Geographic Legislative Constituencies: A Defense

Noncomparative Constituency Localism

Noncomparative constituency localism is a popular view, then, that is ostensibly under pressure from criticisms that have recently been directed against the geographic definition of constituencies. This article ultimately argues that this pressure is ostensible only and that the localist view in fact remains tenable. But developing such an argument would be pointless if there were nothing to say in favour of noncomparative constituency localism in the first place. Accordingly, this section’s objective is to, first of all, underscore that noncomparative constituency localism is an a priori compelling view, for a number of different reasons.

A first reason why the localist view is attractive is that geographically defined constituencies sustain practices of partisan voter mobilisation. Geographic constituencies engender that constituency members are gathered (or concentrated) within a clearly demarcated area, which in turn makes it relatively easy and inexpensive for legislative candidates or parties to conduct campaign activities aimed at mobilising voters. Geographically concentrated electorates make it cheap and easy, for example, to hold local political meetings, to conduct door-to-door campaigns, to distribute political advertisement in broadcast media, etc. As Rehfeld writes, there is an important sense in which geographically defined constituencies “enable political campaigning” (2005, 174). Such campaigning and mobilisation activities are also normatively valuable. Town hall meetings, door-to-door campaigns, political advertisements in local media, and other forms of partisan mobilisation arguably motivate voters to turn out on Election Day. Furthermore, these activities help convey information to voters about the candidates or parties that stand in the election. Moments of partisan voter mobilisation also enable citizens to exact substantial policy pledges from their future representatives (cf. Beerbohm, 2016), and thus noncomparative constituency localism is first of all attractive because geographic constituencies sustain partisan voter mobilisation. As Melissa Williams expresses the point, retaining a “geographic basis of constituency definition . . . is important for the purpose of maintaining clear and close communication between representatives and constituents, and for constituting "town hall" meetings” (1998, 234). Seconding this line of thought, Rehfeld also writes: “This, then, forms a strong justification for local constituencies: They enable political campaigning, and the transferal of information between the candidates and their voters” (2005, 174).

A further reason why noncomparative constituency localism is attractive is that in addition to supporting partisan mobilisation, geographic constituencies also sustain the practice of constituency service—that is, the practice that legislators offer guidance to citizens on particular administrative procedures and/or confer with administrative bodies on behalf of individual citizens. The fundamental reason for this is that geographically defined constituencies render potential voters identifiable to legislators. As James Wilson has recently pointed out, it is an effect of geographically defined constituencies that legislators “can know (or can find out) that a citizen is a resident of their district” (2019, 199) and hence a potential voter in future elections. Local constituencies thus eliminate some of the risks that constituency service might otherwise hold for a legislator. Enabling legislators to verify that the people demanding constituency services are indeed resident in the legislator’s own district, geographic constituencies eliminate the risk of legislators expending time and energy on servicing individuals who belong to a different constituency and thus cannot electorally reward the legislator’s service. In its turn, this likely means that geographic constituencies facilitate or catalyse the provision of constituency service—at least more so, as Wilson notes, than do electoral systems in which “representatives do not [necessarily] know who their constituents are” and so “have no particular responsibility (and, perhaps, little motivation) to provide targeted constituent service” (2019, 199).

At the same time, it is arguable that constituency service is a desirable practice. In particular, its availability creates opportunities for citizens to contest administrative processes and decisions, thus aiding to make it public knowledge that administrative bodies exercise power on behalf of citizens and that public authority ultimately rests with citizens. As Wilson has put the point, “citizen authority is better respected when citizens have the opportunity to raise individual concerns about administrative process,” (2019, 200) and in a democracy such respect for citizen authority is presumably valuable. In a democratic political order, it presumably is important that citizens are generally recognised as the ultimate source of public authority, and to the extent that constituency service promotes that recognition, that practice is arguably desirable or, in Wilson’s words, “valuable” (2019 , 199). Geographic constituencies, for their part, facilitate this valuable practice, and thus it is a second argument in favour of noncomparative constituency localism that the geographic definition of constituencies sustains the provision of constituency service. Indeed, it is as Williams suggests (and as Wilson’s arguments strongly imply): maintaining a geographic basis of constituency definition is in part “desirable . . . because of the importance to many citizens of constituent service.” (Williams 1998, 234)

