Explaining Who Pays Grass-Roots Bribes

Abstract

Explanations of corruption are methodologically determined. Studies at the national levels usually concentrate on capital-intensive corruption, but understanding the experience of individuals at the grass roots requires a different approach.

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Explanations of corruption are methodologically determined. Studies at the national levels usually concentrate on capital-intensive corruption, but understanding the experience of individuals at the grass roots requires a different approach. To infer the behaviour of individual citizens from national characteristics implies that individual behaviour is determined by the incentives and constraints imposed by their country’s institutions. If that were the case, then almost everyone living in a county with a negative rating on the Corruption of Perceptions Index would pay a bribe and nobody would do so in countries placed at the top of the Index. However, this is not the case.

To understand why individuals living in the same country differ in their experience of public services requires information about individuals as well as their national context. A sample survey of a nation’s citizens provides appropriate data for testing why some people pay bribes while others do not. However, studying people within a single country cannot take into account differences between national contexts in how public services are delivered. In default of comparative survey data, explanations must be limited to inferences or speculations. It is an ecological fallacy to project onto everyone in a society conclusions drawn from the analysis of national institutions (cf. Robinson 1950). In a complementary way, it is an individualist fallacy to explain the behaviour of individuals solely in terms of their personal characteristics and disregard how political institutions influence what happens to individuals in the process of governance (Scheuch 1966).

The experience that individuals have when contacting grass-roots officials reflects differences between people, between national contexts and between specific public services. The need to contact public services such as education differs with stages in an individual’s life cycle. The institutions that deliver these services differ in their national history and contemporary context. In this chapter, the first section sets out theoretical hypotheses about which individual differences cause some individuals to pay a bribe while others do not. Controlling for variations in national context, the hypotheses are then tested in the second section. The third section goes further: it tests how differences between public services such as health care and policing affect whether an individual pays a bribe for one service but not another.

1 Theories of Who Pays Bribes

The Global Corruption Barometer (GCB) shows that the probability of an individual paying a bribe varies greatly between individuals within countries where bribery is relatively common and where it is infrequent. It also provides data that can be used to test four theoretical explanations about whether an individual pays a bribe. This is likely to vary according to an individual’s contacts with services, perceptions of institutions, political efficacy and socio-economic status.

Individual differences. Contact with public officials is a necessary condition of an individual being asked to pay a bribe. Annually, the great majority of citizens contact a public service and the proportion doing so varies little between continents. Since public services differ in how they are delivered and operate independently of each other, the fact that an individual receives one service by the book still leaves her or him vulnerable in encounters with other services. Because of variations in bribery between services, the number of times an individual contacts a single service is less important than the number of different services contacted. The more public services that an individual contacts, the more likely an individual is to pay a bribe.

When individuals contact public officials, they have expectations influenced by their perceptions of whether they operate by the book or by crook and this will increase their sensitivity to hints by officials that a bribe is needed to get what they want when it is wanted. If there is a widespread perception that officials behave corruptly, people will think it is normal to do so. As Gunnar Myrdal (1968: 409) put it, ‘If everybody seems corrupt, why shouldn’t I be corrupt?’ The more an individual perceives public officials as corrupt, the more likely an individual is to pay a bribe.

In a system of good governance, individuals do not need to know how to make an effective claim, because public officials are trained to ask claimants the questions relevant to determine whether they are entitled to receive a benefit. However, in a system in which an official may hint at the need to pay a bribe, individuals themselves need a sufficient sense of political efficacy to demand that an official give them what they are entitled to without a bribe. If an official is unwilling to do so, people who are more politically confident may turn to someone they know for help in pulling a hook to get what they want. People who have a low sense of political efficacy are more likely to pay a bribe if a public official wants it, because they lack the confidence to resist the demand. The hypothesis that follows is: The greater an individual’s sense of political efficacy, the less likely an individual is to pay a bribe.

