SECOND SECTION

CASE OF YUMAK AND SADAK v. TURKEY

( Application no. 10226/03 )

JUDGMENT

STRASBOURG

30 January 2007

THIS CASE WAS REFERRED TO THE GRAND CHAMBER

WHICH DELIVERED JUDGMENT IN THE CASE ON

8 JULY 2008

This judgment will become final in the circumstances set out in Article 44 § 2 of the Convention. It may be subject to editorial revision.

In the case of Yumak and Sadak v. Turkey ,

The European Court of Human Rights (Second Section) , sitting as a Chamber composed of:

Mr J.-P. Costa , President , Mr I. Cabral Barreto , Mr R. Türmen , Mr M. Ugrekhelidze , Mrs A. Mularoni , Mrs E. Fura-Sandström, Mr D. Popović , judges , and Mrs S. Dollé , Section Registrar ,

Having deliberated in private on 9 May 2006 and 4 January 2007 ,

Delivers the following judgment, which was adopted on the last ‑ mentioned date:

PROCEDURE

1 . The case originated in an application (no. 10226/03) against the Republic of Turkey lodged with the Court under Article 34 of the Convention for the Protection of Human Rights and Fundamental Freedoms (“the Convention”) by two Turkish nationals, Mr Mehmet Yumak and Mr Resul Sadak (“the applicants”), on 1 March 2003 .

2 . The applicants were granted legal aid .

3 . The y alleged that the national electoral threshold of 10% for parliamentary elections interfered with the free expression of the opinion of the people in the choice of the legislature. They relied on Article 3 of Protocol No. 1.

4 . By a decision of 9 May 2006 the Chamber declared the application partly admissible.

5 . The applicants and the Government each filed further written observations (Rule 59 § 1).

6 . A hearing took place in public in the Human Rights Building , Strasbourg , on 5 September 2006 (Rule 59 § 3).

There appeared before the Court:

(a) for the Government Mr A.M. Özmen , Co- Agent , M r M . H. Ünler , M rs V. Sirmen , M rs Y. Renda , Mrs A. Özdemir, Mrs Ü. Yeğengil ,] Adviser s ;

(b) for the applicants Mr T. Elçi , Counsel , Mrs S. Turan , Adviser .

The Court heard addresses by Mr Özmen and Mr Elçi.

THE FACTS

I. THE CIRCUMSTANCES OF THE CASE

7 . The applicants were born in 1962 and 1959 respectively and live in Şırnak . They stood for election in the parliamentary elections of 3 November 2002 as candidates of the People ’ s Democratic Party (DEHAP) in the province of Şırnak , but neither of them was elected.

A. The parliamentary elections of 3 November 2002

8 . Following the 1999 earthquakes Turkey went through two serious economic crises in November 2000 and February 2001. There then followed a political crisis, due firstly to the state of health of the then Prime Minister and secondly to the numerous internal divisions within the governing coalition , a grouping of three political parties.

9 . It was in that context that on 31 July 2002 the Grand National Assembly of Turkey (“the National Assembly”) decided to bring forward the date of the next parliamentary elections to 3 November 2002.

10 . In early September three left-wing political parties, HADEP, EMEP and SDP, decided to form a “Labour, Peace and Democracy Block” and to form a new political party, DEHAP. The applicants began their electoral campaign as the new party ’ s leading candidates in the province of Şırnak .

11 . The results of the elections of 3 November 2002 in the province of Şırnak gave the DEHAP list 47,449 of the 103,111 votes cast, a score of about 45.95%. However, as the party had not succeeded in passing the national threshold of 10%, the applicants were not elected. The three seats allocated to Şırnak province were shared as follows: two seats for the AKP ( Adalet ve Kalkınma – the Justice and Development Party, a party of the conservative right), which had polled 14.05% (14,460 votes), and one seat for Mr Tatar, an independent candid ate who had polled 9.69% (9,914 votes).

12 . Of the eighteen parties which had taken part in the elections only the AKP and the CHP ( Cumhuriyet Halk Partisi – the People ’ s Republican Party, a left-wing party) succeeded in passing the 10% threshold. With 34.26% of the votes cast, the AKP won 363 seats, 66% of those in the National Assembly. The CHP, which polled 19.4%, obtained 178 seats, or 33% of the total. Nine independent candidates were also elected.

13 . The results of these elections were generally interpreted as a huge political upheaval. Not only did the proportion of the electorate not represented in parliament reach a record level in Turkey (approximately 45%) but in addition the abstention rate (22% of registered voters) exceeded 20% for the first time since 1980. As a result, the National Assembly which emerged from the elections was the least representative since 1946, the year in which a multi - party system was first introduced. Moreover, for the first time since 1954, only two parties were represented in parliament.

14 . To explain the National Assembly ’ s unrepresentativity , some commentators [1] have referred to the cumulative effect of a number of factors over and above the existence of a high national threshold. For example, because of the protest vote phenomenon linked to the economic and political crisis, the five parties which had obtained seats in the 1999 parliamentary elections were unable to reach the 10% threshold in 2002 and were accordingly deprived of representation in parliament. Similarly, electoral fragmentation had an effect on the results in that numerous attempts to form pre-electoral coalitions had come to nothing.

B. The general context and the electoral system

15 . The electoral system is one of the subjects which have been the most debated in Turkey ; it still remains highly controversial.

