The Links between appointments, policy-making and patronage in government-supporting parties relationships

The Links between Appointments, Policy-making and Patronage in Government-supporting Parties Relationships Jean BLONDEL European University Institute Working Paper n.101 Barcelona 1995 The question of the nature of party government has only recently begun to attract serious attention. This is surprising, given the importance of parties, not just in liberal regimes, but also in many authoritarian systems. This state of affairs is now changing. Thanks in particular to the work of R.C. Katz in the 1980s, there is now greater awareness of the fact that there are many types of partygovernment and indeed that there are dimensions according to which different types of party government can be classified1. By following the direction indicated by Katz, one can begin to determine both the factors (institutional and behavioural) which account for the existence of a particular type of relationship in a country and the fields in which or the planes on which this relationship takes place. The factors which account for the existence of different types of party government are the institutional arrangements and in particular the presence or absence of a rigid separation of powers system, the party system, the party structure and especially the degree of centralisation and of cohesion of parties, and the character of the leadership, including the existence of a split between party and governmental leadership. There are three planes or fields on which the relationships take place. These relate to matters of appointment, both of ministers and of members of party elites, to the distribution of patronage, and, above all, to the elaboration and implementation of policies2. While the planes on which government-party relationships take place are distinct, they are also linked. For instance, parties press for ministerial appointments at least in part in order to achieve policy results as well as to benefit more easily from patronage; patronage and policy-making also seem likely to give rise to trade-offs. The links may not always be obvious in practice; nor is the direction of the linkage necessarily the same everywhere. Thus parties may press for ministerial appointments and for policies simultaneously but independently, without having to obtain one (appointments) in order to achieve the other (policies). Moreover, even when links exist, they may be general and not specific: while some appointments may be made in order to ensure that particular policies are carried out, others may be made simply to keep an eye on governmental policy overall. At the limit the relationship between the two elements may be so broad that the link becomes truly tenuous. Only a detailed empirical investigation can therefore reveal the precise extent of the links between the three aspects of government-supporting party relationships; as such an empirical investigation should take place in a comparative perspective, however, the aim of this paper is to examine the forms, directions, and importance of the links which can be expected to exist. THE FORMS OR TYPES OF LINKS BETWEEN APPOINTMENTS, PATRONAGE, AND POLICY-MAKING Appointments, policy-making, and patronage distribution can be linked theoretically in three different ways in a pairwise manner; in practice, however, five different types of links need to be considered as some situations have to be distinguished from others in terms of their origins and of the chronological sequence of the linkage. Links between appointments and policy-making are thus of three types: first, parties may put forward the name of ministers in order to ensure that certain policies are subsequently carried out; second, ministers may nominate or influence the nomination of top party men and women in their parties in order to reduce the chances of internal opposition within these parties; and, third, parties may be prepared not to demand seats in the government in exchange for policy concessions. The fourth type of link relates patronage to appointments. As parties and not governments ask for favours, the only linkage which can exist in this respect is the one which relates favours made by governments to appointments made in the government at the behest of the parties. The link between appointments and patronage is also of one type only: it is basically a trade-off between the government, which wishes to see given policies adopted, and parties, which wish to obtain favours. We already noticed that the links may be so general that they are almost non-existent. Assuming that they are sufficiently specific to be regarded as building a chain between parties and governments, these links can be distinguished in three ways. The first distinction concerns the actors who are involved. In one of the five cases, the exchange relationship takes place primarily between the parties which support the government: this occurs when some of these parties are more interested in seeing the policies which they advocate being carried out than in having seats in the government: they exchange one type of benefit (policies) for another (participation in the government). In the other four cases, on the other hand, the link is between the government and the supporting party or parties. Second, the links also differ in terms of who initiates the move, which may be the party or the government. When ministerial appointments are made by the party in order to achieve certain policy results or to facilitate the distribution of favours, the link is party-led; when the government influences top party appointments or when patronage is distributed in order to smooth up the passage of some governmental policies, the link is government-led. Third, the links differ in their basic character: in two cases, the link takes the form of a trade off; in the other three, one element depends on and follows sequentially the other. The link between policy-making and patronage is based on an exchange; so is the situation which occurs when a party refrains from joining the