The Ethics of Political Alliance

Abstract

Usually pictured in relations of opposition, political parties are sometimes inclined to make alliances. This article examines the ethical questions such arrangements give rise to. It considers first the formal characteristics of an alliance as a distinctive form of association, moving on to examine what reasons for alliance are good reasons. Intrinsic arguments that invoke epistemic or democratic criteria, and instrumental arguments that cite areas of shared programme or imperfect institutions, are weighed in turn, with the latter judged to be more consistent with the partisan ethos. The final section examines the normative standards to which alliances should be held once formed.

‘[P]rogressive alliances are now essential, not just because that is the only way we can beat the Tories, but because that is the way we will make better decisions and take more of the country with us. Frankly, I want to be in government with Caroline Lucas, not against her – and certainly not in permanent opposition’.

Clive Lewis, MP, British Labour Party ¹

As the fortunes of political parties wax and wane, new lines of political conflict arise, and with them new possibilities for co-operation. Challenged by a powerful new rival, long-standing adversaries may find reasons to co-ordinate, now aware of what they have in common. Faced with a newly salient set of public concerns, parties may seek partners to bolster their position, or to secure their political breakthrough. Electoral realignment, and the institutional rigidities it can expose, creates fertile conditions for the pursuit of alliance.

But what normative considerations might guide such relations? On what grounds is it reasonable for parties to make common cause, and what standards should such arrangements be held to? The ethics of political alliance is a topic largely unstudied. While the political theory of partisanship is growing, its concerns to date have been elsewhere: on the norms of adversarialism on the one hand, on intraparty ties on the other. ² Political science has a wealth of material on electoral pacts, yet has tended to focus overwhelmingly on just one manifestation of alliance – the government coalition – and has made normative considerations a secondary concern, generally restricted to questions of voter choice.

This neglect of the normativity of alliance is problematic on several counts. Pacts are for one thing quite common, across a range of political settings. ³ The forms they take may vary, but a marginal phenomenon they are not. Moreover, they are typically significant events in the life of a party, as times of great possibility, also of risk. They promise to augment a party’s capacity to effect change, yet threaten to dilute what it stands for. And beyond those directly involved, alliances have systemic effects for the electoral field. Both for those on the inside and the outside, they are generally consequential and contentious. An analysis of the normativity of alliance has the potential to extend our theories of democracy in an important new direction, thereby contributing to the political theory of institutions. ⁴ Equally, it may be revealing in an empirical sense. Only by reflecting systematically on the arguments that alliance can support will we grasp the real-world disagreements that decisions to enter, maintain or exit such arrangements tend to produce.

The article begins by examining the concept of alliance and its normative dimensions. It distinguishes between the externalities and internalities of alliance relations, focusing the discussion on the latter – on the implications of alliance for those directly involved. The following sections identify two kinds of argument that may be advanced to justify alliance participation – one that treats it as an intrinsic good, another that defends it for instrumental reasons. Based on an analysis of such arguments in more than one form, the suggestion is that alliances can be justified on instrumental grounds, and only these. In the final section, the article moves to consider the relations between allies once their pact is established, examining the norms by which arrangements of this kind may be circumscribed.

As the opening quotation suggests, Britain is one country in which the prospect of alliance formation is a live question. With increasing frequency, the idea of a ‘progressive alliance’ has been proposed by partisans of the left and centre-left, for reasons that mix opposition to one party in particular with notions of the intrinsic appeal of partnership among like-minded parties. Similar discussions have arisen in other European countries, including Spain, France, Germany, Norway and Sweden, in the context of the fragmentation of bipartisan politics, increasingly frequent grand coalitions, emerging challenger parties, and / or the strengthening of the far Right. ⁵ Alliances are also a recurrent theme in non-Western democracies, notably in Argentina and India. The article seeks to bring to light some of the normative considerations in play, and in so doing advance a constructive dialogue between the political science and political theory of partisanship. The analysis should further shed light on the more general question of the normativity of alliance, relevant to contexts beyond party democracy.

ALLIANCE, AND POLITICAL ALLIANCE

As a form of collective action, an alliance has a number of distinguishing features. Given that these structural properties afford some normative possibilities and not others, our analysis must begin with an overview. ⁶

Prime among the defining characteristics of an alliance is that its constituent parts are pre-existing groups. An alliance is an arrangement of collective entities that predate the formation of the alliance. Depending on the numbers involved, the relations may be bilateral or multilateral, but in all cases the parties to the pact come together as collectives. ⁷ An alliance cannot be adequately grasped in terms of the individuals that populate it: it is first and foremost an association of associations.

