Home Is Where the Heart Is
An analysis of local-orientation in parliamentary
questions
Bachelor thesis Political Science
Name: Olaf Leeuwis
Student number: S1849913
First reader: Dr. Tim Mickler
Second reader: Dr. Frits Meijerink
2
Table of Contents
Introduction ... 3
Literature review ... 4
Representation and representational foci ... 4
Research design and method ... 8
Methodology ... 8
Operationalization ... 9
Results ... 12
Conclusion ... 17
3
Introduction
Representation is generally considered to be one of the core functions of parliaments (Von Beyme, 2000, p. 72). There are various normative and theoretical views about what this function of representation entails. For instance, some scholars prescribe representatives to be more national focused, whilst other scholars prefer representatives to be more focused on their constituency. One can always have a normative debate about what ought to be the proper focus of representation, but at the end of the day, it is an empirical question how representatives really behave and what they focus on.
In this thesis, I will examine what factors influence legislators to be more national focused or more constituency, local focused. First of all, because in recent decades, the discussion about a supposed ‘gap’ between legislators and citizens (or elite and the people) has become hot topic. This gap has not only fueled the rise of populism since the beginning of this millennium, populist parties themselves have signaled the gap to the people through an elite-people dichotomy. This gap between legislators and citizens has led some scholars to call it a ‘crisis of legitimacy’. When members of parliament’s are more local focused, they might contribute to closing this gap because citizens not only trust local candidates more, the burden to share perceived societal problems also becomes smaller. Second of all, this thesis will contribute to the overall transparency of legislators. Voters have a right to know what, why and how an MP spends his time on. To know that a an MP is more locally focused through the use of parliamentary questions, may for many voters be a reason to vote for him/her or not.
Most research has dealt with measuring factors that influence the variation in foci of representation in political systems that have electoral districts. However, we do not know much about what factors influence the foci of representation in parliamentary systems that do not have any districts at all. This makes my thesis scientifically relevant.
This thesis will use parliamentary questions as the main tool of representation, because this tool is almost completely free from party-restraint and can therefore be used by MPs for more personal aspirations. My research question will thus be:
What explains the variation of local-orientated parliamentary questions?
In this next part, I will review the literature on parliamentary questions and the two main foci of representation. I will start with a short literary overview of what representational foci are. Besides, I will outline the various factors that influence MPs to be either more national or more constituency focused. From these factors, I will deduce 8 testable hypotheses for my thesis. The section after the literature review will focus on the method and research design.
4 Besides, I will elaborate more on the Dutch and why it suits the analysis. The part after that will discuss the results: the data will support only 1 of the 8 hypotheses because only the factor of whether an MP is born in his residential region can explain whether an MP has asked a local-orientated PQ.
Literature review
Representation and representational foci
Through the centuries, there have been theoretical and political debates about what the function of a representative (or member of parliament or legislator) entails and what he ought to focus on in parliament. One view is the view of Edmund Burke: ‘Parliament is not a congress of ambassadors from different and hostile interests (…) but Parliament is a deliberative assembly of one nation, with one interest, that of the whole (Burke, 1777). In these sentences, Burke makes a clear distinction between two possible foci of representation: a local focus on the one hand and a more national focus on the other hand (Eulau, 1962, p. 270). Normatively, Burke makes a clear suggestion about which role representatives ought to fulfill: the role by which legislators are focused on the national interest rather than on separate, hostile interests.
Scholars have built upon these roles/foci of representation and because these foci are not mutually exclusive (Eulau, 1962, p. 270), scholars have aimed to explain why some representatives tend to be more national focused or some to be more local focused. Various factors have been researched that might explain the variation in foci of representation present in various parliamentary activities. The emphasis will lie on the activity of parliamentary questioning by reason of my research goal. Below, I will discuss these factors and I will deduce testable hypotheses from the discussed literature.
The first two factors are about electoral incentives to cultivate personal votes. These incentives vary among cases and electoral systems, and these also vary within a case and an electoral system. As possible explanations for constituency-centered behavior, these two factors are in line with the assumption that MPs are continually making efforts to gain electoral ground through signaling their stances and parliamentary efforts. This is what Martin (2011a, 2011b) calls a ‘personal vote earning strategy’, which is opposite of more ‘party vote-earning strategies’ (Martin, 2011b, p. 474, 475). Constituency-centered behavior can be a personal-vote earning strategy, in both district-systems and district-less systems.
