Development, social norms, and assignment to task

Abstract

Economic development involves a structural transformation in the way people are allocated to tasks. There is a shift from self-provision to market exchange, facilitating specialization. There is also a shift from self-employment to wage employment in large firms and organizations, driven by innovation and increasing returns to scale. Changes in allocation mechanisms require changes in norms and attitudes. Because different labor assignment domains coexist, conflicts arise among norms that apply to different domains, possibly resulting in dysfunctional outcomes. I argue that religion, humanism, and schools have all played an important historical role in fostering the changes in social norms and attitudes that are needed to accompany structural changes in the way economies allocate workers to tasks.

Much has been written about the relationship between economic development and market institutions. The literature has emphasized the need for societies to evolve from personalized moral obligations (e.g., to kith and kin, from vassal to feudal lord) to generalized morality and the respect of arms-length contracts (1–5). Institutional evolution is often portrayed as the replacement of certain social norms with others, recognizing that any enforcement mechanism only works when compliance is sufficiently widespread. Hence, the development of market institutions is typically depicted as the replacement of social norms emphasizing loyalty to a group or family with social norms emphasizing the respect of contractual obligations resulting from market transactions (e.g., contracts regarding labor, credit, insurance, service provision, warranty, and the like). These insights have been immensely influential, and they are not in dispute.

The purpose of this paper is twofold. The first objective is to propose a simple way of thinking about which changes in social norms are needed for development to take place. Rather than focusing on the rarefied confines of high finance and its need for sophisticated institutions, I focus on the social norms that affect the majority of the population. The organizing concept behind this approach is the relationship between economic growth and institutional mechanisms for allocating workers to tasks. I illustrate how looking at social norms from this vantage point casts new light on various phenomena often discussed in the context of development.

The second objective of the paper is to demonstrate that the transition in social norms required for development is not a simple one. First, the domain of application of ancient norms shrinks but does not entirely disappear as economies grow. Economic agents thus cannot abandon one set of norms for another. Rather, they must adapt the domain over which different sets of norms apply. Second, multiple sets of partially conflicting norms are required for development, some supporting markets and entrepreneurship and others supporting wage employment and the functioning of large organizations. Development is characterized by various shifts in the respective domains of these norms of behavior. Third, the overlap among domains of application of different norms results in strategic complementarity or substitution among different norms across their domain of application. This, in turn, molds the process of institutional change, potentially generating outcomes that are inefficient and/or inequitable. I provide some examples.

In this paper, I take social norms to mean shared understandings about behavior and attitudes. Sometimes, social norms are self-enforcing: Knowledge that they are sufficiently widely shared makes it optimal for each individual to abide by them. In other situations, they need supporting enforcement mechanisms of social sanctions. These sanctions can take many forms and need not rely on the legal system. Fairness is one possible manifestation of social norms. Recent experimental literature has indeed shown that people are willing to punish others for behavior they do not regard as fair (6, 7). What is regarded as fair varies across societies, however. In some, not paying a landlord if a distant relative needs money is considered fair, whereas it is not acceptable in others. The purpose of this paper is to identify the norms and norm shifts that play an important role in the process of development from self-subsistence to family-based enterprises and to large firms. The issue of enforcement is considered in detail elsewhere (8).

Development and Assignment to Tasks

Allocation of workers to the right tasks has a significant impact on economic performance. This is true in general but particularly so during the development process, when innovations in technology, institutions, and organizations change the nature of the tasks to be performed. The recognition that development implies a structural transformation of the economy from self-subsistence to market-oriented production goes back at least to the work of Rosenstein-Rodan (9) and Lewis (10), and it forms the cornerstone of the work of Hirschman (11) and Chenery et al. (12), among others. Recent models of the structural transformation process include, for instance, the work of Murphy et al. (13), Matsuyama (14), and Fafchamps (15). The faster development is, the faster tasks change and the more important the reallocation of workers to new tasks becomes.

It is easy to forget that labor markets are not the only way of allocating workers to tasks. Workers can also be allocated to tasks within firms and organizations, typically through command and control, or within households or kin groups, typically via gift exchange. The allocative function of the labor market therefore depends on how much production takes place within firms and organizations as well as within households.

The labor market matters for the allocation of workers to enterprises. However, even in this case, the allocative role of the labor market depends on whether labor contracts are of short or long duration. If labor contracts are short, workers are allocated to tasks through the labor market. In contrast, if labor contracts are long, workers are allocated to tasks through command and control: If a task is needed, a worker from within the firm is allocated to it; if a task is no longer required, the worker is reallocated to another task.

The allocation of workers to tasks can also be organized by markets for goods and services. Consider microenterprises using no wage workers. In this case, employment creation and firm creation coincide. Workers are “told” what to do by the demand for their products: If demand is there, the enterprise thrives, but if demand is absent, the enterprise closes down or is never created. The supply of inputs and raw materials also dictates which firms are viable.

It follows that the importance of the labor market as an allocation mechanism depends on the size distribution of firms. When firms are small, workers are primarily self-employed and are allocated to tasks by markets for goods and services and their effect on enterprise creation and survival. In contrast, when firms are large, workers are initially allocated to tasks via the labor market; however, once in a permanent job, hierarchical allocation takes over.