16 . The elections of 1950, 1954 and 1957 – in which the majority representation system was used – were unable to ensure an institutional balance between the majority in parliament and the opposition. This imbalance was one of the main reasons for the 1960 coup d ’ état. Following the intervention of the armed forces parliament adopted proportional representation, using the D ’ Hon d t method, to strengthen pluralism and the political system. As a result, the elections in 1965 and 1969 produced stable majorities in the National Assembly while enabling small parties to be represented. However, in the elections of 1973 and 1977 the main political movements were unable to establish stable governments, although they had wide electoral support. That period of government instability was marked by the formation of one coalition after another, each made fragile by the disproportionate influence of the small parties on government policy.

17 . Following the military regime of the years 1980 to 1983 Law no. 2839 on the election of members of the National Assembly, enacted on 13 June 1983 , re-established proportional representation, with two electoral thresholds. To the 10% national threshold was added a provincial threshold (the number of electors divided by the number of seats to be filled in each constituency); in 1995 the Constitutional Court declared the provincial threshold null and void. In the 1983 parliamentary elections the Motherland Party (ANAP) obtained an absolute majority in parliament.

18 . The parliamentary elections of 29 November 1987 likewise enabled the ANAP, with 36.31% of the vote, to form a stable parliamentary majority. Two other parties also won seats. About 19% of votes were cast in favour of parties which ultimately failed to reach the 10% threshold. In the elections of 20 October 1991 five parties gained seats in parliament. This result was due in particular to the fact that three small political parties (M Ç P, IDP and HEP) had taken part in the elections under the banner of other political parties with the aim of circumventing section 16 of Law no. 2839, which makes it illegal to form joint lists before elections. The proportion of the votes cast in favour of parties not represented in the new parliament thus fell to 0.5%. The Government was based on a coalition of two parties. In those elections the eighteen candidates of the HEP (People ’ s Labour Party – pro-Kurdish) were elected to parliament on the list of the (social-democratic) SHP party; they later resigned from the SHP to join the ranks of their own party, the HEP.

19 . In the general election of 24 December 1995 five parties gained seats in parliament. However, as none of them had a parliamentary majority, a coalition was formed. The proportion of the votes cast in favour of parties not represented in parliament came to 14%.

20 . The 1999 parliamentary elections again resulted in no party having a parliamentary majority. Five political parties won seats in the National Assembly. A coalition of three parties formed a government. The proportion of the votes cast in favour of parties not represented in parliament came to 18%.

21 . At present, numerous proposals to correct the effects of the 10% threshold have been put forward , both in parliament and by leading figures of civil society.

II. RELEVANT DOMESTIC AND INTERNATIONAL LAW AND PRACTICE

A. Relevant domestic texts

1. The Constitution

“Electoral laws must strike a balance between fair representation and governmental stability.”

23 . Article 80 of the Constitution provides:

“Members of the Grand National Assembly of Turkey shall represent the whole nation and not the regions or persons which have elected them.”

2. The electoral system

24 . Law no. 2839 on the election of members of the National Assembly, published in the Official Gazette on 13 June 1983, lays down the rules of the system for parliamentary elections.

25 . The Turkish National Assembly has 550 members, elected in 85 constituencies in a single round of voting. They take place throughout the national territory, on the same day, under the proportional representation system. The suffrage is free, equal, universal and secret. Counting the votes and recording the results is done in public. Each province forms one electoral constituency.

“ ... [P] olitical parties may not present joint lists ... ”

27 . Section 33 of Law no. 2839 (as amended on 23 May 1987) provides:

“In a general election parties may not win seats unless they obtain, nationally, more than 10% of the votes validly cast ... An independent candidate standing for election on the list of a political party may be elected only if the list of the party concerned obtains sufficient votes to take it over the 10% national threshold ... ”

28 . In allocating seats the D ’ Hondt system of proportional representation is used. That method – under which the votes cast for each list are first divided by a series of whole numbers (1, 2, 3, 4, 5 etc.) and seats then allocated to the lists which have the highest quotients – tends to favour the majority party .

3. Constitutional case-law

29 . In a judgment of 18 November 1995 (E. 1995/54, K. 1995/59) the Constitutional Court had the opportunity to rule on the constitutionality of section 34/A of Law no. 2839. That section, which referred to section 33 of the same law, also imposed the electoral threshold of 10% for the allocation of the seats for Assembly members elected in the “national constituency”.

30 . The Constitutional Court declared the provisions establishing the national constituency null and void, but held that the 10% national threshold could be regarded as compatible with Article 67 of the Constitution.

The relevant passages of the judgment read as follows:

“ ... [T]he Constitution defines the Turkish State as a Republic ... The constitutional structure of the State, which is based on national sovereignty, is a product of the nation ’ s will, mediated through free elections. That choice, emphasised in the various Articles of the Constitution, is set forth clearly and precisely in Article 67, entitled ‘ The right to vote, to be elected and to engage in political activities ’ . Paragraph 6 of Article 67, as amended, provides that electoral laws must be framed in such a way as to strike a balance between the principles of ‘ fair representation ’ and ‘ governmental stability ’ . The aim is to ensure that the electors ’ will is reflected as far as possible [ in ] the legislature. ... [In order to] choose the system whose methods are most conducive to the expression of the collective will and the taking of collective decisions in the legislature, ... enacting the appropriate legislation in the light of the country ’ s specific circumstances and the requirements of the Constitution, it is necessary to opt for [ the system] which is most compatible with the Constitution or to reject any system incompatible with it.

