Local Political Decision-Making: A Case of Rationality or Appropriateness?

Abstract

The party and trustee principles are two representational styles used to describe how politicians make decisions. Swedish politicians have historically relied more on the party principle than the trustee principle. This article studies the decision-making practices of local Swedish politicians by exploring to what extent they rely on these principles when making decisions on two issues that diverge in political dignity: tax level and organisational change in the municipal administration. The study draws on new institutional theory, in which theories from rational choice and sociological institutionalism were used for modelling and performing a large study. The results indicate that, although Swedish politicians still rely on the party principle when making decisions, there is a significant difference as to what extent they do so in regard to the two policy issues. The trustee principle is more frequently used when deciding on organisational change than on tax levels. This result is valid for all Swedish parties, except for a relatively new political party at the extreme right of the ideological spectrum. Future research of the two decision-making principles in relation to other policy issues, as well as research that delves deeper into the deviant results of the different political parties, is encouraged.

Key Words:

• Party principle
• trustee principle
• decision-making
• new institutionalism

Introduction

Who do elected politicians follow when making political decisions – the party or their own preferences? The answer to this question varies between countries and decision-making institutions and sometimes also between policy issues. When it comes to Sweden, politicians have been prone to follow the party principle in parliamentary voting, rather than the trustee principle (Esiassion and Holmberg 1996; Mattson 2005).

The party principle is related to a representation style in which the elected politicians’ loyalty is primarily to the party he or she represents, rather than to the voters or their own beliefs. Following the party principle thus indicates that the politician follows the party standpoint when making decisions regarding different policy issues. The trustee principle, on the other hand, is related to a representational style in which the politician is seen as a trusted elected capable of making the right decisions according to own personal beliefs and preferences (Bäck, 2000).¹

Institutional changes in the Swedish system may suggest a weakened loyalty to the party, but the strength of the party principle in Sweden remains (Mattson 2005). For example, it has been argued that the increase in personal voting would decrease the loyalty to the party when making decisions (see Skoog 2011), but previous research shows that is not the case (see Nielsen 2007). On the contrary, the party cohesion in decision-making in Sweden has increased. A significant factor that has been highlighted to explain this culture is the collective benefit of portraying the picture of a cohesive party to the voters. A unified party, with consistent voting in line with the party standpoint, offers transparency for the voters by providing them with information on how the party and its members of parliament (MPs) act if elected. Party cohesion, rather than fragmentation, increases an individual MP to remain in parliament (Mattson 2005).

There is, however, significant variation in how and when an MP follows the party and trustee principles. First, there is a difference among the Swedish parties in which MPs in parties to the left (the Left Party and the Social Democratic Party) and the Conservative Party tend to follow the party principle more than MPs in other political parties (Esiassion and Holmberg 1996), which might be explained by differences in party cultures (see Barrling-Hermansson 2004). Second, Swedish politicians tend to rely on the party principle more than the trustee principle in situations related to crucial policy issues, such as taxation, and especially so on the local level (Karlsson 2011; Bäck and Larsson 2008).

Empirical studies on voting behaviour reveal that Swedish politicians primarily make decisions in line with the party standpoint rather than the trustee principle (Karlsson and Gilljam 2014a). Even so, when asked about their decision-making, Swedish politicians, at the local level, refer to both principles. There is thus a discrepancy between how the politicians state that they want to conduct their political work and how they actually do. To understand this paradox, one must explore, explain and compare the compliance to the party and trustee principles further. In particular, one needs to examine how the party and trustee principles play out in different decision-making contexts regarding different policy issues.

New institutional theories have proven to successfully explain political decision-making (see e.g. Kiewet and McCubbins 1991; Lewis 2005; Pollack 2007), and some of these can be connected to the party and trustee principles in their explanations (see e.g. Karlsson 2012).

Sociological institutional theories suggest that individuals, such as politicians, act in accordance with what is appropriate in their roles as decision makers rather than their own personal beliefs (see March and Olsen 1984, 1989). Such explanations can, at least in the Swedish case, be linked to the party principle. Decision-making is explained as a product of politicians doing what they consider as appropriate; that is, to do as the party says.

Rational Choice Institutionalism (RCI), on the other hand, perceives individuals as rational decision makers who act according to their own preferences when making these decisions, except when the sanctions for doing so are too high (see e.g. North 1990; Hindmoor 2006). The perspective can thus be linked to the trustee principle and the understanding of the representative making accountable decisions in line with the personal opinion, rather than someone else’s opinion. However, when the sanctions for following the trustee principle are high, it is rational for the individuals (i.e. the politicians) to abandon the trustee principle and instead follow the party principle due to high personal costs (such as conflicts within the party, exclusion, etc.).