The first factor is the electoral incentives shaped by institutional variables, with the electoral system in particular. Due to the fact that electoral systems and their accompanied electoral incentives vary, scholars have made comparative analysis and case studies on what
5 the effect of the electoral system is on the degree of constituency focus. As an example, a comparative analysis between six electoral systems by Heitshusen et al. (2005) suggests that the degree of constituency focus mainly depends on the electoral safety and incentives the electoral system offers (Heitshusen et al., 2005, p. 42). For instance, the degree of constituency focus in closed-list proportional-systems falls behind this focus in other electoral systems because the electorate cannot vote for individual candidates and therefore personal vote-maximization efforts become theoretically useless. However, when looking at the cases the Netherlands and Italy, the degree of constituency focus seems to be lower, but not absent (Thomassen & Andeweg, 2004; Russo, 2011). In both countries, the electoral system (proportional representation) offers no institutional reasons to be more local focused, due to the absence of districts (The Netherlands) and due to a closed list (Italy). The former article concluded this on the basis of self-reported behavior and the latter article on content analysis of PQ’s. However, even in closed-list PR systems (where voters cannot vote for individual candidates), such as Italy, local candidates matter because voters turn out to vote who otherwise would abstain and through shifts in distribution of votes that would turn out anyway (Fiva & Halse, 2016, p. 22). Hence, the type of electoral system has an effect on the degree of local-focused behavior by MPs, however, even in electoral systems where the institutional incentives are few or absent, MPs may still engage in constituency-focused behavior. Nonetheless, not much research has been done on party-centered electoral systems and its institutional effect upon representational foci.
The second factor is the intra-system electoral incentives, shaped by electoral uncertainty and career-prospects. MPs and candidates have to compete against each other both to be re-selected and to be re-elected. This competition creates electoral uncertainty and therefore personal-vote earning strategies (such as local-focused behavior) may be beneficial for an MP. The factor varies because some MPs have to face higher levels of electoral uncertainty than other MPs and some MPs lack these electoral incentives altogether because they are to retire. For instance in the German case the amount of parliamentary questions is at its peak with MPs who are at the end of their first term (Bailer & Ohmura, 2017). The article attributes this result to the career prospects of MPs, ‘showing themselves’ in order to get re-elected. The authors also find that senior MPs are on the other hand asking fewer PQs, which they attribute to this lack of career prospects. The premise of this article is that the longer MPs are in parliament, the more certain they are of re-election due to shown efforts. However, this latter finding is not supported in Zittel et al. (2019) when looking at the absolute number of constituency-focused PQs and seniority in the same parliament. A similar conclusion is drawn
6 in the case of the House of Commons: MPs who represent marginal constituencies ask more parliamentary questions than MPs who represent electorally safe constituencies (Kellermann, 2016, p. 102). Based on an analysis of 6 parliaments, this overall conclusion is also supported: MPs with a very low degree of electoral safety in general tend to be much more constituency focused (Heitshusen et al., 2005, p. 40). I therefore will deduce the following hypothesis concerning electoral safety:
H1: More electorally vulnerable MPs will ask more local-orientated parliamentary questions than less electorally vulnerable MPs.
Besides, because seniority has been linked to both electoral safety and personal vote-earning strategies, I will deduce the following hypothesis:
H2: More senior MPs will ask fewer local-orientated parliamentary questions than more senior MPs.
In Zittel et al. (2019), the authors examine what explains constituency-focused PQs of MPs in the German Bundestag. The overall factor that influences German MPs to do so is their localness. Localness in essence means the local ties an MP has with his constituency, due to either biographical ties or political ties, such as having been born there and having held a local political office (Zittel et al., 2019, p. 689). Therefore, in the line of personal-vote seeking behavior, the authors conclude that MPs ‘signal’ constituency-orientated behavior to their constituents in order to gain ground and be re-elected (Zittel et al., 2019, p. 703). In the Italian case, the same conclusion was drawn: ‘(..) being born and living in the constituency of election more than doubles the number of constituency-orientated questions (Russo, 2011, p. 299). Another type of parliamentary activity that is caused by localness is pork barreling, though it should be stated that this possibility is rare among parliaments. Pork barreling happens when politicians favor their own constituency more when attributing funds. In the case of Norway, when holding executive office, MPs tend to engage in pork barrel politics because of the fact they remain loyal to their own region where they have held local office (Fiva & Halse, 2016, p. 25). A plausible reason for this is that representatives that have been active in local politics are more informed about local projects and local political issues.
7 Because political ties and biographical ties influence the variation in representational foci, I will deduce the following two hypotheses:
H3: MPs who have political ties to their residential region ask more local-orientated parliamentary questions in comparison with MPs who do not.
H4: MPs who were born in their residential region ask more local-orientated parliamentary questions in comparison with MPs who were not.
