Abstract

This paper explores some causes and the consequences of constituency service in a proportional representational system. Turkey provides an interesting case here. First, its province-size constituencies generate a significant size of personal votes for a candidate, though nominated in the party list. Second, as in other non-western democracies, it is the service and allocational aspects of responsiveness that compose the most frequent demands from constituents to parliamentarians in Turkey. There exists, however, conflicting evidence regarding what priority Turkish parliamentarians give to constituency service. The analysis of the data collected through a questionnaire survey highlights several features of constituency service in Turkey. First, in terms of daily activities, parliamentarians spend for constituency service the largest part of their time at their own disposal. Second, for reelection purposes, parliamentarians consider their individual activities to be almost as important as party popularity. Third, the demand for and supply of constituency service depend on different politico-economic structures of the constituency. The smaller the constituency size, the larger will be the demand for and the supply of constituency service. The economically less developed constituencies tend to generate more demand than the more developed constituencies but do not significantly motivate the parliamentarians to provide constituency service. In addition, the Turkish parliamentarian?s turnover rate is much higher than in other democratic countries. Therefore even reelected parliamentarians have to take constituency service seriously. The statistical test shows no significant difference in the practices of constituency service between newly-elected and reelected parliamentarians.

Texte intégral

Public support to the legislature, which is one of the two keys for an institutionalized legislature, largely depends on the effectiveness of representation. Representation has been generally understood to encompass 1) policy responsiveness, 2) service responsiveness, 3) allocation responsiveness, and 4) symbolic responsiveness. Traditionally, most of the literature on representation has been devoted to policy responsiveness. It was not until the late 1970s that the service and allocation aspects of representation especially in single-member constituencies came to receive attention in the literature (Cain et al. 1984, Jewell 1985, Cain et al. 1987, Norton et al. 1993).

[2] The service and allocational aspects of representation should be also important for some plural-member constituencies for at least two reasons. First, proportional representational systems, which require the electorates to vote for a party, do not encourage representatives to establish closer ties with their constituents than majoritarian electoral systems do. If a proportional-representation system assumes relatively small constituencies, however, connections between representatives and constituents will be closer. This is because voters can easily recognize the few candidates in the party list and also because representatives find it geographically easier than in large-sized constituencies, to reach constituents. Second, for non- Western countries it has been argued that the service and allocational aspects of responsiveness constituted the most important functions of their legislatures (Mezey 1979, Jewel 1985).

[3] This paper1, based on a questionnaire, surveys some causes and consequences of constituency service in a proportional representational system. Turkey provides an interesting case here since, first of all, its province-size (at largest) constituencies, which ranged from two-member to six-member (See Appendix I), bring the personal as well as the partisan variable into electoral politics2. Second, as in other non-Western democracies, it is the service and allocational aspects of responsiveness that compose the most frequent demands of constituents to parliamentarians in Turkey (Kim et al. 1984: 95, Kalayc?o?lu 1986: 322, Kalayc?o?lu 1995: 48-54, Kele? et al. 1993: 37).

[4] Third, on the other hand, however, there is conflicting evidence about what priority Turkish parliamentarians give to constituency service. Authors (Kim et al. 1984) in the latter half of the 1970s found out that Turkish parliamentarians, unlike other non-Western counterparts, spent less time for constituency service than for those activities directly or indirectly related to legislation. Out of 101 Turkish parliamentarians who were asked about the first most frequent activity, 39 responded ?explaining policies to voters? and 30 responded ?proposing, debating, and amending bills?. Responses related to constituency service included only 22 for ?interceding with civil servants on constituents behalf? and 7 for ?obtaining government resources for my district? (Kim et al. 1984: 73). However, Kalayc?o?lu in his 1984 and 1988 surveys, found out that, parliamentarians on average spent as much as half of their time listening to demands from constituents (Kalayc?o?lu 1995: 48-54)3. The difference in time spent for constituency service may have emerged partly because Kalayc?o?lu's field surveys were conducted shortly after the transition from the military to the civilian government in December 1983. Legislative activities were novel to most of the newly elected parliamentarians, since then the politicians were banned from politics until 1987. It is necessary therefore to recast the issue in present light.

[5] The rest of the paper consists of research design, results, and conclusions sections. After the data and the analytical methods are explained in the research design section, the results section investigates 1) the time allocated for constituency service among the total activities of the average parliamentarian, 2) motives for constituency service, 3) politico-economic variables determining the demand and supply of constituency service in each constituency. The conclusions section summarizes the major findings and discusses their implications.

[6] The questionnaire text was written in Turkish by the author and then shown in August 1994 to leading members of the major political parties4 and to high-ranking bureaucrats for comment. After being revised, the text was finally checked by a high-ranking bureaucrat. The questionnaire was mainly composed of seven-point Likert scale indexes while including a few multiple-choice questions. If the respondent did not find any of the given scales or alternatives appropriate, he/she could choose ?other? and could write down his/her answer.

[7] The questionnaire survey was conducted in Ankara between June and October 1995. This was the time when Turkey was experiencing the first coalition government - bringing together the center-right DYP (True Path Party) and the center-left SHP (Social Democratic Populist Party) - after the 1980 military intervention. The parliamentarians were elected in October 1991 for the 19th legislative period of the unicameral Turkish Grand National Assembly (TBMM). While the Turkish legislative period is defined as five years, early general elections were held in December 1995. During the questionnaire survey, however, neither the government nor Parliament had decided to call early elections. Both the 1995 and 1999 general elections resulted in a fragmented parliament. The period of coalition governments lasted until the pro-Islamic AKP (Justice and Development Party) won the November 2002 general election to form a single-party government.

