) objectives, featuring the relative power weights of the various actors and the negotiating arenas existing in a society (see Acemoglu and Robinson, Reference Acemoglu and Robinson2000, Reference Acemoglu and Robinson2006a). In the new framework, an efficiency-improving institutional change will not occur if the interests of the state elite are aligned with those of a group of people opposed to that change. This is especially likely if organisations and interest groups that benefit from the current (inefficient) arrangements are able to shape the polity according to their wishes, thus making it still more difficult for change to come in the future. Ideology can also play a role, such as when agents or groups construct rationalisations aimed at vindicating the prevailing rules and structures, thereby accounting for their poor objective performance (North, Reference North1990: 99).

2 The Political Coase Theorem

It is useful to stress the analogy between some of the aforementioned problems undermining the applicability of the Coase theorem in a decentralised context, on the one hand, and the problems that arise when a centralised state exists that fulfils important functions, on the other hand. As argued by Acemoglu (Reference Acemoglu2003), an extension of the reasoning underpinning the Coase theorem to the political sphere would mean that political and economic transactions should push towards efficient policies and institutions (i.e., those that achieve the best outcomes for the society), regardless of who, or which social group, holds political power (p. 621).Footnote ¹⁴ In actual practice, however, politicians and powerful social groups do not typically strike deals with the rest of society for the purpose of adopting policies and establishing institutions that maximise aggregate production or social welfare while ensuring the redistribution of part of the gains to themselves. In short, they do not ‘predate efficiently’. At the root of this recurrent state failure or political failure are ‘inherent commitment problems associated with political power’ (p. 622; see also Besley and Coate, Reference Besley and Coate1998).

More precisely, while the validity of the so-called political Coase theorem rests on the ability of the state to write enforceable contracts, the state, or the social groups controlling it, cannot actually commit to not using their power to renege on their promises, or to not modifying the terms of the contract, if their selfish interests dictate that they should do so. Because contracts written by the state are by definition unenforceable – there is no state above the state – an inherent commitment problem thus arises that prevents efficient outcomes from materialising. On the other hand, if the rulers relinquish their power, the citizens cannot commit to making side payments to them in the future, since they will no longer possess the political power to enforce the agreed-upon promises. This is the political form of the double commitment problem mentioned above. In this view, problems of credible commitment ultimately appear as problems about the future exercise of political power.

Because the relationship between the state and the citizens is repeated, it is possible to conceive some commitment based on reputation and supported by the threat of future punishment (Acemoglu, Reference Acemoglu2003: 623). This is the function of political constitutions and other institutions (e.g., the delegation of monetary policy to a politically independent central bank) that are intended to provide checks and balances on the power of the rulers. The commitment problem cannot be completely overcome, however, and observation of the reality in many developing countries, in particular, confirms that states often choose inefficient and even disastrous policies and institutions because these choices ‘are not made for the benefit of society as a whole, but for the benefit of those who control political power’ (p. 648). A powerful ruler can thus flout a constitution that prescribes political mandates of a finite duration or that violates the independence of the central bank.

A general conclusion of the political economy literature addressing the problem of credible commitment is that the degree of inequality in the distribution of the costs of reforms, or of the benefits from the status quo, plays a critical role in determining whether reforms will be feasible or not. More precisely, when the inequality of power, or of the benefits from status quo policies or existing institutions is quite large, elites are prompted to oppose reforms in order to be able to hold on to their power and the associated rents in the future. Because the initial inequality in wealth and power is high, it is then difficult to commit in future to making everyone better off without making the elites worse off (Khemani, Reference Khemani, Dixit and Hamilton2020: 5). Only through revolution or through a change in the composition of the elites and their preferences can key reforms become possible.Footnote ¹⁵

