Taken from: Governance and Democracy in the Commonwealth
Caribbean: An introduction by
Patrick Emmanuel pp 41-50
the Commonwealth Caribbean, problems have arisen both over the nature of the
electoral systems in principle and with regard to the practical administration
of various parts of them, in particular over such matters as relative
constituency size, the setting of constituency boundaries, voter registration
and balloting. General elections, even by-elections, have been marked by serious
conflict, including violence. The boycotting of elections by major parties in
Trinidad and Tobago (1971) and Jamaica (1983) stemmed from major disagreements
over the use of voting machines, and charges of corrupt registration practices,
in Guyana, all the countries practice the first-past-the post (FPP) electoral
system, their Constitutions and electoral laws generally providing for the
division of national space into constituencies, from which a single
representative, gaining the highest number of votes, or being the sole
candidate, is elected. In all these cases, except Jamaica, there are no
constitutional limits on the (maximum) number of constituencies, which as shown,
is alterable by the elected Houses of Parliament. In Jamaica, however, section
67 (subsection 1) of the Constitution places an upper limit of sixty on the
number of elected members.
In Guyana, under a list system of proportional representation, the Constitution provides for the election of fifty-three representatives on the basis of each elector being entitled to cast a single vote in favour of any one of the party lists of candidates. The fifty-three seats are allocated among the parties “in such a manner that the proportion that the number of such seats allocated to each list bears to the number of votes cast in favour of the list is nearly as may be the same for each list” (S. 160:1C). Parliament is empowered to provide, inter alia, for the manner in which lists of candidates are prepared, including additional names to fill casual vacancies in lieu of by-elections; for the precise manner in which the number of seats to be allocated to each list is calculated; and for the extraction from the lists and declaration of the names of candidates elected (S. 160:3). Incidentally, Parliament may also provide that party lists of candidates may be combined for the purpose of allocation of seats, but not for the purpose of voting (S. 160: 3d).
The election of the President is simultaneous with the
Parliament. Each list of candidates is required to designate a candidate for the
Presidency and, unless there is only one such candidate, the candidate on the
list which receives the largest number of votes becomes President. In the event
of a tie between two or more lists the Chairman of the Elections Commission
shall publicly choose one of the candidates “by lot”, and that person is
declared President (S. 177).
When compared with PR methods, the FPP electoral method has the major
systemic defect of disproportionality between proportions of seats won and votes
received. This is a world-wide experience rather than a merely Caribbean one.
The Caribbean experience in this regard is illustrated by looking at the
proportions of seats and votes received by competing parties in samples of
elections in several of the FPP countries (Table 3).
The table shows an almost unbroken pattern of disproportionality
in the relationship between seats and votes won, always in favour of the winning
party. Losing parties (and independents) relatedly are disadvantaged as they
routinely receive shares of seats significantly lower than their shares of
votes. During the period covered there were six occasions on which the winning
party received less than 50 per cent of the votes.
But when data for specific losing parties are
examined, the level of disadvantage is even more disturbing.
In recent times there have been three instances (Antigua 1984, St. Vincent
1989 and Trinidad 1981) where a party won no seats but polled vote shares of
23%, 30% and 22% respectively. A party, also won one seat while receiving 31% of
total votes (Antigua 1989) and one seat with a national vote share of 36%
Cases of Disproportionality
Country Election Party No. of Seats (%) %Votes
1984 UPM 0 23
1989 UNDP 1(6) 31
1989 SKLP 2(18) 38
St.Vincent 1989 SVLP 0 30
Trinidad 1981 ONR 0 22
In the remaining cases, vote shares of 39%, 41% and 38% realized three or
two seats seats.
the manifest uneveness of the FPP method has not engendered any very strong and
persistent demands for the introduction of the PR system throughout the region.
Supporters of the FPP ii argue, in effect,
that what its opponents see as its major defect, its disproportionality, is
indeed its major advantage, as the consequence of disproportionality is
effective governing majorities. The record in fact shows that ruling parties
have had quite comfortable parliamentary majorities, with the one-seat
majorities obtained in Dominica1990 Montserrat 1991 and St. Lucia 1987, being
It has been only in Trinidad and Tobago that the case for
electoral reform, involving the introduction of PR, has had widespread advocacy.