[8] The 1983 electoral law (Law No. 2839) provides the basic frame for the present electoral system in Turkey. Applied first in the 1983 transition election, the law reflected the 1982 constitutional regime and brought about change from the abolished 1961 electoral law. First, the law set two hurdles to eliminate minor parties. One was the requirement of gaining 10% of the total national votes to obtain seats in the parliament. The party could not have the right to seats unless it cleared the national 10% minimum requirement. The other was the same constituency minimum requirement as was used from 1961 to 1965. Second, less importantly, the districting system was changed so that rural provinces would be more over-represented relative to urban provinces. One seat was allocated to each province without regard for population size, and then the remaining seats were allocated in proportion to the population, with at least one seat to each province. If the province was entitled to more than seven seats, then the provinces were divided into plural constituencies that had no less than three seats.

[9] Even after the transition to the civilian government, the incumbent?s meddling with the electoral system for political expediencies continued. Prior to the 1987 general election, ?contingency seats? were introduced for the provinces that had plural four, five, or six-member constituencies. In these constituencies, one contingency seat was to be allocated to the first party in the constituency. Incumbent ANAP thus tried to and actually did secure the majority in the face of declining popularity. Prior to the 1995 general election, the incumbent as well as other mainstream parties apparently feared that the emerging pro-Islamic party might gain the parliamentary majority. As a result of amendments on the electoral law and of their repeals, 1) the constituency minimum requirement was lifted, 2) the contingency seats were abolished, and 3) the ceiling on constituency size was raised to 18 from previous six so that the system became more proportional in effect. This system remains valid to this day5.

[10] The questionnaire was mailed in June 1995 to 430 parliamentarians, which equaled the total 450 members of parliament minus those who had passed away and minus those who had been expelled from Parliament including the DEP (Democracy Party) members6. The parliamentarians were asked to send the fulfilled questionnaire to the researcher in the enclosed envelope. Follow-up work was done in October 1995 to remind by phone the parliamentarians and/or their secretaries of the questionnaire. Since the data was not collected by interview, it was not possible to verify the answers given in the questionnaire sheet. The data therefore cannot be interpreted as purely descriptive. There were 85 respondents, i.e. a 19.8 % response ratio7, consisting of 42 members in governing parties, 42 in opposition, and one member whose party was unknown (Table 1).

Table 1: Respondents by Party and Ministerial Experience

Party in

Ministerial experience in the 19th legislative period

Sample?

Sample %8

Population9

?

Population %

Government

Subtotal

42

50.0

233

52.6

?Yes?

14

16.7

 79

17.8

?No?

27

32.1

154

34.8

No answer

 1

 1.2

?

?

Opposition

42

50.0

210

47.4

Subtotal

84

100.0

443

100.0

No answer

 1

?

?

?

Total

85

?

443

?

Sources: Compiled by the author from TBMM Albümü, 19. Dönem, 1. Bask?, 1992, Ankara, 1992 for the party distribution of the parliamentary seats and from Sanal (1995: 219-223) for the ministerial experience of parliamentarians.

 [11] Out of the 42 respondents from governing parties, 14 had experienced ministerial posts and 27 had not while the remaining one did not answer the relevant question. Despite the small response ratio, the sample distribution of parliamentarians by party and ministerial experience are considerably similar to the population distributions, except for the Eastern region due to the expulsion of DEP parliamentarians. The sample parliamentarians by region (Table 2) are also distributed in much the same manner as the population parliamentarians, with a single exception in the Southeast region. Data was then put into the non-parametric statistical analysis since nominal and ordinal scales used in the questionnaire did not allow the normal-distribution assumption on the population.

Table 2: Respondents by Region

Region

Sample

Population10

N

(%)11

N

(%)

West

29

36.7

131

30.5

South

12

15.2

60

14.0

Central

21

26.6

107

24.9

North

8

10.1

45

10.5

East

9

11.4

87

20.2

Subtotal

79

100.0

430

100.0

No answer

6

?

?

?

Total

85

?

430

?

Source: Compiled by the author from TBMM Albümü, 19. Dönem, 1. Bask?, 1992, Ankara, 1992.

[12] From the results of the questionnaire survey, this section reveals 1) various kinds of constituency service and its proportion to the total activities of parliamentarians, 2) parliamentarians? motives for engaging in constituency service, and 3) political and economic variables that affect the demand for and the supply of constituency service.

[13] Within the time at their own disposal, parliamentarians were busier with meeting constituents than with legislative activities. Table 3 shows that the two largest shares of their time were spent either in the General Assembly or in party group meetings. These activities, however, are almost obligations as parliamentarians or as party members. Then what came next was ?visiting the constituency?, followed by ?meeting constituents for their personal problems? if we exclude another such obligatory activity as ?participating in parliamentary committee meetings?. As a result, parliamentarians allocated very limited time to preparing bills and questions12.