In every society, there are multiple inequalities in the socio-economic resources of individuals and in their theoretical implications for whether people pay bribes. Theories of social status postulate that inequalities in resources have a major influence on how individuals are treated by others in society. In a developing country, public officials tend to have a higher status than the people they deal with because they have more education and job security. If some may use their higher status to extract bribes from lower-status people, this is in keeping with Hunt‘s theory (2007): ‘Corruption hits people when they are down’. Reciprocally, grass-roots officials may hesitate to ask higher-status people for a bribe because the latter will be knowledgeable about their rights. The hypothesis is: The lower an individual’s social status, the more likely an individual is to pay a bribe.

Microeconomic theories postulate that socio-economic status has the opposite effect. Profit-maximizing public officials are expected to be more likely to extract bribes from people with higher incomes because they can pay more (Becker 1968). Moreover, higher-income people should be more willing to pay a bribe since its relative cost is much less than for a poor person and the higher value of their time makes it rational to pay to jump a queue for hospital treatment or to get a permit promptly. From a rational-choice perspective: The higher an individual’s economic status, the more likely an individual is to pay a bribe.

In feminist theory, gender can influence whether an individual is exposed to bribery (Wängnerud 2015). Especially in developing countries, social status tends to be male-dominated and men are more likely to be public officials. Insofar as women are more likely to perform caring roles within the family, they are also more likely to be in contact with heavily used social services. Insofar as women have a lower status in a society this predicts: Women are more likely than men to pay a bribe.

Individuals in context. Of the contextual influences on corruption tested in Chapter 3, four were statistically significant. In this chapter, we test the robustness of their effect by seeing whether they remain significant after controlling for within-country individual differences.

The impact of historical context on individuals is ignored by most theories of individual behaviour. Individuals cannot choose the country in which they are born; it is given. People are socialized from a very early age to become part of their society, whether it is governed honestly or corruptly. If you grow up in a country in which bureaucratic institutions have long been established, you do not expect public officials to demand bribes for their services, whereas socialization in a country where bribery is part of everyday life can have the opposite effect. The hypothesized effect is: If an individual has been socialized in a country that bureaucratized early, an individual is less likely to pay a bribe.

Communist regimes represent the other extreme from commitment to the Weberian idea of bureaucracy. Marxist–Leninist doctrines give overriding authority to the Communist party to decide how services are administered. The result is: If an individual has been socialized in a country that has been under Communist rule, an individual is more likely to pay a bribe. In other words, governance is corrupt by the standards of a rule-of-law modern state. Where Communist rule was introduced after early bureaucratization, the interaction of these two opposing influences should reduce individual payment of bribes.

Legacies favouring corruption are receding into the past, and the socialization effect of old regimes is especially weak for younger generations. In no longer Communist regimes, a steadily increasing proportion of adults have been born or socialized into the use of public services after the fall of the Berlin Wall. Democratic institutions are the only institutions that they know. In theory, these institutions make it possible for individuals to hold accountable at the ballot box politicians who are responsible for corruption. The hypothesis is: If governors can be held accountable through democratic institutions, individuals are less likely to pay a bribe.

Foreign aid is significantly associated with capital-intensive corruption in low-income countries where it is a disproportionate amount of national income. Because foreign aid is paid to the national government, high-ranking politicians and officials may be the first to benefit from its corrupt management. The effect of high-level corruption may then trickle down to the grass roots. Aid projects that finance building roads, hospitals and schools can distribute contracts and jobs to regional and local officials who can take a percentage of what is spent for their private benefit without regard to rules governing the procurement of these services. These behaviours can encourage a tolerance of grass-roots officials demanding bribes from individuals who have directly or indirectly benefitted financially from the trickle-down effects of aid. The more important foreign aid is in a national economy, the more likely an individual is to pay a bribe.

2 Both Levels Matter

Our basic model of who pays bribes is: f(bribery) =Individual differences +National context. Logically, the combined effect of individual and contextual hypotheses can have opposing effects, with one set increasing bribery and the other reducing it. Alternatively, taking individual attitudes and behaviour into account could reduce the size of the effect of national context to insignificance. Third, their effects may be additive, with each significantly exerting influence after controlling for the effect of the other. Insofar as integrating individual and contextual influences shows that each is significant, this provides a fuller explanation of who pays bribes than models relying exclusively on studies of individual effects within a single country or cross-national studies that cannot account for within-nation differences of individuals.

