What are formal and informal institutions?

Searches were conducted on the JSTOR database for the term ‘formal institution’.Footnote ¹ There are a few appearances in the 19th century. Many are unclear about the meaning of ‘formal’. Sometimes ‘institution’ appeared as a verb. Before 1950 the term ‘formal institution’ (or its plural) is infrequent in articles in English on the JSTOR database, with only 81 journal articles in which it appeared. ‘Informal institution’ (or its plural) appears in only five articles.

Moving on, from 1950 to 1999 inclusive, the term ‘formal institution’ appeared in 2,446 articles and ‘informal institution’ appeared in 570 articles. Usage of both terms increased dramatically thereafter. So far in the present century, the term ‘formal institution’ has appeared in 5,558 articles and ‘informal institution’ has appeared in 3,057 articles. Both terms appear in 1,693 articles. Academic discussion of ‘formal’ and ‘informal’ institutions is largely a 21st century phenomenon. But many works fail to propose a clear definition of the formal/informal distinction.

From about 1989, North made frequent use of the terms ‘formal’ and ‘informal’ in an institutional context. His work was a major stimulus behind the 21st century explosion in the usage of these terms. But his terminology raises questions as well as providing some answers. North wrote mostly of ‘informal constraints’ and sometimes of ‘informal norms’. He typically deployed the word ‘formal’ in ‘formal rules’ and ‘formal institutions’.Footnote ² Why did he associate the word ‘formal’ with rules and ‘informal’ with constraints? North (Reference North1990: 3–4) indicated that constraints are the broader category. He did not claim that constraints and rules are mutually exclusive. This might suggest that many constraints are not rules. On the contrary, constraints typically serve as rules and many rules are constraints. ‘No trespassing’ is both a rule and a constraint. Constraints can be enabling. Rules that constrain drivers to keep on one side of the road help drivers to reach their destinations safely. North's reasons for associating ‘formal’ with ‘rules’ and ‘informal’ with ‘constraints’ are unclear (Voigt and Kiwit, Reference Voigt, Kiwit and Giersch1998: 87 n.). His distinction between ‘formal’ and ‘informal’ institutions has nothing fundamentally to do with any difference between ‘rules’ and ‘constraints’. It would have been better if he had not added further terminological complications.

Once North (Reference North1990: 46) saw the ‘difference between informal and formal constraints’ as ‘one of degree’. He envisaged ‘a continuum from taboos, customs, and traditions at one end to written constitutions at the other … from unwritten traditions and customs to written laws’. As far as I am aware, North did not continue with this ‘continuum’ idea. While we need not propose a continuum, definitional demarcations are rarely sharp. Fuzziness and boundary cases exist.

North (Reference North1991: 97; Reference North1994: 360; Reference North, Drobak and Nye1997: 6) gave examples of ‘informal constraints’ as sanctions, taboos, customs, traditions, conventions, codes of conduct, and norms of behaviour. He saw ‘formal rules’ as ‘constitutions, laws, property rights’.Footnote ³ North (Reference North1992: 4) similarly wrote that institutions are ‘composed of formal rules (statute law, common law, regulations)’ alongside ‘informal constraints conventions, norms of behavior’. Hence, for North, the term ‘formal’ is typically linked with legal matters. He never defined ‘formal’ pre-eminently as ‘written’, and he frequently linked ‘formal’ with legal enforcement.

In correspondence with the present author, North (Reference North2002) clarified his position. He wrote: ‘Formal rules are enforced by courts and things like that. Informal norms are enforced usually by your peers or others who impose costs on you if you do not live up to them.’ Here he focused on the means of enforcement. The key criterion to demarcate the formal from the informal is whether, respectively, there is enforcement by a legal court (or something like it), or enforcement without such an institution. If North had said this more clearly at the outset it would have saved some confusion.

For the philosopher and institutional analyst Chrys Mantzavinos (Reference Mantzavinos2001: 83–4), as for North and others, formal institutions are enforced by law, whereas informal institutions ‘do not need for their enforcement the state’. This is helpful, but some clarification is required. Some laws, such as driving on the same side of the road, are largely self-enforcing, and require state enforcement in exceptional cases only. Many laws are followed because people accept their moral validity. Consequently, formal rules should be defined in terms of potential, and not necessarily actual, enforcement by the state. Crucially, formal rules are potentially enforceable by the state, including some that are often self-enforcing.

There are frequent cases where laws are enforced by institutions that are not part of the state legal system. Unestablished religious institutions may enforce marriage laws. Private companies may enforce employment laws. And so on. By reasonable definition, if they are laws and the state is potentially involved as an enforcer of last resort, these rules remain formal. Otherwise they are informal.

At the start, North (Reference North1989a: 239; Reference North1993: 20) emphasised that ‘formal rules … must be complemented by informal constraints … that supplement them and reduce enforcement costs’. He added: ‘If the formal rules and informal constraints are inconsistent with each other the resulting tension is going to induce political instability.’ Consequently, he wrote of ‘complementary’ versus ‘competing’ formal–informal relationships.

