5 - THE IMPACT OF FEDERALISM (Nam)

THE IMPACT OF FEDERALISM

18. Kent Weaver

18. Kent Weaver is professor of public policy and government at George-town University and a senior fellow in the Governance Studies Pro-gram at the Brookings Institution. He has written widely on institutional and electoral reform issues in industrial democracies, and is the editor of The Collapse of Canada? (1992) and the coeditor of Do Institutions Matter? Government Capabilities in the United States and Abroad

(1993). This essay originally appeared in the April 2002 issue of the

Journal of Democracy.

Federalism and electoral rules are usually seen as two distinct mecha-nisms for managing societal conflict in general and territorial conflicts in particular. Electoral rules that provide some element of proportion-ality in legislative elections, for example, are frequently viewed as a “consociational” mechanism that allows many different groups to ex-press their interests rather than polarizing them around a single dominant cleavage. Such electoral rules can be particularly effective at protecting minority interests when they are combined with other con-sociational features, including oversized coalitions with informal norms of cabinet selection (in parliamentary systems) that reach across major cleavages.

Federalism, on the other hand, is viewed as a quite different mecha-nism for limiting the rule of national majorities: Instead of building consensus across cleavage lines at the national level, conflict is man-aged by devolving decisions to geographic subunits, where decision making in turn may be managed through either majoritarian or major-ity-limiting mechanisms. Devolution may in fact leave regional minor-ity groups—such as Spanish speakers in Catalonia, francophones in Ontario and the Canadian prairie provinces, or Catholics in Northern Ireland—less protected than they would be if the central government were to take a more active role.

As this simple example suggests, the interaction between federalism

R. Kent Weaver

57

and electoral rules needs to be carefully examined. Electoral systems can vary in an almost infinite number of ways. In reviewing the institu-tional effects of different sets of electoral rules, it is helpful to begin with the polar opposite cases—single-member-plurality (SMP) systems, also known as first-past-the-post, and closed-list proportional represen-tation (PR)—and then discuss intermediate cases. The focus here will be on legislative elections rather than selection rules for the chief ex-ecutive in systems where the latter is separately elected.

Single-member plurality. SMP electoral rules, as the name implies, involve one-person legislative districts in which the candidate who receives the most votes is declared elected, regardless of whether that vote share is a majority or not. The very large academic literature on electoral systems has identified a number of consequences associated with SMP electoral rules:

• SMP tends to restrict the number of political parties that are po-tential contenders for office, and creates a tendency toward a two-party system—for example, in the United Kingdom, the United States, and New Zealand (prior to 1996).

• SMP facilitates the creation of single-party majority governments by turning pluralities of votes into a majority of legislative seats, al-though this outcome is by no means guaranteed. Furthermore, it is less likely to occur when territorially based cleavages allow minor parties to gain a significant share of legislative seats because their supporters are concentrated in specific regions.

• SMP leads to some votes being “worth more” than others in vote-to-seat conversions. In particular, relatively small parties whose support is diffused across the whole political system without a territorial strong-hold are likely to be severely punished, while parties of a similar size with geographically concentrated support may get a bonus.

• SMP tends to lead to centrist, nonideological, pragmatic, “broker-age” politics as parties compete for the “median voter.” Yet significant policy swings may occur when there is a change in government if par-ties move away from median-voter positions on some issues.

• SMP increases incentives for strategic voting, since a vote for a minor-party candidate may be wasted, leading to a voter’s least pre-ferred candidate getting elected.

Critics of SMP have noted that it may have a number of additional (mostly harmful) consequences. When used at the national level, it may exacerbate regional cleavages, since it tends to exaggerate the advan-tage enjoyed by the largest political party in a region in vote-seat conversions and to punish smaller parties. In combination with the ten-dency of SMP to promote single-party majority governments (in parliamentary systems), there is a danger that in a country with strong territorially based cleavages one region may end up dramatically over-represented in the governing party, while another region may end up

58 The Impact of Federalism

with virtually no representatives in the governing-party caucus or the cabinet. Perceptions of regional exclusion are therefore reinforced.