Table 3: Time Spent Annually for Each Activity13

Activities

Index14 mean

Obligatory

Participating in the General Assembly

6.06

Participating in parliamentary group meetings

5.96

Participating in the parliamentary committee

5.32

Voluntary

Visiting the constituency

5.35

Meeting constituents for their personal problems

5.19

Preparing bills

3.13

Preparing written and oral questions

3.08

Source: the author?s questionnaire survey.

[14] The frequency of demand from constituents was closely related to the possibility that parliamentarians will find any form of solution to their problems (Table 4). Spearman?s rhos between the demands brought and demands solved ranged from .311 to .590. It is either because parliamentarians tried to solve those problems that they were most frequently asked to or because constituents brought to parliamentarians those problems which could be possibly solved by parliamentarians, or because of both reasons. The results also show that parliamentarians were more capable in allocation responsiveness than in service and policy responsiveness. Parliamentarians were thus more responsive to such demands as job-creating investments in the constituency (?=.590) and infrastructural services of the constituency (?=.554) than to those demands related to constituents' personal problems (?=.321), complaints of the local government (?=.425), complaints to the central government (?=.311), and legislative activities (?=.448).

Table 4: Frequency of Demands from Constituents15

Demands

Index16Brought (A)

Index Solved (B)

Spearman?s rho for (A) and (B)17

Their personal problems

5.92

3.71

.321

Infrastructural services18 of the constituency

4.79

4.06

.554

Complaints to the central government

4.51

3.25

.311

Job?creating investment in the constituency

4.14

3.10

.590

Complaints of the local government

3.56

3.12

.425

Legislative activity

2.43

3.49

.448

Source: The author?s questionnaire survey.

[15] What are some of the most important motives behind constituency service that take up much of parliamentarians' time of their own disposal? This study assumes that parliamentarians are motivated for constituency service by their primary objective of reelection. It will be possible to analyze the clearer relationship between constituency service and re-electoral motives by dividing the reelection process into the party-list nomination stage and the election stage. This section also sheds light on the role of media that publicize parliamentarians? constituency service activities.

[16] In all parties but the social democratic CHP (Republican People?s Party), it is the party leader and/or the party central committee that decides on the ranking of candidates on the party list. Kalayc?o?lu (1995: 46) argued that ?the political party leaderships always try to balance the local popularity against the personal loyalty of the candidates to the party leader?.

[17] Table 5 shows, however, that the balance is tilted slightly in favor of popularity against loyalty, according to parliamentarians? perception. Parliamentarians regarded ?support from the party rank and file? as the most important variable for being nominated in the party list. ?Support from the party executive? came second in importance. The third and fourth most important variables (?securing investment and public services to the constituency? and ?meeting constituents to cope with their demands?) were both related to constituency service.

Table 5: Variables Affecting MPs' Nomination in the Party List19

Variables

Constituency service relevant

Index20 mean

Support from the party rank and file

X

5.67

Support from the party executive

5.27

Securing investment and public services to the constituency

X

4.78

Meeting constituents to cope with their demands

X

4.71

One's legislative activities

4.34

Speaking to or writing an article for the press

3.87

Asking written and oral questions

3.21

Source: The author?s questionnaire survey.

[18] Among variables affecting election (Table 6) that follows the party list nomination, parliamentarians regarded support to the party from public opinion as the first most important. Public support to the party, however, was beyond the control of individual parliamentarians. The second and the third most important variables were related to constituency service, namely, ?securing investment and public services to the constituency? and ?meeting constituents to cope with their demands?. These variables involved efforts mostly achievable by individual parliamentarians. Thus, among variables within the reach of individual parliamentarians, constituency service was regarded as having the largest effect on the reelection of each parliamentarian.

Table 6: Variables Affecting MPs in Election21

Variables

Constituency service relevant

Index22 mean

Support to the party from public opinion

5.96

Securing investment and public services to the constituency

X

5.16

Meeting constituents to cope with their demands

X

5.15

Party members working for election

4.84

One's legislative activities

4.55

Speaking to or writing an article for the press

4.51

Asking written and oral questions

2.99

Source: The author?s questionnaire survey.

[19] In order to link constituency service to reelection, parliamentarians have to publicize their constituency service activities to their constituents. How do parliamentarians tell their constituents that it was them who have helped to bring some of the investments and public services to the constituency? Will it be rewarding to spend a significant part of their time for meeting constituents to cope with their demands although those people whom parliamentarians can meet in person constitute only a negligent proportion of the total constituents?

Table 7: The Media by which Constituents Learn Their MPs' Activities23

Types of media

Index24 mean

Circulation

Constituents who have met MPs

5.24

Local

Party organizations

4.83

Local

Local newspapers

4.52

Local

Television

4.22

National

National newspapers

3.58

National

Pamphlets

3.50

Local

Radio

2.98

National

Magazines

2.84

National

Source: The author?s questionnaire survey.

[20] This study contends that parliamentarians expect various forms of media to publicize their activities. The media can be broadly grouped into local and national ones in terms of their areas of dissemination and impact. The local media were perceived generally to be more important than the national ones. In particular, constituents who had met parliamentarians were regarded as the most important among not only local but all forms of media. Parliamentarians may have expected constituents who have met them to let other constituents know their activities. In other words, the effort for parliamentarians to cope with constituents? personal demands can pay off if such contacts will have spill-over effects on parliamentarians? publicity.