Besides electoral systems, electoral uncertainty and localness, scholars have examined whether the geographic distance to the capital matters for the variation in constituency focus. This has been measured both in distance and in a dichotomous way: periphery versus the center. The literature finds strong evidence for this. For instance in the Italian case, Russo finds evidence that Southern Italian MPs are more likely than Northern Italian MPs to be involved in constituency service through PQ’s (Russo, 2011, p. 299). In this case, the south is the peripheral part and the north is the center part of the country. The same goes for Ireland: Martin (2011b) finds that besides government party membership, the geographic distance to the political center is the only robust predictor of a local orientation. This finding is also supported in the case of Chile (Aléman et al., 2018, p. 239) and in 6 other legislatures when measuring travel time: ‘(..) once an MPs has expended much of her time traveling to the constituency, he is more likely to focus exclusively on those activities, having already expended sunk costs to get there’ (Heitshusen et al., 2005, p. 40). Because there is strong evidence that the distance between center and constituency matters, I will deduce the following hypothesis:
H5: MPs who reside in the periphery ask more local-orientated parliamentary questions in comparison with MPs who reside in the center.
Another factor is the MPs party membership and whether his/her party is in government. Martin (2011b) focuses on the latter and expects that: ceteris paribus, MPs from a government party may have a larger number of committee assignments, resulting that committee work pulls backbenchers away from representing constituency interests such as through parliamentary questioning (Martin, 2011b, p. 481). Apparently, in the case of Ireland this effect is the other way round: the percentage of local questions is 8 percent higher for MPs from government parties than MPs from opposition parties. Still, there is a significant effect of party-membership upon the variation in local questions. Russo (2011) focuses on the former in the case of Italy. A clear finding is that regional parties (Lega Nord, Alleanza Nazionale) are more focused on their constituencies through PQs than MPs from non-regional parties (Russo, 2011, p. 297).
8 According to the author, this may be strengthened due to the fact these parties were in the opposition at the time, which was also Martins expectation but which was not the case. Hence, we do know the party family (regional parties) and whether the party is in government has an effect on representational foci, but in the case of the latter, we do not know in which way. Therefore, I will deduce the following two hypotheses for this latter factor:
H6a: MPs who are member of a government party ask more local-orientated parliamentary questions.
H6b: MPs who are member of an opposition party ask more local-orientated parliamentary questions.
Besides, because regional parties tend to be more regional-orientated and therefore MPs from more regional-orientated parties are more likely to ask local-orientated PQs, I will deduce the following hypothesis:
H7: MPs from more regional-orientated parties ask more local-orientated parliamentary questions.
Lastly, some research has linked gender to the content of parliamentary questions. From a research on PQs in the House of Commons, it became clear that female members of the HoC were much more likely to ask parliamentary questions which contained the words ‘women’ and ‘gender’ than their male colleagues (Bird, 2005). This gives the expectation that female MPs are more national-policy focused and I will therefore deduce the following hypothesis:
H8: Male MPs ask more local-orientated parliamentary questions than female MPs.
Research design and method
Methodology
In order to test the hypotheses and to answer the research question, I have conducted this analysis through a single-case study. Based on the literature review, a range of explanatory factors were linked to the variation in representational foci. It therefore seemed that representational foci are a complex phenomenon. Case studies are particularly useful as a research design when there is a complexity of variables because this type of study allows a researcher to apply a more holistic approach. A holistic approach in essence is an approach, which aims to explain certain phenomena by taking as much as possible factors from its environment into the analysis, due to the phenomena’s complex nature (Swanborn, 2010, p. 18). However, case studies should only be ‘holistic’ when ‘we simply do not know which
9 variables are relevant for the model and which variables are not’, only then it is wise not to be too selective in your choice of factors (Swanborn, 2010, p. 20). Based on the literature review, it was clear that some variables were found to significantly influence the variation in representational foci, whilst in some other cases these same variables did not influence this variation. Thus, Swanborn’s criterion applied to my study, making it suitable to apply a holistic approach and to conduct a case study (Bryman, 2016, p. 61; Swanborn, 2010, p. 25).
The case chosen for a case study can be classified into various types, such as a typical or an extreme case (Seawright & Gerring, 2008). The most appropriate type for my thesis was a least likely case. First of all, because this type has the advantage that when the data support our hypotheses, then the outcome provides leverage for our confidence in the validity of these theoretical propositions as for the fact these relationships were found even in adverse circumstances (Levy, 2008, p. 12). Second of all, this in turn gives confidence for making cross-case generalizations because ‘if it makes it there, the theory can make it everywhere’ (Blanchard, 2017). In my thesis, a case that would constitute a least likely case would be a parliament in which all MPs represent the same national district instead of separate districts. In such a case, we would not expect there to be MPs who are more local focused because the electoral system lacks institutional features that force geographical representation (Latner & McGann, 2005, p. 710; Thomassen & Andeweg, 2004, p. 48). Besides, the case should have a party-centered rather than a candidate-centered political system, because the former lacks the incentives to cultivate a personal vote, such as through a local focus (Grofman, 2005).