The impact of a representative democracy is visible in various fields. The effect of unfair systems adopted with the intention of ensuring stability is to hamper social developments. ... Where representation is concerned, the importance attached to fairness is the main condition for governmental stability. Fairness ensures stability. However, the idea of stability, in the absence of fairness, creates instability. The principle of ‘ fair representation ’ with which the Constitution requires [compliance] consists in free, equal, secret and universal [suffrage], with one round of voting and public access to the counting of votes and the recording of results, and produces a number of representatives proportional to the number of votes obtained. The principle of ‘ governmental stability ’ is perceived as a reference to methods designed to reflect votes [within] the legislature so as to guarantee the strength of the executive power. The ‘ governmental stability ’ which it is sought to ensure through the threshold ( described as a ‘ hurdle ’ ), just like ‘ fair representation ’ ... , is protected by the Constitution. In election s ... importance must be attached to combining these two principles, which seem antinomi c in certain situations, in such a way [as to ensure] that they counterbalance and complement each other ...

In order to achieve the goal of ‘ governmental stability ’ , set forth in the Constitution, a national [threshold] has been introduced ...

Clearly, the [threshold] of 10% of the votes cast nationally laid down in section 33 of Law no. 2839 ... came into force with the approval of the legislature. Electoral systems must be compatible with constitutional principles ... , and it is inevitable that some of these systems should contain strict rules. Thresholds which result from the nature of the systems and [are expressed] in percentages, and [which] at national level restrict the right to vote and to be elected, are applicable [and] acceptable ... provided that they do not exceed normal limits ... The [threshold] of 10% is compatible with the principles of governmental stability and fair representation ... ”

Three judges of the Constitutional Court disagreed with the arguments of the majority, considering that the 10% national threshold was incompatible with Article 67 of the Constitution.

31 . In the same judgment, however, the Constitutional Court declared null and void an electoral threshold of 25% for the allocation of seats within provinces (provincial threshold). Holding that such a threshold was inconsistent with the principle of fair representation, it observed:

“Although a national threshold is imposed in parliamentary elections in accordance with the principle of ‘ governmental stability ’ , imposing in addition a threshold for each electoral constituency is incompatible with the principle of ‘ fair representation ’ .”

B. Relevant Council of Europe documents

1. Report of the ad hoc Committee of the Parliamentary Assembly of the Council of Europe

32 . The Government referred to the report of the Ad hoc Committee for the Observation of Parliamentary Elections in Turkey (3 November 2002), produced on 20 December 2002. The relevant parts of the report read as follows:

“As widely reported by the media, two parties only out of 18 found their way into the new TBMM; the AKP (Justice and Development) and CHP (Republican People ’ s Party), leaving out all other parties, which had been represented so far in the parliament because they could not meet the 10% threshold. The party in government until the elections received only 1% of the votes. Economic and corruption problems were determining in the elections.

A clear and absolute majority has emerged with 362 seats for the AKP, 179 seats for the opposition and 9 seats for independent members. (These independent members are elected in small towns where they have a good reputation.) It should be recalled that AKP had 59 seats in the previous parliament, and the CHP three (1999 elections).

This situation might create probably greater stability in the country by avoiding complicated and unstable coalitions. On Monday 4 November 2002 the Turkish stock exchange went up by 6.1%.

However, it also means that approximately 44% of the voters have no representation in the Parliament.

The results must thus be considered as a clear protest vote against the establishment as a whole, since none of the three parties in the old governing coalition got enough votes for a single seat!”

2. The Code of good practice in electoral matters

33 . The Council of Europe has not issued any binding standards for electoral thresholds. The question has not been raised in the organisation ’ s standard-setting texts. On the other hand, the Code of good practice in electoral matters, adopted by the Venice Commission, makes recommendations on the subject (see Venice Commission, “Code of good practice in electoral matters: Guidelines and explanatory report”, Opinion no. 190/2002). As a general principle, the Code requires suffrage to be direct , but in the case of a bicameral parliament it permits one of the Chambers to be elected by indirect suffrage. As for the electoral system to be used, the Code ’ s guidelines state that any system may be chosen.

3. The Parliamentary Assembly ’ s Resolution 1380 (2004)

34 . Paragraphs 6 and 23 of Resolution 1380 (2004) on “Honouring of obligations and commitments by Turkey ”, adopted by the Parliamentary Assembly of the Council of Europe on 22 June 2004 , are worded as follows:

“6. With regard to pluralist democracy, the Assembly recognises that Turkey is a functioning democracy with a multiparty system, free elections and separation of powers. The frequency with which political parties are dissolved is nevertheless a real source of concern and the Assembly hopes that in future the constitutional changes of October 2001 and those introduced by the March 2002 legislation on political parties will limit the use of such an extreme measure as dissolution. The Assembly also considers that requiring parties to win at least 10% of the votes cast nationally before they can be represented in parliament is excessive and that the voting arrangements for Turkish citizens living abroad should be changed.

...

