- 1 Literally “expert councillor”, a Prussian administrative position reached on the basis of education (...)
- 2 Literally “expert privy councillor” referring to the top official with specialist training.
1For a long time “expertise” has been considered to be the main basis of bureaucratic power in modern democracies. For Max Weber the expert, trained bureaucrat is indispensable to modern government and this indispensability leads to the “continually growing power position” of the state bureaucracy in modern politics (Weber, 1972, p. 836). “The power position of all bureaucrats rests on knowledge” (Weber, 1972, p. 854–855). The “specialist trained vortragender Rat1 is superior in technical matters to the minister” (Weber, 1972, p. 856). “There is the continual question: who will govern the bureaucratic apparatus? The domination of the apparatus by the non-expert remains only possible to a limited degree. Over the longer term the expert Fach-Geheimrat2 mostly has the upper hand in relation to the inexpert politician” (Weber, 1972, p. 128–129).
2While M. Weber makes many qualifications to these observations, and we will come back to these later, the basic idea of a conflict between the expert bureaucrat and the inexpert politician has been a staple of subsequent literature on bureaucracy. The idea of the potential conflict between “generalist” politicians and “specialist” bureaucrats had long been discussed in traditional public administration accounts of bureaucracy (Ridley, 1968; Judge, 1981). Peter Self (1977, p. 204) points to the common belief that hierarchy and specialization are incompatible, exemplified above all in Victor A. Thompson’s (1961) argument that managers are inadequately qualified “to control the specialists who work under their control”. The same basic point, that specialist knowledge gives the bureaucrat power over those generalist politicians who are meant to give them direction, has been recast in different ways over the years by different brands of social science theorising. These range from the Marxist Burnhamite view of the rise of the manager (Burnham, 1942) to the sociological power-dependence approach popular in the 1960s and 1970s (Crozier and Friedberg, 1981), in which the power position of the bureaucrat arose from the dependence of the politician on the expertise of the bureaucrat, through to the neo-Marxist arguments of the 1970s and 1980s where, according to Nicos A. Poulantzas (1975), the role of technocrats was critical in resolving conflicts and contradictions within different sectors of the bourgeoisie and capital, to the literature on corporatism (Schmitter, 1974).
3Even the currently more fashionable Foucauldian notions of “governmentality” attribute a powerful position to expertise, though the concept appears to refer to broader processes of analysis but still appears to make claims about certain kinds of professionals’ dominance in some areas of state activity. For example, with Nikolas Rose’s discussion of “psychological expertise”, “dependence of power upon a claim to rationality opens up a vast and auspicious territory that expertise – authority grounded in a claim to truthful knowledge and efficacious technique – can colonize: the role of psychologists in the legal and penal complex and in industry provides obvious examples” (Rose, 1996, p. 99). Principal-agent approaches, originating in economic thought, argue that significant problems in bureaucratic performance result from an “information asymmetry” in the relationship between the bureaucrat as agent and the politician as principal which lead to the agent’s ability to “shirk” and avoid direct control by the principal (Moe, 1984).
4Given the centrality of the issue to classical and modern bureaucracy studies, it is surprising there is not more work examining the proposition that expertise brings power and subverts hierarchies rather than generally accepting it as a likelihood or a proposition within policy-making (for an exception and a review: Brint, 1990). One of the reasons, perhaps, that expertise as a bureaucratic resource is difficult to examine is that the term “expertise” itself is not easily defined. Another problem is that working out whether a particular group or individual is powerful is hard enough, working out whether they are powerful because of one particular characteristic –here their expertise– rather than another, even more so. And then there are the large number of variables surrounding the deployment of expertise that might be expected to shape whether expertise gave the bureaucrat power in that particular circumstance: was the bureaucrat the sole source of expertise or was s/he one expert among many, did the experts agree or disagree, were the experts under contract or otherwise influenced by the bureaucrats, and were the “expert” bureaucrats also politicised, possibly using their expertise to justify the conclusions reached or desired by their political masters?
5Such considerations make it difficult to devise a set of case studies that control for the array of variables that might affect the impact of bureaucratic expertise on public policy. In this paper, I would like to approach the question of the impact of bureaucratic expertise from the perspective of evaluating three tableaux of bureaucratic expertise. Tableaux (or tableaux vivants), popular in Victorian times, were stylised poses of actors –stage as opposed to political actors– used to create or recreate scenes from works of art, history, myth or the imagination. They tended to have a didactic purpose –to bring culture or moral arguments to those that watched– before they became synonymous with nudity and erotic poses after the beginning of the xxth century and died out. The tableaux to be discussed in this paper will be described in some more detail later on, but a short description can be given here. They are based on three separate relationships between bureaucrats and experts; where bureaucrats themselves are experts, where bureaucrats have experts at their disposal and where bureaucrats are themselves subordinated to experts. The first tableau is that of the bureaucrat as expert, much as we would expect to be able to derive it from Max Weber’s account of bureaucracy, or even that of Terry Moe. The second is that of the bureaucrat as mobilizer of expertise, somebody who stands between the politicians and the experts. The third tableau is that of the bureaucrat as servant of experts.
6These tableaux will be illustrated with cases from a cross-national study covering France, Germany, Sweden, the USA, the UK and the European Union. The study focuses on 52 items of subordinate legislation (décrets and arrêtés in France). The criteria for selection and the context of the study are set out in the Appendix. The cases were not selected specifically for a study of expertise, but rather for a comparative study of how bureaucrats go about making policy. However, the cases that were selected show wide variation in the variables associated with the importance of expertise –above all how “technical” the legislation is and how far its development required some degree of specialist knowledge. While the cases are certainly not a random sample of government legislation, if expertise is associated with bureaucratic power, we would at least expect to find some evidence of the relationship in this sample of cases.
