What is an example of a country that exist the proportional representation?

1.1 Tactical Voting Beyond Single-member Plurality

Reacting to Duverger's apparent belief that tactical voting was unknown under PR systems, Leys (1959) and Sartori (1968) argued that tactical voting under PR was no different in kind from that found under plurality, differing only in the degree to which it came into play—and, hence, in the degree to which it tended to reduce the number of viable parties. Sartori's notion of a continuum of systems, from strong (in which tactical voting acts forcefully to depress the number of parties) to feeble (in which tactical voting is largely absent and thus puts little downward pressure on the number of parties), is now standard in the literature.

Subsequent work has made the Leys–Sartori viewpoint more precise, both theoretically and empirically. Steven Reed, for example, has extended Duverger's Law to Japan's former electoral system (used from 1925 to 1993, with a brief postwar gap), which gave voters a single vote but used multimember districts. To illustrate Reed's extension, consider a district with three seats. Candidates expected to place fifth or lower in such districts were sometimes seen as hopeless and were thus deserted by their supporters, essentially reducing the competition to a four-candidate fight for three seats. Reed discovered similar tendencies for elections in four-seat districts to reduce to five-candidate contests, and for elections in five-seat districts to reduce to six-candidate contests. More generally, under a wide range of electoral systems, the rule is that tactical voting tends to reduce competition in an M-seat district to, at most, M+1 candidates or lists (Cox 1997).

Although political scientists' main reason for interest in tactical voting is its tendency to concentrate votes locally, hence to restrict party systems nationally, further research into four questions has clarified some limits of this line of reasoning. These questions are considered in the next four subsections.

Electoral System

The electoral system is one of the main sources of institutional diversity among democratic countries. The most common distinction is between proportional representation systems and plurality–majoritarian regimes. Under proportional representation, a party's vote share determines their share of electoral seats – an arrangement that tends to produce multiparty coalition governments. By contrast, in plurality systems the party with the most votes typically controls policy alone. It is not surprising that the electoral system affects incentives for choosing certain types of exchange-rate policies, even though the findings have been somewhat contradictory. Some research finds that democratic politics with majoritarian electoral systems are more likely to choose fixed exchange rates and experience less exchange-rate volatility after cabinet dissolutions than those with proportional systems. In contrast, other authors find that industrial countries with majoritarian systems experience higher exchange-rate volatility than countries with proportional electoral systems, whereas another study concludes that electoral systems do not affect currency markets in emerging markets.

The range of estimates of whether and how the electoral system affects exchange rates once more suggests the need to consider how this institutional feature interacts with other aspects of the political system. The existing literature suggests that the level of opposition influence is one such conditioning factor. For instance, one article observed that countries with both majoritarian systems and low opposition influence are least likely to fix, whereas proportional representation (PR) systems in which the opposition exerts a lot of influence are most inclined to adopt some type of fixed exchange rate. The electoral system thus influences some countries’ exchange-rate policy, but the direction and strength of influence depend on other political-economic factors.

3.1 Electoral Competition: Theory

To isolate electoral effects, let us consider a “pure” model of electoral competition as one in which two parties vie for control of government by publicly committing to a policy before voters place their votes, and after the election, the winning party cannot renege on its policy commitment. Further, let us assume that each party cares only about winning, and has no interest in policy other than as an instrument for winning the election.

Naturally, the electoral rules matter for the nature of electoral competition, and perhaps the most important distinction along these lines is between models of proportional representation and majoritarian systems; see Persson and Tabellini (2002, ch. 8), for an introduction. If the country is divided into districts, then a majoritarian system is one in which one representative is elected from each district, and then control of the government goes to the party that wins the largest number of districts.k A proportional system in one in which each district has multiple representatives, who are chosen from each party in proportion to the party vote in that district. A special case is the case in which the whole country is a single district; either way, if there are enough representatives, the nationwide share of each party in total representatives is equal to its nationwide share of votes. Both systems are common; The United States, the United Kingdom, and Canada are examples of countries with majoritarian systems; Germany, Brazil and Turkey are examples with proportional systems.