23. The Assembly therefore invites Turkey , as part of its authorities ’ current reform process, to:

...

ii. amend the electoral code to lower the 10% threshold and enable Turkish citizens living abroad to vote without having to present themselves at the frontier;

... ”

C. Comparative law

35 . Although there is no uniform classification of types of ballot and electoral systems, it is usual to distinguish three main types: majority vote systems, proportional systems and mixed systems. In majority vote systems the winner is the candidate or list of candidates obtaining the majority of the votes in the decisive round of voting. This type of ballot makes it possible to vote in governments with clear parliamentary majorities, but at the same time it militates against the representation of minority political parties . Thus, f or example, in the United Kingdom the use over many decades of a single round of voting in a single-member majority-vote system (“first past the post”), combined with the existence of two dominant political parties , has had the effect of giving few seats to other parties in relation to the number of votes that they obtain. There are other similar cases, in France for instance, where there is a majority-vote system spread over two rounds of vot ing . At the opposite extreme, the aim of the proportional representation system is to ensure that the votes cast are reflected in a proportional number of seats. Proportional representation is generally considered to be the fairest system because it tends to reflect more closely the various political forces. However, the disadvantage of proportional representation is that it tends to lead to fragmentation among those seeking electoral support and thus makes it more difficult to establish stable parliamentary majorities.

36 . Currently, proportional systems are the most widely used in Europe . By way of example, Denmark , Spain , Estonia , Ireland , Luxembourg , Malta , Moldova , Norway , Poland , Portugal , the Czech Republic , Romania , Sweden , Bulgaria and Turkey have opted for one or other variant of proportional representation. There are also mixed systems containing various combinations of the two types of ballot (in Italy , Lithuania , Russia , Ukraine and Germany , for example ).

37 . In order to ensure stable majorities in legislatures elected by proportional representation, statutory electoral thresholds are often used. Thresholds are “limits, fixed or variable, defined in terms of the electoral result, which determine the share of a list or candidate in the distribution of seats”. However, the role played by thresholds varies in accordance with the level at which they are set and the party system in each country. A low threshold excludes only very small groupings, which makes it more difficult to form stable majorities, whereas in cases where the party system is highly fragmented a high threshold deprives many voters of representation.

38 . A mong the member States of the Council of Europe which use one or other variant of proportional representation in the context of a mixed system, and which set an electoral threshold , the following examples may be found . In Sweden a party must gain 4% of the votes cast nationally or 12% of the votes cast in the base constituency in which the seat is to be allocated. In Bulgaria a national threshold of 4% is imposed. In Liechtenstein it is necessary to pick up 8% of the votes cast nationally. In Denmark parties must either pick up 2% of the votes cast nationally or obtain a particular number of votes in two of the country ’ s three geographical zones. In the Netherlands there is a national threshold fixed at 0.67% of the votes cast.

39 . As a general rule, the threshold fixed does not apply as such to coalitions, which must pass higher thresholds. In the Czech Republic , for example, the threshold for one party is 5%, whereas in the case of a coalition it is raised by 5% for each of the constituent parties. In Romania the base threshold of 5% is raised by 3%, and only a further 1% for coalitions with three or more members. In Poland the electoral threshold varies between 5% for local lists and 8% for national lists; for a coalition the threshold is set at 8% whatever the number of constituent parties. Following the same logic, the threshold for independent candidates is lower – 3% in Moldova , for example.

THE LAW

I. ALLEGED VIOLATION OF ARTICLE 3 OF PROTOCOL N o . 1

40 . The applicants alleged that the imposition of an electoral threshold of 10% in parliamentary elections interfered with the free expression of the opinion of the people in the choice of the legi slature. They relied on Article 3 of Protocol No. 1, which provides:

“The High Contracting Parties undertake to hold free elections at reasonable intervals by secret ballot, under conditions which will ensure the free expression of the opinion of the people in the choice of the legislature.”

A. Arguments of the parties

1. The applicants

41 . The applicants submitted in the first place that the electoral threshold was based on the particular situation in Turkey after the 1980 military regime and that its aim was to depoliticise society by installing an authoritarian government.

42 . Secondly, they rejected the argument that the threshold served the legitimate aim of ensuring governmental stability. A study of the historical background in Turkey showed that an electoral system without a threshold could also enable solid governments to be formed. Observing that a proportional system without a threshold had been used in the parliamentary elections of 1965, 1969, 1973 and 1977, they emphasised that after the first two of those elections it had proved possible to form single-party governments. Moreover, during the period 1983-2006 Turkey had had only three single-party governments, even though the threshold had then been in force. Imposing such a high threshold did not serve any legitimate aim.

43 . The applicants contended that it was difficult to defend the view that the exceptional measure in question strengthened representative democracy. Such a high national threshold made representation very unfair and led to a crisis of legitimacy for the government, since parliament ought to be the free tribune of any democracy. Clearly, a parliament whose composition reflected only about 55% of the votes cast was not capable of supplying the representative legitimacy on which any democracy is based.

44 . The national threshold of 10% was also disproportionate and arbitrary, and impaired the very esse nce of the right guaranteed by Article 3 of Protocol No. 1. It deprived a large proportion of the population of the possibility of being represented in parliament. In the parliamentary elections of 1987, 1991, 1995 and 1999 the proportion of the votes cast in favour of parties not represented in parliament had been, respectively, 19.4% (about 4.5 million votes), 0.5% (about 140,000 votes), 14% (about 4 million votes) and 18.3% (about 6 million votes ) . The results of the 2002 election had led to a “crisis of representation”, since 45.3% of the votes – that is, about 14.5 million votes – had not been taken into consideration and were not reflected in the composition of parliament.

