‘Governance’ is a protean term.Footnote 2 In one definition it can simply refer to processes of government, whatever form they take. More recently, it has come to be used to denote ways of governing that are non-hierarchical and involve networks of actors, both public and private, determining policy through negotiation, bargaining and participation. One way of taming this promiscuous usage is to think of ‘governance’ as defining a spectrum of meaning. On the one hand, policy-making may incline towards the hierarchical and authoritative; on the other, it may incline towards the participatory and negotiated, the choice between the two depending on the issues involved, the purposes of the actors and the resources those actors bring to the decision. Even when we consider processes located at the negotiating end of the spectrum – for example industrial self-regulation through voluntary agreements – there are still important questions about how far bargaining and persuasion in voluntary agreements depend upon the ‘shadow of hierarchy’ in which the prospect of authoritative intervention in the background affects decision-making through anticipated reactions.
Examples of new modes of governance include regulatory networks negotiating standards, for example in respect of environmental protection or the operation of financial regulation. Agreements emerging from such processes need not take the form of hard law, but instead comprise codes of conduct and voluntary agreements. Even when decisions do lead to law, the informal processes of negotiation leading to legislative decision may well be decisive in substantive terms, setting the terms of the policy that legislators eventually adopt. An example of negotiated standard-setting in this sense would be the Auto/Oil Project of the European Commission, which was a tripartite research project between the oil and automobile industries and the Commission to find the most effective and efficient ways of reducing vehicle emissions, although in that case the negotiating process only weakly foreshadowed the legislative process.Footnote 3 Other examples at the European level include negotiations on energy efficiency, road safety, conditions for temporary workers and financial and cyber fraud.Footnote 4
A second version of new modes of governance includes policy coordination through benchmarking and comparative analysis of policy performance. The most obvious example of governance in this mode is the open method of coordination in the European Union, by which member states agree to submit data to peer review and analysis, with the aim of improving policy effectiveness. This is a soft policy instrument that has been developed both for treaty-based sectors like economic and monetary integration, in the Economic Policy Guidelines, for example, but also in areas of social and educational policy where treaty-based competence is weak. In these processes, member states set common policy guidelines, which are incorporated into national action plans, and performance against those action plans is subject to review by the Commission and a high level Council-Commission committee, with the aim of fostering best practice.Footnote 5
As well as these examples, the term ‘new modes of governance’ can also refer to the delegation by legislatures to independent regulatory authorities of the responsibility to set standards and define the terms of compliance to those standards by industry and other private actors. The most obvious example is the responsibility of central banks, including the European Central Bank, to set interest rate policy, but independent agencies are found across many countries in such areas as environmental protection, occupational health and safety, public procurement and the administration of public services such as health and education. Although some of these agencies go back a number of years, for example in relation to the regulation of broadcasting in the UK or competition policy in Germany, Mark Thatcher has shown that the 1980s and 1990s saw a significant wave of innovation of such bodies in Britain, France, Germany and Italy.Footnote 6
What binds these various examples together is the principle that decision-making under new modes of governance escapes direct control by democratically elected representatives. Important policy decisions either may not require legislation at all, since policy aims can be accomplished through voluntary agreement, or, where legislation is eventually required, the terms and character of that legislation may be set by the negotiating processes that have preceded it and so we have legislation in the shadow of negotiation. Even where the legislative framework mandates independent agencies to undertake certain prescribed tasks in specific ways, the realities of implementation often mean that what happens in practice depends upon negotiated agreement among key actors.
