Research brief: Some considerations* on designing an electoral system for electing the Chief Executive of Hong Kong by universal suffrage

 In Chief Executive & Legislative Council Elections, Constitutional Reform, Constitutional Reform, Governance & Institutional Design
,

(*These considerations become irrelevant if the nomination procedure of CE elections limits Hong Kong people’s equal rights to elect and to be elected, and the CE is not allowed to have political party affiliation.)

 

  • We have repeatedly emphasized that a particular electoral system is chosen not for its own sake but for achieving certain desirable political outcomes.
  • Thus, the first-order task of designing or choosing an appropriate electoral system for electing the Chief Executive (CE) of Hong Kong by universal suffrage is to examine what CE electoral systems will produce what political consequences.
  • Since Hong Kong’s political system is essentially a presidential system, research on countries adopting presidentialism will hopefully provide some insights for our reference.

 

Key research findings on presidential electoral systems

  • There is a consensus that presidentialism is not parliamentarism plus an elected chief executive (i.e. presidentialism cannot and should not be thought of as just like parliamentarism except for its manner of selecting a chief executive); rather, presidentialism constitutes a political system in its own right and should be analyzed separately.
  • One dimension of the consensus is that presidential elections and legislative elections in presidential systems are not independent of one another; instead, they interact and produce political consequences different from those of parliamentary systems.
  • While it has been pointed out that legislative elections may affect presidential elections, little research has been done on this.
  • Research has mainly focused on the impacts of presidential elections on legislative elections.
    • Because of the importance of the office of the presidency in a presidential system, it has been found that presidential elections have strong impacts on legislative elections.
    • Specifically, institutions of presidential elections affect
      • legislative fragmentation – presidential coattails on legislative fragmentation
      • partisan support for the president in the legislature

 

Presidential coattails on legislative fragmentation

  • The impact of presidential coattails on legislative fragmentation varies along two institutional arrangements of presidential elections, namely the electoral formula for electing presidents and the electoral cycle.
    • Electoral formula for electing presidents – plurality rule vs majority run-off rule
      • It affects legislative fragmentation through its impact on the number of presidential candidates.
      • The plurality rule reduces the number of presidential candidates, which in turn deflates the number of parties in the legislature (even when the legislature is returned by proportional representation elections) – the deflationary effect of the plurality presidential election rule on party fragmentation.
      • The majority run-off rule proliferates the number of presidential candidates, which in turn inflates (or at least does not deflate) the number of parties in the legislature (provided that the legislature is returned by proportional representation elections) – the inflationary effect of the majority run-off presidential election rule on party fragmentation.[1]
    • Electoral cycle – the temporal proximity of presidential and legislative elections
      • The electoral cycle attenuates the strength of the inflationary/deflationary coattails.
      • The temporal proximity of presidential and legislative elections varies along a continuum ranging from maximally proximate to minimally proximate
      • Maximal proximity denotes that the presidential and legislative elections are held concurrently.
      • Minimal proximity denotes that a legislative election is held at the presidential midterm.
      • In general, the more temporally proximate the presidential and legislative elections are, the stronger the coattail impacts will be.
      • Thus, the presidential coattails exert the strongest impact on reducing or increasing legislative fragmentation when presidential and legislative elections are held concurrently.
      • Conversely, the further the presidential and legislative elections are temporally apart, the weaker the coattail impacts will be.
      • Thus, the presidential coattails exert the weakest impact on reducing or increasing legislative fragmentation when a legislative election is held at the presidential midterm.
      • In between the two extremes when presidential elections and legislative elections are held non-concurrently but are relatively close, it has been found that:

– the coattails of presidential elections preceding a legislative election are less strong than the coattails of presidential elections subsequent to a legislative election.