Institutions are structures that affect the behavior of actors in collective units (see Goldmann, Pedersen, and Østerud 1997; Åström 2011). This definition relates to Peter’s (2012) characterisation of institutions as a structural feature of society, being relatively stable over time and affecting individual behaviour. The RCI and Sociological Institutionalism (SI) have different positions on the scale between agency and structure. RCI emphasises agency more than SI does, but they both do share the overall structure focus that is prominent in new institutionalism in general. Following this, the party principle and trustee principle can be perceived as two competing institutions that are closely related to the notions proposed by RCI and SI, respectively. This makes the case for political decision-making appropriate for the purpose of applying and testing institutional theories.

The aim of this article is to explain local Swedish politicians’ decision-making by exploring to what extent they rely on the party principle and the trustee principle when making decisions on two issues that diverge in political dignity: tax levels/activity fees and organisational change in the municipal administration. Tax levels/activity fees represent a policy issue characterised as crucial, whereas an organisational change in the municipal administration is not (see discussion further down). Furthermore, the article also has the theoretical aim of showing both how institutional theories, in particular RCI and SI, can be used as a theoretical foundation of the two principles and to test the explanations of these theories against each other in this decision-making case.

A large N study of politicians at the local level in Sweden, with surveys sent out to all Swedish politicians on this level, was conducted to answer two research questions: (1) Is there a difference in how politicians follow the party and the trustee principles when deciding on tax levels/activity fees and organisational change, respectively? This question explores and tests whether previous assumptions about local politicians being more prone to follow the party when deciding on important policy issues are supported. Furthermore, the empirical study tests and validates the institutional theories’ predictions in this particular decision-making case.

Previous research has, as mentioned above, shown that MPs belonging to the Left, the Social Democratic and the Conservative Parties follow the party principle more than politicians in other Swedish parties. Therefore, the second question reads thusly: (2) Is there a difference between parties in how politicians follow the party and trustee principles when deciding on tax levels/activity fees and organisational change, respectively?

This large empirical material (N = 8,523) enables us to explore the results of previous studies on an extensive empirical material on elected politicians at the local level in Sweden. Explaining and comparing the party and trustee principles in regard to the two policy issues will contribute to our understanding of how these important principles (or rules in the terminology of new institutionalism) are applied in different situations and enable us to assess the results of previous studies mentioned above. The application of the new institutional theories will contribute to the understanding of how these theories can be used to explain different decision-making phenomena. Thus, the study will provide both empirical and theoretical contributions.

The next section outlines how new institutional theory can be used to explain the representation principles. Here also is included the operationalisation regarding decision-making hypotheses. This is then followed by the methods section, in which methodological considerations on how to test the decision-making hypotheses are presented, mainly in the form of non-parametric variance analysis. The results are finally presented and discussed in relation to the two aforementioned research questions in the article’s concluding section.

A new institutional framework for explaining the representation principles

There are several new institutional approaches that differ in the way they view the social world (Peters 2012). Two of the more prominent of these new institutional approaches are RCI and SI.² RCI models its explanations of institutional phenomena from the perspective of rational actors who are goal oriented and follow their personal preferences in order to obtain utility when facing a decision-making situation (North 1990). SI, on the other hand, explains decision-making by emphasising the cognitive way actors internalise their institutional surroundings and ascribe their meaning, which becomes the foundation for their actions (DiMaggio and Powell 1991).

Both of these approaches have generated theories that successfully explain the decision-making of politicians in legislature, such as the theories used by Strøm, Müller, and Bergman (2003) in their study of delegation and accountability in parliamentary democracies from an RCI perspective, as well as Goetze and Rittberger’s (2010) sociological analysis of the European Parliament. While these studies differ from each other in how they explain what drives human decision-making and action, as described briefly above, they also differ, as Peters (2012) points out, with regard to one fundamental issue in institutional theory: how actors and institutions relate to each other.

RCI sees institutions as exogenous to the actors. This means that institutions do not shape the preferences of the actors but merely provide them with incentives to a behaviour that will result in a certain utility within a specific institutional setting. Institutions serve as structural guidelines for the actor to act upon, but the preferences are not changed and thus remain rather intact. SI, on the other hand, sees institutions as endogenous to the actors, which means that the actor internalises the institutions and the meaning ascribed to them. This results in preferences shaped not only by the actor’s values and beliefs but also by institutions (Peters 2012).