When used in territorial subunits, SMP electoral rules may allow the majority social group in that territory to govern alone, thereby excluding ethnic, linguistic, or religious minorities from political power (as in North-ern Ireland prior to the imposition of direct rule from London). SMP may also allow a political party that is ambivalent or even hostile to national unity to gain a strong power base through control of a single-party major-ity government in a state or province, which can then be used to promote autonomist or even separatist policies and popular sentiment.

It should be noted that SMP is something of a dinosaur: Most parlia-mentary systems in the advanced industrial world—even those without severe territorially based cleavages—have moved away from strict single-member first-past-the-post systems, generally to some sort of proportional representation. New Zealand, for example, recently moved to a German-style system of mixed-member proportional representation (discussed below). Even the United Kingdom has adopted elements of proportionality for elections to the Scottish Parliament, the Welsh As-sembly, and the European Parliament, and a debate has been opened about introducing it for House of Commons elections.

Proportional representation. At the other end of the spectrum of electoral systems are those using proportional representation. Ballots are cast in multimember districts for a party list rather than an indi-vidual candidate, and the party receives seats based on its share of the vote in that district. There are of course many variations among propor-tional representation systems, most notably in three areas: the number of seats filled in each electoral district (district magnitude); whether or not there is a legal minimum threshold that a party must pass in order to win any seats; and whether voters have any choice over individual candidates or must simply endorse a pre-ordered party list. Overall, proportional representation generally has the following effects:

• PR increases the number of parties that are likely to compete in elections, although the magnitude of the increase depends on the spe-cific electoral rules, notably the threshold and the size of electoral districts.

• Due to the proliferation of parties, PR increases the number of “non-throwaway” choices available to voters, although the number again varies depending on district magnitude and legal thresholds.
• PR makes it extremely unlikely that a single party will gain a majority of seats in the legislature, thus making either minority or coa-lition governments the norm in parliamentary systems.

• PR therefore increases the probability that the composition of a gov-ernment and its policies will be decided not by the election but after it, during negotiations among party leaders. This outcome may be mitigated, however, by pre-electoral alliances and common electoral programs.

R. Kent Weaver

59

• In most cases, PR weakens the ability of legislators in parties that operate nationwide to act (and to be seen by constituents to act) as regional representatives, since they are likely to be bound by party discipline. PR also weakens individual legislators’ incentive to act as providers of constituency services, since the ability of voters to reward individual MPs is lessened (although this effect is weaker in open-list systems).

• PR tends to lead to programmatic parties that seek to make dis-tinctive appeals to relatively narrow shares of the electorate.

Proportional representation may also interact with federalism in im-portant ways. For example, we might expect that federalism would reinforce PR’s tendency to produce a profusion of parties. The combi-nation might especially spur movement toward territorially focused parties, since the entry barriers for small parties are relatively low and the chance that a modest-sized party may share power at one or both levels of government is higher than under SMP.

Intermediate and mixed systems. A number of intermediate alterna-tives exist between SMP and PR. The alternative vote (AV), used in elections for the Australian House of Representatives, uses single-mem-ber districts but compels voters to list not just a single choice but a ranked set of choices; votes for the least-favored candidates are redis-tributed in stages until one candidate wins a majority. The two-round majority system used in elections for the French Chamber of Deputies also employs single-member districts and in most cases produces major-ity support for a single candidate. Both systems are likely to stimulate the existence of more than two parties both within and across districts, but fewer than under PR rules. The single transferable vote (STV) oper-ates in multimember districts as in PR, but voters rank individual candidates without regard to party, thus weakening party leaders’ con-trol over the selection of legislators. Because of the practical difficulties of vote counting and allocation under STV, however, both the district magnitude and the number of parties are only modestly greater than under SMP.

