Revised June 1999

MALTA: STV IN A TWO-PARTY SYSTEM

Wolfgang Hirczy de Miño - University of Houston

and

John C. Lane - State University of New York at Buffalo


ABSTRACT

Although STV is often held up as a voting system that maximizes voter choice and does not presuppose partisan candidacies, political parties play a key role in Malta and are firmly entrenched at every level of electoral politics, including the mass public. There are virtually no formal barriers that impede candidacies by independents or third parties, yet only two, the Nationalist Party (PN) and the Malta Labor Party (MLP) have been represented in parliament for the past thirty years. In the Maltese setting, STV primarily serves to allow for competition within parties. The incidence of cross-party transfers is minimal. The two major parties routinely over-nominate. Even though the system is highly proportional, a minor deviation in the vote-seat ratio led to a major political crisis in 1981, when a party whose candidates had received an absolute majority of first-preference votes wound up as a loser in parliamentary representation. The ensuing impasse was resolved through a constitutional amendment guaranteeing that the party winning the popular (i.e. first-preference) vote will form the Government, by giving it as many additional seats as are necessary to achieve a parliamentary majority. First-preference votes alone, rather than the number of candidates initially elected, may thus determine which party will control the Government in Malta.


Malta, a small island nation in the middle of the Mediterranean Sea, was formerly a crown colony of Britain. It became a sovereign and independent state within the Commonwealth in 1964 and a republic in 1974. The country's experience with elections to representative institutions, however, predates its independence by more than a century. The first vote by secret ballot, for a council of government, took place as early as 1849 and STV has been in continual use since 1921. However, various representative institutions were often suspended by the colonial authorities in the years that followed and the elected bodies lacked comprehensive powers prior to country's independence.

The Malta Independence Constitution of 1964 established the country as a liberal parliamentary democracy. It provides for protection of fundamental human rights, separation of powers between executive, legislative, and judicial branches, and regular elections based on universal suffrage. Given its small size (320km2 and a population of about 375,000), Malta does not have a federal structure and only introduced locally elected town and village councils with very limited powers in 1993.

Malta has a unicameral parliament, known as the House of Representatives. Not counting any 'bonus' seats, its size is presently fixed at 65 members. From 13 districts of five members each, the members are "elected upon the principle of proportional representation by means of the single transferable vote." This electoral formula is constitutionally entrenched and has been used in all 20 elections conducted since 1921. The maximum length of a parliamentary term is five years.

Formally, executive authority is vested in the President of Malta who is elected by parliament for a five-year term (the current incumbent is Dr Ugo Mifsud Bonnici, elected in 1994). While his or her duties are largely nominal, they include the power to appoint as Prime Minister "the member of the House of Representatives who, in his judgment, is best able to command the support of a majority of that House". Real political power is exercised by the Prime Minister and Cabinet, consistent with the Westminster model. The President signs bills into law, dissolves parliament, and discharges other responsibilities, but except in rare constitutional matters must act on the advice of the Government.

Selection of cabinet members is a prerogative of the PM, but all of them must be drawn from the parliament. The Government presently consists of the PM and fourteen Cabinet Ministers, assisted by five Parliamentary Secretaries. Dr Fenech Adami was Prime Minister from 1987 to 1996 and returned to that office in 1998, following the premature end of Dr Alfred Sant's Labour Party government.



INSTITUTIONAL FEATURES OF MALTA'S ELECTION SYSTEM

A notable aspect of Maltese electoral politics is the co-existence of an electoral formula (STV) whose distinguishing emphasis is on separate individual candidacies, with the pre-eminence of political parties from the very beginning. Over the years the constitution, laws and regulations have increasingly acknowledged and promoted the role of political parties as central actors in the electoral process. Thus, the constitution requires that the political party affiliation of candidates be indicated on the ballot paper and since 1987 it requires that the aggregate number of first-preference votes for political party candidates, and not simply the success of individual candidates, may determine how many seats a party will have in parliament. These provisions and a host of others which create a pervasive accommodation of political parties in the electoral process, will be described in some detail on the pages that follow.

There are two sources of law that define the legal framework for the electoral process: Malta's Constitution and the General Elections Act of 1991. An Electoral Commission, which administers elections and maintains the Electoral Register through its Electoral Office, is further given discretionary authority to establish procedures, amend forms, and to make additional rules as deemed necessary. In various ways, as described below, it shapes the conditions and the conduct of elections. The seven members of the Electoral Commission are chosen by the Prime Minister, after consultations with the Leader of the Opposition.

Districting and district size

Malta is presently divided into 13 electoral districts from which five MPs each are to be elected by STV. The actual number of districts and seats is determined by Parliament within constitutional parameters which stipulate that the number of districts be between 9 and 15 and the number of representatives per district between five and seven. footnote 1 The average number of eligible voters per district has by now grown to over 21,600. With a district size of five (ignoring the possibility of bonus seats), each seat in 1998 required approximately 4,300 valid votes. Although each MP is personally elected on the basis of preference votes, there is no one-on-one constituency representation because of the multi-member districts and secrecy of the vote. The quota needed for election pursuant to the Droop formula varies somewhat because of slight variations among the districts in the number of voters and in turnout. In the 1998 elections the quota in the 13 districts ranged from 3,257 to 3,642 votes (see Figure 1).

The Constitution mandates that the Electoral Commission review the boundaries of the electoral districts at intervals of no less than two and no more than five years. The Commission may alter such boundaries as it deems necessary to assure that the ratios of electorate size and members returned to Parliament from the districts are as equal as reasonable practicable. Factors such as geographical contiguity and variation in population density may be taken into account, but the net effect of the deviation must be within a margin of 5%. As Figure 1 indicates, the once large disparities in district size have disappeared since the 5% requirement went into effect in 1976. The various proposals by the Electoral Commission for boundary changes have, over the years, regularly brought complaints from political parties and parliamentarians as involving partisan maneuvers. Under present constitutional rules, parliament can request a reconsideration of newly determined district boundaries but lacks the power to override the Commission's final determination.

Voter eligibility and registration

The Constitution stipulates that to become a registered voters a person must be a citizen of Malta, at least 18 years old, and a resident in Malta for at least six months. (However, permanent foreign residents are eligible to vote in local council elections.) Grounds for exclusion are mental incapacity or infirmity, imprisonment, and conviction of an election-related offense.

Voter registration is automatic and individuals need not make application to be registered. The Electoral Commission compiles and maintains the Electoral Register, updates it continually and publishes it twice a year in the Government Gazette, with all new registrants specifically identified. The Electoral Register is a public record and is also released in electronic form.

All voters must vote in the district where they are registered. A voting document is issued to each voter before each election, and served in person by the police or a designated agent of the Electoral Commission. The voting document contains a photograph of the voter in addition to name, place of residence, and date of birth to properly verify the voters' identity.