This composite structure marks an alliance not only at the time of its foundation but is something the arrangement is expected to maintain, since the relationship is intended to be reversible. In contrast to a merger, an alliance should preserve the distinct identity of the associations that comprise it. It rests on the possibility that the units forming it can be disaggregated at a future date: the parts must be divisible. The timescale may be clearly specified, or the alliance may be indefinite in duration, but there is a common understanding that the decision to come together should be possible to undo. This is a function of the fact that each party to an alliance is understood to be involved because it has chosen to be. The possibility of exiting the arrangement and reverting to the independence previously enjoyed expresses the voluntary basis of the association.

The intended reversibility of an alliance is expressed in its institutional structure. As a compound of existing units that is intended to preserve them intact, an alliance involves a plurality of decision-making procedures. While new ones may be introduced to govern the alliance, these will exist in parallel with whatever institutional forms preceded it. One of the challenges faced by the agents of an alliance is balancing their pursuit of a common agenda with the maintenance of distinct institutions, including related mechanisms of participation and accountability.

As an association of associations, an alliance entails second-order ties. The individuals involved have a primary set of commitments corresponding to the groups that form the alliance. The nature of these primary commitments we may bracket for now, but it is crucial that the terms of an alliance are constrained by a prior set of associative ties. The question of consistency between the two becomes a persistent theme in the life of an alliance. The reasons for entering the pact, and the actions pursued within it, require ongoing reconciliation with the terms of each primary association.

An alliance is furthermore a form of co-operation that is non-clandestine. Unlike a conspiracy or a cartel, the parties to an alliance make no effort to conceal their co-operation. ⁸ Typically they formally announce the formation of the alliance, since public recognition of the existence of the pact may be intended to contribute to the ends for which it was formed. It may be expected to change the public estimation of the agents involved, and dissuade others from acting in certain ways. But even if expressed more informally, as an ongoing intention to co-operate, it is an overt arrangement nonetheless. An important implication is that an alliance always contains an element of risk for those pursuing it, since its existence cannot simply be denied when reactions to it are unfavourable. An alliance is an object of scrutiny, and an unpopular alliance can deprive its members of support.

An alliance, then, is an association through which two or more collective agents commit to some form of public co-ordination more or less extended in time, while retaining their separate identities both as moral associations and institutional entities. Allies are less close than those we might call political friends, because their relation is indirect, being mediated by the primary associations of which each is a part. ⁹ Their ties are also looser given the expectation that, however long in duration, they are ultimately impermanent. If friends are forever, allies are for the time being. Conversely, allies are more than simply peers, since they have common goals they pursue together, and of which others will be aware due to the public nature of their relationship.

These formal features of alliance find electoral expression in the form of pacts between political parties. Involving co-ordination between two or more parties for an extended period, these are a feature of democratic politics in many institutional settings. These practices take place on a spectrum ranging from mere non-aggression (for example, two parties choosing not to campaign against each other on certain issues) to the development of shared programmes for government. ¹⁰ In the relevant research in comparative politics, the study of alliances has tended to be equated with the study of ruling coalitions. ¹¹ More recently, however, greater attention has been paid to pacts formed prior to elections, between parties that may or may not be in government. ¹² This broadening of study usefully reminds us that pacts are possible at all points of the electoral cycle. ¹³

If we ask what normative questions are raised by alliance, existing scholarship offers few answers. Certainly this is true as regards the ethics of political alliance. Pacts between parties have tended to be approached as a problem of bargaining. ¹⁴ In line with a general focus on coalition-formation, to the extent that normative issues have been raised in the scholarship these have centred on the informational question of how far post-electoral alliances could reasonably be inferred from pre-electoral behaviour. ¹⁵ ‘Could voters know what they were voting for?’ is taken as the crucial normative question. As important as this is as a democratic indicator, pacts raise challenging questions even when such criteria of information are satisfied.

Emerging scholarship on the philosophy of partisanship has so far added little in this area. The questions raised by alliances are not reducible to those of partisan association more generally. The formal features of alliance, notably its status as an association of associations intended to be divisible at any moment, give it a character distinct from the party. While interparty relations have received some attention, including questions of partisan compromise, these have been framed as a problem of adversarialism. ¹⁶ Such work considers the normative constraints on how parties oppose each other, not those on how they co-operate. The intermediary status of allies, as distinct from both friends and adversaries, is still to be analysed by theorists of partisanship.

There is also limited guidance to be taken from other fields. The ethics of military alliance – that is, between states in conflict – has received surprisingly little scholarly attention. Though alliances are a widespread phenomenon of interstate politics, international relations (IR) theory rarely dwells on the questions of principle they raise. Alliances are instead approached as a problem of order, with the focus on how they may be a source of stability in an interstate system governed by the balance of power. ¹⁷ Research in the field of business ethics tends to be equally thin in this area, notwithstanding the occasional suggestive analysis. ¹⁸ And it is not just a question of volume of course, since the considerations arising in these cognate fields are not necessarily transferable to the analysis of political alliance. When states (or firms) make alliances, this is typically understood as a question of reconciling interests. In the case of political parties, normative commitments ¹⁹ – the principled goals a party uses to define itself – are at stake. As many have observed, finding agreement between competing normative commitments is very different from finding it on competing interests. ²⁰ It follows that the formation of alliances between parties cannot be approached as a situation of bargaining like any other.