Complexities in vote counting under AV and STV, and the added expense of two-round elections, have limited the appeal of these inter-mediate options. Instead, electoral reform in recent years has focused on the development of mixed electoral systems, with a combination of single-member districts and seats awarded by proportional representa-tion. Two attributes of such systems are critical in determining the incentives for party fragmentation and the prospects for single-party majority government: 1) the ratio between the number of single-mem-ber seats and PR seats; and 2) whether PR seats are awarded as a completely separate tier from the SMP tier or in order to compensate parties that did poorly in vote-seat conversions in the SMP seats. Where the two types of seats are relatively equal in number and awarded on a

60 The Impact of Federalism

compensatory basis, as in Germany and New Zealand, something close to full proportionality in translation of seats into votes is achieved. Indeed, such systems are generally known as a mixed-member propor-tional (MMP) systems.1

While MMP tends to yield fewer parties than most pure proportional systems (especially when combined with high electoral thresholds as in Germany), the prospects for single-party majority government under MMP are nonetheless minimal. Where the percentage of PR seats is small and they are awarded parallel to SMP seats rather than on a com-pensatory basis, the number of political parties is likely to be smaller and the prospects for single-party majority government higher. This option is generally referred to as mixed-member majoritarian (MMM).

Patterns of Federalism

Before examining how electoral rules affect party systems and gov-ernance in federations, it is important to get a general lay of the land. The Table on the following pages provides a preliminary list of elec-toral rules for national lower and upper chambers, national executives, and provincial legislatures in a number of federal systems. Several pat-terns and nonpatterns are evident in the Table. Most generally, there is no clear cross-national relationship linking a federal system to a par-ticular set of electoral rules at either the national or subnational levels. Federalism coexists with a variety of electoral systems—with separa-tion of powers or parliamentarism, with varying degrees of proportion-ality, and with concurrent or non-concurrent elections—at both the na-tional and territorial levels.

Not only is federalism consistent with many electoral systems, but the Table shows that it is also consistent with a wide variation in the “effec-tive number of parties” in the lower chambers of national legislatures.2 Previous research suggests that institutional variables (notably district magnitudes, electoral thresholds, and the presence or absence of concur-rent presidential elections decided by plurality), as well as the degree of ethnolinguistic fractionalization in a country, have a very strong impact on the effective number of parties: Both electoral rules that do not dis-criminate against relatively small parties and more complex patterns of social fragmentation expand the number of legislative parties.

We might also expect federalism to have an impact on the effective number of parties. In particular, federalism may stimulate the growth of province- or region-specific parties that compete and win seats feder-ally, thereby fragmenting party representation in the national legisla-ture. This effect is likely to vary, however, depending on whether or not federal and provincial elections take place simultaneously. In general, we would expect concurrent elections to stimulate votes for parties that operate and have some chance of winning at both levels of government.

R. Kent Weaver

61

Concurrent elections may also allow more efficiency in campaign orga-nization and advertising.

Yet there is little evidence that federalism has an independent effect on the degree of fragmentation in a legislature. Indeed, two federations with concurrent federal and state elections—the United States and Bra-zil—currently have among the lowest and highest effective number of legislative parties, respectively. Federations with nonconcurrent fed-eral and subfederal elections also vary dramatically in their effective number of parties.

There is a very strong association, evident in the Table, between federalism and having a two-chamber national legislature, with the ter-ritorial subunits generally playing a distinctive role in the upper chamber. While unicameralism is common at both the national and subnational levels worldwide, all of the countries shown in the Table are bicameral at the national level, as are virtually all federations worldwide—the tiny Federated States of Micronesia is a rare exception. Canada is also a partial exception, since its upper chamber is a vestigial body appointed by the federal prime minister, and it is no accident that the Canadian Senate has persistently been seen as a problematic political institution. The United Kingdom, with its hereditary and appointive House of Lords—also a vestigial body—is another partial exception.

Upper chambers in national legislatures vary widely both in their selection rules and in the degree to which their powers approach or equal those of lower chambers. Despite this diversity, electoral rules in upper chambers almost always reinforce the role of those legislators as representatives of provincial interests by having them selected on a province-wide basis, even where the lower-chamber elections occur in much smaller electoral districts. The most common situation is that territorial subunits serve as the electoral districts for upper-chamber elections, either in single-member districts as in the United States or in multimember districts as in Australia and Switzerland.