The political scientists Gretchen Helmke and Steven Levitsky (Reference Helmke and Levitsky2004) also stressed the importance of ‘informal institutions’. They classified forms of interaction between the formal and informal. Working in two dimensions, they (p. 728) considered ‘the degree to which formal and informal institutional outcomes converge’. They asked whether following informal rules produces a similar or different result from that following a strict adherence to formal rules. When the outcomes ‘are not substantively different, formal and informal institutions converge’. Otherwise, divergence occurs. ‘The second dimension is the effectiveness of the relevant formal institutions, that is, the extent to which rules and procedures that exist on paper are enforced and complied with in practice. … Where formal rules and procedures are ineffective, actors believe the probability of enforcement (and hence the expected cost of violation) will be low.’ Hence they created a taxonomy of formal–informal interactions. With convergent outcomes and ‘effective formal institutions’ the formal–informal relationship is ‘complementary’. With convergent outcomes and ‘ineffective formal institutions’ the relationship is ‘substitutive’. With divergent outcomes and ‘effective formal institutions’ the relationship is ‘accommodating’. With divergent outcomes and ‘ineffective formal institutions’ the relationship is ‘competing’. Hence, they usefully distinguish between complementary, accommodating, competing, and substitutive relations.

Helmke and Levitsky (Reference Helmke and Levitsky2004: 727, emphasis in original) wrote:

This is a thoughtful attempt to define informal and formal institutions. But it has some defects. First, taxonomic definitions should be parsimonious. It would be advisable to remove the words ‘usually unwritten’. Many informal rules are written down, such as in books on the rules of language, etiquette, and table manners. The words ‘usually unwritten’ are redundant and potentially misleading. Second, ‘officially sanctioned channels’ is too vague. Language is widely regarded as an institution (e.g., Searle, Reference Searle1995, Reference Searle2005). Arguably, it is best regarded as informal too. But state agencies often attempt to interfere with languages in various ways. In all countries, official policies favour one or more languages over others, in their administrative and education systems. In some countries (such as France) official attempts have been made to exclude some words, promote others, and to regularise spellings. The imperial Chinese state imposed the Mandarin language and a single writing system on a linguistically diverse population (Scheidel, Reference Scheidel2019: 312). Do these official interventions make all these rules formal? If so, then ‘formal’ rules become much more widespread. ‘Officially sanctioned’ is best removed from the definition of ‘formal’. With these modifications, the Helmke and Levitsky definitions become closer to those inspired by North, where ‘formal’ means ‘legal’ and ‘informal’ means ‘non-legal’.Footnote ⁴

As English dictionaries reveal, the word ‘formal’ has many meanings. These include ‘ceremonial’, ‘official’, ‘proper’, and so on, as well as ‘lawful’. It can mean ‘explicit’, but this does not necessarily imply ‘written’. With writers on institutional research, particularly because of the impact of North, the association of ‘formal’ with law is stronger. But this legal issue is often muddied by the addition of other terms.

Given the many meanings of ‘formal’, we cannot in this case refer to a single, prominent usage for the purpose of definition. Features like ‘written’ or ‘explicit’ have no obvious precedence over ‘legal’ as the defined meaning of ‘formal’. North did more than anyone to establish the formal/informal distinction in modern institutional analysis. He largely associated ‘formal’ with ‘legal’. Given his influence, any plea to change these meanings is unlikely to be successful.

It would be best to follow the example of North and others, and to link the ‘formal’ with the legal, and not necessarily with the written. Consequently, informal rules are rules that are not laws. Another option would be to define formal rules as both ‘legal’ and ‘written’. But the ‘legal’ part of the definition would be doing most of the work of demarcation. A great number of written rules are not laws, and they would remain ‘informal’. Some early legal systems, in which laws were unwritten, would be pushed out, but that would do no great harm in the modern context. Adding ‘written’ to the definition of ‘formal’ unnecessarily complicates the picture, without adding any demarcatory value.

The North-inspired definition of the formal/informal distinction does not rule out important subdivisions within each category. For example, some informal rules are against the law, while others are not. Informal rules differ greatly in their modes enforcement, in their strictness, and whether they are written or unwritten. Some formal rules are often ignored. Others are followed with diligence. Some formal rules have high moral salience. Others are matters of mere convenience or convention. Countless subdivisions are possible, and some of these may be useful.

Some important implications

Henceforth in this paper we associate ‘formal’ with law and potential legal enforcement. Informal (i.e., non-legal) rules are not laws, and are enforced apart from the legal system (somehow defined). Non-legal is not the same as illegal. Informal rules, being non-legal, may or may not comply with the law. Hence we need to distinguish between at least two kinds of informal institution or rule – between those that comply with the law and those that are illegal.

The prevalence of illegal informal rules is likely to be greater if the (formal) system of law is corrupt, or otherwise less effective (Keefer and Shirley, Reference Keefer, Shirley and Ménard2000). Sometimes, informal rules substitute for ineffective formal rules. It is widely recognised that problems of poor governance and corruption are especially severe in less-developed countries. Businesses may evade registration, taxation, safety standards, employment laws, and so on. This is sometimes referred to as the ‘shadow economy’ or the ‘informal economy’ (Gërxhani, Reference Gërxhani2004; OECD, 2002; Schneider and Enste, Reference Schneider and Enste2000).

Klarita Gërxhani and Stanisław Cichocki (Reference Gërxhani and Cichocki2023: 657) argued that ‘a proper functioning of the economy needs the formal and informal institutions to be complementary and well enforced’. As Gërxhani and Cichocki accepted, the idea that formal and informal institutions should be complementary and ‘well enforced’ would be problematic and questionable if many informal rules were harmful or illegal. If they were, the strengthening and enforcement of informal rules could act against or undermine the power of formal (legal) rules.

