WikiSummary, the Social Science Summary Database

Powell: Elections as instruments of democracy

From WikiSummary, the Free Social Science Summary Database


Powell. 2000. Elections as instruments of democracy. New Haven: Yale UP.

This summary needs cleanup to raise it to WikiSummary's standards. Can you help us improve it? Please volunteer.


Elections are key way for citizens in a democracy to communicate with representatives. Different rules about elections and concentration of assembly power can create either majoritarian or proportional systems of democracy. Theoretically, we expect a tradeoff between majoritarian and proportional aspects of democracy. Empirically, both types of systems seem to do a relatively good job of achieving the goals for which they are designed. However, proportional systems seem to do a better job both representing diverse interests (based on committee power) and aligning government and policy makers with the median voter's ideal point.

What is responsiveness?

PR and majoritarian systems have a different view of the relation between voters and politicians. Majoritarianism encourages legislative mandates and the implementation of majority/plurality will (goal of accountability). Proportionalism is an attempt for policy to be responsive to as large a coalition as possible (goal of representation).

Role of elections

Majoritarian vision

  • Accountability: limited demands on electorate; possible rejection affects incumbents
  • Mandates for majority government to enact policy

Proportional vision

  • Voters choose agents to bargain for them
  • Proportional representation between interests of society and elected political groups

Two dimensions of voting are relevant.

  1. voting for a representative agent or for collective government
  2. retrospective versus prospective voting

Election threshold and consequences (see Cox 1997)

  • Number of parties: higher the threshold the fewer the number of parties.
  • Seat-vote disproportionality: higher threshold leads to greater disproportionality
  • M+1 rule: number of effective parties = number of members/district + 1 (Cox 1997)

Concentration vs. dispersion of assembly power

Examine role of committees in legislative process (see Strom 1990)

  • Countries fall into three groups: rules facilitate opposition influence; rules encourage some dispersal of influence; rules support centralized legislative control

Dispersion of power outside of assembly: independent executives; separate legislative chamber; federal system; judicial review

  • Majoritarian systems tend to score higher on concentration of power.

Closeness to median citizen

  • Survey used to locate median citizen; expert opinion places political parties on left-right dimension. We might think that majoritarian systems would do better on this issue because of tendency of two parties to converge to median voter (see Downs and Hotelling)
  • But counterintuitively, PR systems have greater congruence between government/party policies and the public than majoritarian systems.
  • Where there is little post-election bargaining, the government tends to be further from the median than when there is open, unconstrained, post-election bargaining.

Additional Summary from a contributor: (should be merged with above)

SUMMARY: Powell examines over 150 elections in 20 democracies over 25 years. While both majoritarian democracy and proportional representation have their merits, proportional representation better reflects the populace's needs and better represents the voters' wishes. He concludes PR is more democratic. But also notes that concentration-dispersion involves tradeoffs that make a perfect set of electoral arrangements impossible.

Elections as instruments of democracy

  • Elections are not only instrument of democracy but they are a defining feature
    • They build connections between wishes of citizens and behavior of policymakers (p. 14)
  • Majoritarianism uses elections to bring power of the people to bear on policymakers.
    • Rules of representation encourage election of legislative majorities that can control the exec. (p. 21)
    • Rules for making authoritative public policy concentrate political power in hands of this "govt"
  • Proportionalism an alternative positive, democratic ideal to overcome "limiting majorities"
    • Election rules encourage equitable representation of multiple parties
    • Decision rules encourage dispersion of power

Concentrated power

  • Elections are seen as tools to control policymakers
  • Elected officials can make policy
  • Citizens know who is responsible; can reward/punish
  • Concentrated power necessary for majoritarianism provides a clear winning policy coalition

Dispersed power

  • Elections are more indirectly related to policy making
  • They bring many different agents together to bargain; a shifting coalition
  • Why disperse power?
    • Elections are "clumsy" elections may not reflect actual majority preferences (!)
    • Preferences of all citizens, not just winning position, taken into account

Assuming full participation, how do elections influence government? From the voter's perspective:

  • Target of choice voter is either considering a representative agent who will negotiate best for them (delegate), or voting for a government with power to make decisions (trustee)
  • Temporal dimension choosing retrospective or prospective voting

Different roles of elections under majoritarian control/concentrated power:

Accountability evaluating incumbent government. Conditional requirement is clarity of responsibility.

  • Citizens have opportunity to change policymakers.
  • Some (Riker) say the electorate can't really choose positive policy direction effectively.
  • Anticipation of possible rejection shapes policies.
  • Requires little knowledge on part of the electorate.

Electoral mandates not strictly focused on incumbents. What does opposition have to offer?

  • Voters must be able to identify possible future government.
  • Winning party must be able to dominate policymaking after election

Conditions for citizen influence/dispersed power:

  • Authorized representation in policymaking
  • Retrospective-prospective distinction less important here. Citizen represented by multi-stage process:
  • At election stage voters must be able to identify a candidate that has their confidence
    • Number of representatives must represent the number of voters who chose them
  • At post-election stage, assumes flexibility in forming coalitions
    • Distance between voter and his/her authorized representative greatest at this time

Test by looking at responsiveness in selecting governments and policymakers:

  • Citizen preferences and party positions voting indicates citizen preferences but only within a limited range of choices. So he also uses a Left-Right scale.
  • Representational congruence optimum relationship between voters and govt (meaning elections are good instruments of democracy)


Majoritarian versus proportional designs determined by looking at election rules and policy-making rules:

Election rules

  • Duverger's Law: single-member district plurality tends to lead to two major parties
  • Strategic voting: citizens may vote for second choice if they know their first choice won't win
    • Anticipating this, parties may not bother running candidates with poor prospects of winning
  • In single member districts, we expect only two major candidates.
    • The greater the district magnitude, the more candidates.
    • The greater the magnitude, the less the consequences of a strategic mistake
  • Effective threshold (Lijphart): how many votes must a party win to get national representation?
  • Election outcomes:
    • Majoritarian vision favors high threshold, making representation difficult for small parties
    • Proportional vision uses low threshold
    • Data for his 20 countries support these expectations in terms of effective number of parties and vote-seat disproportionality

Read With: Lijphart; Richie; Anderson & Guillory

Critiques: He perhaps does not credibly account for the outlook of the losers (Anderson &Guillory 1997 for example) and does not prove his assumption that one outcome is the best outcome