Noncomparative constituency localism is not only attractive, though, for reasons of partisan mobilisation and constituency service; it is a compelling view also for reasons that have to do with deliberation. For one thing, it is a feature of geographically defined constituencies that they sustain deliberation amongst coconstituents. Geographic constituencies engender that voters who live close to each other generally belong to the same legislative constituency. This proximity in turn means that coconstituents have ample opportunities to meet and thus communicate with each other, which is essential on all accounts for the practice of deliberation. One can define deliberation in a “thick fashion” as, for instance, Joshua Cohen (1989), who conceives of deliberation as a conversation between free and formally equal persons in which the participants appeal to mutually acceptable reasons and aim at reaching a consensus. Or else, one might prefer a “thinner” definition of deliberation that characterises it as “mutual communication that involves weighing and reflecting on preferences, values, and interests relevant to common concerns” (Mansbridge 2015, 27). But either way, it is clear that deliberation involves communication between citizens. Local constituencies, for their part, enable communication between members of the individual constituencies (by ensuring that neighbours belong to the same constituencies) and so, geographic constituencies arguably contribute to sustaining deliberation amongst coconstituents. There is, of course, the caveat that geographic constituencies will rarely, if ever, enable communication and deliberation between all members of a constituency. Given the large number of voters who are typically contained within a modern-day constituency, this is simply impossible. But the geographic rapprochement of coconstituents does enable communication and hence deliberation within the many subgroups that will be nested within each geographic constituency. As Rehfeld puts the point, we might “in each neighborhood . . . see real deliberation about candidates and policies. . . . [I]ndeed today many citizens attend candidate meetings that are held nearby and discuss politics with their neighbors” (2005, 172). Such practices of deliberation are also normatively worthwhile. It has compellingly been argued that deliberation promotes substantially good policy making (Landemore 2012), legitimacy in state action (Cohen 1989), and the expression of citizens’ mutual respect for each other (Gutman and Thompson 2004). And so, a third argument for noncomparative constituency localism is that geographic constituencies contribute to intraconstituency deliberation.

That said, it is important to note that geographic constituencies sustain deliberation not only amongst constituents but also within the legislature. To see this, notice that a geographic definition of constituencies requires legislators to attract votes from voters who reside within particular areas of the country. This creates strong incentives for legislators to heed and carry into the legislature the political preferences, values, and interests that are characteristic of the locally demarcated electorates to whom they owe their election and are electorally accountable (cf. Rehfeld 2005, 151–53). At the same time, it is likely that geographically differentiated electorates within a state have at least somewhat disparate interests, values, and preferences. Different locales typically differ in terms of landscape, climate, and infrastructure, and so living conditions usually differ from one place to another. Different locales are moreover likely to be sites for different kinds of economic activity, which results in the inhabitants having dissimilar socioeconomic profiles. Different areas within a country may additionally be homelands for culturally, linguistically, or religiously distinct groups. Thus, residents of different geographic areas are for a host of reasons likely to have disparate needs, interests, and political preferences. Geographically defined constituencies, for their part, are likely—because of how they structure the incentives for legislators—to transport this diversity into the legislature as members articulate and defend the diverse views of their own local constituents. And this diversity of perspectives is, in its turn, a key precondition for deliberation—regardless whether deliberation is conceived thickly or more thinly.

On a thick conception, deliberation refers to conversation that is in part characterised by an orientation towards consensus building. As Cohen emphasises with respect to his own, thick conception, “deliberation aims to arrive at a rationally motivated consensus” (1989, 23), and hence, it is conceptually necessary for deliberation in the thick sense that there is a starting point of perspectival diversity between those who participate in the deliberation; if deliberation is the communicative process of consensus construction, it is conceptually impossible that the deliberators already agree with each other prior to the process. Rather, they must come to the table with initially disparate opinions, and viewpoint diversity thus represents a precondition for deliberation conceived thickly. Indeed, Cohen himself stresses that initially, “members [of a deliberative venture] have diverse preferences, convictions, and ideals” (1989, 21).

