Forerunners vs. latecomers—institutional competition in the German federalism during the COVID crisis

Breide, Lukas; Budzinski, Oliver; Grebel, Thomas; Mendelsohn, Juliane Article — Published Version Forerunners vs. latecomers—institutional competition in the German federalism during the COVID crisis European Journal of Law and Economics Provided in Cooperation with: Springer Nature Suggested Citation: Breide, Lukas; Budzinski, Oliver; Grebel, Thomas; Mendelsohn, Juliane (2025) : Forerunners vs. latecomers—institutional competition in the German federalism during the COVID crisis, European Journal of Law and Economics, ISSN 1572-9990, Springer US, New York, NY, Vol. 59, Iss. 1, pp. 101-132, https://doi.org/10.1007/s10657-025-09832-4 This Version is available at: https://hdl.handle.net/10419/323359 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. http://creativecommons.org/licenses/by/4.0/ Vol.:(0123456789) European Journal of Law and Economics (2025) 59:101–132 https://doi.org/10.1007/s10657-025-09832-4 Forerunners vs. latecomers—institutional competition intheGerman federalism duringtheCOVID crisis LukasBreide1· OliverBudzinski1· ThomasGrebel1· JulianeMendelsohn1 Accepted: 8 January 2025 / Published online: 7 February 2025 © The Author(s) 2025 Abstract During the COVID-19 pandemic, political competition among the premiers of Germany’s 16 federal states intensified, with leaders striving to position themselves as proactive forerunners in managing the crisis. This paper examines the timing and determination of these state leaders in announcing, legislating, and enforcing COVID-19 policies, with attention to regional contexts and specific determinants influencing their actions. Utilizing multiple distinct databases, we conduct a survival analysis to assess each state’s political response in relative terms. Our findings reveal that state leaders who were early advocates in public announcements and discourse did not necessarily lead in formal legislation or enforcement of COVID-19 measures. This study provides a nuanced view of political competition in crisis governance, highlighting the divergence between political rhetoric and tangible policy action across Germany’s federal landscape. Keywords Political competition· Institutions· COVID· Survival analysis JEL Classification D7· H7· H11· H12· R5 * Thomas Grebel [email protected] Lukas Breide [email protected] Oliver Budzinski oliver[email protected] Juliane Mendelsohn [email protected] 1 Institute ofEconomics, Ilmenau University ofTechnology, Ehrenbergstraße 29, 98693Ilmenau, Thuringia, Germany 102 European Journal of Law and Economics (2025) 59:101–132 1 Introduction Political competition within federal systems fosters both innovation and imitation among state leaders, particularly during crises like the COVID-19 pandemic (Mistur etal., 2023). Federalism enables forerunners, who initiate novel political responses, and latecomers, who adapt strategies proven effective by others, to shape public policy through a balance of proactive and reactive approaches. This dynamic, akin to competition in goods markets, suggests that both strategies can be effective, with forerunners seeking first-mover advantages despite inherent risks, such as unpredictable public reception (Ruggeri et al., 2024; Barberá etal., 2019), while latecomers benefit from observing the challenges faced by early adopters, allowing them to avoid missteps in policy implementation. The urgency of a pandemic amplifies this interplay, as states within a federal system must rapidly respond to external shocks (Segatto etal., 2022). The decision to act as a “forerunner” —to introduce measures that are perceived as necessary or effective—can enhance a leader’s political standing if subsequent events validate those actions. Conversely, positioning oneself as a “latecomer” by waiting cautiously before implementing potentially restrictive policies can mitigate risks to political reputation. A nuanced strategy seen during COVID-19 was the practice of early policy announcements without immediate implementation, allowing leaders to appear responsive without the immediate political costs associated with enforcement (Segatto etal., 2022; Ruggeri etal., 2024). While the broader dynamics of federal competition have long been observed, the COVID-19 pandemic provides a unique empirical opportunity to examine forerunners and latecomers in Germany’s federal system, particularly across the domains of public announcements, legislation, and enforcement. Unlike centralized systems, the German federal structure allows states (“Länder”) to implement region-specific actions within a national legislative framework (Hegele & Schnabel, 2021). Although centralized responses may offer immediate coherence during crises, federal systems like Germany’s are argued to be more adaptable in addressing regional variations and uncertainties over time (Bolton & Farrell, 1990; Cepaluni etal., 2022). For example, rural areas, where social costs of contact restrictions are lower, may benefit from distinct policy adjustments compared to densely populated urban centers (Carozzi etal., 2024; Congleton, 2023). The German Infection Control Act—falling under the federal government’s concurrent legislative powers (Art. 74(1) Nr. 19, German Basic Law) – explicitly permits federal states to implement such context-tailored regulations within the framework of federal guidelines. However, by April 2021, the German Bundesregierung shifted from a primarily statecoordinated pandemic response to a more centralized regulatory framework, aiming to reduce friction both among the states and between federal and state governments. (Färber, 2021). Existing literature on federalism and crisis management, such as Congleton (2023), highlights the flexibility federal systems offer in tailoring responses. Opposing views, such as those of Cepaluni et al. (2022), emphasize that high state capacity may be the essential factor enabling leaders to implement stringent 103 European Journal of Law and Economics (2025) 59:101–132 measures quickly, regardless of federal or unitary system structures. The shortterm advantages of centralized responses may decline as considerations of longterm social, economic, and psychological impacts necessitate decentralized adjustments. In the case of Germany, as pointed out, the process was reversed. The ambiguity as to whether a centralized or a decentralized approach to crisis management is preferable is due to the inherent uncertainty of an unprecedented pandemic (Bjørnskov & Voigt, 2022; Kollman etal., 2000). Leaders face uncertainty about the most effective responses, leading to experimentation, policy innovation, and an evolving selection process (Garzarelli & Keeton, 2018; Samad etal., 2022). They must navigate a complex set of influences, including their own ideological leanings, the prevailing public opinion, the priorities of their respective parties, and the dynamics of their political coalitions. These factors give rise to yardstick competition (Harrison, 1978; Bolton & Farrell, 1990; Besley & Case, 1992; Färber, 2021). Leaders seek to benchmark their policies against those of neighboring states, thereby fostering a spirit of adaptive competition. This study