The theoretical roots of RCI can be found in rational choice theory (RCT). According to Monroe (2001), the traditional economic concept of rationality found in RCT and the bounded rationality³ found in RCI share the following four fundamental assumptions: (1) individual action, (2) goal-orientation, (3) conscious choice and (4) self-interest. Individual action draws on methodological individualism; that is, the claim that social scientific explanations must contain statements of individual actions. Societies or groups cannot act as such on their own without consideration for and analysis of the individuals who form these societies and groups (Rothstein 2003).

Goal-orientation refers to the individual’s knowledge of her/his goals and the according act to fulfil them (Riker 1995). Conscious choice refers to the assumptions upon which individuals make decisions based on obtained information and in line with their own preferences. This eliminates choice based on emotions, frustration, routine or any other non-rational factor (Johnston 1991). Self-interest indicates that the individual knows what his/her preferences are and behaves consistently with these preferences. Hence, the individual strives to obtain utility based on these preferences (Ostrom and Ostrom 1971).

According to the RCI perspective, institutions are structures that lower the transaction costs for these rational actors to reach agreements and thus get the most out of their own maximising of utility in the long run. Institutions ease the process of making consistent decisions and therefore also maximise personal utility. They provide decision-making individuals with information on available alternatives (however, this does not mean all possible alternatives) and incentive structures for the individual preferences that are suitable to maximise the particular institutional surroundings (North 1990).

If we rely on the RCI we should predict that individuals act in such way that they benefit themselves, to gain some kind of reward or to avoid sanction, depending on the incentives provided by the institution. If the preferences of the individual decision maker conflict with the constraints of the institution, we should expect the individual to make decisions in accordance with the individual preference when the cost of doing so is low. Furthermore, the individual should be expected to make decisions in line with the incentives provided by the institution when the cost of following the individual preference is high. When the cost for following preferences is low, individuals make decisions in line with their individual preferences, and when the cost for following preferences is high, individuals make decisions in line with the incentives provided by the institution.

The most prominent feature of SI, on the other hand, is the role culture is ascribed when explaining and understanding institutions; it places no focus on the rational calculations of individuals. Instead, SI highlights that institutional forms and procedures are culture-specific practices, which are taken for granted and provide the individuals with meaning. The individuals assimilate these culture-specific practices through processes that are not necessarily means-end efficient (as stated by RCI) but culturally bound. The way institutions affect decision-making behaviour is primarily cognitively explained through SI. Automatised schemas, scripts, routines and classifications of the individual internalise the institutional surroundings (DiMaggio and Powell 1991).

Individuals act not upon their self-interests, but on what they perceive as being appropriate in the role they see themselves having. March and Olsen (1984, 1989) addressed this as part of the logic of appropriateness. When making decisions, the individual tries to pair a situation with the action that is perceived as eligible in their role. What that role is, as well as what constitutes appropriate action for that role, is formed by institutions and is socialised by the individual (March and Olsen 1989). Because of this logic, the impact of rules, and especially norms, are important factors in understanding the interaction between institutions and individuals. Norms in terms of decision-making describe how the individual acts according to what’s perceived to be appropriate in her/his role, regardless of personal preferences.

Since several types of different norms may influence behaviour (see Kashima and Lewis 2000), further clarification of norms is necessary. In this article, I refer to norms as social norms (cf. personal norms, see Schwartz and Howard 1980) and as such also as injunctive social norms. Injunctive norms are ‘perceptions concerning what should or ought to be done with respect to performing a given behavior’ (Fishbein and Ajzen 2010, 131). Individuals make decisions according to the institutional rules that have been internalised as appropriate for the specific role they identify themselves as having.

How do RCI and SI assumptions on how institutions affect individual behaviour differ in their predictions when applied to decision-making in the case of Swedish politicians? The answer to this question can be found in whether the institution of party pressure, following the party principle when making decisions, is internalised or not. Before this is outlined, the two policy questions have to be examined further.

The two policy issues differ in the way they are regulated by Swedish political parties, and thus the pressure for the local politicians to conform to party standpoint varies. Tax levels/activity fees are correlated with party positions on the economic left-right scale and a policy issue where Swedish parties have strongly stipulated positions (Pettersson-Lidbom 2008). The left-right position is thus strong in Sweden and, as such, is one of the strongest variables in predicting which political party Swedish citizens vote for in elections (Holmberg 2002). Hence, the party pressure on the politician belonging to the party is strong and so also is the pressure to follow the party principle rather than the trustee principle when making decisions in these matters.