Multivariate statistical analyses take into account the independent effect of each hypothesized influence after controlling for the effect of all other variables (Table 1). As we are testing what divides people who have and have not paid a bribe, we use logistic regression and treat countries as cluster variables (Rabe-Hesketh and Skrondal 2012). Since our sample of countries is far smaller than that of individuals, we consider contextual-level variables significant if they have ap-value below 0.10. At the individual level, we consider variables significant if they have a p-value below 0.001. The substantive effect of significant variables is shown in the column reporting shifts in predicted probabilities. Shifts estimate the change in the likelihood of paying a bribe when an independent variable changes from its minimum to its maximum value and the effect of all other variables is held constant. Appendix Table A.1 gives details of each variable.

Table 1 Combined individual and institutional influences on bribery

Contact most important for individuals. Services differ in how they are delivered and the opportunities offered for bribery. It is, therefore, appropriate to test whether contacting more services increases the likelihood of an individual paying at least one bribe. This is the case. After controlling for all other influences, the 14% who contact four or five major services during the year have a 45% greater chance of paying a bribe than those with no contact. The median group contacts three services annually. Of this group, one in four is likely to pay a bribe but they are not inveterate payers of bribes. Instead, their experiences are mixed: Contact sometimes leads to a service being provided by the bureaucratic book and sometimes money changes hands.

The GCB measures the perception of corruption in nine different major institutions (see Fig. 1 in Chapter 4). In total 37% think most or all MPs are corrupt and 36% think that most or all government officials are corrupt. Our measure combines perceptions of government officials, who are responsible for delivering public services, and of MPs, who are expected to represent the interests of people receiving services. The more that individuals see public officials as corrupt, the more likely this belief will become a self-fulfilling prophecy. When confronted by an official demanding a bribe, those who think all officials are corrupt are 10% more likely to pay a bribe than those who think that none want cash in hand. Although the tendency is significant, most people who perceive officials as corrupt do not pay a bribe because officials do not make this a condition of getting their service. This disjunction challenges relying on perceptions of corruption as proxy evidence of the actual payment of bribes (see Rose and Peiffer 2015: Table 5.3).

If people think bribery is wrong and public institutions tend to be corrupt, this is an incentive to avoid contact with services in order to escape the risk of having to pay a bribe because they lack the political efficacy required to make public officials give them what they are entitled to. To assess political efficacy, people were asked whether they agreed or disagreed with the statement: Ordinary people can make a difference in the fight against corruption. One in five believed actions of people like themselves could make a difference, one-third explicitly disagreed with this belief and almost half had no opinion. Notwithstanding the theoretical logic about how political efficacy ought to affect how people deal with public services, the hypothesis associated with political efficacy receives no significant statistical support (Table 1). Those who think that ordinary people can make a difference in fighting corruption are just as likely to surrender to demands from corrupt officials as people who feel helpless.

Major inequalities in status are measured in a multiplicity of ways. Occupational class is inappropriate for characterizing GCB respondents, since a majority live in developing countries in which many people do not have a steady occupation. Income is inappropriate, because of the lack of a single currency that can be used for international comparisons. Moreover, many GCB respondents avoid poverty by eking out their irregular cash earnings by producing much of what they consume. Education is a major source of status in every country and inequalities are particularly marked in developing countries. Countries differ in the number of years of free education offered, and within every country younger people have more opportunities for education than their elders. The GCB survey identifies three levels: no education or only an elementary level; secondary schooling; and higher education.

Education makes no difference for the payment of bribes. Educated people not only know what services they are entitled to but also know from experience that an official demanding a bribe has no more respect for their status than for his or her bureaucratic obligation to deliver services by the book. The lack of influence of socio-economic status is confirmed by analysis of the 2013 GCB survey, which included a subjective measure of income: whether people view their household income as average, below average or above the average in their society. When this income measure was included in a multilevel statistical analysis, it had no significant effect on the likelihood of paying a bribe (Rose and Peiffer 2015: 68ff).