Earlier, Gërxhani (Reference Gërxhani2004: 294) reported that ‘there is a general agreement that, in the long run, the informal sector should be reduced in size or formalized’. This suggests a reduction of illegality in the ‘informal sector’. Since then there may have been a shift of opinion among development researchers. In their introduction to a forum on African ‘informal economies’, Kate Meagher and Ilda Lindell (Reference Meagher and Lindell2013: 58) reported that ‘informal economic activity has burgeoned across the developing world, particularly in Africa’. Efforts to limit the informal sector have largely failed.Footnote ⁵ In response, policymaking efforts have shifted ‘away from an emphasis on eliminating or absorbing the informal economy toward policy discourses of collaborative interaction, expressed in such terms as “hybrid governance,” “coproduction,” and “formal-informal linkages”’. The focus has become ‘the politics of inclusion’ whereby ‘informal’ (and often illegal) activities are seen in part as an understandable response to state failure. With the maxim of ‘inclusion’, the illegal is tolerated under cloak of the ‘informal’. This is important not simply because of moral qualms about illegality. It is also crucial for non-normative reasons. Since Max Weber, the perceived legitimacy of legal rules has been seen as crucial for their support. While in small-scale customary groups, legitimation may derive from local arrangements (Murtazashvili and Murtazashvili, Reference Murtazashvili and Murtazashvili2016); a large ‘informal’ sector may lack the mores and ideology to cultivate appropriate perceptions of legitimacy. In which case, toleration of an expanding informal sector (harbouring illegal activity) may exacerbate the problem of (partial) state failure.

Several authors have noted the importance of state capacity, including fiscal capacity, alongside appropriate constraints on the executive (Bardhan, Reference Bardhan2016; Besley and Persson, Reference Besley and Persson2011; Ogilvie, Reference Ogilvie2022; Ricciuti et al. Reference Ricciuti, Savoia and Sen2019; Savoia and Sen, Reference Savoia and Sen2023). Tolerating the illegal as part of the ‘informal’ may run against the enhancement of those elements of state capacity required to help promote economic development.

It is far beyond the scope of the present essay to assess the possibility of Africa or elsewhere finding a new path to economic development that dispenses with the need to build an effective system of state administration, involving higher standards of governance, politico-legal checks and balances, and the comprehensive rule of law (Waldron, Reference Waldron, Zalta and Nodelman2023). If that were achieved, it would be unprecedented. The point here is that narratives dominated by ‘informality’ and ‘inclusion’ may sidestep key questions of precedence and feasibility. How can a state function effectively while it accepts widespread illegality? Such toleration would heighten systematic corruption. The question is raised here, but it is not answered. We note simply that the loose rhetoric of ‘informality’ may push this issue into the background. It needs to be addressed more squarely.

The importance of a good definition of formal institutions

We turn to work by economists Claudia R. Williamson and Carrie B. Kerekes on formal and informal institutions, where they assess the differential impacts of these two main types of rule system (Williamson, Reference Williamson2009; Williamson and Kerekes, Reference Williamson and Kerekes2011). Williamson and Kerekes (Reference Williamson and Kerekes2011: 546) wrote: ‘Informal institutions are those rules that shape human behavior but are outside of government and are not part of a written legal framework.’ Their statements conform to the Northian view that informal rules are, by definition, non-legal in character. Williamson (Reference Williamson2009: 372) wrote: ‘Formal institutions are defined on political constraints on government behaviour enforced by legal institutions. Formal rules encompass constitutional constraints, statutory rules, and other political constraints.’ Williamson and Kerekes (Reference Williamson and Kerekes2011: 538) wrote similarly: ‘We define formal institutions as political constraints on government behavior and informal institutions as private constraints, such as norms or customs.’ These statements narrow considerably the meaning of ‘formal institutions’ to laws that constrain government.

Their modification of the formal/informal dichotomy had a major impact on the empirical design and results of their two studies. Williamson and Kerekes drew in part from the work of Edward L. Glaeser et al. (Reference Glaeser, La Porta and Lopez-De-Silanes2004) who also focused on political institutions that constrain governments. But Glaeser et al. (Reference Glaeser, La Porta and Lopez-De-Silanes2004) did not distinguish between ‘formal’ and ‘informal’ institutions. Their concern instead was to assess whether higher levels of education lead to democracy and other constraints on government, or whether there is stronger causation in the opposite direction. This is an important issue, but it is much narrower in scope than trying to assess the general differential impact of formal versus informal institutions. Williamson (Reference Williamson2009: 374) and Williamson and Kerekes (Reference Williamson and Kerekes2011: 545) took four types of formally constituted constraints on government from Glaeser et al. (Reference Glaeser, La Porta and Lopez-De-Silanes2004), namely ‘plurality, proportional representation, judicial independence, and constitutional review’. Williamson and Kerekes used these criteria to assess the strength of formal institutions. The strengths of informal institutions in each country were assessed in terms of cultural traits, including levels of trust. Accordingly, they developed indicative measures of the strengths of formal and informal institutions.

Williamson (Reference Williamson2009) compared her measure of the strength of formal institutions with GDP per capita (purchasing power parity) for the year 2000. The dependent variable in Williamson and Kerekes (Reference Williamson and Kerekes2011) is different – it concerns property rights. For reasons of brevity, I concentrate here on the GDP per capita analysis. Figure 1 displays the Williamson and Kerekes measure of the strength of formal institutions, plotted against levels of GDP per capita, as explored by Williamson (Reference Williamson2009) in 38 countries.

Figure 1.

GDP per capita and the strength of ‘formal institutions’.