The development process affects the mix of labor allocation mechanisms by shifting three boundaries: (i) the boundary between what is self-provided and what is purchased from the market, (ii) the boundary between what is provided by microenterprises and what is provided by large firms, and (iii) the duration of employment contracts. The size and allocative role of the labor market depend on where an economy is along these three dimensions.

Self-Provision and Gift Exchange

At low levels of economic development, self-provision is the rule. This is still true in many rural areas of the developing world today, where many goods and services are self-provided within the household. Examples include food crops; milk and other animal products; firewood and water; food processing and meal preparation; house construction, maintenance, and repair; child care and elderly care; insurance against various shocks; and entertainment. What is not self-provided within a single household is obtained via gift exchange among households. Examples include access to land; personal services, such as haircuts; accommodation for visitors; transfers at weddings and funerals; and various labor pooling arrangements.

For goods and services that are allocated through self-provision, there is no role for market exchange. The allocation of workers to tasks is based on gift exchange or reciprocal exchange among mostly coresident, and often related, individuals. When, as is often the case in poor rural areas, households are large and integrated, either vertically or horizontally, the range of goods and services that is covered by self-provision can be quite large.*

Gift exchange can also regulate some of the transactions taking place among households. A good example is when members of a lineage work together to construct or repair a house. Another common example is when a vulnerable member of society (e.g., elderly, orphan) is integrated into a preexisting household for the purpose of providing elderly care or child care.

Within households, the allocation of workers to tasks is determined partly by social norms and partly by intrahousehold bargaining (16–18). The formation of a new household marks the creation of a new self-provision production unit. It follows that the allocation of adults to tasks is ultimately regulated by the marriage market. For instance, the assortative matching of spouses on wealth or education generates a correlation between their productive potential and tends to reproduce social differentiation across generations (19). The resulting allocation is typically inequitable (e.g., the rich marry the rich). This negatively affects social mobility, especially in agrarian societies, where the most important productive asset is land.

Parents often play a role in selecting a mate, although the extent of their involvement varies over time and across space (18). Whether or not they participate in the selection of a spouse, parents always have a profound influence on the start-up human and physical capital that the new household has at its disposal for self-provision purposes. In agrarian societies, parents are expected to endow the starting household with land and livestock to produce its own food, and possibly with a house and furnishings. The extent to which a new household is able to provide for itself adequately thus depends on the social and economic background of the bride and groom.

Nothing in this process, apart from parental goodwill, guarantees that workers are allocated to tasks at which they are most productive. Using data from rural Pakistan, Fafchamps and Quisumbing (20), for instance, show that the allocation of tasks among household members only partly reflects comparative advantage; much of what people do is determined by social norms and spheres of responsibility within the household. They find evidence of static returns to specialization: Given that most household tasks are not a full-time occupation, it is typical for tasks to be divided up among household members, with little team work unless the task (e.g., farming) requires many hands. They also find that people often change tasks within a given gender-specific sphere. They interpret this finding as suggesting that dynamic returns to specialization within the household are small (i.e., whatever skill people learn to perform their assigned task, it can be learned relatively quickly). Two exceptions stand out: farm work and wage employment, two tasks that are more market oriented. If a household member is observed at either of those tasks in a given year, it is extremely likely that the person will be observed performing the same task years later. The same is not true for household chores, which tend to be rotated among members of large households.

Working for the Market

Self-provision means that individuals are not specialized: They undertake many different tasks at different times of the day or year (21). Because the range of skills they can acquire is limited, they are not necessarily very good at what they do. Gains from specialization can be achieved when workers focus on a smaller range of activities at which they become really proficient. For this to be possible, they must provide the good or service not only to their immediate friends and relatives but to a larger number of people. This means offering the goods and services through the market.

If people are geographically isolated, this often is not possible because of transport costs. It is therefore not surprising that urban and periurban areas have less self-provision and more market provision. This point is examined in detail by Fafchamps and Shilpi (21, 22) and Fafchamps and Wahba (23). Fig. 1, taken from a study by Fafchamps and Shilpi (22), illustrates some of the patterns observed in empirical data. The further away households live from urban centers, the less engaged they become in nonfarm wage employment and the more engaged they become in agricultural self-employment. Agricultural wage employment initially rises with distance from towns but eventually falls as commercial farming becomes unprofitable. Nonfarm self-employment is more prominent in the vicinity of towns and market centers, although the difference is not statistically significant in Fig. 1.

Fig. 1.

Employment and distance from towns. The x axis represents the distance from town in hours, and the y axis represents share of work. Nonfarm wage employment (Upper Left), nonfarm self-employment (Upper Right), agricultural wage employment (Lower Left), and agricultural self-employment (Lower Right). Reprinted from ref. 22, with permission from Taylor & Francis Ltd.