By contrast, on a thin conception of deliberation, the quest for consensus is removed from the notion of deliberation. Here the notion refers more broadly to any form of mutual communication in which participants thoughtfully consider and weigh preferences, interests, and values in relation to some common concern. But here, too, viewpoint diversity remains an important basis for deliberation, for the thin conception stresses that deliberation involves an element of intellectual care or mindful consideration. As Jane Mansbridge emphasises, the thin or minimalist conception seeks to “capture the elements of care and thoughtful consideration [that are] central to the . . . meanings that in ordinary language adhere to the term deliberation” (2015, 29), and for people to engage in such careful reflection and dialogue, they typically need to be exposed to diverse viewpoints. For people to discuss an issue thoughtfully, as the thin conception demands, they typically need to be challenged on their preconceived notions. They must be prodded to revisit their certainties about the interests and values that are at stake, the interests and preferences that are legitimate, and so on. Such prodding is relatively rare, though, in discussions between like-minded people and principally occurs in exchanges between people who at least initially have diverse views, interests, and political judgments. Accordingly, thin deliberation is in practical terms much facilitated by the presence of initially diverse views amongst deliberators, and as such it is arguable that perspectival diversity is a practical precondition for thin deliberation as much as it is for thick deliberation. As Anne Phillips (1998, 159) has expressed the point, deliberation cannot get off the ground unless deliberators are initially divided on what is important, just, or politically expedient, on whose interests are affected in what ways and so forth (cf. Sunstein 1991, 33–34). Geographic constituencies, for their part, help to ensure that such diversity is present in the legislative chamber, through the representation of diverse geographical perspectives. ⁷ And hence, there is an argument that geographic constituencies are noncomparatively desirable, not only because they support intraconstituency deliberation but also—perhaps especially—because they enable deliberation at the level of the legislature, where government budgets are passed and legislation is actually crafted and enacted. ⁸

Finally, there is an equality-centred case for noncomparative constituency localism, in that geographically defined constituencies help to equalise citizen’s opportunities for political influence. For a better understanding of this, two points are important. Firstly, it is a feature of geographic constituencies that they disperse legislator attention across the demos. Geographic constituencies engender that incumbent legislators must vie for the votes of specific, geographically delimited electorates, and as such they create incentives for legislators not to direct their joint attention onto the very same segments within the demos, for instance, onto the inhabitants of areas where people are generally well off, well educated, and most likely to be electorally active. ⁹ Rather, legislators will want to spread their attention, each aiming at the specific voter area they depend on for re-election, and so, geographic constituencies help to bring it about that legislator attention is distributed more-or-less evenly across the demos, with all citizens enjoying roughly the same degree of legislator attention.

Secondly, it is important to see that citizens of a democracy can exercise political influence in two broad ways: they can either participate in elections or they can exercise influence through nonelectoral channels, for example, by taking part in marches, writing letters to members of the legislature, and so on. For citizens to be able to exercise influence in the latter, nonelectoral ways, though, it is typically necessary that they enjoy some minimum attention on the part of the legislators they address. As an example, consider the opportunity to influence politicians by the means of letter writing. This opportunity will be meaningfully available to citizens only if their letters are actually read by the relevant legislators—or, more realistically, their staff—and the legislators (or staff) register the messages. They need not agree with or adopt the position of the message. But the legislators (or their staff) do need to present some basic receptivity or attention to the citizens’ communications, otherwise there is no meaningful sense in which letter writing forms an opportunity to influence politics. This also applies to other important ways of exercising nonelectoral political influence. For citizens to be able to exercise political influence through marches or direct conversations with legislators, say, it is practically indispensable that the members of the legislature have some minimum attention to spare for the citizens in question. Accordingly, there is a general sense in which important opportunities for political influence are exercisable, or available to citizens, only to the extent that citizens benefit from, or enjoy, a certain degree of legislator attention. Geographic constituencies, for their part, help to ensure that such attention is available to citizens across all parts of the demos, and so there is an argument that the geographic definition of constituencies helps to realise equal opportunity for political influence, especially equal opportunity for nonelectoral political influence. As detailed below in section 5, it cannot be said that geographic constituencies robustly guarantee equal opportunities for political influence. Against this strong claim, there is notably too much evidence that democracies in practice afford greater opportunities for political influence to the affluent than they do to the poor. But the considerations above do provide solid reasons for thinking that geographic constituencies facilitate this objective.

That said, it is clear as well that equal opportunity for political influence is a valid objective. The existence of equal opportunity for political influence helps to render procedurally fair the process of political decision-making and thereby arguably helps to render legitimate the political outcomes that eventually issue from it (Rawls 1973, 221). Equal opportunities for political influence may additionally be an effective way to publicise citizens’ equal moral status (Christiano 2008, ch. 3), while also being instrumental for bringing about conditions of social equality more generally—that is to say, a condition where no one permanently enjoys more power or better status than others (Kolodny 2014, 308). Equal opportunity for political influence is thus in many ways an attractive objective, and hence it is a final argument in favour of noncomparative constituency localism that geographically defined constituencies help to realise this very objective. ¹⁰

Bias, Disproportionality, and Gerrymandering

Noncomparative constituency localism is a popular view, then, that has much going for it. Yet, in recent years a number of arguments have arisen that are critical of geographic constituencies, potentially calling into question the notion that they are noncomparatively desirable. For example, it has been argued that geographic constituencies are normatively suspect as they are systematically biased against certain voter groups and hazard the legislative pursuit of the common good. How challenging are such critiques for noncomparative constituency localism? Do they dispositively undercut the localist view?