Interest in new modes of governance is stimulated by the observation that non-hierarchical methods of decision-making have become more important both nationally and in the European Union. It is, of course, easy to overstate both the newness of new governance and its quantitative importance. In relation to newness, many political systems have relied over several years upon self-regulation to secure the governance of the professions, ranging from medicine, through law to academia. In these cases, legislation has been empowering, bestowing upon self-regulatory organizations the authority to make rules, rather than coercive or constraining. Similarly, corporatist arrangements in the making of economic policy relied in their heyday upon negotiated agreement between the state and the social partners. In terms of quantitative importance, although networked negotiation has risen in importance, this is still compatible with the bulk of policy initiatives taking a traditional legislative form, a fact identified by Héritier in relation to EU policy initiatives, where out of 926 measures only 99 could be strictly construed as new modes of governance.Footnote 7
European integration is both a cause and an occasion of new modes of governance. European integration has been a cause not least because the Union lacks control of significant public expenditure, so that its repertoire of instruments is limited to regulation (in the broad sense) preceded by negotiation, together with voluntary agreements and the fostering of networks among specialist bodies and civil society organizations. In particular, where it is difficult to secure treaty-based competence in particular areas – even when there is genuine problem interdependence – new modes of governance may be the only way to achieve action. Thus, the White Paper on European governance and successor initiatives have been important in fostering the interest in the new modes.Footnote 8 Beate Kohler-Koch has drawn attention to the confluence of interests and currents of thought that went into the Commission's emphasis upon network governance in the White Paper, including the need to avoid treaty changes, the wish to avoid treading on the toes of other important institutional actors, especially the Council, the crisis of the Santer Commission and the then-fashionable thinking about the value of civil society organizations in a context where parliamentary democracy appeared to be losing its legitimacy.Footnote 9 However, the move towards new modes of governance, though associated with European integration, cannot always be ascribed to it. There are general pressures of public administration across all developed democracies, leading to the creation of independent agencies and regulatory authorities, reflecting theories of new public management.
New modes of governance raise serious questions of political legitimacy, since the principal source of legitimacy in a democratic society is the procedurally based legitimacy that comes with elections and elected representatives giving parliamentary support to government policies and programmes. To the extent to which new modes of governance remove important decisions from the sphere of representative control, they appear to lack procedural legitimacy. This general worry is augmented by a number of other concerns. While being participatory, in that decisions are made through persuasion and negotiation, new modes of governance may be exclusionary in other respects, since those selected to participate may only be a sub-set of relevantly affected interests and opinion. Such a defective participatory structure may lead to biased outcomes that do not reflect a fair and reasonable balance among competing interests and social values. Policy processes and policy goals may be poorly or vaguely specified in new modes of governance, losing sight of the transparency that is supposed to be the hallmark of a democracy. Doubts have also been raised as to whether the supposed cost advantages of new modes of governance, avoiding the time and burden of legislation, really do exist when the accounting is done correctly.
Although these new modes of governance are not unique to the European Union, it might be said that the legitimacy problems to which they give rise exacerbate more general concerns about political legitimacy in the EU. Within the EU, negotiated rule-making involves not only actors in horizontal networks of governance but also agreement through the vertical relations of member states together with their domestic constituents. The complexity of multilevel governance is enhanced by the combination of these horizontal and vertical features, with consequent potential for lack of transparency. For this reason, although such modes of governance raise questions of political accountability in general, the significance of those questions is heightened within the EU. Indeed, in some ways, the importance and extensive use of new modes of governance in the EU is a major element in the perception that the EU suffers a democratic deficit.
Héritier and Lehmkuhl note that these problems of legitimacy are in effect problems of political accountability.Footnote 10 They concede that new modes of governance may enhance technical efficiency, enabling those with expertise to shape policies. On the other hand, the specialized character of expertise, as well as the institutionalization of a narrow range of interests, means that actors will be insensitive to spillover or external effects, where adopting a policy in one sector will have adverse effects on a policy problem in another sector. In this respect, Hériter and Lehmkuhl contrast new modes of governance with conventional decision-making in the democratic nation-state, where the choice among competing policy objectives falls to elected representatives accountable to an electorate, so that trade-offs between interests and values can be determined in a way that is consistent with a unitary concept of the public interest. They ask whether new modes of governance may not need the shadow of democratic hierarchy to be legitimate as well as the institutionalization of other forms of interest-balancing, including multi-stakeholder representation and control through the public sphere and civil society. In short, from their point of view, the task is to define the institutional settings within which the obligations of political accountability can be discharged.
In this article, I consider this critique of new modes of governance in the light of the principle of political accountability. Clearly, a proper understanding of the critique involves clarifying the nature of political accountability and its rationale as a standard of evaluation for a political system. I begin by questioning the extent to which new modes of governance can be legitimated by being more clearly under the shadow of a democratic hierarchy, particularly if elections are conceived as the principal instrument of accountability. The problem, I suggest, is not simply institutional, but also conceptual. Conceiving accountability purely in terms of elections ignores the deliberative dimensions of political accountability, dimensions that can only be understood once we introduce the idea of public reason, in which policy-making is thought of as the exercise of practical reasoning within a collective context. Party competition in the democratic nation-state can be seen as one way by which obligations of political accountability are discharged, but we should not conflate the normative ideal with the extent of its practical realization. Once we see the importance of public reason, we might even see that the practices of EU policy-making offer the prospect of democratic gain as well as the danger of democratic loss.
