Standing on the shoulders of giants or science? Lessons from ordoliberalism

James Buchanan would have celebrated his 100th birthday in 2019. This serves as an inspiration to look at the future of public choice and the question of how much normativity public choice can bear. In our analysis we draw parallels between public choice and German ordoliberalism (and its source in the Freiburg School of Economics). We argue that the reception of ordoliberalism exemplifies easy-to-grasp pitfalls that should be taken seriously. We anchor the future agenda of public choice in a solid individualist perspective. Similar to ordoliberalism, public choice will have to clarify its relation to normative economics. The effects of rules and institutions and their working properties should be thoroughly analyzed empirically. The role of ideas is important for the normative foundation of both public choice/ constitutional economics and ordoliberalism, and is rooted in normative individualism. It provides a benchmark by which rules and institutions can be judged as favorable.


Introduction
Public choice and constitutional economics have been under attack recently. Nancy MacLean's (2017), Democracy in Chains, for example, characterizes James Buchanan and the group of scholars around him as an antidemocratic crowd. Munger (2018) writes in his review essay that MacLean aims at showing the "centrality of James Buchanan and his work in a gigantic conspiracy designed to end democracy in America" (p. 381). Several reviews of MacLean's book convincingly demonstrate that such an account is utterly wrong (in addition to Munger 2018, see Fleury & Marciano, 2018or Boettke, 2019. Nevertheless, MacLean's book has affected the thinking of public choice scholars to date, be it only to the extent that they feel the need to escape the defensive stance MacLean's narrative has seemingly put them in (see Boettke, 2023).
For many years, a group of scholars has aimed at distinguishing themselves from the research program of public choice and constitutional economics, by calling their research program political economics (see Alt & Alesina, 1996;Persson & Tabellini, 2000;Besley, 2007). They have disassociated themselves from public choice or constitutional economics even though their efforts have been to develop the economic analysis of politics further and their approach has been more or less the same to that of public choice and constitutional economics.
Moreover, there are some signs that recognition of public choice in the broader economics and political science literatures may have declined since the beginning of the 2000s. Frey (2022), for example, finds a decline in use of the term "public choice" from that period. Frey and Moser (2021) discuss three possible explanations for this phenomenon. The first explanation states that public choice has become so successful that it has been taken up into mainstream economics and political science. On the one hand, this explanation may be dismissed given the dominance of the traditional neo-keynesian synthesis (including welfare economics) in economic policy. On the other hand, it is clear that public choice considerations have definitely arrived in the economic mainstream. Until recently, it appeared that there was an indispensable necessity to study institutions and rules in several areas of economic research, in microeconomics (e.g., mechanism design, industrial economics, public economics) as well as in macroeconomics (e.g., central bank independence, fiscal rules).
Their second explanation states that an approach vanishes, because the longer an approach exists, the more difficult it is for that approach to offer new insights. The third explanation attributes the decline to a "Zeitgeist" in which problem-solving capacity is easily attributed to the state, in particular after the financial crisis of 2008/2009 or regarding climate change. A critical perspective on politics, as it is offered by public choice, is unpopular in such an environment. This explanation says nothing about the importance of public choice. On the contrary, it is probably more important given the general public's believe in the problem solving capacity of the state.
Fourth, the identification revolution in empirical economics may pose special challenges to analyzing many institutional questions with which public choice is given the frequent modesty of institutional variation over time, the lack of natural experiments, and the difficulties of implementing field experiments. 1 Even in laboratory experiments, institutional analysis is often difficult. Of course, public choice is interested in more than just institutions; empirical papers in public choice are abundant; and experimental public choice has existed for many years. Yet, institutions matter more in public choice and in constitutional economics than in standard economics or even political science. Identifying a causal effect of institutions is thus of particular import.