The seminal model of trade policy determination through electoral competition is Mayer (1984), whose main argument is built on a straight HO model injected into the proportional representation paradigm.l Home is a small open economy that produces two goods, the imported good 1 and the exported numeraire 2, using labor and capital. Both factors are homogeneous and mobile across industries. There are a continuum of citizens in Home, and each citizen i has a unit endowment of labor and an exogenous endowment ki of capital. The government must choose a level for the tariff t on imports of good 1. The production income of citizen i is w + rki, where w is the wage and r is the price of capital services, and so i’s production income relative to the average is ϕi = (w + rki)/(wL + rK), where L and K are the aggregate Home endowments of the two factors. All Home citizens have identical and homothetic preferences represented by the indirect utility function U(p,y) for a good-1 price p and income of y.

Tariff revenue is redistributed to Home citizens lump sum; to clear away nuisances in the algebra, it is assumed that citizen i’s tariff revenue rebate is proportional to her share of production income, so that her total income will be given by yi = w + rki + ϕiT, where T is the aggregate tariff revenue. Consequently, the utility achieved by i can be written as Ui(t) = ϕi(t)U(π(1 + t),Y (t)), where π is the world price of good 1 and Y (t) is aggregate income in Home evaluated at domestic prices, including tariff revenue. The first factor is i’s share in total welfare, and the second is total welfare. Total welfare is maximized at t = 0 because Home is a small open economy, so if dϕi(t)/dt > 0 at t = 0, then citizen i would prefer a positive tariff to free trade, and if dϕi(t)/dt < 0 at t = 0 she would prefer a negative tariff, an import subsidy, to free trade.

The sign of dϕi(t)/dt is determined by i’s position in the asset distribution as well as by Home's factor endowment. A change in tariffs affects the factor prices by Stolper–Samuelson logic, since it changes the domestic price of good 1, and this is how it affects ϕi. In particular:

(3)dϕidt=wL(wL+rK)2(1+t)r(k−ki)(ŵ−rˆ)pˆ,

where k ≡ K/L and a hat indicates a proportional rate of change (so for example ŵ=1wdw/dt). Thus, dϕi(t)/dt > 0 if ŵ>rˆand k > ki, or if both inequalities are reversed; otherwise, dϕi(t)/dt < 0. Since the tariff raises the domestic price of good 1, by Stolper–Samuelson, ŵ>rˆiff good 1 is labor-intensive, and since good 1 is imported, that would mean that Home is capital abundant. Thus, citizen i's share of national income is increasing in the tariff, and so i will want a positive tariff, if the country is capital-abundant and i has below-average capital, or if the country is capital scarce and i has above-average capital. Otherwise, citizen i’s share is decreasing in the tariff and she will prefer a negative tariff.

Now, consider the “pure electoral” game sketched above. Party A and B must each choose a value of t, simultaneously; these choices then become public, as an irrevocable policy commitment; the voters see these choices, and each votes for the party that offers the tariff value that voter prefers; the party with the most votes wins, and implements the tariff to which it had committed. The winning party receives a positive rent from power, while the losing party receives a payoff of zero. It is easy to see that, provided the utility relation Ui(t) is a single-peaked function of t, there is a unique equilibrium in pure strategies: Both parties commit to the most preferred tariff level tm of the median voter, the voter whose capital endowment is greater than that of exactly one half of the population.m Couple this observation with the assumption that in each country the median voter has less capital than the average voter (a safe assumption, borne out in every data set on income and asset ownership), and the model yields two crisp predictions. First, in high-income countries, tariffs will be positive because the median voter is a worker who wants to raise the price of labor-intensive imports, while in low-income countries, tariffs will be negative because the median voter is a worker who wants to lower the price of capital-intensive imports. Second, this will occur only in countries with a skewed income distribution; if the median voter has the average capital endowment, the outcome will be free trade. As a result, in high-income countries, a rise in the gap between mean and median income will increase the tariff, but in low-income economies it will have the opposite effect. These predictions are testable, as we will discuss in the next section.

