“Our countries have different traditions in dealings between state, industry, trade unions and social groups, but we share a conviction that traditional conflicts at the workplace must be overcome. This, above all, means rekindling a spirit of community and solidarity, strengthening partnership and dialogue between all groups in society and developing a new consensus for change and reform.” Tony Blair and Gerhard Schröder (1998)
With these words written in the programmatic paper, The Third Way, Tony Blair and Gerhard Schröder, the sitting Prime Minister of the United Kingdom and Chancellor of Germany, the way for a couple of structural reforms was paved (Blair and Schröder 1998). As a result, in Germany we saw– in particular with respect to the labor market –the most comprehensive social security reform since the World War II.
However, such kind of reforms seldom take place and politicians often turn out to be conservative preserver rather than risk taking gambler. Consequently, the Berlin-based journalist Anna Sauerbrey (2017) points out, “To most Germans the politicians are boring straight arrows – which is how we want them.” Beside the fact that preserving a good status quo can be beneficial, conservatism can be problematic if needed structural reforms are not initiated. Hence, the reluctance to implement reforms is criticized by a number of economists, as potential gains in welfare and economic growth might not be realized. 1 Therefore, the emergence of, and resistance to, reforms is one of the fundamental questions of modern political economy (see Alesina and Drazen 1991; Alesina, Ardagna, and Trebbi, 2006; or Fernandez and Rodrik, 1991). Here, the reasons for the absence of reform are, for instance, seen in the blockage by veto players whose agreement is required for a change of the status quo (Tsebelis, 1995) or in powerful interest groups involved in wars of attrition (Alesina and Drazen, 1991, p1170f) as well as in reform logjams initiated (Haggard and Webb, 1993, p143f).
In an additional branch of the literature, uncertainty and variations of risk aversion are seen as reasons for the resistance to reform. For instance, in the work of Fernandez and Rodrik (1991, 1146f) uncertainty in the distribution of the winners and losers of a proposed reform leads to a strong status quo bias and, hence, resistance to reform. The authors assume that voters are, on average, risk-neutral and stick to the median voter model. Consequently, this also implies risk-neutrality for politicians on average. 2
Another explanation for reform resistances are based on results from behavioral science 3: For instance, in the prospect theory of Kahneman and Tversky (1991, 1039f) individuals evaluate the options to act concerning the expected change in their wealth situation in comparison to a reference point. In the model of Kahneman and Tversky, potential profits have a lower impact on individual decisions than potential losses. In the political context, this can result in a resistance to change. 4 Thus, the model of Kahneman and Tversky shows that it can be perfectly rational to have a status quo bias – if people are risk averse.
A potential explanation for the absence of structural reforms could be founded in deviations in the risk preferences of voters, politicians, and bureaucrats. Voters, politicians, and bureaucrats interact in a multiple principal-agent relationship (see Table 1). Dixit (2006) as well as Dixit, Grossman, and Helpman (1997) analyze the principal-agent problem between politicians and bureaucrats as their agents, while Kau and Rubin (1979), Kalt and Zupan (1984), and Peltzman (1985) analyze if U.S. Congressional representatives, as political agents, serve the interests of their principal narrowly. Merville and Osborne (1990, 40) come the conclusion that the U.S. Constitution can be seen as a “poor agency contract” and see political agents as either risk averse or risk neutral. If the risk appetite of politicians (and bureaucrats) turns out to be smaller than the risk appetite of the voter, this could lead to reluctance for reform from the voter’s perspective.
However, despite the multitude of arguments to explain the absence of structural reforms, politicians actually seem to reform all the time. For instance, between 2002 and 2013, 1,569 legal acts passed the German Bundestag. 5 One explanation for the coexistence of the huge number of legislative acts, on the one hand, and the complaints by many economists, on the other, is that the majority of these reforms can be seen as incremental, if not symbolic, rather than structural. This gradualism is well discussed by Dewatripont and Roland (1992a, 1992b, 1995).
Now the puzzle is: are politicians and bureaucrats actually risk neutral or risk averse as the literature suggests and, if so, why do so many gradual reforms take place? In our short note presented here, we attempt to disentangle the puzzle. Based on a unique dataset containing three surveys, we investigate empirically how the risk preferences of German voters, politicians and bureaucrats look like and if risk aversion is one of the reasons for the resistance to reform structurally. The remainder of this paper is structured as follows: Section 2 presents our data and Section 3 the empirical results. Section 4 offers some concluding remarks and discussion.