As a first step in organizing how we think of the normative stakes of alliance, it is useful to distinguish between the externalities and internalities arising. The former concern the costs and benefits of an alliance for those who did not choose to incur them by direct involvement. The decision of two groups to come together is typically significant for third parties, whether these are non-participating associations or unaligned individuals. Both parties and unaligned citizens are relevant actors in the case of political alliance. To illustrate the concerns arising, consider the following charges: that pacts entail a reduction of choice for the citizen, who then has fewer parties to join or vote for; that they blur the contours of political opinion, such that governments and/or oppositions no longer have an identifiable programmatic orientation; or that a particular alliance is unwelcome, for example because it strengthens undesirable forces. For political agents typically defined by their relations of conflict, practices of non-combat always risk attracting the allegation of collusion. Clearly such outcomes, if real, may be balanced by countervailing considerations. ²¹ But one sees the potential for a pact to carry decisive implications for the citizen and the democratic system as a whole – to pose important externalities, that is.

The internalities of alliance concern its implications for agents inside the arrangement. Participation can have unequal effects for those involved, both as associations and as individuals. At a minimum, pacts may be more advantageous to some parties than others. More critically, a dominant partner may exploit weaker members. Because of the concessions that co-operation demands – for example, the sharing of confidential information – parties to an alliance put themselves in a situation of vulnerability, from which some may emerge better than others. Likewise, the consequences of alliance formation may be asymmetrical for individuals. Alliances may deny some partisans (for example, those in certain districts) the opportunity to pursue struggles they have long been committed to. Such arrangements may also empower leaders at the expense of the wider membership, since effective co-ordination between parties may depend on centralizing decision-making in each. Negotiation may require secrecy, thereby privileging a small circle of representatives. ²² Even in relatively benign cases, the compromises necessary to maintain the unity of an alliance always threaten to denude partisanship of its basis in principle. When commitments professed by the primary association(s) are suspended or discarded, not only are certain reasons for affiliation abandoned, but the credibility of the association may be impaired. ²³

The following discussion examines such internalities of the arrangement. Since the externalities of an associative practice are generally far-reaching, focusing on the internalities makes for a more manageable research problem, and arguably directs attention to the most crucial aspects. The citizens directly involved in an alliance are those most directly concerned: at stake is the fate of the projects they have committed to and helped to shape. If an alliance cannot be justified to those participating, the normative implications for third parties pale in significance.

THE INTRINSIC VALUE OF ALLIANCE?

Political pacts raise challenging questions for their prospective partners. This is true even if one adopts a stripped-back conception of the party as nothing more than a network of elites seeking office. ²⁴ Alliances in this image raise problems concerning the distribution of burdens and benefits in the maintenance of co-ordination, even if the partisan has considerable flexibility in meeting the demands of co-ordination. But the challenges are all the greater if we approach partisanship as a principled activity in which partisans have commitments they cannot dispense with lightly. ²⁵ If these commitments are what give the activity its very meaning, and are not simply adornments for public consumption, then the potential arises for a stubborn clash between the primary commitments of partisanship and the second-order commitments of alliance. Here I adopt the latter conception of partisanship, and with it the assumption that partisanship of this kind is valuable. Arguments for alliance cannot be indifferent then to the related implications for the ethics of partisanship.

If alliance participation is typically challenging, on what grounds can it reasonably be advocated? In the following I examine some salient arguments. I draw on reasons that have been put forward in real-world debates, including by parliamentarians, activists and commentators. ²⁶ In contrast to the paucity of scholarship in this area, there are some sophisticated interventions that can structure the analysis. Typically they are given as reasons for a particular alliance rather than to defend the very idea, but they can be reformulated in more general terms.

There are broadly two families of argument I shall consider: (1) those that treat political alliance as intrinsically desirable, and (2) those that treat it as instrumentally useful. In the first category are reasons for participation that should hold irrespective of whether it serves electoral success or other measures of partisan achievement: they are non-consequentialist reasons that rest on the idea that the parties in question are ‘better together’. The second category comprises reasons that hold only to the extent that in given circumstances they advance the party’s cause. This section examines the former. ²⁷

Let us first consider an epistemic argument:

1) Alliance between parties is intrinsically valuable because it enriches the political outlook of each.

‘No party has a monopoly on wisdom’ is the intuition behind this view. Partisans, it says, are more clear-headed when nested in a larger formation. An umbrella association of this sort allows the attainment of a more enlightened perspective. ²⁸ In the terms introduced earlier, the secondary ties that shape the alliance mark an improvement on the primary commitments of each party taken separately.