45 . The applicants also stressed the question of regional representation. They asserted that the parties from the south-eastern part of the country did not have a single member of parliament, although they could count on about two million votes. They submitted in that connection that the electoral threshold had been fixed in particular to block the representation of the Kurdish people of the region. In addition, whereas DEHAP was the leading party in thirteen provincial constituencies and the second strongest in two more, it had not obtained a single seat in parliament.

46 . Lastly, the applicants submitted that the electoral threshold of 10% was very high in comparison with the thresholds which applied in other European systems. They argued that there was no good reason to impose a minimum of 10% nationally and that such an obstacle was fundamentally at variance with representative democracy.

2. The Government

47 . Referring to the principles established in the Mathieu-Mohin and Clerfayt v. Belgium case (judgment of 2 March 1987, Series A no. 113), the Government submitted that Article 3 of Protocol No. 1 did not set forth an absolute right to vote and that the Contracting States should be left a wide margin of appreciation with regard to the fixing of electoral thresholds.

48 . They observed that Article 3 of Protocol No. 1 did not include expres sions such as “everyone” or “no one shall ”, arguing that this seemed to indicate merely an undertaking on the part of the High Contracting Parties “to hold free elections at reasonable intervals by secret ballot”.

49 . Article 3 of Protocol No. 1 guaranteed in principle the right to vote and the right to stand for election to the legislature. Consequently, it provided for the organisation of free elections without imposing any particular electoral system. In addition, these elections had to be held by secret ballot and at “reasonable” intervals. Admittedly, the elections had to be held under conditions calculated to ensure “the free expression of the opinion of the people”. That concept meant that no constraint or pressure was to be brought to bear on electors to influence their choice of candidate; it also implied, essentially, the principle of equal treatment for all citizens in the exercise of their right to vote and their right to stand for election.

50 . As regards the Turkish electoral system, the Government explained that Law no. 2839 had introduced the proportional system with a national threshold of 10%. That system had made it possible to form majorities in the aftermath of the elections in 1983, 1987, 1991, 1995, 1999 and 2002. Thanks to the threshold, it had been possible after three of those elections to form a government from the representatives of a single majority party. That meant that the threshold served a legitimate aim, namely ensuring governmental stability, and that there was a consensus in favour of keeping it. Moreover, in its judgment of 18 November 1995 the Constitutional Court had held that the threshold was not an obstacle to “fair representation”, a principle enshrined in the Constitution since 1995.

52 . The Government pointed out that if DEHAP had succeeded in crossing the 10% threshold it would have won seats in parliament, like the AKP and the CHP, which had obtained 34.26% and 19.4% of the votes respectively.

53 . They further submitted that in domestic law there was nothing to prevent political parties from forming coalitions in order to get through the 10% barrier . DEHAP could have organised a coalition with the other political parties who had presented candidates in the elections on 3 November 2002 and thereby gain ed seats in the Grand National Assembly. In that connection they emphasised that independents, who had obtained 1% of the votes, had won nine seats .

54 . The Government further observed that the CHP – the second strongest party in parliament after the 2002 elections – had been unable to cross the threshold in the 1999 parliamentary elections. That showed that a political party which did not get over the hurdle at any particular election could do so at a later one and thus obtain members ’ seats.

55 . Moreover, the Government emphasised that between 1961 and 1980, during which period proportional representation without any electoral threshold was the practice followed, Turkey had had twenty different governments, whereas during the period 1983 to 2006, during which the 10% threshold had been in force, there had been six – three coalitions and three single-party governments. That clearly showed that the threshold ensured political stability , which had a crucial influence on the country ’ s economy.

B. The Court ’ s assessment

1. General principles

57 . Article 3 of Protocol No. 1 seems at first sight different from the other provisions of the Convention and its Protocols which guarantee rights, as it is phrased in terms of the obligation of the High Contracting Parties to hold elections which ensure the free expression of the opinion of the people rather than in terms of a particular right or freedom.

58 . However, having regard to the travaux préparatoires of Article 3 of the Protocol and the way the provision has been interpreted in the context of the Convention as a whole, the Court has established that Article 3 of Protocol No. 1 guarantees individual rights, including the right to vote and the right to stand for election (see Mathieu-Mohin and Clerfayt , cited above, pp. 22 ‑ 23, §§ 46 ‑ 51). In fact, it has taken the view that this wording, of a type which does not have its like elsewhere, can be explained by the desire to give greater solemnity to the commitment undertaken by the Contracting States and emphasise that this is a sphere in which they are under an obligation to take positive measures and not just refrain from interference ( ibid. , § 50).

59 . The Court , which has frequently pointed out the importance of the democratic principles underl ying the interpretation and application of the Convention (see, among other authorities, United Communist Party of Turkey and Others v. Turkey , judgment of 30 January 1998, Reports of Judgments and Decisions 1998 ‑ I, § 45), emphasises that the rights guaranteed by Article 3 of Protocol No. 1 are crucial to establishing and maintaining the foundations of a meaningful democracy governed by the rule of law.

60 . Nonetheless, the rights enshrined in Article 3 of Protocol No. 1 are not absolute. There is room for implied limitations, and Contracting States must be given a margin of appreciation in this sphere.