7Not all cases fit very neatly into one of the three tableaux –for instance some mobilizers of expertise are themselves experts– so any one case might correspond to more than one tableau. The tableaux are presented in descending order according to the strength of role one would expect each tableau to give to bureaucrats –the bureaucrat as expert (found in 31 of the 52 regulations), the bureaucrat as mobiliser of expertise (15 regulations) and the bureaucrat as the servant of experts (18 regulations). I will then look at the cases used to illustrate these tableaux and ask two main questions. First, how frequently does each tableau give a reasonable representation of what actually happened? Answers to this question must be treated with caution because the cases are few in number and cannot be treated as random. However, to the extent that the tableaux can be found to offer reasonable approximations to reality, we at least gain some reassurance that they help us understand the role of expertise. The second question to be posed is how much power arose from the deployment of expertise in the cases that fit each of the tableaux? Thus the method is a way of interrogating the 52 case studies derived from interviews with 96 officials involved in developing the regulations and analysis of the available documentary material –including legislative drafts, consultation documents, official and press reports. The tableaux allow one to address two sets of basic questions: what type of expertise do bureaucrats have at their disposal and what policy-making power does expertise give bureaucrats?
8Before going on to the empirical evidence a couple of reservations need to be offered. First, the interviews and other case study material refer to a specific time period, between 2006 and 2008, and they reflect conditions in the early xxist century. While there is little or no evidence from earlier periods with which to compare it, in the conclusion I will explore briefly the question of whether the conclusions reached are likely to have no validity beyond this period. Second, this paper must of necessity exclude analysis of one central source of expertise in the policy-making system: interest groups. They are excluded despite the fact that such groups play important roles in the stories of how many of the regulations in the sample came about and despite the fact that groups’ specialist or expert knowledge of policy areas and their ability to present technical cases make them of relevance to a study of expertise in the policy process. While the research gathered a significant amount of information about bureaucrats’ approach to groups, they are excluded from this paper as the material gathered for this research focussed on the bureaucratic processes of producing regulations rather than the wider political process of how groups developed and presented their cases. Some of the implications of the findings of this research for the interest group-expertise issue will be explored in the conclusion, but there is insufficient evidence to present this satisfactorily in any of the tableaux.
9The question still remains: how do we define expertise? The terminology surrounding the issue is rather loose and difficult to pin down. Expertise is not simply, as in the principal-agent formulation, a matter of one person having more information than another. Knowledge, say, of the precise number of people worldwide who have caught swine flu, does not necessarily give an official greater expertise than a politician, or indeed another official, who may not have these facts at their fingertips. Expertise is about knowledge of relationships and consequences as well as facts. We are unlikely to be able to come up with a generally accepted definition of expertise, but can offer a working definition that should suffice to explore a series of central propositions about bureaucracy and expertise. Expertise is a high level of familiarity with a body of knowledge and/or experience that is neither widely shared nor simply acquired. This broad definition allows one to regard a variety of different types of knowledge as “expertise”. Rather than seek to refine abstract definitions of expertise, let me point to four separate forms of familiarity with bodies of knowledge that bureaucrats appeared to have, or have at their disposal, in my cross-national study of 52 items of secondary legislation in six jurisdictions. Such different kinds of expertise were found on different occasions :
-
Scientific expertise refers knowledge of a set of abstract concepts, theories governing relationships between these concepts and a range of techniques to apply the insights of this body of knowledge to policy problems. Thus, for example, the economists who modelled the costs and benefits from protecting habitats of endangered species of fish in the United States were experts, as was the British veterinarian who decided which particular surgical procedure on animals was a “mutilation” and therefore prohibited, and the epidemiologists who evaluated proposed changes in the regulations governing the restrictions on the movement of birds in France during the bird flu scare also had scientific expertise;
-
Policy expertise refers to knowledge of the range of policies and instruments, current and past, proposed and enacted, governing a particular policy area as well as knowledge of how they work. Experts on this basis would include, for instance, the Swedish farmers agency official who knew the types of issue that generated litigation between the government and farmers in the administration of subsidies or the US official who knew how student exchange schemes worked and could make recommendations;
-
Process expertise refers to knowledge of the complex processes that have to be followed to ensure a proposal can be put into effect. This would include the German environment ministry official who had an understanding of the different perspective of the different ministries in the Länder with which he had to negotiate or the US official who knew how to point a Congressional inquiry in the direction of supporting a cherished idea so that his boss would order him to carry it out;
-
Instrument expertise refers to knowledge of how to put a law together, what makes a “good” law, what is permissible and what is not, what could be challenged and what is likely to stand. This would include the UK departmental lawyers who actually draft regulations or the German Ministry of Justice or French Conseil d’État officials who advise on the constitutionality and legality of proposed regulations –the influence of this kind of expertise on policy development can be very strong (Page, 2009).
10The tableaux will be examined from the perspective of the first two kinds of expertise only – scientific and policy expertise. This is in part for reasons of space, in part because these two forms of expertise are the most commonly associated with bureaucratic power and in part because process and instrument expertise were generally less easy to detect in the research.
11In this tableau the bureaucrat is the expert for a particular policy issue because s/he has some particular form of technical training relevant to the matter in hand or has particular expertise that comes of familiarity with the policy area. One might expect that bureaucracies in different countries fall into the category of those made up largely of generalists, people who have no particular training relevant to the job they do and frequently moved to different fields, and specialists who, by training or experience, are especially knowledgeable about a particular policy area. In fact, if we examine the 52 cases in the six jurisdictions and ask whether the officials directly involved in developing them had scientific or policy expertise, here defined as a technical training in the subject of the regulation or greater than five years experience in the policy field, such expertise is found in just over half of them. It must be admitted that I have used a strict definition of expertise of both kinds since it is taken to refer to detailed or technical knowledge directly relating to the matter in hand and so, for example, the official developing a highway planning programme in the EU Commission was not considered a specialist even though he had previously worked in roads but in an unrelated area. This was certainly his own evaluation of his expert status, “I did not have anything to do with [this programme] before that. I was in [a] road safety unit”.