In several countries, provinces play a direct role in selecting mem-bers of the national upper chamber. This can take several forms. In the case of the German Bundesrat, the Länder delegations are direct ap-pointees of the Land governments. Provincial legislatures (rather than executives) choose most members of the Indian Rajya Sabha; autono-mous community legislatures choose a minority of members of the Spanish Senate; and the Belgian Senate is a hybrid of directly elected members, those delegated by regional councils, and additional mem-bers appointed by the first two groups. This suggests a further conclusion about upper chambers in federations: They are more likely than lower chambers to have their members selected in several quite distinct ways.

Moreover, the representation of territorial subunits in upper chambers is frequently disproportionate to population, either with complete equality across units, as in the United States and Australia, or with rules that divide

TABLE—LEGISLATIVE CHARACTERISTICS IN SELEC

COUNTRY	NATIONAL LOWER CHAMBER		NATIONAL UPPER CHAMBER		NATION
	ELECTORAL RULES	EFFECTIVE	ELECTORAL RULES	DISPROPORTIONALITY
		NO. OF		IN REPRESENTATION
		LEGISLATIVE		OF TERRITORIAL
		PARTIES		SUBUNITS
Australia	AV	2.38	STV by states; equal	Moderate	Leader
			representation of states		with pl
					majority
Belgium	PR	7.01	Combination of directly	Low	Leader
			elected, indirectly elected		coalitio
			by linguistic community		majority
			councils, and co-opted
Brazil	PR; very highly	8.69	PR by	Extremely High	Preside
	disproportional rep-		representation of states		elected
	resentation of states				majorit
Canada	SMP in highly	1.69	Vestigial body is	Moderate	Leader
	disproportional		appointed by federal		pluralit
	districts		Prime Minister		of seats
					party c
Germany	MMP	3.16	Appointed by Land	Moderate	Leader
		states; equal	governments; weighted		with pl
			representation of Länder		majority

ABBREVIATIONS FOR ELECTORAL RULES:

AV: Alternative vote

MMM: Mixed-member majoritarian

MMP: Mixed-member proportional

SMP: Single-member plurality

STV: Single transferable vote

TABLE—LEGISLATIVE CHARACTERISTICS IN SELECTED

COUNTRY	NATIONAL LOWER CHAMBER		NATIONAL UPPER CHAMBER		NATIO
	ELECTORAL RULES	EFFECTIVE	ELECTORAL RULES	DISPROPORTIONALITY
		NO. OF		IN REPRESENTATION
		LEGISLATIVE		OF TERRITORIAL
		PARTIES		SUBUNITS
India	SMP	1.69	Mostly elected by state	Low	Leader
			legislatures by STV for fixed		with pl
			terms; minority nominated		majorit
			by President; weighted
			representation of states
Spain	PR with many	2.81	208 directly elected; 44	Moderate	Leader
	small districts		appointed by parliaments		with p
			of autonomous		majorit
			communities; weighted
			representation; fixed terms
Switzerland	PR by cantons,	5.26	Equal representation of can-	High	Colleg
	but effectively		tons, half representation of		chosen
	SMP in single-		demi-cantons; cantons set		term b
	seat cantons		their own selection rules		legisla
United	SMP	2.09	Vestigial hereditary and	N/A	Leader
Kingdom			appointive body		with p
					majorit
United	SMP in highly	1.95	Two seat SMP with	Very High	Presid
States	proportional		staggered elections		plurali
	districts				highly
					state-f

Sources: Data on effective number of legislative parties are from Octavio Amorim Neto and Gary W Number of Parties,” American Journal of Political Science 41 (January 1997): 149–74. Data on disp in Alfred Stepan, “Federalism and Democracy: Beyond the U.S. Model,” Journal of Democracy 10 as low disproportionality, 0.16 to 0.40 as moderate, 0.41 to 0.50 as high, and 0.51 or greater as ver

64 The Impact of Federalism

states into a small number of categories that tend to overrepresent the smallest and underrepresent the largest units, as in Germany. The degree of disproportionality in representing territorial units in upper chambers var-ies broadly, ranging from close to perfect proportionality in Belgium, Aus-tria, and India to extreme disproportionality in Brazil and Argentina.