The gains from specialization that are made possible by physical proximity are a contributing factor to agglomeration effects. Consumers also derive some direct utility from the expanded range of choice made possible by close proximity to varied goods and services produced by specialized producers. In their simplest form, these gains from specialization trigger an entrepreneurship revolution by which individuals learn skills that enable them to produce for the market. Specialization does not imply that firms are large, however. To take a well-known example, that of Adam Smith's pin factory, specialization in specific tasks makes learning possible and increases productivity. However, the different tasks, such as packing or cutting thread, need not be performed within the same firm; they can be performed in separate firms. In that case, the gains from specialization are achieved through the operation of the market.

Agricultural trade is a good example of specialization through the market. It is possible to conceive of a vertically integrated distribution system whereby the same firm buys grain from farmers and sells to final consumers. However, in developing countries, the different tasks required to bring grain from producers to consumers are typically undertaken by a multitude of independent entrepreneurs (e.g., collectors, wholesalers, brokers, retailers, hawkers, millers, transporters). This entrepreneurship revolution underlies the rapid urbanization of much of the developing world, a process often described as the growth of the nonfarm sector or the “informal” sector. Self-provision is replaced with market provision, but it does not necessarily translate into the emergence of large firms. This point is made clear in Fig. 2 below [taken from the study by Fafchamps and Shilpi (21)]. This figure shows that firm size only shows a weak positive correlation with urban proximity within individual sectors. This is particularly noticeable in manufacturing and in trade, transport, and restaurants, where most microenterprises are found.

Fig. 2.

Firm size and distance from towns. The x axis represents distance from towns, and the y axis represents the ordered probit coefficient. Agriculture (Upper Left); manufacturing and construction (Upper Center); private services (Upper Right); public services (Lower Left); trade, transport, and restaurants (Lower Center); and all sectors (Lower Right). Reprinted from ref. 21.

The social norms required for the emergence of small entrepreneurship are not the same as those governing self-provision. In market exchange, reciprocity is immediate (i.e., the consumer pays for the good or service). In self-provision, there is no immediate payment. Receiving from others generates obligations to reciprocate, but these obligations are delayed and contingent. In market exchange, insurance (e.g., risk sharing) is separate from compensation for effort (e.g., price). This means that people discover hard budget constraints, which is the way by which the market ensures reciprocity and compensation for effort. It is therefore not surprising to observe that simple transaction forms dominate (e.g., cash-and-carry). Producers and consumers have only just emerged from a world of self-provision and are still partially immersed in it for some important services, such as insurance. Consequently, the norms of behavior to which they are accustomed are not those that hold currency in a market allocation world.

In terms of efficiency, the allocation of workers to tasks through self-employment is fraught with difficulty. Job creation is equivalent to the creation of a new enterprise, often of a very small size. Constraints to the creation of new firms, such as limited access to credit or an unsuitable regulatory framework, can therefore impinge on the efficient allocation of individuals to tasks. These constraints have attracted a massive literature that need not be summarized here.

Worker-entrepreneurs do not always know where their comparative advantage lies; hence, there is a high proportion of firm exit. A large-scale study of Zimbabwe, for instance (24), concludes that one third of microenterprises exit in any given year. Given such a high rate of business failure, it is perhaps not surprising that banks and other financial intermediaries are reluctant to lend to microenterprise start-ups. The data show that microenterprise start-ups are nearly all financed with the worker-entrepreneurs’ own funds and with informal loans from friends and relatives.

Hierarchies and Employment Contracts

With the application of science to technology comes the industrial revolution. Innovations in technology (e.g., machines, electrical power) and organizational methods (e.g., accounting, invoicing) enable firms and organizations (e.g., civil service, hospitals) to grow. As firms and organizations grow, wage employment develops.

The social norms required for an effective labor force are not the same as those governing entrepreneurship. In wage employment, opportunism is discouraged; in entrepreneurship, it is essential. In wage employment, discipline is required; in entrepreneurship, personal initiative is essential. The switch from self-employment for the market to wage employment thus entails a change in social norms and expectations of behavior as large as those required to go from self-provision to market provision.

In large firms and organizations, labor management becomes important. Delegation of authority is essential to allow local problem solving; hence, the need for intermediate management personnel. Hierarchies also need specialized personnel to process information (e.g., accounts, reports, memos). These are skilled tasks, and they involve an element of trust. To the extent that skills and trustworthiness are not perfectly observable, screening and monitoring workers are costly.

It follows that casual labor contracts are not appropriate for management and clerical (white collar) workers. Once a suitable worker has been identified, keeping this worker economizes on screening for skills and loyalty. Those workers who are unsuitable or misbehave are either laid off or their contract is not renewed. The fear of losing the job helps discipline workers (25).

This leads to the development of so-called “permanent” employment contracts. In developed economies, these contracts are so pervasive that they are regarded as “normal” employment. Permanent employment contracts dramatically shift the burden of risk from worker to employer. As a result, most people prefer permanent employment to facing the stress and uncertainty of self-employment or casual employment. Evidence of this can be found in queues of school and college graduates waiting for wage employment or civil service jobs (26, 27).