Table 1. Scientific and policy expertise of policy bureaucrats.
|
Scientific
|
Policy
|
Neither
|
nb
|
USA
|
0a
|
10
|
0
|
10
|
Sweden
|
0
|
6
|
1
|
7
|
France
|
7
|
5
|
2
|
10
|
Germany
|
1
|
3
|
3
|
6
|
EU
|
2
|
1
|
5
|
7
|
UK
|
1
|
2
|
10
|
12
|
TOTAL
|
11
|
27
|
21
|
52
|
a See text for discussion of limitations of US data.
b Numbers sometimes less than sum of other columns as multiple codings possible.
12Table 1 presents the degree to which the policy bureaucrats directly involved in drawing up the regulations had scientific training. It shows three main points –of course only in relation to the 52 cases as we have no way of extrapolating to the whole of the bureaucracies of each country. First, that scientific training is, France excepted, somewhat unusual as a qualification for the bureaucrats handling the regulations. The two EU regulations where the bureaucrats were qualified in their fields were regulations covering veterinary medicines and fishery management; the German regulation was a financial regulation handled by an economist (Diplom Volkswirt) and the UK regulation concerned housing handled by an official with postgraduate training in planning. Second, it shows that expertise in policy is reasonably widespread –over half of regulations are written by officials with some policy expertise. Third it suggests that there are differences among our 52 regulations that might be explained by different national approaches to the treatment of expertise. In the United Kingdom, it is highly likely that officials handling the regulations will have no formal training directly relevant to the subject area of their work and will not have stayed very long on the job to develop policy expertise. Rather the approach to expertise appears to be based on the assumption that administration is a transferable skill that can be applied effectively in a host of different areas as officials tend to move from one position to the next in under five years (Page and Jenkins, 2005). This was most clearly the case with an official who described her skills as those of a “project manager”:
“I’m in the Project Delivery Unit [...] I do projects that affect three ministerial departments –CPS, Home Office and Ministry of Justice. A year ago, I was doing three separate projects –videocommunication, prison records and [performance assessment]. Now I’m just focussed on this project.”
13The European Union officials interviewed, somewhat surprisingly given the presumed “technicality” of EU measures, were similarly unspecialised. One UK official in Brussels could have been speaking of Whitehall when he explained how he came to be working on a particular regulation “Well, I’m just a bod. I have no particular unit responsibility, but I’ve been around long enough to understand how things work and this was something I could do.”
14By contrast the French regulations were most likely to be handled by some kind of specialist, whether by virtue of scientific training or policy expertise. The importance of scientific education was particularly distinctive. An official dealing with a food regulation pointed out that he had a diploma in food science (diplôme de technicien) and had worked in a similar area for decades:
“What I have learned about [doing laws] I have learned on the job. I came into the job and was titularisé. I did not do a concours but a professional examination. I’m a dinosaur. I have been here 30 years. I’ve moved a couple of times from sweet things, chocolate was one of them, to cereals as well as things to do with épiceries and vinegar.”
15A veterinarian handling an animal welfare regulation pointed out that scientific training and preparation for an administrative career can be combined in the French system:
“After you have passed the concours to study in one of the five veterinary schools in France [...] you do a concours to get into the École national du service vétérinaire in Lyon. You don’t get a juridical education but you have enough to do your work for most of the time”
16And even officials without formal fonctionnaire status tended to have formal training as well as expertise. An official responsible for the transport of hazardous materials pointed out:
“I’m not actually a fonctionnaire.I am detaché from a transport firm [...] I was a conseiller de sécurité there. I was not initially an expert for hazardous materials [...] I did the examen de conseiller de sécurité and got to know all aspects of security. I studied logistics and transport and have a diplôme d’ingénieur.”
17In the two cases where there was neither policy nor technical expertise, the regulations were handled by officials from the grands corps, one by an énarque and the other by a polytechnicien. As the énarque put it when asked about his specialisation:
“On se spécialise dans l’État. The [specialist] services [...] are the people who have the specialist knowledge. When you get very technical passages, they are the ones that take them away and dissect them –the rules concerning [a particular technical issue]. They are the ones who can see whether the textes are connected to reality.”
18The German pattern was somewhat mixed, with half of the six regulations written by officials with some expertise, mainly policy expertise. In both Sweden and the USA, the expertise that officials tended to bring to their work was policy expertise, although the reasons for each require explanation. In Sweden policy officials tend to stay a long time in the same job and this helps explain the near ubiquity of policy expertise in Sweden. As a lawyer who co-wrote a regulation with a policy official put it:
“Desk officers can have been in the job for 6 months or they can have been there for 10 years. [Mrs H] has been in the job for 10 years. The non-lawyers are here for life: they have permanent positions. She is one of these. I dealt with her and the Deputy Director of her division.”
19In the US cases the policy expertise in part resulted from the fact that officials commonly stay in the same post for several years and develop policy expertise. As an official responsible for an information security regulation put it when asked if the regulation was straightforward: “I’ll say it was straightforward. I’ve been doing these for 25 years”. In addition regulations in the US were more likely to involve more than the one, two or three people who take the lead role in writing a regulation in most of the European regulations. Many significant regulations are drafted by teams, and regulations are more likely to have involved closely officials who have experience of the policy area. Thus the US entry in Table 1 needs to be qualified. I did not interview all officials involved in the collective writing of the regulation and have relied upon the accounts of those I did interview about the expertise at the disposal of the teams. One official handling major immigration regulation, for example, pointed out:
“The workgroup had [officials from a range of departments and agencies on it]. The workgroup conceptualised the [initial draft of the rule]. [Did you show the draft to the lawyers?] You see the guy who was in the room just as you came in? He is the Head of Litigation in the Department. Those people look at it for legal vulnerabilities. They’ll also be on working groups with us and review drafts. And we need lots of help to make it all read reasonably well.”