There are probably two distinct reasons for the strong affinity be-tween federalism and disproportional upper chambers in the national legislature. The first is that federal systems frequently coexist with ma-jor ethnic, religious, or linguistic divisions, and minority groups seek protection not just through consociational mechanisms (if they exist) but also through the “limited government” checks and balances of an upper chamber in which minority regions (if not minority groups) are overrepresented. The second reason is that many federations (notably Australia, Canada, and the United States) were formed from the union of preexisting political units. Politicians from those units did not want to see their influence extinguished in the new national institutions, and those from smaller units were particularly concerned that their jurisdic-tions not be dominated by the larger units in the new federation.

Electoral rules for each chamber are almost always uniform across territorial subunits. In other words, representatives from all subunits are elected following the same sets of rules, even though they may not represent equal numbers of voters. There are exceptions, perhaps the best known of which is the use of the single transferable vote in North-ern Ireland for elections to the House of Commons, while single-member plurality is used in other areas. In the United States, many states moved to direct election of senators before the passage of the Seventeenth Amendment to the Constitution: Thus two different systems existed simultaneously during this transition. In Switzerland, most members of the lower chamber are elected according to PR, but plurality elections are used in five small cantons. In addition, the cantons set their own election rules for the Council of States, although all currently use pro-portional representation. Nonetheless, asymmetrical electoral rules for specific territorial subunits in upper chambers are definitely the excep-tion rather than the rule.

The extent to which legislators in upper chambers actually function as representatives of regional interests varies substantially, however, even if they are elected from province-wide districts. Party discipline is fairly high in most of these bodies, meaning that representation of regional in-terests takes place primarily in the formulation of legislation rather than in votes on its adoption, as well as through the development of regionally focused parties. In Australia, for example, the fact that a different electoral system is used for the Senate (single transferable vote rather than the alter-native vote, as in the House of Representatives) has had a clear impact on policy—the governing party or coalition has never had a Senate majority in recent years, and has therefore been forced to negotiate with smaller

R. Kent Weaver

65

parties to win approval for its legislative program. But the combination of party discipline, the prevalence of class and ideological cleavages over regional ones, and the absence of regionalist parties means that Australian senators do not act as representatives of the interests of their states.

Finally, it should be noted that because representation ratios for ter-ritorial subunits tend to be embedded in constitutions, over time small rural subunits (and interests) may become increasingly overrepresented relative to urban interests.

Subnational Legislatures

Several patterns (and nonpatterns) are also evident when one exam-ines legislatures in territorial subunits. First, as noted earlier, there is, at a minimum, a strong “family resemblance” between legislative-cham-ber electoral rules at the national level and those at the subnational level. The most obvious cross-level similarity is in the choice of a par-liamentary system versus the separation of executive and legislative powers: Countries almost always make the same choice at both levels. Furthermore, within this broad category, there is a tendency to make very similar choices in terms of degree of proportionality, openness or closedness of lists in PR systems, and so on. In some cases, most notably Brazil, there is a common electoral law that operates at both the na-tional and state levels. More frequently, however, these similarities appear to result simply from common institutional inheritances, and perhaps from a reluctance on the part of political elites to make elec-toral rules too difficult for ordinary voters to understand.