For production workers, things are a little different. Some large firms have permanent workers for management and clerical tasks but casual workers in production. This is fine if production work is relatively unskilled or if there is a large supply of people with the required skills (e.g., sugar plantation, garment firms). However, if production workers with the required skills are hard to locate, finding workers involves effort. The firm may also end up training its workforce because the production process requires task-specific skills. In either of these cases, once a suitable worker has been found or trained, the firm wishes to retain him or her. This leads to permanent employment contracts for production workers as well. The immediate corollary is that if tasks are more standardized and available jobs and workers are easier to match, the need for permanent contracts is reduced (e.g., agricultural wage work is often of short duration).

Large firms and organizations face difficult contract enforcement issues. The literature has typically couched this issue in terms of shirking (i.e., a worker not putting in the required level of effort). When a worker shirks, the employer loses the difference between the worker's potential and realized marginal value product. This difference is seldom larger than the wage itself; thus, the damage caused appears circumscribed. In practice, one bad worker can cause damages to the organization that much exceed their wage or marginal value product (e.g., through pilferage, damage to vehicles and equipment, adverse morale effect on other workers).

How large the damage can be depends on the organization of work within the firm or organization. For instance, an unqualified worker can cause more damage if the firm has expensive machinery than if it has simple tools. A similar reasoning applies for work on an assembly line: Poor work at one point of the chain can reduce the quality of the finished product (28).^†

Because workers are typically insolvent with respect to large damages of this kind, employers have to bear the cost. Because the threat of firing need not be sufficient to induce workers to exert the required level of care, we expect employers to be primarily concerned about worker discipline. If many workers are ignorant of modern organization of work, employers are likely to be even more concerned; for instance, it has been shown that more resources are devoted to worker management in sub-Saharan African manufacturing enterprises than in similar enterprises in middle-income countries (29).

Shifting Boundaries and Shifting Norms

Economic development induces shifts in the boundaries among the domains of application of the different labor allocation mechanisms. As a result, the allocative role of the market changes constantly, on the one hand, taking over tasks that were previously the realm of self-provision and gift exchange but, on the other hand, seeing its allocative rule superseded by command and control within expanding organizations. The domains never disappear; even in the most market-oriented economies, there are still goods and services that are self-provided, but their range of applicability is constantly changing.

Negotiating these changes can be confusing for those involved because a set of optimal norms and attitudes is associated with each allocation mechanism. For tasks assigned through self-provision and gift exchange, allocation is ultimately based on reciprocity: I help you today because I expect you to help me in a yet-to-be-defined way later. There is no hard budget constraint to facilitate risk sharing and redistribution (e.g., the needy do not “pay” for the services they require). The kind of social norms that sustain household self-provision and gift exchange nevertheless emphasize the obligation to reciprocate for gifts received, if one is in a position to do so. Hence, making a gift to someone today can be a way of obligating this person to help you later, a bit like the brave reciprocity of Alexrod (30). Comola and Fafchamps (31) find evidence in Tanzania consistent with such norms. Schecter (32) shows that transfers to other villagers are a condition for the safe enjoyment of rights over land. Self-provision and gift exchange leave much room for altruism and emotions to shape exchange, and thus the assignment of tasks among individuals. Self-provision also encompasses a hierarchical element, with the head of household playing a role similar to that of the firm manager.

In contrast, market allocation is based on clear, and often instantaneous, reciprocation in the form of money. This leaves little room for risk sharing and no room for redistribution: Money is required to obtain goods and services. Although individuals procure goods and services from each other as they would in a gift exchange economy, the workings of the market do not require moral obligations extending beyond the instantaneous transaction and the role of emotions and altruism is minimized.

In market exchange, hard budget constraints serve as an obstacle to free riding: Individuals cannot consume if they do not contribute to society something that is valuable and worth paying for. In gift exchange, there is no such hard budget constraint, such that, in principle, some people can receive without giving in return. Preventing free riding is thus more difficult in gift exchange. Two types of norms can be seen as a way of minimizing free riding. The first is the moral obligation to reciprocate when one can. This has already been discussed. The second is a sharing, or equity, norm. This basically requires that standards of living remain basically in line with each other: One person cannot rise above all others. If someone has good fortune, he/she is expected to share this good fortune with others so as to raise the standards of living of the group as a whole. Equity norms of this kind limit free riding: It is not possible to take from the group without giving in return. However, they do so in a redistributive fashion (i.e., they show no apparent concern for possible discrepancies between someone's standard of living and that person's economic contribution to society). This is in sharp contrast to the hard budget constraint imposed by the market, which establishes a clear link (i.e., monetary income) between consumption and contribution to society.

With economic development, individuals learn to rely on the market to obtain goods and services they previously obtained through gift exchange (e.g., meal, haircut). They also learn to offer through the market goods and services that they previously supplied within the self-provision sphere. For instance, a woman may begin selling meals to people whose occupation keeps them away from home at lunchtime. The task of food preparation moves from one domain, with its rules and obligations, to another domain and another set of rules. What is needed on the supply side is an eye for arbitrage (i.e., asking oneself the question, “Is there a market need for goods and services some people can no longer secure through gift exchange?”) Sometimes, the answer is surprising, as when people begin selling sexual favors, but the underlying process is the same.

