Voters, Politicians and Clientelism: a Dublin Survey


ãLEE KOMITO 1989, Administration 37 (2): 171-196


The examination of clientelism has been a major theme in Irish politics and administration. People usually understand clientelism as referring to exchanges in the electoral arena: politicians intervene, on behalf of voters, in the administrative process, and, in return, voters reward politicians with votes. If most citizens do not recognize the term, they recognize the phenomenon: politicians using their personal influence to obtain state benefits for constituents and, in return, constituents providing their votes. Politicians are viewed as brokers, mediating between the state and the public.

Discussions of clientelism are often couched in moralistic terms, with some arguing that voters are being ‘fooled’, and others saying that clientelism trivializes Irish politics. Clientelism has been linked with rural values, in which people accustomed to face-to-face interaction are afraid of bureaucracy and so seek assistance from politicians. Others have seen it as a relic of colonial domination: in order to deal with an unyielding and foreign administration, middle-men who might have special influence were sought. It is often seen as, at best, old-fashioned, and at worst, immoral: people should be able to directly obtain what they deserve, and politicians should not capitalize on people’s ignorance in order to get themselves elected.

Academic analyses often also imbue Irish clientelism with an aura of corruption or suspicion: as though clientelism should happen only in societies ‘less-developed’ than Ireland. Since Chubb characterized politicians as ‘persecuting civil servants’, the clientelist model has been used to explain those aspects of Irish politics which differ from ‘modern’ political systems. Clientelism, it seemed, was a feature of Irish politics that served no constructive purpose, and should disappear. Similar themes manifest themselves in government studies. The Public Service Organisation Review Group [the Devlin Report] identified political interventions as a major problem in the Public Service, and suggested various remedies.

It is indicative of the ambivalent attitudes to clientelism that, despite years of discussion, little changed. In recent years, clientelism has become a political, as well as academic, issue: not only do journalists write articles about ‘messenger-boy’ TDs, but politicians themselves argue that the present system, in which politicians must spend all of their time in constituency service, is counter-productive.

Given these concerns, and the extent to which rhetoric and analysis overlap, it may be useful to review what is known about clientelism in Ireland, and analyze a survey that provides useful data on clientelist activity. Part of the difficulty has been that the clientelist model so closely matches the popular vision of Irish politics that it is difficult to distinguish between metaphor and analytic concept. A first step is to realize that the clientelism attributed to Irish politics is only one type of clientelist exchange, one in which the electoral process is used as the medium of exchange. This type of clientelism has been termed ‘electoral clientelism’. Electoral clientelism presumes two sets of dyadic interactions: one between voter and politician, and another between politician and public official. In early discussions of Irish clientelism, neither dyad was examined very closely, since both exchanges were presumed to be rather self-evident. Since then, the exchanges between politicians and public officials have received closer examination (e.g., Collins 1987; Higgins 1982; Komito 1984,1985; Roche 1982). However, exchanges between voters and politicians have not, as of yet, received similar attention (with the important exception of Kelly, 1987). What does the exchange between politician and voter actually involve? How wide-spread is it? Must all voters become ‘clients’ of some politician or other? If not, what are the social and economic characteristics of those who become clients? There is very little concrete information on any of these questions. This article analyzes a survey of Dublin voters taken in the early 1970’s, which provides some information about these issues.

Clientelism is often discussed as an ‘abnormality’ that requires explanation. This may be sufficient for popular discussion, but not for use as an analytic concept. As a minimal definition, clientelism is an exchange between two individuals of unequal standing, in which each provides something which the other cannot directly obtain. In the case of electoral clientelism, the exchange between citizen and politician is an exchange of votes for state benefits.

Anthropological studies of clientelism originated with studies of peasant communities (e.g., Mintz and Wolf 1959; Pitt-Rivers 1961; Wolf 1956). These studies described long-lasting personal bonds between non-kin, in which individuals of unequal status provided resources to which the other party had no access. Such relationships exist in many parts of the world: the Mediterranean, Latin America, Africa, and Asia (see Eisenstadt and Roniger 1984; Powell 1977). All of these relationships exhibit some common characteristics: they are between people of unequal socio-economic status, they are personal and face-to-face, they are voluntary, and they persist over time (for a sample of various definitions, see Eisenstadt and Roniger 1984; Powell 1977; Clapham 1982; Scott 1977; Graziano 1975). These exchanges are often infused with special moral values, suggesting a non-economic bond between the parties. Fictive kinship, or ‘godfatherhood’ gives an added moral dimension to the personal relationship, which serves to disguise the inequality which creates the need for such exchanges. Pitt-Rivers’ (1961) characterization of them as ‘lop-sided friendships’ evokes the mixture of economic transaction and moral value which seems to characterize clientelism.