Data source: Williamson (Reference Williamson2009).

Six countries are within the ellipse at the top left of Figure 1, indicating high levels of GDP per capita and low estimated levels of the strength of formal institutions. These countries are the Netherlands plus the Nordic nations of Denmark, Finland, Iceland, Norway, and Sweden. Nine countries are located within the long, dashed oblong, with rounded corners, at the bottom of the figure. They all have low levels of development but vary enormously in the reported strength of their formal institutions. They are, moving from the left, Nigeria, Egypt, Jordan, Indonesia, and Bangladesh, followed by four clustered far to the right, namely Pakistan, Philippines, Uganda, and Zimbabwe. The figure shows negligible correlation between the adopted measure of formal institutional strength and GDP per capita. By implication, for economic development, the explanatory burden shifts instead onto the strength of informal institutions. As Williamson (Reference Williamson2009) concluded: ‘informal institutions rule’.

But questions can be raised about a methodology that deems formal institutions in the Nordic countries and the Netherlands to be very weak, and roughly equivalent in strength to those in (say) Nigeria. The idea that Pakistan, Philippines, Uganda, and Zimbabwe have among the highest levels of formal institutional strength should also be queried. These challengeable classifications and conclusions stem partly from the narrow focus on government constraints as defining the set of formal institutions. The choices of definitions and datasets matter.

Figure 2 displays a different measure of the strength of formal institutions in the same 38 countries, plotted against the same levels of GDP per capita. If formal institutions are defined more broadly as involving laws, then one possible measure of their strength could be some index of the rule of law. There are several datasets on this, a few of which go back to the year 2000, to which Williamson's (Reference Williamson2009) GDP data apply. The GlobalEconomy.com index for the rule of law goes back to 2000 and, in its publisher's words, ‘captures perceptions of the extent to which agents have confidence in and abide by the rules of society, and in particular the quality of contract enforcement, property rights, the police, and the courts, as well as the likelihood of crime and violence’. By this measure, all five Nordic countries plus the Netherlands were in the top 10 in 2000.

Figure 2.

GDP per capita and the rule of law.

Data sources: Williamson (Reference Williamson2009) and https://www.theglobaleconomy.com/rankings/wb_ruleoflaw/. Retrieved 6 September 2023.

Twelve countries are located within the box at the top right of Figure 2, indicating high levels of GDP per capita and the strong rule of law. They all had a GDP per capita level of at least $25,000 and a rule of law score exceeding 1.5. They include the Nordic nations and the Netherlands, as in the preceding figure, with the addition of Australia, Austria, Canada, Ireland, the UK, and the US.

Nine countries are located within the dashed oblong at the bottom of Figure 2. They vary less in terms of the rule of law index: they are all at relatively low levels. They are, moving from left to right, Zimbabwe, Nigeria, Pakistan, Bangladesh, Uganda, Egypt, Indonesia, Philippines, and Jordan. Overall, Figure 2 shows a strong correlation between the rule of law measure and GDP per capita. While Williamson found little if any correlation using her measure of ‘formal institutions’ there is a strong positive correlation between her GDP per capita data for 2000 and the GlobalEconomy.com rule of law index for the same year.

Although measures of the rule of law by other agencies are different, the broad picture is often similar. At the time of the publication of the Williamson (Reference Williamson2009) paper, the World Justice Project had not fully developed its index of the rule of law. Since then, they have assessed the rule of law over eight dimensions, including ‘constraints on government powers’. The other dimensions include fundamental rights, the absence of corruption, open government, regulatory enforcement, and civil justice. Their data exclude Iceland, but in 2015 the remaining four Nordic countries plus the Netherlands were their top five places in their overall rule of law index. The same five countries were in the 2015 component ranking for ‘constraints on government powers’. This picture is very different from that reported in the Williamson (Reference Williamson2009) study.Footnote ⁶

By prominent studies of formal (legal) strength, none of the Nordic countries have ‘weak’ formal institutions. The same is true for the Netherlands. Even if we concentrate on ‘constraints on government powers’, reports by the World Justice Project broadly concur. The conclusion that ‘informal institutions rule’ is challenged. But informal institutions are still very important. The implementation of formal (legal) rules always depends to a degree on supportive informal norms and rules (Helmke and Levitsky, Reference Helmke and Levitsky2004). Formal institutions always rely on informal institutions to be effective. Informal institutions always matter. But formal (legal) rules matter as well.Footnote ⁷

North (Reference North1994: 367) opined that ‘long-run economic growth entails the development of the rule of law’. Much evidence supports his claim (Dam, Reference Dam2006; Haggard et al., Reference Haggard, MacIntyre and Tiede2008; Koyama and Rubin, Reference Koyama and Rubin2022: 39 ff.). While informal (customary) institutions are also important, the rule of law signals the dominance and effectiveness of formal (legal) rules.

Worthwhile indicators of the rule of law concern de facto operation and enforcement, as well as de jure legislation. Cases of ineffective formal legislation are well known (Aldashev et al., Reference Aldashev, Chaara, Platteau and Wahhaj2012). For example, discrimination by caste in India is illegal, yet it persists (Mosse, Reference Mosse2018). Many Latin American countries adopted a constitution similar to that in the US, but the enforcement of many rules has been ineffective, and consequently outcomes have been quite different (North, Reference North1990: 101). For formal rules to work, they must cater for diverse interest groups. Supportive informal rules (or cultural norms) and effective monitoring or persuasion by a critical mass of law-abiding actors are required. As Ian Ayres and John Braithwaite (Reference Ayres and Braithwaite1992) put it somewhat paradoxically, state regulation of business can operate well when there is ‘enforced self-regulation’.