For the first criterion several countries appeared. These were South Africa, Fiji, Mozambique, The Netherlands, Moldova, Slovakia, Israel and Serbia. In all these countries, members from the lower house all represent the national district. Of these countries, the party focus of MPs in the Netherlands is one of the strongest, making it also suitable for the second
criterion (André et al, 2016, p. 47). The Dutch Tweede Kamer is elected through an open-list
proportional representation system by which there is only one, nation-wide electoral district.
In this way, there is no formal bond between Tweede Kamerleden and a particular district (Thomassen & Andeweg, 2004, p. 53). Besides, most voters vote for the party leader, which is the number one on the party list, and therefore ‘preference votes’ hardly matter in the Dutch case (Gallagher & Mitchell, 2005, p. 18). Based on this, Dutch MPs were expected to be much more national and party focused due to this lack of formal bond.
10
Operationalization
The dependent variable was the ratio between local-orientated parliamentary questions and national-orientated parliamentary questions asked by an MP. This dependent variable ranged from 0 to 10, by which 0 means that all the questions by the MP were national and by which 10 means that all the questions were focused. A parliamentary question was considered local-orientated when it at least contained one of the following characteristics (Baumgartner et al., 2006):
a. Did the member mention a geographic constituency specifically?
b. Did the member mention a geographical location that the coder can confirm is within the geographical constituency of the member?
c. Did the member mention a constituent or a particular case surrounding an individual, reasonably assumed to be a constituent?
d. Did the member mention a particular building or facility that the coder can confirm to be located in the geographical constituency of the member?
e. Did the member mention a particular organization or business that the coder can confirm to be located in the geographical constituency of the member?
f. Did the member mention an event, such as a local festival, specifically taking place in the geographical constituency of the member?
In addition, I supplemented this list with the following characteristic, because Dutch MPs’ PQs often link to news articles which contain local-orientated information. Without characteristic
g, some sets of PQ would have been coded as not-local, whilst the subject and aim of the set
would actually have been local-orientated.
g. Did the member add a news article to the set of PQs that mentions the above characteristics?
Dutch MPs submit their written parliamentary questions that concern the same topic in one document. For instance, an MP might ask several questions about a local highway and submit all these questions in one document and then sent this document to a minister. Because the PQs in such a document share the same subject, a set of parliamentary questions equaled one parliamentary question for this variable. The (sets of) parliamentary questions that were analyzed were the last 10 sets of PQs (until the 28th of November 2019) the MPs has asked to the ministers. These sets were retrieved from tweedekamer.nl and in total 1419 sets of
11 Place on the party list measured the overall political uncertainty (H1), because the lower on the list means the higher level of electoral uncertainty and vice versa. To account for the fact that some party lists are much longer than others, this variable was measured in relative terms. The following sum was used to equalize the party lists: (place on list/amount of places)*10. The outcome ranged from 0 to 10, by which 0 means the MPs is at the top of the list and 10 means he is at the bottom of the list. The data for this variable were retrieved from kiesraad.nl.
The seniority variable (H2) was measured on the basis of a member’s amount of terms he has served in parliament. The data for this variable were retrieved from tweedekamer.nl.
On the basis of an MP’s pre-parliamentary political career, local political ties (H3) was measured. An MP has political ties with his residential region, when he has held local political office in this region. This the case when the MPs has been a local councilmember (raadslid), provincial council member (statenlid) or when he has been member of the local executive (wethouder or burgemeester) or provincial executive (gedeputeerde or commissaris van de
Koning) in his residential region. This variable was either 0 or 1, by which 0 means the MP
has no formal local political office held and 1 means he has. Data on MPs’ local political ties were retrieved from parlement.com and/or tweedekamer.nl.
Local biographical ties (H4) was measured on the basis of the birthplace of an MP. The MP was ranked 1 when he was born in his residential region/province or 0 when he was not born in his residential region/province. The data for this variable were retrieved from tweedekamer.nl, where short MPs’ biographies can be found.
A member’s party was also obtained from tweedekamer.nl. For the government-opposition variable (H6), a member of a government-party was member of either Democraten
66, Christen-Democratisch Appèl, Volkspartij voor Vrijheid en Democratie or ChristenUnie.