In this way, I hope to show that studying new modes of governance in the EU is not only important in its own right, but can also improve our understanding of important political principles and values. Institutions and practices encode political values, but we should not make the mistake of identifying any particular set of institutional arrangements with the values encoded. To take an extreme example, political leaders facing ostracism in ancient Athens were subject to a particular form of political accountability; modern politicians facing re-election are subject to a different form of political accountability. No doubt in changing the institutional form there has been a change in the character of political accountability between the two cases, but there is a recognizable sense in which answerability to constituents is a feature both of ostracism and election, even if the consequences of failure in the second case are less severe.Footnote 11 In short, political values need to be studied in their institutional embodiment, but should not be identified with that institutional embodiment.
ACCOUNTABILITY AND ELECTORAL COMPETITION
In its broadest sense, political accountability means that those exercising political authority have to give reasons for the public action they undertake. Accountability is usually thought of as comprising two elements: explanation and sanction. It has been expressed as follows: ‘accountability is a relationship between an actor and a forum, in which the actor has an obligation to explain and justify his or her conduct, the forum can pose questions and pass judgment, and the actor may face consequences’.Footnote 12 Thus, the first element of accountability is that those who hold the reins of political power in a society have to give an account – explain and justify – their use of that power in terms that are in principle acceptable to the members of society or their representatives. Within a democracy, the political authorities cannot appeal for their legitimacy just to their having decided a policy. They have to act in ways and for purposes that can be publicly defended. To be sure, these defences may well be controversial, but that is simply because the condition of politics – the need for common action under conditions of political disagreement – means that no policy will receive unanimous consent. There is, nonetheless, a difference between the use of political authority with no attempt to be accountable and a use of political authority that involves explaining and justifying what is done.
This leads to the second element of political accountability, namely the opportunity for sanction if the explanation is judged by the forum to be inadequate. The most obvious example of such sanctions within democracies is elections. Democratic institutions tie those wielding power to the need to secure a winning share of the vote, or at least a share of the vote that potentially makes elected representatives eligible for public office and the exercise of political power. This protective function of the electoral connection, theorized by James Mill, is supposed to have a number of consequences.Footnote 13 It provides an incentive for those seeking office to explain what they will do with political power if they succeed in achieving it, and an incentive for those who have held office retrospectively to justify their use of political power. Thus, in setting up a competition between those seeking office, it forces rival candidates to campaign against one another in such a way that there is accountability to the electorate.
As well as sanctioning devices, providing some form of political control by the electorate on the conduct of the government, elections can also be seen as authorizing devices, mandating an elected body of representatives to undertake political action. In common with the logic of representation generally, these analytically distinct elements of authorization and sanction combine to form a particular conception of representation.Footnote 14 For example, Manin, Przeworski and Stokes distinguish between a mandate view of representation and an accountability view and define the latter as the claim that elections serve to hold governments responsible for their past actions, but they go on to say that mandate and accountability can be combined, since electors can use information about the past to choose a better government for the future.Footnote 15 The conception here is that accountability is instrumental to authorization, linked through the judgements of electors.
Within the member states in the European Union elections provide the forum in which alternative political programmes are presented within competing ideological frameworks drawing upon the thinking of the main party families in Europe. Elections are not simply authorizing devices (though they are certainly that); they also provide an incentive for parties to adopt programmes and positions that are responsive to popular preferences in a context in which parties explain the rationale of the positions they adopt. Competition for public office is supposed to provide the incentive for parties to explain the position they hold on salient issues of the day. The primary role of citizens within electoral democracy is through their votes to judge the plausibility of the claims made by competing political parties.
How might new modes of governance fit into these institutional arrangements? One obvious possibility is that the authority given to actors in policy networks and independent regulatory agencies should be seen as a form of delegated authority. Parliaments stand in a relationship of accountability to the electorate and actors in new modes of government stand in a relationship of accountability to parliaments. The policy task then is to ensure that the accountability of actors in new modes of governance to parliament is rendered transparent and effective. New modes of governance become politically legitimate to the extent to which they are nested in institutions underwritten by the democratic authority of parliaments. If parliaments were able to exercise greater control over the decisions that emerge from new modes of governance, those decisions would have enhanced political legitimacy by virtue of the fact that elected representatives were themselves subject to requirements of political accountability in elections.