The shift from market provision to wage employment calls for another change in norms and attitudes. For people to be efficiently assigned to tasks by the market, they must respond rapidly to arbitrage opportunities: Somebody needs goods or a service, they can supply it, and they do. Swift action is needed; otherwise, the opportunity will be seized by someone else. In contrast, workers in long-term employment are supposed to perform the tasks they have been assigned via command and control.^‡ They are not expected to set their own tasks in response to arbitrage opportunities.

For instance, a civil servant may meet an impatient businessman who needs an import license urgently. This creates an arbitrage opportunity: By providing extra effort, or by delaying work on other tasks, the civil servant could provide the service in exchange for a fee. For a self-employed entrepreneur, this would be the natural thing to do. For a civil servant, it is a violation of the employment contract. Corruption of this kind is equivalent to applying norms from one domain, the market domain, to another domain, the wage employment domain.

Nepotism is another form of confusion among the norms applying to different domains: In the realm of gift-exchange, reciprocating favors is a social obligation; in wage employment, diverting from one's assigned task to reciprocate favors violates the employment contract. The development process therefore creates a lot of potential for clashes among the norms applying to different domains. Furthermore, people have to learn not one but two sets of very different norms, one regulating interactions through the market and another regulating hierarchies. It is therefore hardly a surprise if the development process is often associated with confusion in norms (i.e., when the new norms from one domain, the market, are applied to another, wage employment). Not only do people have to learn new social norms, but they must also learn which norms apply when in an environment that is perpetually changing and role models are few.

To accommodate emerging market opportunities without losing all the risk sharing and redistributive benefits from gift exchange, new norms and mechanisms must be invented. This is the world of monetary gifts and informal loans. Take migrant remittances, for instance. Without the emergence of wage work opportunities elsewhere, in all likelihood, the migrant would have been working on his father's farm until marriage. If the marginal return to labor is sufficiently low on the farm, it is more efficient to let the dependent male work for a wage elsewhere. However, for the family, this means losing out on opportunities for gift exchange within the household. In replacement, the migrant sends remittances in the form of money. These remittances are a monetized form of gift exchange, a transfer that serves a redistributive and insurance purpose. The gift exchange domain has to adapt to the new market reality, and traditional norms of reciprocity and of respect for parental and lineage authority are reinvented in another guise. What this example illustrates is that the shift in boundaries among domains also forces a reorganization of norms within each domain. What is often seen as “traditional” (e.g., risk sharing via monetary transfers and informal loans) seldom is. Rather, the norms and attitudes surrounding informal risk sharing are likely to be relatively recent inventions put in place in response to the emergence of market opportunities.

Coexistence and Leakages

I have argued that in any modern economy, all three domains of self-provision, market work, and permanent employment must coexist. I have pointed out that their coexistence generates tension, and potential confusion, regarding the range of applicability of different norms. This coexistence also generates endless possibilities for one domain to exploit, or be exploited by, another, creating leakages or cross-subsidization from one domain to another.

There are many examples of such leakages from one domain to the other. Employers, for instance, can harness household labor in different ways. They can offer employment contracts that allow the hired worker, typically a prime-aged male, to bring in his spouse and children to assist him. Such contracts were common in European mining in the 19th century, for instance. Alternatively, the employer can offer the worker a piece of land that reduces the worker's cost of living if his household uses the land for food self-provision (33). In both cases, the employer reduces labor costs by gaining indirect access to family labor, including child labor.

A similar process of cross-subsidization from one domain to another occurs in the putting-out system. Here, the worker is self-employed but takes orders that can be fulfilled with family help, including child labor. Contract farming can be seen in a similar light: The farmer is a self-employed provider responding to a market signal, but he typically relies on family labor to complete his contractual obligation.

It is also possible for one domain to portray itself as another domain altogether. For instance, the hierarchical relationship between employer and employee is not too different from the relation of authority between father and son in a patriarchal society. Employers may sometimes find it in their interest to adopt the language of the family domain in the way they approach their employees. In so doing, they may succeed in channeling toward the firm the moral norms of loyalty they have learned in the family domain. This is the realm of paternalistic employers. For this strategy to be successful, the employer probably needs to behave paternalistically toward employees in other areas as well (e.g., housing, health care, protection).

Of course, leakages can also flow in the other direction, as when the domain of gift exchange and self-provision diverts market relationships for other purposes. For instance, an entrepreneur may divert business funds to cover a family emergency, even if this means paying suppliers late. Similarly, a wage worker may divert goods or funds from the firm to repay favors received, or may abuse his office to generate funds to repay these favors.

