Political Equality and Political Sufficiency

1 Introduction

Most democratic theorists support a high degree of political equality, arguing that all citizens should be treated equally in the democratic process. But some democratic theorists confuse political equality with what I call ‘political sufficiency’, namely, that all citizens should be at or above a certain level of a democratic resource, even if some citizens have more of that resource than others.

Conflating political equality and political sufficiency has significant implications for democratic theorists defending or criticising political equality. Thomas Christiano advocates political equality when his arguments sometimes only require political sufficiency. Sometimes, rewording his arguments would entail the same conclusion. But sometimes his arguments genuinely imply political sufficiency, I show, not political equality. Different defences of political egalitarianism may be needed.

Meanwhile, David Estlund criticises political equality, but I show that what he really objects to is not political equality as such: it is political equality below sufficiency levels. Clarifying the distinction between political equality and political sufficiency can thus have normative payoffs.

This paper also has significant methodological implications, for political theorists and philosophers more generally. We can easily draw faulty inferences by treating two things as one (a form of conceptual stretching) and then ignoring the effect of one of these two things, on its own (omitted variable bias) or in interaction with the other (an interaction effect). (Conceptual stretching involves expanding a concept too far, e.g. calling all left-wingers ‘communists’ or all right-wingers ‘fascists’. Omitted variable bias means that an explanation overlooks a factor which is actually relevant, e.g. explaining children’s height in terms of gender and age only, while ignoring diet. To understand interaction effects, consider alcohol and pills: perhaps drinking lots of alcohol makes you vomit; perhaps taking lots of pills makes you ill; but doing both kills you. The interaction creates a whole new chemical reaction whose effect is far worse than simply vomiting plus feeling ill.) Democratic theorists, I show, often discuss political equality or inequality with examples that also involve insufficiency, and although they think it is equality or inequality that drives their conclusions, it is actually insufficiency, either on its own or in interaction with (in)equality. Such errors are not limited to democratic theory.

These comments may sound as if political equality and political sufficiency are rivals. That is sometimes true, and this matters particularly when democratic theorists recommend political equality while overlooking political sufficiency. But, importantly, political equality and political sufficiency can also be allies: theorists committed solely to political equality are open to ‘levelling-down’ objections, accepting low-quality procedures for the sake of equality. Political egalitarians can thus strengthen their position by specifying the sufficient level of the resource to be equalised. Political egalitarians should not reject sufficientarian arguments because of problems such as how to agree on what is sufficient. These difficulties are significant, but egalitarians too must address them.

The equality/sufficiency distinction, of course, comes from the distributive justice literature. Even there, interestingly, sufficientarianism is discussed far less than egalitarianism (as noted by Shields 2016, p. 8, and Fourie 2017, p. 11). True, the ideas – albeit rarely the labels – are sometimes applied in democratic theory. Mansbridge (1980, p. 245) criticises participatory democrats who propose an equal voice in politics when the educational benefits of participation may merely require some voice in politics. Beitz (1989, pp. 207–209) and Cohen (1996, pp. 109–110) argue that election campaigns should not cap spending at a maximum level, but provide a minimum, sufficient level for all candidates. And Kymlicka (1995, pp. 146–147) distinguishes the view that a minority group’s members should have legislative representation in proportion to their share of the population, and the view that they need ‘a threshold number of representatives, sufficient to ensure that the group’s views and interests are effectively expressed’. This threshold number may be less than, the same as, or more than the number required by political equality.

Nonetheless, the distinction is not prominent in democratic theory. Section 2 thus clarifies the distinction – and offers useful tools for distributive justice theorists too. Section 3 discusses sufficientarian conclusions which are wrongly inferred from egalitarian premises, using Mark Warren’s work on democratic corruption. Section 4 discusses egalitarian conclusions which are wrongly inferred from sufficientarian premises, using Thomas Christano’s defence of political equality. Section 5 considers an objection involving political power as a positional good. Section 6 discusses anti-egalitarian conclusions which are wrongly derived from sufficientarian premises, using David Estlund’s critique of political equality. Political egalitarians, I show, need political sufficiency to deflect levelling-down objections. Section 7 highlights the methodological importance of identifying and differentiating key variables, for political theorists and philosophers more generally.

2 Conceptual Framework

This section offers conceptual distinctions which are fundamental to this paper and which should benefit not only democratic theorists but also analysts of distributive justice. The latter, I suggest, often overlook or elide distinctions which help us understand sufficiency and which thus help us see when our conclusions should be sufficientarian, egalitarian, prioritarian, and so on.

Understandably, the distributive justice literature on equality and sufficiency usually treats egalitarianism and sufficientarianism as positions (e.g. Christiano 2007, pp. 42–45, 59–60; Widerquist 2010, p. 474; Kanschik 2015, pp. 82–83; an exception is Ypi 2012, p. 89). I prefer to start lower, with egalitarian and sufficientarian claims, without implying that egalitarian and sufficientarian positions always and only entail egalitarian or sufficientarian claims respectively: sufficientarians also support equality, and egalitarians need sufficiency to deflect levelling-down objections. My starting point is particularly important for democratic theory, since democracy is so closely linked to political equality (Beetham 1999, p. 5). I wish to broaden our conceptual armoury and help political egalitarians avert levelling-down objections.

Egalitarian claims have the form All Ps should M the same amount of X, where P is a subject (usually an individual or group, but potentially a region, house, horse, etc.), M is a verb (e.g. have, receive), and X is an object (e.g. money, opportunity). For ease of exposition, I will henceforth replace M with ‘have’. Egalitarian claims thus take the form All Ps should have the same amount of X.