20In this tableau does expertise give bureaucrats power? It is not possible to assess degrees of power particularly easily. Power, in this sense, cannot simply be assessed on the basis of whether an expert bureaucrat wrote a regulation differently from the way in which a non-expert might have been expected to have written it. If bureaucratic expertise gives political power we would expect to find that issues that raise controversy, whether major or minor controversies, are decided on the basis of bureaucratic expertise. Establishing the contribution of a particular piece of expert advice to the outcome of a policy is unlikely to be a precise science. The task is all the more difficult in a context of secondary legislation where the numbers of people involved are few and the process, issues and positions are frequently undocumented. Therefore one has little alternative but to place significant weight on the accounts of the process given by the officials involved in drafting them. This is not quite the same as relying on the bureaucrats’ assessments of their power in the way that an opinion survey might yield. Rather I take the accounts of the people most intimately involved in the process, follow them up through public and published records where available –on the more politically contentious issues such as the EU fisheries regulation there tends to more such evidence–and make an assessment of their contribution based on this account.
21On the basis of this kind of evidence, scientific expertise did not, in the handful of regulations where officials possessed it, put them in the driving seat in the sense that the policy outcomes reflected positions put forward by the bureaucrats with technical expertise over positions put forward by politicians, other officials or outside interests. One reason for this is that where scientific expertise is relevant for the construction of a regulation, bureaucrats are not invariably monopoly holders of expertise. Where a policy involves “technical” issues in which the bureaucrat-expert might be expected to have power, there are also likely to be other experts involved. Thus, for example, in the EU fisheries regulation where Commission officials were trained in marine biology, so too were others including the officials from some of the member states with whom they had to negotiate for the Council regulation, as well as the fisheries scientists who provided them with the estimates on which their proposals were based (see below). In the French case, similarly, the official overseeing the development of a system for assessing levels of social support required for different kinds of disability –the guide barème– was involved in a process in which teams of specialists largely drawn from outside the civil service examined the different methods that could be used to assess disabilities; the official at the Institut national de l’origine et de la qualité (INAO) responsible for changing regulations governing appellation contrôlée reacted to officials from the wine producing region who were similarly scientifically trained.
22A second reason for scientific expertise not leading to dominance in the policy process is that, as Peter Self put it, “no amount of expert evidence will [as a rule] point logically and unambiguously to a given conclusion” (Self, 1977, p. 205). As studies of the role of natural science in policy-making tend to suggest (Collingridge and Reeve, 1986; Sarewitz, 2004; Pielke, 2007), where there is policy controversy the role of science in settling it is usually very limited. A third reason is suggested by Peter Self: “scientific training often makes an expert scrupulously objective about the policy implications of his knowledge, and it is not unusual for pure scientists to take an aloof attitude towards administrative decision making” (Self, 1977, p. 207). The scientists who advised the EU Commission, as well as officials working within the Commission itself, went on record to complain that scientists were being expected to make judgements beyond the science:
“Scientists were found to be under systematic pressure to ‘inflate the natural science boundary’, by which we mean various efforts to expand the range of issues that can be resolved legitimately through the methods and investigations available to them. Fisheries scientists are being asked more and more to expand their models to deal explicitly with allocation problems, and to address problems and concepts more directly suited to the social and economic sciences, such as requests for fisheries-based rather than stock-based advice. The scientists are resisting these pressures because they prefer to stick to questions that they are well suited to answer, given their scientific and methodological training.” (Schwach et al., 2007, p. 800)
23While Peter Self (1977, p. 207) goes on to argue that scientists might also, on the other hand, be tempted to make “policy judgements which considerably exceed the relevance of their skills”, there was rather little evidence of this in this tableau. As I will argue below (4. Tableau 3), this does not mean that science never trumps political or administrative criteria in decision making, but rather that it does not do so on its own.
24The power that policy expertise gives officials appears in this tableau to be highly variable. The fact that the Swedish bureaucrats responsible for the regulations in this sample tended to have been in the same job for a long time, giving them a form of policy expertise, did not give them any noticeable opportunities to set the agenda or shape decision making in a system where political advisers and appointees have the upper hand. For example, in describing one of the most significant, and technical, regulations in the Swedish sample, the official –technically trained in a related field but not defined in Table 1 as a scientific expert– pointed out that the issue the regulation addressed:
“opened up [...] discussions between the Social Democrats [in government] and the Left [Party] and the Greens [who supported them but were not in a coalition with them]. [These lasted] for around three or four months. This was done at the political level but we put forward calculations and factual material. [...] We gave broad technical background material. But the decision was a political one.”
25In other cases the role of the bureaucrat as policy expert was decisive. In several of the US cases the bureaucrats acted as policy entrepreneurs (Kingdon, 1984). For instance, a substantial change in the “green card” arrangements for immigrants to the USA came about because officials in the Department for Labor became convinced of the need for it:
“We knew we had to do something. It is the outshot of [and earlier] regulation in which we inserted a placeholder to say we were going to come back and do something about fraud. [...] We knew for a long time there was a problem [...] This procedure [...] gets you a lawful permit to stay. We had information here that was telling us that there was fraud. The Department knew, the Administration knew. Everyone knew [...] We took the lead in conceptualising the arguments, but it was a huge legal brief. [...] When we published the draft rule we got comments back, mainly ‘you guys ought to be shot’. [...] You have to let the policy leadership people know it is contentious. We’re all big boys and girls, we have our eyes wide open and you fully brief those folk. This is big stuff, not just the Labor Secretary but he had the Administration behind him.”