2.1 Survey among the electorate (voters)
Annually, in the German Socio-Economic Panel Study (SOEP), about 20,000 respondents are asked questions about their socio-economic status, attitudes and values (Siedler et al., 2009; Wagner, Frick, and Schupp, 2007). One question of the survey focuses on risk appetite, respectively, risk aversion: On a scale of 0 (not at all willing to take risks) to 10 (very willing to take risks) respondents indicated the degrees of their risk appetite, respectively risk aversion (see Figure 1). The questions posed in the SOEP have been validated multiple times and replicated in other surveys around the world. For validation see Dohmen et al. (2011) and Frey et al. (2017).
Because only German citizens at the age of 18 years or older have the right to vote for the Federal Parliament (Bundestag) for the analysis presented here we only analyze the SOEP respondents who are German citizens and 18 years or older.
2.2 Survey among members of parliament (politicians)
In December 2011, Hess et al. (2013a) surveyed risk attitudes of members of the German Parliament (Bundestag). The question was to the same which is asked in the SOEP-Survey among the general population. To keep the questionnaire short only very few questions on standard demographics were included in addition to the questions on the risk attitudes. The survey was announced with an email two days before it arrived via postal mail and for each returned questionnaire five euros were donated to the German Children’s Cancer Foundation (Deutsche Kinderkrebsstiftung). A more detailed description of the survey can be found in Hess et al. (2013a), (2013b).
Of the 620 members of parliament who received the postal survey questionnaire, 175 responded. This is a high rate of response given (1) members of parliament are a very busy group of professionals (e. g. ISG, 2011, 10) and (2) response rates for mail surveys are usually low at about ten percent or less. The data allow conclusions about all members of parliament, as the socio-demographic composition of this sample corresponds by and large to that of the parliament overall: The 175 respondents were on average 52.6 years and 30 percent were women and 70 percent men. The members of parliament were on average 49.3 years old and 33 percent were women and 67 percent were men. Further information can be found in Hess et al. (2013b).
2.3 Survey among higher officials from ministries (bureaucrats)
In June and July 2014, Haucap, Thomas, and Wagner (2015) conducted an Email survey among higher officials from ministries on the federal- and state-levels in Germany. Included was a question about their risk-appetite. The question was again the same as in the SOEP-Survey. Because the bureaucrats were surveyed online the questionnaire was as short as possible. The core of the questionnaire was a question about the importance of economists in the context of scientific policy advice. 6 In addition, a question about the governmental level of the professional affiliation (federal or state-level) was asked.
Of 106 higher officials from ministries, 60 were federal and 46 state. The data on the risk-appetite were not yet analyzed and published till now.
The results based on the three surveys are summarized in Table 2. The survey of politicians in 2011 shows that members of the German Federal Parliament (Bundestag) are risk-seekers: On average, they indicate a risk-appetite of 6.4 on the scale from zero to ten. This is 1.9 points more than the voters. 7 In comparison to the 2012 result for the general population, the difference decreases, but remains, at 1.6 points, remarkable. In terms of the principal agent relationship between voters and politicians, the agent (politician) has a much higher tendency to take risks than the principal (voter). It is unlikely that this is due to strategically distorted responses on the part of the members of parliament, as particular risk attitudes do not seem to be socially desirable or undesirable. As the surveys were strictly anonymous, no link to specific members of parliament or political parties can be made.
With respect to the bureaucrats, in 2014 officials from ministries indicated, on average, a risk-appetite of 5.7. This is 0.7 points less than politicians in 2011. In terms of the principal agent relationship between politicians and bureaucrats, the agent (bureaucrat) has a lower tendency to take risks than the principal (politician). However, bureaucrats still have a greater risk-appetite than the electorate: in comparison to the 2011 survey results, it is 1.2 points higher. In comparison to the survey results of 2014, it is still 1.1 points higher.
However, the difference in risk attitudes might appear due to the differences in the sample composition rather than differences in attitudes between the three groups. In comparison to the voters the groups of politicians and bureaucrats differ in terms of socio-demographics. For instance, members of parliament and officials from ministries are older than the average of the population, while the share of male parliamentarians and bureaucrats far exceeds the share of males in the population of the voters (see Table 3). 8
However, for our main question, if the risk preferences of the voters (principal) are well represented in parliament (and in ministries) (agents), the socio-demographic reasons for the deviations in risk-appetite is not relevant. Those reasons are interesting from a more general perspective. Are the differences in risk appetite between voters, parliamentarians, and bureaucrats driven by the non-representative demographic structure of the parliamentarians and bureaucrats or are the differences due to a higher risk-appetite of people who decide to become a parliamentarian or bureaucrat.