Nomination of candidates

There are no significant barriers for any eligible voter to stand as a candidate. It only takes self-nomination plus four supporting signatures and a deposit of 40 Maltese Liri (ca. $100), which is refunded if the candidate attains at least 10% of the quota in the district. Political parties identify their candidates by providing the Commission with lists of candidates contesting the election in their interests. This list serves as conclusive evidence that the candidate belongs to the respective party.

Candidates may contemporaneously stand in two districts and such dual candidacies have become quite common in recent years, as Figure 2 attests. Whenever a dual candidate wins in both districts, he or she must give up one of the seats. This happens frequently enough to make the subsequent "casual elections" to fill the vacated seats a regular occurrence following a general election. The procedures for such elections are described below.

Ballot form and structure

The Constitution mandates that voting shall be by secret ballot and requires that the ballot paper be drawn up in such a way as to enable illiterates to distinguish between the political parties to which candidates belong. The Election Act specifies that the parties be denoted by different colors; they may also use different icons. Since 1976 candidates have been listed alphabetically (by last name) within separate party blocs; in earlier elections, all candidates appeared on the ballot in alphabetical order regardless of party. The length of the ballot necessarily varies with the number of candidates; in one district in 1962 it contained 38 individual candidates.

Voting procedure

All voters must vote in the ballot booth designated on their voting document and present this document to prove their identity. Voters are handed a paper ballot authenticated with an official stamp, which they mark in the secrecy of the voting booth, fold, and then deposit in the ballot box in view of the Assistant Commissioners.

The voter must place the number 1 in the line of his or her preferred candidate and is encouraged to continue expressing lower-order preferences by placing consecutive numbers next to the names of additional candidates, but is not required to do so. X-voting and assignment of the same number to two candidates are not permissible. Instructions for voting are found on the back of the ballot.

Special assistance is available for the blind, physically infirm, or illiterate. Subject to a number of conditions protecting the integrity of the process, Assistant Commissioners may record the preferences for such individuals on the ballot on their behalf. They shall ask the voter 'Which of the candidates do you most desire to see elected?' and place the number 1 accordingly. Next they are to ask: 'For which of the candidates do you desire to express a second choice?' and so on until the voter no longer wishes to express a preference. At the request of the voter, the Assistant Commissioner shall read out the list of available candidates in the order in which they appear or - if the voter so desires - the names of the candidates of one particular political party only, omitting the names of others. Special rules also exist for voting in retirement homes and transportation to and from polling places. To facilitate participation by eligible voters temporarily residing abroad, low-cost air and sea transportation is available (Government Press Release No. 1533; October 2, 1996).

Counting rules and procedures

There are three distinguishing features of Malta's version of STV which should be noted: (1) the transfer of transferable votes to continuing candidates in proportion to their respective strength in the pool of transferable ballots examined at any one particular count; (2) the use of so-called casual elections (count-backs) to fill vacancies; and (3) the dual character of the first-preference vote which, since 1987, is not only a candidate-preference vote but, in given circumstances, also a party vote determining control of government. Each deserves a brief explanation.

Counting and transfer rules. Any candidate obtaining the requisite number of votes (the so-called "quota") in the first count is declared elected. The quota is determined by dividing the total number of valid votes cast by six (the number of candidates to be elected from each district plus 1) increased by one, ignoring any fractions.

In the second count, the second preferences recorded on all ballots of the elected candidates are examined and transferred in proportion to their respective share in the elected candidate's pool of ballots, after the number of votes needed to meet the quota of the elected candidate has been subtracted. Footnote 2 If there are several surpluses, the larger one is dealt with first.

In subsequent counts, the same procedure is used except that the proportional allocation is based on the ballots examined in the last count. Ballots credited to a candidate in an earlier count are not reexamined unless the candidate is excluded and his/her votes are to be transferred to other continuing candidates. Candidates are eliminated, and the votes previously credited to them transferred to those still eligible to receive transferred votes, when no candidate meets the quota in any particular count. The candidate with the lowest number of votes is excluded first. Several candidates are excluded on the same count if their combined number of votes is less than that of the next-higher candidate.

If upon exclusion of a candidate or candidates the number of continuing candidates is one more than the number of seats remaining to be filled, the candidate(s) with the higher number of votes is(are) deemed elected and no further transfers take place. If the number of continuing candidates (i.e. those eligible to receive transfers) drops to the number of remaining seats to be filled, no further counts and transfers will take place and the remaining candidate(s) is(are) declared elected. The number of counts needed to fill all seats is largely a function of the number of candidacies in the district. It ranged from a low of two in 1921 to a high of 37 in 1962. (See Table 1).


TABLE 1: NUMBER OF REQUIRED COUNTS, BY YEAR AND DISTRICT, 1921 - 1998

+-------------------------------------------------------------------------+
|      |                          District                                |
|      |  1st  2nd  3rd  4th  5th  6th  7th  8th  9th 10th 11th 12th 13th |
| Years|------------------------------------------------------------------+
| 1921 |    6    9    7    8    9    7    6    2                          |
| 1924 |   13    9   11    9   12   10    9    6                          |
| 1927 |    8    9    8    8    9   10    8    8                          |
| 1932 |    7    9    9    8    9   10    7    6                          |
| 1939 |   10   10                                                        |
| 1945 |    3    7                                                        |
| 1947 |   13   11   10   16   15   12   13   13                          |
| 1950 |   18   15   18   20   23   23   21   23                          |
| 1951 |   17   12   10   11   17   13   16   11                          |
| 1953 |   24   12   14   20   17   17   23   17                          |
| 1955 |   19   15   12   10   13   12   20   22                          |
| 1962 |   28   23   23   24   23   31   37   26   28   35                |
| 1966 |   29   15   14   18   13   27   29   29   23   35                |
| 1971 |   18   17   13    8   10   14   12   19   17   26                |
| 1976 |   19    8   10   10   14   12   16   16   20   15   14   13   23 |
| 1981 |   21   10   12   10   11   12   18   13   16   16   17   18   21 |
| 1987 |   22   17   12   10   13    9   18   13   17   19   19   13   14 |
| 1992 |   21   17   19   14   13   12   21   10   19   20   17   19   13 |
| 1996 |   17   19   17   19   11   13   18   15   20   21   16   18   10 |
| 1998 |   18   24   19   20   16   14   16   24   25   23   21   20   14 |
+------+------------------------------------------------------------------+

Rules for filling "casual vacancies". When a legislative seat becomes vacant there is no going back to the voters in Malta, in contrast to the use of by-elections in the UK or Ireland, another country using STV. Instead, vacated seats are filled by means of a recount of the vacating candidate's ballots papers. Any candidate who stood in the district in the General Election and was not elected is eligible for the election to the vacated seat. Functionally, the reuse of the ballots to select a replacement MP is similar to use of STV in a single-member district. Instead of the formula, a straight 50% rule is used as a quota. Although candidates can be elected with fewer votes than the quota in the General Election, a mere plurality will not suffice in the casual election.