So much so that the 1972 Constitution Commission balancing the merits of both
FPP and PR methods actually recommended the adoption of a mixed system by which
one-half the seats in (a unicameral Parliament) would be chosen by the FPP
method and the other one-half would be allocated on the basis of proportions of
votes parties received in the FPP voting.
In this regard it is of considerable interest to note that
Guyana’s Constitution provides for the introduction of just such a mixed FPP-PR
system if Parliament so decides:
“Parliament may make provision for the division
of Guyana into such number of electoral areas, not being more than half the
number of the said fifty-three elected members of the Assembly, as Parliament
may prescribe and for the election in each such area of one member of the
Assembly, each elector having for this purpose one vote in addition to the vote
which he may cast in favour of a list of candidates ...“(Art. 160:2).
Thus far, however, the
Guyanese Parliament has not opted to introduce the mixed system and elections
and seat allocation there have continued to be under the PR system only.
In Trinidad and Tobago, the proposal of the first
Constitution Commission for the introduction of a mixed FPP and PR system was
strongly rejected by the then PNM Government and the new republican Constitution
of 1976 contains a clause (section 73:1) which provides that: “The election of
members of the House or Representatives shall be by secret ballot and in
accordance with the first-past-the post system”. The 1987 Commission fully
supported retention of the FPP system for elections to the House.
Commission however advanced a quite novel proposal that in an enlarged Senate,
the quota of 12 opposition senators should be allocated “in proportion to the
votes cast at the general election for those parties not forming the Government ...“
with provisions for the minimum number of seats contested and votes
received by such parties (pp. 81-82). But in a minority statement, three of the
Commissioners rejected proportionality, arguing that “any differences and
divisions that exist in our society today will only be exacerbated by the
introduction of any type of proportional representation, however diluted or
disguised” (p. 96).
Oddly, the minority statement also observed that “proportionality, i.e., the relationship between seats obtained and votes cast, does not vary substantially between the first-past-the-post and proportional representation system” (p. 96). However, a study of the outcomes of 101 general elections in ten Caribbean states (see Tables 3 and 4) categorically refutes this assertion (Emmanuel, 1992). Indeed, in the 1991 general elections in Trinidad and Tobago, the distribution of votes and seats amongst the three major parties was as follows:
Party % Vote Seats Won (%)
PNM 45.1 21(58.3)
UNC 29.1 13 (36.1)
Source: Elections and Boundaries Commission
PNM won a comfortable seat majority with only 45.1 per cent of the vote, while NAR with almost one-quarter of the vote won only 2 seats. UNC, like PNM, also took a greater share of seats (36.1 per cent) than its share of total votes (29.1 per cent). It will be recalled, as shown in Table 4, that in 1981 ONR received 22 per cent of the votes but no seats. At the same time ULF with 15.2 per cent votes, won 8 seats (22.2 percent).
If the 1991 elections in Trinidad and Tobago were held
along the lines of mixed FPP and PR proposed by the 1972 Commission, seats in a
House of 72 members would have been allocated as follows:
Trinidad and Tobago: Seat Allocation by FPP & PR System
Party (% Vote) FPP Seats PR Seats Total
PNM (45.1) 21 16 37
UNC (29.1) 13 11 24
NAR (24.0) 2 9 11
Under this mixed system, the 36 additional seats are
allocated in direct proportion to votes cast for the parties by the FPP voting.
The mixed system would preserve a majority for PNM, i.e. 37 of 72 seats, but
this is reduced from the margin, 21 of 36, afforded by the FPP system only. On
the other hand, a pure PR system alone could have resulted in a distribution of
PNM 16, UNC 11 and NAR 9 (assuming of course that in a pure PR election, the
party vote shares would have been the same as occurred in the actual FPP
election). Thus the PR system alone would have denied any party majority and
called for the very coalition-based government which is regarded as a major
defect of PR.