It is worth pausing to reflect the fragility of these conclusions, however. First, the political-economy game really does require the 2 × 2 structure of the economy. With more than two goods, there is no meaningful sense in which there is a median voter, and generically no equilibrium (Plott, 1967). Indeed, even in the two-good model the equilibrium makes sense only if the election is held as a referendum on trade policy, with no other issues relevant.

Second, the equilibrium is not robust to adding tiny amounts of noise to voters’ preferences. To see this, consider a variant based on Lindbeck and Weibull (1993). This is a model of pure electoral competition in the meaning suggested above, but has the added feature that each voter votes for the party that offers her the highest utility, where utility has two components: (i) a real-income component that is derived from general equilibrium given the policy and the voter's utility function, as above, and (ii) a second component that is an exogenous preference for one party or the other. It could be based on ethnic affinity, the charisma of the party leader, or some element of the party's history. Regardless of the source, suppose that each voter has an exogenous preference for party A given by an idiosyncratic value μ ∈ ℜ. For voters of each economic type, Lindbeck and Weibull (1993) allow for μ to be drawn from a mean-zero distribution. In the Mayer model, the only way voters differ is by capital ownership, so suppose that for each group of voters who share the same value k~of capital endowment, the values of μ are distributed by a probability density function f(μ;k~)with mean zero. If we restrict attention to densities that imply a very small variance to the μ values, then the model is the same as the Mayer model but with a tiny amount of noise added to voters’ preferences. In other words, one can think of letting f(0,k~)become arbitrarily large for each value of k~, so that voters barely care at all about anything other than policy.

Lindbeck and Weibull (1993) show that equilibrium of the electoral competition in this case results in both parties choosing the same policy, and the policy maximizes the weighted sum of utilities for Home voters, where the weight on voters with endowment k~is given by f(0;k~).n Now, if the density does not depend on k~(so that f(μ;k~)=f(μ;K/L)∀k~), this clearly indicates social welfare maximization, which is a stark departure from the median-voter model. Further, by varying the densities to make f(0;k~)greater for some values of k~than for others, the most-preferred tariff corresponding to any k~value can be arbitrarily closely approached in equilibrium. The fact that this can be done for an arbitrarily small amount of noise means that the median voter equilibrium fails a basic stability test.

The solution is most plausibly to build idiosyncratic partisan preference shocks into the model from the beginning, as is standard in the rest of the literature (for example, Persson and Tabellini, 2002, ch. 8 and Strömberg, 2008), rather than trying to preserve the median-voter result for its own sake.

An important extension of the pure electoral model is explored by Ponzetto (2011), who focusses on the effect of voters’ incomplete information on trade policy. A small open economy has several industries, each using a specific factor and labor; all consumers have identical additively separable utility functions so that partial-equilibrium analysis can be applied (the model is the same as is used by Grossman and Helpman, 1994; see Section 3.3.2). Two parties compete for votes by choice of trade policy vector. Each voter who owns a specific factor wants trade protection for her own industry, but prefers free trade in all other industries. Voters also have partisan preferences in the Lindbeck and Weibull (1993) mold. The trick is that it is difficult for each voter to learn what each policy proposal is; each proposed tariff by each party is learned with a probability less than one by any given voter, with a higher probability for the voter's own industry (this is given a microfoundation based on information networks). Since a voter cannot be swayed by a tariff proposal of which she is unaware, and since each voter is more likely to be aware of tariff proposals for her own industry, each party knows that by offering a tariff to industry i it will win more votes from voters in i than it will lose from other industries. Consequently, all tariffs are positive in equilibrium. Further, if information about an industry becomes more evenly distributed (so that the difference between the rate at which people in that industry and people outside that industry become informed about it declines), then the equilibrium tariff in that industry falls. This effect, dubbed the “Dracula” effect (meaning the influence of sunlight on protectionism), is interpreted as reflecting the influence of the media on trade policy; as noted in Section 2.4, the evidence suggests that this interpretation holds water in the real world.