Gender does have a statistically significant influence on bribery: Women are less likely to pay a bribe than men. However, the impact is insubstantial. After controlling for other influences, the likelihood of a woman paying a bribe is only one percentage point less than for a man. A preliminary analysis also tested the effect of age; the result is not reported here since it had no significant effect on whether any bribe was paid.

Context matters too. The multilevel analysis confirms the independent effect of national context on bribery. After controlling for the effect of within-country individual differences, all four contextual influences that significantly affected a country’s national Corruption Perceptions Index (CPI) rating influence individual behaviour too. Moreover, the impact of context on individual behaviour is similar in size to the impact on a country’s national CPI rating (cf. Table 1, Fig. 1 and Table 1 in Chapter 3).

The significance of national context is strikingly illustrated by legacies from a country’s political past having a strong effect on individuals born when their country had a different political system. The impact of a country beginning to bureaucratize before 1914 is significant because it started a path-dependent process that has led to good governance today. After controlling for individual differences, bureaucratization begun more than a century ago reduces the likelihood of paying a bribe by 11%.

The impact of a Communist legacy is contingent. When all countries experiencing a Communist regime are grouped together, it has no significant effect on bribery. This is because countries that are no longer Communist do not share similar histories. Successor states of the Soviet Union, for example had no experience of early bureaucratization before becoming subject to governance in accord with Marxist–Leninist principles that relied on political favouritism rather than bureaucratic impartiality. By contrast, a majority of Central and East European countries experienced early bureaucratization as part of the German or Austro-Hungarian empires. Even though no GCB respondents had been born when bureaucratization was introduced before 1914, they nonetheless experience its path-dependent effects today.

The interaction between being bureaucratized early and then being subject to Communist rule cuts two ways. On the one hand, dealing with officials in a political system that had been bureaucratized early reduces the likelihood of having to pay a bribe by 11 percentage points. However, if the political system was subsequently turned into a Communist regime after the end of the Second World War, the interaction of this negative experience and the positive pre-1914 legacy raises the chance of paying a bribe by 18 points. The greater strength of the Communism legacy reflects the thoroughness with which so-called Socialist legality was imposed in the wake of the Second World War. Communist commissars purged senior officials who had been trained to administer services by the book; they were replaced by officials whose first loyalty was to follow the twists and turns of the Communist party line.

Since the Freedom House democracy index does not discriminate between old and new democracies, its impact shows that recent changes in institutions can have a significant effect independent of historical legacies. Compared to countries with the lowest rating for democracy, citizens in countries that have the highest rating for democracy are 20% less likely to pay a bribe. This calls into question the argument of many democratic theorists that it takes generations for new democratic institutions to influence the way in which public officials behave and how ordinary citizens respond.

A second contemporary influence, the proportion of a country’s national income coming from foreign aid, has a big trickle-down effect on individuals using services that aid helps finance. Even though the intention of aid is to improve the quality of public services and governance, in countries where aid is most important the likelihood of an individual having to pay a bribe is 20% higher than where national governments receive no foreign aid.

The combined effect. Multilevel statistical analysis confirms that where you live is as important for paying a bribe as who you are. How individuals react to the experience of contacting public officials cannot be fully understood by relying exclusively on measures of individual attitudes and socio-economic characteristics. Four characteristics of national context are significant. While necessary, contextual influences are not sufficient to account for who pays bribes. Three measures of individual differences also have a significant impact on the payment of bribes (Table 1).

Fig. 1

(Sources Predicted probabilities of impact of multi-level logit analysis reported in Table 1 significance level: characteristics < 0.001. Institutional level: < 0.05)

Impact of significant influences on paying bribe

From an anti-corruption perspective, the ideal system of good governance would be found in a country that began bureaucratization early, that is fully democratic today, receives no foreign aid and has never been subject to Communist rule. This combination makes the delivery of public services by the book normal. However, there are only seven countries in our GCB database that combine all four characteristics conducive to good governance. This small number reflects its under-representation of countries where bribery is very low (see Appendix Fig. A.1). It is due to many of Transparency International’s national chapters in Western Europe convincing TI’s Berlin headquarters not to bother asking about bribery in their countries when at most only a dozen or two respondents among the thousand persons interviewed will have paid a bribe.