An analytic distinction has been drawn between patronage and brokerage. Patronage involves providing valued goods that are directly controlled, they are in someone’s ‘giving’. One who does not directly control the resources, but has special influence over, or contact with, those who do, is a broker (Boissevain 1974; Paine 1971). In both cases, the element of monopoly is crucial. It is a broker’s or patron’s exclusive access to valued resources which makes clients depend on them and which permits brokers and patrons to ‘charge’ for their services (cf. Silverman 1965), as well as the power to define the terms of the exchange (Paine 1974)

One mode of clientelist exchange uses the medium of elections: the exchange between voter and politician. The citizen has his vote, which he exchanges for the resources which the politician can provide: access to the benefits distributed by the state. The client’s bond is a personal loyalty to the politician, and the party or policies which the politician represents is irrelevant for the client’s support. It is this type of clientelism, often called ‘electoral clientelism’, which most people presume exists in Ireland.

Irish Clientelism

Irish politics appears to fit the clientelist model. Election rhetoric revolves around special influence and past favours as politicians compete to help the voters, and voters believe that the assistance of politicians is the best guarantee for receiving state benefits. The stereotyped view of Irish politics is the personal exchange between politician and voter in which the politician uses his influence to obtain state benefits for the constituent, and the constituent provides electoral support in return.

The agenda for clientelist research in Ireland was set by Chubb’s (1963) description of politicians as local men who looked after their constituent’s interests by ‘going about persecuting civil servants’. Chubb suggested that the Irish politician’s primary task was to mediate between his local constituents and the state’s administrative apparatus. Voters wanted state services, and politicians helped or appeared to help people obtain those services. Within this framework have come descriptions of local politics from Cork (Bax 1976), and Donegal (Sacks 1976), and a study including material from Kildare (Carty 1981). These studies emphasize the personal contacts of politicians, as well as the diffuse economic and moral bonds between politician and voter. People use their vote to reward those politicians who demonstrated their ‘pull’. The explanation for clientelist politics focuses on rural values: ‘The countryman, coming out of a small community, places a strong value upon face-to-face relations with people, and [the politician is] the countryman’s personal emissary to an anonymous state’ (Sacks 1976:50-51).

These studies provide excellent descriptions of Irish politics, but there are a number of issues which remain ambiguous. If clientelist politics results from rural values, then it would, presumably, disappear as modernizing influences spread throughout the country. But others have suggested that clientelist beliefs are the result of structural factors (Gibbon and Higgins 1974; O’Connell 1982; Garvin 1982; Komito 1984). If so, then clientelism will exist in urban as well as rural areas until there are economic and social changes in Irish society.

In these debates, there seems to have been little discussion about what clientelism actually is. In fact, there is little concrete data on how clientelist exchanges manifest themselves. What is the precise relationship between voters and politicians? The use of the term ‘clientelism’ implies a link between voter and politician; do such links dominate Irish politics? Studies have tended to treat all clients as an amorphous aggregate, whereas clients clearly differ in the resources which they need and the resources they can offer in exchange. Socio-economic status, position in the community, and political participation all distinguish one client from another. One must specify the kinds of resources over which influence is desired, what is being exchanged between politician and bureaucrat, what resources each side brings to the exchange, and the clients to which it is directed.

Existing studies provide little information on such issues. In so far as the relationship between voter and politician has been discussed, it has been somewhat impressionistic, with a few stories used to represent the overall pattern. With the exception of Kelly (1987), studies have not examined individual transactions between voters and politicians. Nor do surveys such as the Irish Times’ pre-election surveys or the RTE 1976 survey throw much light on the relationship between voters and politicians. While the Civic attitudes survey (Raven and Whelan 1976) has some material, it has not been analyzed in terms of brokerage. A survey carried out by the Institute of Public Administration, which provides information on exchanges between politicians and voters, is thus of particular interest.

Dublin Survey

In 1972, the Institute of Public Administration undertook a survey to elicit material on civic attitudes and political knowledge among Dubliners, seen as especially significant in the wake of the abolition of Dublin Corporation when, in 1969, it refused to strike a rate. Although the questions asked in the survey were not explicitly designed to elicit information on clientelism, they do offer insight into clientelist beliefs and strategies. Although the survey is over fifteen years old, there is so little survey research on this topic that any information can only help fill out the very incomplete picture that presently exists.