This line of accountability from actors in new governance to parliaments to the electorate looks more persuasive on paper than it is in practice. One problem is that accountability induces blame-shifting and blame-avoidance behaviour by office-holders, so that it may be impossible for even well-informed members of the electorate to know how well or badly a particular group of actors has performed. For example, a natural division of role between politicians and regulatory agencies is between the policy responsibilities of politicians and the operational responsibilities of agencies. However, the distinction between operational and policy responsibility is not clear cut, and so an elector may not know whether it is the policy or its implementation that is at fault, and therefore whether it is the politicians or the agency that is liable to blame.
This identification problem is made worse by the fact that many policy issues – particularly in fields like financial regulation, biotechnology development, environmental protection and energy production – have highly technical components, the understanding of which is essential to good policy-making. The ability of politicians to engage with such issues is variable. There are politicians who are able to specialize, and in parliamentary systems in which committee structures are strong, this specialization can be honed to strengthen accountability. However, the incentive for many politicians is to have a shallow knowledge of several of topics rather than deep knowledge of a few, and hence there will be an asymmetry of information between actors in new governance, with their expertise, and the politicians to whom they are supposedly accountable. Indeed, one of the original justifications for new modes of governance appeals to the problems to which specialist sectoral knowledge gives rise.
In addition, electoral competition has its own logic, which may prevent sufficient political attention from being given to particular sectors of policy. Electoral competition defines an agenda of conflict that may organize important issues of public concern out of the political agenda. For roughly 25 years in the period between 1945 and 1970, this happened to environmental issues, as mainstream political parties competed with one another while sharing a broadly productivist set of assumptions in which there was a neglect of environmental damage and the destruction of natural habitats and landscapes. Even if the electorate approves of a party's or party coalition's performance in office in general, this does not mean that adequate attention is being given to particular issues, particularly when accountability for performance in a particular sector has broader implications.
To see the underlying problem here, it is useful to take the most favourable interpretation of the claim that electoral competition is a way of stimulating political responsiveness by politicians to their electorates. Researchers in the Comparative Manifestos Project have argued that, on the basis of their comparative evidence, party competition and the movement of parties over time can be understood primarily in terms of a one-dimensional representation on a left–right spectrum.Footnote 16 In the light of these findings, Ian Budge and colleagues have argued for a median mandate theory of elections.Footnote 17 The core of this claim is that the electoral connection secures a close correspondence between the ideological position of the median voter on the left–right spectrum and the ideological position of winning parties or winning coalitions in the legislature. Moreover, legislative composition tracks ideological change in the electorate over time, so that elections can be seen as devices by which mandates are granted or withdrawn. A particularly striking feature of this result is that these findings obtain across a variety of electoral systems, although proportional systems generally have a better correspondence than majoritarian (properly ‘pluritarian’) ones.Footnote 18
Let us grant for the sake of argument the empirical validity of these findings. Does it follow that elections are efficient instruments of political accountability? If we were to define political accountability as median correspondence in a one-dimensional ideological or issue space, then elections would be efficient instruments. Median correspondence need not even imply that political parties and candidates explain and justify their actions, rather than simply announcing (what they take to be) a winning position. So, median correspondence does not satisfy the procedural and conceptual requirements that public reasoning involve logical and evidential validity as well as a public orientation. The upshot of this argument is that median correspondence bears only a contingent relationship to the answerability required of political accountability.
One reply at this point is to say that, whatever its defects in terms of answerability, party competition does provide a mechanism for dealing with the spillover effects from one policy sector to another, which Héritier and Lehmkuhl see as central to the problem of democratic accountability and new governance.Footnote 19 Indeed, one might see party programmes and positioning in ideological space as a way of dealing with such problems. An ideological commitment is not just a commitment to a particular package of policies. Rather, it is a statement of principles about the orientation that will be brought to dealing with the full range of policies that governments confront, including the spillovers. From this point of view, the electoral connection makes a valuable contribution to accountability, since it authorizes elected representatives to deal with spillovers.