To shed some light on the second question, in a first step we simply adjust the comparison group. All the Federal bureaucrats surveyed hold a university degree (and most parliamentarians hold such a degree) and it is well known that higher education comes with an above-average risk appetite (Dohmen et al. 2011). Moreover, most bureaucrats and parliamentarians are males. Thus, we compare the risk preferences of politicians and bureaucrats with the risk preferences of only the male and high educated among the voters.
Table 4 shows the results: In the SOEP survey, males with higher education have a risk-attitude-score of 5.2 in 2011 and 5.1 in 2014. However, with risk-attitudes-score of 6.4 (2011) politicians still have a remarkably higher risk appetite than the adjusted comparison group. Same applies for the group for the bureaucrats 5.7 (2014) on a lower level.
The results of t-tests show that the differences between politicians and bureaucrats, on the one hand, and the voters, on the other, are statistically significant. Hence, it is unlikely that the differences between risk-appetites are found just by chance. In other words: the risks appetites of politicians and bureaucrats are so high that we could find significant differences to the electorate even in small samples, which imply large sampling errors for politicians and bureaucrats. 9
To investigate the question if the difference in risk attitudes might be due to the differences in the sample composition rather than differences in attitudes in more depth, we use a simple regression technique to control for age and gender. However, this could be done for politicians only because bureaucrats were not asked to indicate their age and gender. The results are shown in Table 5. The dependent variable was the risk attitude. Independent variables were the status as politician with the voters as the reference category. Control variables were the age and gender as female with male as the reference category.
The result shows that the difference in the risk preferences of politicians and voters is driven, on one hand, by demographic differences in terms of age and gender. However, on the other hand, the differences in risk appetite are additionally driven by (pure) differences in risk-appetites in terms of attitudes.
As a result, politicians are more risk loving than the electorate and more risk loving than a comparable educational group among the general population. Hence, politicians are, with respect to risk appetite, not perfect representatives of the electorate. However, as they have a higher risk appetite than the electorate (as the bureaucrats have), risk preferences of decision makers cannot be seen as an explanation for the resistance to structural reform.
4 Conclusion and discussion
As, from the political economy perspective, politicians are often not successful in implementing structural reforms, we investigate if the resistance to reforms is due to the differences in the risk preferences of voters, politicians, and bureaucrats.
Based on the empirical results of three surveys among 20,000 German households, 175 members of the German Parliament (Bundestag), and 106 higher officials from German ministries, this cannot be argued. In the principal agent relationship between the voter (principal) and the politician (agent), the agent has a much greater tendency to take risks than the principal. This hints at a greater tendency to reform in comparison to the risk preferences of the electorate. The results are inverted in the principal agent relationship between the politician (principal) and the bureaucrat (agent). Here, the agent has a lower risk preference than the principal. As bureaucrats are needed in order to prepare and administer a reform, this could mitigate the risk appetite of politicians. However, as bureaucrats also have a higher risk appetite than the electorate, this still cannot be seen as an explanation for the resistance to reform. In contrast, the common assumption that politicians act in their own self-interest, thereby avoiding risks, cannot be proven by the empirical results presented here.
Our result of risk loving politicians is compatible to the common experience in western democracies, where politicians regularly implement reforms that are more or less all risky, at least with respect to their political consequences. However, most of the legal acts passing the legislative process cannot be seen as structural reforms that are game changers (for example, the introduction of completely new taxation schemes or public health insurance). Hence, the majority of these reforms can be seen as incremental, if not symbolic, rather than structural. This “conservatism,” which is criticized by a number of economists, is not an indication of risk aversion. Hence, it must be caused by other reasons, such as interventions by veto players (Tsebelis, 1995), wars of attrition by powerful interest groups (Alesina and Drazen 1991), or reform logjams initiated (Haggard and Webb, 1993).