government to achieve certain policy results. On the contrary, the other forms of links establish a chain: appointments by the parties or the government lead to policy decisions and, in the case of ministerial nominations made by the parties, to patronage distribution as well. GENERAL HYPOTHESES ABOUT THE LINKS BETWEEN APPOINTMENTS, PATRONAGE, AND POLICY-MAKING On the basis of the general characteristics of institutional arrangements, party systems, party structure, and party leadership, a number of hypotheses can be formulated as to whether links will be found among the three aspects of government-supporting party relationships and how these links will be shaped. First, it is hypothesised that interparty trade-offs in which a party refrains from joining the government in exchange for certain policies will occur only in limitcases, as they are based on the assumption that parties are prepared to forego exercising power, admittedly in exchange for some policy outcomes. These cases will tend to occur where the party structure or the party system have unusual characteristics, for instance where the whole political elite is tends to be egalitarian or when when small parties or independent members give support to a government composed of representatives of one or more large parties. Second, it is hypothesised that party-led links -ministerial appointments made by the parties in order to achieve policy results or in order to obtain patronage advantages (or both)- will occur in situations in which the parties are strong enough to be able to impose their wishes on the government and yet also feel that they must remain on their guard as they are in a relatively hostile or at best highly competitive environment. Both conditions have to obtain as, if parties are so strong that they dominate the scene, they will achieve appointments, patronage, and policy results without needing to link one (policy results or patronage) to the other (appointments). These links will therefore tend to exist 1) in coalition governments with no truly dominant party, 2) in single party governments if the party is divided into clearly defined factions, or 3) in single party systems or dominant single party systems if the party views itself as surrounded by hostile ‘non’ or ‘anti-party’ elements against which it needs to be ‘vigilant’, a situation which has typically characterised many Communist systems. Conversely, such party-led links between ministerial appointments and policymaking and/or patronage will tend not to develop 1) where a single party majority government fights, in a united manner, a strong opposition party, 2) where a single party minority government relies on the external support of other parties, or 3) in presidential governments, unless the president is a rather weak compromise candidate among a number of parties. It is hypothesised that government-led links will emerge in two different kinds of situations. Links between party appointments made or proposed by the government and policy-making are expected to occur when a relatively strong government is confronted with a cohesive party, but a party which is sufficiently lively and therefore distinct from the government to be able to generate opposition to the executive within itself as, otherwise, the government would not need to influence appointments to ensure that its policies are carried out. Such a situation will arise therefore 1) in single party majority governments or 2) in coalitions in which one party is dominant or 3) in presidential systems in which there is a single, dominant, or at least cohesive majority party which is not merely the puppet of the government leader, however. Thus the link between appointments made in the party and policy-making may exist in both authoritarian and liberal systems. It exists in authoritarian systems in which a charismatic or personalised leader exercises great influence on the party but does not dominate it entirely: it therefore takes place more often when the leader inherits a party than when he creates it. This link may also exist in liberal systems in which the parties have a large popular support but where party competition has an ‘adversary’ rather than a consociational character and one party normally wins the election outright, as is typically the case in Britain and many Commonwealth countries. Victory at the polls by a party means that the whole party leadership group becomes the government: while, in a sense, the party ‘appoints’ the government in such a situation, it is more realistic to describe this case as a transfer, en bloc, of the leadership group to the government: the party leadership as such disappears2. However, as the government is anxious to see its proposals accepted, it typically uses its influence to ensure that any new party leadership subsequently put in place will support government policies: ministers will therefore endeavour to control party appointments though the link may often be general rather than specific. Finally, it is hypothesised that policy-making will tend to be linked to patronage where, on the contrary, there is substantial distance between the government and the supporting party and neither can impose its will on the other. This situation characterises many constitutional presidential systems, since these are based on separation of powers arrangements which are widely regarded as resulting in a ‘congressional party’ being distinct and indeed autonomous from a ‘presidential party’3: the main way in which the executive can then endeavour to see its initiatives accepted is through trade offs, many of which take place among policies, but some of which also take place between policies and favours: in such cases, the link can be very close, in part because neither of the partners can exercise control over the other. A similar situation can also occur in parliamentary government if the party system is rather inchoate and the executive lacks therefore a secure majority, provided this executive has a relatively autonomous power base: patronage may then be used as a means of obtaining