However, there are limits to the extent to which the control of spillovers from one policy sector to another can be dealt with by legislative means. There are certainly areas of public policy, of which environmental regulation is the most obvious, in which the management of spillover is central, since environmental problems arise as the unintended by-products of otherwise legitimate activities, including trade, agriculture and transport.Footnote 20 One way of dealing with such problems is to seek to integrate environmental concerns into a wide range of public policies, thus hoping to deal with the causes of environmental problems rather than simply remedying or mitigating their effects. However, greater parliamentary control is unlikely to be a favoured instrument in this context. Reviewing the evidence of environmental policy integration in 30 OECD countries since the 1970s, Jacob, Volkery and Lenschow describe administrative instruments as being the favoured tools. The reorganization of parliamentary oversight is not featured.Footnote 21
In short, elections are necessarily general in scope. To the extent to which the electoral connection secures accountability it does it through a judgement on the overall performance of governments and political parties. To construe accountability purely as electoral competition is to neglect the role of evidence, argument and persuasion in the policy process, which is necessarily sector specific.Footnote 22 Unbundling political accountability in respect of the rationale of public decisions requires a notion of political accountability in terms of public reason and public reasoning.
POLITICAL ACCOUNTABILITY AND PUBLIC REASON
I have already noted that, in its broadest sense, accountability means that the political authorities have to give reasons for their public action and that there is a difference between the use of political authority with no attempt to be accountable and a use of political authority that involves explaining and justifying what is done. Political accountability thus involves some form of reasoning. As reasoning about public action, it can be described as public reasoning and its conduct is an exercise in public reason.
In using the notions of public reason and public reasoning, I hope to avoid theoretically controversial matters as much as possible. That is to say, the terms are used to identify the practices that are associated with those in authority offering explanations and justifications for the way in which political power is used. On this account, political accountability emerges from practical reasoning, where practical reasoning is distinguished by its being action-guiding rather than an inference to a credible belief about how things are. Whereas a theoretical use of inference yields a conclusion from a body of evidence, a practical use of reason yields a determination to a particular course of action. Political accountability is therefore the task of laying out the chain of reasoning intended to lead from the recognition of a policy problem to a collective course of action. In this sense, political accountability involves the use of public reason, understood as reasoning in the public domain about the actions the polity is to undertake. This definition is descriptive. Issues about how well or badly public reasoning is conducted is a matter for evaluation in the light of explicit standards.
However, there is a potential theoretical controversy buried in this descriptive account. There would be no point in examining public reasoning if reasoning and discourse did not matter in politics and explanations and justifications associated with political accountability were simply cheap talk. Many schools of political analysis – ranging from Pareto, through Namier, to contemporary rational choice theory – have contested the claim that institutionalized reasoning matters in politics, preferring to see political discourse merely as an instrumental device deployed in the pursuit of interests. There is not enough space here to contest this claim properly, so I shall merely state a position and say that, in my opinion, Quentin Skinner is right to point out that in so far as legitimation is an important activity in politics, the discourse by which political programmes and projects are legitimated will place limits on the range of such programmes and projects that powerful political actors can adopt.Footnote 23 Similarly, to see democratic decision-making as captured solely by the aggregation of preferences into a social choice function is to underplay the role of practical reasoning in shaping those preferences and the susceptibility of that practical reasoning to be influenced by discussion and deliberation in the political process. This is not to say that there is no aggregative moment in decision-making – clearly there is in many situations – but it is to say that we misleadingly abstract from an important element of political choice if we ignore the public reasoning as a practice of political accountability.
This discursive conception of political life has received considerable impetus in recent years from theories of ‘deliberative democracy’.Footnote 24 However, deliberative democratic theory echoes traditions of political thought with a long and distinguished pedigree. John Stuart Mill, for example, thought of parliament as a congress of opinions, in which all sections of society were able to voice their view and be challenged in discussion.Footnote 25 Sir Ernest Barker argued that when faced with the need for a common course of action, people pooled their minds and put forward points of view for others to consider.Footnote 26 Within these traditions the notion of public reasoning is central.