Similar exercises could be tried using measures of state capacity, instead of the rule of law. But all measures have problems, and state capacity is not short of them. One major difficulty is that military capacity is often a major component of state capacity, and that is less relevant for our purposes (Ogilvie, Reference Ogilvie2022). The main point here is not to establish the best measure, but to show that different measures and definitions can lead to dramatically different results.

Problems with the meanings of law and culture

We mention culture and, in fear of raising too many difficulties, move quickly on to law. Famously, culture has many meanings (Kroeber and Kluckhohn, Reference Kroeber and Kluckhohn1952). It often refers to a set of prevalent rules, norms, values, and beliefs. We cannot go into all the nuances here. Instead, we briefly raise the issue of how the concept of culture might relate to institutional terms. A key problem is to distinguish the concept of culture from that of an institution, where both refer to rules and norms. One option is to use the word culture to describe a set of traits found in several institutions in a society (Hodgson, Reference Hodgson2001: 298–300). These might include cultural traits such as individualism, hierarchy, and masculinity, as defined by Geert Hofstede (Reference Hofstede1984) or others. Hence one might claim, for example, that the US has an individualistic culture, and this trait is reflected in some of its institutions. This might be contrasted with more group-oriented cultures in (say) East Asia. With this option, the distinction between formal and informal rules or institutions can exist alongside the concept of culture, and each term has a distinct meaning.

As another option, the institutional economist Svetozar Pejovich (Reference Pejovich1999: 166) proposed that informal institutions are ‘traditions, customs, moral values, religious beliefs, and all other norms of behavior that have passed the test of time … informal institutions are the part of a community's heritage that we call culture’. Alberto Alesina and Paola Giuliano (Reference Alesina and Giuliano2015: 902) followed this and made the further stipulation that the term institution is confined to ‘formal institutions (formal legal systems, formal regulation)’ and culture is applied to ‘values and beliefs’ and ‘informal rules’. Here the concept of informal institutions is replaced by culture. Another alternative would be to retain the ‘informal institutions’ wording but measure it using cultural variables. More discussion is needed on the options.

We now turn to the meaning of the concept of law. This is debated, including by legal theorists. At one end of the debate, law is associated with the state. Hence, for William Blackstone ([1765–69] Reference Blackstone1893: 44) in his famous Commentaries on the Laws of England, law is ‘a rule of civil conduct prescribed by the supreme power of the state’. A very different option is to associate law with custom in general. For example, James C. Carter (Reference Carter1907: 173), a president of the American Bar Association, wrote ‘law … is custom, and like custom, self-existing and irrepealable’. But if law consisted simply of custom, then any distinction between formal and informal rules based on whether they are enforceable by law would fail.

It is sometimes useful to reduce generality and concentrate on the modern period. We are concerned with law as it relates to populous, highly organised, and complex societies. Once we consider the problems of enforcement in complex systems with many agents, and the motivational reasons why individuals might obey the law, then arguably some public body like the state is required to ensure enforcement. As Ronald Coase (Reference Coase1988: 10) put it, with ‘a vast number of people with very different interests … the establishment and administration of a private legal system would be very difficult’. Hence trading agents depend ‘on the legal system of the State’. We broadly construe the state as a realm of public ordering based on institutionalised authority. Such a state establishes a monopoly of force within a territory. A state's commitment to operate within legal constraints can help to enhance its own legitimacy as a wielder of power (Weber [1921] Reference Weber1968: vol. 1: 212 ff.). Although law may be built upon custom and rely on informal rules, it also relies on the powers of state enforcement.

Although custom is a source of law, this does not mean that ‘customary law’ is law in the above sense. Furthermore, common law (necessarily involving an institutionalised judiciary and legislature) is much more than customary law (Hasnas, Reference Hasnas2005). Furthermore, customary mechanisms (although necessary to sustain small-group, face-to-face interactions and mutual trustFootnote ⁸) are insufficient to explain adherence to large-scale, complex systems of law (Engel, Reference Engel2008; Tyler, Reference Tyler2006).

In history, legal systems often dealt with cases where customs were inadequate or in conflict. Accordingly, the essence of law resides in its transcendence of custom, particularly when breaches or conflicts of customary conventions arise (Diamond, Reference Diamond1935; Hodgson, Reference Hodgson2015: ch. 3; Redfield, Reference Redfield1957; Seagle, Reference Seagle1941).

The picture is further complicated by the doctrine of ‘legal pluralism’ that upholds that societies contain groups with rival rule systems and normative orders (Griffiths, Reference Griffiths1986; Merry, Reference Merry1988). Examples include the imperialist imposition of European legal systems upon substrata of still functioning indigenous legal traditions in the colonial era. One problem is that theorists of legal pluralism regard any normative order or system of social control as ‘legal’ in character (Tamanaha, Reference Tamanaha1993). This broad conception of law goes beyond the notion that modern law is legitimated and (potentially) enforced by the state. It would include a host of customary and other stipulations. If law is broadened to include customs, then we may be inclined to overlook or downplay what is distinctive about modern legal systems.