26However, in this case as in the contrary Swedish case, it was not the policy expertise that was decisive in giving the officials the level of influence that they had, but rather a range of other features including the institutional characteristics of the system in which they worked, the nature of the political issue in hand as well as the level of political support and opposition surrounding the specific measure concerned. Policy expertise can certainly be relevant in understanding policy-making power, but its impact is highly contingent (Brint, 1990).
27In our second tableau we move to a reticulist (Friend et al., 1974) or nodality (Hood, 1982) conception of the power of bureaucracy as shaped by expertise. Bureaucrats have the position and sometimes also the budget to get experts to talk to them or even work for them and they form the link between expert knowledge and power. Bureaucrats might be involved in formally commissioning expert advice, say from consultants or from scientific research units and universities, or in seeking some kind of less formal consultation. Moreover, those who might be commissioned or consulted could be from inside the bureaucracy as well as outside. This could take the form, for example, of engaging economists to develop a cost-benefit analysis or even asking the opinion of someone who used to handle a particular policy matter and developed a degree of expertise in it, but who has moved to another post. Here lies the difficulty in staging this particular tableau. The informal and internal mobilisation of expertise is exceptionally hard to track, requiring knowledge of every email, meeting and phone call made to anyone who could conceivably be considered as an expert, and the research on which this paper draws simply does not contain that level of detail. In some cases the consultation of formal scientific expertise was evident, as with the development of a US regulation aimed at the safe transportation of Lithium batteries which was developed by the Pipeline and Hazardous Material Safety Administration non-scientist officials advised by scientists from the Office of Hazardous Materials Technology. In other cases it was not.
28This tableau therefore concentrates on formal or written approaches to experts, whether internal or external, for their views and advice. Moreover, since we are interested in this tableau in the ability of bureaucrats to use the expertise placed at their disposal to shape policy, we will not look at approaches to experts or expert bodies for adjudication or validation, as, for example fits the role of the Agence française de sécurité sanitaire des aliments in the French cases in the sample, in the third tableau –although it must be added that the distinction between advice and instruction on which the distinction between the two tableaux is based is not always clearcut, as will be discussed in Tableau 3.
29If we look closely at the 15 regulations in this tableau where bureaucrats commissioned or sought expert advice, perhaps the most striking feature of this mobilization is that the bureaucrats themselves often had little choice about mobilising expertise. Expertise was commissioned as a result of a procedural requirement in six cases and appeared to have little bearing on the substance of the regulation. This meant, for the most part, hiring the services of an economist to make sure that the regulatory impact statement, often a statutory requirement, was met and this had little effect on the content of the regulation. As a UK official put it:
“We finished the draft of the regulation and attached the results of the [public] consultation and it was supported by an impact assessment. Our comments were on the impact assessment. For this we passed it over to the economist. [They need to put together a statement that Better Regulation people can then approve]. When I first did these eight years ago it was a matter of sitting down yourself and working something out one afternoon –not any longer. Now you have a bureaucracy. You have to fill in the forms and there are bits you cannot leave blank, you have to use the format [and fill in all the bits on the form]. The Better Regulation people will test what you put down and see if it is up to scratch and they will come back to make sure it is.”
30In two further cases expert advice was sought as a result of a procedural requirement but had more significant impacts on the content of the regulation. In the case of an EU fisheries regulation the choice was limited. Commissioning the International Council for the Exploration of the Sea (ICES) research into fish stocks is a routine and reflects a statutory duty on the Commission to take “into account available scientific, technical and economic advice” when developing proposals for adoption by the Council for the conservation of fish stocks through the Total Allowable Catch (TAC) process (Council Regulation (EC) No. 2371/2002 of 20 December 2002). The discretion of the US official putting together a regulation protecting fish habitats was somewhat greater since the commissioning of fish biologists and economists lent significant weight to a controversial proposal that could easily have been defeated, yet such advice had to be included in any such regulation under the 1973 Endangered Species Act: any regulation designating areas as critical habitats had to use natural scientific and economic advice –designations had to be made “on the basis of the best scientific data available and after taking into consideration the economic impact, and any other relevant impact, of specifying any particular area as critical habitat.”
31While six cases were essentially procedural uses of expert advice and not much more, nine were not. In a couple of these nine the expert advice sought was rather peripheral to the development of the regulation: in a German farming regulation the official thought he would “ask an economist what it would do for the market” in case anyone raised that as an objection and in another an human resources consultant was hired to produce a “good practice” guide that was associated with the regulation. In three of the remaining seven cases where expert advice was sought and had an impact on the resulting regulation, the force of the expertise came essentially independent of the ministerial bureaucrats involved. For example in two French cases, one regulation related to the educational qualifications required to practice as an osteopath, the process that led to the legislation was dominated by a working group set up by the minister composed of medical doctors and physiotherapists and the recommendations were accepted by the minister of health largely unmediated by ministerial officials. In the case of a regulation updating the guide barème, the regulation that governs the assessment of different kinds of physical disability for eligibility for different levels of social support –separate groupes de travail and subgroups were created for each form of disability, on which were represented at least one representative of claimants. The role of the ministry bureaucrat was not to interpret the expert advice: the groups produced recommendations which they evaluated themselves and the job of the ministry officials was to translate the material into the form of a regulation. In a third French regulation the recommendations of an expert group whose advice was sought was similarly translated directly into a regulation –in designing an avian flu regulation it was important to know the locations of migrating birds:
“L’office national de la chasse et de la faune sauvage (ONCFS) came up with the list of communes that contained the wetlands where migrating birds bred, fed, rested. Here [in this office] we don’t know where they are so we had to ask them.”
- 3 However, a central problem for the scientists developing the ICES forecasts is the degree to which (...)