We now turn to majoritarian systems, and to discuss these we need a model with at least three districts. The simplest example can be constructed by dividing up the voters in the Mayer model into multiple districts, a majority of which must be won by a party in order to control the government. Karabay and McLaren (2004) point out that in such a model, the equilibrium tariff rate is the most preferred rate of the median voter in the median district (when the districts are ranked by the most-preferred tariff of their median voters). As a result, depending on how voters are allocated to districts, the equilibrium tariff can be anywhere from the national 25th percentile voter's to the 75th percentile voter's most preferred tariff. This, of course, suffers from the problems of the original Mayer model.

Grossman and Helpman (2005) present a three-district majoritarian model with legislative bargaining after the election, which becomes a pure electoral competition model as a special case.o There are three industries that produce with specific capital together with labor, plus a numeraire sector that produces only with labor. Each district has some of each industry's capital, and the districts are symmetric in that each district has the largest share of one industry's capital and smaller shares of the other two; the pattern of shares is the same for all three districts, with the identities of the three industries permuted. The outcome of the political game will be a vector of tariffs for these three industries. Voters have idiosyncratic partisan preference shocks, and each district has an aggregate partisan preference shock revealed after the campaigning is over, but the ex ante distributions of these shocks is the same for each district. In that setting, the equilibrium of the pure electoral competition is free trade, which is a stark contrast with the same model when postelection legislative bargaining is allowed, as we shall see.

Clearly, in the pure electoral-competition version of Grossman and Helpman (2005), the majoritarian electoral system has no effect on policy. This is because of the symmetry of the setup. However, a large part of the interest in majoritarian systems stems from effects that arise when the districts are not ex ante identical in their partisan tendencies. The basic model of fiscal and redistributive policy in a majoritarian democracy in Persson and Tabellini (2002, ch. 8) has two districts with partisan biases, each biassed toward one of the two parties, plus a third that is biassed toward neither, which we might call a “swing” district. Since the swing district is more likely to be pivotal (in other words, to be the district that determines the outcome of the election), both parties commit to redistribution toward that district. In addition, since neither party derives much electoral benefit from public good provision for the partisan districts, only the benefit from public goods to the swing district matters, and consequently less public good is provided than under proportional representation (and less than optimal). Because districts differ in their partisan bias ex ante, majoritarian elections make a big difference to the policy in equilibrium. A rich model of the US electoral contest in Strömberg (2008) allows for probabilistic voting and uncertain outcomes but states with different levels of ex ante partisan bias. The analysis shows that a state with a higher probability of being pivotal receives more campaign resources, a prediction confirmed in the empirical portion of the paper.

These differences between partisan states and swing states in a majoritarian system can of course matter for trade policy as well. Muûls and Petropoulou (2013) present a stylized model of the electoral college with four types of voter: those who always vote for Democrats (type D); those who always vote for Republicans (type R); those who vote for the candidate most likely to provide tariff protection (type P); and those who vote for the candidate most likely to provide free trade (type F). The allocation of these voters can differ across states. In particular, ωps∈[0,1]is the difference between D voters’ share of state s and R voters’ share. If |ωps|is close to 1, then trade policy will not have much effect on the outcome of state s’s electoral vote, because the state has a strong partisan leaning in one direction or the other, but if |ωps|is close to 0, trade policy can possibly swing the state's election. (Although each voter's choice is deterministic, the aggregate outcome is still random because the number of voters that turn out to vote can vary randomly.) The model has an unusual feature in that it assumes away any commitment by politicians. A newly-elected president can choose free trade or protectionism in her first term, which can then signal her true trade-policy preferences to voters who must choose whether or not to support her for a second and constitutionally final term. In an incumbent's second term she will simply implement her most preferred trade policy.