However, it could be the case that modern democracies are doing better than many believe: The political process selects above average risk loving politicians (and bureaucrats) who are, in general, willing to take risks and action. In turn, the democratic institutions, different interests, and different pressure groups control them. Although bureaucrats have an above-average risk appetite (compared to the electorate), they are less risk loving than politicians and may act as another “control device.” During times of populist campaigns, the election process can put forth candidates with very high risk appetites, but the constitutions of democracies turn out to be rather effective, if they inhibit hazardous actions and measures by political rookies and gamblers by checks and balances.
The authors are grateful to the anonymous referees, the editor of the journal as well as to Adam Mark Lederer (Berlin) and Wolfgang Schwarzbauer (Wien) for useful hints and comments.
Alesina, A., S. Ardagna and F. Trebbi (2006): Who Adjusts and When? the Political Economy of Reforms, IMF Staff Papers 53, Special Issue, 1–29. Google Scholar
Alesina, A. and A. Drazen (1991): Why are Stabilizations Delayed?, American Economic Review 81, 1170–1188. Google Scholar
Blair, T. and G. Schröder (1998): The Third Way, Friedrich Ebert Foundation, download (14.07.2017): http://library.fes.de/pdf-files/bueros/suedafrika/02828.pdf
Deutscher Bundestag (2017): Kleine Anfrage Der Abgeordneten Jutta Krellmann, Klaus Ernst, Susanna Karawanskij, Weiterer Abgeordneter Und Der Fraktion DIE LINKE, Bundestagsdrucksache 18/11087. Google Scholar
Dewatripont, M. and G. Roland (1995): The Design of Reform Packages under Uncertainty, American Economic Review 85, 1207–1223. Google Scholar
Dixit, A. (2006): Predatory States and Failing States: An Agency Perspective, Princeton University, Department of Economics, Center for Economic Policy Studies Working Paper, No 71. Google Scholar
Dixit, A., G. M. Grossman and E. Helpman (1997): Common Agency and Coordination: General Theory and Application to Government Policy Making, The Journal of Political Economy 105, 752–769. CrossrefGoogle Scholar
Dohmen, T., A. Falk, D. Huffman, U. Sunde, J. Schupp and G. G. Wagner (2011): Individual Risk Attitudes: Measurement, Determinants and Behavioral Consequences, Journal of the European Economic Association 9, 522–550. Web of ScienceCrossrefGoogle Scholar
Fernandez, R. and D. Rodrik (1991): Resistance to Reform: Status Quo Bias in the Presence of Individual-Specific Uncertainty, American Economic Review 81, 1146–1155. Google Scholar
Frey, R., A. Pedroni, R. Mata, J. Rieskamp and R. Hertwig (2017): Risk Preference Shares the Psychometric Structure of Major Psychological Traits, Science Advances 3, e1701381, . CrossrefPubMedGoogle Scholar
Göbel, M., A. Schneider and T. Thomas (2010): Consumption Behavior and the Aspiration for Conformity and Consistency, Journal of Neuroscience, Psychology, and Economics 3, 83–94. CrossrefGoogle Scholar
Hess, M., C. Von Scheve, J. Schupp and G. G. Wagner (2013b): Members of German Federal Parliament More Risk-Loving than General Population, In, DIW Economic Bulletin 3, 20–24. Google Scholar
Hess, M., C. Von Scheve, J. Schupp and G. G. Wagner (2013a): Sind Politiker Risikofreudiger Als Das Volk? Eine Empirische Studie Zu Mitgliedern Des Deutschen Bundestags. SOEP Paper 545. Google Scholar
ISG (Institut für Sozialforschung und Gesellschaftspolitik GmbH) (2011), „Studie Zur Wahrnehmung Und Berücksichtigung Von Wachstums Und Wohlstandsindikatoren“ Im Auftrag Der Enquete-Kommission „Wachstum, Wohlstand, Lebensqualität“ Des Deutschen Bundestages (Kommissionsmaterialien M-17(26)11), Berlin. Google Scholar
Kalt, J. P. and M. A. Zupan (1984): Capture and Ideology in the Economic Theory of Politics, American Economic Review 74, 279–300. Google Scholar
Peltzman, S. (1985): An Economic Interpretation of the History of Congressional Voting in the Twentieth Century, American Economic Review 75, 656–675. Google Scholar
Rabin, M. (1998): Psychology and Economics, Journal of Economic Literature 36, 11–46. Google Scholar
Sauerbrey, A. (2017): Will Russia Try to Hack Germany?, New York Times (International Edition), July 22, 2017. Google Scholar
Schäfer, W. (2010): Politische Kartelle, Review of Economics (Jahrbuch Für Wirtschaftswissenschaften) 61, 213–224. Google Scholar
Stadelmann, D. and M. Portmann (2017): Testing the Median Voter Model and Moving beyond Its Limits: Do Personal Characteristics Explain Legislative Shirking?, Social Science Quarterly, published online first: 15 February 2017. Google Scholar
Stadelmann, D., M. Portmann and R. Eichenberger (2016): Preference Representation and the Influence of Political Parties in Majoritarian Vs. Proportional Systems: An Empirical Test, British Journal of Political Science, published online first: 5 December 2016. Google Scholar
Tsebelis, G. (1995): Decision Making in Political Systems: Veto Players in Presidentialism, Parliamentarism, Multicameralism and Multipartyism, British Journal of Political Science 25, 289–325. CrossrefGoogle Scholar
Wagner, G. G., J. R. Frick and J. Schupp (2007): The German Socio-Economic Panel Study (SOEP) – Scope, Evolution and Enhancements, Schmollers Jahrbuch 127, 139–170. Google Scholar
However, on the other hand there might be potential welfare losses if harmful reforms are implemented. Schäfer (2010) argues that in particular this can take place if political cartels exist and hence political competition is limited. In addition, Hielscher (2016) empirically analyses that economic gains derived from structural reforms are sometimes questionable.