[21] The preceding analysis showed that Turkish parliamentarians in general perceived constituency service to be critical for reelection and acted accordingly. The question remains, however, whether this is a ubiquitous phenomenon in Turkey. In other words, it is possible to expect the political and economic structures of the constituency to affect the amount, as well as the salience of constituency service. Some structural characteristics of the constituency may more strongly predispose constituents to demand constituency service while other characteristics will make parliamentarians take constituency service more seriously. This paper hypothesizes that at least three structural variables (two being political and the other being economic) are particularly important: 1) constituency size, 2) the level of economic development in the constituency, and 3) the length of incumbency.

Constituency size

[22] Constituency size may well influence the supply of constituency service. In Turkey?s proportional representation system, the number of seats in each constituency ranged from two to six for the 19th TBMM. In terms of reelection, the smaller the number of seats in the constituency, the more important will be the personal appeals of the candidates. Conversely, the larger the number of the seats the more important will be the popularity of the parties of the candidates. It can be hypothesized therefore that parliamentarians from small constituencies are more motivated for and more involved in constituency service than are parliamentarians from large constituencies.

[23] Constituency size may well reflect other features of constituency environments such as socioeconomic development and cultural values. For instance, constituency size was moderately correlated with the constituency?s level of economic development, measured by the per capita income for 1993 of the province25 in which the constituency was located (Spearman?s correlation coefficient = .301, p = .02). In the following, the partial correlation analysis was therefore adopted to control the effect of socioeconomic development on constituency size and vice versa.

[24] The ?supply? part of Table 8 seems to support the above hypothesis. Constituency size, when the per capita income effect was controlled for, was negatively correlated with all indicators of the supply (composed of motives and practice) of constituency service. The smaller the constituency size was, the more importance did parliamentarians attach to constituency service. Relatedly, the smaller the constituency size was, the greater part of their time did parliamentarians spend for meeting constituents, either in Ankara or in their constituencies, and visiting their constituencies.

Table 8: Partial Correlation of Constituency Features with Service26

Constituency service

Constituency features (controlled effect)  ?Constituency size (1991)27

Constituency features (controlled effect)  ?Per capita income (1993)28

Supply ? Motives?Perceived importance for election29/Meeting constituents to cope with their demands

-.253**

.074

Supply ? Motives?Perceived importance for election/Securing investment and public services to the constituency

-.218*

-.025

Supply ? Practice?Time spent30/Meeting constituents for their personal problems

-.255**

-.123

Supply ? Practice?Time spent/Visiting the constituency

-.226*

-.234*

Demand ? Visitors31

-.126

-.137

Demand ? Letters & faxes32

.127

-.139

Demand ? Telephone calls33

-.273**

-.283**

Sources: For constituency service, the author?s questionnaire survey. For constituency size, see Appendix 1. For per capita incomes, see Appendix 2.

[25] Constituency size may also affect demands for constituency service. For small constituency makes it easier for constituents to recognize their representatives. The ?demand? part of table 8 shows that constituency size was negatively correlated with the number of telephone calls from constituents although not with the other two demand-related variables that were more costly for constituents and whose cost varied across constituencies at different levels of development and distance to the capital. The recognition of parliamentarians seems to enhance the likelihood of a constituent making a phone call but not that of his or her writing a letter or paying a visit to Ankara, where MPs usually stay.

[26] The level of economic development may be closely related to demands for constituency service. It will be recalled that major demands for constituency service included jobs, infrastructure, and investment. If the constituency is economically less developed, these demands will be naturally higher there than in economically more developed constituencies. The ?demand? part of Table 8 shows that the per capita constituency income, after the constituency size effect has been controlled, was negatively correlated with the number of telephone calls, which was the cheapest and the easiest way to express constituents? demands to their representatives34.  Per capita income, on the other hand, was not correlated with any supply variables of constituency service except for parliamentarians? ?time spent for visiting the constituency?35. Thus, a lower level of economic development will make more people ask for constituency service but it will probably not directly increase the importance of personal votes.

[27] In his article on constituency service in Britain, Norton (1990) argued that new parliamentarians, who were less secure than others in the following election, might spend a larger part of their time for constituency service than others do. He found between two successive elections a larger number of incremental votes cast for the more recently elected parliamentarians than for others. This led him to believe that new parliamentarians had provided more constituency service than older incumbents had.

[28] The Wilcoxon test applied to the data for the current study showed, however, that in Turkey newly elected parliamentarians did not necessarily provide more constituency service than reelected parliamentarians did. Between the newly elected parliamentarians and the reelected parliamentarians there were not any statistically significant differences in the medians of constituency service indices (see Table 3), i. e., ?visiting the constituency? (p = .56) and ?meeting constituents for their personal problems? (p = .52).

[29] This result is largely due to the difference between the system of single-member constituency in Britain and that of plural-member constituency in Turkey. In a plural-member constituency any incumbent, though holding competitive advantages over new candidates in the constituency, has to compete, first, with other incumbents from his own party for a higher rank in the party list. This competition will be more intense when his party has done well in the preceding election thus producing a larger number of incumbents from his party in the constituency. Second, even if he has secured a higher position in the party list, he must compete with other incumbents from other parties. To be a single incumbent in the constituency from his party, though opening access to the highest rank in the party list, will narrow his chances in the struggle with incumbents from other parties. In brief, being an incumbent provides less security for reelection in a plural-member constituency than in a single-member constituency.