If there are enough P-type voters crammed into enough low-|ωps|states, then in equilibrium a free-trade-preferring president will, with positive probability, choose trade protection in her first term in order to dupe those protectionist voters into supporting reelection. Note that this can be an outcome even if there are more F-type, free-trade-loving voters than the P-type, provided that the F-type voters are concentrated in states that are sufficiently partisan.

Ma and McLaren (2015) study a pure-electoral-competition version that works out more simply. Two parties compete for electoral votes by committing to policy positions; each voter has an idiosyncratic partisan bias, and each state has an aggregate partisan bias, except for one state, the “swing state,” whose bias is zero. In the baseline model with no aggregate uncertainty, in any equilibrium in pure strategies, both parties commit to the policy vector that maximizes swing-state welfare, placing a weight of zero on nonswing-state welfare. Adding a probabalistic element to the voting so that the bias of the swing state is not exactly zero and is not known precisely before the election creates a swing-state bias that converges to the extreme of the baseline model as the aggregate uncertainty becomes small, so that the effective weight on nonswing-state welfare converges to zero in the limit. Fitting the model to data on tariffs suggest that US politicians put a weight on nonswing-state welfare about 0.7 as large as on swing-state welfare.

3 Stability or Change?

Although interest groups are an enduring and persistent part of democratic politics, they frequently face significant changes in their political environment. In parliamentary systems with proportional representation, for example, new political movements can quickly metamorphize into new political parties. The lobbying organizations for that cause may have to switch quickly from protest politics to campaign politics. In the American context, with its stable two-party system, the challenge more recently has been the unprecedented growth of interest groups. Looking back on his career, one member of the House remarked ‘When I first came to Congress there were five major financial trade groups, but now there are at least five times that. Now if you're trying to satisfy all the trade groups, it's pretty hard to do.’

For present-day lobbyists the most significant tactical decision may be figuring out what to do just to get noticed. With so many lobbyists working on the same issue, how do the lobbyists from a single organization convince legislators and their staffers that their organization is one to consult with and to work up drafts of legislation with? The competition between lobbyists has put an even greater premium on expertise. Although well-connected lobbyists are always helpful, policy expertise is a particularly important trait of effective lobbyists.

An interesting question that has emerged is whether the new technologies are changing interest group politics. The new technologies have certainly made it easier for interest group organizations to reach their members and attentive publics. Today communication can be virtually instantaneous. If events change rapidly lobbyists in the capital can mobilize their members quickly to let policy makers know that voters back home have noticed what's going on and they are upset about it. Interest groups have also taken advantage of the Internet, establishing their own web sites to serve as vehicles for soliciting members, advertising their views, and facilitating the retrieval of documents by sympathesizers around the country. Another use of the Web is that it facilitates the organization of new groups. Small constituencies with modest resources can use the economies of the Internet to reach those concerned with their issue and to generate communication among themselves.

Yet for all the changes that have taken place in government and in telecommunications, the fundamental nature of lobbying remains relatively stable. At its core lobbyists work assiduously to mobilize their followers and to bring information to the attention of policy makers. Although how those tasks are carried out changes over time, these changes have not revolutionized the process.

As political science has become increasingly sophisticated and the discipline has placed greater emphasis on precise measurement of complex phenomena, lobbying has remained a particularly difficult nut to crack. Assessing the effectiveness of different types of lobbying or different lobbying campaigns continues to be a formidable methodological problem. We know a lot more about many facets of interest group politics, especially in understanding how and why people join organizations. Research on lobbying, however, has made only limited progress.

10 Legislative Institutions

A number of debates have centered on the question of what kinds of legislative institution are best for a democratic society. What choice we make here will depend heavily on the kind of underlying ethical justification of democracy, our conception of citizenship, and our empirical understanding of political institutions and how they function. The most basic types of formal political representation available are single-member district representation, proportional representation, and group representation. In addition, many societies have opted for multicameral legislative institutions. In some cases, combinations of the above forms have been tried.