61 . The scope of that margin in the present case has given rise to considerable debate. The Court re-affirms that the margin of appreciation in this area is wide (see Mathieu-Mohin and Clerfayt , cited above, § 52; and, more recently, Matthews v. the United Kingdom [GC], no. 24833/94, § 63, ECHR 1999 ‑ I; Labita v. Italy [GC], no. 26772/95, § 201, ECHR 2000 ‑ IV; Podkolzina v. Latvia , no. 46726/99, § 33, ECHR 2002 ‑ II ; and Hirst v. the United Kingdom (no. 2) [GC], no. 74025/01, § 61, ECHR 2005 ‑ IX). The same applies to determination of the type of ballot through which the expression of the opinion of the people is mediated, whether proportional representation, majority voting or some other system (see Matthews , cited above, § 63). In that connection, Article 3 of the Protocol goes no further than prescribing “free” elections held at “reasonable intervals” “by secret ballot” and “under conditions which will ensure the free expression of the opinion of the people”. Subject to that reservation, it does not create any “obligation to introduce a specific system” such as proportional representation or majority voting with one or two ballots (see Mathieu-Mohin and Clerfayt , cited above, § 54).

The rules in this area vary in accordance with the historical and political factors specific to each State; the large variety of situations provided for in the electoral legislation of numerous member States of the Council of Europe shows the diversity of the possible options. For the purposes of applying Article 3 of the Protocol, any electoral legislation must be assessed in the light of the political evolution of the country concerned, so that features that would be unacceptable in the context of one system may be justified in the context of another (see Py v. France , no. 662 89/01, § 46, ECHR 2005 ‑ I (extracts)) , at least so long as the chosen system provides for conditions which will ensure the “free expression of the opinion of the people in the choice of the legislature”.

62 . Moreover, it should not be forgotten that electoral systems seek to fulfil objectives which are scarcely compatible with each other: on the one hand to reflect fairly faithfully the opinions of the people, and on the other, to channel currents of thought so as to promote the emergence of a sufficiently clear and coherent political will. In these circumstances the phrase “conditions which will ensure the free expression of the opinion of the people in the choice of the legislature ” implies essentially – apart from freedom of expression (already protected under Article 10 of the Convention) – the principle of equality of treatment of all citizens in the exercise of their right to vote and their right to stand for election. It does not follow, however, that all votes must necessarily have equal weight as regards the outcome of the election or that all candidates must have equal chances of victory. Thus no electoral system can eliminate “wasted votes” (see Mathieu -Mohin and Clerfayt , cited above, § 54).

63 . And yet none of the above criteria should in principle be considered more valid than any other, provided that it guarantees the expression of the opinion of the people through free, fair and regular elections.

64 . The Convention institutions have always considered electoral thresholds in the context of the margin of appreciation left to member States, noting that in this sphere States enjoy considerable latitude (see Federación nacionalista Canaria v. Spain (dec.), no. 5 6618/00, ECHR 2001 ‑ VI ; Etienne Tete v. France , no. 11123/84, Commission decision of 9 December 1987, Decisions and Reports (DR) 54, p. 52 ; Marcel Fournier v. France , no. 11406/85, Commission decision of 10 March 1988 ; and Silvius Magnago and Südtiroler Volkspartei v. Italy , no. 25035/94, Commission decision of 15 April 1996, DR 85, p. 112).

65 . However, it is for the Court to determine in the last resort whether the requirements of Article 3 of Protocol No. 1 have been complied with. It has to satisfy itself that limitations do not curtail the rights in question to such an extent as to impair their very essence, and deprive them of their effectiveness; that they are imposed in pursuit of a legitimate aim and that the means employed are not disproportionate (see Mathieu-Mohin and Clerfayt , cited above, § 52). In particular, any such conditions must not thwart the free expression of the people in the choice of the legislature – in other words, they must reflect, or not run counter to, the concern to maintain the integrity and effectiveness of an electoral procedure aimed at identifying the will of the people through universal suffrage (see Hilbe v. Liechtenstein (dec.), no. 31981/96, ECHR 1999 ‑ VI, and Melnichenko v. Ukraine , no. 17707/0 2, § 56, ECHR 2004 ‑ X). Any departure from the principle of universal suffrage risks undermining the democratic validity of the legislature thus elected and the laws which it promulgates. Exclusion of any groups or categories of the general population must accordingly be reconcilable with the underlying purposes of Article 3 of Protocol No. 1 (see, mutatis mutandis , Aziz v. Cyprus , no. 69949/01, § 28, ECHR 2004 ‑ V). Equally, once the wishes of the people have been freely and democratically expressed, no subsequent amendment to the organisation of the electoral system may call that choice into question, except in the presence of compelling grounds for the democratic order (see Lykourezos v. Greece , no. 33 5 54/03, § 52, ECHR 2006 ‑ ... ).

2. Application of the above principles in the present case

66 . In the applicants ’ submission, the fact that they were not elected to the National Assembly, despite the score of 45.95% of the votes cast in the constituency of Şırnak achieved in the parliamentary elections of 3 November 2002 by DEHAP, the party on whose list they had stood for election, was incompatible with Article 3 of Protocol No. 1. They explained that their party, which had polled 6.22% of the national vote, had failed to reach the electoral threshold of 10% an d had accordingly been deprived of parliamentary representation.

67 . However, the Court notes that the national threshold concerned is the product of an electoral rule which determines how the seats in parliament are to be shared nationally among the different lists and different candidates. Its effect is to deprive of parliamentary representation those political parties which fail to cross it. It is provided for in section 33 of Law no. 2839 and was introduced well before the elections of 3 November 2002, so that the applicants could have foreseen that if their party failed to get over the hurdle complained of in those elections they would not be able to win any seats in parliament regardless of the number of votes they obtained in their constituency (see, by converse implication, Lykourezos , cited above, § 55).