Organisational change in the municipal administration, on the other hand, as well as political organisation in general is not an issue that correlates with party position on the economic left-right scale and does not define the left-right position cleavage between Swedish political parties (see Holmberg 2002; Karlsson 2010). Furthermore, the parties themselves do not hold clear and consistent positions in this policy area across the different municipals and regions in Sweden (see e.g. Lidström 2009). Hence, the Swedish parties do not as strongly regulate this policy issue and the pressure to conform to the party standpoint when making decisions in these matters is thus substantially weaker than in the case of tax levels/activity fees.

If the politicians do not internalise the institutional rules and if institutions are entities exogenous to these actors, as assumed in RCI, we should expect the degree to which the politicians follow the party principle to create more variation. When applying RCI logic, Swedish politicians should only do as the party says when there is either a high benefit for following the party standpoint or a high cost for not following the party standpoint (cf. rational arguments in Hindmoor 2006).

Consequently, there should be a difference in how the politicians act when making decisions about tax levels/activity fees, since this policy issue is more regulated by the party. The act of following the trustee principle, when it conflicts with the party standpoint, would therefore also be more sanctioned than in the situation of making decisions regarding organisational change in the municipal administration. Based on this, the following hypothesis emerges:

Hypothesis 1: At times when the party standpoint conflicts with their own preferences, Swedish politicians follow the party principle more when they make decisions on tax levels/activity fees than in decisions on organisational change in the municipal administration.

The role of Swedish politicians is to follow the party principle rather than the trustee principle, and they do so more than politicians in other countries (Barrling-Hermansson 2004). If we apply SI logic, we should expect that Swedish politicians should internalise the institution of the party principle (a social norm) rather than the trustee principle. If the politicians internalise the informal institution of the party principle, as predicted when applying SI assumptions of appropriateness, there should be no significant difference in how these politicians act when making decisions in the two policy issues. They should follow the party principle, even when it conflicts with their own personal views and even when the pressure to conform to party views is low. The role of Swedish politicians is to follow the party principle (as pointed out by Barrling-Hermansson 2004) and doing what is appropriate to follow this social norm.

In situations in which there is no official party standpoint, the politicians should also turn to the party for advice, if we follow SI assumptions. These situations have a high degree of uncertainty, and as DiMaggio and Powell (1983, 1991) have pointed out, times of uncertainty result in mimicking behaviour in which individuals and organisations do as those who are important and successful others in the same field do. In the case of Swedish politicians at the local level, we can predict that, if we follow the SI assumptions, politicians turn to the party when making decisions on questions characterised by uncertainty, since they mimic the behaviour of important others (i.e. their party colleagues). When related to decision-making in the two policy issues examined here, the following hypothesis can thus be drawn from the SI explanations above:

Hypothesis 2: At times when the party standpoint conflicts with their own preferences, Swedish politicians follow the party principle to the same extent when they make decisions on tax levels/activity fees as in decisions on an organisational change in the municipal administration.

There is a debate in SI regarding whether it is possible to make predictive claims. I agree with the modelling of, for example, Villadsen, Rosenberg Hansen, and Mols (2010) that such predictive claims can be made. From these two hypotheses we can see that the assumptions of RCI and SI lead to different predictions when applied to the chosen case. A summary of the decision-making predictions drawn from RCI and SI is presented in Table 1.

Table 1. Decision-making predictions for behavioural assumptions of self-interest and social norms in relation to the party and the trustee principles

		Decision-making predictions
		Self-interest (RCI)	Social norms (SI)
Party pressure	High	Follow party principle	Follow party principle
	Low	Follow trustee principle	Follow party principle

Methods

A questionnaire methodology was used to collect data. The questionnaire⁴ was sent out to all 9,725 Swedish politicians at the local level; that is to every member of a municipal council (in Swedish: kommunfullmäktige) in Sweden. The number of respondents that answered both questions in the questionnaire was 8,523 out of 9,725, which equals a response percentage of about 88%.