The winners are determined as follows: The Commission opens the parcel of the ballots credited to the vacating MP and starts by crediting them to the first available valid preference (skipping excluded candidates). All candidates other than those validly nominated for the vacant seat are deemed to be excluded. In a case where there is only one validly nominated candidate, he/she will be declared elected if he/she has at least half of the number of papers in the parcel of the vacating MP. In any case where the number of nominees is two or higher, the Commission will exclude from the poll the candidate with the lowest number of votes and transfer them to the next available valid candidate. The process of excluding the weakest candidate, and of transferring his/her votes to the next available valid preference continues until there is only one candidate left. This candidate will be elected if he/she has accumulated at least half of the votes in the parcel. If this condition is met earlier and a candidate also has more votes than all other candidates taken together, the candidate shall be declared elected without any further transfers. If no candidate secures election or none is validly nominated, the vacancy is filled by co-optation by the Members of the House.

Second-Tier Adjustment (allocation of bonus seats). With the relatively small district magnitude of five seats, fairly sizable disproportionalities between vote and seat percentages are unavoidable. In most elections, these discrepancies at the district level have canceled themselves out on a national basis. However, the 1981 election produced the "perverse" result of the Nationalist Party obtaining a majority (50.9%) of the first-preference votes nationally but nevertheless failing to obtain a majority of the parliamentary seats. The MLP with only 49.1% of the vote won 52.3% of the seats and formed the government.

To prevent a recurrence of such a result, a constitutional amendment was adopted in 1987 to assure that the party winning a majority of the first-preference votes will also be given a parliamentary majority by allocating so many additional seats as are necessary to give it a one-seat majority. Winning is defined as more than 50% of the valid votes in the aggregate at the first count of all the votes. A second amendment was added in 1996 to assure a party winning only a plurality of the popular vote of a parliamentary majority, but only in a situation where more than two parties compete for votes and only two manage to get candidates elected to parliament. (For the text of these provisions, see Section 52(1) of the Constitution.)

In both instances, any additional seats are to be filled by those candidates of the party in question who had the highest number of votes (second-highest, and so on in descending order) at the last count but failed to get elected, irrespective of the district in which they stood. Because of these provisions the marking of the first preference on the ballot is functionally equivalent to x-voting for a party in a party-list system, while at the same time it serves as a vehicle for a personalized candidate-centered vote.

Local government

Local government was introduced by the Nationalist government through the 1993 Local Councils Act. There are 67 local councils with a membership of between 5 and 13, depending on the community's size. Council members are elected by the same STV method as prevails at the national level. Elections are held every three years with a third of the councils up for reelection each year. The question of whether political parties should field candidates became a matter of some dispute. The MLP (and the Church) argued for non-partisan elections at the local level whereas the PN preferred contests by party candidates. Eventually, there were in every community candidates under the PN banner and others as standing as independents, although many of the independents were in fact MLP members or supporters. The MLP government (1996-98) planned to introduce legislation to make council elections non-partisan ones; but it failed to do so and by the March 1999 local elections the party had begun to field candidates under its own banner.

The councils are in charge of handling some services within their communities, such as minor beautification projects and street maintenance; but they are not free to levy their own taxes and their budgets are entirely determined by annual grants from the central government. Probably as a consequence of these limited powers and financial resources, the voter turnout for local elections has lagged far behind that for national elections. In the first election cycle only 65 percent of the eligible voters cast valid votes and this declined to 61 percent in the second cycle of elections. The local elections may prove to be a new avenue for entry into national politics, as 40 persons whose only prior candidacy had been at the local level contested 46 seats in the 1996 election and won seven of them.

Having described the major features of the institutional framework of elections in Malta we will now turn to the dynamic, behavioral factors and the actual conduct of campaigns and elections that represent the intersection of the two. Electoral behavior is certainly shaped by institutions and the applicable rules (and we will identify additional ones where appropriate), but the conduct of the players is not pre-determined. We will first focus on party behavior (strategies and tactics), then on the candidates' moves and choices given the structure of opportunity and constraint, and finally on voters and some characteristics of mass political behavior.

DYNAMIC ASPECTS OF MALTESE ELECTIONS

Party behavior

Notwithstanding the personalized nature of voting, and notwithstanding the intense competition among individual candidates, election campaigns have always been very much a party affair in Malta. This is hardly surprising, given the fact that the outcome of the election directly determines which party will form and run the Government.

Each party presents an electoral manifesto and runs a campaign focused on issues and leadership. The campaigns tend to stress the differences in political aims and dwell on the past performance of the parties. The respective party leaders -- their past, their policies, their personalities -- are always a prominent feature of electoral campaigns. Mass media coverage, and the open-air rallies which are held in various places over the country prior to an election, give particular prominence to each party's leader.

Both major parties have their own newspapers and their own television and radio station. Consequently, much of what the public encounters in the political reporting and analysis in the news media has a decidedly partisan slant. Negative campaigning and strategic release of incriminating material or allegations about politicians belonging to other parties are standard weapons in the campaigns. The characteristics of modern Maltese campaigns do not differ fundamentally from those seen elsewhere although the small size of the country arguably facilitates traditional personal contact campaigning, such as door-to-door canvassing by the candidates and street parties.

The parties control candidates through endorsement, but there appear to be few internal constraints. Party organizations actively seek out potential candidates, especially professionals such as physicians and lawyers who are likely to have their own local following. The major parties maintain mechanisms for formal approval of candidates, enabling them to exclude persons who are unacceptable on personal or political grounds. It appears to be rather rare that a party will reject a person who shows an interest in becoming one of its candidates.

The two dominant parties routinely over-nominate. Both the PN and the MLP have offered five or more candidates in five-member districts in about 95% of the cases; and there are frequently ten or more candidates from the major parties in a district. It is rare for a party to win four of the five seats; the most common result is a 3:2 split of the seats; and none of the major parties has ever accomplished a sweep of five seats. Figure 3 shows that both PN and MLP have consistently fielded between two and four candidates for each seat that they have actually won over the past thirty years. Because there is a long-standing pattern of most transfers benefiting other candidates of the same party, there is little risk in offering substantially more candidates than a party could possibly hope to elect in a district.

In fact, a greater range and diversity of candidates may benefit a large party because it puts it in a position to attract voters through the candidates' own personal networks of family and friends. It also allows them to cater to any conceivable taste or predilection among its potential constituency as long as a suitable candidate can be found.

A number of legal rules constrain the campaign behavior of political parties and candidates. Although there are statutory limits on expenditures by candidates, the political parties are not required to disclose the sources of their funds nor their expenditures. A host of other legal restrictions are imposed to safeguard the electoral process: No rallies or other public meetings may be held on election day and the day preceding it; even political advertising and reporting in the press are limited during this period. No person may even 'loiter' in the proximity of the polling-place on election day. Any shop, club, or other public establishment within fifty meters is to be kept closed on the day fixed for voting; and a number of provisions prohibit intimidation, obstructions, and breaches of the peace.