There is a useful distinction between the ‘deliberative’ and the ‘democracy’ parts of deliberative democracy. To focus on the deliberative part is to understand that the public reasoning that underlies political accountability has to be a form of reasoning. That is to say, public reasoning has to be more than the uttering of sounds or marks on paper or a computer screen. The sounds and marks have to have meaning, they have to be linked in ways that makes their sequences intelligible, and it has to be possible to make reliable inferences from the sounds or marks. In short, public reasoning has to conform to minimum standards of meaningfulness, inference and evidence. Chambers has drawn attention to this aspect of public reason, labelling it the ‘Socratic’ as distinct from the ‘democratic’ element, drawing attention to the way in which debate requires people to define their position clearly, respond to counter-arguments and state openly the principles and chains of argument to which they are appealing.Footnote 27
Although this Socratic element of deliberation is important in its own right, it can be linked to the basic principles of democracy if we see democracy as being a form of governance combining the principles of political equality, common interests and the fallibility of human judgement.Footnote 28 This last feature, the fallibility of judgement, is particularly important. Within a framework of practical reasoning, both decision premises and the weighing of relevant considerations need to be adequate. Bounded rationality means that no one individual will be in a position to gather together the relevant information and evidence to make a publicly defensible position. As John Stuart Mill pointed out, the case against benevolent despotism is as much an epistemic one as a prudential one. The benevolent despot would need to be ‘informed correctly, in considerable detail, of the conduct and working of every branch of administration’, and such are the faculties and energies required that we could not imagine anyone consenting to the office ‘unless as a refuge from intolerable evils, and a transitional preparation for something beyond’.Footnote 29 Conjoined with the assumption of human fallibility and bounded rationality, then, the Socratic element in public reasoning becomes an element of democratic practice as the public discussion of policy issues.
Regarded purely descriptively, public reasoning of this sort does not need to bear an intrinsic relationship to any particular set of political values. One could imagine an elite in society immune from any serious political contestation publicly setting out an account of how government policy would bolster their own privileged position, addressing their account to their political clients, with the publicity merely being a way of adding insult to injury for the rest of the population. However, the point of institutionalizing public reasoning in a form accountable to the vast bulk of the population is to place constraints on the specification of interests that can be acceptably advanced as relevant to political decision. From this aspect, public reasoning is reasoning from a public point of view, in the sense that it involves the good or interest of the public. Thus, the democratic component of public reasoning is to be found in the claim that the protection and advancement of common interests is itself a constitutive ideal of democracy.
This element of public reasoning can be related to the ideal of constitutional, as well as democratic, government.Footnote 30 Constitutional government is government in accordance with the principles of the rule of law. The rule of law requires that authoritative rules governing social life should have certain properties, most notably that they should be non-arbitrary, prospective and impersonal.Footnote 31 Non-arbitrariness, according to the rule of law, is primarily intended to restrict the discretion of the public authorities in the making and enforcing of legislation and policy, though it can also be seen as a way of preventing unauthorized actors usurping otherwise legitimate power to their own ends. Thus, taking the idea of non-arbitrary government seriously means more than that public policies should avoid logical inconsistency or whimsy. It also means that public policy should be given a rationale consistent with an underlying conception of the public good or the public interest, an idea that has been discussed in the revival of republican political theory.Footnote 32
Under these constraints, public reasoning will need to be conducted in such a way that participants show how the point of view they advance is consistent with a defensible conception of the public good. This is not to say that they cannot be an appeal to partial or sectional interests. The public good does not have to be conceived in unitary or homogeneous terms. However, any appeal to particular or sectional interests has to show how those interests, distinct though they may be, have a legitimate claim upon others joined in social cooperation.
Common interests and fallibility come together once we realize that within a society of any degree of complexity, the public point of view will be given plural definitions. There will be competing conceptions of the public interest and of the justice or injustice of distributing costs and benefits in particular ways. These differences are a result of the ‘reasonable pluralism’ that Rawls characterizes.Footnote 33 The free exercise of human reason will typically bring about a variety of judgements as to the good and its means of attainment. Indeed, it can be argued that one of the crucial rationales for taking public reasoning and deliberation seriously is that these differences of judgement need to be pooled and examined within a well-functioning democracy.
Political accountability as public reasoning, therefore, yields two standards of evaluation for a political order or mode of governance. The first derives from the Socratic component and is a standard of political understanding. According to this standard, public reasoning needs to be conducted in such a way that the reasoning can be tested as to its intellectual robustness. This in turn requires a public examination of the evidence that is used in forming decision premises, the inferences derived from that evidence and the way in which relevant considerations are balanced. The second standard is one that relates to the public orientation of a decision process. Are interests being weighed in a form that is consistent with a defensible conception of the public good, or are private interests being allowed to usurp public purposes? How far are relevant opinions and values allowed a say in the process of governance? With these two tests in place, we can now return to the institutional and policy issues raised by new modes of governance.
NEW MODES OF GOVERNANCE
In order to address the institutional and policy questions, consider a hypothetical democratic debate about new modes of governance. Let us assume that the parties to the debate are well informed and public spirited, as well as being aware of the general circumstances of politics, in particular that any society is a mixture of cooperation and conflict where differences of viewpoint have to be resolved under conditions of bounded rationality and constrained generosity.Footnote 34 What arguments could validly be offered in such a debate for permitting decision-making on important areas of public policy by private actors, voluntary agreements or independent agencies?