Constructs of self-interest and norms were operationalised by relating the questions of decision-making to the perceived party pressure of the respondents. Here, answers through which respondents report that they will follow the party principle in a policy issue, even though they personally don’t agree with this position, were calculated as conforming to party pressure norms. Likewise, answers through which respondents report that they will make decisions according to the trustee principle, even when their opinion conflicts with the party’s position in a policy issue, were calculated as self-interest. In the questionnaire, these two alternatives were operationalised as opposite extremes on the same scale for the respective questions of organisational change in the municipal administration and municipal tax levels/activity fees. The scale was a four-grade ordinal scale, and the respondents were informed to report which of the two statements they agree with more when asked how they make decisions in a policy issue, as shown in the Appendix.

Frequencies were calculated to determine to what extent the Swedish politicians reported that they make decisions according to the party in these matters. To see if there was a significant difference, variance analysis was conducted in the form of a non-parametric pair-sampled test, the Wilcoxon Signed Rank Test (see Greene and D’Oliveira 1999). We used the Wilcoxon test in the light of the ordinal quality as well as the range of only four grades of the questionnaire. The results from these calculations were then compared to the hypotheses of the study.

To explore the second research question, whether there is a difference among political parties in how their representatives relate to the party and the trustee principles, the sample was split by party affiliation, and the frequencies and the variance analysis were performed on this split sample.

At this point, a few things are important to consider. First, the issue of whether there is a difference between the decision-making in the two policy matters is not the only important result. Since the scale is constructed as presenting opposite extremes, with self-interest on the lower end and norms on the higher end of the scale, it is also important to notice and analyse what the respondents answered. By doing so and examining the distribution of the answers, we can see to what degree the politicians tend to follow the party principle (their party principle) or the trustee principle (the social norm) when making decisions, something that is not possible by only looking at differences between the two conditions in the Wilcoxon test.

Second, there was a possibility that the results would support neither RCI nor SI because of the scale. A result in that direction would indicate that neither of the presented theories may be suitable for explaining this decision-making or that the party pressure would be unimportant when making these decisions.

A valid point to discuss is that the questionnaire only measured politicians’ reports on their decision-making behaviour. Self-reported behaviour can be problematic, since discrepancies may arise between how respondents describe their actions and how they actually act. Still, reported behaviour and actual behaviour have proven to correlate when asking about more specific behaviour, and not only general behaviour (see Fishbein and Ajzen 2010). The questions are directed towards two specific policy issues to strengthen the validity of findings.

Results

First, frequencies were calculated for both questions regarding the two policy issues in order to determine to what extent the Swedish politicians at the local level follow the party and the trustee principles. For the question on decision-making concerning organisational change in the municipal administration, 10.3% of the respondents answered that they totally agreed with statement A; that is, they follow the party’s standpoint when making these decisions in cases of conflict between party standpoint and personal preference. Furthermore, 44.1% answered that they partly agreed with this statement. On the other side of the spectrum, 36% of the respondents reported that they partly agree with statement B, that they follow their personal preference when making these decisions when there is a conflict between party standpoint and personal preference. Finally, 9.6% of the respondents answered that they totally agree with statement B.

For the question on decision-making concerning municipal tax levels/activity fees, 19.1% of the respondents answered that they totally agreed with statement A, that they follow the party’s standpoint when making these decisions when there is a conflict between party standpoint and personal preference. Furthermore, 43.6% answered that they partly agreed with this statement. On the other hand, 29.3% of the respondents reported that they partly agreed with statement B, that they followed their personal preference when making these decisions when there is a conflict between party standpoint and personal preference. Finally, 8% of the respondents answered that they totally agree with statement B. These results are shown, together with the frequencies for each policy issue, in Table 2.

Table 2. Frequencies of respondents’ answers about decision-making concerning organisational change in the municipal administration and municipal tax levels/activity fees

	Totally agree with A	Partly agree with A	Partly agree with B	Totally agree with B	n	Missing	Total
An organisational change in the municipal administration
Frequency	880	3,780	3,090	826	8,576	1,149	9,725
Per cent of respondents	10.3	44.1	36.0	9.6
Cumulative per cent	10.3	54.4	90.4	100.0
Tax levels/activity fees
Frequency	1,629	3,714	2,496	684	8,523	1,202	9,725
Per cent of respondents	19.1	43.6	29.3	8.0
Cumulative per cent	19.1	62.7	92.0	100.0

A Wilcoxon test was conducted to evaluate whether there is a significant difference in the way the Swedish politicians at the local level followed the party standpoint or their own preferences when making decisions about different issues. The results indicated a significant difference, z = −22.711, p < .001. The mean of the ranks in favour of following your own preferences, if it conflicts with the party standpoint when making decisions on organisational change in the municipal administration, was 1536.69, while the mean of the ranks in favour of following your own preferences, if it conflicts with the party standpoint, when making decisions on municipal tax levels and activity fees, was 1485.60.