These MPs were coded 1 and MPs from the opposition parties were coded 0. MPs who have left a government-fraction were marked as member of the opposition.
The regional-focus of an MP’s party (H7) was based on data from the Manifesto Project. This research group derives policy positions of individual parties through the coding of party manifestos. These positions are grouped in several topics, such as regionalism. A party is more regional-focused when its manifesto is coded positive about decentralization, such as support for local and regional customs, mentions of the subsidiarity-principle and mentions of special favors for sub-national areas (variable 301; Werner et al., 2014, p. 18). For instance, the CDA was coded 3,18 because its manifesto was relatively positive about regionalism whilst the Forum for Democracy-manifesto was coded 0 because it did not
12 mention anything positive about regionalism at all. The data for this variable came directly from the Manifesto Project dataset, which has also coded the manifestos from the Dutch 2017-election. Though, this data had limitations. Manifesto-data only give a minor indication of MPs’ stances on regionalism. That is, this data measures party manifestos and not direct party-fractions’ or individual MPs’ stances on this topic. However due to a lack of data of the latter two, aggregate manifesto-data were used.
For the gender variable (H8), a member was coded as 0 when the MP was male and 1 when the MP was female. The data for this variable were retrieved from tweedekamer.nl.
One control variable was added. This concerns the absolute size of the party-fraction. Existing research has found a significant positive effect of the size of fraction upon the severity of party discipline (in the case of the European Parliament, see Bailer, 2011). Therefore, it was possible that this would incentivize MPs from bigger fractions to ask local-orientated PQs because a PQ is an instrument which is relatively unconstrained by party leaders (Aléman et al., 2018, p. 228). However, because this relationship had never been researched, this variable was best suited as a control variable. The up-to-date sizes of party fractions was obtained from tweedekamer.nl.
Results
To test the hypotheses and to answer the research question, one statistical analysis was conducted. On first glance, an OLS regression analysis was the most suited method to analyze the data. The reason for this was the substantive variety of variables (8 variables) and due to the fact that the dependent variable was a continuous variable (Field, 2009, p. 210, 220). However, the data have to meet several assumptions of this model in order to deduce valid conclusions and to make generalizations from the outcome (Field, 2009, p. 251). The data that were collected, did not meet some of these assumptions. One assumption that was not met, was the assumption of homoscedasticity: this means that the residuals of the predictors should have the same variance on each level. This posed limitations to our model and its possibility of generalizations. That is, when the outcome(s) of the analysis is significant, it cannot be generalized beyond the sample (Field, 2009, p. 251). Not only did the data not meet this assumption, neither was the dependent variable suitable enough for this model due to its rather constrained variability. The dependent variable is constrained when the data collected only varies between a limited range, whilst the overall measure ranges more substantively. In my data, the overall measure runs from 0 to 10. However, as can be seen in table 1 below in which a summary of the collected data is displayed, the data of the dependent variable only range from
13 0 to 4 which made the data bounded. Therefore, the collected data did not meet Field’s assumption of variable type (Field, 2018, p. 388).
Table 1. Descriptive statistics
N Minimum Maximum Mean Standard Deviation Ratio local-orientated PQ’s 149 0 4 0,611 0,867
Ratio place on party list 149 0,13 9,78 2,411 1,613
Seniority 149 1 7 1,87 1,259
Local political ties 149 0 1 0,51 0,502 Local biographical ties 149 0 1 0,42 0,496 Periphery or center 149 0 1 0,42 0,495 Government or opposition 149 0 1 0,50 0,502 Party stance regionalism 149 0 3,18 1,92 1,089
Gender 149 0 1 0,34 0,474
Fraction size 149 1 32 17,99 9,159
Due to this, it was more suited for my analysis to convert the outcome variable from a continuous variable to a dichotomous variable, by which 0 means the MP has not asked any local-orientated PQ and 1 means the MP has asked 1 or more local-orientated PQs. When the outcome variable is dichotomous, the most suitable analysis is a binary logistic regression analysis (Field, 2018, p. 879). For the analysis, the most appropriate method of entering the predictors into the model was through hierarchal regression (Field, 2009, p. 212). This method is best when some predictors are more important than others in predicting the dependent variable based on the existing literature. Amongst several case, some variables were repeatedly found to have a significant effect upon representational foci, these were local biographical ties, local political ties and whether the MP lives in the center or in the periphery. The first model contained these most important variables, the second model added the less important ones plus the new control variable. The outcome of the model can be found below in table 2.