contributes to the literature by examining how state leaders within Germany’s federal system navigated their roles as either forerunners or latecomers in response to the COVID-19 pandemic. We analyze political actions across three dimensions—public announcements, legislative action, and enforcement— using a unique dataset compiled from LexisNexis, Juris, and Destatis. This data enables us to quantify competitive dynamics in policy-making and categorize state leaders according to their strategic approaches: “procrastinators,” who make early policy announcements yet delay legislative action; “armchair activists,” who rapidly legislate but exhibit leniency in enforcement; and “silent policymakers,” who effectively lead in both legislative and enforcement efforts without prioritizing media presence. As such, our analysis complements the “credit and blame”game (Greer etal., 2022) by using more granular data and digging deeper into the underlying politico-competitive mechanisms. We show that on the level of states and their leaders, strategies looking at the popularity of response instruments also play an important role as well. Closest to our research is the study by Broschek (2022), which compares three response strategies adopted by Canadian provinces during the pandemic: laissezfaire (although no Canadian province adopted this strategy), mitigation (six provinces), and containment (four provinces). Using a process-tracing approach, the paper tests several hypotheses about whether leadership contributed to the success or failure of the pandemic response during the second and third waves, and concludes that the containment strategy is superior to combat the pandemic and that political leadership plays an important role in both successes and failures. Broschek (2022) calls for more research on the role and behavior of political leaders during such crises. Our paper follows the call for more analysis of the role of leaders and focuses on the mechanisms of federal political competition, where leaders have an eye on public opinion and reelection probabilities and thus may strategically deviate from announcements, law codification, and enforcement. In short, we address the (strategic) inconsistency between (media) announcements, law codification, and law enforcement in a federal system. To the best of our knowledge, no previous work has addressed this issue. 104 European Journal of Law and Economics (2025) 59:101–132 The remainder of this paper is structured as follows: Section2 provides an overview of the German federal system and pandemic response framework, and Section3 presents descriptive statistics on forerunners and latecomers across announcement, legislation, and enforcement. Section4 models the propensity of state leaders to enforce COVID policies, with regression results in Sect.5. Section6 concludes. 2 Political competition inafederal system Germany is a federation of sixteen states, each of which enjoy a large scope of individual sovereignty. While the COVID-crises was a national health crisis and, thus, coordinated at the federal level, the implementation and enforcement of individual measures happened at state and regional levels. The individual states thus enjoyed much political discretion. In particular during the beginning of the pandemic this created a diverse array of political statements being made and rules for the individual citizens being implemented. 2.1 The scope forpolitical competition In order to analyze the competition leading to forerunners and latecomers, it helps to provide an overview of the legal framework that determines the vertical competition first between the federal state (Bund) and the individual states (Länder), as this also determines the scope of horizontal competition between the individual 16 federal states. We also show the shifts and changes in the amount of executive discretion afforded to the individual 16 states during the pandemic, thus providing a timeline for the scope of political competition. Being a federal country, both the federal state (Bund) and states (Länder) enjoy legislative and administrative competences. In relation to the states, the federal level either has exclusive, concurrent, or ancillary legislative powers. According to the constitution, federal legislation takes precedence over state legislation (Art. 31 German Basic Law). For dangerous and infectious diseases there exists a so-called concurrent legislative power (in German: konkurrierende Gesetzgebung). Art 74 para. 1 Nr. 19 of the German Basic Law describes this as the competence for “measures to combat human and animal diseases which pose a danger to the public or communicable”.1 This means that the states can pass legislation only until and insofar as the federal state has not done so. With the German Act on the Prevention and Control of Infectious Diseases (Infektionsschutzgesetz, InfSchG) of 20002 the federal state has made 1 For an official version and translation of the Federal Basic Law (Grundgesetz), see http:// geset zeiminter net. de/ gg/. 2 Infection Protection Act of July 20, 2000 (BGBl. I p. 1045), which was last amended by Article 4 of the law of March 18 2022 (BGBl. I p. 473). The law was introduced as Article 1G of. July 20, 2000 I 1045 (SeuchRNeuG) from the Bundestag decided with the consent of the Federal Council. According to Article 5, Paragraph 1, Sentence 1, this law comes into force on January 1, 2001, Sections37 and 38 came into force with effect from July 26, 2000. 105 European Journal of Law and Economics (2025) 59:101–132 use of this power or competence, thus precluding the federal states from passing any formal laws or legal acts. The individual states, however, implement the provisions contained in this act and can pass regulations (in German:“Verordnungen”, i.e. administrative orders) to do so. During the COVID-pandemic, this federal act (InfSchG) saw many rounds of amendments and is the primary source of law governing all measures and decisions passed in relation to the pandemic. In addition, the federal state (Bund) has passed other acts and regulations (pertaining to travel restrictions, labour conditions, protective vaccinations) as well.3 The federal InfSchG contains several substantive measures or rules, such as social distancing, the wearing of masks, the closing of schools, to mitigate the effects of the pandemic and control the spread of the virus. Within the framework of the federal law, the states enforce these measures by first passing specific regulations (executive laws). At the outset of the pandemic, the federal states enjoyed considerable discretion in doing so – in defining the specific measures, their scope and the time-line of their implementation. During the course of the pandemic, however, the scope for the state regulations decreased significantly. The execution or enforcement and thus the practical implementation of these regulations (executive orders) was not done by the states (Länder) themselves but by smaller administrative units, by counties and districts (in city states such as Berlin, Hamburg and Bremen). Several states also passed ancillary regulations on sanctions (fines) and interpretations of the primary regulation.4 The discretion given to the individual states (Länder) changed significantly during the course of the pandemic as ever more uniform rules were set on the federal level which also included implementation guidelines and guidance on which interests to consider when exercising discretion. An assessment/opinion of the legal service of the federal parliament from 29March, 2021 found that the federal state has all-encompassing and far-reaching legislative competences in relation to all measures pertaining to the COVID-pandemic and could also specify the implementation of these measures in full detail. This means that the powers and discretion of the states could have been “reduced to zero”, including measures related to schools and health facilities. By 22April, 2021, the Federal Infection Protection Act (InfSchG) was so detailed and expansive that no significant competition among the individual states could still take place. From this point onward, the competition for announcing and implementing preventative measures was largely replaced by the roll-out of the vaccination. Most of the restrictive measures on federal and state level have now been lifted. The legal framework and the introduction of vaccinations thus defines the timeline and scope of our analysis. Having clarified the scope for action, we now turn to the competitive mechanisms and incentives within the German federal system. 3 COVID-19-Schutzmaßnahmen-Ausnahmenverordnung (SchAusnahmV); Coronavirus-Einreiseverordnung (CoronaEinreiseV); SARS-CoV-2-Arbeitsschutzverordnung (Corona-ArbSchV); Verordnung zum Anspruch auf Schutzimpfung gegen das Coronavirus SARS-CoV-2 (CoronaImpfV); Verordnung zum Anspruch auf Testung in Bezug auf einen direkten Erregernachweis des Coronavirus SARS-CoV-2 (TestV). 4 Find a list here: https:// www. twobi rds. com/ de/ insig hts/ 2021/ germa ny/ covid19veror dnung enundverfu egung enbl. 106 European Journal of Law and Economics (2025) 59:101–132 2.2 Leadership intimesofcrisis The dynamics of political competition in a federal system are multidimensional and include both horizontal and vertical effects. Focusing on the situation of state-level leaders in the German federalism and their horizontal competition among each other reveals a number of interesting competition mechanisms that are relevant for our paper.5 General remarks are made in the following before the special characteristics and effects of an unexpected crisis like the pandemic with its inherent uncertainty are introduced in the second half of this section. First, leading state-level politicians (in particular, state premiers), compete for voters with all the typical caveats and imperfections of the political process in a representative democracy, i.e. imperfect knowledge, asymmetric information, rational ignorance, etc. Here, providing high-quality policy solutions for relevant problems of the population naturally represents a channel of competition. Second, state-level leaders stand in competition with their fellow state-level leaders as voters may observe and assess (both imperfectly) what happens in other German states. Thus, also the relative quality of their policies plays a role. Third, media attention is a relative parameter of political competition. By winning the attention of their own and the other states’ population, state leaders may improve their re-election chances and, at the same time, qualify for a career on the federal level by gaining popularity across the states. Depending on their individual preferences, political leaders may be driven by differing weights of ideology (political beliefs about the best solutions) and career concerns like the probabilities of getting re-elected or getting into attractive political or other (post-politician) positions in the future. Note that ideology and re-election desire may also correlate in the sense that without getting re-elected, no politician can implement political solutions according to his or her ideology. Notoriously, it is difficult or impossible to disentangle these elements of an individual politician’s utility function. In this paper, we put a special emphasis on this dimension as it also sheds some light on the motivation of politicians. We assume that state-level political leaders instrumentalize all three dimensions to position themselves favorably in this horizontal political competition. Strategy choice is not trivial in a multi-dimensional strategy space under imperfect information, wherefore we expect that different political leaders will view different strategies as individually optimal for them due to their individual utility functions and due to different weighing of the pros and cons and the dimensions sketched above. In times of unexpected crisis, the knowledge problem is particularly relevant as there is usually no blueprint for successfully dealing with the crisis problems. In other words, it is ex ante unclear what a “high quality political solution” may be. At the same time, the population expects their leaders to prove themselves as successful crisis managers. The individual weight of ideology and political beliefs in the utility function of political leaders may actually increase in the face of the knowledge 5 In order to keep our paper compact, we do not discuss the huge body of general economics of federalism and political economy in a federal and democratic state. For a recent theory-driven approach with reference also to generally relevant literature see, inter alia, Congleton (2023). 107 European Journal of Law and Economics (2025) 59:101–132 problem since a reliance on one’s own beliefs may fill the void of (external, empirical, available) problem-solving knowledge. Alternatively, and depending on the individual leader, the uncertainty may delay or even paralyze decision making because of the fear of doing the “wrong” thing. Uncertainty also relates to the question of whether voters will actually use the next election as a referendum on the COVID19 policies of the respective political leaders. During the pandemic heydays, this is clearly one of the most important issues, but the next elections may be a long way off, and the pandemic turbulence may (or may not) have receded in importance for voters’ decisions by then. In times of crisis, however, it is unclear ex ante how long the crisis will last and how important the behavior during the pandemic time will be for the next election. This is unknown both to the politicians and the voters. However, we assume that political leaders are interested in their current popularity—to varying degrees (see above)—and it will matter to them how they fare in the eyes of the electorate when they decide what to do. Given the widespread uncertainty on length and course of the pandemic, we think that this assumption is reasonable. This is especially true because the popularity gained through increased media coverage–essentially a popularity that makes a regional politician gets better known at the federal level—is likely to last beyond the pandemic and will be very helpful in ascending in higher office in the future—regardless of whether the pandemic lasts and remains important to voters or not. In such a scenario, for which the COVID-19 pandemic is a representative case, the strategy of acting as a forerunner comes with advantages