To examine this question, consider first a set of three critiques that respectively take aim at geographic constituencies’ purported bias, their alleged propensity to drive vote-seat disproportionality, and their vulnerability to gerrymandering.

With respect to the first charge, I evoke a critique pressed especially by Thomas Pogge who vividly argues that the use of local constituencies turns dispersed voter groups into local electoral minorities, rendering inter alia “Blacks, Hispanics, Asians, homosexuals, environmentalists, and single mothers . . . utterly incapable of voting even a single one of their own into the legislature.” Meanwhile, Pogge continues, no such problem exists for “geographically concentrated minorities such as the Mormons,” who are consistently “able to secure legislative representation” for themselves. There is no such problem either for “the preponderant Anglo-Saxon Protestant majority.” The electoral difficulty affects only those voter groups that happen to be geographically dispersed, and such a structural inequality in groups’ prospects for electoral success is normatively improper—or so argues Pogge who qualifies it as “unfair” (2002, 27).

The second critique concerning disproportionality, meanwhile, is pressed by Andrew Reeve and Alan Ware, who together argue that geographic constituencies cause parties to receive legislative seat shares that are not proportional to their share of obtained votes. When constituencies are local, the two theorists reason, it is more-or-less inevitable that the supporters of some parties form minorities within some (or even all) constituencies. In turn, this means that the relevant parties do not win seats in those constituencies and the parties hence receive a legislative seat share that is proportionally smaller than their overall share of votes. As Reeve and Ware write, in a local-constituency system it is “virtually impossible for constituency boundaries to be drawn so that there is an exact correspondence between the distribution of the total vote and the representatives actually elected.” Geographically defined constituencies are hence bound to drive vote-seat disproportionality. And according to Reeve and Ware, this is a normative liability that constitutes a “major objection to territorial subunits as the basis of representation” (1992, 120).

However, Reeve and Ware are not only worried that geographic constituencies cause vote-seat disproportionality; they are also concerned—and this is the third critique—that geographic constituencies invite undue, partisan manipulation. In particular, they argue that when constituencies are defined in geographic terms, there is an unavoidable necessity to make decisions about the precise location of the borders separating individual constituencies from each other. These decisions, though, affect the population structure of the individual constituencies and thus the chances that different candidates have of becoming elected there. As such, political parties have stakes in the placement of district borders, and Reeve and Ware hence charge that geographic constituencies foster normatively unsound forms of partisan “manipulation,” notably the kind of noxious gerrymandering practices that have “a long history in the United States” (1992, 121).

Each of the above charges seems serious. If geographic constituencies are biased against particular voter groups, drive vote-seat disproportionality, and also prepare the ground for gerrymandering, that seems to militate very powerfully against the geographic definition of constituencies and the view that such constituencies are noncomparatively desirable. However, a closer inspection reveals that none of the charges have much merit.

Consider first the case of bias. When Pogge complains that geographic constituencies are biased against geographically dispersed voter groups, he explicitly refers to the geographic constituencies in the United States, which return only one legislator each. But a constituency magnitude of one is no necessary feature of geographic constituencies. Geographically defined constituencies can just as well return two, seven, or any number of legislators chosen by the electoral system designers. And once the constituency magnitude exceeds one, the bias that Pogge worries about abates. When constituencies return more than one legislator each, it is not only the largest voter group within each constituency that can elect its preferred candidate for legislative office; on the assumption of plurality selection rules, ¹¹ the largest minority voter group can do so as well. Depending on the exact magnitude value, this may also be possible for voter groups that are even smaller. For instance, if the constituency magnitude is three, it is possible also for the second-largest minority group to vote its preferred candidate into office, and with even larger magnitude values, electoral success becomes available for even further segments within the constituency. Dispersed voter groups thus become less and less electorally disadvantaged the more the constituency magnitude increases. Under high-constituency magnitudes, dispersed groups actually have relatively robust opportunities to be electorally successful, and it is not clear therefore that the bias against dispersed voter groups is, indeed, a feature of geographic constituencies per se.