Yet family resemblances have been strained in recent years by sev-eral factors that have led to increasingly diverse electoral systems within individual countries. One particularly important factor behind this trend is the rise of a supranational parliamentary institution, the European Parliament. As a result of an EU directive to use proportional represen-tation for elections to the European Parliament, countries like the United Kingdom that had formerly used SMP found themselves hosting on a regular basis elections using a different set of rules. Second, moves toward devolution, especially in Western Europe, have created a need for new electoral arrangements—sometimes with different cleavage-man-agement tasks at the subnational level that appear to call for different electoral rules than those which operate at the national level. Third, the growth of mixed systems, along with widespread perceptions among electoral experts that these rules might be an effective “middle road” between SMP and PR, has also fueled increased diversity of rules within countries. Fourth, perceptions in countries like Japan and Italy that existing electoral rules have had negative consequences have spurred a search for new arrangements that are not always pursued at all levels of government. Thus the United Kingdom, for example, currently uses SMP

66 The Impact of Federalism

in elections for the House of Commons, PR for the European Parliament (except in Northern Ireland, where the single transferable vote is used for both), STV for the Northern Ireland Assembly, and MMP for the Scottish Parliament and the Welsh Assembly.

A second pattern evident across subnational legislatures is the strong degree of uniformity in the electoral rules of individual subunits within countries. Again, there are exceptions: Louisiana with its two-round majority elections, some experiments with the alternative vote and the single transferable vote in western Canadian provinces prior to the 1960s, differences across Australian states in whether all choices must be filled out in their alternative-vote ballots, the use of straight proportional representation rather than a mixed-member system in elections for three Land legislatures in Germany, and so on. But uniformity is the norm, even in multinational federations that are asymmetrical in the constitu-tional powers that they assign to specific territorial subunits.

A third very clear pattern is the low degree of synchronization be-tween national elections and state and local elections. In general, we would expect simultaneous national and subnational elections to foster closer cooperation between candidates and political apparatuses at the two levels, as well as less of a disjunction between party systems and the messages conveyed by parties operating at the two levels. The low degree of inter-level electoral synchronization observed here appears to be largely a by-product of the fact that most federations in the ad-vanced industrial countries are parliamentary systems at both levels, with electoral cycles at each level varying (within fixed outer limits) based on calculations of electoral advantage by those holding the reins of government and their capacity to maintain a legislative majority. Parliamentarism and federalism need not lead to unsynchronized elec-tions, however. This is shown by the case of Sweden, where electoral rules require elections by a specific date even if an intermediate, un-planned election has been held since the last scheduled balloting. This system has preserved a high degree of certainty in electoral timing and the complete synchronization of national and subnational elections.

Implications for Governability

Electoral rules may have several implications for the governability of federations. As noted earlier, single-member-plurality electoral rules may severely punish parties with broad national appeal but not enough support to win pluralities in any region, while disproportionately re-warding regionally concentrated parties if they are able to win plurality victories throughout their regions. An even more serious problem is associated with plurality electoral rules in countries with severe re-gional cleavages: Because the system requires winning pluralities in individual districts, even political parties that win a majority of seats in

R. Kent Weaver

67

the legislature overall may be shut out entirely from regions where their popular support is relatively weak.

This effect is particularly troubling when it involves the governing party, for weak representation of a region in the caucus of the governing party may feed a vicious cycle in which a region’s perceived underrep-resentation in government further exacerbates regional alienation. Per-haps the clearest and most persistent case of such regional underrepre-sentation in government is found in Canada. The Canadian electoral system has produced severe underrepresentation of one or more regions in the governing party much of the time, most notably of Quebec within the Conservative Party prior to 1984 and of the Western provinces within the Liberal Party. Overall, countries with proportional or semipropor-tional electoral systems, such as Germany and Spain, are far less likely to experience severe regional underrepresentation in the governing-party caucus than countries with single-member-plurality (the United Kingdom, Canada) or alternative-vote (Australia) systems.

Several factors in addition to electoral rules also seem to affect the prospects for severe underrepresentation in government. One is the size of the subnational jurisdiction: Extremely small jurisdictions, with only one or a few members of the legislature (for example, the city-state of Bremen in Germany), are far less likely than larger jurisdictions to achieve proportionality in electoral outcomes. A second factor, not sur-prisingly, is the degree to which cleavages have a strong regional component. In Australia, despite the alternative-vote electoral system, severe regional underrepresentation is a phenomenon confined largely to very low-population jurisdictions, because cleavages are stronger along class and ideological lines than by region.