and disadvantages. A common expectation of political leaders in times of crisis is the effect that the respective politicians in charge take initiative and provide a clear direction in the crisis that the population can follow. Thus, forerunners can be expected to be praised by the press and the public for their decisive action. They may be seen as strong leaders who, in spite of obstacles, move forward with determination, set the course for their state, communicate clearly and often with the citizens, and also enjoy increased popularity in this phase because they “don’t talk, they act”. However, while consequent and clear measures may be popular ex ante, this may change when the population actually suffers from them and when, in the course of time, doubts emerge that these interventions were necessary or the “right” ones. It is possible that voters will think that political leaders have overreacted because of the panic nature of the situation.6 If that happens, the initial popularity for forerunners turns out to be short-lived and the midterm effects may be negative. Latecomers, on the other hand, may represent a natural reaction to the uncertainty of the situation and the lack of (knowledge about) optimal solutions, thus promoting rational imitation or even herd behavior. Nevertheless, these political leaders remain in competition with others and may be perceived as too cautious, lacking in leadership and indecisive in the initial phase of the crisis. Their cautious approach to restricting civil liberties may get sharply criticized and there is pressure to follow the forerunners. However, if the more cautious approach turns out to have saved the population from (perceived or real) unnecessary hard interventions into their daily life and welfare, then the latecomer may gain reputation 6 See e.g. Jones & Baumgartner (2005). 108 European Journal of Law and Economics (2025) 59:101–132 as a “thoughtful person” providing measured policy responses. Thus, the choice whether to act immediately or rather wait is subject to uncertainty. The picture gets even more complicated when mixed strategies are considered. In reality, political leaders do not have to decide to act as a forerunner with all subsequent effects. Instead, they could choose to be a forerunner in announcing consequent interventions and measures but then be careful to codify them into law and/or actually enforce them. By splitting strategies—a forerunner in announcements combined with a latecomer in codification and enforcement—the advantage of media attention (which always focuses on the first calls for actions of a new measure) may be combined with minimizing the danger of getting slated by their own population suffering harder interventions than the citizens in other states. Vice versa, other political leaders may find it attractive to be comparatively silent at first but then prove to be persons of action by being forerunners in codification and enforcement of measures, thus avoiding possible characterizations as being “just loudspeakers” or “populists”. If codification – i.e.the implementation of a law/regulation introducing a certain instrument or measure—can be seen as a signal, then splitting strategies between codification and enforcement may also make sense since only the enforcement will actually hurt people by restricting their life and reducing their welfare. The distinction between legislation (the codification of specific measures and their modes of implementation in state-level regulations) and enforcement (of these measures) is, to the best of our knowledge, a novel contribution to the discussion. Thus, it is important to emphasize that the distinction between legislation and enforcement is not identical to the stringency or degree of enforcement of a policy measure. Being the first to legislate and/or enforce does not automatically imply the adoption of particularly strict policies. During periods of exploding infection rates, there is a relevant likelihood of a race toward stricter policies (among forerunners), whereas during periods of declining infection rates—perhaps accompanied by rising vaccination rates—there may well be a race toward more lenient policies (among forerunners). It is important to emphasize that the quality of the policies—here: avoiding deaths and controlling infection waves—plays the most important role. However, especially under the assumption of imperfect knowledge about the “right” anti-crisis policy— the scope for strategic horizontal competition may be used by political leaders. The following empirical investigation is meant to shed light on the existence of forerunners and latecomers as well as mixed strategies during the pandemic in Germany. Of course, we control for the main goal—the combat of death tolls and infection rates. Before doing so, we provide a brief overview of the data used in our study. 3 Data anddescriptive statistics As pointed out above, forerunners and latecomers can be identified from different perspectives. (1) who is first (last) in the media addressing a crisis-related aspect, (2) who is first (last) to cast a policy measure into law, and (3) who serves as a role model in law enforcement. Using different data sources, we will try to identify the forerunners and the latecomers in the three categories: first (last) in media, 115 European Journal of Law and Economics (2025) 59:101–132 3.4 Law enforcement andits contextual factors Of the 24 aggregated measures by the statistical office reported in Table 7 of the appendix, we only use 21 categories of protective policies. Some of the measures had not been used by anyone and therefore do not provide any further information about federal states’ pandemic history. We have chosen the period from March 1, 2020 to January 15, 2022. Before this period, no COVID policy had been implemented as the virus had not provoked any political action. By the end of 2021, the omicron variant started to spread. Since it was perceived as a weaker variant, the virus and pandemic lost more and more of its perilous nature, especially as a large majority of the population had received its second vaccination. Moreover, the invasion of Russia in Ukraine shrunk the media-relevance of the COVID pandemic even further. Hence, we ended up with the information about 21 policy categories from 400 NUTS-3 regions from 16 states for 686 days, i.e., in total 5,762,400 data points. As we are interested in forerunners and latecomers, we only look at the events when a policy was introduced. Within the period under consideration, we identified 47,301 moments when a COVID policy was introduced. Table 4 shows the variables that we calculated from the available information. Because many policies were introduced, abolished, and reintroduced several times, we count more than just a single introduction of a given policy in a certain county. Every policy was introduced at least 4 times, some even 25 times. The statistical office provides further information on the NUTS-3 level which we will also exploit. The remaining variables in the table are industry structure, which ranges between 10 and 30% and measures the share in total sales of near-personal services (Hotels and restaurants, Education, Health and Social