Sufficientarian claims have the form All Ps should have at least F amount of X, where F is an absolute threshold (such as 10 units of a resource, or 5 units for a child and 10 for an adult). Sufficientarianism is a minimum absolute-threshold argument; maximum absolute-threshold arguments state that All Ps should have no more than G amount of X (on maximum absolute-threshold arguments, e.g. limiting our emissions or consumption, see Spengler 2016, pp. 925–936). Some maximum absolute-threshold statements are equivalent to minimum ones, as discussed later. (Maximum absolute-threshold statements should not be confused with ‘upper limit sufficientarianism’, where no distributive principles apply once sufficiency is met – see Shields 2016, pp. 22–23.)

What I will call ‘relativarian’ claims have the form All Ps should have at least H amount of X, where H is a relative threshold (such as half the mean amount of X, or 10 units under the maximum amount of X).

All three ideas are found in Rousseau’s Social Contract. Key political resources should be distributed equally: each citizen has one vote, and no law-making power is delegated to representatives. Material resources must lie within absolute and relative thresholds. There are two absolute thresholds, one minimum and one maximum: no citizen should be so poor as to be corrupted by desiring luxury, or so rich as to be corrupted by its possession. As regards relative thresholds, no citizen should be so much richer than anyone else that he could corrupt others, or so much poorer that he could lose his independence through bribery. These relative thresholds sit somewhere between the minimum and maximum absolute thresholds (Rousseau 1762/1997, 2.11 [book 2 chapter 11], p. 78; 3.4, p. 91; 3.15, pp. 114–115; 4.2, p. 124). (As is the convention for history of political thought, I have retained Rousseau’s gender-language to avoid glossing over his sexism.)

My main focus is on sufficiency and equality, not ‘relativity’. But two quick points are worth noting. One is that while sufficientarian claims are non-comparative (the position of other Ps is irrelevant when deciding whether a given P is over an absolute threshold F), relativarian claims are comparative (the position of other Ps affects whether a given P is over the relative threshold G). The second point is that the distinction between sufficientarianism and relativarianism may be blurred by conventional use of the term ‘enough’. Yet having enough means different things in absolute and relative terms. If people need ‘enough’ education for a fair chance in the job market, this could imply absolute standards, such as being able to read and write, and/or it could imply relative standards, such as not being so far below the average level of education that one is significantly disadvantaged. The idea of relativity thus strengthens our conceptual armoury.

Seven clarifications are needed. First, and most important, we must distinguish two separate aspects of equality, which Larry Temkin (2001) calls ‘equality as universality’ and ‘equality as comparability’. Equality as universality involves all Ps, equality as comparability refers to the same amount of X. Importantly, egalitarian, sufficientarian, and relativarian statements all entail equality as universality – indeed, most normative positions use it as a closure principle (Raz 1986, p. 220), while disagreeing about who is a P. But only egalitarian statements make a further commitment to distribute the resource comparably, that is, the same amount to all Ps.

Blurring these two ideas makes some theorists wrongly depict ideas as egalitarian rather than sufficientarian, as we will see. Talking of ‘enough’ does not make a claim sufficientarian, just as saying ‘equally’ does not make a claim egalitarian (Raz 1986, p. 221). ‘Everyone equally should have a right to political power’ could mean everyone having a right to political power, everyone having the same right to political power, or everyone having a right to equal political power. I thus use ‘universalist’ when discussing equality as universality, and ‘equality’ only when discussing equality as comparability.

Second, there are many types of political equality and sufficiency, reflecting the many parts of the political process (e.g. agenda-setting, deliberation, decision-making), and the many different P, M, F, G, H and X components. Few democratic theories are egalitarian or sufficientarian in all respects. (For distributive justice theorists discussing non-sufficientarians accepting sufficientarian ideas, see Axelsen and Nielsen 2015, pp. 407, 425–426; see also Rondel 2016, p. 155.) But someone might call themself a political egalitarian or a political sufficientarian if they have a presumption in favour of that principle or if core parts of their theory embody that principle.

Third, because the political process has several stages, egalitarianism in one respect may be sufficientarianism in another. For example, equality in votes (one person, one vote) would only be sufficiency of influence if everyone could influence outcomes but those with more voice have more chance of influence. Equal votes may even lead to insufficient influence for people in a permanent minority (Guinier 1994).

Fourth, equality and sufficiency are analytically distinct in theory; they may or may not overlap in practice. Both equality and sufficiency are transgressed where some Ps have at least F amount of X and some have less than F. This is a key reason why equality and sufficiency are often confused, as we will see.

Fifth, what is sufficient for one person may be insufficient for another, since individuals differ in their ability to convert resources. My initial example of sufficiency thus specified 5 units for a child and 10 for an adult.

Sixth, ‘sufficiency’ need not mean a bare minimum. The distributive justice literature, understandably, often depicts sufficiency as the bare minimum for subsistence (e.g. Ypi 2012, p. 108; García-Gibson 2016). But in this paper, sufficiency means any specified positive level of the resource, whether that level is low or high (see likewise Burroughs 2016, p. 230; Christiano 2007, p. 60; Gilabert 2012, p. 5; Shields 2016, p. 27; Swift 2001, p. 122; Waldron 1993, pp. 250–251, 262, 270). A sufficiency of time for deliberation might mean enough time to cover the ground adequately or enough time to probe the issues in detail. So, we should ask not only ‘equality of what?’, but also ‘sufficient for what?’. Similarly, depending on their aims, political sufficientarians could simply require one person, one vote; or they could also require agenda-setting powers, recall powers, and so on. My paper leaves open where the sufficiency threshold is.