One particularly pernicious form of leakage is when the moral obligations to one domain are so strong that they become parasitic on other domains and lead to inefficient outcomes. There is one such configuration that has attracted some attention from the literature. It concerns sharing norms and respect toward traditional authorities (34).^§ By taxing market or wage-earning activity, these norms can act as an impediment to investment and firm expansion. For instance, Fafchamps et al. (35) find that cash grants to small enterprises in Ghana do not translate into more working capital and higher profits; rather, it is suspected that these grants are partly appropriated by other household members for consumption purposes. Goldberg (36) reports the results of an experiment in which Malawian participants win a lottery, either publicly or privately. She shows that when participants win publicly, they spend the money faster in what appears to be less useful consumption. Corruption and nepotism are other widely cited examples of leakages from one domain to another that have been blamed for slow development.

Leakages toward the gift exchange domain can thus potentially harm the development of market specialization and the growth of large firms. They are also likely manifestations of the readjustment that societies have to undergo to permit market development and firm growth. The gift exchange domain is capable of providing insurance, something that markets in early stages of development are often unable to do. Until market solutions are found to the population's very real need of protection against shocks, we probably have to tolerate a certain amount of leakages from the market and enterprise domain to the family domain.

In many countries, particularly in sub-Saharan Africa, gift exchange also serves a redistributive function. As we have discussed earlier, this function can be construed as a way of deterring free riding. However, it also means that societies in which the moral norms of gift exchange are still strong are particularly resentful of inequalities in standards of living, and they can apply considerable pressure on the wealthy to redistribute through gifts and the exchange of favors. Corruption at the top echelons of government can often be seen as a way for the rich and powerful to create the cash they need to pacify their kith and kin.

Development and the Evolution of Social Norms

Much of the literature on institutional change and development has discussed the issue as if society needs to shed one set of norms and behaviors and adopt a new one. We have argued that this description is probably too stark: first, development calls for at least two different sets of new norms and behaviors, one for market self-employment and one for long-term wage employment, and, second, the domain of applicability of self-provision/gift exchange shrinks but does not disappear entirely with development, such that the attached norms cannot entirely disappear either. What this suggests is that development requires people to learn multiple sets of new norms, and also which norms are applicable when.

From Self-Provision to Market Provision

From a historical perspective, the transition from self-provision to market provision probably comes first, although there are important early examples of large organizations, such as churches and armies.

Seen from the perspective of gift exchange, market exchange is inherently immoral. It does not allow for risk sharing in transactions. It provides no safeguard against inequitable outcomes. It does not help build the long-term interpersonal relationships that are the hallmark of gift exchange, and thus fails to elicit the same emotional bond. Thus, market exchange initially lacks legitimacy.

The commonly offered solution is to introduce promarket laws and courts to enforce them. There is no denying that the formal legal system can help enforce contracts. It can do so directly (i.e., by allowing economic agents to sue on breach of contract). It can also prop up various institutions that indirectly support market exchange, such as standards and weights, land registration, or a free press combined with libel laws. It can also put indirect pressure on customary legal institutions to reform, as discussed by Aldashev et al. (34).

Most market transactions, however, remain below the radar of formal legal institutions, either because they are too small ever to be brought to court or because the parties involved have no assets to seize. Without the protection of the law, these market transactions are at the mercy of opportunistic behavior. The end result is what I have elsewhere called a “flea market” economy, that is, market exchange reduced to its simplest expression: cash-and-carry only, no warranty, and little quality control (37). Such forms of exchange enable some individual specialization, but they make it virtually impossible for enterprises to grow beyond the microenterprise format. The end result is a bloated informal sector, as can be seen on the roadsides of many developing countries today.

To transcend these limitations, obligations arising from market exchange must receive moral validation. In the literature, this has been discussed under the umbrella concept of generalized morality, that is, a sense of moral obligations that applies, if not to all mankind, at least to a much larger group than just kith and kin (4). Market development calls for a new moral code that recognizes strong obligations extending beyond kith and kin. The transition from specific to generalized morality has attracted considerable attention. Using three behavioral experiments administered in 15 diverse populations across the world, Henrich et al. (38) show that market integration (measured as the percentage of purchased calories) positively covaries with fairness in exchange among unrelated groups. This suggests that concepts of fairness adjust to modes of exchange.

By deterring opportunistic behavior in anonymous transactions, generalized morality enables the emergence of generalized trust. Using data from bribery experiments, Barr and Serra (39) show that they could predict who among participating undergraduates would act corruptly with reference to the level of corruption in their home country. These results suggest that norms of fairness and malfeasance vary systematically across human societies and that individuals behave in a less trustworthy manner in anonymous exchange in societies where malfeasance is more widespread. Societies that develop a generalized morality should thus have a comparative advantage in market development, hence facilitating specialization and growth. Evidence to this effect is given by Algan and Cahuc (40), who provide empirical evidence of a systematic correlation between trust and growth by using the inherited trust of descendents of US immigrants as a time-varying measure of inherited trust in their country of origin.