Single-member district representation returns single representatives of geographically defined areas containing roughly equal populations to the legislature and is present most prominently in the USA and the UK. The most common form of proportional representation is party-list proportional representation. In a simple form of such a scheme, a number of parties compete for election to a legislature that is not divided into geographical districts. Parties acquire seats in the legislature as a proportion of the total number of votes they receive in the population as a whole. Group representation occurs when the society is divided into nongeographically defined groups such as ethnic or linguistic groups or even functional groups such as workers, farmers, and capitalists and returns representatives to a legislature from each of them.

Many have argued in favor of single member district legislation on the grounds that it has appeared to them to lead to more stable government than other forms of representation. The thought is that proportional representation tends to fragment the citizenry into opposing homogeneous camps that adhere rigidly to their party lines and that are continually vying for control over the government. Since there are many parties and they are unwilling to compromise with each other, governments formed from coalitions of parties tend to fall apart rather quickly. The post-War experience of governments in Italy appears to confirm this hypothesis. Single-member district representation, in contrast, is said to enhance the stability of governments by virtue of its favoring a two-party system of government. Each election cycle then determines which party is to stay in power for some length of time.

Charles Beitz (1989) argues that single-member district representation encourages moderation in party programs offered for citizens to consider. This results from the tendency of this kind of representation towards two-party systems. In a two-party system with majority rule, it is argued, each party must appeal to the median voter in the political spectrum. Hence they must moderate their programs to appeal to the median voter. Furthermore, they encourage compromise among groups since they must try to appeal to a lot of other groups in order to become part of one of the two leading parties. These tendencies encourage moderation and compromise in citizens to the extent that political parties, and interest groups, hold these qualities up as necessary to functioning well in a democracy.

In criticism, advocates of proportional and group representation have argued that single-member district representation tends to muffle the voices and ignore the interests of minority groups in the society. Minority interests and views tend to be articulated in background negotiations and in ways that muffle their distinctiveness. Furthermore, representatives of minority interests and views often have a difficult time becoming elected at all in single-member district systems, so it has been charged that minority views and interests are often systematically underrepresented. Sometimes these problems are dealt with by redrawing the boundaries of districts in a way that ensures greater minority representation. The efforts are invariably controversial since there is considerable disagreement about the criteria for apportionment. In proportional representation, by contrast, representatives of different groups are seated in the legislature in proportion to citizens' choices. Minorities need not make their demands conform to the basic dichotomy of views and interests that characterize single-member district systems so their views are more articulated and distinctive as well as better represented.

Another criticism of single-member district representation is that it encourages parties to pursue dubious electoral campaign strategies. The need to appeal to a large, diverse, and somewhat amorphous sector of the population can very often be best met by using ambiguous, vague, and often irrelevant appeals to the citizens. Thus, instead of encouraging reasonable compromise, the scheme tends to support tendencies towards ignorance, superficiality, and fatuousness in political campaigns and in the citizenry. It encourages political leaders to take care of the real issues of politics in back rooms while they appeal to citizens by means of smoke and mirrors. Of course, those who agree in the main with the elitist-type theories will see nothing wrong in this; indeed, they may well champion this effect. Proportional representation requires that parties be relatively clear and up-front about their proposals, so those who believe that democracy is ethically grounded in the appeal to equality tend to favor proportional representation.

Advocates of group representation such as Iris Marion Young (1993) have argued that some historically disenfranchised groups may still not do very well under proportional representation. They may not be able to organize and articulate their views as easily as other groups. Also, minority groups can still be systematically defeated in the legislature and their interests may be consistently set back even if they do have some representation. For these groups, some have argued that the only way to protect their interests is legally to ensure that they have adequate and even disproportionate representation.

One worry about group representation is that it tends to freeze some aspects of the agenda that might be better left to the choice of citizens. For instance, consider a population that is divided into linguistic groups for a long time, and suppose that only some citizens continue to think of linguistic conflict as important. In the circumstances, such schemes may tend to be biased in an arbitrary way that favors the views or interests of those who do think of linguistic conflict as important.

What is an example of proportional representation?

Is there proportional representation in the US?

Does UK use proportional representation?

Is proportional representation used in Australia?