After the General Elections parties also play a key role in decisions regarding the vacating of a seat by double-winners. The major discretionary component here is the question of which seat should be resigned, and which other candidates will thereby be favored as a replacement. In 1996 these decisions by party officials resulted in the election of two additional female candidates. (See "Parties decide on would-be MPs fate," The Malta Independent, November 3, 1996, p.6). Parties also play the decisive role in selecting the Speaker and suitable politicians for co-optation.

Candidate behavior

As already indicated a number of restrictions constrain the scope of permissible electioneering. Candidates are subject to spending caps and must report all election-related expenditures. However, the limit of 600 Maltese Liri (ca. $1,600) per candidate appears to be widely ignored and various campaign finance regulations remain unenforced. Footnote 3

Strategic and tactical choices faced by prospective candidates include whether to compete as independents or to seek party endorsement, the district in which to stand, whether to stand in only one or in two, and how to conduct their personal campaigns within the legal and political framework. Some of these choices, however, are highly restricted. Thus, an independent candidacy is a virtual invitation to electoral defeat; and party organizations have a voice in determining the district(s) in which a candidate will compete.

There are rather strict legal rules regarding influence peddling, promises, quid-pro-quos and inducements. Yet all candidates, would-be candidates and incumbents must continually make decisions about how to best promote themselves and cultivate clientelistic relationships and support networks. Many voters expect candidates to promise to be helpful in obtaining some of the numerous favors which a government in Malta can provide. More than a third of the work force is employed by the government directly or by government-controlled enterprises; there is a sizable supply of public housing and a host of licensing laws -- all of which provide tempting targets of political favoritism. One idealistic, and unsuccessful, MLP candidate in 1987 complained that "the vote has become a negotiable instrument - a bill of exchange - rather than something that is based on principles and the party program. If someone asks me to try to get them a job or a flat or whatever, I often get very irritated." Footnote 4 (Weekend Chronicle, February 21, 1987)

Are Maltese elections and election campaigns candidate or party-centered? This may be a false dichotomy, because they are both. Not only may one not come at the expense of the other, mobilization and campaigning by parties and their candidates may be mutually reinforcing. Although for individual candidates intra-party competition is the bigger challenge, they may in turn make their own party more competitive (relative to the other major party) in sparing no effort in search of supporters. This dynamic, and the premium it places on additional votes, may be responsible, in part, for near-universal voter participation.

Voter behavior

Participation

The Maltese electorate is characterized by high partisanship and extremely high levels of participation in parliamentary elections (Hirczy 1995). Turnout, measured as the percentage of registered voters casting valid votes, has remained stable in recent elections at near 95%, as Table 2 shows. Footnote 5> Figure 4 illustrates the difference in voter turnout between various districts over the years and the virtual disappearance of such differences since the 1960's.

TABLE 2: REGISTERED VOTERS, BALLOTS CAST AND VALID VOTES, 1947 - 1998

+-----+--------------------+------------+------------------------+-----------+
|     |   Reg.    Ballots  |Ballots Cast|  Valid  Invalid   %    |Valid Votes|
|Years|  Voters    Cast    | As  % of   |  Votes  Votes  Invalid | As  % of  |
|     |                    |Reg.Voters  |                        |Reg.Voters |
+-----+--------------------+------------+------------------------+-----------+
|1947 |  140,703   106,141 |     75.44  |  105,494    647   0.61 |     74.98 |
|1950 |  144,516   106,820 |     73.92  |  106,129    691   0.65 |     73.44 |
|1951 |  151,979   113,366 |     74.59  |  112,625    741   0.65 |     74.11 |
|1953 |  148,478   119,333 |     80.37  |  118,453    880   0.74 |     79.78 |
|1955 |  149,380   121,243 |     81.16  |  120,655    588   0.48 |     80.77 |
|1962 |  166,936   151,533 |     90.77  |  150,606    927   0.61 |     90.22 |
|1966 |  161,490   144,873 |     89.71  |  143,347  1,526   1.05 |     88.77 |
|1971 |  181,768   168,913 |     92.93  |  168,059    854   0.51 |     92.46 |
|1976 |  217,724   206,605 |     94.89  |  205,440  1,165   0.56 |     94.36 |
|1981 |  238,239   225,466 |     94.64  |  224,151  1,315   0.58 |     94.09 |
|1987 |  246,292   236,719 |     96.11  |  235,168  1,551   0.66 |     95.48 |
|1992 |  259,310   249,141 |     96.08  |  247,139  2,002   0.80 |     95.31 |
|1996 |  274,113   264,037 |     96.32  |  261,224  2,813   1.07 |     95.30 |
|1998 |  281,078   268,159 |     95.40  |  264,492  3,667   1.37 |     94.10 |
+-----+--------------------+------------+------------------------+-----------+

The percentage of invalid ballots is very low (see Table 2), which is noteworthy for two reasons: (1) because the voting system is relatively complex, and (2) because the rate of spoiled ballots is much higher in high-turnout countries where voting is mandatory. High turnout in Malta appears to reflect genuine high motivation of the electorate to bring their preferences to bear on election outcomes rather than forced participation brought about by external pressure, which appears to prompt vote spoiling as a form of protest in other nations.

Party and candidate choice

It is important to again emphasize that, at the mass level of electoral behavior, the first-preference vote serves a dual purpose, namely to select a government in addition to allowing the voter to express a preference for the best-liked candidate. The first choice is a vote to affect the complexion of government, both the legislative and the executive branch. The voters know that their verdict is decisive, at least as long as third parties have little chance of breaking the party duopoly.

While the contention that the Maltese electorate is highly partisan is not based on systematic empirical survey data, there is little doubt about it. One does not have to rely on qualitative evidence only (though such evidence is very persuasive indeed). Voters' recorded ranking patterns speak clearly: Although their voting system allows for cross-party transfers, most voters do not exercise that option. Instead, they stop ranking candidates when the supply of their party's nominees is exhausted, making their vote non-transferable and thus a vote of no value, rather than have their vote benefit a candidate of another party.

Vote transfer patterns

A statistical analysis of what Michael Gallagher (1978) has termed "party solidarity" in vote transfer patterns shows that the percentage of cross-party transfers is about 1% on ballots where the preference in the preceding round was for a MLP or PN candidate. Footnote 6 The number of such defections were larger in earlier years when there were various minor parties with close affinities to either the MLP or PN. The defections reached their current minuscule proportion since the arrival of two-party competition in 1976. (See Fig. 5.) Where no candidate of the same party is still eligible to receive transfers, more than 99% of the ballots become non-transferable (this is a measure of "party exclusivity"). This pattern does not hold, however, for the small number of ballots on which the last-examined preference was for a minor party like the Alternattiva Demokratika; this may in part be because they typically field only two candidates per district.