Seventh, I have deliberately presented equality and sufficiency minimalistically. Many writers are more expansive. For example, Robert Huseby treats sufficiency as ‘a telic principle … [which] is meant to apply to all human beings at all points in time’ (Huseby 2010, p. 179, emphasis removed). This is a legitimate conception of sufficiency but is only one specification of the concept of sufficiency. Meanwhile, my sufficiency claim is similar to Paula Casal’s ‘positive thesis’ (that people should be at/above some threshold), whereas sufficientarianism as a position involves more, for example, a ‘negative thesis’ (inequalities above this threshold are irrelevant) or a ‘shift thesis’ (inequalities above this threshold are less important). (The positive and negative theses are from Casal 2007, pp. 298–299; the shift thesis is from Shields 2016, p. 30.)

This has major normative implications. Egalitarians who reject particular conceptions of sufficiency can and should uphold the general concept of sufficiency. Starting with a specific conception of sufficiency, or sufficientarianism as a position, hides egalitarians’ need for sufficiency to avoid levelling-down objections. (See, similarly, the view that ‘The commonly assumed opposition between prioritarianism and sufficientism is, in an important way, a false dichotomy; we need not choose between either giving priority to the worse off or giving priority to the badly off, since we might instead give priority to both’- Brown 2005, pp. 202–203.)

So, egalitarians should take sufficiency seriously because sometimes egalitarianism requires sufficiency, not just equality. It thus helps to start with sufficiency as a claim; starting with sufficientarianism as a position makes egalitarians less well disposed to sufficiency claims.

My minimalist definition is not as minimal as Harry Frankfurt’s classic definition of sufficiency, that ‘each should have enough’ (1987, p. 21, emphasis removed; see likewise Brighouse and Swift 2006, p. 471, and Axelsen and Nielsen 2015, p. 406). This needs fleshing out, however, not least because egalitarians might say that people with less of a resource do not have enough. It thus helps to add that sufficientarianism means having enough in absolute terms – more than F amount of a resource, where F is an absolute threshold. By contrast, Adam Swift also moves between absolute and relative thresholds when discussing sufficiency. He discusses ‘enough’ first in relative terms (2001, pp. 111–112), then in absolute terms (2001, pp. 121–122). Both positions are viable but we should not conflate them, hence my distinction between sufficientarianism and relativarianism.

This stance partly reflects my methodological commitment to defining terms pragmatically. There are many ways of defining terms, but some are more useful than others. Defining sufficiency as ‘having enough’ is legitimate: indeed, this is our conventional understanding. But it is not useful for most political theorists, because so many theories have requirements about what is enough – and we would still need a term for distributions above an absolute threshold. Perhaps that could be called ‘thresholdarianism’, but the philosophical language in this area is already ugly enough. There are good reasons, then, to define sufficientarianism as I have done here.

Each distinction matters for this paper. Distinctions 5 and 6 show that ‘sufficiency’ is broader than the distributive justice literature usually implies. Distinctions 1–4 and 7 show why egalitarians can accept some sufficiency claims. Indeed, I argue later, egalitarians need sufficiency claims to deflect levelling-down objections. If you’re an egalitarian, how come you’re so opposed to sufficiency?

3 Mark Warren: Inferring Political Sufficiency from Political Equality

Overlooking the equality/sufficiency distinction can foster faulty normative inferences. I start with sufficientarian conclusions wrongly drawn from egalitarian premises, as with Mark Warren’s account of democracy and corruption.

Corruption is an inherently normative idea, so Warren updates traditional (pre-democratic) ideas of corruption by drawing on democratic norms. He depicts the basic democratic norm in terms of equality (or, a variant, proportionality): ‘every individual potentially affected by a decision should have an equal opportunity to influence the decision’ (Warren 2004, p. 333, emphasis removed) or ‘proportional to his or her stake in the outcome’ (Warren 2006 p. 804, emphasis removed; see similarly Warren 2006, p. 386).

Warren then applies these democratic norms to corruption. He argues that the most common idea of corruption today, the misuse of public office for private gain, is inadequate for contemporary democracies. Egalitarian and proportional democratic norms give us new insights about corruption.

However, Warren moves from egalitarian and proportional democratic norms to an idea of corruption based on insufficiency: ‘the norm violated by corruption is that of inclusion in collective decisions and actions of all affected’, such that ‘the very logic of corruption involves exclusion’, and ‘every form of corruption involves exclusion’ (Warren 2004, p. 333, emphasis removed). Similarly, ‘political corruption attacks democracy by excluding people from decisions that affect them’ (Warren 2006, p. 804, emphasis in the original). The new conception is even called ‘corruption as duplicitous exclusion’ (Warren 2004, pp. 329, 332, 2006).

But, strictly speaking, Warren’s own norms – equality or proportionality – suggest that corruption should be defined not in terms of people being excluded, but people having less weight than they should have, whether or not they are excluded.

Indeed, most of us would agree that democracy can still be corrupted where a duplicitous exchange of resources leads to affected parties having unequal or disproportionately low influence, while still being included. Imagine that a quarry opens in the countryside near a village. Lorries from the quarry will disturb villagers, through noise, smell and so on. Politicians are asked to decide whether to restrict the hours in which lorries can go to or from the quarry. The politicians consult with villagers and the quarry-owners. The villagers conclude that while they would ideally not want any lorries, the business is legitimate. The villagers thus ask that lorries run from 9 am to 5 pm, on weekdays only. The quarry-owners conclude that while they would ideally want lorries to run at all hours of all days, the villagers’ concerns are legitimate. The quarry-owners agree with the villagers’ proposal, and the politicians enact this.

Now imagine a corrupt version of this scenario: the quarry-owners bribe the politicians. The politicians refuse to accept the quarry-owners’ ideal scenario, running lorries at all hours of all days, but agree to let lorries run from 6 am to 8 pm, six days a week.