The survey examined the Dublin Corporation area comprising central Dublin and some suburban areas. Out of 141 electoral wards, 24 were chosen, and 42 names were chosen from the electoral register in each ward. The sample size was thus 1,008, and 701 interview schedules were completed. A preliminary account of the survey results can be found in Litton (1973). The raw data was subsequently made available to the author for analysis, and was recoded. Unfortunately, due to an interviewer’s error, over 200 of the responses could not be used, leaving a total of 499 valid cases. While 500 cases is a less than desirable number, it has been possible to obtain a number of relevant correlations regarding clientelist beliefs and behavior.[1]

At the time of the survey, there was little data available on the demography of Dublin; it was difficult to ensure a representative coverage in the selection of wards throughout the city. In retrospect, the survey achieved good coverage, as shown by a comparison between the number of actual electoral wards in each social area of the NESC (1981:91-103) survey and the number of wards and respondents per social area sampled by the IPA survey:



total wards

IPA sample

percent of IPA


per area

wards per area

respondents per area





Area 1

25 (18%)

4 (16.6%)


Area 2

24 (17%)

2 (8.3%)


Area 3

24 (17%)

5 (20.8%)


Area 4

15 (11%)

3 (12.5%)


Area 5

31 (22%)

6 (25%)


Area 6

22 (16%)

3 (12.5%)







141 wards

24 wards




The inner city area (Areas 1 and 2) was undersampled and Corporation housing estates (Area 5) were oversampled; but the selection of wards generally parallels the social areas suggested by the NESC study. The survey results also parallel the distribution of occupations in Dublin as described in the 1971 Census of population (CSO 1975:147). Both surveys categorized occupations according to the same modified Hall-Jones scale (CSO 1975:vii; see also Reid 1977:44-45). The IPA categories could be collapsed into the following:[2]








Professional [AB]



Other non-manual [C1]



Skilled working-class [C2]



Semi-skilled and unskilled [DE]




As with the geographical sampling, there was a slight over-sampling of the more deprived socio-economic groups. The survey used occupational classifications, and these classifications will be used as short hand labels for the different socio-economic groups.[3] Wives were assigned the occupational status of their husbands. The IPA survey found the same overlap between socio-economic status and residence as the NESC (1981) study found. The inner city and Corporation estate residents are largely working-class, while residents of suburban areas are largely middle class.

The occupational distinctions are paralleled by other measures, such as housing density. The survey asked respondents to indicated how many rooms in the residential dwelling, and then how many people lived in the dwelling. If one divides number of individuals by number of rooms, ones can obtain a rough measure of housing density. The average housing density was 1.04 (n=496), with expected contrasts between inner city and public housing versus private housing. This is most striking if one uses the NESC areas, comparing the Corporation estates of Area Five with the post-1960’s private housing of Area Six:



























The IPA survey shows other differences in community life. For instance, approximately one-third of inner city and Corporation estate respondents had lived in their residence for over 20 years; only 20 percent of suburban residents responded similarly. Nearly 50 percent of inner city and Corporation estate residents could name over 20 of their neighbours, while only one-third of suburban respondents could do so. Only 17 percent of inner city and Corporation estate residents said that there was no one in their neighbourhood that could go to for assistance, while 27 percent of suburban residents would not go to anyone in the area. Thus, the stereotype of isolated middle class households versus close-knit working class communities is largely accurate.

The local electoral wards mirror the social and economic divisions. The survey shows that each area has quite different priorities regarding which state services are important. Asked which local service they considered most important, 48 percent of inner-city residents (Areas 1 and 2) selected housing; only 37 percent of Corporation estate residents who, by definition, already had houses (Area 5) chose housing (and were more concerned with amenities). The private home owners (Areas 4 and 6) and would-be owners (flat dwellers in Area 3) were least concerned with public housing (only 26 percent), but were very concerned about infrastructure (such as refuse, water, sewerage, and town planning), emphasizing the problems in providing services to new estates. These different concerns are reflected in the different services which politicians would be asked, as brokers, to provide.[4]

Brokerage Expectations

The IPA survey asked respondents to choose what role they considered the most important one for councillors to assume (see p. 192). They could have indicated that councillors should ‘intercede with the corporation for the people, to get them housing, planning permission, etc.’, or they could have indicated that councillors should ‘run the city according to some overall policy which the people approved at election time’. Sixty percent of 463 respondents chose the brokerage role, while only 40 percent indicated policy-making.[5]

It is not unusual that voters would want politicians who would look after their concerns; this is a feature of the politician’s role throughout the world. As Mezey (1979: 159) notes:

  ‘If you were asked to name the one activity that legislators are most likely to perform regardless of the country in which they serve, your best response would describe the job of receiving and dealing with particularized demands coming from constituency-based individuals and groups. Members of every type of legislature say that they are subjected to an incessant flow of such demands, and they indicate that coping with them requires a substantial portion of their time and resources.’

Descriptions of politics virtually anywhere in the world sound similar to the rounds of clinics and constituency work of Irish politicians. For instance, a survey of voter’s expectations in Kenya, Korea, and Turkey suggested that, in all cases, over forty percent believed that legislators should help constituents who have personal problems with the government or obtain benefits for the constituency (from Mezey 1979: 165). Such a comparison may be misleading, since politicians in many other countries have the power to intervene successfully; it is still not clear that Irish politicians possess such power.