the passage of proposals which the government initiates4. The rationale for these hypotheses needs to be presented in somewhat greater detail if one is to assess the extent to which specific links exist between appointment practices, patronage distribution, and policy-making; in many cases, as we saw, actors may not need links between the three aspects of the relationship to achieve their desired results, while, in other cases, they may too weak to be able to establish a link. The rest of this paper is thus devoted to the examination of the conditions under which links are likely to exist and to develop. POLICY SUPPORT WITHOUT PARTICIPATION Policy support without participation in the government is ostensibly special, since, in a first phase at least, it is a relationship between the parties involved in supporting the government rather than one between government and parties: the decision to participate or not has obviously to take place before the government is set up and in order to enable the government to be set up. Over time, however, the government becomes involved, as it may well -indeed almost certainly willperiodically negociate with the parties which give their support without participating in order to ensure that that support remains firm. The hypothesis is that support without participation occurs in limit-cases only, since parties can normally be expected to want to maximise power as well as influence policy-making influence. Indeed, while it has been claimed that it can be rational for parties to abstain from joining the government, it has also to be noted that it is rational for a party to want to achieve power in order to see to it that the policies which it prefers are adopted: parties should therefore seek ministerial positions because they favour certain policies5. Yet it has been shown that there are cases of ‘power-shyness’, to use K. Strom’s imagery and this ‘power-shyness’ is reflected in at least a proportion of the very large number of cases of minority governments6. These have not been studied systematically so far, almost certainly because minority governments do not fit easily within any of the theories accounting for coalitions7. These situations of ‘power-shyness’ must be regarded as exceptional, however. To begin with, not all cases of minority governments constitute real instances of trade-offs between ministerial power and policy-making. First, many minority governments are borderline majority, the government party having obtained 45 percent or more of the seats in parliament8. Full discipline and total collaboration of the other parties is required if the government is to be brought down or its policies are to be rejected: in a multi-party system, this situation is likely to be relatively rare; indeed, one might hypothesise that, the larger the number of ‘non-governmental’ parties, the greater the probability that a minority government will be able to survive. Second, small parties can be to an extent blackmailed into giving support if they are ideologically placed in such a way that they can only bring down the government or contribute to the defeat of its policies by joining with parties which are ideologically very distant from them. Finally, when the parties are small, there may well be behind-the-scenes deals involving individual legislators: such cases resemble therefore those in which a link exists between policies and patronage and which we shall examine later. Thus, the only true cases of support without participation are those where a party of substantial size sustains a government which comes markedly short of a majority. These cases are mainly concentrated in Denmark and Norway, although they have also occurred occasionally in Sweden, in Italy, and in some Commonwealth countries (Britain, Canada). In Norway and Denmark, support without participation can be regarded as rational, as K. Strom points out, since the electoral costs of joining the government are not compensated by large benefits: there is almost no patronage to be distributed and political behaviour is so consensual that even opposition parties, let alone those which support the government without joining it, share in the elaboration and development of policies in and around parliament9. Little is therefore gained by participating in the government. The characteristics of the Norwegian and Danish systems analysed by K. Strom are nonetheless exceptional, as the equation between costs and benefits is usually different. Part of that difference stems from the underlying implication that members of the party elite are not very eager to become ministers: this point result from the conclusion, drawn by K. Strom for Norway, that there is little parliamentary hierarchy in that country10; one might even go further and note that ministers may be so harassed by a legislature which interferes with the activities of the government (as the Danish legislature over European Community affairs) that, in such a case, one might exercise more power as a member of the legislature than as a member of the government. These remarks suggest that, the more egalitarian the social and political elite, the less the personal aspects of power are likely to weigh in these party decisions on participation. It seems therefore fair to regard situations of support without participation as limit-cases. They occur, first, in the context of small supporting parties, especially if these are ill-placed on the ideological continuum to mix their votes with those of the rest of the opposition. They also occur where a party gains little from participation as much is decided consensually, patronage is limited, and the political elite tends to be egalitarian. Finally, they occur where the party or parties which support the government without joining it feel that they have primarily to be concerned with their electoral future, the expectation being that the government will not remain in office for very long: a similar game can be played by the government party or parties as well, as these may blackmail the supporting parties by threatening them to