The first model which contained the three most important variables can predict only a small percentage of the variation of the dependent variable: 3,9 percent measured by Cox and Snell’s R2 and 5,2 percent measured by the Nagelkerke R2. This means that the predictors are
14 close to being useless at predicting the outcome variable (Field, 2018, p. 882). The R2 figures
in the second model are a bit higher, but based on these, the model is still very limited in its ability of predicting the variation of whether an MP has asked a local-orientated PQ or not. Table 2. Logistic regression analysis of whether an MP has asked a local PQ
Model 1 Model 2
(Constant) - 0,602 - 1,453
(0,285) (0,749)
Local political ties - 0,078 - 0,117
(0,347) (0,358)
Local biographical ties 0,688* 0,684
(0,347) (0,354) Periphery or center 0,340 0,344 (0,354) (0,365) Place on list 0,066 (0,135) Seniority - 0,001 (0,150) Government or opposition - 0,650 (0,602)
Party stance on regionalism 0,243
(0,215) Gender 0,155 (0,364) Fraction size 0,028 (0,030) -2LL 199,165 196,707
Cox and Snell’s R2 0,039 0,054
Nagelkerke R2 0,052 0,073
N 149 149
Note: binary logistic regression coefficients with standard errors in parentheses. ***p < 0,001, **p < 0,01, * p < 0,05
The first two variables (local political ties and local biographical ties) have been put in both models. The fact that an MP has held local political office has in both models a negative but insignificant prediction of whether an MP has asked a local-orientated PQ. The data show that the amount of MPs that have asked a local-orientated PQ or not, is almost evenly distributed among this variable: 32 MPs that have no political ties to their region, have asked a local PQ and 35 MPs that have political ties, have asked a local PQ. Besides, the amount of MPs that have asked no local PQ is the same for both groups: 41 MPs that have no political ties and 41
15 MPs who do. The fact that this factor has no significant effect upon the dependent variable is rather a surprise when looking at the substantive evidence the literature gives for this effect in other cases. The logic here is that MPs who have been politically active in their province, must in some way be much more informed local matters and thus more inclined to ask PQs about such matters (Fiva & Halse, 2016, p. 25).
Whether the MP was born in his residential province was, however, a significant, positive predictor of whether the MP has asked a local-orientated question in the first model. Meaning that MPs who were born in their province ask significantly more local-orientated PQs than MPs who were not born in the province where they currently are living. A possible reason for this is that when an MP represents a region which he/she has biographical ties with, this might result in a greater motivation to ask PQs about matters in this region (Zittel et al., 2019, p. 689). This is significant possibly because the data show that 52 MPs who had no biographical ties, have asked no local PQ whilst 28 MPs who do have biographical ties, have asked no PQ either. Still, the amount of MPs that do have asked a local PQ, is almost equally distributed amongst these two groups: 35 MPs that do and 32 that do not have biographical ties. Though, when the second model accounted for the other variables, the outcome was not significant anymore in predicting the outcome variable. Hence, to some extent, the data therefore support the 4th hypothesis.
A third and last variable that was found to be a significant factor amongst several cases (such as Italy and Ireland, see Russo, 2011; Martin, 2011b), was whether the MP resided in the peripheral or in the center party of the country. Again, surprisingly this variable does not predict the variation in the dependent variable in neither the first nor the second model. Looking more closely at the data, the results did indicate that more than half of the MPs from the periphery had at least asked a question about matters in their province whilst this figure is lower for MPs from the center (32 of 62 MPs from the periphery and 35 of 87 MPs from the center). Despite this, the outcome of the logistic regression analysis does not support the 5th hypothesis. This outcome was contrary to self-reported behavior by MPs (Thomassen & Andeweg, 2004). The percentage of MPs from the periphery that considered it very important to represent their region’s interests, was much higher than the MPs from the center (74 percent and 40 percent respectively, see Thomassen & Andeweg, 2004, p. 59). The data however did not support this divergence because the focus on the center and the periphery seem to be on an almost equal level (35 and 32 MPs have asked PQs about the center and periphery respectively).
The outcome did not support the 1st hypothesis either. When controlling for more factors in the second model, the factor of electoral uncertainty was no significant predictor of
16 whether an MP has asked a local-orientated PQ. Hence, the place on the party list does not seem to influence MPs to have a local focus or not. This outcome is less surprising than the former two. As discussed in the methodology-section, the Netherlands constituted a least likely case because its political system is considered to be more party-centered than candidate-centered. One reason for this is that Dutch MPs are mainly dependent on the party-vote rather than on personal votes in order to get re-elected because of the fact that preference votes hardly matter when electing the Dutch lower house (Gallagher & Mitchell, 2005, p. 18). Therefore, party vote- seeking may triumph personal vote-seeking such as through gaining ground in a certain region. Even an MP who was fully dependent on personal votes (Maurits von Martels, whose place on the party-list was 44 of 45) did not excel in asking PQs about his own province where he was born and has held local political office (only 1 out of 10 PQs was local-orientated)
Despite the fact that more senior MPs are more probable to be put higher on the list than they were earlier on and hence more secure that the party vote reaches their place on the list, seniority was not a significant predictor of the dependent variable’s variation. In table 2 it is clear that more senior MPs do not seem to ask fewer local-orientated PQs.