The argument can take more than one form. In one version it may be suggested that alliance participation facilitates the emergence of new perspectives superior to the initial programmes of the parties involved. Alliance, it may be said, is conducive to an enlightened consensus, ²⁹ whereby members willingly revise their preferences on the issues at hand as they come to see the merits of the better arguments. Alliance formation becomes analogous to a process of deliberation. The perspective is subsidiary to our concerns, however, as it leads quickly from an endorsement of alliance formation to an endorsement of merger. If the pooling of resources allows this type of epistemic advance, why should an alliance be defended as anything other than a stepping stone to the formation of a ‘great’ party, one that offers the same without problems of co-ordination, divided commitments and the like?

In an alternative and more relevant view, alliance participation is valuable not because it encourages the development of new perspectives but because, without fundamentally altering or supplementing the initial programmes of the parties involved, it cultivates in these parties a more reflexive disposition towards their commitments. ³⁰ By fostering the encounter of diverse views and encouraging compromise between them, an alliance helps curb the excesses to which each constituent party may be prone, thus raising the quality of their decisions. According to this view, single parties cultivate a blinkered form of partisanship. They encourage partisans to regard themselves as having privileged insight into moral and political truths. As one advocate of alliance puts it, ‘nearly always we position ourselves, our own little tribe, as better than the other progressive tribes, and grumble “if only they would see the light and vote for us!”’ ³¹ This conviction of ‘our’ superiority is misplaced. Political collectives should be maintained in a condition of alliance therefore, while merger should be avoided so that these tendencies towards ‘tribalism’ do not reappear on a larger scale.

What should one make of this argument for the intrinsic value of alliance? Clearly one may wonder whether such pacts are necessary for the epistemic advantages described. The clash of opinions at the wider political level, between parties in an adversarial relationship, is potentially conducive too, as the research on deliberative systems may suggest. ³² One may also wonder whether the positive epistemic effects of alliance participation do not heavily depend on the nature of the parties involved. Presumably not all alliances can be expected to be enriching in this way, but only those whose individual programmes already have something going for them.

But key to assessing this argument I believe is to reflect on the idea of partisanship it asks us to accept. It would seem to rest on seeing the primary association to which partisans are committed (their party, that is) as something incomplete, ill-equipped to stand on its own. It assumes that partisanship is a form of partiality, a distorted outlook that produces imperfect reasoning, and which needs tempering by its articulation with other such perspectives. Such assumptions track some familiar ways of thinking about political commitment. Mill’s discussion of those ‘one-eye men’ who, due to their stubborn attachment to their views are able to grasp only partial truths, is an expression of the thought that, though commitment may be welcome, it is always epistemically incomplete. Truth lies at best somewhere ‘between’ such views. ³³

The problem with this argument is that it radically diminishes partisanship as a reasonable endeavour. The worth of alliance is established, but at the expense of the worth of partisanship. It is one thing for the partisan to acknowledge that the commitments she holds could be mistaken, or that they are insufficiently demonstrable in their truth for all to be persuaded of them. Arguably, democratic partisanship relies on accepting, at some level, the contestability of commitments. It is quite another thing, however, to accept that one’s commitments are so imperfect as to make it intrinsically valuable to pair them with others. Such a view requires partisans to take an unreasonably sceptical view of the goals and principles espoused by their primary association – the very things for which they have reason to align with the party in the first place. Doing so would belittle the political conflicts that they, and their predecessors in the party, have committed to.

To be sure, it may be that some partisans can accept this sceptical view while still finding reasons to act. Or it may be that some can endorse this argument for tactical reasons. Emphasizing the intrinsic value of alliance may, in some cases, serve the instrumental purpose of boosting a party’s reputation, particularly in a climate of anti-partisanship. But it is not, I suggest, a position that can reasonably be accepted by most partisans as an accurate description of the nature of their commitment to their party, nor one they can easily attribute to the previous generations of partisans whose work they continue.

Let us consider then a different line of argument for why alliance formation is intrinsically valuable. This is the argument from democratic inclusion.

2) When parties form alliances, they give voice to more citizens, and thereby express the ideal of democratic inclusion better than any single party.

In this line of thought, parties are representative agents whose purpose is to connect citizens – their interests, preferences or identities – with the decision making of political institutions. They give voice to those who, in a modern democracy, are inevitably at some distance from legislative and executive power. Each, moreover, has natural constituencies they are uniquely well placed to represent. The argument for alliance follows from this. Because each party has special links to a different set of citizens, aggregating several into a bloc allows more constituencies to be heard, and gives the viewpoints represented a better chance of influencing political debate and decisions. ³⁴

As before, the question arises why this should be an argument for alliance rather than merger. Could some kind of great party not simply speak for more citizens? The preference for alliance is likely to rest on the assumption that organic attachments between parties and their constituencies have developed over an extended period, and are rooted in deep-seated features of social structure (group interest, identity, and so on). They are not likely to be recreated in any adequate way by the formation of a new super-party. ³⁵ Political collectives should be maintained in a condition of alliance so that these links are not broken, and the prospects of the democratic empowerment and inclusion are not lost in the process of reconstitution.