In most member states of the United Nations, the national political context falls far short of what is required for good governance. More than three-quarters started bureaucratization late. Fewer than half the countries evaluated by Freedom House are rated as fully democratic. Moreover, among developing countries, foreign aid is often more than 1% of their national income. When tax revenue is relatively low and difficult to collect, foreign aid constitutes a substantially larger proportion of the revenue controlled by the country’s central government.

Political leaders in countries lacking a legacy of early bureaucratization cannot be blamed for what happened long before they were born. Nor can contemporary corruption be blamed on a colonial legacy, since statistical analysis shows that it has no significant influence on bribery in countries that have been independent for half a century or more. Two significant influences on corruption are contemporary conditions for which today’s governors are responsible: the refusal of governors to be democratically accountable and the diversion to their private benefit of foreign aid money meant to finance the country’s postcolonial development.

3 Differences Between Services

The public services that low-level public officials deliver at the grass roots differ from each other in what is delivered, how it is delivered and who benefits. For example, policing is different from health care in the laws police enforce, their authority to tell people what they must do, and, by contrast with the reliance on doctors for treatment, many people do not want to have contact with the police. There are also substantial differences in the extent to which people have contact with different services: Individuals are four more times as likely to contact health services during a year as the courts and half again as likely to contact schools as the police. There are also differences in the likelihood that being in contact makes a person vulnerable to paying a bribe. Even though police services are much less used than social services, the likelihood of having to pay a bribe to the police, once contact is made, is much greater than for social services (see Table 2 in Chapter 4).

The delivery of grass-roots services is a two-step process. Individuals must first make contact with a service. This divides people into two groups: those who are not at risk of paying a bribe because they have no contact and those who face a risk, whether small or large, by being in contact with public officials. With the exception of health, in the year before being interviewed a majority of GCB respondents had no contact with the courts, the police, officials issuing permits and schools. They thus were not at risk of paying a bribe. For those who do contact a service, the second step is getting what you want by the book, by hook or by crook. Individuals in contact with public services divide into two groups: the majority who do not pay a bribe to get what they want and a minority who pay a bribe at least once.

Service-specific hypotheses. For each service, the likelihood of an individual making contact or paying a bribe may be influenced by service-specific features, such as the supply of hospital places, as well as by pervasive features of national context. Laws set out the specific conditions that entitle or obligate individuals to make contact with a public service. Education is doubly affected; laws not only entitle children to receive a free education but also obligate parents to see that their children attend school. Laws about the health service do not compel individuals to have periodic check-ups or follow a healthy daily routine; they specify the conditions that entitle people to receive health care financed in whole or part by public funds. Contacting local officials to secure a permit or license depends on what people want to do. If an individual wants to secure a passport for foreign travel, contact is obligatory, but foreign travel is a matter of choice. Contact with the police is contingent on individual characteristics and behaviour, and on informal as well as formal standards of police behaviour.

Social roles arising from age and gender may influence the likelihood of individuals contacting a service. Contact with education services follows from adults being parents of school-age children. This is most likely to be the case when people are in the age range between the mid-20s and the early 50s. Social roles are different for health. Even if the whole of a country’s population is eligible for publicly funded health care, people over 65 are much more likely to need health care. Gender can also affect contact with health care, since women are a disproportionately large part of the older population. Since crime rates are higher among youths, age can influence the extent to which young people unwillingly come into contact with the police.

Gender can affect the likelihood of men and women contacting specific services. During the life cycle, women are more likely to need health care because they live longer in old age, when health needs are greatest, and because of needing maternity services when younger. In many societies, women are more likely to have the role of a carer for other members of their family. Looking after a child’s education and health will bring women into contact with two public services and looking after the care of ageing parents or relatives can add to contact with health services. In more traditional societies, dealing with local government officials may be considered a man’s role. Engagement with the police can be gender-biased towards men rather than women (Kevane 2004). (Hypothesis 1 Social role.) If an individual’s social role is relevant to a specific service, she or he will be more likely to contact it. A corollary of this hypothesis is that, after controlling for contact, differences in age and gender will have no effect on the payment of bribes.