  • Some figures may help illustrate the magnitudes of the presidential coattails on legislative fragmentation.
  • Statistical models using data of all democratic legislative and presidential elections between 1946 and 2000 show the following predictions:[2]
    • When a parliamentary system with 4 parties (effective number of legislative parties) turns into a presidential system under the scenario of a two-candidiate presidential race (under the plurality rule), all else being equal, the models predict that:
      • the number of legislative parties will reduce to about 3 when the presidential election and the legislative election are held concurrently.
      • the number of legislative parties will be in the range of 3.4-3.8 when the legislative election is held at 1/4 of the presidential term.
      • the number of legislative parties will be in the range of 3.9-4.7 when the legislative election is held at the presidential midterm.
    • Conversely, when the presidential election is a 6-candidate race (under the majority-runoff rule), the models predict that:
      • the number of legislative parties will increase to about 7.5 when the presidential election and the legislative election are held concurrently.
      • the number of legislative parties will be in the range of 6.7-7.3 when the legislative election is held at 1/4 of the presidential term.
      • the number of legislative parties will be in the range of 5.9-7.2 when the legislative election is held at the presidential midterm.
  • An earlier study using data from 16 Latin American countries have the following projections:
    • Changing a presidential system with the plurality-concurrent electoral format to one with the majority-runoff-concurrent electoral format will increase the level of multipartism by 1.5 times.
    • Changing a presidential system with the plurality-concurrent electoral format to one with the plurality-non-concurrent format will increase the level of mulipartism by 1.43 times.[3]
  • Regarding the impacts of presidential electoral rules on the number of presidential candidates, data from 54 elections across 13 countries show that “the mean effective number of presidential candidates… was 3.8 under [majority run-off], as opposed to 2.8 under plurality.”[4]

 

Partisan support for the president in the legislature

  • The importance of the office of the presidency affects voters’ voting behavior in legislative elections.
  • In presidential systems, there is a tendency for voters to cast their votes in legislative elections according to their evaluation of the president and her/his party (unless legislative electoral rules encourage legislators to campaign on bases other than their party labels).
  • Since presidential popularity varies along a president’s term, it can be expected that the electoral prospect of the president’s party in legislative elections and hence the share of its seats in the legislature will depend on when the legislative election is held in the president’s term. In short, the electoral cycle affects the share of seats of a president’s party and hence his/her partisan support in the legislature.
  • Empirical research on the relationship between the electoral cycle and presidential partisan support in the legislature shows the following:[5]
    • Presidents tend to enjoy their greatest popularity early in their terms during their “honeymoon” period, and to experience decreasing levels of popular support thereafter.
    • When presidential elections and legislative elections are held concurrently, voters tend to cast straight partisan ballots for both president and legislature, with the result of parties running strong presidential candidates being rewarded in terms of their share of seats in the legislature.
    • It is estimated that the party of the winning presidential candidate can expect, on average, to increase its share of legislative seats by around 7% under concurrent elections.
    • If a non-concurrent legislative election is held immediately after the election of a new president, the expected increase in the share of legislative seats for the new president’s party jumps to 19% (as the originally wavering voters may tend to vote for the new president’s party).
    • However, in legislative elections held after this point, the expected gains in legislative seats for presidential parties decline in a linear fashion over time, dropping by more than 4% from the “honeymoon high” for each tenth of the president’s term.
    • When a legislative election is held at the presidential midterm, the expected change in the share of seats of the president’s party is negative.
    • An even greater loss of seats is expected if a legislative election is held later in the presidential term.
    • In short, the later a legislative election is held in a president’s term, the less partisan support in the legislature a president will enjoy.

 

Political consequences of presidential electoral systems and their implications for electoral design

  • Thus, electoral cycles and presidential electoral formulas have bearings on the effectiveness of governance, legislative representativeness, and accountability of presidential systems.
  • They bear on legislative representativeness because the less fragmented a legislature is, the less representative it may be.
  • They bear on the effectiveness of governance because the more fragmented a legislature is,
    • the lower the expected share of legislative seats the party of a president will hold (the presidential popularity cycle impacts on this further), and hence
    • the less likely the president will enjoy majority legislative support.
  • At the same time, a fragmented legislature will make it more difficult for the president to solicit legislative support or to form a legislative coalition.
  • The overall outcome will be the president’s ability to act being seriously undermined.
  • The electoral cycles and presidential electoral formulas also bear on accountability because the more fragmented a legislature is,
    • the more likely that a no-majority situation will appear, and hence
    • the more a president will rely on a grand legislative coalition or a shifting coalition to get his/her policy initiatives through in the legislature.
  • But when everyone is responsible for policy, the actual outcome will be that nobody is responsible!
  • It follows from the above that if our electoral design objective is to make the legislature of presidential systems more representative, apart from adopting a PR formula for electing it, the plurality rule for electing the president should be avoided.
  • Conversely, if our objective is to make the government of presidential systems with proportional representation legislative electoral rules more effective, the plurality rule for electing the president should be adopted and the elections of the legislature and the presidency should be held concurrently or closely enough (so as to maximinize the seat gain of the president’s party in legislative elections through utilizing the “honeymoon” of the presidential popularity cycle).
  • Likewise, if our objective is to enhance the accountability of the presidential system (in the sense of providing voters with a clear sense of who is responsible for what), the same electoral arrangement as in the preceding point should be adopted.
  • Through deflating the number of legislative parties and enhancing the chance of electoral success of the president’s party in legislative elections, the plurality-concurrent (or close enough) format helps manufacture a majority in the legislature, which in turn will improve the effectiveness of governance and accountability of the system.
  • Finally, it is noteworthy that the research findings presented in this brief cast serious doubt on the received wisdom of adopting the majority-runoff rule for electing presidents.
  • One of major arguments in favor of the majority-runoff rule is that it will ensure the election of a president with a mandate from a majority of voters and hence enhance his/her ability to act. However, as mentioned before, to the extent that the majority-runoff rule contributes to fragmentation of the legislature, it may simply undermine the president’s ability to act rather than strengthen his/her hand!