In the corrupt scenario, citizens have not been excluded. If they had been, lorries would have run at all hours of all days. The citizens still have some voice: their views are listened to, and they have influenced the outcome. But they have less influence than they should, due to a duplicitous exchange of resources. We should call this corrupt too.

Warren thus moves too quickly from an egalitarian/proportional norm to a conclusion involving sufficiency. Insufficiency is an extreme form of corruption, and is all too common, alas. But Warren’s democratic norms seem instead to imply that corruption involves unequal or disproportional influence. ‘Corruption as duplicitous exclusion’ should be ‘corruption as duplicitous inequality’ or ‘corruption as duplicitous disproportionality’. Warren’s formulation is certainly catchier. But it is not quite right.

Warren’s definition thus permits duplicitous violations of political equality, provided everyone is still included. By not adequately distinguishing between political sufficiency and political equality, his definition opens the door to anti-democratic actions.

A methodological point of considerable importance for this paper concerns Warren’s use of examples involving both inequality and exclusion: actions where corrupt actors achieve considerable influence while others have no influence at all. He then assumes that what makes this corrupt is the exclusion, not the inequality. I have suggested that what makes this corrupt is actually inequality; exclusion is just an extreme manifestation of inequality, and Warren thus wrongly depicts corruption in terms of exclusion, when inequality is the key. Warren’s arguments thus suffer from omitted variable bias: he thinks exclusion is the key, but inequality actually drives the result.

Clearly, if our examples involve two different factors without us realising it, we can easily focus on one and overlook the other (as with omitted variable bias: our model does not include a factor which is actually influencing the results, so we do not see its importance). Consider this question, adapted from Isaiah Berlin (1969, p. 130): is the average US citizen freer than the average Chinese citizen? Our answers will depend on what we mean by ‘free’, but there is a risk of conceptual stretching here. Berlin himself distinguishes between freedom (what we are allowed to do) and ability (what we can do): I am free but unable to play Beethoven’s piano sonatas, and able but not free to steal food (1969, pp. 122–123). Many left-wingers disagree: freedom includes ability, and poverty limits freedom. Unless we are clear about whether or not freedom includes ability, we may argue at cross-purposes, disagreeing about how free US and Chinese citizens are because some of us include ability while others do not. In short, treating two or more things as if they are one will affect our answers. This methodological slip is worryingly easy to make and potentially very damaging, as the next section explains.

4 Thomas Christiano: Inferring Political Equality from Political Sufficiency

I now address the most significant problem: egalitarian conclusions wrongly inferred from sufficientarian premises. I exemplify this using the work of Thomas Christiano. His democratic theory certainly places most emphasis on political equality, in relation to justice (1996, p. 59; 2008, p. 4), giving everyone an equal say (2000, p. 540; see also 2003, p. 195, 2004, p. 276, 2008, pp. 88, 95–96, 201–202), and the institutions which embody this (1996, pp. 224–240, 265–295, 1997, pp. 256–261). He explicitly rejects the view, compatible with sufficientarianism, that ‘everyone’s complaints ought to be taken seriously but some of these ought to be taken more seriously than others’ (2001, p. 212).

I will argue that some of Christiano’s arguments do, actually, point in sufficientarian directions – and, crucially, that he sometimes invalidly produces egalitarian conclusions from sufficientarian premises. I offer three such areas where Christiano’s egalitarian conclusions are invalid. First, where he can save his egalitarian conclusions by upgrading his sufficientarian premises to egalitarian ones. Second, where his egalitarian conclusions must be downgraded to sufficientarian ones. Third, where saving his egalitarian conclusions requires a different defence of political egalitarianism.

The first example of conflating sufficiency and equality is Christiano’s explanation of why societal institutions must publicly and visibly ‘embody the equal advancement of interests’. We have ‘fundamental interests in being publicly treated as equals’, Christiano writes, and this requires ‘equality in voting power, equality of opportunities to run for office, and ideally equality of opportunities to participate in the processes of negotiation and discussion that lead up to voting’. But he then tacitly moves to sufficiency:

If someone’s judgment is not permitted a hearing in society, then the interests described above will be set back. Anyone who is excluded from participation in discussion and debate can see that his or her interests are not being taken seriously and may legitimately infer that his or her moral standing is being treated as less than that of others. So justice, which requires public equality, demands equal respect for the judgment of each (2004, pp. 275–276; see also 2008, p. 88).

Like Warren, Christiano makes a crucial methodological slip here. Because he does not distinguish between equality and sufficiency, his example unwittingly involves inequality where some people are also below sufficiency. But it is not immediately clear whether we object to the inequality and/or the sufficiency. Christiano understandably infers that the problem is the inequality. However, inequality where everyone is above sufficiency will sometimes be less troubling than inequality where some people are below it. Egalitarian conclusions may not follow as easily from the former situation. Either the premise or the conclusion should be altered. In short, Christiano has not justified the ‘so’ of the final sentence. The danger is that Christiano’s inference is being driven at least partly by insufficiency. It is a bit like using examples of liberals who overlook community, then criticising ‘liberalism’, when only certain liberals overlook community, and many non-liberals also overlook community. The criticism should be about overlooking community, not liberalism.

Fortunately, Christiano’s error can be averted by upgrading the sufficientarian sentences into egalitarian ones:

If someone’s judgment is not permitted an equal hearing in society, then the interests described above will be set back. Anyone whose views are given less weight in discussion and debate can see that his or her interests are not being taken as seriously as others and may legitimately infer that his or her moral standing is being treated as less than that of others. So justice, which requires public equality, demands equal respect for the judgment of each.

The italicised changes better justify the egalitarian conclusion in the final sentence. The second sentence may need expanding: for example, racist interests might legitimately be given less weight. But overall, the amended version supports political equality better.