68 . The Court would further point out that, unlike other Convention provisions, Article 3 of Protocol No. 1 does not specify or limit the aims which a restriction must be intended to serve, and it accepts that the measure complained of is calculated to prevent excessive and debilitating parliamentary fragmentation and thus strengthen governmental stability, regard being had in particular to the period of instability Turkey went through in the 1970s (see paragraph 16 above).

69 . As regards the proportionality of the measure, the Court must examine this question in the light of the criteria established in its case-law and tak e due account of the political and historical context in Turkey, without losing sight of the fact that rules that would be unacceptable in the context of one system may be justified in the context of another.

70 . The Government argued that the measure was proportionate and was largely a matter which fell within their margin of appreciation. They submitted in particular that the applicants could have been elected if they had been independent candidates or if DEHAP had formed a coalition with larger parties before the election.

71 . Regarding the argument grounded on the possibility of standing as an independent candidate, the Court emphasises the irreplaceable contribution made by parties to political debate, in which they can be distinguished from other political actors such as independent candidates, who in general are locally based. In representative democracies political parties represent the different shades of opinion to be found within a country ’ s population, thus contributing to “the free expression of the opinion of the people” (see , in particular, United Communist Party of Turkey and Others , cited above, §§ 44 and 45).

74 . However, an analysis of the results of the parliamentary elections held since the adoption of the thre shold (see paragraphs 14 and 17 ‑ 20 above) shows that it cannot as such block the emergence of political alternatives within society. Equally, the Court notes with interest the Government ’ s argument that the threshold is intended to give small groupings the opportunity to establish themselves nationally and thus form part of a national political project.

75 . It should also be pointed out that Article 67 § 6 of the Constitution (see paragraph 22 above) requires electoral laws to strike a balance between the principles of fair representation and governmental stability. In its judgment of 18 November 1995 the Constitutional Court examined the rationale for the existence of the threshold as a corrective counterbalance to the general principle of proportionality whereby excessive and debilitating parliamentary fragmentation could be avoided. While accepting that thresholds restricted “the right to vote and to be elected”, it considered them acceptable provided that they did not exceed normal limits. Consequently, it held that the 10% threshold was compatible with the constitutional principles concerned (see paragraphs 29 ‑ 30 above).

76 . Admittedly, in view of the extreme diversity of electoral systems adopted by the Contracting States, and taking into account the fact that many countries using one or other variant of proportional representation have national thresholds for election t o parliament (see paragraphs 35 ‑ 39 above), the Court must accept that in the present case the Turkish authorities (both judicial and legislative) – but also Turkish politicians – are best placed to assess the choice of an appropriate electoral system , and it cannot propose an ideal solution which would correct the shortcomings of the Turkish electoral system. The fact remains, however, that the 10% national threshold applied in Turkey appears to be the highest in comparison with the thresholds adopted in other European systems.

77 . Consequently, while noting that it would be desirable for the threshold complained of to be lowered and/or for corrective counterbalances to be introduced to ensure optimal representation of the various political tendencies without sacrificing the objective sought (the establishment of stable parliamentary majorities), the Court considers that it is important in this area to leave sufficient latitude to the national decision-makers. In that connection, it also attaches importance to the fact that the electoral system, including the threshold in question, is the subject of much debate within Turkish society and that numerous proposals of ways to correct the threshold ’ s effects are being made both in parliament and among leading figures of civil society (see paragraph 21 above). What is more, as early as 1995 the Constitutional Court stressed that the constitutional principles of fair representation and governmental stability necessarily had to be combined in such a way as to balance and comple ment each other (see paragraphs 29 and 30 above).

78 . In the light of the above conclusions, the Court does not consider that Turkey has overstepped its wide margin of appr eciation with regard to Article 3 of Protocol No. 1, notwithstanding the high level of the threshold complained of.

79 . Accordingly, there has been no violation of Article 3 of Protocol No. 1.

FOR THESE REASONS, THE COURT

Holds by five votes to two that the re h as been no violation of Article 3 of Protocol No. 1.

Done in French , and notified in writing on 30 January 2007 pursuant to Rule 77 §§ 2 and 3 of the Rules of Court.

S. Dollé J.-P. Costa Registrar President

In accordance with Article 45 § 2 of the Convention and Rule 74 § 2 of the Rules of Court, the joint dissenting opinion of Mr Cabral Barreto and Mrs Mularoni is annexed to this judgment.

J.-P. C . S. D .

JOINT DISSENTING OPINION OF JUDGES CABRAL BARRETO AND MULARONI

(Translation)

We cannot agree with the majority ’ s finding that there has been no violation of Article 3 of Protocol No. 1.

We consider it useful to summarise the general principles applied in the case-law of the Convention institutions on that provision, which are recapitulated in paragraphs 57 to 65 of the judgment:

(1) Article 3 of Protocol No. 1 guarantees individual rights, including the right to vote and the right to stand for election;

(2) the rights guaranteed by Article 3 of Protocol No. 1 are crucial to establishing and maintaining the foundations of a meaningful democracy governed by the rule of law;

(3) Contracting States must be allowed a margin of appreciation in this matter, at least so long as the chosen system provides for conditions which will ensure the free expression of the opinion of the people in the choice of the legislature ;

(4) it is for the Court to determine in the last resort whether the requirements of Article 3 of Protocol No. 1 have been complied with; it has to satisfy itself that limitations do not curtail the rights in question to such an extent as to impair their very essence, and deprive them of their effectiveness; that they are imposed in pursuit of a legitimate aim; and that the means employed are not disproportionate.

