European policy analysis 

About: European policy analysis is an academic journal published by Wiley. The journal publishes majorly in the area(s): Medicine & Politics. It has an ISSN identifier of 2380-6567. Over the lifetime, 39 publications have been published receiving 64 citations. The journal is also known as: EPA.

The influence of democracy, governance and government policies on the COVID‐19 pandemic mortality

Abstract: Abstract The paper analyses the determinants of mortality during the Corona pandemic. In a first step, possible causes are subjected to a global comparison.The focus is on political, institutional, economic, demographic and health policy factors. It is shown that, contrary to the assumption, democratic countries have a higher Covid‐19 mortality. In a second step, the developed democracies are then analysed to explain this puzzle. Here, more detailed information is used than in the global comparison. It turns out that, measured by the Oxford Stringency Index, government action has been largely unsuccessful. Also, the party‐political composition of the government does not play a role.The most important factors are a country's governance structures, the level of corona incidence, a country's burden of high‐risk patients and its health system. In addition, cultural factors and the vaccination rate seem to have an influence on mortality.

Analyzing national policy styles empirically using the Sustainable Governance Indicators (SGI): insights into long‐term patterns of policy‐making

Abstract: The concept of national policy styles differs between states in respect of whether their governments react to policy changes in an anticipatory or reactive fashion and whether they seek to achieve consensus with societal actors or impose decisions on them. To date, this conceptualization has been applied to a limited number of states and produced only a small set of case studies due to the absence of large-n data. We assess whether the dimensions on strategic planning and public consultation of the Bertelsmann Sustainable Governance Indicators (SGI) provide conceptually sound and empirically insightful indicators of national policy styles. Our explorative analysis reveals that the SGI are useful for operationalizing the concept of national policy styles and could advance the debate on it. Our analysis shows that differences exist between countries in terms of their policy styles, and that the policy styles remained stable in most countries between 2014 and 2020. 国家政策风格 (national policy styles) 的概念在不同国家中存在差异, 具体表现为政府是否以预期性或反应性的方式响应政策变革, 以及政府是否试图与社会行动者达成共识或对其施加决策。迄今为止, 这一概念已被应用于一小部分国家, 并且仅产生了为数不多的案例研究 (因缺乏大样本数据) 。我们评估了可持续治理指标 (SGI) 的战略规划和公共咨询维度是否对国家政策风格提供理论健全且具有实证意义的指标。我们的探究性分析显示, SGI有助于对国家政策风格的概念进行操作化, 并且能促进相关辩论。我们的分析显示, 各国在政策风格上存在差异, 并且大多数国家在2014-2020年间的政策风格保持稳定。 El concepto de estilos de política nacional difiere entre estados con respecto a si sus gobiernos reaccionan a los cambios de política de manera anticipatoria o reactiva y si buscan lograr el consenso con los actores sociales o imponerles decisiones. Hasta la fecha, esta conceptualización se ha aplicado a un número limitado de estados y ha producido solo un pequeño conjunto de estudios de casos debido a la ausencia de datos de n grande. Evaluamos si las dimensiones sobre planificación estratégica y consulta pública de los Indicadores de Gobernanza Sostenible (SGI) proporcionan indicadores conceptualmente sólidos y empíricamente perspicaces de los estilos de política nacional. Nuestro análisis exploratorio revela que los SGI son útiles para operacionalizar el concepto de estilos de política nacional y podrían avanzar en el debate al respecto. Nuestro análisis muestra que existen diferencias entre los países en términos de sus estilos de política y que los estilos de política se mantuvieron estables en la mayoría de los países entre 2014 y 2020. Policy analysis is a discipline characterized by a strong interest in conceptual and theoretical debates. One of them concerns national policy styles, a concept originally formulated by Richardson et al. (1982), which contends that long-term patterns can be observed in the manner in which policies are formulated and implemented in different countries (see also Richardson, 1982). Richardson et al.'s (1982) conceptualization of national policy styles distinguishes between two dimensions. The first is how government approaches problem-solving, which can be anticipatory or reactive. The second refers to the relationship between government and other actors in the policy process, which can be characterized by the government seeking to reach consensus with organized groups or imposing decisions on them. This perspective has gained considerable attention in the last few years as policy scholars have grappled with issues such as explaining the different responses of countries to the COVID-19 pandemic (Capano et al., 2020), and whether these strategies are fixed and/or similar to the way in which these states have dealt with other issues in the past (Howlett, 2021). In order to explain these tendencies, various scholars have attempted to breathe new life into the study of policy styles (Howlett & Tosun, 2019a, 2021); among them, Jeremy Richardson, who was one of the inventors of the concept (Mazey & Richardson, 2020; Richardson, 2018a, 2018b). Most of these empirical applications are case studies that first detail the various ways in which specific countries habitually deal with ongoing issues, then juxtapose them with others in a comparative, case-based analysis. One of the reasons why the literature has been dominated by case studies on policy styles is that