The same applies to the claim that ‘if one excludes a person or group of persons from having a say, their interests will not be properly taken account of’ (2003, p. 191, emphasis added). This can be amended to ‘if a person or group of persons has a less than equal say, their interests may not be properly taken account of’. (I say ‘may’ because the result is contingent: if that person or group is in a majority, their interests might still be well catered for, while a group with above-average say might still have no effect if they are in a permanent minority.)

I now discuss a second case, where Christiano’s conflation of sufficientarian and egalitarian claims cannot be resolved in an egalitarian direction, as just discussed, but requires an egalitarian conclusion to be downgraded to a sufficientarian one. Consider this claim:

each person has a fundamental interest in being treated as a person with equal moral standing among his fellow citizens. To be treated in a way that entirely ignores one’s way of perceiving how one is treated constitutes a serious loss of status for a person in a society. A person whose judgment about that society is never taken seriously by others is treated in effect like a child or a madman. Such a person is denied recognition of his or her moral personality. … This is a disastrous loss of moral standing (2004, p. 273; see also 2001, p. 206).

The egalitarian first sentence is stronger than the ensuing sufficientarian sentences. Yet being totally ignored is more degrading than being listened to less seriously than others, but still being heard. Indeed, I doubt that people are treated like children or the insane if their views are listened to, and sometimes taken seriously, but not always as seriously as others’ views. This is probably the situation for most people in current democracies; a ‘disastrous loss of moral standing’ does not always result.

The word ‘equal’ in ‘equal moral standing’ should thus be removed: nothing else in this passage entails equality, just sufficiency. Indeed, Christiano adds that one has ‘a deep interest in having one’s moral standing among one’s fellows clearly recognised and affirmed’ (2004, p. 273). That too sounds sufficientarian, not necessarily egalitarian. The conclusion of this discussion should be sufficientarian too. Egalitarian conclusions would require different premises. (The next section asks if anything less than equal moral standing amounts to no moral standing.)

I now turn to the third and final set of cases. One reason why Christiano’s conflations of sufficiency and equality are problematic is that he often makes empirical claims about the effect of political arrangements on self-respect, claims which look questionable once the sufficiency/equality distinction is made clear.

Consider this claim: ‘When an individual’s views are ignored or not given any weight, this undermines [a citizen’s] sense of self-respect, in which he has a deep interest’ (Christiano 1997, p. 259, emphasis added; see also 1996, p. 72). This is one justification for ‘the intrinsic importance of equality in public deliberation’, and Christiano concludes that ‘each person has a just claim to an equal share of the resources for making decisions and contributing to the public deliberations’ (1997, pp. 259, 261). This has major political implications, but the egalitarian conclusion also requires us to show deep losses of self-respect for citizens who are included but whose views are merely considered less strongly than others.

But – crucially – such situations of inequality above the sufficiency level will not necessarily cause deep losses of self-respect. For example, the single-member plurality electoral system, or first-past-the-post, breaks some principles of political equality (Christiano 1997, p. 234). Many citizens do not know this, but some do. They include democratic theorists, including me, who criticise these inequalities – without suffering deep losses of self-respect.

Political inequality can cause losses of self-respect, but not as often as political inequality which is also below the level of sufficiency, it seems. Christiano’s sufficientarian premise does not support his egalitarian conclusion. Again, an egalitarian justification is required, or a sufficientarian conclusion.

Here, though, I do not think Christiano’s premises can just be restated in egalitarian ways, as was the case with my first example above. Rather, like my second example above, a sufficientarian conclusion seems to be all that is justified. Empirically, there are cases where inequality simply is not as damaging as Christiano implies.

In such places, egalitarian conclusions probably require a different argumentative strategy – defending equality on intrinsic grounds, not instrumental/empirical ones. Criticising inequality because of its instrumental effect on self-respect leaves Christiano open to empirical objections like those above, since political inequality does not always undermine self-respect as he claims. Intrinsic defences of political equality are stronger: political inequality is wrong in itself. It is essentially unjust; it is not how people should be treated, whether it bothers them or not. Christiano sometimes uses such arguments elsewhere (e.g. 1997, pp. 245, 251–252, 2003, p. 188, 2004, p. 269, 2008, pp. 4, 95).

These three sets of cases pose three different challenges to Christiano’s defence of political equality. Some of his sufficientarian premises can be upgraded to permit egalitarian conclusions. Some of his egalitarian conclusions must be downgraded to sufficientarian ones. And sometimes his argumentative strategy needs to change direction significantly. Taking the equality/sufficiency distinction seriously thus has significant normative implications.

Finally, I will address the underlying methodological issue, which has major ramifications for political theory more generally. As with Warren, conflating equality and sufficiency makes it hard to see when the real problem is inequality and/or insufficiency. On several occasions, Christiano’s concern seems to be inequality where some are below sufficiency: ‘those who are excluded’ can see that ‘their interests are not treated as equally worthy of advancement’ and feel ‘as if they have a lesser moral standing’ (2004, p. 276). But to reject inequality in general, Christiano should not place so much weight on examples of inequality which involve insufficiency. Otherwise, readers might accept egalitarian conclusions because they dislike particular kinds of inequality, where some people are below sufficiency.