Services, and Arts, Entertainment and Recreation); total population, and population density; the number of installed policies in all other states as a measure of imitation (technically, a measure of spatial correlation); the number of employees in partial employment (i.e. Kurzarbeit or in short: KU). Before our empirical exercise, we present a simple model – in line with our theoretical reflections – that motivates the propensity of regional leaders to implement a COVID policy. Table 4 Summary statistics on NUTS-3-level Overall, 400 NUTS-3-regions have been identified. As we only take the information about the day a region introduces a policy, the number of observations decreases to 47,301 Variables N Mean Stand. dev. Min Max Number of policies 47,301 8.364 5.528 1 21 Number of previous policies 47,301 4.0 3.6 0.0 25.0 Industry Structure 47,301 0.20 0.00 0.10 0.30 Total population (NUTS-3, in thousands) 47,301 210.8 231.6 34.0 3664.0 Population density (per square kilometer) 47,301 553.6 712.8 36.0 4790.0 Deaths 47,301 2.8 2.1 0.0 33.3 Number of policies in remaining states 47,301 1.7 0.7 0.0 3.2 Number of employees in KU 47,301 3076.0 8847.0 0.0 255,368.0 116 European Journal of Law and Economics (2025) 59:101–132 4 The role ofpolicy makers In this section, we model the enforcement behavior of politicians, i.e. the circumstances under which they decide to introduce one of the 21 protection measures. During the COVID pandemic, policy makers had to decide on the kind and whether or not to introduce protective policies. As a gauge, the infection rate, which we label i in the following, was agreed on. In Germany’s federal system, each state had some scope in the decision about the ‘right’ timing of a further measure. Forerunners are expected to be among the first to introduce a required policy measure, compared to latecomers. However, calendar time is not the right indicator to determine who among state leaders can be considered a forerunner or a latecomer. The mere progression of time is inconsequential in the implementation of a policy intervention in the absence of a change in infection rate: no infection, no COVID policy, regardless of the time that passes. A leader can only be considered a forerunner if he or she decides to introduce a protective policy more decisively than others at a given level of infection, that is, in a specific pandemic situation. In other words, leadership is demonstrated by the determination to make an unpopular but necessary decision at the right (infection) time. The decision itself depends on many determinants, for which we have to control as argued in the theoretical part above, the decision to implement a protective policy at a given infection rate i has to be considered as a random variable. Therefore, we consider the propensity to adopt the policy m of the leader l as a random variable 𝜆(i) while controlling for contextual, regional factors. For this reason, and in contrast to traditional survival studies, we do not use calendar time as the duration parameter, but the actual infection rate i.7 4.1 Propensity toimplement COVID policies As we look for the probability that a political leader introduces a protective measure given infection rate i and contextual factors, it is statistically tantamount to use survival analysis. The cumulative density of a policy measure that has not yet been implemented at infection rate i is P(I>i)=1−F(i)=S(i) . Hence, the probability of state leader l to introduce a COVID-policy m is8: with fm l(i) as the probability density function of leader l with respect to measure m. Equation (1) is therefore equivalent to the hazard function. It is the instantaneous rate of introducing a policy at survival-time i. In simpler terms, the denominator in Eq.1) states the probability that a certain protective measure m has not yet been introduced at the given infection rate i; the numerator states the probability (1) 𝜆 m l(i)= lim di→∞ P(i≤I<i+di|I≥i) di = fm l (i) 1−Fm l (i ) 7 A further advantage of this procedure is that all time-variant variables, which we discuss below, are time-invariant with respect to survival-time i, though they are time-variant with respect to calender timet. 8 Conceptually, our model is aligned to the work by Agarwal and Gort (2002). 117 European Journal of Law and Economics (2025) 59:101–132 that measure m will be (re-)introduced by leader l instantly given i. It expresses the conditional expectation. We hypothesize that, in addition to the infection rate, the probability of political action depends on the state leaders l’s baseline determination Dl(0) , regional calendar-time-invariant factors Rl (e.g. industry structure, population density) as well as calendar-time-variant determinants, i.e the current state of the COVID pandemic Il(t) such as the intensity of restrictions experienced so far, the number of deaths, or the vaccination rate; furthermore, the contemporaneous economic situation El(t) (the greater the economic damage of previous COVID-policies, the greater the reluctance of policymakers to increase restrictions) plays an important role and last not least, under strong uncertainty, the extent to which policymakers imitate other leaders L −l( t ) due to lack of better knowledge. 4.2 Leaders’ baseline determination For readability we leave out superand subscripts m and l, in the following. We assume leaders to have an individual and constant baseline determination to impose restrictions for protective reasons. This is what makes a leader either a forerunner or a latecomer in enforcement. It is assumed as a fixed personality trait that determines the leader’s baseline determination Dl(0) to introduce a restrictive COVID policy measure by evaluating the trade off between citizens’ freedom and the protection of their health. The determination changes as the pandemic progresses. The change can be expressed as: where the determination Dl(i) depends on: which is, the leader l’s baseline determination Dl(0) , the determination Dl(i−1) given the previous infection state i−1 , the current pandemic situation Il(t) , regional time-invariant specificities ( Rl ), the region’s economic situation ( El(t) , i.e. number of workers in subsidized part-time employment, henceforth: Kurzarbeit, KU), and the current policy actions of remaining states L−l(t) . To allow for behavioral uncertainty, we include a disturbance 𝜀∼N(0, 𝜎) – as previously discussed in the introductory section of this paper. Whereas the leaders’ baseline determination Dl(0) does not change, the course of the pandemic will change his/her actual determination. When the situation gets more severe and the human cost of the pandemic increases (deaths), the propensity to introduce further measures will increase ( ⇒ 𝜕f ∕ 𝜕I l( t ) > 0 ), though making the population suffer more and more. Regional specificities will also impel leaders to act accordingly. For instance, a region’s industry structure with a high share in services that involves high-frequent human interaction will increase the leader’s propensity to pass further COVID (2) ΔDl(i)=Dl(i)−Dl(i−1) (3) Dl (i)=f ( D l (0),D l (i−1),I l (t),R l ,E l (t),L −l (t),𝜀(i) ) 118 European Journal of Law and Economics (2025) 59:101–132 policies ( ⇒𝜕f∕𝜕Rl>0 ), although the sign of the derivative is unclear. A high population density requiring more immediate policy actions may make a leader more reluctant to impose restrictions, as he/she faces substantial political (decline in reputation) as well as socio-economic cost (reduction of citizens’ well-being). When the pandemic starts to affect a region’s economy, negatively, the leader’s willingness to increase the burdening pressure of further policies will decline ( ⇒𝜕f∕𝜕El(t)>0 ). An aspect that also will influence a leader’s decision making process is to imitate one’s neighbors, that is, looking at one’s peers and act alike ( ⇒ 𝜕f∕𝜕L −l t > 0 ) (=imitation). 