Federalism may also lead to the development of territorially based subnational parties, usually founded on linguistic differences or strong regional cultural identities. In Spain, for example, the growth of region-ally oriented parties followed the creation of Spain’s various autonomous communities, although such parties dominate only in the two regions with exceptionally strong identities (Catalonia and the Basque region). Regionalist parties like the South Tyrolean People’s Party in the South Tyrol province of Italy and the Convergencia i Unio electoral coalition in Catalonia frequently participate in, and sometimes dominate, subnational assemblies in parts of Europe where minority languages are dominant.

Of more importance for the question of governability, however, are situations in which such parties not only represent the interests of a particular territorially based interest group but seek to alter national borders, either through secession to form a new state or irredentism to join another state. I will refer to such parties as “antisystem” parties. The boundary line between regionalist and separatist or irredentist parties is not always clear in practice, of course: The Social Democratic and Labour Party in Northern Ireland, for example, supports eventual union with the

68 The Impact of Federalism

Irish Republic but eschews violence and favors consensual approaches, while Sinn Fein, the political wing of the Irish Republican Army, has only recently moved toward a more gradualist position. Italy’s Northern League, which has vacillated between support for Northern separation and support for federalism, is another ambiguous case.

In general, we might expect antisystem parties to be more likely to emerge under PR (especially where electoral thresholds are low), because the barriers to entry faced by new parties are relatively weak. Yet evidence from the industrialized Western countries does not permit any firm con-clusions on this issue. Belgium’s extreme form of proportional representa-tion has facilitated the emergence of the Flemish nationalist Vlaams Blok as a significant presence in both the federal and Flanders legislatures, but it is not a part of the government in either. In Scotland, the SMP-based elections to the British Parliament have consistently discriminated against the Scottish Nationalist Party. The Nationalists did indeed fare much bet-ter in the 1999 elections to the Scottish Parliament, which were held under MMP rules. Thus SMP electoral rules do appear to have dampened both votes and vote-to-seat conversions for the Scottish Nationalists.

Evidence regarding the impact of Canada’s SMP electoral rules on the growth of antisystem parties suggests that such rules can cut both ways. In the 1960s and early 1970s, Canada’s system clearly discrimi-nated against political parties promoting sovereignty for Quebec. In 1976, however, those same rules gave the Parti Québécois control of Quebec’s National Assembly thanks to a plurality of the overall vote and an electoral platform that emphasized “good government” over at-taining sovereignty. Since that time, the Parti Québécois has won three more National Assembly majorities, but never a majority of the popular vote. Indeed, SMP electoral rules particularly benefit the Parti Québécois, since under this system, the votes of immense majorities for the Liberals in Montreal-area constituencies, with many anglophones and persons whose first language is neither French nor English, are “wasted.” In the November 1998 provincial election, the Parti Québécois won 76 of 125 seats in the Quebec National Assembly, even though the Quebec Liberal Party won 1 percent more of the popular vote. In federal politics, the pro-sovereignty Bloc Québécois won the second largest number of seats in the House of Commons in the 1993 federal election, using its concen-tration of plurality victories (it ran candidates only in Quebec) to win 18 percent of all Commons seats—54 of Quebec’s 75—with only 13.5 per-cent of the national vote. In short, SMP electoral rules may discourage secessionist parties in their early stages, but they may provide such parties a major boost once they attain sufficient size to start winning a substantial number of victories in individual constituencies.

The Italian situation of electoral coalitions within a mixed-member-majoritarian system may in fact be the most likely to promote anti-system parties. Relatively small parties that do not fit clearly on one

R. Kent Weaver

69

side of the left-right divide but have a concentrated regional base are able to win single-member districts in the plurality part of the election contest. Because there are strong incentives to build a broad coalition that can win a majority of seats in both chambers of the Italian Parlia-ment, these small regional parties may be in a strong position either to provide the crucial majority-forming partner for one of the two fairly evenly divided coalitions, or to win good deals in preelection pacts and thus share power if their coalition wins. This is in fact what happened with the Northern League in both 1994 and 2001.