An examination of the distribution of beliefs and expectations provides insight into these issues. The distribution of brokerage beliefs is by no means uniform throughout the population, and social and economic divisions are clear-cut. As few as 24 percent of middle class heads of households respondents (AB) believed councillors should be brokers, while 70 percent of semi-skilled and unskilled working class household heads (DE) opted for brokerage (n=278). In between were C1 (54 percent) and C2 (52 percent) groups. If wives were also included, the differences were lessened, but were still significant:


Household Heads
































All respondents

























It would seem to be the extremes (AB and DE) that are significant; with little difference between non-manual and skilled working-class categories. However, an investigation of respondent’s father’s occupation, some measure of class origin, shows a more striking contrast:








































It is clear that those with a middle-class background diverge markedly from those with working-class or rural backgrounds. Thus, the generalization that rural dwellers and urban blue-collar workers are more likely to see politicians as brokers seems accurate. Furthermore, socialization has a lasting influence on beliefs which is not altered by subsequent changes in socio-economic circumstances. The distinction between current status and origin is especially relevant in Dublin, since most farmers’ offspring who have come to Dublin are middle class: in this survey, 86 percent have white-collar jobs, 77 percent live in newly built suburbs, and 86 percent rent or own housing in the private sector.

In addition to socio-economic status, residential status was another important measure of brokerage beliefs. Those who rented from the Corporation were likely to want politicians to be brokers, while those who owned their own houses were less so. This is not simply a reflection of socio-economic status, even though middle-class respondents are more likely to own their own house, and working-class respondents most likely to depend on public housing. Working-class respondents who own their own house are less likely to desire brokerage than other working-class respondents, and middle-class respondents who depend on public housing are more likely than other middle-class respondents to desire brokerage. It seems clear that a dependence on state-supported housing is an important determinant of belief, regardless of any other factors.

Dublin’s socio-economic groups are, by virtue of housing policies, constituted into distinct residential groupings (see NESC (1981), Hourihan (1978) for details); it is not surprising that brokerage beliefs vary accordingly. In the inner city (comprising, roughly, the inner canal ring, and social areas 1 and 2 in the NESC survey) and Corporation estates (social area 5), 67 percent of respondents wanted councillors to be brokers. On the other hand, in the private housing estates (areas 3, 4, and 6), only 51 percent opted for brokerage:













































Some other factors which might have been expected to be relevant in predicting beliefs were not. Age was not significant, but gender was:




Household Head











This is relevant across all socio-economic groups. In each category, there is a difference of about 20 percentage points:













Household Head











Brokerage Strategies

In addition to asking people to describe the politician’s role, people also chose who they would go to if they needed assistance. People preferred politicians over civil servants as brokers: 37 percent of the respondents would select a national-level politician and 20 percent would select a local politician (a total of 57 percent), while 43 percent would go to a local or national bureaucrat.[7] The overall dependence on the TD as preferred broker is startling. He or she is not only an important first option in obtaining help, but he is also a crucial second option. Forty-two percent of those first going to an official, and 57 percent of those who first went to a local politician would, as their second choice, then pick a TD.

Whereas class origin is a better predictor of brokerage beliefs than current occupation, it is current status, rather than origin, that determines brokerage strategies. Thirty percent of working class respondents would go to a civil servant, while 51 percent of middle class respondents would go to a civil servant. Almost half of working class respondents would to go a TD, Senator, or Minister, as compared with a quarter of the middle class respondents. Social background does not lessen these distinctions, which suggesting that those who came from working-class or rural backgrounds and now have middle class occupations have contradictory views: they expect politicians to be brokers, and yet will actually go to civil servants rather than politicians when they need assistance.

Occupation is not the sole determinant of brokerage expectations or strategies. Housing status is independently linked with variations in political perceptions: those who live in public housing tend to view politicians as brokers. While only 51% of home owners want politicians to be brokers, 69 percent of Corporation renters want brokers. If one excludes owners who live in Corporation estates (assuming this represents Corporation renters who have bought out their tenancy), the percentage of home owners who want brokers drops to 45 percent. Housing operates independently of occupational status: 58 percent of those few middle class respondents in public housing still expect brokerage, while only 48 percent of working-class respondents in private housing want brokerage.

Generally, in the inner city and the Corporation estates, most residents are likely to feel that politicians should be brokers. The dependence on the TD is highest in Corporation estates (Area 5). Here, 51 percent would first go to a TD, whereas only 28 percent of those in the other areas would follow that strategy. Of those going to a TD in a Corporation estate, 31 percent would go nowhere else if dissatisfied, as compared with only 9 percent of those going to a TD in the other areas. Thus, not only would most voters in Corporation estates go to a TD, they wouldn’t go anywhere else. It is more complicated in middle-class areas. While most residents of middle-class areas do not expect brokerage services, those who rent (whether privately or from the Corporation) do.[8] No such ambiguity exists regarding strategies for obtaining services; area of residence is the dominant factor in both working-class and middle-class areas. Inner city and Corporation estate residents choose to go to politicians, while suburban residents approach bureaucrats.