In fact, there are strong reasons to suggest that such bias has nothing to do at all with the geographic nature of constituencies. To appreciate this, consider a hypothetical constituency scheme in which it is different age groups, rather than geographic groups, that form the legislative constituencies. So long as each constituency elects a considerable number of legislators—say ten, using plurality selection rules—it will genuinely be possible for a broad variety of voter groups to elect their preferred candidates for legislative office. But if the age-based constituencies return just one legislator each, it is only the majoritarian voter groups within the relevant age brackets that can succeed in electing preferred candidates for Parliament. Voter groups that straddle many different age brackets, by contrast, will form persistent electoral minorities and thus find themselves at an electoral disadvantage—just as do geographically dispersed voter groups in single-member, local constituencies. Accordingly, the bias that Pogge worries about is not really a consequence of constituencies being defined geographically. It is the result of constituencies having a small magnitude, and Pogge’s bias critique is thus empirically inaccurate.

But not only is Pogge’s charge empirically inaccurate, it is also normatively unpersuasive. In particular, the charge turns on a notion that it is somehow bad or unfair if different voter groups have unequal prospects for electoral success. However, this is a controversial idea, with prominent theorists arguing that equal chances for electoral success for different voter groups is no desideratum—at least not from the perspective of political or social equality (cf. Beitz 1989, 133–40; Wilson 2019, 197–98; Kolodny 2014, 321–26). ¹² Yet, no explanation is provided in the bias critique as to why equal prospects for electoral success would be a desideratum, or why its absence would constitute a deficiency. The critique merely stipulates this to be the case. This lack of justification does not, of course, render the idea untrue, but it does render unconvincing the normative premise in the bias objection; there simply is no reason for anyone to accept this premise. The objection’s empirical premise, meanwhile, is importantly inaccurate. So, on closer inspection it is clear that the bias objection actually poses no substantial challenge to noncomparative constituency localism.

Partly similar considerations apply to the second charge of disproportionality. The critique’s premise that vote-seat disproportionality is objectionable is not explicitly motivated by Reeve and Ware, but it arguably is acceptable nonetheless, as vote-seat disproportionality betrays a failure to offer impartial treatment to the parties running in the election. Where parties’ vote and seat shares are not proportional to each other, parties can obtain an equal number of additional votes and yet receive differentiated numbers of additional legislative seats, which arguably represents an undue departure from standards of impartiality.

However, the situation is different with regard to the predictive claim that geographic constituencies drive vote-seat disproportionality, as can be seen by contemplating a constituency scheme in which the constituencies are defined in terms of geography and the magnitude is one. Here it is only the votes cast for the winning candidate or party in each constituency that positively produce a seat for a party in the legislature. Ballots cast for other candidates and parties are all “wasted” and thus the electoral system overall effects a disjuncture between parties’ vote share and seat share. A party may, for instance, receive the second-highest number of votes in numerous constituencies and yet receive a legislative-seat share that is smaller than its overall vote share. But if the districts’ magnitude is now increased to ten, say, the vote-seat disproportionality diminishes. A party that receives the second-highest number of votes in a constituency now receives one (or more) seats for that constituency. This may be true for smaller parties, too, and thus vote-seat disproportionality can be expected to decrease generally, across all parties. Indeed, it is a general consensus among empirical electoral systems researchers that “proportionality increases as district magnitude increases” (Gallagher and Mitchell 2005, 14; cf. Taagepera and Shugart 1989; Lijphart 1990) and the disproportionality critique is thus flatly mistaken about the causal sources of disproportionality. Vote-seat disproportionality is caused by low constituency magnitudes, not the geographic nature of constituencies. Accordingly, the charge of disproportionality is fundamentally unwarranted (even though its normative premise is fundamentally sound) and as such it does not pose any challenge either to noncomparative constituency localism.

With respect to the third and final critique, a similar point can be made again. Provided that gerrymandering refers to the definition of constituency boundaries for partisan advantage, it is virtually impossible to gerrymander constituencies whose magnitude is larger than one. Magnitudes higher than one make it extremely difficult to predict the electoral consequences that will follow from particular district maps and to anticipate what map is preferable from a partisan point of view. Accordingly, “once district magnitude is two or more, gerrymandering becomes virtually impossible” (Williams 1998, 219).