The answers regarding decision-making in the two policy issues were also split by party affiliation to determine whether the results were coherent for all Swedish parties represented in the municipal councils that also have representation in the national parliament. Parties having representation in municipal councils but no representation in the national parliament were excluded from the analysis. The results are shown in Table 3.

Table 3. Frequencies of respondents’ answers, split by party affiliation, about decision-making in the policy issues of organisational change in the municipal administration and tax levels/activity fees

	Decision-making on organisational change							Decision-making on tax levels/activity fees
	Totally agree with A	Partly agree with A	Partly agree with B	Totally agree with B	n	Missing	Total	Totally agree with A	Partly agree with A	Partly agree with B	Totally agree with B	n	Missing	Total
Left party
Frequency	35	190	190	61	476	52	528	87	185	150	54	476	52	528
Per cent of respondents	7.4	39.9	39.9	12.8				18.3	38.9	31.5	11.3
Cumulative per cent	7.4	47.3	87.2	100.0				18.3	57.1	88.7	100.0
Social democratic party
Frequency	440	1,647	890	182	3,159	396	3,555	776	1,506	714	148	3,144	411	3,555
Per cent of respondents	13.9	52.1	28.2	5.8				24.7	47.9	22.7	4.7
Cumulative per cent	13.9	66.1	94.2	100.0				24.7	72.6	95.3	100.0
Green party
Frequency	34	153	227	58	472	58	530	69	186	181	32	468	62	530
Per cent of respondents	7.2	32.4	48.1	12.3				14.7	39.7	38.7	6.8
Cumulative per cent	7.2	39.6	87.7	100.0				14.7	54.5	93.2	100.0
Centre party
Frequency	43	325	446	130	944	103	1,047	101	377	361	99	938	109	1,047
Per cent of respondents	4.6	34.4	47.2	13.8				10.8	40.2	38.5	10.6
Cumulative per cent	4.6	39.0	86.2	100.0				10.8	51.0	89.4	100.0
Christian democratic party
Frequency	21	160	154	41	376	65	441	33	170	129	36	368	73	441
Per cent of respondents	5.6	42.2	41.0	10.9				9.0	46.2	35.1	9.8
Cumulative per cent	5.6	48.1	89.1	100.0				9.0	55.2	90.2	100.0
Liberal party
Frequency	37	217	291	84	629	89	718	87	259	216	61	623	95	718
Per cent of respondents	5.9	34.5	46.3	13.4				14.0	41.6	34.7	9.8
Cumulative per cent	5.9	40.4	86.6	100.0				14.0	55.5	90.2	100.0
Conservative party
Frequency	203	877	675	180	1,935	271	2,206	382	818	547	174	1,921	285	2,206
Per cent of respondents	10.5	45.3	34.9	9.3				19.9	42.6	28.5	9.1
Cumulative per cent	10.5	55.8	90.7	100.0				19.9	62.5	90.9	100.0
Swedish democrats
Frequency	42	105	95	24	266	57	323	44	103	89	31	267	56	323
Per cent of respondents	15.8	39.5	35.7	9.0				16.5	38.6	33.3	11.6
Cumulative per cent	15.8	55.3	91.0	100.0				16.5	55.1	88.4	100.0

The results show that all parties, except the Swedish Democrats, follow the same pattern, since the politicians report that they tend to follow the party standpoint more often when making decisions on tax levels and activity fees than on organisational changes in the municipal administration. To further test this, Wilcoxon tests were calculated for the respondents’ answers, split by party affiliation. The results show that there is a significant difference in the way the politicians report that they make decisions in the two policy issues for all parties, except the Swedish Democrats. Furthermore, the results for all parties except the Swedish Democrats showed greater mean ranks and the sum of ranks in favour of following their own preferences, if it conflicts with the party standpoint, when making decisions on organisational change in the municipal administration compared to making decisions on tax levels and activity fees. The full results of the Wilcoxon tests are shown in Table 4.