Other reasons than unequal initial endowments can explain political opposition to reforms. These have already been mentioned in our discussion of the failures of the Coase theorem. In particular, uncertainty may exist over the distribution of the costs and benefits from reforms, due to a lack of solid technical evidence, to heterogeneous preferences leading to different evaluations of the pro and contra of a change, or to ideological resistance to available evidence. Regarding the latter source of uncertainty, it is worth noting that propagation of ideological prejudices against a reform may actually be driven by powerful groups whose interests would be harmed by its implementation. In all these cases, commitment problems will necessarily emerge and will block change (Majumdar and Mukand, Reference Majumdar and Mukand2004). Moreover, a bias towards the status quo may be caused by the fact that some of the individual winners and losers from the reform cannot be identified beforehand. For instance, it may be difficult to predict ex ante which precise sectors and firms would benefit from trade liberalisation reforms. When individuals ignore how they will fare under a reform, aggregate support for it is likely to be lower than it would have been under complete information. This is true even when individuals are rational, forward-looking, and risk-neutral, and when there is no aggregate uncertainty. Under these conditions it may happen that, once enacted, reforms will receive adequate political support, but they would nevertheless fail to carry the day ex ante (Fernandez and Rodrick, Reference Fernandez and Rodrick1991).

Clearly, one important objective of the IDP is to reduce uncertainties and information asymmetries so as to ease the reform process by creating conditions conducive to an effective application of the Coase theorem.

3 Norms and Preferences for Public Goods

More recent strands of the political economy literature have turned their attention to problems beyond the credible commitment issue. One of them approaches resistance to reforms through the angle of norms, while the other refers to preference for public goods. The starting point of the norm approach is the idea that if a contract among rival interest groups is not enforceable or justiciable, it has to be self-enforcing. Here lies a prisoner’s dilemma, since each group can extract a private gain by reneging. Consider the case where several competing interest groups, acting as multiple principals, attempt to control a government institution or department (the common agency). Each of these groups has an interest in offering the agency a little bit more than what others are offering in order to advance its interest at their expense (Dixit, Reference Dixit1996, Reference Dixit2003). Thus, Dixit (Reference Dixit, Basu and Cordella2018) argues that persistent and endemic corruption should be seen as a prisoner’s dilemma in which, say, business groups are ‘givers’ of bribes. Although all of these groups would benefit from reduced bribes when taken as a whole, they stick to their current practice because they believe that this is the best they can do given the behaviour of the others. Absent some coordination mechanism or norm to regulate corruption, the high-level corruption equilibrium will persist.

That such mechanisms or norms are hard to come by is attested to by the general failure of anti-corruption policies, such as those embedded in the creation of the anti-corruption agencies that are so much favoured by the donor community (Olken and Pande, Reference Olken and Pande2012; Khemani, Reference Khemani, Dixit and Hamilton2020: 7–8). These failures tend to be attributed to the lack of legitimacy of political leaders. A leadership’s legitimacy, if one follows Akerlof (Reference Akerlof2017), is functionally equivalent to agents monitoring each other with a view to ensuring compliance, implying the existence of some norm among peers regarding performing duties that are prescribed by a (legitimate) leader. The problem, however, is that in many instances political leaders are themselves corrupt – they extract rents from public office, indulge in vote buying and the like – or they are perceived as corrupt by the population. Therefore, they are not in a position to implement reforms aimed at combatting not only big cases of embezzlement but also more ordinary forms of petty corruption. Clearly, to break the vicious cycle of corruption, there is a need for exceptional and strong leaders, such as Kemal Atatürk in Turkey or Habib Bourguiba in Tunisia (Platteau, Reference Platteau, Basu, Ghatak, Kletzer, Mundle and Verhoogen2022). When available, these leaders can serve as role models capable of starting a virtuous mechanism of ‘clean politics’.