While the surveys show that people prefer politicians to bureaucrats, this may say more about dissatisfaction with bureaucrats than preference for politicians. The survey suggests that people distrust politicians’ motives, and people may prefer politicians to bureaucrats only as a lesser evil, rather than as any positive preference for politicians. After people indicated what local politicians should do, they were then asked what politicians actually did. Only 53 percent said that politicians actually engaged in policy or brokerage activities; 25 percent thought local politicians were actually just trying to make a name for themselves, and 22 percent thought politicians were just trying to get things for themselves and their friends. Those most likely to want politicians to be brokers were also most likely to distrust politicians’ motives. For example, 39 percent of semiskilled and unskilled workers (category DE) thought that politicians were just getting themselves publicity. Both the inner canal area and the Corporation area respondents felt the same (35 percent and 31 percent, respectively).

To summarize, it is clear that no single factor explains brokerage beliefs and strategies. Three factors were closely related with peoples’ preference for politicians as brokers: occupation, area of residence, and housing conditions. Although the three factors are clearly inter-related, it seems that they each are relevant, independent of either of the other two. To put it quite simply, if one is poor, renting a Corporation house, or living in the inner city or Corporation estates, one is more likely to expect politicians to be brokers, and to go to politicians rather than bureaucrats. The common thread of the three factors is dependency: the greater the need for state services, the greater the dependence on politicians as brokers. These factors tend to reinforce each other, creating two distinct groups, one which expects brokerage and one which does not. The areas in which people want brokers are likely to be those areas where individuals most need state assistance.


One of the greatest lacks in research on clientelism has been the lack of information on voter-politician interactions: what is the actual frequency of clientelist contacts? It is on this issue that so many previous studies have provided only impressions or politician’s estimates. The IPA survey asked respondents whether they had occasion to be in touch an official or politician ‘after a complaint or inquiry’. This would include people who had been active on policy issues, as well as individual benefits, but this question gives a rough indication of clientelist activity.[9]

Given all the previous discussions of brokerage and clientelism, one would expect a relatively large number of people to have contacted politicians. In the IPA survey, a large number of people indicated that they wanted politicians to be brokers, and a large number also indicated that they would go to a politician rather than an official if they had a complaint or problem. Contrary to expectation, only 17 percent of respondents indicated they had contacted politicians at one time or another. This means that less one in five people have ever had occasion to contact a politician. This is a remarkably small number, considering the widespread belief in brokerage. Regardless of people’s beliefs, it would appear that a relatively small number have ever needed to pursue that option in order to obtain state resources.

As might be expected, the poorer respondents are most likely to contact politicians. Although the DE categories comprise under 30 percent of the population, they account for almost 40 percent of the contacts which politicians have with the public:








Politician’s Contacts




39.2 =


Total Population




28.0 =



Politicians see a disproportionate number of semi and unskilled working-class voters, and markedly fewer of all other occupations.[10] If these figures are presented as a percentage of each category who contacted politician, official, or no one, the differences between each category are striking:





























The relatively high number of contacts from middle-class respondents is worthy of further investigation, since it seems to contradict expectations. The reason for this becomes clear when brokerage activity is examined in terms of housing. Contacts with politicians were most frequent in inner city and public housing areas (Corporation estates), and generally low amongst middle class private householders, with the exception of the new suburbs:



















































Here we find the reason for the high level of contacts amongst the middle-class category noted above. Of the middle-class respondents who reported contacts with politicians, about half lived in Corporation estates, and the other half[11] lived in newly built suburban estates. None of them lived in older, more established middle-class areas. In newly built middle class estates, householders have need of brokerage as they try to force builders to finish off estates, as well as force the local authority to provide community amenities.

The highest frequency of brokerage contacts came from three categories: those living in the inner city or Corporation estates (22 percent), those renting from the Corporation (24 percent), and those classified as semi-skilled or unskilled manual workers (24 percent).[12]  Brokerage exchanges are most frequent among those segments of the community who are most dependent on state assistance. Does this high level of dependence translate into votes for politicians, or support for them? Is there any link between respondents’ claimed voting behavior and their brokerage beliefs, strategies, or frequency of contacts? This is, in clientelist systems elsewhere, the crucial element: the voter repays, with his vote, the politician who has assisted him.

Respondents were asked in they had voted in the previous election. There was no correlation between brokerage beliefs and voting; in other words, people are not more or less likely to vote if they believe politicians should act as brokers. However, voting behavior correlates somewhat with brokerage strategies (though not with a significance of under .05), and very well with brokerage contacts:


                    Brokerage Contacts and Voting












Did not vote











It would seem that politicians who believe that people vote on the basis of brokerage contacts are correct.

To what extent can reports of voting be taken as accurate? According to government reports, 76.6 percent of electorate voted in the 1973 general election, and 77 percent in the 1969 general election (Stationary Office 1970, 1973). The 1969 election is the most relevant, since this was the most recent election at the time of the survey. The turnout in the Dublin area was 72.3 percent, with constituencies ranging from as low as 68.5 percent to as high as 76.5 percent. Thus, the reported figure matches actual voting statistics.