 

Lessons for Hong Kong

  • One noteworthy feature of Hong Kong’s electoral system is that the terms of the Chief Executive (CE) and the Legislative Council (LegCo) are of different lengths.
  • The terms of CE and LegCo are 5 years and 4 years respectively.
  • This means that the temporal proximity of CE and LegCo elections displays an irregular pattern.
  • At one time, the two elections are concurrent.
  • At other times, the LegCo election is held at the fourth year, or the third, or the second, or the first year of the CE’s term.
  • Assuming that the CE is allowed to have party affiliation and also that the popularity cycle of Hong Kong’s CE is the same as elsewhere, the irregular pattern of the temporal proximity of CE and LegCo elections suggests that the prospects for partisan support in LegCo will be quite different for alternate CEs.
  • Yet this difference stems not from the configuration of electoral support but is completely the result of institutional design.
  • The irregular pattern also makes predicting the deflationary/inflationary impact of different CE electoral rules on legislative fragmentation difficult.
  • It is likely that the deflationary or inflationary impact will take a relatively long period of time to consolidate.
  • Notwithstanding all the above, however, Hong Kong may still want to consider adopting the plurality rule for electing the CE if improving the governance of Hong Kong is the objective.
  • In the long run, in order to make our system more rational, we should consider amending the terms of the CE and LegCo to make them of equal length.

 

Prepared by Winston Ng (吳君韻)

Hong Kong Democratic Foundation (香港民主促進會)

19 April 2013

 

 

Endnotes:

[1] For a brief review on how electoral forumlas affect the number of presidential candidates, see John Carey (1998). Regarding the mechanism of presidential coattails, see also John Carey (1998) as well as Matt Golder (2006) and Mark Jones (1994).

[2] Heather Stoll (2011).

[3] Mark Jones (1994)

[4] John Carey (1998), p. 97.

[5] The following findings are quoted or paraphrased from John Carey (1998), pp. 102-103. The primary source of the findings is from Matthew Shugart (1995).

 

 

References:

Amorim Neto, Octavio and Cox, Gary. 1997. Electoral institutions: cleavage structures, and the number of parties. American Journal of Political Science 41(1):149-74.

Carey, John. 1998. Constitutional choices and the performance of presidential regimes. Journal of Social Sciences and Philosophy 88(3): 93-122.

Golder, Matt. 2006. Presidential coattails and legislative fragmentation. American Journal of Political Science 50(1): 34-48.

Hicken, Allen and Stoll, Heather. 2013. Are all presidents created equal? Presidential powers and the shadow of presidential elections. Comparative Political Studies 46(3): 291-319

Jones, Mark. 1994. Presidential election laws and multipartism in Latin America. Politic Research Quarterly 47(1): 41-57.

Shugart, Mattew. 1995. The electoral cycle and institutional sources of divided presidential government. The American Political Science Review 89(2): 327-343.

Stoll, Heather. 2011. Presidential Coattails: A Closer Look. Unpublished manuscript. Earlier versions of this paper were presented at 2011 Southern California Political Institutions Conference, University of California, Los Angeles, 16 September, and 2011 National Conference of the Midwest Political Science Association, Chicago Illinois, 31 March–3 April.

 

 

 

Facebook Comments
Recent Posts
The Professional Commons’s response on Healthcare Reform Second Stage Consultation 
Developing country park – from absurdity to reality
Taking stock of the local agricultural industry is more than a matter of market values (Chinese version only)
Our dignified autonomy: farmers for the umbrella movement (Chinese version only)
China survival tip: where in the world are Guangdong’s hazardous chemical wastes? (Chinese version only)
Political expectations of the new China-HK model in post-political reform Hong Kong (Chinese version only)

Start typing and press Enter to search

Home
Home