it is not clear a priori what would constitute appropriate data for comparing a large number of countries over time. This results from a conceptual ambiguity since it is not immediately apparent if a national policy style—the “dependent variable”—is an “output,” a characteristic choice of policy instruments, or the process through which choices are made. Nor is it clear what are the “independent variables” that should be examined in order to understand how policy styles emerge and evolve. While policy styles are evidently linked to institutional factors that constrain the options available to decision makers, how these should be measured or operationalized remains problematic. But another reason for the dominance of case studies is that no attempts have been made to develop a database that explicitly seeks to capture the characteristics of national policy styles and how they evolve over time. In this study, we assess the suitability of the Sustainable Governance Indicators (SGI) maintained by the Bertelsmann Foundation as a means for operationalizing national policy styles for a large number of countries. The SGI were not developed for this purpose. But since, at a more abstract level, national policy styles concern which countries operate according to the logic of “government” and which according to that of “governance” (Richardson, 2018b), the SGI offer some indicators that qualify as proxies for gauging the two dimensions of national policy styles. More precisely, and as we will discuss in depth in later sections, the SGI on “strategic planning” and “public consultation” align closely with the concept of national policy styles as conceptualized by Richardson et al. (1982). Our findings show that the two SGI indicators selected produce scores that vary across countries. Interestingly, for most countries, they exhibit very little or no cross-temporal variation during the observation period, which is from 2014 to 2020. We believe that this supports our view on national policy styles as a concept that stresses the entrenched institutional features of political systems and how they operate. Nonetheless, the SGI data also show that changes in the indicators can occur in the short- to medium term, and this finding requires enhanced scholarly attention (see Tosun et al., 2022). The remainder of this study unfolds as follows. We first briefly define the concept of national policy styles and then turn to the SGI and their approach to measuring governance arrangements. Subsequently, we provide an empirical assessment of the two SGI indicators, which we consider to approximate the two dimensions of national policy styles. Then, we return to the concept and draw conclusions based on the empirical findings obtained. In the final section, we offer some concluding remarks. Policy-making represents one of the key functions of any political system, irrespective of the nature of the political regime in place. Therefore, it comes as little surprise that, for decades, scholarship in comparative public policy has been interested in determining if there are any distinct patterns of policy-making, so-called “standard operating procedures” (Richardson, 2018b, 215), in different countries and sectors of activity (see Knill & Tosun, 2020). This has been the focus of much work in comparative public policy, which has generally found limited variation in generalized processes and outcomes that are affected by common factors such as globalization, transnational policy learning, and diffusion (Drezner, 2001; Stone, 2004). However, while this view once generated a “convergence thesis,” implying that national and local differences would inevitably dissolve and all governments would move toward common ways of doing things and common processes for arriving at those outputs, the persistence of distinctive national efforts in many areas has eroded confidence in this (Bennett 1991; Weiss 1999). Whether, to what extent, and why governments might develop distinctive “styles” of policy-making and why these differences might persist (Simmons et al., 1974) have been outstanding questions since the first major study of the subject carried out in the volume by Richardson (1982). Identifying such styles and where they exist is thus an essential task for comparative policy analysis. While it is clear such styles affect the shape and character of policy outputs and outcomes in important ways and must be resistant to the more general forces of exogenous events, most other aspects of them are unknown or understudied. It is thus unclear, for example, how much a certain style affects policy choices and if solutions emerge from carefully crafted formulation processes or are more heavily influenced by other processes such as political or electoral bargaining, as the perspective of policy design hypothesizes (Howlett & Mukherjee, 2014, 2021). The extent to which evidence and information influence such policy choices could have an important impact on the character and effectiveness of a policy, but might not or at least might have to coexist with “traditional ways of doing things” in the civil service or bureaucracy, for example (Biesbroek et al., 2018; Knill, 2001; Peters, 2021). This is not purely a historical or speculative enterprise. As studies of the COVID-19 pandemic have noted, policy styles can be a helpful perspective for explaining different national responses to this health crisis (Bakir, 2020; Capano et al., 2020; Howlett, 2021; Mazey & Richardson, 2020; Mei, 2020; Migone, 2020; Zahariadis et al., 2021, 2022). Both contemporary and historical studies examining such styles have tended to focus on