Although there is evidence of a correlation between social norms and market development, the question arises of how generalized morality gets instilled in a population immersed in the ethics of gift exchange. Historically, world religions seem to have repeatedly played this role by broadening emotional attachment to kith and kin to a much larger group.^¶

By building bridges between strangers, world religions seem to have enabled the emergence of generalized morality among groups of converts not related by blood, and possibly living far apart. During the medieval period, the Catholic Church did play this role quite forcefully. It has been pointed out, for instance, that medieval universities almost exclusively taught Roman Law; as a result, the promarket and procontract legal principles that pervade Roman Law progressively penetrated Europe.^‖ It has been illustrated that similar results can be achieved by other world religions: in one study by Grief (41), Judaism is the market-friendly religion; in another study by Grief (42), it is Catholicism; and in a study by Ensminger (43), it is Islam. Geertz et al. (44) provide another example in which the market-supporting role is not fulfilled by whole community of believers but only by specific Muslim brotherhoods. Using experimental data from diverse populations, Henrich et al (38) find contemporary evidence that participation in a world religion is associated with fairness in ephemeral exchange. They conclude that “modern prosociality [generalized morality] is not solely the product of an innate psychology, but also reflects norms and institutions that have emerged over the course of human history” (38).

In more recent times, generalized morality has received the support of nonreligious philosophies, such as humanism, enlightenment philosophies, liberalism, or communism. All these recognize that moral obligations resulting from market exchange have the same value irrespective of the person with whom the exchange took place. If anything, humanism is a better foundation for generalized morality than religion because it extends moral obligations similarly to all mankind, without distinction of faith. The spread of Napoleon's civil code to many European countries is an illustration of this universalist approach.

The framework proposed here refines the above concepts because it makes us realize that the legitimization of market transactions is not just done once and for all; it must be repeated as the boundary between self-provision and market provision keeps evolving. As a result, market exchange is often seen as favoring “loose morals.” For instance, society may have difficulties recognizing that women have a right to interact with the market directly, and not just through the authority of their husband or father. If this is found unacceptable, new norms, such as purdah, may arise that make such exchange impossible. The marketization of sexual relationships is illegal in many societies. This is one boundary that these societies do not wish to cross, preferring to confine the exchange of (and control over) sexual favors within the family.

Finally, the coexistence of the market and gift exchange domains enables individuals to escape the moral obligations they have incurred in the gift exchange realm, a point that was made most clearly by Scott (45). Hoff and Sen (46) make a similar point when they note that demands for in-kind transfers (e.g., hosting kin and providing them urban jobs) encourages people to sever links with kin members once they enter the modern sector. In response, the kin system may erect exit barriers even when these barriers reduce aggregate welfare. Azariadis and Stachurski (47) provide other examples.

Although these authors emphasize the temptation for the rich and the successful to escape their obligation to help less fortunate members of society, markets also undermine the control that heads of households have on their male and female dependents. New norms, such as purdah, may come to the fore as a way of reasserting control over women.**

A rise in bride price can similarly be seen as a way for elders to curtail the economic freedom achieved by young men through the market. Goldstein and Udry (49) provide yet another example of distortions that arise in the transition from self-provision to market, in this case, the difficulty for farmers to fallow land despite falling fertility for fear that kin will request land transfers.

From Markets to Hierarchies

As I have pointed out, the attitudes and moral outlook that are required for hierarchies to perform well are quite different from those required for market exchange, especially unsophisticated forms of market exchange that lack a strong moral compass and are rife with opportunism. We have already discussed the issues that notably arise here: shirking, corruption, pilferage, and absenteeism. The attitudes required for hierarchies are also distinct from those nurtured by a gift exchange morality. Wage employment does not provide insurance against illness or unemployment, and it is associated with large differences in standards of living.

When wage work lacks legitimacy, loyalty to the employer is weak and it is difficult to effectively resolve the agency problems that naturally arise in hierarchies effectively. As a result, large organizations do not work well: Orders are not followed, workers at all levels divert organizational time and resources to private gain, and recrimination against employers is common.

Vestiges of risk-sharing principles survive even in advanced market economies. Using standards of fairness elicited by telephone surveys, for instance, Kahneman et al. (50) report that it is acceptable for firms to raise prices or cut wages when profits are threatened but it is deemed unfair to exploit shifts in demand by raising prices or cutting wages. Similarly, Young and Burke (51) show that fairness norms make it difficult for sharecropping contracts to adjust contracts to local market conditions.

When social norms expressed in fairness principles come into conflict with market forces, lack of legitimacy increases the potential for moral hazard by reducing the guilt and shame people feel when they breach contractual obligations (e.g., shirk, pay late). This may explain why the legitimacy of wage employment often improves with the provision of health and unemployment insurance, especially if such insurance is organized through the employment contract, as has been the rule in many developed economies. Efforts to redress the worst effects of inequality also help legitimize the employer-employee relationship. Paternalistic capitalism uses some of the same ideas, albeit in a different way.

Early examples of large hierarchies illustrate the difficulty of keeping discipline: Armies were notorious for either abusing the local population or rebelling against their leader. It should be said that employers were also notorious for not paying soldiers. Aboard Renaissance merchant ships, discipline was obtained through extremely strict rules and harsh punishments. The first corporate entities, the East India Company and the like, were famous for imperfect discipline among their staff, many of whom customarily engaged in regional trade and side deals against the interests of their employer (52).