However, the point is not merely that gerrymandering is likely only when geographic constituencies have a magnitude of one. It is rather that the constituencies’ geographic nature does not seem to be causally effective at all in the generation of gerrymandering risks. If constituencies are based not on geography but on age, and they each have a magnitude of one, it is likely that the decision to delimit constituencies at one age limit rather than another affects the constituencies’ political profiles as well as eventual electoral results. The viewpoints and political preferences that are dominant within a constituency are probably not the same if the constituency comprises citizens who are aged between 20 and 29, or if it instead consists of citizens who are aged 25 to 34. Districters will thus be tempted to define the age limits of constituencies with partisan considerations in mind. But if so, the risk of gerrymandering has very little to do with the geographic nature of constituencies; rather, it attaches to single-member constituencies, and as such it is an unwarranted allegation that geographic constituencies invite gerrymandering. It is single-member constituencies that do so, and noncomparative constituency localism thus also remains unharmed by the charge of gerrymandering.

Injustice, Inefficiency, Common-Good Neglect, and Limited Agenda Control

That noncomparative constituency localism is left unscathed by the critiques from bias, disproportionality, and gerrymandering is not the end of the story, though. There are further charges against geographic constituencies.

For one thing, there is a bundle of criticisms that all relate to the behaviour that geographic constituencies risk sparking amongst the members of the legislature. In particular, the thought here is that geographic constituencies give legislators a powerful electoral incentive to attend to the interests that are peculiar to their own local constituents, causing elected representatives to prioritise the securing of parochial advantages for their own, local electorates. And to critics, this is unacceptable for a number of reasons. Pogge (2002, 49), for instance, alleges that legislators’ pursuit of parochial advantages may lead to legislative resource allocation decisions that hamper the efficient achievement of stated policy goals. Pogge (2002, 49) also argues that the pursuit of parochial benefits may lead to allocation decisions that fall short of standards of justice. And other commentators further worry that legislators’ preoccupation with parochial benefits may weaken the legislatures’ ability to discern the common, national good and to pursue goals that are in the interest of all citizens. Notably, this is the concern of Kris Deschouwer and Phillipe van Parijs, who in a discussion of the Belgian electoral system remark that it is a risk with geographic constituencies that legislators might “simply express . . . the demands of their own . . . group” and lack “the spirit of accommodation that is needed for a divided society to be smoothly governed” (2013, 126, cf. 113). However, this concern is also expressed by others, for example Reeve and Ware, who caution that “territorial representation is likely to . . . make it more difficult for common interests to be weighed adequately in relation to particular geographic ones” (1992, 119).

Moreover, there is one last critique against geographic constituencies that deems them responsible for limiting citizens’ agenda control. As pressed by Thomas Christiano, this charge holds that when constituencies are defined geographically, it is “likely that geographical issues will become important in campaigns,” simply because of the way local constituencies structure the incentives and attention of legislators as well as candidates. But in Christiano’s way of thinking, this unduly curtails ordinary citizens’ control over the political agenda. In his view, “the power to decide whether geographic considerations are important is an agenda setting power that should be held by citizens,” and so Christiano charges that geographic constituencies “take power out of the hands of citizens” (1996, 227).

Contrary to the critiques from bias, disproportionality, and gerrymandering, this further set of complaints about geographic constituencies is warranted. Each line of critique invokes a normative benchmark that is basically sound—though not always well explained. The arguments’ empirical/predictive claims are also sound. True, one might remark in relation to the critiques from risk of injustice, risk of inefficiency, and risk of common-good neglect that the risks abate when geographic constituencies return but a part of the legislature and some legislators are elected in a separate, at-large constituency, as is the case in Germany and other democracies with so-called mixed electoral systems. In other words, one might reason that the risks of injustice, inefficiency, and common-good neglect dissolve when geographic constituencies are not the only basis for legislative elections but are integrated instead into a mixed-members assembly. ¹³ But even if true, this reference to mixed-member assemblies does not disprove or cast doubt on the central terms of the critiques in question. The case of mixed-member assemblies arguably shows that some undesirable consequences of geographic constituencies can be controlled and limited with judiciously chosen counterbalancing institutions—in particular a second, national tier within the legislature. However, this is no reason to doubt or reject the criticisms’ more basic claim that geographic constituencies pose risks in the first place of injustice, inefficiency, and/or common-good neglect. These claims remain valid despite the potentially counterbalancing effects of a national tier, and hence it remains legitimate to charge that local constituencies pose risks of inefficiency, injustice, and common-good neglect, and that they also limit citizens’ agenda control.