Finally, there are sociological aspects worth taking into account. The first generation of public choice scholars have passed away 2 and a second generation is ageing. 3 Younger public choice scholars, who regularly attend conferences or who cover topics in the area of public choice might have different ideas as to how their research is part of a particular agenda in a particular field.
In this paper, we consider the development of public choice from a totally different perspective. We discuss these uncomfortable questions by using German ordoliberalism as an illustrative example-from its roots in the Freiburg School of Economics and the thoughts of Walter Eucken in the 1930s to its normative deadlock at the end of the twentieth century.
This may be helpful in several respects: Firstly, as Vanberg (1988) argues, ordoliberalism is relatively close to constitutional economics and public choice (see also Feld & Köhler, 2011. Ordoliberalism sprung up in the beginning of the 1930s and, independently from the Old Chicago School with Frank Knight and Henry Simons, developed its own concept of rule-based economic policy (Köhler & Kolev, 2013). The key idea of the Freiburg School is to improve the living conditions of citizens by changing constitutional rules and reforming institutions, rather than by political interventions. With this rule-based approach and the analytical double-layered perspective of the order of rules and the order of actions, the Freiburg School is surprisingly close to the Virginia School of Economics and in particular to James Buchanan's approach to political economy.
Secondly, similar to public choice, ordoliberalism has come under attack during the Eurozone crisis. Due to its emphasis on austerity, it is blamed for causing people's hardship in member countries of the European Union (EU) during the debt crisis (Blyth, 2013;Feld et al., 2015;Wyplosz, 2017). The ordoliberalization of the European Union in general has become a major concern to some in recent times (Nedergaard, 2020).
Thirdly, the normative foundation of ordoliberalism and its normative orientation in giving economic policy advice has been criticized for a long time (Kirchgässner, 1988). This discussion offers parallels to MacLean's (2017) narrative. Today, the Freiburg School of Economics is often perceived as a biased normative agenda with scholars standing on the shoulders of Walter Eucken repeating his "ideas". Thus, both research programs may have an unsettled conflict with normativity as discussed in public choice (Voigt, 2011a(Voigt, , 2011b. In summarizing the current discussion, we want to focus our attention on the normativity issue that is a looming threat for the future development of public choice. As Paul Romer (2015) recently reminded us in his paper on "mathiness", the ability of economists 1 3 to form a consensus is undermined by those who disguise normative convictions as science by conflating politics and science into "academic politics", i.e., the abuse of scientific methods to further a political agenda. In his wide-ranging discussion of economic methodology, Rodrik (2015) cautions that economists' statements often contain hidden value judgements. The problem of ideological bias is particularly relevant when economists are called upon to evaluate public policies.
Neither public choice nor ordoliberalism have to fully refrain from normativity though. They should stick to normative and methodological individualism as key methods of finding improvements that make all citizens better off by a step-by-step variation of rules and institutions in accordance to their conjoint constitutional interest and by informed and free consent. Advising citizens with better institutions for realizing mutual gains is surely a positive endeavor. Finding legitimation must be nothing but a matter of individual consent. 4 With this return of a contractarian perspective as the ultimate criterion for legitimacy, we want to give a clear rejection to those who see the future of public choice within a doctrine. This is the central lesson to be drawn from Cold War ordoliberalism.
The remainder of the paper is organized as follows: Sect. 2 will briefly sketch the deadlock of German ordoliberalism at the end of the Cold War, Sect. 3 will summarize the most important lessons we have learned, and Sect. 4 will conclude.