Elsewhere in the region, notwithstanding the long
experience of FPP disproportionality, PR has had few advocates. The Constitution
Commission in Barbados briefly considered the issue and concluded that “the
introduction of such a system [of PR] could create other problems which did not
seem to justify a recommendation for such a system” (p. 28). However, in a
minority report, one Commissioner argued for a system of “Fixed Proportional
Representation” designed to “produce a viable government and a strong
opposition and discourage nuisance parties”. Under this system, in a House of
30 members, a minimum of 18 seats would be allocated to the party gaining the
most votes and a minimum of 9 to the next best supported party, with the
remaining 3 seats allocated “in accordance with the results of the election
provided that in no event the Government party will hold seats in excess of
two-thirds [i.e 20] of the total number of seats” (p. 194).
A key proviso is that a party must receive at least 4 per
cent of the votes to qualify for seat allocation. However, as described, the
system anticipates only two qualifying parties and does not explicitly provide
for sub-division of opposition seats where as in 1991, two opposition parties (BLP
and NDP) would be eligible for seats.
But the major problem
with this proposal is in fixing a minimum number of seats allocated to the
winning party, irrespective of the vote share gained vis-a-vis all other
parties. This defect results from a preoccupation with securing comfortable
majorities rather than democratic proportionality, which would allow a wider
variety of party interests a chance to be represented in Parliament.
shown, there are constitutional requirements in the countries practising the FPP
electoral system, that all constituencies be as nearly as practicable of equal
size, having regard to the influence of such factors as population density,
geographical barriers and administrative boundaries. Approximate equality of
constituency size is essential for satisfying the important principle of equal
weight of each ballot. Wide differences in size may also exacerbate the already
inherent disproportionality between shares of votes and seats seen to
characterise the FPP system.
The following Table (7) applies the yardstick adopted in
Trinidad and Tobago and Barbados, that no constituency should be less than 90
per cent nor more than 110 per cent of the average size of constituencies, to
illustrate the pattern of inequality existent in several countries in recent
The table shows, by country (and election year), the number
of constituencies, or seats (Column A), the total electoral registration (B),
average constituency size (C), the numbers of constituencies that were more than
110% of the average (D), and less than 90% of the average (E), and the totals
for (D) and (E).
Separate data are provided for the smaller territories
in the five cases where constituencies exist on more than one land mass. In
these instances, populations are small and average constituency size is less
than on the main land mass.
Trinidad and Tobago is the only case in which all
constituencies fell within the set range. Barbados and mainland St. Vincent are
the other noteworthy cases, with only a few constituencies, i.e., four out of
twenty-eight and two out of thirteen respectively, slightly outside the limits.
In all other cases, however, there are considerable
disparities, as reflected in the Column (F) totals, relative to Column (A),
total seats. All constituencies in St. Kitts fell outside the 90-110 percentage
range. In other territories, the proportions (of constituencies) falling outside
this range are Dominica and Grenada - 86
per cent, Antigua -81 per cent, St. Lucia -
71 per cent, Belize and Jamaica - 50
per cent, and Montserrat - 43 per cent. Contrastingly, in Barbados the
proportion is 14 per cent and in St. Vincent - 15
The data show a disconcerting disparity of
constituency size, with exceptions of course. The extent of disparity varies
within countries, some constituencies being only slightly larger or smaller than
the outer limits, while others are considerably larger or smaller. Table 8
provides data for the largest and smallest constituencies in each main island
and Belize for the election years being used.
The ratios of smallest and largest constituencies are
lowest in the cases of Trinidad, Barbados and St. Vincent, but are quite high in
St. Kitts, Grenada, Antigua, Montserrat, Dominica and St. Lucia. Jamaica and
Belize are intermediate cases.
It has of course been pointed that in all the cases the constitutional stipulation regarding equality of constituency size is qualified by the need to have regard for geographical, population, communication and administrative factors. These qualifications, reasonably applied, will help explain the various levels of disparity found in some cases.