Some general deductions can be made from this material. Cultural explanations of clientelism attribute the same beliefs to all members of a society. This survey shows that brokerage beliefs and values vary within the population. Clientelist belief and practice have both a regional and class dimension. Furthermore, at least as a determinant of brokerage belief, socialization is a stronger determinant than experience. People whose father had either working-class or agricultural occupations are far more likely than those with middle-class parents to believe in brokerage. The stereotype that rural people are more likely to believe in brokerage than urban people, and working-class more than middle-class appears to be true. Moreover, these beliefs do not change, even when the respondent’s current occupation or region no longer matches his or her parent’s.

People decide who they would contact, however, on the basis of their experience rather than upbringing. That is to say, a belief that politicians should or should not be brokers does not determine brokerage strategies. Strategies also follow class and regional lines, but dependence on state resources is the common factor for all. It is those people who need something from the state, regardless of whether it be a Corporation house or a middle-class housing estate to be finished off, who would go to politicians rather than officials. Clearly, people who think they need assistance feel that they have a better chance of getting that assistance from a politician, not an official. This corresponds with studies such as Roche (1982) and Kelly (1987), suggesting that it is a lack of response from the bureaucracy which directs people to politicians.

Regarding the frequency of actual brokerage contacts, economic dependence remains the crucial determinant here as well.   Those who need Corporation housing, or those who need their housing estate finished off, or those who need social assistance are likely to have contacted politicians. Some groups, such as middle-class private house owners are also likely to have contacted officials, unlike their working-class counterparts. There is also a strong suggestion that those who have had brokerage contacts are much more likely to vote, thus validating a politician’s concern that he should provide brokerage services.

One of the most striking results of this survey is the low frequency of brokerage contacts. The figures for brokerage beliefs and strategies would led one to expect, in accord with previous ethnographies, a high frequency of brokerage contacts. A figure of less than one in five seems hardly sufficient to support a ‘clientelist political system’, as the term would normally be applied cross-culturally. The existence of such beliefs without material foundation is a puzzle that requires investigation, and is quite at variance with expectations. Previous ethnographies seem to assume that since people believe in brokerage, brokerage exists. Yet less than one in five have ever contacted a politician, even once. Even in Corporation housing estates, it is less than one in four. It would seem that clientelist exchanges do not constitute the primary mode by which state resources are allocated, they are more of an ‘addenda’ to the existing formal structure of allocations.

Should Irish politics be described as clientelist? If one used the term loosely, in a metaphoric way, one could ascribe clientelist ideology to Irish politics: voters do believe that politicians should provide brokerage, and that politicians should provide special assistance for voters. However, such a view confuses matters. As Mezey (1979) has noted, most voters throughout the world want politicians to provide special assistance. Ireland is different only in that the electoral and political system forces politicians to meet that demand. Since people who have contacted politicians are more likely to vote at election time, clinics and a reputation of ‘availability’ and ‘influence’ provide some additional marginal votes. In an electoral system like Ireland’s, where politicians are very vulnerable to the voters, such minor assistance may be significant.

Using the term ‘clientelism’ as it is usually used when studying other societies, Ireland is not clientelist. Relatively few people seem to engage in exchanges with politicians. Furthermore, the exchanges which do take place may well be once-off transactions. Even if such a transaction created an enduring voter loyalty, this would still involve less than 20 percent of the population. Thus, despite the prevalence of clientelist ideology in political life, it would confuse more than clarify if Ireland were described as clientelist. Not only does it depart from common social science usage (which means that comparison with other societies becomes impossible), it obscures the most interesting aspect of Irish politics: the lack of congruence between belief and practice. Why should the disparity between belief and practice be so great? Why such a strong belief that politicians should be brokers, even when relatively few people go to politicians?

One possible clue which the survey provides is the class and regional dimension in both brokerage belief and strategy: those from a rural or from an urban working-class background are most likely to believe in brokerage. Sixty-one percent of those with an agricultural background believe politicians should be brokers and yet have had no contacts with politicians. The figure is only thirty-seven percent for professional backgrounds, and forty-two percent for non-manual backgrounds, fifty-two percent for skilled working-class. In contrast, the figure is sixty percent for semi- and unskilled working-class. Thus, the discrepancy between belief and practice is most pronounced for those groups already identified as most likely to believe politicians should be brokers. To phrase it another way, seventy-six percent of those with agricultural backgrounds who have had no contact with politicians or officials believe, none the less, that politicians should brokers. The figure is forty-eight percent for professional, fifty-one for non-manual, sixty-one for skilled working class, and seventy-seven for semi- and unskilled. Thus, a significant number of people with agricultural or working class backgrounds believe politicians should be brokers, even though they have not needed brokerage themselves. It is possible that such groups feel they might, eventually, need assistance, but the survey does not tap that issue. Perhaps, in this area, cultural explanations deserve further investigation.