The empirical evidence thus leads to the conclusion that the supporters of the two major parties are strong partisans, while those willing to express a preference for alternative candidates are not. Ostensibly, the supporters of MLP and PN would rather let their ballot go naught, than support a competitor.


ASSESSING MALTA'S ELECTORAL ARRANGEMENTS

As many election scholars have pointed out, assessing the performance of different electoral systems is problematic because there are multiple evaluative considerations and criteria, all of which rest on political and/or normative premises. Moreover, many of the values electoral systems are expected to further will not only compete with, but are in direct opposition to each other. A further caveat concerns the implications of such assessments. Any prescriptions flowing from them have to be reconciled with the political reality of feasibility constraints. Many proposed electoral reforms are entirely unrealistic because they would require the support of those who benefit from the status quo. Footnote 7

Without attempting to resolve these problems, we will address the major dimensions of electoral system evaluation in the Maltese case, marshaling the relevant empirical evidence and historical record which properly forms the basis for any normatively-informed debate.

Ballot access and domain of choice

Although Malta has a pure two-party system at the institutional level, and is unique in this respect in all of Europe, the electoral system cannot be faulted for restricting the number of options. In fact, the opposite is true. The threshold requirements for becoming a candidate are very low. Time and again minor parties and independent candidates do indeed appear on the ballot and thus diversify the menu of choice. In many elections and districts, voters have between one and two dozen candidates to choose from, and may rank all or some of them in one of hundreds, if not thousands, of different orders of preference. The typical outcome of the contests, namely the election of PN and MLP candidates only, thus do not reflect a lack of opportunity or alternatives.

Alternation in government

Malta's electoral system has worked reasonably well in assuring change in government as demonstrated by the record of election outcomes (see Figure 6). This is not tantamount to saying, however, that the system has always allowed popular majorities to effect a change in government. On several occasions, a majority of first-preference votes did not translate into a majority of elected MPs. The adoption of constitutional amendments in 1987 and 1996 were to correct distortions of parties' electoral strength through additional seats in particular circumstances. (Table 3) One consequence of these amendments is that when bonus seats are assigned, the benefiting party will be given only a one-seat majority, a feature that contributed to the premature end of the MLP government in 1998, when a single member of its parliamentary contingent voted against the government.


Table 3: Government and Opposition, 1947 - 1998

+-------------------------------------------------------------------------+
|     Parliamentary Seats| Parliamentary Seats | The Governing   |Popular |
|            of          |        of           |    Parties'     |        |
|     Governing Parties: | Opposition Parties: | Percentage Share| Vote   |
+------------------------|                     |    of      of   |        |
|Year |  MLP   PN   MWP  |  MLP   PN  MWP Other|  Seats    Votes |Majority|
+-------------------------------------------------------------------------+
|1947 |   24             |         7         9 |   60.0     59.9 |        |
|1950 |          12      |   11        11    6 |   30.0     29.6 |   No   |
|1951 |          15    7 |   14              4 |   55.0     54.3 |        |
|1953 |          18    3 |   19                |   52.5     49.9 |   No   |
|1955 |   23             |        17           |   57.5     56.7 |        |
|1962 |          25      |   16              9 |   50.0     42.0 |   No   |
|1966 |          28      |   22                |   56.0     47.9 |   No   |
|1971 |   28             |        27           |   50.9     50.9 |        |
|1976 |   34             |        31           |   52.3     51.5 |        |
|1981 |   34             |        31           |   52.3     49.1 |   No   |
|1987 |          35 *    |   34                |   50.7     50.9 | Bonus *|
|1992 |          34      |   31                |   52.3     51.8 |        |
|1996 |   35 *           |        34           |   50.7     50.7 | Bonus *|
|1998 |          35      |   30                |   53.8     51.8 |        |
+-----+------------------+---------------------+-----------------+--------+

Note: An asterisk (*) indicates that without the award of four bonus seats to the winning party, the majority of parliamentary seats would have been won by the party that obtained only a minority of the first-preference votes.

Party pluralism

There have been brief periods during which some parties in addition to the PN and MLP competed in elections with some success. One was the immediate post-war period, from 1947 to 1953, when a new mass electorate was created, the PN was in disarray and the MLP was split; the other was the 1962 election which was conducted under the shadow of a major conflict between the MLP and the Catholic Church. All of these minor parties went into quick decline and were short-lived.


Table 4: Parties' Share of First-Preference Votes, 1947 - 1998

+----------------------------------------------------------------------------+
|     |MLP  |PN   |MWP  |DAP  |CON  |PCP |CWP |DNP |GOZO|JON |AD  |Oth.|Ind. |
|_____|_____|_____|_____|_____|_ ___|____|____|____|____|____|____|____|_____|
|1947 |59.9 |18.1 |     |13.3 |     |    |    |    |5.2 |3.5 |    |    |0.14 |
|1950 |28.6 |29.6 |23.2 | 6.0 |10.0 |    |    |    |    |0.8 |    |    |1.84 |
|1951 |35.7 |35.5 |18.8 |     | 8.1 |    |    |    |    |0.9 |    |    |1.07 |
|1953 |44.6 |38.1 |11.8 |     | 1.2 |4.3 |    |    |    |    |    |    |     |
|1955 |56.7 |40.2 |     |     |     |3.0 |    |    |    |    |    |    |0.04 |
|1962 |33.9 |42.0 |     |     |     |4.8 |9.5 |9.3 |    |    |    |0.5 |0.08 |
|1966 |43.1 |47.9 |     |     |     |1.4 |6.1 |1.3 |    |    |    |    |0.27 |
|1971 |50.8 |48.1 |     |     |     |1.0 |    |    |    |    |    |    |0.06 |
|1976 |51.5 |48.5 |     |     |     |    |    |    |    |    |    |    |0.02 |
|1981 |49.1 |50.9 |     |     |     |    |    |    |    |    |    |    |0.01 |
|1987 |48.9 |50.9 |     |     |     |    |    |    |    |    |    |0.2 |0.01 |
|1992 |46.5 |51.8 |     |     |     |    |    |    |    |    |1.7 |    |0.04 |
|1996 |50.7 |47.8 |     |     |     |    |    |    |    |    |1.5 |    |0.02 |
|1998 |47.0 |51.8 |     |     |     |    |    |    |    |    |1.2 |    |0.01 |
+----------------------------------------------------------------------------+

MLP = Malta Labour Party; PN = Nationalist Party; MWP = Maltese Worker's Party; DAP = Democratic Action Party; CON = Constitutionalist Party; PCP = Progressive Constitutionalist Party; CWP = Christian Workers' Party; DNP = Democratic Action Party; GOZ = Gozo Party; JON = Jones Party; AD = Alternattiva Demokratika; Oth. = Other parties; Ind. = Independent candidates.