It is extremely difficult to define and develop measures of political violence that are reliable cross-nationally. Moreover, expectations about the effects of electoral rules on political violence are far from obvious. One plausible hypothesis holds that political minorities and groups outside the political mainstream are driven to political violence be-cause under SMP electoral rules they see no prospect for achieving their objectives through normal political channels.

Evidence for this proposition is quite limited, however. Arend Lijphart, for example, finds in a study of 36 democracies that, once levels of economic development, social pluralism, and population size are controlled for, relationships between “consensus democracy” insti-tutions and the level of political violence are quite weak, with much of the variance accounted for by two outliers: the United Kingdom (be-cause of Northern Ireland) and Jamaica (because of violence surround-ing the 1980 election).3 Majoritarian electoral rules for the Northern Ireland Parliament in Stormont, along with widespread gerrymandering practices designed to marginalize Catholic influence, have received at least part of the blame for the violence in Northern Ireland. Yet many other factors, such as widespread employment discrimination, seem at least as important as electoral rules. And the argument is even less per-suasive in other countries. The movement to gain additional autonomy for Quebec has been almost entirely devoid of violence, except for a brief outburst in 1968–70. Political violence has been far more evident in Belgium (with PR elections) than in Canada (with SMP elections). Italy, too, underwent a bout of political violence in the 1960s in the German-speaking South Tyrol region. In short, factors other than majoritarian electoral rules are probably far more important in explain-ing why a minority group turns to political violence. And changes in areas other than electoral reforms (for example, increased autonomy for provincial governments in jurisdictions where minority groups predomi-nate) are probably more important in containing it.

Lessons for New Democracies

This review of electoral systems and their effects on governability in European and North American federations suggests that federalism is

70 The Impact of Federalism

compatible with a variety of electoral rules at both the national and territorial levels. Countries that choose more proportional systems for their national legislatures generally make similar choices for territorial legislatures as well. That said, however, many countries in Europe are developing increasingly diverse sets of electoral rules across different levels of governments.

Single-member-plurality electoral rules do have some distinctive risks in terms of governing federations. In particular, when compared to PR systems, SMP electoral rules appear significantly more likely to create a high risk of territorial exclusion from the governing-party caucus at the national level in parliamentary systems. The effect of SMP electoral rules in fostering antisystem secessionist or irredentist parties is more ambiguous, however. Such rules appear to raise the barriers to anti-system parties that enjoy modest levels of support, but to favor them once their voter support has reached around 35 to 40 percent. Yet both of these effects are highly contingent on other factors—notably on the existence of strong and persistent regionally based cleavages. Finally, the effects of electoral arrangements on stimulating political violence appear to be weak at best.

Overall, the experience of the industrialized democracies suggests no single “best design” of electoral rules for governing federations that should be adopted by new democracies. Past institutional choices in-fluence the range of options that elites are likely to see as viable, as does the need to make election administration simple and less suscep-tible to fraud, and to make the system intelligible to voters. Perhaps most important, the precise conflict-management tasks faced by new democracies vary widely across countries. Elites in new democracies thus will—and should—make very different choices in designing elec-toral systems. What is essential is that their choices be well-adapted to national circumstances and able to be perceived as fair over time, rather than based upon short-term calculations of electoral self-interest.

NOTES

1. See Matthew Soberg Shugart and Martin Wattenberg, eds., Mixed-Member Electoral Systems: The Best of Both Worlds? (New York: Oxford University Press, 2001).

2. On the definition and calculation of the effective number of parties, see Rein Taagepera and Matthew Soberg Shugart, Seats and Votes: The Effects and Deter-minants of Electoral Systems (New Haven: Yale University Press, 1989), 78–80.

3. Arend Lijphart, Patterns of Democracy: Government Forms and Perfor-mance in Thirty-Six Countries (New Haven: Yale University Press, 1999), 271.