When there is both inequality and insufficiency, is the fundamental problem insufficiency alone, or is it the combination of inequality and insufficiency? This is hard to answer in the abstract: the answer depends on the resources involved, the reasons for the inequality/insufficiency, and so on. The question is especially hard to answer in democratic contexts. To assess the relative importance of equality and sufficiency, we ideally need to compare at least three situations: (a) inequality with everyone above sufficiency; (b) inequality with some above and some below sufficiency; and (c) equality with everyone below sufficiency. Yet category (c) is unusual in democracy: for example, a situation where everyone was equal in having no voting rights would be anarchy, not democracy. But a purely aggregative Schumpeterian democracy with no democratic deliberation at all would be an example of category (c). For many of us, this level of deliberation is insufficient. Either way, it is less demeaning and less intrinsically unjust than the category (a) equivalent, where everyone has the right to deliberate but some have more opportunities to do so than others. And it is much less demeaning and much less intrinsically unjust than the category (b) equivalent, where some people have the right to deliberate and others do not.

Now, it is an open question how generalisable this test is. But this test does initially suggest that inequality where everyone is above the sufficiency level is less worrying than inequality where some people are below it. It also suggests, at least in this case, that insufficiency alone is not as worrying as a combination of inequality and insufficiency. If we are all in the same boat, it may not matter if it leaks (as long as it does not sink).

This is why it is dangerous to justify political equality using examples which involve both inequality and insufficiency. The extra variable, insufficiency, distorts the results, because it interacts with inequality, the variable Christiano is most interested in. To make such strong conclusions about the importance of equality, we need different examples, featuring inequality in situations where everyone is above sufficiency levels.

The above criticisms do not fundamentally undermine Christiano’s democratic theory. It remains a powerful and influential exposition of a strongly egalitarian view of democracy. But some of Christiano’s premises should change if he is to reach egalitarian conclusions; some of his egalitarian conclusions should be downgraded to sufficientarian ones; and political equality should be defended intrinsically more than instrumentally. These points, of course, presuppose that one has made an initial conceptual distinction between political equality and political sufficiency – the central aim of this paper.

5 Political Power as a Positional Good

Before moving on, I must consider an objection to the previous section’s arguments, one involving political resources as positional goods.^[1] A good is positional when its benefits depend on how much of it other people also have (Hausman and McPherson 2006, p. 181). As more people move into a secluded part of the countryside, for example, the benefit of living there decreases. ‘If everyone stands on tiptoe, no one sees better’, in Fred Hirsch’s words (2005, p. 5). More formally, Harry Brighouse and Adam Swift write: ‘The absolute value of a positional good depends … on how much of it one has compared to others. … Insofar as goods are positional, relative amount determines absolute value’ (2006, pp. 474–475).

Clearly, political power can be at least partly positional (see too Ypi 2012, pp. 124–127, on global political power as a positional good). Imagine a proposal to give university graduates three votes and all other citizens two votes. Everyone would have greater choice and voice than under one-person, one-vote arrangements. People could vote for the same candidate more than once or spread their votes: for example, a university graduate could vote for a social democrat twice and a green candidate once, or for a conservative, a populist and a libertarian – more choice, more voice. But would everyone really be better off? Positionally, non-graduates would have less power than before. For Christiano, this would be so demeaning as to cause a disastrous loss of moral standing. Inequality causes insufficiency, on this view. David Axelsen and Lasse Nielsen make a similar point, as I will shortly explain.

This threatens my distinction between political equality and political sufficiency. In politics, one might argue, only equal treatment is enough treatment: if someone has less of a political resource than others, she does not have enough. Equality and sufficiency go hand in hand, it seems. Brighouse and Swift characterise this as sufficientarian: in politics, ‘only a fair chance is enough of a chance’, and ‘only an equal chance is a fair chance’ (2006, p. 476; for similar views, see Foley 1994, p. 1241, and Ypi 2012, pp. 109–130).

One way to think about positional goods, then, is that positionality involves the relation between two scales: your relative achievement on one scale determines your absolute achievement on another (Sen 1983, p. 156). For our purposes, this means that the partly positional nature of political resources can bring absolute thresholds into play in a related resource: inequality in one resource causes insufficiency in another.

This undeniably can happen. But my argument only requires that it will not necessarily happen. As shown by the previous section’s example of single-member plurality elections, having less of a political resource than some other people have does not always undermine self-respect, let alone cause insufficient self-respect. Consider, similarly, people too poor to make campaign contributions in American elections. Even if such citizens feel frustrated, I doubt that they would suffer the same loss of moral standing as people excluded from the franchise or prevented from voting, say. In these examples, inequalities where some people are below sufficiency levels cause insufficient self-respect, but inequalities where everyone is at or above sufficiency levels may not have as strong a positional effect; criticising this requires a different argument.

Again, I do not deny that political power can, through positionality, turn inequality into insufficiency. This may drive Christiano’s discussion of equal moral standing (Section 4): perhaps anything less than equal moral standing implies that one has no moral standing, that one is seen as sub-human. Perhaps, but the term ‘moral standing’ is hazy, and this argument needs fleshing out.

Axelsen and Nielsen do flesh out a similar argument, and it shows a direction that Christiano could take. In an important restatement of sufficientarianism, Axelsen and Nielsen characterise sufficiency in terms of ‘freedom from duress’, where duress means that ‘one is under significant pressure in central areas of human life, pressure that would impede any normal human being’s ability to succeed in a similar situation’, and freedom from duress requires ‘sufficient opportunities for succeeding’ (2015, p. 408). They imagine a society where each member of the ‘plebs’ has one vote but each member of the ‘aristocracy’ is then given a second vote. For members of the plebs, ‘their societal status, self-respect, and political influence are worsened in an absolute sense by their relative worsening’. This places the class of plebs under such duress as ‘would impede any normal human being’s ability to succeed in a similar situation’ (2015, pp. 419–20).

Now, that conclusion will not always be true; for example, if there are so few aristocrats that they remain in a permanent minority. I am not defending giving aristocrats extra votes, merely noting that this political inequality does not necessarily lead to insufficiency through positionality.

