As discussed by Arend Lijphart in his analysis of electoral systems in 27 democracies (Lijphart 1994), each of these types of systems can be compared in respect to three basic properties or variables: (1) the specific electoral formula for converting votes to seats, (2) district magnitude or the number of seats to be elected from the an electoral district or constituency, (3) the effective threshold or minimum percentage of the vote that a candidate or party must receive to obtain election.
Taken together, these variables determine how an electoral system converts the total vote each party receives in an election into its number of seats in the legislature. Systems where the proportions of seats awarded to all parties closely reflect their portions of the vote are highly proportional. Conversely, the greater the gap between the proportion of the total vote received by each party and its proportion of seats, the more disproportional the system.
The three variables normally determine the extent of proportionality in the following ways.
Electoral systems can be compared in respect to the extent of their proportionality or disproportionality by calculating one or more indexes that measure the gaps between the proportion of votes and the proportion of seats received by all parties. The most commonly used index is the index of disproportionality developed by Michael Ghallagher (1991). Ghallagher’s index is computed in the following manner:
By this measure, the smaller the ID, the more proportional (i.e. less disproportional) the electoral system. Conversely, the larger the larger the ID the less proportional the system.
Similarly, the more socially homogeneous a given geographic area, the more likely it will support one party, and the lower the index of disproportionality for the entire system. For this reason, SMD systems which elect a large number of representatives from electoral districts that are relatively small in size have much smaller levels of disproportionality than SMD systems that elect a lower fewer representatives from large districts. For an illustration of this relationship, viewers should compare the ID for the present SMD electoral system in Kenya of 210 constituencies with the ID for the simulated SMD system of 350 constituencies.
Because PR systems usually yield results that are significantly more proportional than plurality-majority SMD or MMD systems, PR systems are more inclusive in so far as small parties gain representation in the legislature. However, to the extent that proportionality is enhanced or reduced by threshold, district magnitude, and the geographical distribution of the vote, factors other than PR determine the extent of proportionality and inclusiveness.
Proportionality and inclusiveness are also purchased at the price of accountability to the electorate. Because there are no electoral districts with PR or a relatively small number of large second-tier districts where PR is applied on a regional basis, those elected are not elected to represent specific geographic areas and members of the electorate cannot hold individual legislators accountable to their communities. The use of party lists to determine the rank order of a party’s candidates in PR systems also means that aspiring candidates and legislators are accountable first to their party’s leadership, which determines their position on the list, rather than to the electorate.
To the extent that proportionality and inclusiveness result in many parties being represented in the legislature but where none command a majority, there is a greater likelihood of coalition governments and thus power-sharing among parties than where the number of parties are few. PR systems, however, do not necessarily lead to coalition governments (i.e. South Africa), while some SMD systems frequently do (i.e. India).
The Choice of Electoral Systems
There is no single electoral system that is “optimal” or “best” for all political systems, because political systems vary in respect to the major political groupings that seek power and which shape its politics. These groupings obviously vary as a function of a country’s history and social composition (i.e. its level and type of economic development, its culture, its linguistic, ethnic, and religious makeup).
The basic test is what system is deemed most appropriate by both political elites and ordinary citizens in so far as it constitutes a viable set of rules to sustain the electoral process election after election. By this test, the rules must be both inclusive and promote accountability. If the system is not sufficiently inclusive, groups that believe they are excluded from the process may seek to disrupt or overturn the process. Similarly, if the system is not sufficiently accountable, i.e. if the electorate does not believe that the system promotes accountability to them, then voter turnout will drop and the elections run under the system will be perceived as meaningless.
The choice of what form of system is most appropriate for a country is therefore a decision to be negotiated between competing political elites on behalf of their supporters across the electorate. It is also a decision that requires a high level of consensus and compromise among both elites and ordinary citizens. We present the SDSS for electoral system design as an analytical tool to help elites and citizens reach consensus of which is “best” (i.e. most acceptable) for them.
Copyright © 2001 by Joel D. Barkan, Paul Densham and Gerard Rushton