Like the epistemic argument, though differently, this argument entails seeing the party itself as partial. It perceives parties as grounded in a set of commitments that inevitably appeal only to one constituency among many, and that therefore need to be complemented by other, equally localized, claims. This notion that a party represents a partial constituency, advancing goals that not all can share, corresponds to a long line of theorizing that takes in both normative thought (for example, Kelsen) and sociological study (for example, Lipset and Rokkan).

If this were all that a party is, this rationale for alliance would be compelling. Yet what the view misses is the aim characteristic of partisans to generalize their claims beyond a pre-determined constituency. The logic of partisanship relies on building constituencies, developing justifications for why the good of sectoral groups benefits the wider public. ³⁶ Though partisans may be well aware that their arguments are better received by some groups than others, the structure of their position is intended to be generalizable. This is crucial for the question of alliance, because the aggregation of normative programmes intended to be generalizable is likely to call the coherence of each into question. ³⁷ While parties may superficially empower their followers by aligning them with others in a larger association, they risk depriving them in the process of the coherence of the programmatic goals that first mobilized them. (The experience of many grand coalitions in recent European history suggests that combining parties in this way can undermine their distinctive normative commitments, often to the point that they are publicly viewed as indistinguishable). ³⁸ To suppose the intrinsic appeal of this aggregation of views overlooks the proper claim of each to hold self-standing normative positions.

To be sure, it may be that there are times when party programmes can, at least in part, be reconciled with each other such that coherence of this kind is maintained. ³⁹ Alliances within a ‘party family’, that is, among those whose programmatic views have points of overlap, may successfully sidestep this problem and deliver meaningful democratic inclusion. (I discuss such cases below.) But once we treat the worth of the alliance as conditional in this way on the parties’ goals and compatibility, we are no longer treating it as something desirable in itself.

Arguments, then, for the intrinsic value of political alliances would seem to rely on devaluing partisanship itself as a practice. Both we have considered treat partisanship as a form of partiality. Where co-operation across partisan lines is seen as inherently desirable, partisanship as representing ongoing allegiance to an association of unique commitments loses much of its rationale. Or to put it differently: associations for which alliance relations are intrinsically desirable are no longer best thought of as parties.

THE INSTRUMENTAL VALUE OF ALLIANCE

The possibility remains that political alliance, even if not desirable in itself, may be a valid means of achieving desirable ends. Though partisans may rightly prefer to pursue these as an independent association, there may be circumstances when they are better pursued in co-ordination.

To reiterate, I treat partisanship as related to the collective and committed pursuit of generalizable goals and principles, not just the pursuit of institutional power. Improving the prospects of gaining office will never be a sufficient rationale for forming an alliance. The kind of instrumental argument I have in mind is one that sees alliance as instrumental to furthering the programmatic ends for which a party stands, and while holding office may be advantageous in this regard, it cannot in itself be regarded as the measure of a worthwhile alliance.

Accordingly, consider what we might call the programmatic argument for alliance:

3) Where two or more parties have programmatic goals in common, co-operation is appropriate if it serves to advance these.

The proposition might be illustrated as follows. Two parties share the goal of a publicly owned transport infrastructure. By co-operating in selected ways, be it co-ordinated public statements, legislative voting or electoral campaigning, they may be better able to advance this goal. Because the area of agreement is limited in range, some form of alliance rather than merger is the relevant option. And because the value attached to this co-operation is instrumental rather than intrinsic, there is no expectation that further points of overlap should be engineered: it is simply a matter of working together to advance the ends on which they happen to agree.

Let us set aside for now the possibility that pursuing some goals jointly puts others in jeopardy. ⁴⁰ All things being equal, what should one make of an instrumental approach of this kind? It may be useful to distinguish between two kinds of cross-party agreement: one that stays at the level of policy prescription, and one that extends more deeply to include the value orientations such policies are based on.

The first, more superficial, form of agreement we might call a modus vivendi. ⁴¹ It involves the parties having different supporting justifications for the policies they hold in common, differences which for strategic reasons they agree to overlook. Grand coalitions are one context in which this may be seen. ⁴² One of the problems of co-operation based only on a convergence of this kind is that it encourages the prescriptive elements of a programme to be detached from its diagnostic elements. Whereas one party may advocate public ownership of transport infrastructure on environmental grounds, as one investment among several to reduce air pollution, another may advocate it as a form of economic stimulus. ⁴³ Given that the reasons are different, and possibly at odds, both parties to the alliance would need to set aside the broader principles of their primary association in order to make sense of the pact. ⁴⁴ Such an arrangement puts the programmatic coherence of its parties under strain, even if it involves no compromises that directly contradict them. A modus vivendi seems a problematic foundation for an alliance.