Table 4. Results from Wilcoxon tests, split by party affiliation, when making decisions concerning organisational change in the municipal administration and tax levels/activity fees

	Decisions on organisational change – decisions on tax levels/activity fees
	Left party	Social democratic Party	Green party	Centre party	Christian democratic Party	Liberal party	Conservative party	Swedish democrats
Mean of negative ranks	88.39	551.39	84.15	163.58	63.35	99.77	336.67	40.50
Mean of positive ranks	95.65	561.37	90.15	172.25	59.79	114.65	348.06	38.33
Sum of negative ranks	4,861.50	176,997.50	2861	14,395	2,597.50	4,689	72,383.50	1701
Sum of positive ranks	12,529.50	446,288.50	12,892	43,235	4,783.50	20,064	164,632.50	1380
z	−5.619^a	−13.585^a	−7.991^a	−8.710^a	−3.088^a	−8.663^a	−9.537^a	−0.877^b
p	>0.001	>0.001	>0.001	>0.001	>0.01	>0.001	>0.001	0.38

Notes: ^aBased on negative ranks.

^bBased on positive ranks.

Discussion and conclusions

The aim of this article is to explain local Swedish politicians’ decision-making by exploring to what extent they rely on the party principle and the trustee principle when making decisions on two issues that diverge in political dignity: tax levels/activity fees and organisational change in the municipal administration. By conducting this study, insights into previous research on the relation between political parties and their representatives, regarding institutional theory as well as to the particular case of decision-making of Swedish local politicians in regard to different policy issues, are provided. Several conclusions can be deduced based on the empirical findings.

In line with previous research, the results of this study indicate that Swedish politicians on the local level do follow the party principle to a larger extent than the trustee principle when making decisions. Moreover, the party principle is followed significantly more often when making decisions on important policy issues in this case illustrated by tax levels/activity fees, than on less crucial policy issues, such as an organisational change in the municipal administrations. The assumed difference between the two policy issues is thus prominent, as shown in the variance analysis.

The party discipline in Sweden when deciding on taxes is strong and, as such, is manifested in local politicians’ reports that they follow the party principle rather than the trustee principle. The result of Karlsson and Gilljam (2014a), emphasising the importance of the party principle in actual behaviour, can thus be complemented by the results of this study stating that the more important the policy issue is, the more inclined the politicians are to conform to the party standpoint, regardless of their representation style.

It is also interesting to note that a majority of the politicians, across different political parties, still follow the party principle rather than the trustee principle also when making decisions on a policy issue that is considered less important and less constrained by party pressure. According to previous research (see Esiassion and Holmberg 1996), members of the Left, Social Democratic and Conservative Parties are more inclined to follow the party principle than members of other Swedish parties. The results of this study indicate that this pattern is still robust for the Social Democratic and the Conservative Parties, whose local politicians stand out as being particularly loyal to the party standpoint. The results for the Left party, however, show no signs of leftish representatives being more loyal to the party than others. The Swedish Democrats, a relatively new party in the parliament, stands out as having politicians that follow the party principle to the same extent as the Conservative Party, when deciding on tax levels/activity fees.

Compared to the results of Esiassion and Holmberg (1996), this study shows that there is a large increase in the overall tendency to follow the party principle rather than the trustee principle, when the personal preference conflicts with the party’s preference. The reason the Left Party does not distinguish itself as being particularly party loyal is mainly because other political parties show stronger party discipline than the Left Party does.

An interesting question to explore for further research is thus: why have the Left Party’s politicians seemingly not increased their compliance to the party principle when politicians belonging to other parties have? Here, studies on political culture and institutional theory, for example, may provide interesting answers. Has the party culture in the Left Party changed? Are there institutional changes in the Left Party, which have contributed to a strengthening of the trustee principle?

The result of the variance analysis, when splitting the sample by party affiliation, showed the same significant difference between the decision-making on tax levels/activity fees and an organisational change in the municipal administration, except for politicians belonging to the Swedish Democrats. A majority of the politicians representing this party followed the party standpoint to almost the same extent when making decisions regarding both policy issues. This may be the result of the politicians representing the Swedish Democrats having internalised their role as politicians more than politicians from other Swedish parties. Another explanation may be that the Swedish Democrat party puts more pressure on their representatives to conform than other Swedish parties do. However, the empirical results show that the politicians from the Social Democratic and Conservative Parties follow the party principle to a larger extent in both policy issues than those of the Swedish Democrats, so the aforementioned explanation is not likely. Perhaps the Swedish Democrats being relatively new to the political setting have an influence on their decision-making behaviour. Further research should benefit from delving deeper into explaining this deviation when compared to the other political parties.