[30] There is supporting evidence from the analysis of effects of constituency size in a similar case. Reed (1994: 285) found that in Japan, where the electorate cast a single-non-transferable vote in the plural-member (ranging from three to five members) district, the incumbency effect was stronger in smaller constituencies than in larger ones. In three-member districts 52% of the districts have reelected all their incumbents since 1972 whereas in four-member and five-member districts the rate was 43% and 34% respectively. The larger the constituency size, the more marginal became the incumbent edge.

Table 9: Turnover Rates: Lower Houses of National Legislatures

Country36

Turnover rate37 (%)

Period38

New Zealand

10.4

1946-92

United States1978-9

14.9

1978-92

Japan

15.0

1972-90

Germany

26.039

1953-87

France

30.040

1946-93

Israel*

40.2

1951-9241

Turkey*

70.2

1991

Source: Calculated by the author from the data in Somiet et al. (1994).

[31] In addition, in Turkey the legislative turnover rate is particularly high, being 70.2% (316/450) in 1991 (Table 9). This rate contrasts with 10.4 % of the members of the New Zealand House of Representatives, 14.9 % of the United States Members of Congress, 15% of the Japanese Diet members, 26.0 % of the CDU Bundestag members in Germany, 30% of the French Assemblée Nationale members, and even 40.2 % of the Israeli Knesset members who compete in the single nationwide constituency. This may be partly because the 1982 Constitution prohibited politicians at the time of the 1980 military intervention from political activity for five to ten years. Before this political ban was eventually abolished by constitutional amendment in 1987, many former politicians had either chosen to retire or found their political base taken over by post-1980 politicians recruited from the bureaucracy or business. Thus there remained very few experienced parliamentarians in post-1980 Turkey.

[32] This paper explored some causes and consequences of constituency service in a proportional representational system. Turkey provides an interesting case here. First, its province-size constituencies generate a significant size of personal votes for a candidate, though nominated in the party list.  Second, as in other non-Western democracies, it is the service and allocational aspects of responsiveness that compose the most frequent demands of constituents to parliamentarians in Turkey. There exists, however, conflicting evidence regarding what priority Turkish parliamentarians give to constituency service. It will be important to know how and to what extent parliamentarians are responding to demands from constituents as well as why they do so.  

[33] The analysis of the data collected through a questionnaire survey illuminated three features of constituency service in Turkey. First, in terms of daily activities, parliamentarians spend for constituency service the largest part of their time at their own disposal. This finding falls in between Kim et al.?s survey (1984), conducted in the latter half of the 1970s and Kalayc?o?lu?s surveys (1995), conducted in 1984 and 1988. The present Turkish parliament, after the 1980-83 interlude by military rule, has probably come to revert to that of the 1970s which consisted of many representatives with legislative experience.

[34] Second, for the purpose of reelection, parliamentarians consider their individual activities to be almost as important as the party popularity. Those activities are publicized through the local media, including the press, the most important of which, however, being the constituents that they meet in person.

[35] Third, the demand for and supply of constituency service depend on different politico-economic structures of the constituency. The smaller the constituency size, the larger will be the demand for and the supply of constituency service. This is because in small constituencies parliamentarians are more recognizable by their constituents and the weight of personal votes is larger than in large constituencies. The economically less developed constituencies tend to generate more demand than the more developed constituencies but do not significantly motivate the parliamentarians to provide constituency service.

[36] Thus, even under proportional representation, relatively small constituencies in Turkey render constituency service important for parliamentarians seeking reelection. In addition, the Turkish parliamentarian?s turnover rate is much higher than turnover rates for parliamentarians in other democratic countries. Even reelected parliamentarians therefore will have to take constituency service seriously. The statistical test showed no significant difference in the practices of constituency service between newly-elected and reelected parliamentarians.

Appendix 1: Constituencies for the 1991 general election

Province N

Province Name

Constituency No.

Constituency Seats

1

Adana

1

6

1

Adana

2

3

1

Adana

3

5

2

Ad?yaman

1

4

3

Afyon

1

6

4

A?r?

1

4

5

Amasya

1

3

6

Ankara

1

6

6

Ankara

2

6

6

Ankara

3

6

6

Ankara

4

5

7

Antalya

1

6

7

Antalya

2

3

8

Artvin

1

2

9

Ayd?n

1

6

10

Bal?kesir

1

4

10

Bal?kesir

2

3

11

Bilecik

1

2

12

Bingöl

1

3

13

Bitlis

1

3

14

Bolu

1

5

15

Burdur

1

3

16

Bursa

1

6

16

Bursa

2

6

17

Çanakkale

1

4

18

Çank?r?

1

3

19

Çorum

1

5

20

Denizli

1

6

21

Diyarbak?r

1

5

21

Diyarbak?r

2

3

22

Edirne

1

4

23

Elaz??