When resistance to beneficial reforms is viewed from the standpoint of preference, stress is laid on the fact that citizens may vote in ways that do not properly account for negative consequences. In particular, Stuti Khemani (Reference Khemani, Dixit and Hamilton2020) writes that ‘Citizens’ preferences for emphasizing private benefits over public goods when evaluating leaders, either delivered by co-ethnic leaders, or through vote buying at the time of elections, appears to be a significant factor explaining the persistence of clientelistic politics in democracies in the poor world and its negative consequences for political incentives to pursue reforms’ (p. 11). Citizens actually appear to be cognitively constrained such that they demand policies that confer short-term benefits at the expense of long-term costs (Dal Bo et al., Reference Dal Bo, Dal Bo and Eyster2017). Clientelistic politics may thus cause public goods or services to be under-provided even when they seem to favour the poor disproportionately, like in the case of primary healthcare and anti-poverty programmes or policies (Khemani, Reference Khemani2015; Anderson et al., Reference Anderson, François and Kotwal2015; Bardhan and Mookherjee, Reference Bardhan, Mookherjee, Baland, Bourguignon, Platteau and Verdier2020; Bardhan et al., Reference Bardhan, Mitra, Mookherjee and Nath2020).

Moreover, because the poor often place a high value upon their ethnic affiliation, leaders and their political parties often find it a more effective strategy to supply targeted private benefits to a special clientele than to supply public goods from which many people will benefit (Banerjee and Pande, Reference Banerjee and Pande2007). Following this interpretation, political corruption partly stems from the fact that citizens value honesty in their leaders less than they value their ability to supply public resources on the basis of shared ethnic or caste identity. Stark evidence of this possibility has been adduced by Prakash and co-authors (Reference Prakash, Rockmore and Uppal2019): in India people may not hesitate to cast their votes in favour of politicians accused of criminal actions, even though these politicians generate bad economic outcomes (Prakash et al., Reference Prakash, Rockmore and Uppal2019). In sum, ‘vote buying could be a form of political responsiveness to the demands or needs of voters’ (Khemani, Reference Khemani, Dixit and Hamilton2020: 12).

A key problem with the cognitive thesis, however, is that an alternative explanation exists to account for the poor’s demand for the wrong kind of goods and services. This explanation also relies on a preference characteristic, namely high time discount rates among the poor. Perhaps the cognitive constraint argument is the most relevant when the poor, as they often do, assign a high value to social events and religious festivals. On these occasions, they tend to spend large amounts of money that could have been used for more welfare-enhancing purposes, even in the short or medium term. But even here, rather than invoking cognitive constraints (or preference patterns that have arisen evolutionarily and have become inefficient), one can conceive of other explanations, such as the importance of social identity in the poor’s preference schedule, the need to build up social capital, or the role of the elite in shaping the poor’s preference in a way that legitimates their own power and influence.

The problem of reforms is therefore immensely complex, especially when powerful interests derive enormous advantages from the status quo. Lack of information about the value of reforms and heterogeneous evaluation of its benefits and costs among the parties at stake may also constitute a strong impediment. Worse still, even well-informed citizens may oppose a reform if they are reluctant to change their beliefs and preferences even when shown that their beliefs are wrong (Barrera et al., Reference Barrera, Guriev, Henry and Zuravskaya2020). In the end, the credibility and legitimacy of reformist leaders are a crucial factor for success, but this is not a very reassuring conclusion given that we do not know where credibility and legitimacy come from (Khemani, Reference Khemani, Dixit and Hamilton2020: 16).