This study shows that the party principle can still be considered as the dominant representation style for Swedish politicians at the local level. Consequently, this study validates much of previous research results on representation styles, but also offers insights and explanations to how and when the pressure to conform to the party standpoint varies. A next step for further research is to examine decision-making behaviour in regard to other policy issues and in different contexts for the purpose of exploring the robustness of, and further refine, the results. For example, are the results of this study consistent over time? Regarding what other policy issues are there significant differences in how Swedish politicians at the local level make decisions?

The study contributes also to the theoretical field of new institutionalism. To start with, the study illustrates how new institutional theories can be applied to frame political decision-making and emphasise the structural constraints surrounding the individual decision maker in this process. Moreover, the study contributes to the field by showing how different theoretical assumptions about the individual, and the individual–institution interaction, can be operationalised and empirically measured. Finally, the study contributes to new institutionalism by examining the accuracy of the predictions put forward by the RCI and SI.

According to the empirical results, in this article, the RCI is a more appropriate framework when predicting how local politicians make their decisions. A significant difference in how the local politicians made decisions in the two policy issues was found. The rational hypotheses could not be falsified, and this raises interesting implications for further research. Exploring whether this is just the case for these types of policy issues or for the Swedish political context (with a strong informal institution of following the party principle) can facilitate both validation and falsification of these findings. Identifying contexts in which local politicians make decisions both according to RCI and SI predictions is important for further research, since it may vary between different contexts and policy issues. It is also important to point out that there are more theories in SI than the logic of appropriateness that was used in this study, such as DiMaggio and Powell’s (1983) institutional isomorphism. Perhaps additional perspectives can provide insights into the decision-making regarding other issues.

Disclosure statement

No potential conflict of interest was reported by the author.

Additional information

Funding

This work was supported by the Swedish Research Council Formas [211-2011-1363]; the Swedish Environmental Protection Agency [grant number NV-03842-10].

Notes on contributors

Jens Nilsson

Jens Nilsson is a Ph.D. student in political science at the Department of Business Administration, Technology and Social Sciences at Luleå University of Technology in Sweden. Research interests include institutional theory, decision-making, public administration, local government and political psychology.

Notes

1. There is also the delegate principle. This principle emphasises that politicians should make decisions in line with voters’ preferences in policy issues. For more on this representational principle, see Bäck (2000).

2. Two other prominent new institutional approaches are Historical Institutionalism and Constructive Institutionalism. Constructive institutionalism especially differs epistemologically from both RCI and much of SI, and this epistemological difference is important to note for institutional studies. However, since this epistemological debate is not central to the aim of this study, I recommend reading Peters (2012) for further study regarding these differences.

3. The main difference between RCT and RCI lies in the assumptions of information processing. RCT assumes that individuals have access to full information about the surrounding world when making decisions, which makes structural beings, such as institutions, redundant (Hall and Taylor 1996). Contrary to RCT, RCI assumes that the rationality for making decisions is bounded and the individual is not cognitively capable of processing full information. All available information cannot be taken into account by the individual when making decisions, but evaluation and behaviour stemming from the information they have obtained is still rational and according to their preferences. This is referred to as bounded rationality (see Simon 1957). The assumptions of benefit maximising also differ between RCI and RCT. Whereas RCT assumes that individuals strive to maximise benefit, RCI on the other hand tends to shy away from assumptions of individual maximising of benefit and simply states that individuals strive to obtain benefit, which does not necessarily mean the same as the best marginal utility (North 1990).

4. The questionnaire is for a project based at the Gothenburg School of Public Administration, where several questions on a variety of subjects are asked among Swedish politicians at both the local and the regional levels. For more on this project, see Karlsson and Gilljam (2014b).

Appendix

Questionnaire questions

Below are a few questions on decision-making in the municipal political body of which you are a member.

Question 1.

The questions are constructed as opposites (A and B). To which extent do you agree with A or B when you decide on proposals about an organisational change in the municipal administration?

A: I always follow the party’s standpoint, even when it is contrary to my own opinion.

B: I always follow my own opinion, even when it is contrary to the party’s standpoint.

Totally agree with A ☐ Partly agree with A ☐ Partly agree with B ☐ Totally agree with B☐

Question 2.

The questions are constructed as opposites (A and B). To which extent do you agree with A or B when you decide on proposals about municipal tax levels and activity fees?

A: I always follow the party’s standpoint, even when it is contrary to my own opinion.

B: I always follow my own opinion, even when it is contrary to the party’s standpoint.

Totally agree with A ☐ Partly agree with A ☐ Partly agree with B ☐ Totally agree with B☐