That said, it is not clear that any of these latter, valid complaints dispositively rule out noncomparative constituency localism. If the criticism revolving around injustice were not merely that geographic constituencies carry a risk of injustice in resource allocations, but rather that they demonstrably lead to resource allocations that are systematically and incontrovertibly unjust, then the criticism might indefeasibly undercut noncomparative constituency localism, such that the position can no longer be rescued. But it is not clear that the “risk-of-injustice” complaint—or indeed any other of the four valid charges—is sufficiently grave to form such an indefeasible reason against noncomparative constituency localism. Rather, the valid criticisms seem to form defeasible, pro tanto considerations against noncomparative constituency localism. So, to undermine that position, the criticisms need to outbalance, jointly or individually, the reasons that speak in favour of noncomparative constituency localism. And it is not clear they do.

The considerations from section 2 that support noncomparative constituency localism are first of all normatively very weighty. Deliberation in the legislature and among constituents—equal opportunity of political influence, partisan voter mobilisation, and possibly also constituency service—are major democratic goals that weigh heavily in the thinking of committed democrats. This set of goals is also fairly varied. So, for these goals to be defeated in the overall balance of reasons, the countervailing considerations need to be exceptionally salient and diverse, and it is not clear that the valid criticisms against geographic constituencies satisfy those criteria. That geographical constituencies limit the citizens’ agenda control and pose some risks of injustice, inefficiency, and common-good neglect are certainly important considerations. But without a substantial explanation from local-constituency critics, it is not clear that these considerations are sufficiently grave to tip the balance of reasons against noncomparative constituency localism. The criticisms are perhaps on a par with the considerations speaking in favour of the localist position; but given the normative significance and diversity of the positive considerations, it is not evident that the negative criticisms can prevail and warrant a rejection of noncomparative constituency localism. For that, more seems required.

Are the Benefits of Local Constituencies Exaggerated?

The argument developed so far is that noncomparative constituency localism is a philosophically respectable position, in part because some the criticisms against it are spurious and in part because the challenges that are indeed valid fail to outweigh the diverse and powerful considerations that speak in its favour. But perhaps the latter considerations are exaggerated? Perhaps geographic constituencies are not quite as advantageous as the argument makes them out to be? This is an important challenge that critics of the present argument might press in three distinct ways.

Firstly, critics might challenge the thought that geographic constituencies support intraconstituency deliberation. Picking up on the point that deliberation presupposes an initial diversity of viewpoints, critics might argue that people living in the same area tend to be ideologically homogenous, share similar outlooks, and therefore are unlikely to engage in the give and take of diverse reasons and careful reflection that constitutes deliberation proper—though they might, indeed, discuss with each other. ¹⁴

Secondly, critics might challenge the claim that geographic constituencies sustain equal opportunities for political influence. Drawing on empirical studies by Larry Bartels (2008), Martin Gilens (2012), and others (e.g., Schakel et al. 2020; Elsässer et al. 2021) on democracies’ policy responsiveness to citizen preferences, critics might say that in some democracies (notably the United States), affluent citizens verifiably exercise greater political influence than do poor and middle-class citizens, and that it hence is not the general case that geographic constituencies protect equal opportunity for political influence. Equal opportunity for political influence, the critics might claim, can be sorely lacking despite the use of local legislative constituencies.

Thirdly, critics of the present argument might take aim at the claim that local constituencies create diverse legislatures, thus enabling legislative deliberation. In particular, they might argue that in geographically small countries there is little difference between different locales and that citizens’ place of residence therefore does not reliably track variations in citizens’ interests and political preferences. Citizens of small countries such as Luxembourg or Israel, for instance, certainly have diverse views on what is politically desirable, expedient, and morally right; they might also have starkly different interests. But according to critics, these differences may not have any geographic basis or manifestation, and accordingly they might argue that in geographically small countries local constituencies do not actually help to render the legislature diverse and deliberative.

All these exaggeration objections are important for they have the potential to unsettle the balance of reasons that sustains noncomparative constituency localism. To the extent that the exaggeration objections are warranted, the case for noncomparative constituency localism weakens, and it becomes more arguable that the downsides of geographic constituencies do outweigh their benefits, such that noncomparative constituency localism ultimately becomes unjustifiable. However, closer inspection shows that with the exception of the last variant, the exaggeration complaints are not actually sustainable.