But if there are enough aristocrats, Axelsen and Nielsen’s fears would be realised, ‘in which case a person may become insufficiently free because of relative deprivation – but it is the insufficiency that creates a problem, not the inequality in itself’ (2015, p. 420). This is why Section 4 argued that Christiano saw some inequalities as problematic because of insufficiency. Christiano’s democratic theory might thus profit from Axelsen and Nielsen’s focus on sufficiency as freedom from duress.

So, the partly positional nature of political resources may or may not mean that inequalities cause insufficiency in a related resource. It is beyond the scope of this paper to explore when positionality has this effect. All I need to show is that as regards political resources, we cannot simply say that all that matters is everyone having the same amount.

6 David Estlund: Attacking Political Equality on the Basis of Political Sufficiency

Thus far, I have argued that Mark Warren wrongly draws sufficientarian conclusions from egalitarian premises and that Thomas Christiano wrongly draws egalitarian conclusions from sufficientarian premises. I will now discuss David Estlund, whose powerful challenge to political equality wrongly draws anti-egalitarian conclusions from sufficientarian premises.

This section is particularly important because it shows a different way in which sufficiency matters for egalitarian democratic theorists. Previously, I treated sufficiency and equality as rivals: democratic theorists who do not separate the two risk drawing invalid conclusions. Here, I treat sufficiency and equality as allies: sufficiency helps egalitarians defend themselves against some levelling-down objections.

Estlund dislikes political equality being pursued at excessive cost to political ‘quality’. Using the example of a voucher scheme for financing election campaigns, Estlund shows that for any given egalitarian distribution of campaign finance, a maximin distribution would increase the quality of political debate. Estlund is not arguing that political quality must override political equality. But the two can clash, and rigidly egalitarian arguments might permit low-quality politics (2000, 2008, pp. 195–198).

I agree with Estlund’s general argument but I have concerns about his use of the levelling-down objection. The levelling-down objection runs as follows. Critics of equality sometimes note that if equality is all that counts, it would be better in one respect if (a) everyone had 100 units of money than (b) half have 150 and half have 200. Yet (b) would be preferred by those who seek to maximise aggregate utility, and on Pareto-superior grounds, where a situation is better for at least one person and no worse for others (see e.g. Parfit 2000).

Estlund makes similar criticisms of political egalitarians. ‘Suppose that the only way to achieve equal (available) influence is by leveling down … If this low level is very low, the reduction in the total volume of deliberation might damage its epistemic value.’ Estlund’s example is a debate where the two sides can either have 50 and 60 min respectively, or 2 min each. Another example is equal campaign contributions: if so little money is raised that the issues are poorly discussed, then higher but unequal contributions could be preferable (2000, pp. 136, 148–149, 153–156, 2008, p. 195).

A crucial feature of Estlund’s critique, as I read it, involves political equality trumping other values, such as political quality. Giving political equality so much weight tells us nothing about the right level of the resource to be distributed, as long as everyone has the same. The required level of equal power might be so high as to permit control over all aspects of our lives, or so low as to amount to anarchism. The required level of equal deliberation could be so high as to require an infinite amount of time or so low as to leave almost no time for discussion. The required level of equal campaign contributions might be so high as to bankrupt most campaigns or so low as to prevent campaigning. Insufficient deliberation and insufficient money are the target of Estlund’s levelling-down objection, which we can now see as part of a broader problem: excessive weight on equality permits over-flexibility in the level of another resource.

We should thus distinguish between two forms of levelling-down. Strong levelling-down means equalising a resource below the sufficiency level. Weak levelling-down means equalising a resource at or above this level. As Section 2 explained, the ‘sufficiency’ level need not be the barest minimum needed for life or democracy, say. (I will not address strong levelling-up, where the resource is equalised above a maximum threshold. This may anyway be equivalent to strong levelling-down: levelling up to a very high degree of equal power, for example, may violate sufficiency by leading to transgression of basic interests.) Egalitarians who accept weak but not strong levelling-down can also be called sufficiency-constrained-levelling-down egalitarians (Casal 2007, p. 319).

The distinction between strong and weak levelling-down suggests that Estlund has overstated his case, because some supporters of political equality will not accept strong levelling-down. Faced with Estlund’s debating scenario, where the two speakers could have 50 and 60 min each or just 2 min each, some theorists will grudgingly accept the 50–60 distribution: this inequality is worse than both having 50 min each or 60 min each, but all things considered it is better than only 2 min each (for a similar point, see Christiano and Braynen 2008, pp. 400–401). On the other hand, most theorists would choose 40 min each if possible: both sides would get a decent hearing, and the small extra benefit from a 50–60 distribution would not negate the disrespect shown to or felt by the speaker with less time. Without a good reason for unequal times, such as both speakers agreeing that one of them has a more complex case to present, or agreeing to give the challenger longer (to counter incumbency advantages), 40 min each is better than 50 and 60.

Estlund’s point, of course, is that political equality is not so vital that the quality of the political process can be neglected. He is quite right. Nonetheless, his thought experiment does not show that the problem is political equality, but political equality where some are above sufficiency and others are below. Where everyone is above this line, political equality is less controversial. Most of us would accept levelling down to 40 min even with a small loss of quality.

So, by giving the example of 2-min speeches, which would prevent proper discussion of complex political issues, Estlund has made the reverse error to Christiano. Christiano defends egalitarianism, but sometimes objects to a combination of inequality and insufficiency, not inequality alone. This weakens his egalitarianism, since inequalities where everyone has a sufficiency are less troubling. Estlund attacks rather than defends egalitarianism, but unlike the voucher-scheme example, the 2-min speeches example suggests that the problem is a combination of equality and insufficiency – again, not equality alone. This weakens his anti-egalitarianism, because equality above the sufficiency level is less troubling.