make them responsible for the collapse of the government. Such a strategy can be played only where dissolution is freely in the hands of the government and where ‘strong and stable government’ is widely regarded as an important asset, these two points being characteristics of Commonwealth countries, but not so much of the Scandinavian States11. PARTY-LED LINKS: a) Ministerial appointments leading to policy-making At first sight, the most obvious way in which a link is forged between two aspects of the relationship between government and supporting parties seems to be that which connects appointments to the government to policy-making. These links appear to follow logically from the very idea of representative government, and especially of parliamentary government, modified and in a sense implemented as it is through political parties. Since parties are now regarded with justification as the institutional mechanisms which make it possible for the ‘will of the people’ to shape governmental policies, a corollary seems to be that governmental appointments should be made from within the parties with a view to ensure that party proposals are implemented. Other types of appointment arrangements seem to lead to a more indirect and therefore less certain way of seeing to it that party proposals are implemented: there is a danger that ministers might be more ‘independentminded’ as these, not being closely associated with the party, may feel less bound to follow party views. There may of course be good reasons to make appointments outside the parties: there may for instance be a need for technical experts whom the party or parties concerned may not have. Some issues may be so controversial or electorally so damaging that the party or parties supporting the government may prefer to reject the poisoned gift, a situation which resembles, at the level of given ministries, that of a party refusing power, not so much in this case in order to achieve a desired policy result, but in order to allow a policy through which is felt to be right and yet is unpopular. Such examples are -one hopesrelatively rare: it seems a contrario that a party which wishes to see a given policy adopted will want to appoint ministers who will loyally work for that policy. In the first instance at least, what therefore seems in need of explanation is the absence, rather than the presence of a specific link between ministerial appointments and policy-making. Many situations do not correspond to this straightforward scenario, however. As a matter of fact, the scenario is fully in operation only when government-building includes a degree of suspicion: for a party which appoints ministers in order to ensure the adoption of its policies must have the political muscle to appoint ministers: this is indeed why this type of link has to be regarded as party-led; but the party must also feel that it needs to have loyal representatives on the spot to see its policies implemented. When, then, do these situations occur? They occur when parties have a strong hold on at least part of the electorate and yet do not exercise enough control over the government to be fully secure. This is why this type of arrangement will not normally prevail when a united single party supports the government or in the somewhat analogous case of support given by a dominant single party, provided it is united. In such situations, the unity of the party and its complete or near-complete hold on the government combine to make it unnecessary for the party to appoint ministers with a view to achieving specific some strong executives may need and be able to make top party appointments and exercise in this way some control over party involvement in policy-making: British and Commonwealth examples show that this might be the case even in parliamentary systems, but the precise extent of the links needs to be established by close observation, as, at this point, too, the boundary is difficult to draw between a puppet party and one on which influence has to be exercised by the government. Links exist between appointments and policy-making, between appointments and patronage, between patronage and policy-making: but their incidence must be compared with the bargains and exchanges which take place within each of the three aspects of the relationships which develop between governments and supporting parties. These links must also be weighed in terms of the opportunities which many governments seem to have to shake gradually the supervision which parties wish to exert on them. Parties may be involved in ministerial appointments when the government is formed, but that government usually retains some leeway in policy-making; it is indeed also sometimes involved in party appointments as well, as for instance in Britain, France, or Germany. Parties announce policies in their electoral programmes; these are then more or less diligently followed by the governments which may also initiate other policies. Thus the links may become gradually more tenuous. The ‘zone’ in which links exist is broad, but rather vague: it may also increase or shrink over time, thus adding to the complexity of what is still, in many respects, a terra incognita of government-party relationships. Opportunity for specific and precise links to be forged between two aspects of the government-party relationship cannot therefore be truly assessed before a careful examination of the cases is undertaken. The examination will need to discover the complex meanderings of the boundaries which pass between the zones of great party strength and of great governmental strength, on the one hand, and the zone where links tend to be numerous because each of the actors needs to be able to lean on the other in order to achieve its goals. The links between appointments, patronage, and policy-making flourish where pressure rather than brute force prevails in relationships between governments and the parties which support them. It is therefore not surprising that the zone should be difficult to define; it is also likely to provide an important means of