the national level and contend that policy-making processes need to be studied in light of how the institutional make-up of a government generates the routines and standard operating procedures that constitute its decision making structure and processes. While the debate about the extent to which each policy sector tends to develop its own style remains unresolved (Cairney, 2021; Freeman, 1985; Knill & Tosun, 2020), the working assumption in studies of national policy styles is that similar kinds of regimes develop roughly similar styles.1 But exactly which institutions matter, and when, and how remain central questions in the field. The key contribution to the study of policy styles made by the volume edited by Richardson (1982) was that it proposed a simple yet elegant and parsimonious way of answering these questions. The contributions in that book investigated the contemporary similarities and differences among the countries of Western Europe in terms of differences in their propensity to take anticipatory or reactive decisions and, in either case, whether they tended to do so by seeking to reach consensus with organized groups or by imposing decisions on them, notwithstanding opposition from such groups. This conceptualization of the nature and significance of national policy styles, put forth by Richardson et al. (1982), commenced an intellectual process out of which alternative conceptualizations have emerged and been critically appraised (Howlett & Tosun, 2019a, 2021; Knill & Tosun, 2020; Unger & van Waarden, 1995; Zahariadis et al., 2021, 2022). Although all these alternative versions follow the same logic as the original, some differ in crucial aspects. For example, Howlett and Tosun (2019b) propose a classification of national policy styles that treats the degree of inclusiveness (high vs. low) of a political system as the first dimension and the key actors involved in the process (bureaucrats/experts vs. politicians/public) as the second. The goal of this conceptual modification is to make the concept of policy styles applicable to non-democratic political systems and to facilitate the observation of variation within both democracies and autocracies as well as to capture democratization or autocratization episodes within the respective political regime types (see Pelke & Croissant, 2021). There is consensus that, at least under most conditions, governments tend to react to policy problems in somewhat distinctive ways through their established general approach to problem-solving (Kagan, 2001). The second dimension concerning the significance of the relationships existing between governmental and nongovernmental actors has also resonated particularly well with research in comparative politics, where interest groups and interest group politics and governance more generally represent central explanatory variables and a common theme (Baumgartner et al., 2009; Richardson, 2000). Empirically, however, it should be noted that some weaknesses exist in these foundational works. The kinds and range of cases examined in most work on the subject, for example, are very limited. The original volume investigated in depth the national policy styles of (West) Germany, France, the Netherlands, Norway, Sweden, and the United Kingdom. However, they only identified ideal-typical policy styles for four of them: (1) the rationalist consensus style in Germany (anticipatory and consensual); (2) the British negotiation style (reactive and consensual); (3) the French concerting style (anticipatory and impositional); and (4) the Dutch negotiation and conflict style (reactive and increasingly consensual; transition model). Subsequent empirical testing of the concept also has not provided much support for these findings, even in the countries originally examined in the 1970s and 1980s. And newer empirical assessments of the concept come to similar conclusions. For example, the study by Candel et al. (2020) argues that Europeanization has affected policy styles in Europe, which aligns with the conclusion already reached in the volume edited by Richardson (1982). Recent empirical work by Richardson (2018a, 2018b) on the British policy style contends that the national policy style has shifted back toward the traditional Westminster model of top-down policy-making. Richardson attributes the change in the British policy style to the change in the ideological composition of the government—a finding also reported by Ertugal (2021) for Turkey. More precisely, it is the entering into office of right-wing populist governments that Richardson (2018a, 2018b), Tosun et al. (2022), and some contributions to the volume edited by Zahariadis et al. (2022) associate with a more impositional and less anticipatory policy style. This work is revealing in the sense that it suggests possible routes and causes through which established standard operating procedures change. It is also compatible with analyses focusing on the significance of institutional variables—i.e., policy styles change when foundational institutions change—but the operationalization of policy styles varies from case study to case study, limiting the development of systematic cumulative knowledge. Most indicators included in the SGI concentrate on identifying and measuring “sustainable” policy outcomes, that is, outcomes that take into account the needs and interests of future generations and help governments to assess how these can be achieved (Schraad-Tischler & Seelkopf, 2016). While the SGI also use statistical information, for the indicators of interest to this explorative analysis, the data are coded using information provided