It is different when the area of agreement has a principled basis – when, for example, two parties advocate a publicly owned transport infrastructure because both endorse ownership of utilities as an essential public good. An alliance grounded on such thicker agreement seems more consistent with partisanship as a principled activity. The secondary commitments arising between the alliance partners are then rooted in the primary commitments of each: the alliance is an extension of the partisan project rather than something that calls it into question. Certainly, an alliance may not be justified in every such case, since unwelcome compromises on further issues may be required to secure it (more on this in the next section). There is also the empirical risk that the issue that allows the parties to come together in a pact becomes the dominant one around which they subsequently campaign, in the process marginalizing other more salient concerns. It is not a sufficient reason to form alliance, in other words, but it does suggest that an instrumental defence of alliance is feasible.

Are there further possibilities for an instrumental defence? I suggest there is a fourth argument, which we might call the argument from imperfect institutions:

4) Where parties face a dominant adversary, and the political system perpetuates this imbalance of power, alliance formation is an appropriate response.

Empirically, it is often the case that alliances are against a third party at least as much as an expression of programmatic convergence. ⁴⁵ The stand-out case involves a rising party that is authoritarian and/or illiberal, and thereby poses a direct threat to democratic rights and practices. In view of the far-right populist currents in many Western countries today, such cases are highly topical; ideas of a cordon sanitaire to exclude such parties from government were also a theme in many European countries in the late twentieth century. ⁴⁶ In some ways these cases are theoretically straightforward. Given that authoritarian parties threaten the capacity of any democratic party to achieve its goals, the basis for an alliance among its opponents seems clear. Indeed, beyond their associational attachments, democratic partisans in such contexts arguably have a civic obligation to form an alliance of opposition. ⁴⁷

More challenging for our discussion is the case of the party that, though not overtly hostile to the democratic process, nonetheless poses a major challenge to it. One thinks, for instance, of a party that has become dominant and threatens to remain so, aided by an electoral system that is disproportionate and systematically favours it. ⁴⁸ The design of the system’s institutions, and their susceptibility to the influence of economic interests, may leave the party’s power unchecked.

Again, the rationale for alliance formation in such cases is instrumental rather than intrinsic. Since the aim is to overcome a particular set of challenges, without those challenges an alliance would be worthless. The secondary commitments that define it are at the general level of commitments to democratic practice. An alliance even between programmatically ‘remote’ parties becomes conceivable in such circumstances. ⁴⁹ Among the forms it may take are non-aggression pacts designed to maximize the seats held by the alliance partners in representative institutions (and thereby impede the adversary), or combined efforts to seek structural change to the electoral system. ⁵⁰

Is this a convincing reason to form an alliance? Clearly it is one that can be misused. Efforts to pursue constitutional change at the expense of a third party can evidently themselves be anti-democratic: much depends on the extent to which there are genuine structural problems in the political system. Such efforts can also be strategically unwise, if they allow a third party to gain further support by denouncing its opponents as a cartel. In some cases they may also be unnecessary: citizens may be willing to spontaneously engage in tactical voting to oppose third parties, thereby achieving the same outcomes without the formation of an alliance.

But if one accepts that representative institutions can be highly imperfect, and that coping mechanisms of some kind may be needed, then in at least some cases this will be a promising rationale for alliance. It offers a way for partisans to make credible the commitments they stand for by correcting the institutional failures that threaten to obstruct their pursuit. Rather than calling into question the reasons for which partisans associate with their party, it promises to augment the normative standing of their primary association.

CIRCUMSCRIBING ALLIANCE

Alliances between parties are fraught with risks, as the article has observed. Among those we might class as the internalities of the arrangement are the potential for one or more parties to be exploited, or for individuals within parties to be adversely effected. Yet as the previous section has argued, there may be occasions when an alliance is warranted as a way to achieve defined ends. The challenges cannot simply be avoided. The pressing question then becomes how to conceive the standards to which such arrangements might be held. Whereas the focus hitherto has been on the reasons that can motivate alliance participation, in this final section we examine the ethics of the parties in interaction.