Despite the small size of the sample and the age of the survey, the IPA survey provides useful directions for future research. It reinforces views, previously expressed, that brokerage beliefs have both a class and regional dimension. Beliefs are partially cultural, but largely functional. It suggests a difference between people’s beliefs about brokerage and their own behaviour. Most importantly, it provides some indication of the level of brokerage activity in the population as a whole. Brokerage contacts are, as might be expected, more frequent amongst those who need assistance from the state. However, this seems to apply to middle-class as well as working-class voters; although this needs further investigation.

The most unexpected, and significant, result is the relatively low level of brokerage contacts among the population as a whole. The frequency of brokerage contacts is much less that rhetoric would lead one to expect. Although politicians spend a lot of time on brokerage activities, it seems clear that most people never have recourse to politicians. It is not clear whether this is because they do not need state benefits, or because they deal directly with officials and bureaucrats. What is clear from this survey is that brokerage is a latent strategy which individuals are prepared to utilize, but it is not widespread enough to actually dominate the system of resource allocation. Thus, there remains a disparity between the widespread belief in brokerage and the low level of actual brokerage activities. It seems clear, from this survey, that ‘prizes for votes’ is not a sufficient explanation, and researchers must search further for an understanding of clientelism in Ireland.




Here are some descriptions people gave of the councillors job: which do you think comes nearest to your view of what a councillor should have done?

--To intercede with the corporation for the people, to get them housing, planning permission, etc. [‘brokerage’]

--To run the city according to some overall policy which the people approved at election time. [‘policy’]

--To get a name for themselves.

--To get into power so as they could fix ‘things’ for themselves and their friends.



 Suppose you wanted to make an inquiry or complain about any service provided by the corporation to whom would you have gone before the council was dissolved?

--An Official

--A Councillor

--A TD

--A government department (Local Government, Health or other)

--A government minister

--Other (Please specify)



Categories were collapsed to official (including government department), councillor, and TD (including government minister).



Have you ever been in touch with an official or any other of the following persons after a complaint or inquiry?

a) a Councillor

b) An official

c) An ex-councillor

d) a T.D.

e) a Senator

f) A government Department

g) A government minister

h) Other

i) Never been in touch


The responses were then categorized as politician (answers a,c,d,e, or g), official (answers b and f), or none (answer i). Those answering other were excluded, and those who had contacted both politicians and officials were included under politician.




[1] All cases are unweighted, all tables have a statistical significance of at least .05.

[2]‘Professional’ includes Higher professional, Administrative and Executive, Senior salaried; ‘other non-manual’ includes Lower professional, Intermediate non-manual, Other non-manual; ‘Skilled working class’ and ‘Semi-skilled and unskilled’ include those categories, as named.

[3] The problem of deriving class categories from occupational ones is beyond the scope of this paper. For ease of terminology, these groups will be referred to as middle class, lower middle-class, skilled working class, and so on, without suggesting any automatic one-to-one correspondence between occupational and class categories.

[4] Such differences were also found by Kelly (1987).

[5] This excludes the small number who chose the other two answers of “fixing things for friends” and “getting publicity”.

[6] The category of ‘agricultural’ includes anyone who indicated that their father’s occupation had been agricultural, regardless of the size of the farm, or type of tenancy, as the respondents were not asked for more detailed information.

[7] This roughly corresponds to the RTE (1977) survey. Dublin respondents selected national politicians (33%) and local politicians (11%), for a total of 44%, while 46% chose bureaucrats.

[8] Field research has confirmed this. For example, one politician in a middle class constituency said that the small number of public housing estates in his area took up as much time as all the rest of the constituency, (Komito 1985).

[9] Respondents were given a choice amongst a number of different types of politicians and officials. The categories have been collapsed into just politician or officlal. A small number indicated that they had contacted both politicians and officials; these were classified as political contacts.

[10] Kelly (1987) reports even a greater disparity, with 80 percent coming from categories C2,D,E, and F (agricultural).

[11]Examining the 13 AB respondents who had contacts with politicians, 6 lived in newly built suburban estates of area six and 5 lived in the Corporation estates of area five

[12] Although there is, of course, a great overlap of the categories, they are not identical. Even when controlling for the other two categories, each category was still significant.