The supply of public services in sufficient quantity to meet all needs is taken for granted in societies with a high gross domestic product (GDP) per capita. However, in many developing countries, governments may find it easier to enact laws entitling people to education and health care than to find the money to supply the school places and health facilities needed to meet the demand they have created. The percentage of youths of secondary school age who are actually receiving a secondary education in countries covered by the GCB is as low as 21% in Niger compared to 100% in most developed countries. A good measure of the current supply of health care is the percentage of childbirths in which the mother is attended by a trained provider of healthcare services. While this is invariably the case in high-income countries, the proportion of births with a medical professional present drops to 35% in Nigeria.

When the supply of a service in demand is scarce, a standard bureaucratic practice is to introduce queuing. It can also stimulate some people to want to jump the queue. Making people queue for months for treatment for a painful illness imposes pain or worse. Raising the examination score required to enter university appears fair, but it also increases the number of applicants who cannot get a university place. Long delays can encourage people to find the money to pay a bribe to jump the queue for health care or to pay an education official to admit a youth to university whose examination marks are not high enough to qualify for a place (Aidt 2003). While scarcity in the supply of a service implies that fewer people will make contact, it also implies that bribes are more likely to be paid by those who do have contact. (Hypothesis 2 Supply of social services.) If a social service is in scarce supply, individuals are less likely to have contact but more likely to pay to pay a bribe.

In every society administering laws is not confined to the police and the courts. Regulations requiring official documents such as permits and licenses are an integral part of public services too. Regulations require individuals to register births, license their car if they have one and get a permit to build a house. A surplus of regulations encourages bribery; the more signatures of public officials that are required before a document is issued, the more costly it is in time and effort to obtain them and the more opportunities there are for corrupt officials to collect ‘rents’ (i.e. bribes) as a condition of expediting the delivery of documents. Bribes can also be paid for issuing a permit in contravention of official rules or ignoring the enforcement of a regulation.

Since businesses are specially subject to regulations, we measure the extent of regulation with the Regulatory Quality measure from the World Bank’s World Governance Indicators. Venezuela and Zimbabwe are ranked as having the poorest quality of regulations while Hong Kong and Australia are credited with being the countries with the highest quality of regulations. The efficiency of high-quality regulations is likely to reduce the need for those affected to have lots of contacts as well as the likelihood of paying a bribe. Conversely, poor quality regulations are likely to create more burdensome red tape requirements and more need to contact public officials for approval. (Hypothesis 3 Supply of regulations.) The greater the burden of regulations in a country, the more likely individuals are to contact regulatory services and to pay a bribe.

Public services differ in whether the state is a monopoly supplier or private sector and not-for-profit institutions offer an alternative. In virtually every country there is a choice between alternative suppliers of social services, such as public and private schools or public and private hospitals and doctors (Rosling et al. 2018). In principle, choice offers individuals the opportunity to avoid contact with corrupt public officials by going to a non-state alternative. By contrast, if a service is a monopoly of the state, such as policing or the issuing of permits, individuals have no choice but to deal with the state when engaging with monopoly services. However, this also means that in a system of bad governance engagement with a monopoly service is at the risk of paying a bribe (cf. Hirschman 1970; Klitgaard 1988).

The theoretical rationale for making alternative services available is economic: If there is a choice between public-and private-sector institutions, the service will be more efficient. The political objection is that making alternative services more available increases social inequality, because only those with more resources will have an effective choice between the two types of services. Even in a high-income country, the cost of using private schools and hospitals is beyond the resources of many people. In a developing country where inequalities of income and education are much greater, very few have the resources to choose between public and private social services. Those with fewer resources are more likely to have to use a public service. In a system of bad governance, the choice is likely to be between finding the money to pay a bribe for using a social service or doing without. By contrast, those with more resources are more likely to want permits for their personal and business affairs, and thus more likely to have contact with regulators. If they cannot deal with them by invoking their higher status, the marginal cost of escaping from monopoly obligations by paying a bribe is less than for people with fewer obligations. (Hypothesis 4 Alternative services or Monopoly.) If individuals have more resources, they are less likely to contact a public service when there is an alternative but more likely to contact and pay a bribe for monopoly services.