For a critical discussion of Fernandez and Rodrik (1991), see Ciccone (2004). Beside this specific criticism, the median voter model can be questioned in general as well. In this line, Stadelmann, Portmann, and Eichenberger (2016) as well as Stadelmann and Portmann (2017) analyze empirically if politicians represent the voter’s preferences narrowly. The authors investigate in parts huge differences and state that the explanatory power of the media voter model for political action is rather limited.
For an overview see Heinemann et al. (2008).
In addition, Kahneman and Tversky (1986) find that the way you “frame” options in a decision situation has an important impact of the decision itself. For instance, in the case of an illness, individuals systematically opt for a therapy connected with a 90 percent probability to survive in comparison to a therapy connected with a 10 percent probability to die. In the context of political reforms, this can result in a resistance to opt for reforms, if they are framed from the perspective of losers. Two additional findings from behavioral science can be of special importance in the context of reforms: the confirmatory bias, which describes the tendency of individual to behave consistently over time (Rabin, 1998), and the conformity bias, which shows the tendency of individuals to conform their behavior to that of their peer group (Bernheim, 1994). As a result, in their integrative approach Göbel, Schneider, and Thomas (2010) show that, after taking biases into account after a shock, individuals will adjust their behavior, but with a delay and in suboptimal ways. In the political context, this can lead to suboptimal reform behavior in terms of size and timing.
The question on the risk appetite was part of a larger survey conducted by the Düsseldorf Institute for Competition Economics (DICE) and Econwatch – Society for Policy Analysis in collaboration with the German National Library of Economics (ZBW). It did focus mainly on the question “Who are the economists you do value most for your personal work”. The answers on this question are one pillar of the Economist-Impact-Ranking, published annually since 2013 in the Frankfurter Allgemeine Zeitung.
Kam (2012) shows, focusing on the U.S., that people with risk-loving attitudes are more likely to participate in political meetings, distribute leaflets, and be active in campaigns. The author explains this with the pleasure derived from new experiences and the excitement to be found with political action, which risk-loving people tend to seek more than risk-averse ones.
For a sophisticated analysis of patterns of political selection see Dal Bó et al. (2017).
The results can be confirmed with a simple calculation: We know that 70 % of the parliamentarians are male and 30 % are female. In addition, we know the risk-preferences of the SOEP-participants in 2011 with a university degree (male: 5.23; female: 4.84). If the SOEP-participants with a university degree would have the same gender structure as the parliamentarians, then the risk preference would be 5.23. As parliamentarians have a risk preference of 6.4, their risk appetite can be seen as remarkably high despite socio-demographic differences in the sample. Same applies for bureaucrats: We know that 63 % of the parliamentarians are male and 37 % are female. In addition, we know the risk-preferences of SOEP-participants in 2014 with a university degree (male: 5.17; female: 4.53). If the SOEP-participants with a university degree would have the same gender structure as the bureaucrats the risk preference would be 4.93. As the officials from ministries have a risk preference of 5.65, their risk appetite can be seen as rather high.
About the article
Published Online: 2017-11-30
Published in Print: 2017-11-27