Since Independence the electoral record establishes the lack of popular support, rather than exclusion by institutional means, of alternative parties. Although a third party, Alternativa Demokratika (AD) fielded at least two candidates in each district in recent elections, and less than a fifth of the vote is necessary to win a seat, support for them has been negligible. The highest number of first-preference votes in any district was a mere 416: equal to 2.3% of the valid votes in the district, or 13.7% of its quota). None of the AD candidates accumulated enough vote transfers from other candidates to meet the quota. Independents have done even worse, accumulating in a typical election, like that of 1996, no more than of 43 first-preference votes in the entire country (0.016% of the total valid vote). Indeed, of the 745 legislative seats filled since 1947, only one has been filled by an independent and that was in 1950.8 Footnote 8

In light of the conventional wisdom regarding the effects of different electoral formulae on composition of parliament in terms of parties and on government formation and stability (Blais and Carty 1987), one might say that Malta's two-party system, and single-party government, have persisted despite PR, simply because the mass public prefers it that way. This may, of course, change in the future. Malta used to have a multi-party system and may yet have one again.

Proportionality

Proportionality presupposes the relevance of party as a basis for allocation of political offices and power. Since voters rank candidates, STV does not necessarily require political parties as a basis for allocating seats. Some have even suggested that it is inappropriate to speak of an aggregate-level party vote in the context of the STV system, which is the quintessential candidate-centered system. Yet nothing could be further from the political reality of elections in Malta. Politicians, voters and political analysts in Malta consistently examine and compare the parties' first-count vote totals when discussing the outcome of elections. As in the electoral literature generally, much of this discussion in Malta focuses on the relationship between the parties' respective vote and seat percentages, and particularly the first-count votes which a party's candidates have collectively amassed. This is not surprising, as the parliamentary election is first and foremost about winning control of parliament, and thus over Government, which in a parliamentary system is clearly a party affair. As a mechanism for choosing parliamentarians, STV essentially serves to determine which of a party's candidates shall enter Parliament.

One key test of this proposition was the election of 1981, which resulted in the MLP winning majority control of parliament even though the Nationalists had 'won' the election if all first-preference votes were to be added, and credited to the respective parties, at the national level. This was seen as so unjust and illegitimate by the Nationalists that they refused to accept the outcome, a quintessential requirement for a functioning democratic system. Footnote 9

There were several earlier instances (as in 1953 and 1966) of what Douglas Rae (1967, p.74) has termed a "manufactured majority," i.e., a party winning a majority of the seats and forming a government without having the mandate of a majority of first-count votes. Each of those cases had benefited the Nationalist Party. (See Table 3) However, in 1981 there was not only a legislative majority lacking a vote majority but there was a rival party which actually had obtained a majority of the votes yet was deprived of a legislative majority (an "artificial minority" in Rae's terminology). Footnote 10

The crisis was resolved before the 1987 election was called and the Nationalists' boycott of parliament ended, with a package of constitutional amendments, one of which would prevent the recurrence of such a reversal of winners and losers in future. This important episode in Malta's political history attests to the vital role of parties in Maltese politics and the entrenched nature of thinking in terms of party representation. STV, as it operates in Malta in conjunction with other features of the electoral system, is anything but a nonpartisan way to elect members to a representative body. Instead, it has more in common with closed primaries used by a number of states in the US (although the latter involve competition for a single slot).

Gerrymander or System Peculiarity?

When the Maltese Labor Party won control of government in 1981 with a minority of the aggregate votes, the Nationalists charged that the reversal of winners and losers resulted from the MLP-dominated Election Commission's gerrymandering of district lines. Footnote 11 It is difficult to ascertain the veracity of these charges. Taagepera and Shugart (1989, p. 237n) and Gallagher (1996, p. 3) point to low district magnitude rather than gerrymandering as a significant factor in the outcome. In any case, the constitutional amendments providing for adjustment by allocating additional seats constitute an effective institutional fix, albeit by different means, regardless of whether the vote-seat distortion was/is due to a quirk of the system or deliberate manipulation.

Given the nearly even electoral strength of the two major parties and the small district magnitudes, the occurrence of a discrepancy between vote and seat majorities cannot be ruled out even in the absence of gerrymandering. Indeed, the phenomenon occurred in three of the last four elections, in 1981, 1987 and 1996 and the new constitutional remedy needed to be applied in the last two cases. The manner in which the problem was solved is a good illustration of how perceived or real shortcoming of a particular set of electoral arrangements can be remedied through creative adaptations.

Election administration

One of the arguments against STV is the complexity of the counting procedures. Indeed, the determination of winners took some 47 hours in the 1992 elections. Amendments to the Election Act, however, have since streamlined the process. In 1996 the first-preference vote totals of the parties were publicly known in less than twenty-four hours, thanks to the assiduous work of party agents monitoring the work carried on in the central ballot counting facility. Casual elections, of course, require additional time, but do not affect party control of parliament.

Complexity of the ballot, or the requirement for rank-ordering of candidates by writing in sequential numbers, does not appear to be a problem as is seen in the low proportion of ballot papers ruled invalid. This was the case even in previous decades when there were many illiterate voters and often exceedingly crowded ballot papers.

Safe-guarding the integrity of the process

The role of administering the elections is given to the Electoral Commission and the Election Office it superintends. Striking is not only the large body of meticulous legal requirements to assure transparency and honesty of the election process (including strict accounting for, and double-checking of voters and ballots, cast and unused, and record-keeping), but the assignment of formal watchdog roles to the political parties, or their designated representatives. Their presence in all forums and components of the election process is pervasive. Essentially their role is to check each other and with competitive eagerness they do so with considerable efficiency.

Political parties designate two party agents each as members of the Electoral Commission. They also nominate doctors for the Medical Board charged with deciding voters' competency for purposes of exercising the franchise; party agents are given access to electoral registers and all other records of the Electoral Office; they are authorized to observe ballot printing, packaging, distribution, and sealing of ballot boxes; parties may even affix their own seals in addition to the official one. Parties designate Assistant Electoral Commissioners in all polling places and are involved in keeping the chain of custody of all official documents and of ballot boxes. Party agents are given access to the counting hall and given facilities to observe and assure the integrity of the vote counting process.

Election system reform

No electoral system meets the demands and self-defined needs of all players and constituencies in a political system. Malta is no exception and has its unique brand of electoral reform debate. As is often the case, considerations of political self-interest masquerade as questions of high principle. Although a major problem in Malta's version of STV was addressed in 1987 and 1996 by amending Article 52 of the Constitution, these were reforms that were generally perceived as a stop-gap measure. While it assured the largest party of a parliamentary majority, it could (and did) give parliamentary seats to candidates who failed to obtain the quota in their district and, by virtue of the bonus seats, give some districts six (and possibly more) instead of five representatives. Moreover, the new rules did not provide any assurance on the continuously debated issue of greater proportionality between the vote and seat percentages of the parties. And thus the issue of electoral reform stayed on the political agenda.