Bax, Mart  Harpstrings and Confessions: Machine-Style Politics in the Irish Republic Amsterdam: Van Gorcum, 1976

Boissevain, Jeremy Friends of Friends: Networks, Manipulators and Coalitions Oxford: Basil Blackwell, 1974

Carty, R.K Party and Parish Pump Ontario, Canada: Wilfrid Laurier University Press, 1981

Central Statistics Office Census of Population, 1971. Volume IV, occupations Dublin: Government Stationery Office, 1975

Chubb, Basil ‘Going about persecuting civil servants: the role of the Irish Parliamentary Representative’ Political Studies 10/3 (1963) 272-286

Clapham, Christopher ‘Clientelism and the State’ in Christopher Clapham (ed.) Private Patronage and Public Power  London: Frances Pinter, 1982

Collins, Neil Local Government Managers at Work Dublin: Institute of Public Administration, 1987

Eisenstadt, S.N., and Luis Roniger Patrons, clients and friends Cambridge: Cambridge University Press, 1984

Garvin, Tom ‘Theory, Culture and Fianna Fail: a review’ in Mary Kelly, Liam O’Dowd, James Wickham (eds.)  Power, Conflict and Inequality Dublin: Turoe Press, 1982

Gibbon, Peter and M.D. Higgins ‘Patronage, Tradition and Modernisation: The case of the Irish ‘Gombeenman’ Economic and Social Review 6/1 (1974) 27-44

Graziano, Luigi A Conceptual Framework for the Study of Clientelism. Western Societies Program Occasional Paper No. 2 Ithaca, New York: Center for International Studies, Cornell University, 1975

Higgins, Michael D ‘The Limits of Clientelism: Towards an Assessment of Irish Politics’ in Christopher Clapham (ed.) Private Patronage and Public Power: Political Clientelism in the Modern State  London: Frances Pinter, 1982

Hourihan, Kevin ‘Social Areas in Dublin’ Economic and Social Review 9/4 (1978) 301-328

Kelly, Valerie ‘Focus on Clients: A Reappraisal of the Effectiveness of TDs’ Interventions’ Administration 35/2 (1987) 130-151

Komito, Lee ‘Irish Clientelism: a reappraisal’ Economic and Social Review 15/3 (1984)

Komito, Lee Politics and Clientelism in urban Ireland: information, reputation, and brokerage Ph.D. thesis; Anthropology Department University of Pennsylvania. Ann Arbor, Michigan: University Microfilms, 1985

Litton, Frank ‘Dubliners and Local Government’ Leargas 5/7 1973 Dublin: Institute of Public Administration

Mezey, M.L. Comparative Legislatures Durham, North Carolina: Duke University Press, 1979

Mintz, Sidney W., and Eric R. Wolf ‘An Analysis of Ritual Co-Parenthood (Compadrazgo)’ Southwestern Journal of Anthropology 6/4 (1959) 341-368

National Economic and Social Council Dublin: Problems of Growth and Decay Report No 55 Dublin: Stationery Office, 1981

O’Connell, Declan ‘Sociological Theory and Irish Political Research’ in Mary Kelly, Liam O’Dowd, James Wickham (eds.)  Power, Conflict and Inequality Dublin: Turoe Press, 1982

Paine, Robert ‘A Theory of Patronage and Brokerage’ in Robert Paine (ed.) Patrons and Brokers in the East Artic  Memorial University of Newfoundland: Institute of Social and Economic Research, 1971

Paine, Robert Second Thoughts about Barth’s Models RAI Occasional Paper No. 32 London: Royal Anthropological Institute of Great Britain and Ireland, 1974

Pitt-Rivers, Julian The People of the Sierra Chicago: University of Chicago Press, 1961

Powell, John Duncan ‘Peasant Society and Clientelist Politics’ in Steffan Schmidt et. al., (eds.) Friends, Followers, and Factions  Berkeley: University of California Press, 1977

Radio Telefis Eireann Survey - Politics Dublin: Irish Marketing Surveys, Ltd, 1976

Raven, John, and C.T. Whelan ‘Irish Adults’ Perceptions of Their Civic Institutions and Their Own Role in Relation to Them’ in John Raven, C.T. Whelan, P.A. Pfretzschner, and D.M. Borock (eds.) Political Culture in Ireland: The Views of Two Generations Dublin: Institute of Public Administration, 1976

Reid, Ivan Social Class Differences in Britain: a sourcebook London: Open Books, 1977

Roche, Richard ‘The High Cost of Complaining Irish Style’ Journal of Irish Business and Administrative Research 4/2 (1982) 98-108

Sacks, Paul Donegal Mafia: An Irish Political Machine New Haven: Yale University Press, 1976

Scott, James C ‘Patron-Client Politics and Political Change in Southeast Asia’ in Steffan Schmidt et. al., (eds.) Friends, Followers, and Factions  Berkeley: University of California Press, 1977

Silverman, Sydel F ‘Patronage and Community-National Relationships in Central Italy’ Ethnology 4/2 (1965) 172-189

Stationary Office Election Results and Transfer of Votes in General Election (June 1969) for Nineteenth Dail Dublin: Stationary Office, 1970

Stationary Office Election Results and Transfer of Votes in General Election (February 1973) for Twentieth Dail Dublin: Stationary Office, 1973

Wolf, Eric ‘Aspects of group relations in a complex society: Mexico’ American Anthropologist 58 (1956) 1056-78