32This leaves just four rules where bureaucrats commissioned or requested expert advice and appeared to exercise some discretion over how the expert advice was used. In the EU fisheries regulation, one central element of discretion concerned the degree to which the Commission’s proposals for fish stocks should reflect faithfully the ICES scientific research on fish stocks, or whether they should try to produce advice which, given the European Council’s record of allowing catches well above the levels the scientific advice recommends, is more “realistic” politically.3 In practice, the decision on this particular case was made by the Commissioner who indicated at least in outline how he wanted his officials to interpret the science in this respect. This leaves three regulations where bureaucrats had substantial discretion in using the expert advice to put together a regulation. Two were UK regulations where administrators sought advice from (public sector) policy expert groups: a housing specialist from the Land Registry was engaged to write a report on problems in implementing measures designed to ease the process of buying a house (the Home Information Pack system) and this was used in part to put the case to ministers that the timetable for introducing one part of this reform might be extended; and a specialist group of statisticians working for the police service were asked to provide views on which particular police performance indicators worked well or not.
- 4 The fact that external expertise does not feature prominently in the 52 regulations also calls into (...)
33Interesting though the UK cases may be, perhaps the only really striking example of an official mobilising expertise and guiding how it is used to develop a regulation that produced a significant change was the US endangered species regulation. While in part the policy of protecting fish habitats was mandated by court decisions, the path to creating the regulation that could correspond to the mandates was highly problematic. In a significant act of administrative entrepreneurship the official concerned had to find the funds to hire the scientific and economic specialists required to write the regulation and ensure that it stood a chance of entering the Federal Register despite substantial opposition from private and public sector bodies. Thus as one looks more closely at this tableau, the picture of bureaucrats acting as mobilisers of expertise does not seem to be a picture of bureaucrats using their access to expertise, whether scientific or policy expertise, to increase their own power in the policy process. In a significant number of cases bureaucrats have little discretion over when they should consult experts or over how this advice is used. While we can find two or three examples of bureaucrats using scientific and policy expertise to help steer policy decisions, this tableau does not offer strong support for the idea that the persuasive force of expertise is commonly mobilised by bureaucrats.4
34Bureaucrats not only commission expert advice, they also receive it whether they ask for it or not. The advice can come from external interest groups or individuals, yet as has already been discussed, we must perforce exclude this source of expertise from the analysis. Within government there is a host of institutions –organisations of experts, such as advisory and other committees in the EU under the comitologie arrangements, or individuals given a role in policy-making because of their expertise, such as a Chief Medical Officer. This differs from tableau 2 because the involvement of such bodies is not a result of officials commissioning experts to help them put a policy together, but rather results from an obligation or norm that they should be involved in making policy. Such a tableau does not necessarily challenge the notion that expertise brings power to bureaucrats, rather it might be expected to confirm the general principle that expertise is related to power since it refers to circumstances where bureaucrats’ own expertise –or lack of it– is overshadowed by the expertise of these groups at the heart of policy.
35Such involvement by expert groups within government could be found in 18 cases. One of these produced no reaction from the advisory committee so such committees played a role in 17 cases. The French regulations were distinctive in two respects. First, the reliance on expert advisory bodies was more marked. In seven of the ten French regulations “expert” or specialist advisory institutions were involved in making a regulation and a large portion of the cases of such influence found in the whole study were from France. In three cases the role of the specialist group was to approve a text that was largely derived without their direct intervention. A housing finance regulation had to be submitted to the Conseil supérieur des HLM:
“They just give us advice –we are not compelled to follow their opinion. Some things we have to submit to them by law, we also submit [other] things [...] In this case the consultation was mandatory [...] If it is negative then we have to see whether it is politically important that we don’t make a concession and we then publish their dissenting opinion and stick with what we want or [if we can make a concession] we try and get something that the Conseil supérieur will accept.”
36The Conseil raised no objection in this case. In another case the Commission de cosmétologie of the Agence française de sécurité sanitaire des produits de santé (AFSSAPS) was asked to approve a regulation implementing an EU law on the composition of cosmetics –it did, and without comment– and the Conseil national consultatif des personnes handicapées approved the guide barème regulation without comment (the status of such advisory groups as “experts” will be discussed below).
37In four further French cases such specialist or expert bodies had a significant impact on the development of the regulation in a role not confined to approval. The Agence française de sécurité sanitaire des aliments (AFSSA) was asked to adjudicate in a dispute between the ministry of health and the ministry of the economy over a proposed regulation ending the statute-based regime for controlling the quality of soup and moving to an advisory code. At issue was the possibility that the industry might use the opportunity to increase levels of salt, and the interministerial conflict was resolved in favour of the ministry of the economy and the code of practice regime. The AFSSA also offered an opinion on how quarantine and movement of livestock arrangements during avian flu alerts should be structured. As an official described the process:
“Normally the first step is to [...] try to write a projet d’arrêté that you then give to the AFSSA for its opinion. We did not give them a projet de texte –we asked a set of questions: what would be the different measures needed for different levels of risk from avian flu? Please tell us what are the different risks posed by captive birds when there is avian flu of different levels of risk [see AFSSA opinion of 12 September 2006]. They give us an opinion. This is done in writing and only in writing. It is a very formal process. We have to invoke the AFSSA in writing –the saisine comes though the directeur who has to approve it. The avis comes back down the hierarchy to me. [...] If the AFSSA says ‘we recommend this’ we generally do it. You have to have a very, very, very, very good reason to ignore it.”
38In two further cases –a regulation governing permitted added sugar levels for a particular group of wines and the movement of hazardous materials on the Rhine– specialist groups initiated and shaped the regulation; the Organisme de défense et de gestion (ODG) for the wine growing area concerned and a committee of the Commission centrale pour la navigation du Rhin (CCNR).