The reform proposals of the past few years were vigorously promoted by the Nationalist government. Their proclaimed dual aim was to assure both proportionality and "governability" (the latter meaning one-party control of government) without acknowledgment that both goals are not necessarily compatible. The proposed remedies involved a variety of procedural approaches, including the use of the d'Hondt formula on a nationwide basis. But no agreement could be reached between the two major parties because, while they seemingly agreed that "governability" was more important than proportionality, they differed on such issues as the disposition of votes cast for losing minor party candidates. Electoral reform efforts are likely to be held in abeyance until, perhaps, a new third-party challenge to the current duopoly needs to be addressed.


CONCLUSION

What place does STV play in the family of electoral systems? And one might ask, concluding a single-country study, what place does Malta take within the family of nations (and subnational units) using STV? What lessons can be learned?

The first point to make is that STV in Malta is not merely another instance of a particular type of voting system situated within a context different from those found in other jurisdictions that have it. Much rather, Malta's use of STV represents a particular implementation of STV among many different possible variants, and must be considered as part and parcel of electoral arrangements overall. Malta's use of STV must be viewed in conjunction with other rules governing elections and electoral institutions, most importantly - in Malta - the adjustment mechanism incorporated into the constitution, which assures the party winning the popular first-preference vote control of parliament and thus of Government. Also relevant are the rules governing multiple candidacies, filling of vacated seats, and co-optation. In short, we should conceptualize STV not as an invariant type of voting system to be contrasted with all others, or even PR systems in particular, but as a variable encompassing many different forms and uses of STV.

In assessing the political consequences or implications of a voting system we must, of course, also take account of the unique aspects of the political and societal context within which it operates. The most salient particularity of the Maltese case is that STV co-exists with a two-party system. It is not clear how the electoral system could be responsible for the fact that only two parties are found in the Maltese parliament. What can be stated unequivocally, however, is that the two-party system (i.e. the two major parties) has allowed STV to remain in force, so that it can be said to represent a democratic political choice the people of Malta have made through their Parliament.

We should be wary of deriving general conclusions from the study of a single jurisdiction all too readily. At the minimum, however, we can conclude from looking at the Maltese case that STV can co-exist with a two-party system, i.e. that STV does not necessarily lead to multipartism and the attendant problems of fragmentation and government instability. Still, there is no reason to believe that the Maltese version of STV would stand in the way of electoral success by a new or third party, which might entail such problems.

How well has STV worked for the Maltese? An unqualified answer one way or another is not possible. STV can be said to have failed when the mechanics of the system resulted in the losing party (in first-preference votes in the nation as a whole) gaining control of the government. While such 'perverse' results are not unheard of in other countries, the perceived injustice of this outcome in 1981 was severe enough to lead to a political crisis in Malta and indeed a breakdown of the system, as seen in the Nationalists' boycott of Parliament. The deadlock was broken and the problem fixed by amending the constitution, although criticism about the system, and about further possibilities of 'perverse outcomes,' has not entirely abated.

The lesson to learn here is that under certain conditions, namely when there are only two viable parties that enjoy near-equal support in the electorate, proportionality can be a problem even where the deviation from perfect correspondence of vote and seat shares is very minor. Indeed it is ironic that a special device to correct for the electoral system's distorting effects should have been necessary in the country with the most proportional STV system of all (Farrell, Mackerras and McAllister 1996).

Appropriate caveats about the limits of inferences based on Malta's experience with its version of STV having been made, we can proceed by putting this particular case of STV in a comparative perspective, identifying commonalities and differences.

Unlike Ireland, both Australia and Malta have strong party systems, but Malta is unique in the bi-polar nature of its politics. The two-party system exists not because Maltese voters do not encounter other choices on the ballot, but ostensibly because they want it that way, as seen in the extremely low incidence of cross-party voting and the failure of new political parties to attract a substantial number of voters.

While STV can be hailed as a system that allows preference-ranking irrespective of partisan alignments, ticket-splitting is statistically rare and unimportant in Malta. Even if cross-party transfers were to become more frequent due to a decline in the electorate's partisanship, it would not per se affect the balance of power between the parties, which is determined by first preferences only. A party can no longer 'lose' its majority due to transfer of surplus votes to candidates of other parties.

The most striking contrast with Ireland (beyond the centrality of partisanship in elections) is the absence of any coalition dynamics that intervene between the expressed preferences of the voters and government formation. In Malta, the outcome of the popular vote alone determines which party will take the helm of Government.

Cross-national comparisons of STV around the world are fruitful and illuminating, but they must not obscure a basic limitation. Not only is the number of cases on which to base general statements about this type of electoral system woefully inadequate; even among the small group of countries with STV, the institutional specifics are rather diverse. At the same time, however, the few actual instances of this particular type of voting system are only a small number of many possible variants of STV that one can imagine, most of which have never been tried. While many pieces of scholarly work with inconclusive results end by stating the case for more research, we face a rather unique problem in studying STV: in order to examine what difference it makes in its different shapes and forms we need more institutional experimentation by policymakers. Rather than creating a research opportunity, the scarcity of accumulated experience, and of empirical studies based on it, is a major constraint.




Footnotes

1 There were eight districts from 1921 until 1955; ten for the elections from 1962 and 1971; and 13 since 1976. To elect a ten-member Council of Government in 1939 and 1945 only two districts were created. Five-member districts have been the norm from the beginning; only in 1971, as an exceptional and temporary measure, five of the ten districts were given six members each. 2 The multiplying factor is obtained by dividing the surplus by the number of transferable papers (which excludes the non-transferable ones).

3 After the 1996 election all of the Alternattiva Demokratika candidates (none of whom was likely to have approached the spending limit) announced that they would not file the legally required affirmation of compliance with the law, on the ground that such affirmations by candidates had become a farce.

4 The candidate, Myriam Spiteri Debono, was unsuccessful in the two subsequent elections but was elected Speaker of the House of Representatives in 1996.

5> The voter turnout figures can be affected by the length of time elapsed between the semi-annual publication of the voter register and the date of the election. In 1996, almost six months divided the two dates and the Electoral Commission offered a novel and unofficial category of "potential voters" to account, in part, for registered voters who had died in the interim period. Many newspapers reports used this category to arrive at a voter turnout percentage that was correspondingly higher.

6 The terms "party solidarity" and "party exclusivity" were introduced by Michael Gallagher in "Party Solidarity, Exclusivity and Inter-Party Relationships in Ireland, 1922 - 1977: The Evidence of Transfers," 10 Economic and Social Review (1978), pp. 1-22.