by country experts (for details, see Croissant & Pelke, 2022). At least two country experts are involved in producing a report, in which they assign scores ranging from 1 to 10 to reflect the countries’ achievements in the areas of interest. The dataset comprises 41 European Union (EU) member states and Organization for Economic Cooperation and Development (OECD) countries. Since the launch of the SGI in 2009, the methodology for coding and aggregating the data have changed. Therefore, only the data from 2014 onward are comparable over time, so we use these for the present analysis. The SGI address three pillars, of which the first captures the governments’ policy performance in terms of economic, environmental, and social policies—the three dimensions of the sustainability concept as put forth by the Brundtland Report released in 1987 (for a discussion, see, Tosun & Leininger, 2017; Wong & van der Heijden, 2019). The quality of democracy is the second pillar and comprises the electoral processes, access to information, civil rights and political liberties, and rule of law. This SGI pillar shares several concepts with other democracy indicators such as the one proposed by the Varieties of Democracy project (Coppedge et al., 2015). The third pillar gauges governance capacities in terms of executive capacity and executive accountability (for details on all three pillars, see Croissant & Pelke, 2022). Of these two, the dimension on executive capacity is of primary importance for the purpose of this analysis. Executive capacity includes, inter alia, the measurement for strategic planning and public consultation, which we consider to approximate the two dimensions of the original formulation of national policy styles. Strategic planning maps onto the dimension on the approach to problem-solving, whereas public consultation captures the second dimension on the relationship between the government and other actors in the policy process (see Tosun et al., 2022; Zohlnhöfer & Tosun, 2021). The dimension on executive capacity comprises additional indicators, but these deviate conceptually from the two dimensions of the original formulation of policy styles.2 Modified conceptualizations of national policy styles that give more weight to the participatory aspects (see Howlett & Tosun, 2019b), for example, could be operationalized more accurately for other indicators, but for the original formulation, strategic planning and public consultation are the best possible proxies. In fact, we rely on the SGI data as, to our knowledge, no other database exists that includes proxies for a quantitative analysis of the two-dimensional concept of national policy styles. The SGI data do not only facilitate a country-comparative analysis but also one over time, which makes this dataset a valuable source not only for this analysis but scholarship on national policy styles more broadly. The indicator for strategic planning assesses the influence of strategic planning units and bodies on government decision making. The highest values (9 & 10) are assigned if “strategic planning units and bodies take a long-term view of policy challenges and viable solutions, and they exercise strong influence on government decision making” (Bertelsmann Foundation, 2020, 52). The lowest values (1 & 2) are assigned to a country by the experts, if “in practice, there are no units and bodies taking a long-term view of policy challenges and viable solutions” (Bertelsmann Foundation, 2020, 52). Public consultation captures whether governments consult with societal actors such as trade unions or societal interest groups in a fair and pluralistic manner. The highest values (9 & 10) are assigned if “the government always consults with societal actors in a fair and pluralistic manner” (Bertelsmann Foundation, 2020, 58). If governments rarely consult with societal actors, they are assigned the lowest values (1 & 2). How well do the two SGI indicators differentiate between countries? This is a necessary condition for using them to operationalize national policy styles. The bar graphs in Figures 1 and 2 provide an answer to this question. They present the scores for the individual countries included in the SGI dataset averaged over seven years. Mean Scores for Strategic Planning, 2014–2020 Notes: Own elaboration based on the SGI dataset. Dotted line indicates the median score. Mean Scores for Public Consultation, 2014–2020 Notes: Own elaboration based on the SGI dataset. Dotted line indicates the median score. Figure 1 on strategic planning reveals that Hungary, Iceland, and Romania have the lowest average score (= 3) and Denmark and Finland have the highest average score (= 9). In the case of the first three countries, this means they have strategic planning units and bodies in place that take a long-term view of policy challenges and viable solutions, but only occasionally exert influence on government decision making (Bertelsmann Foundation, 2020, 52). In marked contrast, the government of Finland has been active in strategic planning and also makes use of policy programs that ensure the strategic goals are met. In addition, the government receives advise from the Committee for the Future, which anticipates issues that may materialize in the future (Anckar et al., 2020, 30). In Denmark, the government operates according to strategic policy plans with time horizons of about ten years. The reforms following from these plans are prepared by specialized commissions, such as the infrastructure commission, and their elaboration also involves expert groups, which are recruited from both inside and outside of the government