In many ways the answer to this question is simple. The reasons that could be offered in such a debate are just the reasons that have been offered for adopting new modes of governance in the policy debate in general. New governance enables problems to be dealt with more quickly than legislation, particularly in political systems that are characterized by a joint decision trap or multiple veto players. Voluntary cooperation by the bulk of relevant agents is in any case a necessary condition of successful legislation, because the imposition of standards and norms of behaviour on citizens and corporate actors requires their cooperation given the asymmetry of information between rule-makers and those subject to the rules. And in a democratic society, in which consent is an important value, there should in any case be a preference for voluntary and decentralized measures, in line with the principle of subsidiarity, over the centralized use of authority.
The difficulty, then, is not stating the arguments in favour of voluntary modes of governance; rather the problem is that the members of the democratic assembly require assurance that these benefits will be secured, and that the conditions under which voluntary governance will operate do not create a departure from the public interest. Since, as Andreas Føllesdal discusses in his article, democratic citizens can be conceived as conditional compliers, it makes sense to consider what are the conditions that would secure reasonable consent to the delegation of political authority by such citizens.Footnote 35 Two particular conditions are relevant here: replicability and openness. The test of justified public reasoning by non-elected actors can be partly met if it can be shown that their policy reasoning is both replicable and open to relevant consideration and points of view.Footnote 36
One test of public reason is given by the following question: for any putative piece of public reasoning, how likely is it that any member of the public with sufficient skill and time could follow the train of reasoning that led to a practical policy conclusion without finding logical fault, as distinct from differing in assumptions? The key thought here is that any case of public reasoning ought to be replicable, and in a democracy anyone – in principle – ought to be able to replicate it.
Note that the test is not that someone following the train of reasoning should find it acceptable. That would be a strong test and the Rawlsian claim of the inherent plurality of viewpoints that obtains under conditions of the burdens of judgement means that it cannot in general be met. Given the inherent plurality entailed by human reasoning, we should not expect any particular piece of reasoning to secure widespread consensus. Different individuals may still follow the train of reasoning, without agreeing on it. What is necessary is that they should be able to follow it as a train of reasoning. That is, they ought to be able to say: given those premises and that body of evidence, I can see how someone might come to this conclusion (even though it is not necessarily the conclusion I would come to myself).
In this respect, the pluralism of a democracy and the distinction between explanation and sanction are closely linked. Pluralism enables citizens to say that they appreciate that there is a train of reasoning which they can follow in the sense that they can see how, given a certain set of assumptions, there is an argument that leads from those assumptions to a particular practical determination. However, citizens who think in this way are not committed to the view that any chain of reasoning for which they can appreciate the logic is, in fact, a chain of reasoning to which they should assent. The sanctioning element of accountability then turns on how close in their thinking citizens are to a promulgated set of arguments for a political position. One's political opponents may make sense in their own terms, even if they do not make sense in yours, but it is not a trivial matter to ascertain that they do make sense in their own terms.
The test of replicability embodies a weak form of the principle of political equality, understood as the claim that no one person's judgement in a democracy should be disregarded merely because of who that person is. However, it does not follow from this approach that any citizen can in fact follow the train of reasoning. The political division of labour, the economic division of labour bestowing different skills on different people and the inevitably time-consuming nature of public reasoning all mean that the vast majority of people in a democracy are dependent upon specialist public officials and others, like civil society organizations, to undertake their public reasoning on their behalf. In these circumstances, the test of replicability becomes a test of assurance. What assurance could I have, as a citizen, that the train of reasoning that is being followed is one that I could trace if I had the time?
One traditional answer in a democracy is that one could have this assurance, provided that the representatives you elected were sufficiently vigilant of your interests and sufficiently adept at following processes of public reasoning. As we have already seen, there are grounds to question this assumption, and there are perhaps a number of reasons why in a post-industrial society a role for representatives in substituting for citizens in respect of public reason is not plausible. The alternative traditional answer is some form of Madisonian republicanism, in which the system of checks and balances serves to protect public reasoning from partial and one-sided interests. In other words, we do not look to agents representing us in public reason but we look to the whole system of institutionalized and informal debate for the processes of accountability. This leads to the second test of democratic public reason, namely openness.