Of course, it is likely that views and interests converge among those who live in the same area. If this were not the case at all, it would not be possible to argue—as it was in section 2—that geographic constituencies help to carry locally specific interests and political preferences into the legislature. But it is most likely an overstatement that residence in a particular place produces such homogenous views that deliberation proper becomes impossible. Absent very peculiar empirical conditions, the inhabitants of any area will be of different genders and varying ages; they will work in different economic sectors, if they are part of the work force at all. Some inhabitants of a given area will be able-bodied whereas other won’t; some will be single, some will be married, some will be parents, and so on. So, in almost any locale, there will be a tremendous amount of differentiation between inhabitants that potentially translates into disparate preferences, interests, and contrasting interpretations of the interests and preferences that are shared within the local community. Furthermore, local election results are typically nonunanimous. So, while local populations are likely to have some views in common, it is presumably exaggerated to view them as enclaves that are ideologically homogenous to the point of being incapable of deliberation proper. ¹⁵

The charge of exaggeration against the argument from equal opportunity for political influence, for its part, has the weakness that it misreads the argument it attacks. The objection seeks to cast doubt on the equality argument by citing empirical cases in which democracies elect legislators through geographic constituencies, but equal opportunity for political influence is lacking nonetheless. This would indeed be a powerful rejoinder, if the argument was that geographic constituencies single-handedly realise equal opportunity for political influence. As it is stated, though, the argument from equal opportunities for political influence does not make such a strong claim; it more moderately contends that geographic constituencies help or contribute to sustaining this goal, which is consistent with observing that some democracies lack equal opportunities for political influence despite using geographic constituencies. Furthermore, it would, in fact, be deeply implausible to make the stronger claim presumed by critics. Realising equal opportunity for political influence is obviously not a monocausal affair that depends only on how constituencies are defined. It also depends on factors such as the extent to which private money funds politics, ¹⁶ the extent to which there is a general political culture of regarding all citizen views as equally worth of consideration, and so on. Thus even though the exaggeration objection is empirically accurate—some democracies really do lack equal opportunity despite using geographic constituencies—it does not call into question the argument that geographic constituencies aid the realisation of equal opportunity for political influence.

With small-country objection the situation is different, though. It probably is an accurate analysis that geographic constituencies make no distinct contribution in geographically small states to sustain legislative diversity and deliberation, and so it is necessary to qualify the present defense of noncomparative constituency localism. The argument presented is robust only in the context of relatively large, geographically extended democracies. In small states, by contrast, the reasons speaking in favour of the localist position are weakened, as it is not clear that local constituencies here meaningfully further the presence of diverse perspectives within the legislature, though it probably is arguable that they also here sustain constituent deliberation, partisan mobilisation, equal opportunity of political equality, and constituency service. Yet, this more restricted set of considerations may not suffice to clearly outweigh the drawbacks of geographic constituencies, and in small democracies it is hence less clear than in the case of extended states that noncomparative constituency localism is an all-in-all justifiable position.

Conclusion

Is noncomparative constituency localism philosophically justifiable? This article has argued that it is—at least within certain limits. It has been argued that noncomparative constituency localism is appealing in the first place because geographic constituencies foster partisan voter mobilisation and sustain practices of constituency service, because they help to sustain deliberation among constituents and in the legislature and because they aid the realisation of equal opportunity for political influence. Additionally, the article has defended noncomparative constituency localism against a range of critical arguments that potentially undermine the localist position. It has shown that it is unwarranted to charge geographic constituencies with being biased against geographically dispersed voter groups, with causing vote-seat disproportionality and with inviting gerrymandering practices. With respect to the charges that geographic constituencies limit citizens’ control over the political agenda and entail risks of unjust and inefficient resource allocations as well as risks that legislators neglect the common good, the article has argued that the concerns are valid but insufficient to outbalance the diverse and normatively weighty advantages of local constituencies. Finally, the article has critically considered several variants of the charge that its defense of noncomparative constituency localism exaggerates the benefits of local constituencies. The article has notably deflected the charge that geographic constituencies cannot sustain constituent deliberation as local populations are ideologically too homogenous to engage in deliberate proper, and it has equally deflected an objection that interprets evidence of unequal political influence in existing democracies as proof that geographic constituencies do not actually promote equal opportunity for political influence. However, the article concedes that in geographically small democracies, local constituencies do not necessarily foster legislative deliberation, and hence the article qualifies its defense of noncomparative constituency localism: geographic constituencies are noncomparatively desirable primarily in the context of geographically extended democracies, but perhaps not in the case of small states. By defending this philosophical view, the article has not put forward a particularly avant-garde claim; noncomparative constituency localism is a commonly held view. The article has not established either that geographic legislative constituencies ought to be implemented (or retained) in political practice; such a judgment requires a further analysis looking specifically at how geographic constituencies compare normatively to alternative constituency schemes. But the argument presented here is important nonetheless given that geographic constituencies have of late faced such stiff criticism that it has been fundamentally unclear whether the geographic definition of constituencies is desirable in any sense at all.