1

4

24

Erzincan

1

3

25

Erzurum

1

4

25

Erzurum

2

3

26

Eski?ehir

1

5

27

Gaziantep

1

6

27

Gaziantep

2

3

28

Giresun

1

4

29

Gümü?hane

1

2

30

Hakkari

1

2

31

Hatay

1

5

31

Hatay

2

3

32

Isparta

1

4

33

Içel

1

6

33

Içel

2

3

34

Istanbul

1

6

34

Istanbul

2

6

34

Istanbul

3

6

34

Istanbul

4

6

34

Istanbul

5

6

34

Istanbul

6

6

34

Istanbul

7

6

34

Istanbul

8

5

34

Istanbul

9

3

35

Izmir

1

6

35

Izmir

2

6

35

Izmir

3

4

35

Izmir

4

3

36

Kars

1

5

37

Kastamonu

1

4

38

Kayseri

1

4

38

Kayseri

2

3

39

K?rklareli

1

3

40

K?r?ehir

1

3

41

Kocaeli

1

4

41

Kocaeli

2

3

42

Konya

1

6

42

Konya

2

4

42

Konya

3

3

43

Kütahya

1

5

44

Malatya

1

6

45

Manisa

1

6

45

Manisa

2

3

46

Kahramanmara?

1

4

46

Kahramanmara?

2

3

47

Mardin

1

5

48

Mu?la

1

5

49

Mu?

1

3

50

Nev?ehir

1

3

51

Ni?de

1

3

52

Ordu

1

6

53

Rize

1

3

54

Sakarya

1

6

55

Samsun

1

6

55

Samsun

2

3

56

Siirt

1

3

57

Sinop

1

3

58

Sivas

1

6

59

Tekirda?

1

4

60

Tokat

1

6

61

Trabzon

1

6

62

Tunceli

1

2

63

?anl?urfa

1

5

63

?anl?urfa

2

3

64

U?ak

1

3

65

Van

1

5

66

Yozgat

1

5

67

Zonguldak

1

4

67

Zonguldak

2

3

68

Aksaray

1

3

69

Bayburt

1

2

70

Karaman

1

2

71

K?r?kkale

1

3

72

Batman

1

3

73

Sirnak

1

3

74

Bart?n

1

2

Total

-

-

450

Source: State Institute of Statistics(1992).

Appendix 2: Per capita provincial income in 1993

Province No

Province Name

Per capita income in 1993 (TL at 1987 prices)

1

Adana

1560940

2

Ad?yaman

1064153

3

Afyon

889549

4

A?r?

296990

5

Amasya

977149

6

Ankara

2251196

7

Antalya

1740647

8

Artvin

1763629

9

Ayd?n

1698122

10

Bal?kesir

1540695

11

Bilecik

2508682

12

Bingol

365579

13

Bitlis

439713

14

Bolu

1614151

15

Burdur

1324673

16

Bursa

2356180

17

Çanakkale

2044309

18

Çank?r?

756539

19

Çorum

1230741

20

Denizli

1620637

21

Diyarbak?r

1074895

22

Edirne

1359210

23

Elaz??

1271147

24

Erzincan

731540

25

Erzurum

656647

26

Eski?ehir

1612345

27

Gaziantep

1324697

28

Giresun

781717

29

Gümü?hane

612764

30

Hakkari

300966

31

Hatay

1444647

32

Isparta

1030363

33

?çel

2084310

34

?stanbul

2539805

35

?zmir

2614320

36

Kars

516588

37

Kastamonu

1108650

38

Kayseri

1116395

39

K?rklareli

2308464

40

K?r?ehir

938757

41

Kocaeli

4422665

42

Konya

1138854

43

Kütahya

1319759

44

Malatya

1138317

45

Manisa

2170312

46

Kahramanmara?

942882

47

Mardin

698693

48

Mu?la

2015040

49

Mu?

361796

50

Nev?ehir

1720398

51

Ni?de

1174826

52

Ordu

697467

53

Rize

1247248

54

Sakarya

1320462

55

Samsun

1230829

56

Siirt

732922

57

Sinop

921038

58

Sivas

834956

59

Tekirda?

2104988

60

Tokat

963919

61

Trabzon

1047306

62

Tunceli

614273

63

?anl?urfa

708905

64

U?ak

1092381

65

Van

519622

66

Yozgat

694058

67

Zonguldak

1399660

68

Aksaray

742698

69

Bayburt

516043

70

Karaman

1696645

71

K?r?kkale

1663582

72

Batman

951359

73

??rnak

326248

74

Bart?n

566968

Source: General Directorate of Regional Development and Structural Adjustment, State Planning Organization (1997).

Bibliography

Cain, Bruce E.; Ferejohn, John A.; Fiorina, Morris P. (1984) ?The Constituency Service Basis of the Personal Vote for U.S. Representatives and British Members of Parliament,? American Political Science Review 78 (1), pp. 110-25.

Cain, Bruce E.; Ferejohn, John A.; Fiorina, Morris P. (1987) The Personal Vote: Constituency Service and Electoral Independence, Cambridge (Mass.), Harvard University Press.

General Directorate of Regional Development and Structural Adjustment, State Planning Organization (1997) ?ller ve Bölgeler ?tibariyle Çe?itli Göstergeler [Various Indicators by Province and Region], http:// www.dpt.gov.tr.

Jewell, Malcolm E. (1985) ?Legislators and Constituents in the Representative Process? in Gerhard Loewenberg, Patterson, Samuel C.; Jewell; Malcolm E.  (eds.), Handbook of Legislative Research, Cambridge (Mass.), Harvard University Press, pp. 97-131.

Kalayc?o?lu, Ersin (1986) ?Turk Yasama Sistemi ve Siyasal Temsil? in Kalayc?o?lu, Ersin; Sar?bay, Ali Ya?ar (eds.) Türk Siyasal Hayat?n?n Geli?imi, Istanbul, Beta.