New modes of governance involve processes of interest group discussion and bargaining similar to functional representation in nation-states. The clearest example of institutionalized functional representation is that of the representation of labour and capital in corporatist institutions. In their ideal-typical form corporatist institutions allowed representatives of labour and capital to bargain with one another about investment, wage restraint and more general priorities for the economy. Corporatism can also be understood in a wider sense of the representatives of corporate interests, most notably business and agriculture, being given special status and rights of consultation on matters that concern them explicitly. We should remember that corporatism, even at the level of the nation-state, always attracted criticism from those concerned with democratic values. In particular, there were those who asserted that its institutions usurped the role of parliaments in the determining of the national interest and excluded certain political interests – most notably those concerned with environmental protection – that should have been properly included.
Looking at new modes of governance in the light of this discussion, their problem is not that functional representation cannot conform to requirements of public reasoning. It is rather that those who are not parties to the bargain find it difficult to have assurance that their legitimate interests and points of view are being sufficiently taken into account. The issue is not whether new modes of governance allow interested parties to bargain with one another to mutual advantage; it is whether this bargained advantage is bought at the expense of the public interest.
We can see the way that these issues have been played out in discussions on governance in the EU. One example of this discursive turn can be seen in the Commission White Paper on European governance.Footnote 37 One of the central goals of the White Paper was to put more effective and transparent consultation at the heart of policy-making. The White Paper suggested that what was needed was a reinforced culture of consultation and dialogue. This would involve greater transparency about how consultations are run and a code of conduct that would set minimum standards on how consultation should be managed. One theme in the White Paper is that consultation with civil society organizations, through transparent and well-structured procedures, may go some way towards rectifying the democratic deficit.Footnote 38
Since the White Paper the Commission has taken a number of steps to improve consultation and social dialogue. The Europa website makes plain how the public and interested parties can comment on proposals and policy issues. The European Transparency Initiative is concerned with ways of improving the quality of the Commission's dialogue with civil society groups. There is a Code of Conduct for Interest Representatives, as well as a Commission communication on the minimum standards and principles of consultation.
How well do these initiatives meet the standards that would be implied by political accountability viewed as a process of public reasoning? The answer is that the measures do better on the test of openness than on the test of replicability. To be sure, the publicly available documents on consultation show the positions that different civil society groups have adopted, but they do not impose any obligations, or contain any procedure, by which the reasonableness of submissions could be evaluated, for example in terms of requirements to cite evidence that is publicly assessable and in a form which others could replicate. There is not even a requirement to say whether a submission is intended to rest upon intersubjectively assessable evidence or is merely the expression of an opinion or a viewpoint. An improvement in public reasoning would require more demanding tests on the submissions and arguments that could be advanced, not to prevent arguments being put but to provide some indication of the weight of evidence they were supposed to bear.
CONCLUSION
In this article I have tried to show that if we understand political accountability in the light of the idea of public reason, we have a distinct way of thinking about the legitimacy problems raised by new modes of governance, whereas if we restrict our understanding of political accountability to the sort of responsiveness to citizens secured through elections, then a proper appraisal of new modes of governance will elude us. In democratic societies, public policy, by definition, needs to exhibit some dependence on public opinion, but this is not to say that we can identify a transmission belt moving from a statement of popular preferences to a statement of what policies should be undertaken. This is partly a matter of sectoral expertise and the differential salience of issues, but it is also because practical deliberation is largely concerned with filling out the details of goals that are specified in abstract and general terms. When citizens in a democracy want clean air and clean water, they do not specify their wants in terms of environmental quality standards, even when the specification of such standards is an integral part of achieving the goal.
Specificity is also important in terms of our understanding of the public interest. One appeal in the idea that greater parliamentary oversight is the key to legitimation is the thought that parliaments can apply a unitary conception of the public interest, and this thought has some attraction if we think of party competition in a democracy as being over whole party programmes, the elements of which fit together in a coherent account of how government policy is to be conducted. Yet, the long-standing literature on bounded rationality should caution us against thinking that there can be a well-specified unitary account of the public interest. Instead, what we should look for are sectorally specific procedures and practices that respect the constraints of the public interest, treating the notion of the public interest or common good as a limit rather than a goal. Such a view is consistent with the key democratic values of political equality and the protection of common goods provided those practices and procedures are sufficiently robust in terms of their intellectual justification and openness to the pluralism of viewpoints that is an inherent, and desirable, feature of any democratic society.