7 In Malta there are at present no interest groups promoting electoral reform, nor are there any institutional mechanisms (such as citizen initatives or referenda) to bypass the one-party majorities in the legislature.

8 Actually, this was the only independent who has been elected since 1947. Two independents were elected in the pre-war period when the restricted franchise created a very small electorate.

9 The Nationalists repeatedly called for new elections and did not take their seats in parliament for over a year. Immediately after the election, Prime Minister Mintoff had wanted to resign and call new elections but yielded to the unanimous opposition of his cabinet, as reported by an insider, the then-Attorney-General Edgar Mizzi (Mizzi, 1995, p.436).

10 Rae added that "it would be intolerable if an electoral law robbed leading parties of their majority, and no such electoral system exists." (p. 75) Actually, one example of an "artificial majority" preceding the Maltese experience was the Australian House of Representatives election in 1954.

11 Interestingly, there were no allegations of gerrymandering when the new election district boundaries were presented and debated in parliament. These accusations, with many plausible inferences about gerrymandering, occurred only after the election results were known.



Bibliography

Note: We have included in this bibliography a number of items on STV and Maltese politics although they are not cited in the text of the paper.

Blais, Andre and R. K. Carty. 1987. "The Impact of Electoral Formulae on the Creation of Majority Governments." Electoral Studies 6:209-218.

Bowler, Shaun, and David M. Farrell. 1991. "Party Loyalties in Complex Settings: STV and Party Identification." Political Studies 39:350-362.

Brams and Fishburn. 1984. "Some Logical Defects of the Single Transferable Vote." In: Arend Lijphart and Bernard Grofman, editors. Choosing an Electoral System: Issues and Alternatives. New York: Praeger. 147-151.

Brodie, Edith. 1967. Malta's Road to Independence. Norman.

Craig, James. 1982. "Malta: Mintoff's Election Victory." West European Politics 5:318-320.

Crick, Bernard. 1995. "The Political Case for Electoral Reform: A Quiet and Selfish Word to my Party." Representation. 33:34-40.

Darcy, R. and Michael Marsh. 1994. "Decision Heuristics: Ticket-Splitting and the Irish Voter." Electoral Studies 13(1):38-49.

Doron, Gideon and Richard Kronick. 1977. "Single Transferable Vote: An Example of a Perverse Social Choice Function." American Journal of Political Science 21:303-312

Engstrom, Richard L. 1990. "Cincinnati's 1988 Proportional Representation Initiative." Electoral Studies 9(3):217-225.

Fenech Doninic. 1993. "Malta." European Journal of Political Research 24:459-500.

----- 1992a. "The 1992 Maltese Election" West European Politics 15(4):189-195.

----- 1992b. "Malta." European Journal of Political Research 22:471-473.

----- 1989. "More or Less Democracy" Society (1989)(4):27-30.

----- 1988. "The 1987 Maltese Election: Between Europe and the Mediterranean." West European Politics 7: 133-138.

Farrell, David M., Malcolm Mackerras and Ian McAllister. 1996. "Designing Electoral Institutions: STV Systems and their Consequences." Political Studies 44:24-43.

Frendo, Henry. 1991. Party Politics in a Fortress Colony: The Maltese Experience. Midsea Publications [2nd ed.]

----- 1989. Malta's Quest for Independence. Valletta: Valletta Publishing Co.

Gallagher, Michael, Michael Laver, and Peter Mair. 1992. Representative Government in Western Europe. New York: McGraw-Hill.

Gallagher, Michael. 1986. ''The Political Consequences of the Electoral System in the Republic of Ireland." Electoral Studies 5(31:253-275.

----- 1978. "Party Solidarity, Exclusivity and Inter-Party Relationships in Ireland, 1922 - 1977: The Evidence of Transfers." Economic and Social Review. (10)1-22.

Hallett, George H. 1984. "Proportional Representation with the Single Transferable Vote; A Basic Requirement for Legislative Elections. " In: Arend Lijphart and Bemard Grofman, editors. Choosing an Electoral System: Issues and Alternatives. New York: Praeger. 1984. 113-126.

Hirczy, Wolfgang, 1995a. "Explaining Near-Universal Turnout: The Case of Malta." European Journal of Political Research 27(2):255-272.

----- 1995b. "STV and the Representation of Women." PS: Political Science and Politics 28 (4) (December):71 1-13.

----- 1994 The Impact of Mandatory Voting Laws on Turnout: A Quasi-Experimental Approach." Electoral Studies 13:64-67.

Howe, Stephen. 1987. "The Maltese General Election of: 1987." Electoral Studies 6:235-247.

Katz, Richard S.. 1986. "Intraparty Preference Voting." In: Bernhard Grofman and Arend Lijphart, eds.. Electoral Laws and Their Political Consequences. New York: Agathon. 1986.

----- 1984. "The Single Transferable Vote and Proportional Representation." In: Arend Lijphart and Bernard Grofman, editors. Choosing an Electoral System: Issues and Alternatives. New York: Praeger. 1984. 133-126.

Lakeman, Enid. 1974. How Democracies Vote. London (4th edition).

Lane, John. 1995. ",The Election of Women Under Proportional Representation: The Case of Malta." Democratisation. 2:140-157.

Lijphart, Arend. 1994. Electoral Systems and Party Systems: A Study of Twenty-Seven Democracies, 1945 - 1990. Oxford.

Mair, Peter. 1986. "Districting Choices under the Single-Transferable Vote." In: Bernhard Grofman and Arend Lijphart, eds. Electoral Laws and Their Political Consequences. New York: Agathon. 1986.

Mizzi, Edgar. 1995. Malta in the Making, 1962 - 1987: An Eyewitness Account. Malta.

Monroe, Burt L. 1994. "Understanding Electoral Systems: Beyond Plurality Versus PR." PS 27(4) (December 1994) 677-682.

Procter, L. H.. 1980. "The Acceptance of Proportional Representation in Malta," Parliamentary Affairs 33 :308-320

Rae, Douglas. 1967. The Political Consequences of Electoral Laws. New Haven.

Tideman, Nicolaus. 1995. "The Single Transferable Vote." Journal of Economic Perspectives 9:27-38

Weaver, Leon. 1984. "Semi-Proportional and Proportional Representation Systems in the United States." In: Arend Lijphart and Bernard Grofman, editors. Choosing an Electoral System: Issues and Alternatives. New York: Praeger. 1984, 191-206.

Wright, Jack F. H.. 1986. "Australian Experience with Major Preferential and Quota-Preferential Systems.'' In: Bernard Grofman and Arend Lijphart, editors. Electoral Laws and Their Political Consequences. New York: Agathon. 1986, 124-138.

Zanelli, Remo. 1990. "The Maltese Electoral System and its Distorting Effects." Electoral Studies 9:205-215

Zimmerman, Joseph F.. 1994. "Alternative Voting System for Representative Democracy." PS 27:674-677