Kalayc?o?lu, Ersin (1995) ?The Turkish Grand National Assembly: A Brief Inquiry into the Politics of Representation in Turkey? in Bal?m, Çi?dem; Kalayc?o?lu, Ersin; Karata?, Cevat; Winrow, Gareth; Yasamee, Feroz (eds.), Turkey: Political, Social and Economic Challenges in the 1990s, Leiden, Brill.

Kele?, Ru?en; Hazama, Yasushi (1993) The Policy Process in Turkish Democracy, Tokyo, Institute of Developing Economies.

Kim, Chong Lim; Barkan, Joel D.; Turan, ?lter; Jewell, Malcolm E. (1984) The Legislative Connection: The Politics of Representation in Kenya, Korea, and Turkey, Durham, NC, Duke University Press.

Mezey, Michael L. (1979) Comparative Legislatures, Durham, NC, Duke University Press.

Norton, Philip (1990) ?Constituency Service by Members of Parliament: Does It Contribute to a Personal Vote?? Parliamentary Affairs 43 (2), pp. 196-208.

Norton, Philip; Wood, David M. (1993) Back from Westminster: British Members of Parliament and Their Constituents, Lexington, The University Press of Kentucky.

Reed, Steven R. (1994) ?The Incumbency Advantage in Japan? in Somit, Albert, Wildenmann, Rudolf; Boll, Bernhard (eds.), The Victorious Incumbent: A Threat to Democracy? Dartmouth, Aldershot.

Sabuncu, Yavuz; ?eker, Murat (1996) ?Seçimler? in Cumhuriyet Dönemi Türkiye Ansiklopedisi: Yüzy?l Biterken 14. Cilt, Istanbul, ?letisim, pp. 1142-1168.

Sanal, Turker (1995) Turkiye Cumhuriyeti ve 50 Hukumeti, Ankara, Sim Matbaacilik.

Somit, Albert, Wildenmann, Rudolf; Boll Bernhard (eds.) (1994) The Victorious Incumbent: A Threat to Democracy? Dartmouth, Aldershot.

State Institute of Statistics (1992) Results of General Election of Representatives (Summary Tables) 20.10.1991, Ankara.

Tuncer, Erol (2003) Osmanl??dan Günümüze Seçimler (1877-2002), expanded 2nd print, Ankara, Toplumsal Ekonomik Siyasal Ara?t?rmalar Vakf?-TESAV.