Canonizing on the shoulder of a giant: the deadlock of traditional ordoliberalism
The key idea of the Freiburg School, as mentioned before, was to improve the living conditions of citizens by changing the few constitutional rules and by reforming institutions rather than by political interventions into day-to-day politics (Feld et al., 2015). During the Cold War era, it was generally accepted among economists to make policy recommendations based on external value judgements. Private property, freedom of contract and the need for an institutional framework provided by the state were often taken for granted. In Germany, Eucken's "constitutive principles" of a free market economy (Eucken, 1952(Eucken, /2004 lent themselves to becoming the standard norm, not least because every economist in German-speaking countries knew them from his or her undergraduate studies. This way, critical reflections on the importance of a value-free economic science were delayed.

Gebhard Kirchgässner's critique of traditional ordoliberalism 5
Gebhard Kirchgässner (1988) offers a forceful critique of traditional ordoliberalism from the point of view of public choice theory. 6 Like Romer (2015), he criticizes that many ordoliberal contributions are seemingly objective, but contain hidden value judgements.
Kirchgässner purports that Eucken's successors make economic policy suggestions involving distributive judgements without referring to political decision-making processes. Ordoliberals often evaluate measures of economic policy by asking whether they comply with Eucken's principles. However, the question whether these principles correspond to the wishes of the members of society is not addressed (ibid, pp. 55-58; 62-65). This indicates that ordoliberals base their policy recommendations on some kind of superior knowledge in the sense of philosopher kings (ibid, p. 53). Kirchgässner argues that this approach resembles the concept of a "benevolent dictator" employed by welfare economics (pp. 58-62). To be sure, both Eucken and Böhm assumed that the outcome of the political process should be taken as given by economists (Eucken, 1940(Eucken, /1950Böhm, 1973, p. 20). Seen from this vantage point, traditional ordoliberalism clearly falls short of public choice theory. When ordoliberalism experienced a renaissance in the 1980s and 1990s (Krieger & Nientiedt, 2023), its supporters attempted to canonize Eucken's policy recommendations to gain support for their anti-interventionist position. As Kirchgässner puts it: "That discussions about the theory of ordoliberalism at times take such a dogmatic turn possibly speaks in favor of its founder, Eucken, but not necessarily in favor of his descendants. Rather, it could indicate that this theory has not been developed further since his death" (Kirchgässner, 1988, p. 65). 7 Kirchgässner argues that the dogmatic adherence to Eucken's principles often leads ordoliberal scholars to call for strong, independent politicians with the "courage" to pursue long term political objectives and implement Eucken's policy program (ibid, p. 59). Eucken and Böhm, who were part of the German resistance movement against Hitler, clearly did not subscribe to a totalitarian conception of the state (Dathe, 2010). Still, Kirchgässner's critique reveals a methodological problem that was not taken seriously enough by traditional ordoliberalism. Viktor Vanberg (1988 points to the similarity of the research programs of the Freiburg School to constitutional economics. Central to constitutional economics is the notion that the outcome of economic activity depends on the institutional framework in which the activity takes place; and that this framework is the subject of deliberate choice. Constitutional economics, then, investigates the choice among rules and institutions, rather than interactions within a given framework of rules (Buchanan, 1990).

Viktor Vanberg's modernization of the Freiburg School's approach
At the heart of constitutional economics are two assumptions. First, constitutional economics is committed to methodological individualism, i.e., it assumes that social phenomena can only be explained in terms of the actions of individual human beings. This denies any "organic" understanding of social collectives such as the state. Second, constitutional economics maintains that legitimacy of social arrangements can only be derived from the voluntary consent of the individuals involved-a concept that is referring to normative individualism (Vanberg, 2005, pp. 24).
Since individuals are seen as the only carrier of values, the legal-institutional framework can only be "improved" by making it more responsive to the wishes of the members of society. In the case of markets, the term consumer sovereignty is used to describe the ideal that the interests of consumers should be the driving force of the economic process. Vanberg argues that such a procedural (i.e., process-oriented) criterion can also be specified for political processes: "Citizen sovereignty means that the individuals who constitute the citizenry of a democratic polity are the ultimate sovereigns in whose common interests the polity should be operated … In other words, citizen sovereignty requires that the 'producers of politics', politicians and government bureaucrats, are made most responsive to citizens' common interests" (ibid, p. 42).
In line with constitutional economics, the Freiburg School has argued that the best way of improving the outcome of markets is to improve the legal-institutional framework in which they operate. When assessing the desirability of economic constitutions, they employ a criterion almost identical to consumer sovereignty, referred to as "performancecompetition" (Leistungswettbewerb). Vanberg's argument is that, since citizen sovereignty must be understood as the logical equivalent of consumer sovereignty in the field of politics, representatives of the Freiburg School such as Eucken and Böhm could-in principle-have agreed to citizen sovereignty as the normative criterion against which different political arrangements should be evaluated (Vanberg, 1997, p. 724). The similarity between the Freiburg School and Buchanan's constitutional economics suggests a convergence of these research programs: "There are certainly sufficient affinities to allow for a fruitful dialogue between the tradition of Ordnungstheorie and the yet emerging paradigm of Constitutional Economics" (Vanberg, 1988, p. 28). Voigt (1997Voigt ( , 2011aVoigt ( , 2011b) describes a strictly positive approach to constitutional economics. He draws a distinction between two types of research programs dealing with the economic effects of constitutions. First, a normative type that asks how constitutional rules can be legitimized. The work of Buchanan provides an example of this. Second, a positive type that analyzes the economic implications of different rule regimes and the emergence or modification of constitutional rules. Positive constitutional economics, then, is concerned with the economic outcomes of different political systems-e.g., direct vs representative democracy, majoritarian elections vs proportional representation, presidential vs parliamentary systems, authoritarian vs democratic regimes-as well as the determinants of the emergence of particular rule regimes. Voigt notes that further research is needed with regard to the latter aspect, i.e., the task of "endogenizing constitutions" (Voigt, 2011a, p. 206).