All societies contain plural systems of normative rules, but this does not necessarily amount to plural systems of law. The more dramatic cases of conflicting systems of normative rules, in Africa, India, and elsewhere, are marks of tribal or clan-based societies with high degrees of religious, cultural, and ethnic fragmentation, some of which are arguably impediments to their institutional, legal, and economic development (Alesina et al., Reference Alesina, Baquir and Easterly1999; Dam, Reference Dam2006; Easterly and Levine, Reference Easterly and Levine1997).

In his critique of legal pluralism, the philosopher and legal theorist Klaus Günther (Reference Günther and Fabra-Zamora2020) noted that with a plurality of legal and normative orders, some kind of meta-law, which enables participants to deal with norm conflicts and collisions, either exists or is required for stability and coherence. National and international processes of development crucially involve the provision of encompassing systems of law to deal with disputes. The widening and development of markets and capitalism have promoted this encompassing process (Coase, Reference Coase1988: 10). In medieval England there was a variety of types of courts, relating to the church, feudal manors, guilds, and trade associations (Berman, Reference Berman1983: 333–56; Milgrom et al., Reference Milgrom, North and Weingast1990; Rogers, Reference Rogers1995). Gradually these were drawn into the clutches of the state (Hasnas, Reference Hasnas2005; Juenger, Reference Juenger2000; Pollock and Maitland, Reference Pollock and Maitland1898), partly because of problems of plural, contested authority in circumstances of growing social complexity. Parties found it easier to submit to a single, powerful, legal apparatus. Historically, capitalism has reduced the multiplicity of legal authorities, in favour of the state. Avner Greif and Guido Tabellini (Reference Greif and Tabellini2010) argued persuasively that the formation of a state legal system with associated moral norms was crucial in Europe's economic development.

There are also plentiful boundary cases, where what is and what is not a legal system is contestable or unclear. But we should not abandon the task of definition for this reason. The Celts did not assemble under national state institutions. But they had what some regard as systems of law, largely of a customary nature, and largely unwritten before their conversion to Christianity (Kelly, Reference Kelly1988). They may have been close to the definitional boundary. Sometimes religious institutions dispense and enforce laws, as in parts of medieval Europe and in theocratic states (Fox and Sandler, Reference Fox and Sandler2005). Boundary cases are inevitable in evolving systems with highly varied components.

In modern times the law is associated with the state, although it may derive originally from other sources. A key point is the operation of some strong public authority that is legitimated by religion, by tradition, by status, or by a measure of democracy. This observation can help to provide us with a definition of law that surpasses mere custom. We may thus define (modern) legal systems as necessarily involving organised legislative, judicial, and enforcement institutions, which are empowered by state authority. A law is a rule sustained or enforced by a legal system, as defined.Footnote ⁹

Formal rules are rules that are actually or potentially enforced by legal systems. This definition of law also helps us to distinguish between formal and informal institutions. But many institutions are driven and constrained by both formal rules (laws) and informal rules (which, by definition, are not laws). This applies, for example, to business firms, that operate according to legal rules. Businesses also have an internal culture, consisting of rules that are sometimes not even written down. No institution lacks informal elements. Hence the term ‘formal institution’ needs further thought. It might denote states, legal systems, modern business firms, and so on, where laws make up the dominant elements in its rule structures. Informal institutions might include phenomena like language, etiquette, and many customs or rituals. Most of their rules are not constituted by lawmaking institutions.

Some laws have a general and others a special or local application (Cloe and Marcus, Reference Cloe and Marcus1936). Normally, general law takes priority over special or local law, but that does not make special laws unimportant. Much of general corporate law, for example, resulted from the accumulation of acts of parliament that each set up specific corporations. Special laws and experiences helped to create general laws. In turn, general laws can nurture specific or private rules. Corporations today comply with general corporate legislation, but each contains special rules that are legally enforceable but specific to the corporation in question. Contract law similarly involves general rules that provide room for private rules. A contractual agreement involves some legally enforceable rules that are specific to the contract, such as when and how an item or service is to be delivered. Because of their legal enforceability, these may be regarded as formal rules.

An overemphasis on the formal and legal aspects can overlook the reliance of legal systems themselves on informal rules and norms. As Émile Durkheim ([1893] Reference Durkheim1984: 158) argued: ‘in a contract not everything is contractual’. Whenever a contract exists there are rules and norms that are not necessarily codified in law. The parties to the agreement are forced to rely on institutional rules and standard patterns of behaviour, which cannot for reasons of practicality and complexity be fully established as laws. Some markets will always be missing (Magill and Quinzii, Reference Magill and Quinzii1996). Legal systems are invariably incomplete, and they give scope for custom and culture to do their work (Hodgson, Reference Hodgson2015, Reference Hodgson2020). We always depend on informal rules as well.

On the interaction of formal and informal institutions

We now turn to the question of how formal and informal institutions may interact and evolve. It is widely upheld that the cultural norms and beliefs that make up informal institutions are generally more resistant to change. North (Reference North1990: 45; Reference North1994: 364, 366) noted the ‘tenacious survival ability’ of cultural traits and explained:

The Nobel Laureate Oliver Williamson (Reference Williamson2000: 596–7) considered four levels of social analysis. He saw ‘Level 1’ as having the most ‘socially embedded’ features, including ‘norms, customs, mores, traditions, etc. … Religion plays a large role at this level. … Institutions at this level change very slowly – on the order of centuries or millennia.’ Reflecting on possible reasons for slow change at this level, he conjectured ‘that many of these informal institutions have mainly spontaneous origins – which is to say that deliberative choice of a calculative kind is minimally implicated. Given these evolutionary origins, they … display a great deal of inertia … the resulting institutions have a lasting grip on the way a society conducts itself.’ At other levels he included formal institutions, including the judiciary, polity, and other bureaucracies. He argued that these can change at faster rates than Level 1 institutions.