Based on the literature, it was dubious whether MPs from the government or MPs from the opposition are more likely to ask local-orientated PQs. The data support however neither of them due to insignificant outcomes. The data indicate that the amount of MPs that has asked a local-orientated PQ is roughly the same for both groups: 33 MPs from the opposition and 34 MPs who were member of a government-party.
Some party manifestoes indicate that some parties are much more regional-orientated. This was, however, no significant predictor of whether an MP has asked a local-orientated question or not. In a way, this was a surprise. Based on the Manifesto Project, the CDA is one of the most regional-orientated Dutch parties and this is publicly known. However, all of the MPs who asked 4 local-orientated PQs were from the VVD which is only to a small extent coded positive about regionalism. Hence, though some parties claim to be regional-orientated, this does not come forward when inspecting the content of their MPs’ parliamentary questions. The two variables which were least expected to be solid predictors of the variation in the dependent variable were gender and the size of the fraction. None of these two were a significant predictor of whether an MP has asked a local-orientated question either. Though, as mentioned above, all the MPs who asked four local PQs were from the biggest fraction in the
17
Conclusion
The goal of this thesis was to try to explain what factors influence and can predict why MPs ask parliamentary questions about matters in their residential province or not. Therefore, the research question ‘what explains the variation of local-orientated parliamentary questions?’ was formulated and aimed to answer. In this research, the data consisted of 1419 sets of parliamentary questions and 8 independent variables have been tried to explain the variation in the dependent variable. This data were eventually analyzed through a logistic regression analysis and the outcome of the two run models was that only one variable was a significant predictor of the dependent variable.
The case that was selected was the Dutch case, this case was chosen because of the fact that it constituted a least likely case: the Netherlands has no electoral districts and has a party-centered political system. The fact that the data did not support 7 of the 8, confirms that the Dutch case is indeed a least likely case: not many MPs seem to focus much on their residential region. Furthermore, whether an MP has asked a local-orientated PQ or not seems to be an almost completely random process. Except for the factor of whether an MP was born in his residential region; this can significantly predict whether an MP has asked a local-orientated PQ.
The overall implication of this result stems from the data, which indicate that MPs almost rarely ask PQs about their residential province. Almost half of the MPs had not asked such a PQ at all. This indicates that geographical representation is almost absent in the case of the Netherlands whilst the tool of PQ is generally one of the most accessible means to represent a certain group. The outcome therefore does not support the conclusion that stems from MPs’ own self-reported behavior that geographic representation is ‘anything but absent’ in the Netherlands (Thomassen & Andeweg, 2004, p. 65). Scholars have frequently assumed that one downside of proportional representation systems is that they do not provide sufficient geographical representation (Latner & McGann, 2004). Therefore, this implication can be taken into account in discussions about reforming the Dutch electoral-system, which has become hot topic in recent year thanks to the Remkes-committee that sought to investigate the Dutch state structure and its possible reform. Besides, one major determinant of the quality of representation is MPs’ responsiveness to their constituents (Van de Bovenkamp & Vollaard, 2019), the lack of geographic representation can therefore be worrisome in times when provinces face distinct problems such as earthquakes and dangerous drug waste.
Future research is useful when it comes to analyzing other means which can be used for geographic representation, such as social media or motions, in order to portray a full picture of geographical representation in the Netherlands and its mechanisms. New survey-data from MPs
18 may not be helpful enough to grasp geographic representation because the results are completely at odds with MPs’ own self-reported behavior. The main limitation of this thesis is therefore that PQs only portray a part of the status of geographical representation in the Netherlands; perhaps geographic representation is indeed ‘anything but absent’ when looking at other (extra-)parliamentary means. Likewise, the fact that each MP’s ten last sets of PQs were analyzed, may have distorted the data, because some major societal events were granted great attention in some MPs’ subsequent sets of PQs. Thus for some MPs, the data might have strongly deviated from his/her regular parliamentary behavior and hence random sampling of PQs would have been more appropriate.
19
Literature list
Alemán, E., Micozzi, J. P., & Ramírez, M. M. (2018). The hidden electoral connection: analysing parliamentary questions in the Chilean Congress. The Journal of Legislative
Studies, 24(2), 227-244.