Stefan Voigt's pursuit of a positive research agenda in constitutional economics
What implications does Voigt's analysis have for ordoliberalism? While he remains relatively vague on this subject, his survey papers suggest that he views positive constitutional economics as having clear advantages over the normative type. It should be emphasized that Voigt uses the term "normative" to refer to a certain field of study. He does not argue that normative constitutional economics attempts to "prove" normative statements ("what ought to be"). Therefore, normative constitutional economics as represented by Buchanan is as scientific as its positive counterpart.
Summarizing Kirchgässner, Vanberg and Voigt, the principle of normative individualism places value decisions firmly in the hand of sovereign citizens while methodological individualism offers an adequate starting point for explaining complex phenomena.

Recent developments in German ordoliberalism: contextual economics
After the latest "Methodenstreit" in German economics between 2009 and 2011, the discussion on the future of ordoliberalism has continued (see Feld & Köhler, 2011. Two types of proposals can be distinguished: First, "contextual economics" that combines normative aspects of ordoliberalism with certain ethical principles and sociology; and second, the pursuit of a positive research agenda. It is crucial to recognize that contextual economics relaxes normative and methodological individualism as the constitutive paradigm for research. Behavioral economics has shown that individuals neither act fully rational nor develop and update their preferences independently of their surrounding socioeconomic contexts and institutional conditions and, most notably, other peers. According to Kolev (2018), contextual economics explores the interdependencies between the economic and the social order, while "isolated" economics focuses on the process within these orders. With this idea, adherents of contextual economics try to reintegrate German ordoliberalism into the research program of economics. In our view, adherents of contextual economics can be in distress with their approach for the following reasons: First, their idea to contextualize economics from a plurality of social science perspectives can be interpreted as a "holistic approach" as Kirchgässner (1988) puts it. The claims of holism have hardly been met in the history of science. Behavioral economics offers a junction that contextual economics can take to prevent falling into this holistic loophole.
Second, marrying different perspectives can fall short of theoretical coherence-at least from the perspective of James Buchanan's public choice theory. There is also a certain irony in the fact that Walter Eucken was a pronounced adherent of theoretical coherence in his criticism of the younger German Historical School's lack of theoretical background. Kolev (2018) as well as Goldschmidt et al. (2016) when they define the purpose and aims of the Journal of Contextual Economics address this point.
In contrast, Feld and Köhler (2019) have argued that any modernization of the traditional perspective of German ordoliberalism cannot be achieved by circumventing commonly shared economic principles, first and foremost a concise individualism. In line with Arrow (1994) on social knowledge and methodological individualism, we argue, like Vanberg (2018b) and Wagner (2018), that the individualist approach has to lay the foundations of any modernized ordoliberalism. This should also guide public choice perspectives on both sides of the Atlantic.
Stemming from the history of public choice, James Buchanan himself started his own way out of Chicago by explicitly rejecting any "organismic theory" while embracing a consistent individualism (Wagner, 2018). There is no leeway for a "special path" rooted in his perspective for anything that is not in line with an individualist approach. And even if we are implicitly building on the insights of Buchanan's first article (1949), we do not canonize a normative idea, but a methodological principle, i.e., individualism and a rejection to any form of other context that cannot be explained by an individualist research perspective.
Third, and well connected to the history of economic thought argument, Kolev (2018) roots the motivation of contextual economics in its envisaged goal to re-attach to the youngest historical school that was well-known for its theoretical incoherence as prominently pointed out by Eucken (1934/54). Why not instead follow a large body of literature that supports a consistent theoretical monolithic perspective, if Eucken himself was turning tables in this direction?
Albeit the fact that their contextualization approach might fall short of its own claims due to incoherence, we believe that the other point they make is very much in line and compatible with James Buchanan's public choice perspective: To improve the living conditions of citizens by reforming the constitution with individual freedom and voluntary exchange as central concepts.
This can be concluded from Buchanan's (1964) presidential address at the Southern Economic Association on the question "What should Economists do?". Buchanan nailed down the plea that economics should not be executed as an "applied mathematics" or "managerial science endeavor". In this sense, we can accept that individual choices are often context-dependent. Bearing into this direction, public choice has a great potential to find explanations for phenomena neoclassic theory would classify as rational.