It is important to underline that an alliance is an ethical order, even if its appropriate rationale is never more than instrumental. Even minimal co-operation in the pursuit of shared ends leads partners to cultivate expectations of mutual support. Where there is an explicit alliance agreement, the arrangement can be likened to a contract, with the alliance partners entitled to expect of each other that they follow jointly agreed courses of action. ⁵¹ Coalition governments are typically underpinned by an agreement of this sort. But even when there is no formal contract, the fact that the parties are publicly working together in pursuit of common goals gives an ethical dimension to their ties. They knowingly place each other in a situation of mutual reliance, and by identifying with the shared purposes that define the alliance they lead each other to believe that any sacrifices made were made for good reason. ⁵²

At the same time, the ethical ties should not be overdrawn. An alliance is structured in a way that anticipates its dissolution at any moment. Its parts are maintained as entities divisible from one another: the exit option is real. When an alliance becomes undesirable to one of its constituent parties, they are ultimately in a position to leave. To this extent one should not expect alliances to reproduce the same levels of obligation as partisanship itself.

Clearly the circumstances of an alliance, and the reasons that it forms, may be relevant to the standards applied. When an alliance is pursued in response to an immediate threat – for example, the rise of an anti-democratic party that institutions are powerless to resist – it may be that norms appropriate in less demanding situations should be relaxed. Some may even argue that there are situations of high emergency when the need for an alliance trumps all ethical considerations. But even if one accepts that such situations exist, it is important to reflect on the kind of norms to which an alliance should aspire. The best guide here would seem to be the presuppositions implied by its formal structure.

The challenges identified in the preceding discussion suggest the importance of a procedural norm of continual review by the memberships of the participating parties. An alliance is typically contentious not only at the point of its formation (when the instrumental reasons for involvement will have to be assessed and balanced against countervailing concerns), but over the course of its subsequent evolution. Alliances take shape in a changing environment, and themselves reshape that environment. As an association of instrumental value, an alliance needs continual review to assess how well it continues to serve the second-order commitments that define it, and to ensure that an arrangement established for defined purposes does not evolve into something else. This much is consistent with the reversibility and possibility of exit that the structure of an alliance is designed to maintain. Such a review is also important for establishing whether an alliance maintains the support of the partisans that form it. Given the potential of such pacts to centralize power within parties, periodic reassessment is crucial to giving it an intraparty democratic foundation and to ensuring key social groups (for example, trade unions) remain supportive. ⁵³ An alliance need not create cross-party democratic structures, but it should invite scrutiny from the members of the parties that form it.

Although procedural constraints on the conduct of alliances are vital, they address only some of the problems such arrangements can generate. After all, alliance participation may affect the membership composition of the parties involved, and with it their ideas. ⁵⁴ Seen ex ante, for partisans considering whether to proceed with an alliance, procedural checks threaten simply to legitimize programmatic drift. So it is important to consider substantive constraints that can circumscribe the conduct of the arrangement.

We have already noted the idea that points of programmatic convergence between allied parties should be principled rather than merely policy based. Consistent with this is a norm of coherence in the ends they pursue. It should be possible for the parties to the alliance to narrate its purposes and composition as a function of the primary commitments they hold. This may involve appealing to supra-partisan categories – ‘the left’, ‘the right’, ‘progressive’ and related ideas of a ‘party family’ – as well as highlighting areas of principled convergence. It may also entail open criticism of alliance partners with regard to those issues on which they diverge, so that the conditional nature of their relationship and the primacy of their party commitments is manifest. The pursuit of programmatic coherence is likely in addition to involve identifying some parties as inappropriate alliance partners if areas of convergence are lacking. Even an alliance of last resort aimed at redressing institutional failings will need to work in the first instance within such parameters of programmatic coherence. ⁵⁵

Still, how can an arrangement that demands compromise of its participants avoid unpicking the normative commitments to which each is attached? I suggest an acceptable alliance also depends on a norm of protected commitments, that is, commitments associated with each primary association on which progress must be sought (and on which regress is unacceptable) over the course of the alliance. Such commitments are those taken to be essential to the normative projects of the parties involved, and for the sake of which an alliance is instrumentally useful. They cannot be sacrificed without depriving the party of its normative purpose and undermining what the structure of an alliance is intended to uphold. These are equivalent to the ‘red lines’ of diplomacy. Where such lines are to be drawn, and how they are to be debated, is a matter for the ethics of partisanship rather than alliance. But it seems fundamental to the normativity of alliance relations that each ally should respect the existence of such lines. Each must accept that there are political options unavailable to its partners, and work to achieve a clear understanding of what these protected commitments are and how the alliance can proceed in accordance with them. Typically a formal written agreement may help provide clarity and mutual assurance.