39The fact that some of these bodies fuse representational and expert advisory roles is a second distinctive feature of the French cases: expertise is closely entwined with the representative composition of the “expert” body. Take for example the CCNR working group, as one official described it:
“There are five countries on the CCNR, you get people from the sociétés de transport, transport people, people from the industry. They meet four times a year for three or four days. There are about twenty or twenty-five people –you have chemists there, representatives of the sociétés de classification, representatives of the manufacturers, charbon, chimie– and of the transport companies. Our discussions are aimed at amending the regulations to improve safety and develop the regulation, and we do that every two years. All the rules are changed on dangerous materials every two years.”
40The ODGs represent wine growers and producers in the region. The Commission de cosmétologie of the AFSSAPS mixes expertise and representation in its composition as it includes four ministerial representatives, including one each from the ministries of competition and industry, fifteen “personnalités choisies par le ministre chargé de la santé, en raison de leur compétence en matière de produits cosmétiques”, three “personnalités scientifiques nommées par le ministre chargé de la santé, exerçant dans l’industrie des produits cosmétiques ou la représentant, à titre consultative” and a consumer representative from the “Conseil national de la consommation, à titre consultative”. While the steering group that developed the osteopath regulations was “expert” as it had medical doctors and physiotherapists on it, this was certainly not the view expressed on the banners of the osteopaths demonstrating against the results of their work in Paris in December 2006 (“Ostéos manipulés par les médecins et les kinés” and “Medecin, kiné, ostéo ne font pas le même boulot”).
41Do expert bodies have influence –whether over the bureaucrats who pay attention to their advice or over the whole policy process– because of their expertise? In the cases in the sample where the expert body had influence, whether in France or elsewhere, it is more accurate to say that the status of the expert body gives it influence rather than the expertise on its own. In the UK case where the ministry had to decide which particular procedures performed on farm animals, such as docking tails or clipping wings, were to be permitted on which animals and under what circumstances, the ministerial official summarised:
“You rely on the expertise of your [Veterinary Officer]. [Fred] went and visited a duck farm and saw them do the tagging and made an assessment based on his expertise and assessed the welfare costs and welfare benefits. He suggested it should only be allowed for specific purposes and for 36 hours after hatching.”
- 5 In conjunction with the Federal Aviation Administration, though there were disagreements between th (...)
42This advice was decisive. The role of the ODG, the CCNR, the AFSSAPS and the AFSSA in the French regulations came as a result not directly of the quality of the advice or the expertise of the members of the body, but because the routines of policy-making in these areas meant that such bodies were deferred to either by law or convention. In the US, the National Transportation Safety Board’s (NSTB) report on a fire in the goods transport part of Los Angeles International Airport in 1999 had a significant impact. It led the Pipeline and Hazardous Material Safety Administration (PHMSA)5 to develop a regulation increasing the precautions for transporting lithium batteries by air. Yet while not mandatory for the PHMSA to follow NSTB recommendations, the NTSB regime for air safety required some action or some explanation from the PHMSA even though, ultimately, the NTSB file on this accident remained open –indicating that not all its concerns had been addressed– after the regulation had been passed.
43The legal status of the expert advice in one of the EU cases also protected those developing the policy from strong political intervention. One powerful member state wanted to change the regulation and the official concerned managed to deflect its criticisms by invoking the authority of the expert committee on which the member state concerned was represented:
“I had all my comments with me and could give good answers to the question [raised by the member state concerned]. And I could even offer that I had asked the Committee again to confirm [that I was right] and they had said ‘yes’. I could show that I took it seriously, etc. The proposal was agreed to unanimously in the Standing Committee, even the [representative from the member state concerned] agreed.”
44The number of regulations included in the study and how they were selected must lead us to treat the generaliseability of the research with some caution. However, the proposition that expertise brings power is so central to a range of approaches to understanding policy-making that the weakness of the supporting evidence from this study for that proposition is somewhat surprising. Levels of expertise among bureaucrats are variable. But even when officials who have scientific or policy expertise, or both, are involved, whether their views shape the resulting policy does not depend upon their expertise alone, and expertise does not guarantee influence, or even on its own make their greater influence more likely. As commissioners of expert advice and analysis, officials are somewhat constrained by the norms and rules governing how this advice and analysis is sought and used. It is certainly possible for a bureaucratic entrepreneur to seek out and use commissioned expert advice to help strengthen the case for a potentially vulnerable policy, but it is not a particularly noticeable or common tool used by bureaucratic entrepreneurs, at least in the 52 cases examined here. Of course, bureaucrats advocating or suggesting particular policy options, or presenting a range of policy options, are generally likely to offer factual, technical or specialist evidence to support their briefs, and those political leaders who choose, approve or sanction particular regulations are likely to want to be reassured where necessary that their decisions are “evidence based”. Moreover, it is quite possible that officials serve as intermediaries between experts and policy makers –a form of mediation not explored in this paper. Nevertheless, there is little to support the view that technical or scientific evidence is particularly decisive or offers bureaucrats power.
45If the ability to argue the merits and demerits of particular policy options using evidence gives bureaucrats power, then that power is more accurately described as deriving from the division of labour: bureaucrats have as part of their responsibility, part of their job, to be concerned about issues with which few other people, whether inside government or not, are familiar. To work up policy options, or even to justify decisions delegated to you as a bureaucrat, does not invariably require technical knowledge or even knowledge and experience of past issues and developments in that particular policy area. As Peter Self (1977; p. 201) puts it:
“An effective adviser needs the ability to draw conclusions from complex masses of data which seem intelligible and realistic to the ultimate decision makers. It does not follow, though, that this gift of translation is coupled with high intellectual capacity in the sphere of knowledge which is being translated, for these two kinds of ability are logically and psychologically separable.”