Lessons for the future of public choice
According to Max Weber, our task as scientists is to break the spell of explanations as myths and magic reality (in its sheer unbelievable complexity as we want to further his point) by a few very simple rules that can explain it (Weber, 1919, p. 536). The core of public choice comes with these few simple ideas, as discussed in the two volumes of the recently published Oxford Handbook of Public Choice (Congleton et al., 2019a(Congleton et al., , 2019b. We cannot summarize them all. But we will surely not find an "optimistic faith" in its future if we seek for further complexity in the methodological framework of public choice (Vanberg 2018a).
In addition, and as a reminder (to ourselves), we side with Weber (1919, p. 543) in his plea that political convictions neither belong in the classroom nor do they belong in the syllabus. Nevertheless, the scientific analysis of politics is more important than ever. Weber ardently urges us to "demand this intellectual righteousness from us" to conduct it.
Below we present key insights gained from the methodological debate about ordoliberalism that we believe to be invaluable for the future of public choice: Buchanan (1977Buchanan ( /2001 has often argued that there is only one "principle" in economics that is worth stressing and that the economist's didactic function is one of conveying some understanding of this principle to the public at large: The invisible hand, or to be more specific, the spontaneous exchange of self-interested individuals who choose among alternatives conditional on a given set of rules.

Keep up with rational choice
It is the merit of Buchanan to have applied this principle not only to markets, but to politics and other domains of society. Ever since, it has inspired researchers in the social sciences, especially scholars outside of economics. It is this very principle that has attracted researchers to analyze bureaucracy, society and, of course, politics. No other approach provides a better principle to explain complex phenomena of the order of action than individual behavior.
Therefore, even if it is tempting to follow up on normative approaches, we should avoid any deviation from the core individualist principle of public choice at all cost. Giving in to any approach that does not follow methodological individualism might lead into the trap that Kirchgässner (1988) detected in the renaissance of ordoliberalism.

Think outside your faculty
A strict individualistic approach is hard to sell outside of the economics faculty. But as a matter of fact, public choice poses questions that are the subject of political science, jurisprudence and other social science disciplines. Consequently, public choice has spread out to Law (Torgler, 2022) or Political Science in the US. In Europe, this is the case to a smaller extent. "If we are interested in discovering the unknown, it is best to allow individual scientists free reign in their searches. The jigsaw puzzle that is confronted can best be 'solved' by allowing different persons to look for differing sub-patterns, especially since the 'big picture' has no defined borders" (Buchanan, 1977(Buchanan, /2001. With this application of Michael Polanyi's solution of the coordination problem to the organization of science, Buchanan (1977Buchanan ( /2001 attempted to gain approval for the core principle of rational individuals in other fields. Most noteworthy is, of course, that political science has come a long way. While we embrace a rational interest approach as the second best, it is still important to reach out to other disciplines to enrich public choice. Here again it is individualism as a paradigm that sets the limit of a truly scientific approach in the sense of Weber (1919).