understanding better how governments interact with the political parties which sustain their efforts. WHAT APPEAR TO BE THE MOST WIDELY EXPECTED RELATIONSHIPS LINKING APPOINTMENT, PATRONAGE DISTRIBUTION, AND POLICYMAKING MECHANISMS IN GOVERNMENT-SUPPORTING PARTY RELATIONSHIPS The relationship between governments and supporting parties can take three forms. It is therefore natural to ask: in what ways are these three forms mainly expected to be linked? Does a privileged role of the parties in connection with government appointments tend to be linked to a privileged role of the parties in connection with policy-making? There seems to be at least some evidence that such a link is often made: after all, parties often claim that they wish to take power in order to be able to carry out certain policies. Yet we also know that this does not always occur: many parties send their representatives to the government and are frustrated because their desired policies are not implemented. Meanwhile, a somewhat different view has also been put forward, in particular by Strom (Comp. Pol. St., 1984: 211), as there appear to be circumstances in which a trade-off takes place between ‘seeking power’ and ‘seeking influence’ or, in our terms, between government dependence on appointments and government dependence on policy-making. The point is not so much that cabinet appointments are sought entirely for their own sake, but that they can be regarded as important in a somewhat vague manner, for instance to protect ill-defined interests of the party in the future. Moreover, the distinction between appointments and policy-making aspects helps to account for cases of minority governments in parliamentary systems, in particular where they are numerous, as in Scandinavia. Some supporting parties may well obtain that the government depends on them for appointments, at least to an extent, while others may prefer to ensure that the government depends on them for policy-making only. Two quite distinct hypotheses can be formulated on this basis and it appears valuable that they should be tested. The first suggests what might be called ‘reinforcement’ (of dependent government appointments by dependent government policy-making, for instance); the second suggests what might be called ‘compensation’, at least in part (also of dependent government appointments for dependent government policy-making, for instance). These hypotheses need to be extended to the other two ways in which governments and supporting parties relate to each other, namely autonomy and supporting party dependence. Overall we need therefore to find out whether, or to be more accurate under what circumstances, the ‘reinforcement’ hypothesis or the ‘compensation’ hypothesis will tend to prevail with respect to a) autonomous, b) supporting party-dependent, and c) government-dependent situations. Such an analysis thus suggests a number of cases which need to be investigated; it is partial, however, since it does not take into account the third element of the government-supporting party relationships, namely patronage distribution. This element is rarely referred to in analyses concerned with party government, in part, as we know, because it is rather difficult to define patronage precisely and to describe it accurately; indeed, as was pointed out earlier, there are probably instances where patronage distribution is very low and may be considered to be in practice insignificant. Yet patronage distribution has to be taken into account since the analysis would not be complete otherwise, given that it plays a large part in many countries. What part patronage distribution plays in relation to appointments and to policy-making needs to be clarified, however. Different hypotheses seem permissible in this respect. In some cases, patronage distribution is likely to increase the dependence of governments on parties further, as those parties may be able to extract patronage distribution on top of appointments and policies; in other cases, patronage is likely to constitute a form of compensation for appointments and/or for policies which the supporting parties may otherwise have opposed. Behind this distinction lies the point that patronage may be controlled by the supporting parties in some cases, while, in others, it may remain controlled by the government: consequently, the effect of the existence of patronage is likely to be different depending on whether it is the government or the supporting party which is dependent in the particular context; moreover, it may have in particular a ‘reinforcing’, or a ‘compensating’ character. ‘Reinforcement’ and ‘compensation’ relate to two possible strategies (deliberate or not) with respect to the way in which the three components combine in a given situation. We did notice, however, while analysing the nature of the these components, that two of them, patronage distribution and policymaking, could vary substantially in absolute magnitude and even, at least for patronage distribution, be almost entirely absent. This suggests the existence of another dimension of analysis, at any rate for these two components, which could be described as that of the ‘level of development’ of each component. Normally ‘incompleteness’ can be expected to concern one component only (patronage distribution or policy-making): an extreme and most peculiar case could be characterised, however, by ‘incompleteness’ with respect to both patronage distribution and policy-making: a government could stay in power only in order to be in office. It would then do very little and even distribute no or almost no patronage. Overall, however, the situations of ‘incompleteness’ with respect to one component only are manifestly the most frequent; they give rise to vaying levels of ‘reinforcement’ or of ‘compensation’. We have thus moved somewhat towards being able to examine possible differences in the manner in which the three components of governmentsupporting party relationships appear to be linked; we have in particular begun to discover the direction in which analyses can be