4 The Political Economy Approach: Illustrating State Capture and State Failures

We can now briefly discuss a few interesting applications and illustrations of the argument according to which political elites may drive their societies away from efficient or development-enhancing institutions and policies, as well as of the role of beliefs and preferences. To begin with, we wish to refer to the political theory of economic backwardness, through which Acemoglu and Robinson (Reference Acemoglu and Robinson2006b) show how and in which conditions state elites may deliberately thwart development. Their basic intuition is that political elites face a trade-off between economic gains and power. All else being equal, they prefer technological progress and prosperity-inducing reforms that might increase their economic rents in the future. Yet all else is not equal because such changes can potentially erode their political advantages relative to other groups. They will thus decide to ‘block beneficial economic and institutional change when they are afraid that these changes will destabilise the existing system and make it more likely that they will lose political power and future rents’ (pp. 115–16). The theory predicts a non-monotonic impact of political competition on resistance to development: while political elites that are either subject to intense competition or do not face any competition due to their complete domination of the electoral landscape adopt new technologies, elites occupying an intermediate position between these two extremes will adopt the opposite attitude. This is because with intense political competition, elites prefer to innovate lest they should be replaced, whereas strongly entrenched elites do not fear losing political power, so for them there is no trade-off between economic gains and power. By contrast, elites that are ‘somewhat entrenched’ but are still afraid of being replaced are tempted to block innovation to prevent such replacement from happening (p. 116).

Along the same line, within a framework in which education is both the engine of growth and a determinant of political participation, Bourguignon and Verdier (Reference Bourguignon and Verdier2000) have looked for the conditions under which an educated oligarchy may have an incentive to subsidise the poor’s education and to initiate a democratic transition, thereby promoting the endogenous emergence of a middle class. When these conditions are violated, the oligarchy will undermine universal education. There is evidence that, indeed, the education of ordinary people may be blocked by the elite. For example, in the case of Pakistan, Martin (Reference Martin2016) has observed a tendency among traditional landlords to oppose the (secular) education of their dependents lest they should seek emancipation and develop ‘unrealistic expectations, and thereby cause a shortage of cheap agricultural labour (p. 87). Also, assuming a regime of autocracy, Auriol and Platteau (Reference Auriol and Platteau2017a, Reference Auriol and Platteau2017b) have shown that a ruler may deliberately sacrifice institutional reforms in order to placate the opposition coming from traditional leaders (including religious authorities) who resent that their erstwhile prerogatives will be encroached upon (see also Auriol et al. (Reference Auriol, Platteau and Verdier2021), where the army is added as a strategic actor). This raises the serious issue of how a modern state can manage the divergent interests of popular masses under the influence of traditional authorities that want to preserve the old order based on the community, on the one hand, and urban elites that have been exposed to Western values centred upon the improvement of the individual, on the other hand.Footnote ¹⁶

Turning to the role of beliefs and preferences, Chinese history supplies us with an interesting example of the role of ideology. Since the early fifteenth century, Chinese authorities have developed the idea that China does not need to develop its external trade links to become or remain a major economic and political power. Under the influence of the mandarins, who scorned and distrusted commerce, and unlike the eunuchs, who had the opposite interests and beliefs, the new worldview led the Chinese into believing that their country is better off when it relies on its own forces. Carried over well into the nineteenth century, ‘this deliberate introversion’, a major turning point in Chinese history, proved to be a disaster for the country that was leading the world in many respects several centuries earlier. It could not have come at a worse time, since it disarmed China just as European power was rising (Landes, Reference Landes1998: 96; see also Jones, Reference Jones1981: 168–9).

It bears emphasis that in the above discussion the state has been treated as a single homogeneous actor. This is an obvious simplification, not only because it ignores the distinction between the executive, the legislative, and the judiciary, but also because it bypasses possible conflicts existing inside each of these branches. Thus, serious dysfunction of the state apparatus may arise from tensions or contradictions that may come from a variety of different sources: between a government and its administration, between various departments or levels of the bureaucracy, between various political interest groups, or between appointed and elected leaders (say, between district officers and elected councillors). Note that the latter tensions may actually be ignited and nurtured by holders of central power who are eager to ‘divide to rule’ (for a vivid application of this argument in relation to Pakistan, see Cheema et al., Reference Cheema, Khwaja and Qadir2005; Malik et al., Reference Malik, Mirza and Platteau2022). Technically, whatever the source of misalignment, the problem can often be framed as one of multiple principals with conflicting interests (Dixit, Reference Dixit1996, Reference Dixit2002; Martimort, Reference Martimort1996; Bolton and Dewatripont, Reference Bolton and Dewatripont2005: Chap. 13).Footnote ¹⁷