Anecdotes are false friends
The following point has already been addressed to some extent by point 1 and 2. We argue that just as economic science requires empirical studies for its own progress, so too the arena of politics requires empirical constitutional economics. Why is that? Steven Pinker (2019) recently criticized that many leaders and influencers, including politicians, journalists, intellectuals, and academics, surrendered to "the cognitive bias of assessing the world through anecdotes and images rather than data and facts". The United States have introduced the "Evidence-Based Policymaking Act" in 2018. It was signed and enacted into law on January 14, 2019, as P.S. 115-435. With this act, the federal government wants to build on empirical analysis to support or decline reform proposals and policy alternatives. This act is combined with an open government data act that will make data open by default to social scientists. This requires us as economists and social scientist to even further support empirical analysis as a method for public choice.
Pinker's plea for more "factfulness" in scientific research also supports Voigt and Gutmann's criticism of case studies (2019, p. 445): "Fairly little is known on how quantitative cross-country and qualitative case-study research can best complement each other." Against which criteria if not his own values should a scientist assess these case studies if he was not informed by data and evidence that is replicable? To be clear, we are not against a historical analysis of constitutions or comparative qualitative analysis per se. We simply argue that it cannot be the main source for providing knowledge in science and political decision-making. The plural of anecdotes is data. Causal inference is impossible with anecdote alone.

Beware of the limits of statistics
The literature on constitutional economics studies shows how different rules, institutions and even forms of government and corresponding voting systems affect policies and economic outcomes. Mueller (1996) and Voigt (1997Voigt ( , 1999Voigt ( , 2011a have summarized early studies on constitutions. A central shortcoming in the empirical literature is that many empirical studies that investigate the effect of constitutional rules suffer from endogeneity. Acemoglu (2005) has emphasized this problem discussing Persson and Tabellini's (2003) seminal book in great detail regarding the identification of the causal effects. Nevertheless, despite these shortcomings, Persson and Tabellini's book gave empirical constitutional economics a "big boost", as Voigt (2011b) noted. Building on Acemoglu (2005), Voigt (2011b) expressed his concerns and analyzed "a number of problems" with Persson and Tabellini (2003). Implicitly addressing the endogeneity problem, Voigt (2011aVoigt ( , 2011b) specifies shortcomings "ranging from the delineation of the independent variables" to the question of "how exogenous the independent variables really are". Rockey (2012) tried to overcome this very problem by methodological refinements to the identification strategy of causal effects. Persson and Tabellini (2003) carefully adjusted their study to newer insights. As early as 2003, they emphasized that measuring democracy as a binary variably was perhaps a blunt concept.

Cures to endogeneity: modify the empirical strategy for better inference?
Most research in empirical constitutional economics focuses on the effects of a single explanatory variable. Voigt (2011aVoigt ( , 2011b argues that we should modify this narrow focus and analyze a variety of explanatory variables and their effect on the dependent variable, as far as the theoretical underpinnings allow for such a wider scope. His suggestions also include the dependent variable, of course. Still, the empirical application of this idea is not without ambiguity, especially if we think about panel data. Providing a moderate n-dimension, cross-country studies suffer from large unobserved heterogeneity (in institutions) which cannot be identified. The omitted variable bias comes on the heels of this unobserved heterogeneity with the vast majority of institutional characteristics remaining anything but constant. At the same time, cross-correlated effects increase with the number of explanatory variable and the time dimension. Thus, we believe that it may be better to stick to a simple identification strategy and control for these time series characteristics as sensitively as possible by choosing the "right" estimator (Feld et al., 2020). But even the best modeling cannot cure a weak identification. Thus, it might be better to take advantage of exogenous variation. Here, quasiexperimental studies serve as a possible solution in order to circumvent the shortcomings of the traditional one-single equation OLS regression that was believed to explain a causal effect (see again Torgler, 2022). Kantorowicz and Köppl-Turyna's (2019) study of a reform in Poland serves as a relevant example. They analyze the fiscal effects of electoral systems by exploiting discontinuities in municipalities conditional on a reform that imposed a change of the voting procedure. Although the Polish reform of 2002 has no experimental characteristic, they can explain its effects indirectly by analyzing the differences in vertical fiscal imbalances between the municipalities that were applying different electoral regimes, just as the reform indicated.