Testing service-specific influences. The logic of separately examining contact and bribery at the service level is that hypothesized influences are contingent rather than uniform. An influence can be significant for contacting public services but not for paying a bribe, or vice versa. In addition, the indicators appropriate for independent variables can differ from one service to another. For example, the stage in the life cycle for contacting a public service differs between education and health care. There are also categoric differences for measuring the supply of different services.

To take account of contingencies, we test separately the influence of social roles, the supply of services and institutional alternatives on four different services: health care, education, permits and the police. In doing so, we control for the effect of differences in national context. To test our two-stage model statistically, we employ a Heckman (1979) logistic regression model, which estimates the impact of the hypothesized effects on contact, and by accounting for those influences on contact, the model also identifies the independent influences of hypothesized effects on the payment of a bribe (Table 2).

The first criterion for testing a hypothesis is whether it has a statistically significant effect on contact with particular services and/or paying a bribe. The second criterion is whether a significant influence also has a substantial influence, that is whether the estimated size of its effect is relatively large or small. Third, we compare whether an influence is consistently significant and strong across most services analysed. Given the volume of statistics generated by testing the Heckman model, Table 2 presents results only for significant influences on each service. The full regression statistics are given in Table 3.

Table 2 Comparing influences on contact and bribery by service

Table 3 Heckman analysis of influences on contact and bribery

The importance of social roles related to gender and to age are each contingently significant. For education, middle-aged persons are more likely to have children, and therefore are nine percentage points more likely to have contact with education officials compared to young people who have not yet started a family and compared to older adults whose children have completed their education. Gender is significantly associated with contact with health services while being older has no influence. Whatever their age, women are 5% more likely to contact health services than men. The absence of an age effect is likely to be due to the fact that women are likely to have need of health services at both ends of the life cycle: When women are young maternity services are needed for themselves and for their children and when women are older, health services are needed to deal with conditions of old age.

Since many permits involve once-for-all concerns arising in earlier stages of the life cycle, the impact of age on contact is very strong, 20 percentage points. The police are significantly more likely to be in contact with young people but the margin is slight. Once the impact of age on contact is accounted for, life cycle differences have no significant relationship with paying a bribe for social services and young people are only 3% more likely to pay a bribe to the police. Differences in contact and bribery due to a person’s position in the life cycle are significant but temporary. In the course of the life cycle an individual’s need for contact is likely to change, being biased towards education when younger and health care when older.

The social role of women has a significant influence on contact too. Women are more likely to have contact with social services than men, although the impact of being a woman increases the likelihood of contact with health services by no more than five percentage points. Being female is associated with a slight increase in the likelihood of contacting education officials but the effect falls short of significance (Table 3). Gender also has a predicted impact on contacting obligatory services. Women are 4% less likely to seek permits and 6% less likely to be in contact with the police. After accounting for the impact of gender on contact, there is no significant impact of gender on bribery except for permits, for which men are only 1% more likely to pay bribes than women.

More educated people ought to be better at understanding bureaucratic rules and feel more confident in asking assistance from well-educated doctors and teachers. Having a better understanding of their entitlements may encourage educated people to ignore signals from officials about paying a bribe. By contrast, people with a minimum of education may be intimidated and feel they must pay bribes to get what they want. This generalization can be applied to the use of obligatory services such as the need for permits as well as desirable social services.

Although the inequality between high and low education is clear, it does not lead to a significant difference in the use of public services. Less educated people are just as likely as more educated people to have contact with each of the four services. However, equality in vulnerability to corrupt officials has a cost in countries where corruption is relatively high, since it means that officials are prepared to exploit the great majority of their subjects, regardless of their social or economic status.