4.3 Leaders’ policy action function Assembling all elements from above, we can now formulate the corresponding policy action function (=hazard function): where h(i) is the hazard rate (or determination) of leader l to introduce a further COVID policy at infection rate i, 𝛿l captures the leader’s baseline determination and f denotes the employed hazard function. In our regressions below, we use the Weibull distribution as hazard function: The Weibull distribution allows us to consider different shapes of the development of the baseline hazard rate. It may decrease, increase, or remain the same during the pandemic. Parameter p captures the corresponding trend. 5 Results To start with, we perform a survival analysis employing a non-parametric cumulative hazard function to show states’ determination to introduce a restrictive COVID policy without considering any covariates (contextual variables). Afterwards, we run a parametric survival model while including states’ context. It is crucial to highlight that the discussion here focuses on the role of law enforcement in implementing stringent protective measures at the sub-regional level, particularly within the 400 NUTS-3 regions. In other words, we derive state leaders’ determination by aggregating decisions made at the sub-regional level. 5.1 Non‑parametric cumulative hazard function The cumulative hazard function by states is graphed in Fig.4. The solid black line indicates the baseline hazard to introduce a COVID policy across federal states. The (4) h(i|x)=f(𝛿l+𝛽R R +𝛽I I +𝛽E E +𝛽L L −l) (5) h0(i|x)=p ⋅ i p − 1 exp( 𝛿 l+ 𝛽 R R + 𝛽 I I + 𝛽 E E + 𝛽 L L −l) 119 European Journal of Law and Economics (2025) 59:101–132 bands shaded in light gray mark the inter-quartile range. Upon initial examination, it appears that the four most cautious states, which may be considered forerunners, are BY, NW, BW, and RP. The four least prudent states; HH, BE, BR, and SC, although SC returns to the inter-quartile range at higher infection rates. Since we do not consider any contextual variables in the figure, the ranking of forerunners and latecomers is not very differentiated. States are not homogeneous, as argued in the introduction. For instance, they differ in industry structure and population density, both determinants that may lead policy makers to different conclusions regarding COVID policies. The actual COVID situation itself will call for specific policies. Furthermore, a densely populated region with a high share of vaccinated people allows for less restrictive policies than in other regions that lag behind in immunization coverage; in regions with an economically difficult situation or a high population density, politicians will likely hesitate to introduce a restrictive and thus additionally restrictive policy measure; the uncertainty that is involved in a hitherto unknown pandemic will exacerbate a stringent purposeful policy; therefore, policy Fig. 4 Cumulative Hazard by States (CHF). Note: The solid line in the middle indicates the average baseline hazard ( h0(i) ) as cumulative hazard function dependent on survival i, i.e. the infection rate; gray bands indicate the interquartile range 120 European Journal of Law and Economics (2025) 59:101–132 makers will also look at neighbors how they cope with the pandemic (imitation). We will address all these determinants in the following survival model. 5.2 Survival regression analysis In Table5 we start with a semi-parametric proportionate hazard model (Cox, 1972). The Cox model (Model 1) is presented in the first column of this table. As covariates we included several groups of variables. First, we look at the regional characteristics. An industry structure with a high share in near-person services forces policy makers to pass COVID policies and so does a high population density because it facilitates the spreading of the virus. A higher population density, however, conversely increases the reluctance to introduce further restricting policies – as argued in the introduction. Second, the current COVID situation in the region, such as the vaccination rate, the number of deaths from which a region suffers will influence the decision making process of politicians. The more dead people, the higher leaders’ determination to introduce further measures. The more advanced the vaccination rate, the fewer new strict measures will be introduced. Third, the involved uncertainty in policy making shows up in the degree of imitation. Leaders of a given region will introduce COVID measures, if the remaining ones have done so. Note that the timely association is not on calendar time but on survival time i. In other words, if other regions introduced certain measures at a given infection rate i, so will the respective region under consideration. As the positive coefficient suggests, there is imitation (spatial correlation) among regions to a significant extent. Fourth, we include a naive partisan definition of state governments (Maggetti & Trein, 2021; Russo & Verzichelli, 2016): left, center, and right, where the base category of this categorical variable ‘Political stance’ is ‘center’ government. The three categories are based on whether the governing coalition during the pandemic can be considered left, center, or right. We also took into account changes in government during the pandemic. Fifth, a difficult economic situation makes politicians more reluctant to impose restrictive policies. This is what the negative coefficient for the variable labeled ‘Number employees in KU’ suggests. An increasing ‘Kurzarbeit’, i.e partial employment subsidized by the state, will reduce policy makers’ willingness to further weaken the economy with COVID policy measures. In Model (1), we apply the Cox model, a semi-parametric model assuming a proportional hazard function. In Model (2), we employ an exponential hazard function as a first robustness test of our results. The signs do not change and the coefficients remain stable, except for the first vaccination coefficient that becomes negative and significant. As in Model (1), we introduce state dummies, which are fixed effects that we will utilize subsequently to derive the ranking of states – our main objective in this paper. Additionally, we introduce dummies for the type of the 23 measures. The drawback of the previous models is that they assume a constant baseline hazard which is rather unlikely in the case of the spreading of a virus. An exponential 121 European Journal of Law and Economics (2025) 59:101–132 Table 5 Survival regression models Standard errors in parentheses, *** p<0.01 , ** p<0.05 , * p<0.1 . Note: Each column presents regression coefficients: (1) a semi-parametric Cox regression, parametric survival regressions assuming (2) an Exponential, (3) and (4) a Weibull probability distribution, respectively. Model (4) includes an indicator variable for institutional change (IC), when the federal state restricted the scope of each state. We do not report the constant term in regressions Time variable: infection rate Cox Exponential Weibull Weibull (IC) (1) (2) (3) (4) Regional characteristics Industry structure 0.047 ∗∗∗ 0.059 ∗∗∗ 0.050 ∗∗∗ 0.049 ∗∗∗ (0.002) (0.002) (0.002) (0.002) Population density − 0.218 ∗∗∗ − 0.298 ∗∗∗ − 0.245 ∗∗∗ − 0.231 ∗∗∗ (0.008) (0.011) (0.009) (0.009) Current COVID performance Number of 1st vaccination (= 1st vac) − 0.001 − 0.012 ∗∗∗ − 0.007 ∗∗∗ − 0.048 ∗∗∗ (0.002) (0.003) (0.003) (0.003) Number previous policies − 0.068 ∗∗∗ − 0.078 ∗∗∗ − 0.070 ∗∗∗ − 0.082 ∗∗∗ (0.002) (0.003) (0.002) (0.003) Deaths 26.059 ∗∗∗ 30.498 ∗∗∗ 27.996 ∗∗∗ 36.390 ∗∗∗ (2.278) (4.249) (3.610) (3.868) Policy imitation Number policies remaining states − 0.287 ∗∗∗ − 0.424 ∗∗∗ − 0.319 ∗∗∗ − 0.600 ∗∗∗ (0.012) (0.017) (0.014) (0.016) Policy stance Left-wing − 0.382 ∗∗ − 0.514 ∗∗∗ − 0.446 ∗∗∗ − 0.810 ∗∗∗ (0.167) (0.165) (0.137) (0.136) Right-wing 1.118 ∗∗∗ 1.784 ∗∗∗ 1.465 ∗∗∗ 1.196 ∗∗∗ (0.190) (0.222) (0.184) (0.194) Economic situation Number employees in KU − 0.008 ∗∗∗ − 0.011 ∗∗∗ − 0.009 ∗∗∗ − 0.012 ∗∗∗ (0.001) (0.001) (0.001) (0.001) Institutional change Centrally managed = 1 0.788 ∗∗∗ (0.018) Disproportionate hazard ln(p) − 0.177 ∗∗∗ − 0.159 ∗∗∗ (0.004) (0.004) State dummies Yes Yes Yes Yes Policy dummies Yes Yes Yes Yes Observations 42,209 42,209 42,209 42,209 LL -401,897 -73,921 -72,663 -71,683 122 European Journal of Law and Economics (2025) 59:101–132 spreading, as is the case with COVID will also affect policy makers sensitivity to act and fight the pandemic in order to keep social and economic cost low. Using a Weibull hazard function will allow estimating a disproportionate hazard rate. Model (3) reports the corresponding results with the Weibull hazard function. Again, the estimated coefficients remain robust compared to models (1) and (2). Parameter p indicates the baseline hazard with p=exp [ln(p)]=.83 <1 . The higher the infection rate once reached, the lower the marginal tendency to take additional actions against COVID, ceteris paribus. Bavel etal. (2020) name various factors that explain the decreasing baseline hazard rate when the infection rate i increases: policy measures are only effective if the public complies with them. As infection rates rise, people may become desensitized to the severity of the situation. Continued exposure to high infection rates and the ongoing nature of the pandemic may lead to a normalization effect, where high numbers become part of the “new normal”. As a result, individuals and communities may perceive the situation as less alarming, reducing the perceived urgency to take further action. This public perception in turn has politicians become more reluctant to take further action.9 Moreover, there are several interactions between variables that we do not consider because they are not our main focus, that is, to derive a ranking of state leaders taking into account contextual factors.10 In Model (4), we include an indicator variable for the time after when the COVID policy was centralized end of April 2021 by the Federal Infection Protection Act (InfSchG). In contrast to Cepaluni etal. (2022), who claim that in the short run a high state capacity is better suited to tackle a pandemic, the German case follows the Canadian example—the latter described in Migone (2020): at first, local actors dominate, and gradually the federal state takes over. This eliminated largely competition between states. As the estimate suggests, centralization significantly increased the propensity to enforce restrictive measures. Overall, the results seem to be robust across models. The coefficients have the expected signs: Regional characteristics increase the likelihood of introducing another measure as the degree of social contact implied by the industry structure increases, and decrease the likelihood at high population densities as social acceptability decreases; the higher the vaccination rate, the less likely further protective measures are; the more policy measures already installed (number of previous policies), the lower the likelihood of further action; the more people die, the more likely a further protective measure. In terms of policy imitation, states seem to be less strict than their neighbors, at first sight. It is plausible that policy imitation interacts with the policy stance of the respective region’s government. Right-wing (left-wing) governments act more (less) strictly than centrist governments11; a tense economic situation (Number of employees in KU) reduces the likelihood of another measure 9 See Bavel etal. (2020) or Petherick etal. (2021) for further explanations. 10 To do so, it is not necessary to delve into the interdependencies of contextual factors. 11 Adding up the coefficients of ‘Policy Imitation’ and ‘Policy Stance’ weighted by NUTS-3 subregions leads to an overall positive policy imitation effect. Since we do not intend to study the granular interdependencies of political behavior, but rather the ranking of state leaders during the COVID pandemic, we do not consider further moderation effects, here. This will be left to a follow-up study. 123 European Journal of Law and Economics (2025) 59:101–132 being introduced. The centralization of power by the federal government significantly increased the likelihood of policy measures being implemented. It should be emphasized here that we are not interested in the specificities of policymakers’ political interdependencies in this paper, but only want to control for contextual factors in order to distinguish forerunners from latecomers in policy action. We now focus on the state fixed effects, i.e., state dummies, which indicate the statespecific baseline hazard of enforcing policies related to the COVID pandemic. The respective coefficients allow us to rank states according to their baseline hazard, i.e. the extent to which states are determined to strictly enforce protective measures. Together with the rankings identified with regard to media announcement and law codification, we can now confront these rankings with the relative performance of states in law enforcement – while considering the heterogeneity of states. Figure5 presents the performance of countries in the three competitive fields: Announcers, Codifiers, and Enforcers. The left vertical axis indicates the ranking based on the search in (print) media (LexisNexis database), the middle axis refers to the search in the Juris database, and the right axis to the ranking in law enforcement according to the Destatis data, controlling for state heterogeneity. The latter is based on the state fixed effect, which can be interpreted as the leaders’ baseline determination to enforce restrictive measures. Correspondingly, the forerunners in announcing are BW, BE, SR, and BY; latecomers are MV, SC, BR, and NI. The first four forerunners in codifying COVID-measures are HE, BE, BW, and