3
Thus, they overcome the looming issue of omitted variable bias in cross-country studies by focusing on a single country only but exploiting many observations on the local level. However, most regression discontinuity designs (RDD) are tied to one country with many cross-sectional observations over a limited time horizon. This is surely not optimal for bold research, but in line with a correct RDD approach. Still, Kantorowicz and Köppl-Turyna do a very good job in providing insights to the citizens that were affected by this reform.
To sum up, tackling endogeneity by modifying the empirical modeling can work, sometimes. Endogeneity problems in cross-country studies will however surely remain a problem for empirical political economy and will probably never be fully resolved. Overall, the causality effects of constitutional or institutional rules provide challenges for future research in (empirical) public choice and constitutional economics.

Better causal inference by better identification strategies
Against this background, one can easily find support for Matsusaka's (2019) assessment that "much of the literature" does not meet modern standards for causal inference, in the sense of using randomized treatments and convincing control groups. This is surely not the case because researchers are unaware of these issues. It is simply due to the fact that most institutions that we as constitutional economists are interested in are endogenous.
There are two possible solutions to this problem. The first is to have long time series that run over many overlapping generations. An example is Funk and Gathmann's (2011) paper on the effects of direct democracy in Switzerland from 1890 to 2000, in which they employed cantonal fixed effects in a long macro panel. The longer the time series domain in panel data, the higher the probability that there are variations in institutions and the better it is for inference. However, even the longest time series does not cure omitted variable or endogeneity biases.
The second solution may therefore be found in quasi-experimental variation to estimate causal effects. However, the primary challenge here is to find sources of external variation that are truly random. Matsusaka (2019) highlights Asatryan's (2016) idea in the study of direct democracy in Germany as a positive example.
However, even the best identification does not solve the problem of effect heterogeneity. In a recent paper, Acemoglu et al. (2019) find a significant causal effect of democracy on long-run growth. Democracy is defined as a binary indicator. A shift to democracy leads to a 20% higher level of economic development in the long run. These findings are important as there is a latent belief that democracy might not be the best regime for generating economic prosperity. Posner (2010) pointed out, that "dictatorship will often be optimal for very poor countries".
With their study, Acemoglu et al. (2019) show that these beliefs and doubts do not withstand closer empirical scrutiny as they find remarkably stable coefficients across empirical specifications. Eberhardt (2022) agrees with their conclusions but raises doubts as to the assumptions of Acemoglu et al.'s (2019) modeling. He captures an important aspect that was assumed away: Heterogeneity. Eberhardt (2022) shows how results of the IV-estimator are significantly reduced if one controls for unobserved heterogeneity. The findings remain robust across a "host of empirical estimators with different assumptions about the data generating process", including "a novel instrumentation strategy", as Eberhardt (2022) points out.

Conclusions
Looking back on James Buchanan's 100th birthday, our paper discusses the future of public choice and constitutional economics against the background of the experience of ordoliberalism. We asked how much normativity can public choice and constitutional economics bear to still meet Max Weber's (1919) as well as James Buchanan's (1977Buchanan's ( /2001 standards. We find the limits of normativity in methodological and normative individualism. With these principles we can further public choice on a coherent foundation and do not fall short of a canonization of Buchanan's ideas for distinct rules that should be adopted by society. Illustrating the reception of ordoliberalism (and its source in the Freiburg School of Economics) and its fall into a deadlock of normativity, we send out a warning to social scientists who believe that public choice carries an inherent value for small government, deregulation or even the opposite. The German example should also teach us that "human governance" cannot "be reduced to ethics, law, and commerce, leaving no room for the political insertion of force into society, "as Richard Wagner recently suggested by calling for a "muscular version of liberalism" that he envisages as necessary since "free societies are not self-sustaining and can degenerate without the proper use of force" (Wagner, 2017, p. 14). 8 The Freiburg School's experience teaches us to refrain from any ideas that marry Carl Schmitt with liberalism (Köhler & Nientiedt, 2021). A benevolent liberal dictator remains a dictator.
Anchoring its future agenda in a solid individualist perspective (Vanberg 2018), we are in support of Voigt and Gutmann's (2019) plea to further positive analysis in (empirical) public choice. If we explore the effects of rules and institutions, we should exploit their working properties in an empirical manner-not by anecdotes. With this stance, we suggest to further public choice into a Popperian direction and call not to give in to the sirens of the sociology of knowledge as suggested along the way.