conducted. We need now to look at this direction a little more: we shall do so by examining first the situations in which all three components ‘reinforce’ each other, then the circumstances in which ‘compensation’ is likely to take place, and finally the cases in which the relationship is ‘incomplete’ because one component is not represented or little represented. REINFORCEMENT There is reinforcement when all three components, appointments, patronage distribution, and policy-making, display the same tendency, namely tend jointly towards autonomy, or towards government dependence on the supporting parties, or towards supporting parties dependence on the government. The clearest cases of reinforcement are those which are located in what can be regarded as the areas of the three corners of the triangle, while compensation is more likely to prevail, on the contrary, in cases which are located near the centre of gravity of the triangle. With respect to reinforcement, let us therefore look in succession at the areas at the corners of the triangle which correspond to autonomy, government dependence, and supporting party dependence. As far as autonomy is concerned, we already know that it cannot be complete or that it is asymmetrical, in that governments can be truly autonomous but supporting parties cannot. Overall, these cases are likely to occur when governments and parties are neatly separated. Perhaps the most normal situation is that in which the constitution or the practice (with an elected president or an effective monarch) establishes an autonomous base for the government. A less ‘normal’ case is that of an emerging political system in which the relationship between government and supporting parties is somewhat unclear: this could occur when the party system is completely reconstructed, as happened in Eastern Europe after the fall of communism, or even only partly reconstructed, for instance after the collapse of a military regime. Yet another type of situation of this kind may be in the process of occurring in some industrialised countries and in particular in the United States, as a result of the increased distancing of parties from the executive in the political system (a distance which has of course always existed to an extent). Thus neither instances of party emergence nor instances of party distancing are so unreal that the cases of full or relative ‘autonomy’ should be entirely overlooked. What, then, should be the characteristics of all three components to correspond truly to a situation of autonomy? What is required of appointments is fairly clear and, indeed, clear on both sides: neither set of protagonists should influence the other in this respect. Examples could be monarchical governments of constitutional, but pre-parliamentary systems (e.g. contemporary Morocco), some presidential governments (such as that of the United States), and perhaps some ‘technical’ governments which are sometimes appointed in parliamentary systems. The distribution of patronage on an autonomous manner is more difficult to envisage. It can none the less exist if it takes place at the same time in two different ways. First, the government has to be in a position to distribute patronage not to the parties, but to independent supporters, to notables for instance, with a view to building support for itself where it is low or to diminishing opposition where it is strong. Meanwhile, some patronage should also be given directly to the parties for distribution by the parties themselves (perhaps including parties not supporting the government) in a routine manner and on the basis of rules which the government could not, in practice at least, really alter: legalised party financing could be considered to be a limit-case of such a routine form of patronage distribution. Autonomous or near-autonomous policy-making is more complex to describe. As policy-making often needs legislation and at least some financial backing in order to be implemented, on the one hand, and, on the other, as the implementation process entails the involvement of public servants and therefore at least the passive concurrence of the government, governments and parties seem unlikely to be truly autonomous from each other. Some cases of governmental near-autonomy may exist, however, if, for instance, party influence merely takes the form of suggestions made by one or a small number of party members -including members of the legislatureand there is no direct intervention of the parties as such. Conversely, while governments are unlikely to remain passive when bills are initiated by the legislature, they may remain neutral, for instance on conscience matters or on matters of private law; or the parties may be profoundly divided within themselves and governments may have to refrain from intervening. Here, too, this autonomy is more at the level of individual legislators than at that of the supporting parties as such; but as there is at least some influence of the parties on the legislators, one may consider these cases as examples of near party autonomy. Thus, to exist generally, autonomy on policymaking entails that there be fields in which each of the protagonists is recognised as having, de facto, a monopoly of development. Cases of full or near-autonomy in which there is true reinforcement are likely to be rare, however, especially where parties are truly alive in the society. They do exist, however, in the special circumstances of a strong executive, supported by a strong monarch, faced with relatively strong parties in parliament: historically, the German Empire at the end of the nineteenth century was such an example. Moreover, even if the model corresponds to a somewhat extreme situation, it is valuable as it helps to examine real-world situations and to ask: a) in which type of parties and party systems do appointments tend to be autonomous in both governments and supporting parties; b) when patronage is so routinised that it does not involve the government at all; and c) under what conditions and/or over what fields do policies tend to be developed