Notes
1 The author would like to express his deep appreciation to all the parliamentarians who spared their precious time to respond to the questionnaire. In particular, Ercan Karaka?, Parliamentary Group Vice Chairman of the SHP (Social Democratic Populist Party) ; ?stemihan Talay, Parliamentary Group Vice Chairman of the CHP (Republican People's Party); ?evket Kazan, Parliamentary Group Vice Chairman of the RP (Welfare Party); R?za Müftüo?lu, Parliamentary Group Vice Chairman of the MHP (Nationalist Action Party), as well as Hasan Ça?layan, Director of Press and Information of the DYP (True Path Party) Headquarters, Faruk Abdüllaho?lu from the RP Headquarters, and Avni Çarsancakl? from the ANAP (Motherland Party) Headquarters, kindly accepted interviews with the author while he was preparing for the questionnaire text. The author owes special thanks to the late Assoc. Prof. Dr. Cahit Emre of Ankara University, who helped greatly with the questionnaire follow-ups, and ?smail Gündüz of the Ministry of Interior, who corrected the author?s Turkish in the questionnaire draft. The author also acknowledges the priceless opportunity for documentation work at the Library of the Turkish Grand National Assembly, which has been provided by its General Director Ali R?za Cihan and other members of the library including Sevgi Korkut, ?ahin Akda?, and Nafiz Ertürk.
2 For constituency size, see infra [22]. Another possible source of personal votes is the preferential voting system, which allows electorates to choose from the candidates on the same party list. The system was introduced just before the 1991 general election although it had been written in the 1983 electoral law. A provisional article of the law prevented the system from being applied to the 1983 general election. The law for the 1987 general election also chose not to put the preferential voting system into effect. Even after the first implementation of the system in the 1991 general election, however, it does not seem to have been institutionalized. In the following general elections from 1995 to 2002, the preferential voting system was again not used.
3 His 1984 survey was originally reported in Kalayc?o?lu (1986).
4 They include Ercan Karaka?, Parliamentary Group Vice Chairman of the SHP (Social Democratic Populist Party); ?stemihan Talay, Parliamentary Group Vice Chairman of the CHP (Republican People's Party); ?evket Kazan, Parliamentary Group Vice Chairman of the Refah Party (Welfare Party); R?za Müftüo?lu, Parliamentary Group Vice Chairman of the MHP (Nationalist Action Party), as well as Hasan Ça?layan, Director of Press and Information of the DYP (True Path Party) Headquarters, Faruk Abdüllaho?lu from the Refah Party (Welfare Party) Headquarters, and Avni Çarsancakl? from the ANAP (Motherland Party) Headquarters.
5 For the vicissitude of Turkey?s electoral system, see Tuncer (2003) and Sabuncu and ?eker (1996).
6 DEP members were arrested and lost their parliamentary seats after Parliament lifted their immunity in March 1994 because of the leader's statement sympathetic to the Kurdistan Workers' Party (PKK).
7 This response ratio is not much low for this kind of mail questionnaire.
8 Percentages to the subtotal of the government and opposition members.
9 The calculation of the number of those with ministerial experience in the 19th legislative period required us to use the earliest data available, as of 1 March 1992, in the first edition of the TBMM album for the 19th legislative period.
10 The population, 430 parliamentarians, are the residual of the total 450 members of parliament minus those who had passed away and minus those who had been expelled from Parliament. I mailed the questionnaire to these 430 parliamentarians in June 1995.
11 Percentages to the subtotal of the five regions.
12 Kim et al. (1984)'s multiple questions included ?proposing, debating, and amending bills? whereas the questionnaire separately treated, on the one hand, participating in parliamentary or party meetings, and on the other, preparing bills and questions. By doing so I distinguished (almost) obligatory legislative activities from voluntary ones.
13 Question: ?How much time do you spend in a year for the following duties? For each item, please mark the appropriate number on the scale.?
14 From 1 (=Least) to 7 (=Most). There were no adjectives for 2 to 6.
15 Question for A: ?How much part do the following items form of the total demand of the constituents who come to meet you? For each item, please mark the appropriate number on the scale?. Question for B: ?How much part of the following items for which the constituents wanted to meet you comes to a solution? For each item, please mark the appropriate number on the scale.?
16 From 1 (=Least) to 7 (=Most). There were no adjectives for 2 to 6. Entries are index means.
17 Spearman's rank-sum correlation coefficients (rhos), all of which are statistically significant at the 0.01 level (the level of statistical significance, or the p value, shows the reliability of the correlation coefficient). Coefficients greater than .50 in value are in bold face.
18 Such as roads, water, and electricity.
19 Question: ?For entering the first rows of the party electoral list, how important are the following efforts? For each item, please mark the appropriate number on the scale.?
20 From 1 (=Least) to 7 (=Most). There were no adjectives for 2 to 6.
21 Question: ?For being reelected in the general election, how important are the following efforts, apart from entering the first rows of the party electoral list? For each item, please mark the appropriate number on the scale.?
22 From 1 (=Least) to 7 (=Most). There were no adjectives for 2 to 6.
23 Question: ?From the following communication media items, to what extent do constituents learn of your activities? For each item, please mark the appropriate number on the scale.?
24 From 1 (=Least) to 7 (=Most). There were no adjectives for 2 to 6.
25 A province (il) in the Turkish political system refers to the whole geographical area covering not only the central district (merkez ilçesi) but all the other districts (ilçeler) that fall within the provincial boundary.
26 Entries are Spearman's partial correlation coefficients. I have adopted the partial correlation analysis since constituency size and per capita constituency incomes are moderately correlated (Spearman?s partial correlation coefficient = .301, p = .02). Question: ?While the parliament is in session, with how many constituents per week do you meet in Ankara or in your constituency??. Answers: a=less than 100, b=100-199, c=200-299, d=300-399, e=400-499, f=500-599, g= 600?699, h=700-799, i=800-899, j=900?999, k=1,000 or more. Question: ?While the parliament is in session, how many letters or fax messages arrive from constituents per day??Answers: a= 1-10, b= 11-20, c= 21-30, d= 31-40, e= 41-50, f= 51-75, g= 76-100, h= 101-150, i= 151-200, j= 200 or more. Question: ?While the parliament is in session, how many telephone calls do you receive from your constituents per day?? Answers: a= 1-20, b= 21-40, c= 41-60, d= 61-80, e= 81-100, f= 101-150, g= 151-200, h= 201-300, j= 300 or more. *p<.10.**p<.05.
27 The effect of per capita income is controlled.
28 The effect of constituency size is controlled. The year 1993 falls in the middle between the first year of the 19th legislative period, 1991 and the year in which the questionnaire survey was conducted, 1995. Per capita constituency incomes were approximated by the per capita income of the province in which the constituency was located. At the time of the 1991 general election out of the total 74 provinces, 54 had single constituencies. It means that in nearly two-thirds of the provinces, constituency incomes are equal to provincial incomes. The most populated Istanbul Province, however, has nine constituencies.
29 See Table 6.
30 See Table 4.
31 Per week, either in Ankara or in the constituency (median= ?200-299?).
32 Per day (median= ?11-20?).
33 Per day (median= ?21-40?).
34 The level of economic development was not correlated with either the number of letters and faxes or the number of visitors at a statistically significant level. It can only be speculated that the low literary rates in less developed constituencies make it difficult for the constituents to write letters or fax messages to their representatives while high transportation costs and bad traffic access may be hindering visits to Ankara from less developed constituencies. These positive relationships between the level of economic development and the difficulties in scriptural communication or journey may have cancelled out the potentially negative relationship between the level of economic development and the intensity of demands.
35 In other words, parliamentarians from economically less developed constituencies spend more time visiting their constituencies than other parliamentarians. It is plausible that they are not so much willing as forced to spend a lot of time to visit those constituencies with poor traffic access.
36 Countries with an asterisk (*) have a unicameral legislature.
37 Period average.
38 Starting and ending with general election-years.
39 Christian Democratic Union (CDU) Bundestag members. The same rate was 24.5 % for German Social Democratic Party (SPD) Bundestag members. Both figures exclude Berlin representatives.
40 Excluding parliamentarians who died in office or resigned.
41 Excluding the very early general election of 1961.