MV; the latecomers are BB, BR, HH, and SR. With regard to enforcers, the forerunners are RP, BY, SH, and NI, the latecomers HH, BE, SC, BR. Figure5 illustrates the change in rankings. MV, Fig. 5 Rankings: Announcers, Codifiers, and Enforcers. Note: The three axes represent the rankings of the 16 states concerning the announcement of COVID-policy measures, the codification of these measures, and their enforcement. Dashed (solid black) lines indicate countries with decreasing (increasing) rankings in terms of the three rankings announcement, codification and enforcement 124 European Journal of Law and Economics (2025) 59:101–132 Table 6 Forerunners vs. Latecomers The first half of the table indicates forerunners and latecomers according to the ranking they achieved in the three competitive fields: announcers (a), codifiers (c), and enforcers (e). In the second half of the table, the forerunners and latecomers are listed in order of greatest difference in the pairwise comparison of each state’s ranking. Comparing ranks such as a>c means that the rank in a (announcing) is higher than the rank in c (codifying). Also, see Fig.5) Criterion Forerunners Latecomers Announcers (a) BW BE SR BY NI BR SC MV Codifiers (c) HE BE BW MV SR HH BR BB Enforcers (e) RP BY SH NI BR SC BE HH Behavioral types Forerunners Latecomers procrastinators (a > c) SR BY BB HH SC TH HE MV Armchair activists (c > e) BE HE NW BW NI SR BB BY Slient policymakers (e > a) MV NI RP SH BW HH NW BE 131 European Journal of Law and Economics (2025) 59:101–132 directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/ licenses/by/4.0/. References Agarwal, R., & Gort, M. (2002). Firm and product life cycles and firm survival. American Economic Review, 92(2), 184–190. Barberá, P., Casas, A., Nagler, J., Egan, P. J., Bonneau, R., Jost, J. T., & Tucker, J. A. (2019). Who leads? who follows? measuring issue attention and agenda setting by legislators and the mass public using social media data. American Political Science Review, 113(4), 883–901. Bavel, J. J. V., Baicker, K., Boggio, P. S., Capraro, V., Cichocka, A., Gelfand, M., Sanchez, R. G., Heslenfield, S., Keltner, D., Han, S., Kenrick, D. F., van der Linden, S., Lonergan, E. A., Puryear, C., Rathje, S. G., Reinero, M., Peitz, L. T., Willer, R., Jordan, J. J., … Morris, M. W. (2020). Using social and behavioural science to support Covid-19 pandemic response. Nature Human Behaviour, 4(5), 460–471. Besley, T. J. and Case, A. (1992). Incumbent behavior: Vote seeking, tax setting and yardstick competition. Bjørnskov, C., & Voigt, S. (2022). This time is different?—On the use of emergency measures during the Corona pandemic. European Journal of Law and Economics, 54(1), 63–81. Bolton, P., & Farrell, J. (1990). Decentralization, duplication, and delay. Journal of Political Economy, 98(4), 803–826. Broschek, J. (2022). Federalism, political leadership and the Covid-19 pandemic: Explaining Canada’s tale of two federations. Territory, Politics, Governance, 10(6), 779–798. Carozzi, F., Provenzano, S., & Roth, S. (2024). Urban density and Covid-19: Understanding the US experience. The Annals of regional science, 72(1), 163–194. Cepaluni, G., Dorsch, M. T., & Branyiczki, R. (2022). Political regimes and deaths in the early stages of the Covid-19 pandemic. Journal of public finance and public choice, 37(1), 27–53. Cerqueti, R., Coppier, R., Girardi, A., & Ventura, M. (2022). The sooner the better: Lives saved by the lockdown during the Covid-19 outbreak. The case of Italy. The Econometrics Journal, 25(1), 46–70. Congleton, R. (2023). Federalism and pandemic policies: Variety as the spice of life. Public Choice, 195(1), 73–100. Cox, D. R. (1972). Regression models and life-tables. Journal of the Royal Statistical Society: Series B (Methodological), 34(2), 187–202. Färber, G. (2021). Germany’s fight against Covid-19: The tension between central regulation and decentralised management. In Comparative Federalism and Covid-19, pages 51–69. Routledge. Gaenssle, S., & Budzinski, O. (2021). Stars in social media: New light through old windows? Journal of Media Business Studies, 18(2), 79–105. Gaenssle, S. and Budzinski, O. (2024). 24 The economics of influencers and social media stardom. Walter de Gruyter GmbH & Co KG. Garzarelli, G., & Keeton, L. (2018). Laboratory federalism and intergovernmental grants. Journal of Institutional Economics, 14(5), 949–974. Garzarelli, G., Keeton, L., & Sitoe, A. A. (2022). Rights redistribution and Covid-19 lockdown policy. European Journal of Law and Economics, 54(1), 5–36. Greer, S. L., Rozenblum, S., Falkenbach, M., Löblová, O., Jarman, H., Williams, N., & Wismar, M. (2022). Centralizing and decentralizing governance in the Covid-19 pandemic: The politics of credit and blame. Health Policy, 126(5), 408–417. Harrison, J. L. (1978). Yardstick competition: A prematurely discarded form of regulatory relief. Tul. L. Rev., 53, 465. Hegele, Y., & Schnabel, J. (2021). Federalism and the management of the covid-19 crisis: Centralisation, decentralisation and (non-) coordination. West European Politics, 44(5–6), 1052–1076. Jones, B.D. and Baumgartner, F.R. (2005). The politics of attention: How government prioritizes problems. University of Chicago Press. Kerber, W., & Budzinski, O. (2003). Towards a differentiated analysis of competition of competition laws. Zeitschrift für Wettbewerbsrecht, 1(4), 411–448. 132 European Journal of Law and Economics (2025) 59:101–132 Kollman, K., Miller, J. H., & Page, S. E. (2000). Decentralization and the search for policy solutions. Journal of Law, Economics, and Organization, 16(1), 102–128. Maggetti, M., & Trein, P. (2021). More is less: Partisan ideology, changes of government, and policy integration reforms in the UK. Policy and Society, 40(1), 79–98. Migone, A. R. (2020). Trust, but customize: Federalism’s impact on the Canadian Covid-19 response. Policy and Society, 39(3), 382–402. Mistur, E. M., Givens, J. W., & Matisoff, D. C. (2023). Contagious Covid-19 policies: Policy diffusion during times of crisis. Review of Policy Research, 40(1), 36–62. Petherick, A., Goldszmidt, R., Andrade, E. B., Furst, R., Hale, T., Pott, A., & Wood, A. (2021). A worldwide assessment of changes in adherence to Covid-19 protective behaviours and hypothesized pandemic fatigue. Nature Human Behaviour, 5(9), 1145–1160. Ruggeri, K., Stock, F., Haslam, S. A., Capraro, V., Boggio, P., Ellemers, N., Cichocka, A., Douglas, K. M., Rand, D. G., Van der Linden, S., etal. (2024). A synthesis of evidence for policy from behavioural science during Covid-19. Nature, 625(7993), 134–147. Russo, F., & Verzichelli, L. (2016). Government ideology and party priorities: the determinants of public spending changes in Italy. Italian Political Science Review/Rivista Italiana di Scienza Politica, 46(3), 269–290. Samad, A., Al Jerjawi, K., & Dadich, A. (2022). Crisis leadership: Political leadership during the Covid19 pandemic. Sustainability, 15(1), 266. Segatto, C. I., dos Santos, F. B. P., Bichir, R. M., & Morandi, E. L. (2022). Inequalities and the Covid-19 pandemic in Brazil: Analyzing un-coordinated responses in social assistance and education. Policy and Society, 41(2), 306–320. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.