Similarly, the economist Gérard Roland (Reference Roland2004: 109) drew a distinction between ‘slow-moving’ and ‘fast-moving’ institutions. For him, a ‘prime example of a slow-moving institution is “culture,” including values, beliefs, and social norms. Fast-moving institutions do not necessarily change often but can change more quickly – sometimes nearly overnight. Political institutions can typically be viewed as fast-moving institutions.’ Roland examined possible interactions between these two kinds of institutions.

In line with North, Williamson, and Roland, there are several impressive case studies of the persistence of cultural traits (or informal institutions) and their lasting effects. Emmanuel Todd (Reference Todd1985, Reference Todd1987) showed that different, largely unwritten, family structures and property inheritance patterns persist in different parts of the world (see also Gutman and Voigt, Reference Gutman and Voigt2022). David H. Fischer (Reference Fischer1989) argued that US regional differences in family structure, gender relations, community attitudes, and propensities to violence emanate from different phases of immigration from the seventeenth to the nineteenth centuries, from contrasting cultures in different regions of Britain. Sonya Salamon and Jack T. Kirby (Reference Salamon and Kirby1992) showed that different cultural patterns of farm management, land tenure, and inheritance survived among Illinois farming communities, depending on whether they were of German or British Protestant descent. Richard E. Nisbett and Dov Cohen (Reference Nisbett and Cohen1996) provided experimental evidence of the survival of a potentially violent ‘culture of honor’ in the US South. They argued that it was derived from Scots-Irish immigrants. There are other case studies of long-term cultural persistence (Koyama and Rubin, Reference Koyama and Rubin2022: 79–87).

But sometimes, cultural norms change more rapidly. Helmke and Levitsky (Reference Helmke and Levitsky2004: 732) gave the example of the disappearance in China of the practice of binding the feet of women. It lasted a thousand years, but within a generation it disappeared, due to strong social campaigns against it (Mackie, Reference Mackie1996). The development economist Gani Aldashev et al. (Reference Aldashev, Chaara, Platteau and Wahhaj2012: 797) reported studies, in developing as well as developed economies, that show that ‘custom is far from being static and continuously evolves under the pressure of a changing environment’. Cultural shifts are often pressured by economic or legislative changes. In the US, female participation in the employed workforce rose from 33% in 1950 to 60% in 2000.Footnote ¹⁰ Rising female employment shifted gender-based norms and prompted equal gender rights legislation from the 1960s onwards (Meagher and Shu, Reference Meagher and Shu2019).

Rules and attitudes concerning sexual relations are often resilient components of a social culture. Frequently, they are based on religion. Williamson (Reference Williamson2000: 597) saw religion as deeply ingrained and slow to change. Consider homosexual acts, which have been regarded as sinful by major religions, including Christianity, Islam, and Judaism. According to the survey of British Social Attitudes, in the UK in 1990, 69% of adults thought that homosexual acts were ‘always’ or ‘mostly’ wrong, compared with 19% saying they were ‘rarely wrong’ or ‘not wrong at all’. But by 2010 these figures had shifted dramatically, to 30% saying that they were ‘always’ or ‘mostly’ wrong, compared with 53% saying they were ‘rarely wrong’ or ‘not wrong at all’.Footnote ¹¹ This rapid transformation of previously enduring cultural beliefs may be partly explicable in terms of a series of legislative acts, including reductions in the age of consent for male partners in 1994 and 2000, the ban on gay people serving in the armed forces being repealed in 2000, the instigation of gay civil partnerships in 2004, and the banning of discrimination based on sexual orientation in 2007. It seems that the legislation played an important role in legitimating more tolerant attitudes. There was also an energetic movement for gay rights. In more secular societies, legislation was important in reversing previous, religiously grounded, attitudes.

As noted above, North (Reference North1994: 366) identified sources of legitimation from within culture itself. Legislation that is perceived to be legitimate can have stronger effects, especially in representative democracies under the rule of law. There are exceptions to this, where legislation has had little effect on attitudes. But there are cases where legislation has transformed some attitudes or beliefs. Aldashev et al. (Reference Aldashev, Chaara, Platteau and Wahhaj2012: 798) modelled how ‘the formal law can actually pull custom in its direction, thereby causing a progressive evolution of the prevailing mores’. They called this the ‘magnet effect’. They gave several examples and outline the conditions under which the ‘magnet effect’ can be more powerful.

Causality operates in both directions, from informal attitudes and rules to the formal, and from formal legislation to the informal domain. The legitimating power of legal and other institutional authority is confirmed by multiple studies, including the experiments of Stanley Milgram (Reference Milgram1974) showing how people can be manipulated to obey authority, and Tom Tyler's (Reference Tyler2006) analysis of why people comply with the law. Ever since Max Weber's (Reference Weber1968, vol. 1: 212 ff.) identification of key processes of legitimation, much evidence has been accumulated to suggest that the effective powers of formal (legal) institutions are in good measure due to perceptions of legitimate authority. Laws often have additional motivational powers, deriving from the perceived legitimacy of their authority. Law is not simply another set of rules.