The importance of acknowledging protected commitments can be observed empirically in alliance practice. Studies of alliances between social-democratic and radical-left parties suggest the frequency with which the former, as typically the larger partner, may pull the latter away from its core demands. ⁵⁶ A written pre-agreement of the terms of alliance, even if it cannot remove this risk, gives the parties involved a resource to defend themselves against exploitation by their partners – or highlight it when it occurs. The Norwegian red-green alliance of 2005–13, bringing together the Labour Party (Ap), Socialist Left Party (SV) and the Centre Party (Sp) in opposition to an alliance of the centre-right, is one such pact that experienced a programmatic shift to the right in its later years. The Soria Moria declaration of principles issued in 2005 at the beginning of the alliance can be said both to have retarded this shift and then later offered partisans of the Left Party a reference point for highlighting how the alliance had gone astray. It offered a resource by which the Left could claim responsibility for popular alliance policies, and distance itself from policies it did not support. Such experiences suggest the importance of clearly codifying protected commitments when forming alliances. ⁵⁷

It may be thought this attitude of respect towards the core views of allies is already guaranteed by the emphasis on principled convergence. But agreement between parties will typically require not just that certain points be agreed, but that points of disagreement be negotiated. To recall our earlier example, it is one thing to agree on the goal of a publicly owned transport infrastructure, and to do so for broadly the same reasons, but this says nothing of the issues that divide the parties. These may be reconciled in more than one way: as one author puts it, agreements ‘can be said to be additive if compromise is reached by including aspects of the other’s position in the position ultimately agreed to. It is subtractive when arriving at a compromise also requires that one or both parties remove, as a condition of arriving at a compromise, aspects of their initial positions that are unacceptable to the other side’. ⁵⁸ A norm of protected commitments implies that, where alliances involve subtractive compromise, the commitments subtracted should not include those considered essential to the parties involved. Without such a norm, alliance negotiations would become a matter of mere bargaining, and the partisans asked to respect their outcome would have every reason to feel disillusioned.

What of those additions to a party programme that may be needed to achieve additive compromise? These are no less relevant to guarding against the unplanned transformation of the parties to an alliance. A party that takes on new goals as part of its involvement in a pact acquires priorities that threaten to marginalize its core commitments. It seems important that whatever programmatic shifts of this kind made in the course of forming an alliance can be reversed when the party exits. In practical terms, this might suggest a freeze on amendments to a party’s constitution for the duration of an alliance. Whether they concern the principled goals of the party, or matters of internal organization such as the relations between the leadership, activists or parliamentary group, constitutional changes are by their nature difficult to reverse. An acceptable alliance is one that does not require, as part of an additive compromise between allies, a party to change its constitution.

Clearly this does not imply that a party ought to undo the changes it makes in the course of joining an alliance. This is a matter for intraparty deliberation, and it may well be that some changes adopted for the sake of alliance co-ordination are later endorsed for different reasons. The point is that it should not be the effect of a pact to settle these questions in favour of one position. As an instrumental arrangement to advance shared purposes, an alliance should be configured to minimize its transformative effects on the identity of the agents that form it. Their core programmatic commitments should be protected, both from the deletion of goals that serve these commitments and from the accretion of goals that marginalize them.

CONCLUSION

While pacts between political parties are common, analyses of the normative questions they raise are rare. Like alliances in other domains, from international relations to the world of business, those in party democracy are typically approached as problems of bargaining strategy or system balance. The ethics of alliance are neglected, and not without cost. The distinctive structure of an alliance, as an association of associations that retains divisibility, leads to a unique set of questions beyond those of partisanship itself. The promise of such pacts is to augment the party as a mode of citizenship and an agent of political change. At the same time, such arrangements threaten to dilute its normative purpose. For anyone interested in the reformative and transformative potential of parties, as well as the controversies that partisanship gives rise to, the ethics of alliance demands attention.

As I have argued, it is possible for the committed partisan to endorse the prospect of alliance participation. Though the tension between the two forms of association is real, and creates tendencies both to dissolution on the one hand and merger on the other, it need not be fatal. Yet resolving the tension depends on seeing the worth of alliances as instrumental rather than intrinsic. Rather than as a form of epistemic enrichment or democratic inclusion, they should be valued as a means to advance programmatic ends where principled agreement exists, and as a way to counter the imperfections of political institutions. The norms to which an alliance should be held centre on ensuring that the pursuit of these second-order ends does not undermine the first-order ends that motivate allegiance to the parties in question.

As an effort to introduce new perspectives on an under-theorized practice, this article necessarily leaves a number of issues for future research. I have discussed alliance in largely general terms, but clearly there is more to be said on the different forms it can take. When, for instance, are tightly co-ordinated kinds of co-operation – for example, open primaries to select joint candidates – to be preferred over looser arrangements such as pacts of non-aggression? What kinds of organizational structures, either within or across parties, are most consistent with the norms examined, including the ones best able to counter the elitist tendencies inherent in pacts? What type of institutional recommendations for an electoral system might one distil from the analysis? If alliances between parties have instrumental rather than intrinsic worth, how might institutions be configured to reflect this? And are the normative considerations the same for transnational alliances? I leave these questions for future research. In the meantime I hope to have shown why such questions invite further study. Political alliance is a practice of normative significance. There are bad reasons to form pacts, but also some good ones, and where an alliance is worth having there are standards it should be held to.