André, A., Depauw, S., & Martin, S. (2016). The classification of electoral systems: Bringing legislators back in. Electoral studies, 42, 42-53.
Bailer, S. (2011, June). Parliamentary party group discipline in comparison. In 1st Annual
General Conference of the European Political Science Association, Dublin (Vol. 16).
Bailer, S., & Ohmura, T. (2018). Exploring, Maintaining, and Disengaging—The Three Phases of a Legislator's Life. Legislative Studies Quarterly, 43(3), 493-520.
Baumgartner, F. R., Green-Pedersen, C., & Jones, B. D. (2006). Comparative studies of policy agendas. Journal of European Public Policy, 13(7), 959-974.
Blanchard, P. (2017, March 31). Why the logic of selecting least-likely cases to make stronger
generalizations makes little sense in case study research. Retrieved from
http://ecpr- methods.org/2017/03/why-the-logic-of-selecting-least-likely-cases-to-make-stronger- generalizations-makes-little-sense-in-case-study-research-by-derek-beach-university-of-aarrhus/
Bovenkamp, H. M. V. D., & Vollaard, H. (2019). Strengthening the local representative system: the importance of electoral and non-electoral representation. Local Government
Studies, 45(2), 196-218.
Bird, K. (2005). Gendering parliamentary questions. The British Journal of Politics and
International Relations, 7(3), 353-370.
Bryman, A. (2016). Social research methods. Oxford university press.
Burke, E. (1777). A Letter to the Sheriffs of Bristol on the Affairs of America. The Writings
and Speeches of Edmund Burke, 3, 1774-1780.
Eulau, H. (1962). The legislator as representative: Representational roles. The Legislative
System: Explorations in Legislative Behavior (pp. 267-287). New York: John Wiley & Sons, Inc.
20
Field, A. (2018). Discovering Statistics Using IBM SPSS Statistics, Fifth Edition.
Fiva, J. H., & Halse, A. H. (2016). Local favoritism in at-large proportional representation systems. Journal of Public Economics, 143, 15-26.
Gallagher, M., & Mitchell, P. (Eds.). (2005). The politics of electoral systems. OUP Oxford. Grofman, B. (2005). Comparisons among electoral systems: Distinguishing between localism
and candidate-centered politics. Electoral Studies, 24(4), 735-740.
Heitshusen, V., Young, G., & Wood, D. M. (2005). Electoral Context and MP Constituency Focus in Australia, Canada, Ireland, New Zealand, and the United Kingdom. American
Journal of Political Science, 49(1), 32-45.
Kellermann, M. (2016). Electoral vulnerability, constituency focus, and parliamentary questions in the House of Commons. The British Journal of Politics and International
Relations, 18(1), 90-106.
Latner, M., & McGann, A. (2005). Geographical representation under proportional representation: The cases of Israel and the Netherlands. Electoral studies, 24(4), 709-734.
Levy, J. S. (2008). Case studies: Types, designs, and logics of inference. Conflict
management and peace science, 25(1), 1-18.
Martin, S. (2011a). Parliamentary questions, the behaviour of legislators, and the function of legislatures: An introduction. The Journal of Legislative Studies, 17(3), 259-270. Martin, S. (2011b). Using parliamentary questions to measure constituency focus: An
application to the Irish case. Political studies, 59(2), 472-488.
Russo, F. (2011). The constituency as a focus of representation: Studying the Italian case through the analysis of parliamentary questions. The Journal of Legislative
Studies, 17(3), 290-301.
Seawright, J., & Gerring, J. (2008). Case selection techniques in case study research: A menu of qualitative and quantitative options. Political research quarterly, 61(2), 294-308. Swanborn, P. (2010). Case study research: What, why and how?. Sage.
21
Thomassen, J., & Andeweg, R. B. (2004). Beyond collective representation: individual members of parliament and interest representation in the Netherlands. The Journal of
Legislative Studies, 10(4), 47-69.
Von Beyme, K. (2000). Parliamentary democracy: Democratization, destabilization,
reconsolidation, 1789-1999. Springer.
Werner, A., Lacewell, O., & Volkens, A. (2014). Manifesto Coding Instructions, March 2014. Retrieved from https://manifestoproject. wzb.
eu/down/papers/handbook_2014_version_5. pdf,
Zittel, T., Nyhuis, D., & Baumann, M. (2019). Geographic Representation in Party‐ Dominated Legislatures: A Quantitative Text Analysis of Parliamentary Questions in the German Bundestag. Legislative Studies Quarterly.