Another important source of efficiency losses lies in the fact that the administration is often undermined by the corrupt behaviour of selfish officers and politicians. Consider the following example of the rise and decline of the effectiveness of land documentation and titling in Kenya (see Onoma, Reference Onoma, Mahoney and Thelen2010). Although this was embraced by the postcolonial government after independence in 1963 (the system was initiated by the British in the early 1900s), the associated practices were gradually eroded so that by the early 2000s they had fallen into disrepute. Ominous signs of this institutional degradation were visible before, as attested by the cancellation of hundreds of title deeds by Kenya’s High Court in May 1991, a dramatic move which led one commentator to wonder ‘whether the assumption that a title gives one indefeasible rights to property was still true’ (p. 65). Onoma’s explanation is that the administration, by exchanging fake land documents for money, largely contributed to the erosion and drift, causing a marked fall in the efficacy of land documentation. At the same time, it appears that civil servants acted under the corrosive influence of well-connected politicians who used land documents to obtain political support. For these politicians, ‘issuing and selling land documents became an excellent way of raising cash for electoral campaigns, for buying the support of various individuals and groups, and for dissuading would-be opponents’ (p. 66).

Onoma sums up his diagnosis by stressing that the problem was not just one of weak state capacity, nor was it that the system of land documentation was not well established in the first place. The problem was rather that state agencies became increasingly less willing to use their powers to stop the fraudulent exploitation of land documentation by people close to the state (p. 66). Herein lies a prisoner’s dilemma between the users of the law: an institution serving the common good (an efficient institution) has been created or confirmed, yet key participants find it in their own selfish interests to break the rules and thereby undermine it. At the root of the problem lies a system of political patronage that prompts key political actors to instrumentalise the administration for their own immediate benefit, rather treating it as a vehicle for advancing the national interest.

Other problems have plagued land titling in Kenya, including the usual problem of a lack of proper updating of land records. One of these problems in particular deserves our attention because it vividly illustrates the role of ideology and the political influence of the common farmers rather than the Big Men on whom attention is usually focused. In Kenya, lenders have usually faced great difficulty foreclosing on land mortgages in the face of determined opposition from family and community. The fact of the matter is that ‘the presence of many kin around mortgaged land makes it politically unfeasible to auction the holdings of defaulters’ (Shipton, Reference Shipton, Downs and Reyna1988: 120; see also Migot-Adholla et al., Reference Migot-Adholla, Hazell, Blarel and Place1991: 170; Ensminger, Reference Ensminger, Drobak and Nye1997; Platteau, Reference Platteau2000: 145–7, 153–6). One important politico-ideological reason behind the government’s reluctance to strictly enforce the law after independence was its fear of breaking the fragile political consensus on which national policies rested. In the words of Bates (Reference Bates1989): ‘The cry of land hunger had fed the nationalist rebellion that had brought the government to power. To turn people off the lands that they had fought to capture would be to risk the wrath of the true believers in the nationalist revolution’ (p. 74). The pressure on the government was all the stronger as the official opposition, represented by a radical party (the Kenya African Democratic Union, KADU) lobbied intensively on the land issue (pp. 67–8).

The stylised approach of the political economy approach has the advantage of supplying us with structured treatments of well-defined issues involving politics, as well as useful predictions of the effects of changing components in the environment. However, because the inter-relationships between institutions, politics, and development, and the dynamics of institutional change, are so complex, the approach needs to be complemented by contributions from other disciplines: political science, history, and socio-anthropology in particular. As attested by the above example about Kenya, these contributions are expected to provide a lot of details on the concrete processes of institutional change, as well as the specific problems they raise in particular contexts. The same interdisciplinary perspective and the same concerns with context and institutional details underlie the IDP approach that led to this book.