46This is neither an argument for the principle of “generalism” nor an argument that it makes no difference to the quality of policy whether it is made by someone who knows what they are talking about or by someone who comes to the issue as a novice. Rather it is to suggest that expertise, whether scientific or policy expertise, does not noticeably confer any special benefits in terms of influencing the policy process on those that have it. Policy expertise rather than scientific expertise was more likely, especially in the United States, to place officials in an active role in proposing and developing legislation, but even here the contrast with Sweden, where policy expertise is not noticeably associated with policy influence, suggests that the reasons for such policy influence have more to do with the wider pattern of bureaucrat-politician relations –in Sweden political advisers have a more direct role in initiating and developing administrative regulations than in the United States– than the possession of expertise.
- 6 Not all regulations come under this notice regime.
47The evidence has, however, shown that “expert” advice can be given a privileged position in the policy process in the form of expert advisers whose advice carries weight, expert bodies that potentially exercise a veto over policy proposals and specialist committees that can require bureaucracies to develop policies. In this case, however, it is the status of the expert institutions rather than the content of their expertise as such that accounts for this influence. While it was not possible to offer a detailed tableau of interest group expertise and its influence in the policy process, it is likely to be affected by this issue of status. The French experience of specialist advisory bodies given powers by statute or convention to initiate and shape policies, and their mixed “specialist” and representational composition, suggests that the status of such committees can offer groups powerful routes to policy influence. In a similar way, it is frequently argued that the mandated procedures surrounding the generation of administrative rules in the United States, above all those deriving from the Administrative Procedure Act (APA), make interest group consultation over regulations6 particularly powerful (West, 2004). In the US, group consultation is part of a framework that requires that agencies address the objections of interest groups and proposals can be –and have been– effectively vetoed by the Office of Management and Budget where such objections are not addressed, whether by pointing out why they are unfounded or why they cannot or should not be met. This procedure can place substantial weight on technical and factual argument, and it is the APA regime rather than the force of expertise, leading policy makers to change their minds, that is at the heart of how this regime works.
48Is the finding here, that expertise might not confer the power on bureaucrats that we might expect, one that is limited to the period in which the research was conducted? It might be possible to think of an era when expertise was more important and the period covered in this paper happens to be one when it has become less so. Of course, since the findings all refer to the same period and cannot be compared to earlier periods, it is impossible to say with any certainty. It could well be that examining bureaucracy in different eras would produce a different set of conclusions –in time of war technical innovation could become a high governmental priority and give greater power to experts, or the same might be found in periods where new technologies of service delivery are being developed. The glimpses that we have of the operation of bureaucracies in earlier times do little to encourage the idea that there was a “golden age” when expertise served to provide a basis for bureaucratic power. For example, Kingsley’s (1944, p. 174–175) study of the British civil service in the 1930 s argued that:
“The status of the expert has always been low in England. He has always occupied a subordinate position in the social hierarchy. Experts in the law and the military or naval strategy have enjoyed a unique position. But this is only because they have been identified with the ruling classes, and their ranks filled in part from them. No similar stamp of approval has been placed on other specialties. It is not surprising, therefore, to discover the expert is... regarded without enthusiasm by his administrative colleagues.”
49Evidence from an earlier period that technical expertise is not a route to influence within the higher levels of public and private organisations is also found, albeit in a different theoretical context, in the work of Gouldner (1954) in the United States, and as we shall see below, even Max Weber did not see expertise as a necessary guarantor of political power. Moreover, several of the cases discussed in this paper –for instance the EU limits in fishing– lend in principle, great importance to scientific expertise, yet this does not appear in practice to give experts strong policy-making power. There was little indication that “technical” issues, that is to say, regulations where knowledge of a body of scientific techniques or principles is required to understand their impact, gave bureaucrats power in the policy process. If policy-making sectors that might be considered to favour expertise-based power turn out to do no such thing, then it is quite plausible that the proposition that there were eras in which scientists or other technical experts working as bureaucrats might have been expected to be powerful could also be illusory. This observation should not be taken to suggest that expertise more generally does not bring policy-making power as the research concentrated on bureaucratic roles rather than the wider world of policy expertise through, among others, advisory bodies, interest groups and hired consultants.
50If this argument that expertise does not automatically equate to power in the world of bureaucratic politics is correct, where does this leave the range of theoretical approaches which make the assumption it does? Perhaps one of the reasons for the tenacity of the view that expertise offers bureaucrats power is that some of the classical statements of the nature of bureaucratic power –such as the work of James Burnham (1942) and above all M. Weber (1972)– gave it such emphasis. While there is little purpose in revisiting Burnhamite theory in this respect, since its central premise of adapting Marxian analyses of development have little currency any longer, how far do these conclusions undermine Weberian approaches which remain central to a wide range of contemporary understandings of the role of bureaucracy? Not much is lost from the integrity of Weberian theory if one simply argues that Weber might have overstated the centrality of the Prussian system of education, training and recruitment to modern bureaucratic systems. The rationality of bureaucracy and the predictability and calculability of government actions which make bureaucracy indispensable for the development of modern capitalism is hardly affected by the degree to which bureaucrats are “experts”. While M. Weber does make great play about bureaucracy being Herrschaft kraft Wissens, he also points out elsewhere that an expert, technically trained officialdom is not a sine qua non of bureaucracy: “Technical examinations [...] are not an indispensable concomitant of bureaucratisation. Bureaucracy in France, England and America have long managed to live without it substantially or completely...” (Weber, 1972, p. 756). The observation that the relationship between expertise and bureaucratic power is at best remote and indirect does, however, lead us to reconsider the assumptions that have been conventionally made about the bases of bureaucratic power. Thus, in principal-agent, speak it is less likely to be what the bureaucrat knows, information asymmetry, that shapes politicians’ ability to control what goes on in bureaucratic organizations, but the opportunity for, and motivation of, politicians to take an interest in what is done in their name. It is mistaken to take the popular adage “knowledge is power” for a social science law as far as bureaucratic policy-making is concerned.