It is therefore surely not for the Court to say whether one electoral system is better than another, seeing that any electoral system has advantages and disadvantages, that there is no “perfect” system and that nobody can avoid the phenomenon of “wasted votes”. However, the Court has a duty, in our opinion, to determine in the last resort whether the conditions imposed on the exercise of the right concerned satisfy the requirements of our case-law (see, among other authorities, Mathieu-Mohin and Clerfayt v. Belgium , judgment of 2 March 1987, Series A no. 113). In our view that means that we should consider the electoral system as a whole.

We are perfectly aware that many countries which have adopted proportional representation systems have at the same time laid down thresholds for the election of political parties to parliament, in order to ensure governability. We acknowledge without hesitation that this is a legitimate aim. However, we consider that a problem can arise from the proportionality point of view when the threshold concerned is too high.

All previous cases about electoral thresholds brought to the attention of the Strasbourg institutions have concerned thresholds at a level generally accepted in Europe, that is thresholds of about 5%; it is regrettable that the

majority avoided saying that in the judgment. In the only case of this kind examined by the Court ( Federación nacionalista Canaria v. Spain (dec.), no. 56618/00, ECHR 2001 ‑ VI ), the Court was at pains to emphasise:

“[T]he second paragraph of the first transitional provision of the Canary Islands ’ Statute of Autonomy ... lays down two alternative conditions: either at least 30% of all valid votes must be obtained in an individual constituency or at least 6% of all valid votes must be obtained in the Autonomous Community as a whole. [The Court] considers that a system of that kind, far from hindering election candidates such as those put forward by the applicant federation, affords smaller political groups a certain degree of protection.”

In Turkey the electoral threshold is 10 % nationally. That threshold is considered to be manifestly excessive by the Parliamentary Assembly of the Council of Europe, which in Resolution 1380 (2004) invited Turkey to lower it. That might be sufficient ground for thinking that there is a serious problem under Article 3 of Protocol No. 1.

But as Article 3 of Protocol No. 1 does not impose specific conditions we consider it important to take the Turkish electoral system as a whole.

We note that this system, which sets a very high national threshold for the election of a political party to parliamentary seats, has no corrective counterbalances.

The Government put forward the following two arguments in seeking to persuade the Court that, although the 10% threshold was high in relation to the thresholds generally adopted, the system as a whole was “proportionate”:

(a) the applicants could have been elected if they had been independent candidates;

(b) the applicants could have been elected if DEHAP had entered into a coalition before the election with the larger parties.

On both points we fully share the considerations expressed by the majority in paragraphs 71 to 73 of the judgment: neither argument is persuasive, and the second is even incorrect.

Moreover, at the hearing, the applicants ’ representative mentioned a bill currently the subject of political debate in Turkey which is intended to do away in future with the possibility of standing as an independent candidate in political elections. On that point the Government ’ s representative did not contradict the applicant ’ s representative: it is therefore quite possible that in future the Turkish electoral system will become even more restrictive as regards the possibility of gaining a seat in parliament.

It would admittedly be naïve to take the view that the result of the 2002 election, and in particular the fact that 45.3% of the votes cast were not reflected in the composition of the National Assembly, was solely due to the electoral system: there is no doubt that the electorate wanted to send a clear signal to the parties which had been in the power in the previous parliament. The fact remains, however, that the electoral threshold – twice as high as the European average – and the lack of corrective counterbalances do not help to ensure “the free expression of the opinion of the people in the choice of the legislature”. In addition, the current system does not permit political parties which are very strong at regional level but less so nationally to win seats in parliament. In a large country we consider it very regrettable to prevent political parties which represent millions of voters from entering the national legislature.

One could argue that in majority-vote systems the distribution of seats in relation to the results obtained may sometimes be much more unfavourable than in a proportional representation system which has an electoral threshold (in the present case, a high one). Nevertheless, in majority-vote systems, in principle, all political parties of any importance at national or regional level are represented in parliament, and for us that is decisive for the purposes of Article 3 of Protocol No. 1.

Like the majority, and in accordance with the case-law of the Convention institutions, we consider that in this area States have a very wide margin of appreciation; however, we take the view that in the present case that margin of appreciation was exceeded and that the degree of latitude which the majority have given to the respondent State is excessive.

We remain convinced that this case would warrant examination by the Grand Chamber, as the issues it raises are serious and new.

In our view, the Turkish electoral system, which lays down a national threshold of 10% without any corrective counterbalances, raises such a problem under Article 3 of Protocol No. 1 that there has been a violation of that provision.

Even following the finding of a violation the national legislature would still have a wide margin of appreciation to determine how to amend the electoral legislation to be applied in future elections in the way it judged best for Turkey, while at the same time ensuring better “the free expression of the opinion of the people in the choice of the legislature”.

[1] . For a detailed analysis of the results of the 2002 elections, see Elise Massicard , Les élections du 3 novembre 2002: Une recomposition de la vie politique turque ? , Istanbul , July 2003.