relatively autonomously by governments and by supporting parties, the fields in which parties tend to be autonomous being almost certainly different from those in which governments tend to be autonomous. At this last point, however, we may come close to finding some compensating element, as there may be trade-offs between one type of policy and another. Government dependence on supporting parties may be regarded as being the natural consequence of the theory of party government and indeed as the direct expression of this theory. There is some ambivalence, however, about the extent to which parties ‘should’ control governments, as it is felt, in classic representative government, that the government ought to have some degree of leadership capability (i.e. in effect autonomy) and that the government should be responsible, not to the parties, which are regarded as being somewhat ‘unacceptable’ because they are ‘divisive’, but to parliament, which is supposed to be composed of ‘independent’ legislators (hence the maintenance of the fiction that all forms of ‘imperative mandate’ are wrong). This ambivalence suggests that we should accept a substantial relaxation of the idea of dependence: this is why we are referring here to an ‘area’ of dependence of governments on supporting parties, as in the case of the autonomous ‘area’. In terms of appointments, full government dependence on supporting parties entails that a group entitled to act in the name of the party (the nominator) chooses the ministers (the nominees) from within the party leadership and also decides which positions these ministers will occupy in the government; conversely the party leaders are not selected by the government from within the government. The group entitled to act in the name of the party may be the parliamentary party or it may be the party executive; it may indeed be a smaller body, for instance a few advisers of the party secretary or party president. Government dependence may not be as complete, however: the nominators from the party leadership may for instance nominate ministers who do not belong to this leadership. Moreover, while the ministers are selected by the party, the specific positions held may be decided by the government head or by the government collectively. Conversely, it may be that the government head chooses the ministers, in which case the nominator is not the party, but these ministers (the nominees) are all drawn from among an elite circle within the supporting parties (for instance from among senior parliamentarians of the parties). At this point one comes to a situation which can be described as fusion, at least at the time the appointments are made. This arises especially when the head of the government is the party leader: there is then so to speak transfer from the party to the government and it is difficult to speak of domination of the government by the party, particularly since, over time, the ministers often acquire independence. One important element of the assessment, however, is whether, in the reality of the appointments, the head of the government is truly independent or the party interferes somewhat in the appointments. In terms of patronage, which must be substantial if we are considering cases of genuine reinforcement, there must be dependence on the supporting parties: these determine the content and the beneficiaries of this patronage: the role of the government consists then simply in giving a formal basis to the individual patronage decisions. With respect to policy-making, the area of government dependence on supporting parties has two aspects, positive and negative. The positive aspect refers to the fact that the policies initiated and elaborated by the parties must be implemented by the government; the negative aspect refers to the fact that policies implemented should have been initiated and elaborated by the supporting parties first; cases of governmental initiative should concern only matters which are derivative or consequential on those which the supporting parties had initiated and elaborated. A difficult borderline problem is that of whether a government which is fully dependent on the supporting parties is able to act on its own initiative in privileged fields such as foreign affairs and in particular when a crisis occurs, for instance in foreign affairs but also over economic questions. The area of supporting party dependence on governments appears primarily to correspond to situations in which these parties have been set up or at least markedly expanded by a government already in office or, more commonly, by a leader who sees the need to consolidate support within the population and in particular to mobilise electoral support. These cases concern primarily rather authoritarian or ‘populist’ systems, but, as was pointed out earlier, some parties in liberal countries can fall in this category: the Gaullist party is perhaps the clearest example in postwar Western Europe. With respect to appointments, the situation in a party dependent context is the converse of the one which was found for government dependent cases. Appointments within the government are made without reference to the parties; appointments at the top of the parties are made by the government members and in particular by the government head (nominators) among government members (who are the nominees). The government leader may thus appoint some ministers to top party positions. There is then party dependence as the ministers are appointed to the party positions to ensure that the party fulfils its role of mobilisation of the population in favour of the government and its leader. Patronage can exist in a situation in which the supporting parties are dependent on the government: we noticed earlier that there may be patronage outside the framework of parties altogether and indeed that this patronage may be used to undermine parties. Conversely, patronage may exist to establish new parties and ensure that these remain controlled by the government. Admittedly, the probable long-term effect of such a development will be