Contrary to the importance of supply and demand in economic theory, whether there is a sufficient supply of social services or they are scarce has no significant influence on whether people make contact with social services. A shortage of school places may mobilize parents to badger education officials in hopes that persistence may secure a scarce place, people seeking medical treatment in short supply may resort to methods to get what they want by using hooks rather than bribery. The quality of regulations has no impact on contact with permits offices or the police, however, it does have an influential association with bribery. In an environment of weak regulations, people are more likely to pay a bribe to get what they want, such as the non-enforcement of a law by the police.

The theory that choice and monopoly services will differentiate users according to their resources is not supported by statistical analysis (Table 2). Whereas income can be a suitable indicator of resources in a fully monetized society, the GCB includes many countries in which people may rely on getting what they want by being in the know politically. Since more educated people ought to know better how to work their political system, we use it as our measure of resourcefulness as well as of having a social role. Table 2 shows that having more education does not result in significantly opting out of contact with public services. Instead, it may enable people to remain in the public sector by getting preferential treatment without being more likely to pay a bribe. Nor do educated people more often use monopoly services. If they do, their knowledge does not make them less likely to be subject to paying a bribe than people without their resources.

Contrary to theoretical expectations, the extent to which people perceive political institutions as corrupt has no significant effect on contact with three services (Table 2). The exception is the police: people with a negative view of the police are eight percentage points more likely to be in contact. In other words, experience counts as well as second-hand accounts from television or other media. Because of the overlap in the independent and dependent variables, the number of services an individual contacts cannot be applied to individual services.

Individual controls consistently influence the payment of bribes for specific services. The effect of perceptions on bribery is as high as 13 percentage points for police, even after controlling for the effect of perceptions on contact. Perceptions also have a significant impact on the other three services. The number of services contacted also has a significant effect on the payment of bribes for health care and for permits.

Applying contextual controls to specific services produces a better understanding of the legacy of Communist regimes (Table 2). In countries with this anti-bureaucratic legacy, people are consistently less likely to contact services. The impact of living in a post-Communist country reduces the likelihood of people seeking permits by 24 percentage points. People prefer to evade requirements for documents rather than risk having to pay a bribe to comply with the law. A similar aversion is shown towards the police. Not contacting schools is a reflection of less need to do so because of the drop in birth rates after the fall of the Berlin Wall.

Bribery becomes significant when there is an interaction of the legacy of early bureaucratization in Habsburg and Wilhelmine German lands and a Communist legacy. This consistently increases contact with all four services; however, this can be an ambiguous benefit. The interaction results in an increase of 11 percentage points in the likelihood of paying a bribe to get permits rather than evade this requirement and an increase of 10 points in the likelihood of paying a bribe for health services.

The extent of democratization has no significant effect on contact with particular services, a reminder that the provision of public services is consistent with ‘welfare state authoritarianism’ as well as the satisfaction of popular demands that can be registered in a democratic political system. Consistent with theories associating democracy and good governance, having fully democratic institutions substantially reduces the likelihood of paying bribes for all four services. The impact is greatest for policing; those living in a fully democratic regime are 23% less likely to pay a bribe to the police and 11% less likely to pay a bribe for education. Although a significant amount of foreign aid is meant to be invested in promoting human capital through improved education and health care, aid levels have no significant effect on contact with education, but a high level of foreign aid is associated with greater use of health care. Citizens living in highly aid-dependent countries are 11% more likely to pay a bribe for health care, 24% for education services and 32% more likely to pay a bribe to the police. This finding is consistent with foreign aid being more likely to be given to countries that the CPI characterizes as having the most corruption (Table 1 in Chapter 3) Grass-roots public employees may then emulate their national politicians by demanding bribes for the delivery of retail services.

The focus on specific services shows that people who contact several services may not always have to pay a bribe and that this variable is different across service types (see Fig. 4 in Chapter 4 ). Moreover, it offers evidence to policymakers about what influences people to make more or less use of particular services. The importance of contextual differences in both models emphasizes that a major source of the inequality in the payment of bribes for public services is not within a country but between countries.