Some rules and attitudes that are often seen as cultural and informal did in fact emanate, at least in part, from laws and formal institutions. For example, the historian Martin J. Weiner (Reference Weiner2004) argued that ‘English culture’ was resistant to industrialisation and economic growth. Hence, the rich and powerful sought country estates. But these motives did not spring originally from a culture that favoured country living, but from the survival of feudal institutions in politics and law that made familial landowning an essential means of economic, social, and political advancement. Land ownership was the principal source of status and power. English laws of land tenure included the duty to raise armies. The monarchy needed that system because it provided military forces for use in war. The English aristocracy and gentry sustained and contrived legal measures to keep land within their families, resisting for centuries the efforts of legislators to make more land saleable. Despite the revolutionary events of 1642–51 and 1688–89, the powers of the aristocracy and landed gentry largely remained intact. Despite the growing influence of industrialists and financiers, large landowners and aristocrats retained a powerful presence in the elite. The origins of their power and their landowning culture lie largely in formal institutions in the feudal and post-feudal era (Hodgson, Reference Hodgson2023: 40–1, 79–81, 101–5, 139–45).

Consider another example where legislation helped to create a noted cultural trait. In 1742, a systematic legal code was enacted in feudal Japan. But it was not designed to deal with disputes over property or contracts. Instead, informal settlements or compromises in such disputes were encouraged, sometimes by the threat of state punishment of both disputing parties. Hence, it is possible that customary Japanese aversions to commercial haggling and litigation may result more from draconian legislation to prevent such disputes in the 18th century than from earlier traditional practices or beliefs. The traditional Japanese aversion to litigation seems to have more to do with its evolving legal institutions, and less as something inherent and longstanding in the Japanese culture.Footnote ¹²

Although many informal rules are slow to change, this is not always the case. Some examples show that rapid changes in some cultural traits and moral beliefs are possible, particularly resulting from formal changes by governments and legal authorities, especially when they are perceived as legitimate.

Concluding definitions and remarks

The following definitions are suggested. They are not in alphabetical order but in a constructive sequence:

• Rules. The term rule is understood as an injunction or disposition, that in circumstances X do Y (a positive rule), or in circumstances X do not do Z (a negative rule). A rule can be constitutive (interpret phenomena X as being Y) or regulative (concerning actions). Rules include norms of behaviour and social conventions, as well as legal or formal rules. Contrary to some accounts, rules do not have to be commonly known to be enforceable. In many cases, what is required is the common acceptance of authority. No-one is aware of more than a tiny fraction of legal rules. Yet they are enforceable, partly because many people accept the authority of the legal system.

• Rules and norms. Sue Crawford and Nobel Laureate Elinor Ostrom (Reference Crawford and Ostrom1995: 583–4) differentiated between rules and norms. For them, unlike norms, rules contain the threat of an ‘or else’ component that points to the consequences of non-compliance. This may over-complicate the picture. First, it needs to be made clear whether the ‘or else’ sanctions, such as shunning or ostracism, might be socially rather than legally enforced. Second, in commonplace parlance, the term rule has a broad usage, including the non-threatening (and informal) rules of language, or practical (and informal) rules that are clearly recommendations. It may be better to regard a norm as a kind of rule that does not have a threatening ‘or else’ component. Hence norms would be a subcategory of rules.

• Institutions. Institutions are systems of established and prevalent social rules that structure social interactions. To make rules operative, institutions must involve some shared conceptions. Systems of language, money, law, weights and measures, traffic conventions, table manners, and firms (and arguably all other organisations – see appendix below) are institutions. All institutions are social structures but not all social structures are institutions. For example, the demographic or gender structure of a society is not necessarily an institution.

• Laws. Laws are codifiable rules in institutional systems involving legislatures and courts. These institutions and rules may be legitimated by religion, by tradition, or by a measure of democracy. Some laws may derive from customs and depend on them, but all laws involve legislative and judicial institutions subject to sovereign power.

• Formal and informal institutions. Institutions may be subcategorised as formal or informal. Formal institutions are constituted or maintained through legally decreed procedures, and laws make up the dominant elements in their rule systems. Informal institutions include language, customs, rituals, or table manners. Most of their rules are not constituted by lawmaking institutions. Some rules in informal institutions may be illegal. Mafias and criminal gangs are illegal informal institutions.

• Organisations. An organisation is a special type of institution involving (a) criteria to establish its boundaries and to distinguish its members from its non-members, (b) principles of sovereignty concerning who is in charge, and (c) a structure that delineates responsibilities within the organisation. These conditions imply the existence of social roles or positions that have properties irreducible to those who occupy them.

• Formal organisations. Organisations are formal if they are constituted and maintained by laws and are dominated by legally enforceable rules. All formal organisations include some informal rules, but the dominance of legal formalities makes them formal. Otherwise, they are informal.

The relative importance of different types of institution in economic development is a matter of empirical study. All empirical work requires workable definitions. The discussion of definitions and datasets in the fourth section underlines the importance of a careful and accommodating definition of ‘formal institutions’. The third section highlights the importance of these definitional issues to questions of economic development.

The discussion and definitions above underline the importance of the criterion of whether rules or institutions are constituted and (potentially) enforced by law. That criterion is well established, but it is important to make it explicit and to explore its implications. Arguably, law is not simply an expression of authority but is also a constitutive part of the institutionalised power structure, and a major means through which control is exercised. Law is also a great motivational force. It works not simply through threat of punishment but also because of commitments to what is perceived as legitimate or moral authority. Many other rules do not have this motivational advantage, and obedience to them may depend more on the expected benefits and costs of compliance versus non-compliance. Research using the distinction between formal and informal institutions should take such factors into account.