to strengthen the party gradually and will therefore end up in a reduction of the influence of the government; but such an effect will take time to materialise and, in the intervening period, the government can use the flow of patronage to ensure that it is distributed to those to which it wants it to be given. As for policy-making, it is indeed appreciably easier to conceptualise its development in the context of dependent parties than in the context of autonomy or of government dependence on supporting parties. The case of supporting parties dependent on governments does indeed correspond to situations in which parties are used by these governments primarily to mobilise or control the population but where the origin of policies and consequently of statutes is to be found among ministers, their entourage, and civil servants. In such a context, the supporting parties are expected to pass the bills which are presented to the legislature; these bills are likely to be accepted without amendments or with minor changes only. The supporting parties may be used by governments which need help in the context of implementation, for instance to ensure that the policies are well ‘explained’. As with patronage, variations over time are nonetheless likely to occur: supporting parties may, albeit slowly, move from a point of strict ‘obedience’ to one in which they begin to make suggestions on policy and are thus less fully dependent. Cases of supporting parties being dependent on governments are thus fairly easy to describe: they are probably typically transitional in character; they are therefore probably not really numerous as situations change gradually, but substantially. Meanwhile, autonomous policy-making and patronage distribution occur relatively rarely as well. Finally, government dependence on supporting parties has to be conceived in a relatively broad manner and include cases in which the concept of government ‘dependence’ is appreciably relaxed in relation to all three components. As these conditions are being relaxed, however, one begins also to move away somewhat from what can be regarded as true ‘reinforcement’ towards ‘compensation’, with trade-offs beginning to operate both within each component and among the components of government-supporting party relationships. COMPENSATION There is ‘compensation’ in government-supporting party relationships if one side gains in one area while the other gains in another. The types of compensation are manifestly very many; the level and extent of the compensation mechanisms are also varied. As with cases of reinforcement, the situations which are primarily considered here are those in which all three components figure prominently, that is to say where both patronage distribution and policy-making are at a high level. There can of course be compensation in situations in which one component is markedly reduced or even absent: we will examine these as examples of ‘incomplete’ relationships in the next section. The basic distinction is that between compensation which occurs within each component and that which occurs between components. Compensation within each component means that appointments, patronage distribution, and/or policy-making may to an extent be autonomous, to an extent government dependent, and/or to an extent party dependent. For instance, appointments to the government may be partly made under party influence and partly autonomously; correspondingly, appointments to the party leadership may be made partly under government influence and partly autonomously. The cases of reciprocal influence at the appointment level occur frequently, and variations can be monitored in terms of differences in the extent to which one side appears to be more influential than the other. Reciprocal influence and trade-offs in policymaking are possibly even more frequent; indeed, we already noted that examples of government and supporting party autonomy in different fields could be regarded as instances of trade-offs: the government is ‘allowed’ by the supporting parties to be involved autonomously in a field (e.g. foreign affairs) on the understanding that the supporting parties are autonomous in another field (e.g. some aspects of social welfare). Such trade-offs occur perhaps even more frequently between government-dependent and party-dependent policy developments: the supporting parties come to accept without reservation some policies which the government wishes to push forward on the understanding that the government accepts to follow these parties on other policies. Indeed, as the detailed analysis of policies can show, trade-offs of this kind occur as individual policies are elaborated: often a clause is passed reluctantly in the legislature by the MPs of the parties supporting the government on the condition that another clause, wanted by these parties but which the government originally opposed, is also accepted by the executive. The second broad type of compensation is that which takes place among the components. The supporting parties press for ministerial appointments on the understanding that the appointees will be fairly free to act in the way they wish once in the government: the ‘more noble’ way of presenting this type of arrangement is to state that the party ‘trusts’ its ‘representatives’ in the government to do what is best for the party (and presumably for the country). This ‘trust’ may be genuine, in the sense that the party recognises that it does not have the political credibility in the nation or the expertise to determine what can or should be done; presumably the party guidelines are vague enough to allow for considerable leeway: but the ‘trust’ may also be imposed on the party, if the government leader or some minister is regarded as indispensable and he or she can therefore act almost entirely as they please. These forms of ‘compensation’ must be monitored; they seem impressionistically to occur frequently. Changes may occur over time in the character of the compensation. We noted earlier that ministers may well move from being party dependent to