Whether defending or attacking political equality, then, we should clarify when our focus is political equality and when it is political equality where some or all are below sufficiency levels. Failing to differentiate political equality and political sufficiency risks undermining our ensuing normative claims.

Political egalitarians may thus need to address the sufficient level of the resource being equalised to avoid strong levelling-down objections. Although equality and sufficiency may be competing distributive principles, sufficiency should also be part of an egalitarian theory which avoids equalisation at too low a level.

This is extremely important. Political egalitarians should not reject sufficientarianism because of the difficulty of agreeing on what is an acceptable minimum. (For an example of this egalitarian argument against sufficientarianism, see Foley 1994, p. 1241.) This is a significant difficulty with sufficientarianism, but egalitarians too must address it. ‘Political egalitarianism is a crude and implausible principle,’ writes Estlund (2008, p. 198). If political egalitarianism means defending political equality no matter what, I agree; this applies to egalitarianism more generally (Temkin 2009, pp. 157, 168). Political egalitarians should thus specify absolute thresholds below which political equality is unacceptable. Egalitarians, then, can accept that ‘sufficiency is not enough’ (Casal 2007, p. 296) and that equality is not enough. Without considering sufficiency, equality may mean we do not get enough of a resource. Sufficiency helps equality (Casal 2007, pp. 318–319, 321–326).

Upper thresholds could also be specified. For example, if issues are so complex that political equality can only be achieved by giving every citizen 50 min to state her or his case, a population of 10 million adults would need 950 years for the ensuing discussion. In terms of time for deliberation, this would allow the original participants to be equally treated, at the cost of leaving them equally dead. Time limits are one example of a practical limit on political equality (e.g. Brighouse 1995).

Two brief points are worth adding. First, political sufficiency would also help Estlund’s inegalitarian democratic proposals: his voucher scheme does not guarantee minimum funds for candidates. No system where money is distributed solely by citizen contributions can guarantee sufficiency, although two-party US elections are unlikely to see problems here. Meanwhile, Brighouse’s alternative proposal does provide a minimum threshold – subsidised election broadcasts for all candidates – but arguably at too low a level: capping campaign spending only at the level needed to reach all citizens would not help campaign quality much (Brighouse 2002, p. 68). Or, to put it another way, when we ask ‘sufficient for what?’, the answer – candidates’ messages reaching all voters – is too narrow. Brighouse’s scheme thus risks making the error that Estlund warns against, although a small restatement would avert this problem.

Second, the equality/sufficiency distinction implies that Estlund somewhat overstates the weight that certain democratic theorists place on political equality. He claims that six to eight studies defend political equality so strongly that they would permit insufficiency, but in my view only one study fully fits his description, Knight and Johnson’s (1997). Four studies probably contradict his claim that political egalitarians would prefer 2 min of deliberation each to a 50–60 split.^[2] Six studies partly contradict his claim that political egalitarians neglect political quality.^[3] Nonetheless, such comments are fleeting and often implicit; here and in much of the democratic theory literature, political equality gets far more attention than political quality. Estlund’s critique thus retains considerable force.

In short, Estlund is right that political equality is not all that matters, just as Christiano is right that political equality is extremely important. But sometimes what seems to drive Estlund’s concern is political equality at levels below sufficiency, just as what seems to drive Christiano’s concern is political inequality at levels below sufficiency. Political equality above the sufficiency level is less problematic for Estlund; political inequality above the sufficiency level is (sometimes) less problematic for Christiano.

This further highlights the value of the equality/sufficiency distinction for democratic theorists. Egalitarian democratic theorists cannot simply reject sufficiency: sufficiency helps us defend egalitarianism from levelling-down objections. Much the same could be said of the distributive justice literature more generally.

7 Conclusion

Careful comparison is at the heart of correct inference, whether in the natural and social sciences (e.g. King, Keohane, and Verba 1994, pp. 124–207) or in normative theory (e.g. Kamm 1996, pp. 17–42). If relevant variables are not appropriately handled, errors of inference can result from statistical analysis and controlled experiments in the natural and social sciences, and from examples and thought experiments in normative theory. Understanding interaction is often vital in both areas. For example, to say that hunger causes migraine cannot be quite right: hunger never causes migraine in most people, and only sometimes does so in those disposed to migraine. At least one other variable must be relevant to turn hunger into migraine. Understanding this interaction is the key. If this extra variable is a particular chemical in neurones, say, then migraine is not caused by hunger alone, or that chemical alone, but by the two together.

Something similar applies here. If the equality/sufficiency distinction is overlooked and political sufficiency is not controlled for in our examples and arguments, its independent effect or its interaction with political equality may be overlooked, leading to faulty normative inferences. So, conceptual stretching means that we sometimes criticise political equality or political inequality when what actually drives our objection, without us realising it, is political insufficiency, or political insufficiency interacting with political equality/inequality. This problem is probably widespread in political theory. Conceptual stretching, omitted variable bias and interaction effects thus deserve more attention in thought experiments. (For a similar concern about interaction effects, in thought experiments about retributive justice, see Goodin 1982, p. 9, and Slavny et al. 2021, p. 23.)

As regards democratic theory, it remains to be seen which political resources should be distributed equally, when minimum or maximum absolute or relative thresholds should be applied, and what the thresholds should be. A fully specified democratic theory will surely involve egalitarian and sufficientarian distributional principles (and more) at different points. For now, it is enough to conclude that even when there is a presumption in favour of political equality, this should not be the only way of distributing political resources; and even when a resource is to be distributed equally, we may need to specify the sufficient level of that resource. The conceptual tools developed in this paper, and the methodological insights into conceptual stretching, omitted variable bias and interaction effects, thus enrich our normative analysis.