In medieval times, the Catholic Church was an early example of a successful transnational corporate entity, with secular and political interests in many countries and thousands of employees spread out over a large area. As Algan and Cahuc (40) point out, 70% of medieval university graduates went to work for the Catholic Church. For such a large organization to function and act as a coordinated entity, it was necessary for the management to be able to pass orders and receive reports. This required a common language, a role provided by Latin. Because of distances and travel conditions at the time, written orders had to be given in the form of rules of behavior that could be implemented in a predictable manner; hence, the focus on Roman law (based on general principles) as an instrument of command and control within the Church and the focus on the teaching of law in medieval European universities. Graduates who were not absorbed by the Church went to work for secular administrations, which, at the time, started to emulate the Church's command and control apparatus. What this early example brings to the fore is the role that education can play in fostering attitudes that are conducive to large hierarchical organizations.

It should be pointed out, however, that employment by the Church was not organized primarily through secular wage contracts but rather through specific contracts adjudicated by special canonical courts. Similarly, much government employment in the Middle Ages was not through wage contracts but rather through feudal contracts, privileges, and the like, at least at managerial levels. This suggests that wage contracts were not deemed sufficient to ensure loyalty and performance. Widespread reliance on wage contracts in government employment only dates back to the 18th or 19th century, depending on the country.

Schools play other roles that are important for work in large organizations. They impart cognitive skills, such as literacy and numeracy, that are essential to large organizations (e.g., use of clerical staff and accountants to keep records, pass orders, and gather information). General education is thus a form of vocational training that prepares workers to operate large organizations. However, these cognitive skills could be imparted in ways other than schools operating on a rigid fixed schedule and maintaining a strict discipline. The purpose of the features is undoubtedly to prepare pupils for the regimented work of large organizations.

Not everyone needs to be educated, however, depending on how production is organized. For instance, in Moroccan manufacturing, most production workers in the textile and garment export sectors are illiterate women (53). Given the traditional status of women in Morocco, they are likely to constitute a docile and disciplined labor force even without having been to school, another illustration of one domain (wage work) taking advantage of social norms and attitudes from another domain (self-provision).

Conclusion

It has long been recognized that the development process involves a structural transformation of the economy. I have argued in this paper that one of the characteristics of this structural transformation is a dramatic change in the way people are allocated to tasks. There is a shift from self-provision to market exchange as many goods and services previously self-provided fall under the purview of the market. This transformation process enables workers to specialize and leads to an entrepreneurial revolution, whereby self-employment in market-oriented nonfarm work rises. Allocation of workers to tasks then responds to the demand for goods and services and the supply of raw materials.

To capture fully the returns to innovation in technology and organization, a further shift is required from small to large firms and organizations. This leads to the rise of wage employment. The allocation of workers to tasks then takes place partly via the labor market. Because of information and incentive issues, there are costs associated with screening workers for management, supervision, and clerical tasks. As a result, firms economize on screening costs by offering employment contracts of indefinite duration. If production workers require firm-specific training or skills, they also may be offered permanent labor contracts. This means that part of the allocation of workers to tasks takes place within firms and organizations rather than through the market.

Changes in allocation mechanisms require changes in norms and practices. These changes need not happen in all sectors and regions of the economy at the same time. There is coexistence, or even partial overlap, among the domains of different labor allocation mechanisms. This coexistence creates conflicts among different sets of norms that apply to different domains (e.g., family, market, work place). For economies to develop, people must learn to combine these different sets of attitudes and behaviors in relative harmony. I have provided examples in which the application of norms from one domain to another domain results in dysfunctional outcomes (e.g., corruption, pilferage, nepotism, low quality). I have also showed that leakages from one domain to another can help or hinder market development.

At the end of the paper, I discussed what processes have historically been instrumental in shaping social norms and attitudes in ways that are more conducive to economic development. The early role of religion in market development was noted whenever it helped the emergence of generalized morality, that is, of strong moral obligations toward people who are neither kith nor kin. Religion, however, is not necessary, and may even turn into an impediment, because it can be used to divide people or to resist change and protect vested interests. Humanism provides a stronger foundation to generalized morality.

With respect to wage employment and the rise of large organizations, I noted the role that education has played and continues to play. Not only do schools impart vocational skills that are essential to the day-to-day operation of large organizations, they also convey noncognitive skills that prepare pupils for the routine and discipline of wage employment. Schools can even be used to foster humanism and universal human values. Given this, it is probably no surprise that a strong correlation has been found between schools and economic growth (54, 55).

I have not discussed the fact that different systems of norms to allocate people to tasks have different implications regarding inequality. If tasks are allocated primarily within the household, power relations revert around marriage and roles socially assigned to specific groups (e.g., women, unmarried men, lower castes), issues that I have touched on in some of my work (17–20). If tasks are allocated via the market through self-employment, norms regarding excusable contractual default affect the distribution of wealth, a point I have discussed elsewhere (8). Finally, if tasks are assigned within large firms and organizations, economic power depends on labor relations. Market development may undermine some forms of control while creating or reinforcing others. There is a large Marxist, and non-Marxist, literature on these issues (45, 56) to which space constraints make it impossible to do justice.