and state ministries (Laursen et al., 2020, 38). Figure 1 shows that differences across the individual countries are observable using this indicator, not only for the countries that are located at the extreme ends of the score range but also those in between. When using the median across all country-years, as included in the analysis (which corresponds to a score of 6), to differentiate between countries with high strategic planning and those where it is low, 24 of them fall into the first category and 17 into the second. Thus, the countries with high strategic planning outnumber those with low strategic planning. What is worth noting is that in the original formulation of the concept, Germany and France were associated with an anticipatory policy style (Richardson et al., 1982). However, when operationalizing the concept using the scores for strategic planning as measured by the SGI, both countries fall into the category of low strategic planning. The Netherlands and the United Kingdom also fall into the opposite category to the one set out in Richardson (1982): the SGI data identify them now as countries with high strategic planning. This observation supports our endeavor for measuring national policy styles in a systematic and comparable fashion across countries and over time. We consider the SGI to offer an appropriate measurement instrument. Further, it should be noted we are not the first to obtain an empirical picture that deviates from the ideal-typical policy styles identified by Richardson et al. (1982). For example, the policy styles identified in the volume edited by Bovens et al. (2002) on the basis of country- and sector-comparative, in-depth analyses also differ from the four hypothesized policy styles. The different characterizations of national policy styles reported in the literature reveal the importance of pushing the corresponding research into a direction that pays more attention to operationalization and data production. Similar to the dimension on strategic planning, we can discern significant country differences for public consultation. The most extreme cases are Hungary, which has an average score of 2 only, and Norway and Switzerland, which have an average score of 10. The case of Switzerland is particularly interesting since the involvement of societal groups there results from two defining institutional features. The first is the country's corporatist system, which secures the main interest groups access to the institutional venues where policies are formulated. The second feature refers to Switzerland's strong components of direct democracy (Papadopoulos & Maggetti, 2019). Societal groups that deem themselves and their interests as not being considered sufficiently in policy formulation can credibly threaten the government with organizing a referendum (Armingeon et al., 2020, 54). Numerous studies have shown that such referendums tend to be oriented around the status quo, which means that new policies are likely to be dismissed by the public (see Sciarini, 2018). Since societal actors have veto power over policy decisions, it is imperative to include them in policy-making. Interestingly, in Norway, it is the consensus orientation that guarantees societal groups to have a say in policy-making, rather than the threat of organizing a referendum. Such groups are involved in the policy formulation process by expressing their views on a policy proposal before it is presented to parliament (Sverdrup et al., 2020). In contrast to strategic planning, the range of the average scores for public consultation goes from 2 to 10 (see Figure 2). However, similar to strategic planning, most countries fall into the category of high public consultation. Returning to the original formulation of policy styles, France and the Netherlands were associated with low public consultation, and Germany and the United Kingdom with high consultation. When inspecting Figure 2, we can see that out of these four countries, only France stands out as still having a low level of public consultation—all the other countries have high levels. In this context, it is worth highlighting that the Netherlands scores highest among these three countries in public consultation, followed by Germany and then the United Kingdom. In all likelihood, Hungary has a low score in public consultation in policy-making because the observation period starts after Viktor Orbán's electoral victory in 2010. The Orbán Administration has taken a series of steps to transform the country into an illiberal democracy (Bauer & Becker, 2020; Pirro & Stanley, 2021), which is reflected in the SGI scores. Figures 1 and 2 present the scores for the countries averaged over the observation period (2014–2020) and prima facie reveal the usefulness of these indicators for observing national policy styles using the two dimensions proposed by Richardson et al. (1982). But they also allow us to begin to answer some of the key questions posed at the outset of this study concerning the concept. Figure 3 (for strategic planning) and Figure 4 (for public consultation) use box plots to present the data. Box plots offer an instructive way of inspecting data since they reflect both differences across countries as well as variation over time. For countries that did not experience any changes over time, the box plots indicate the median only. The longer the box, the more the countries’ scores vary over time. The dots in the graphs indicate outlier values that do not fit with the overall variation pattern of the data. Two observations are worth flagging when inspecting the figures. First, we observe temporal stability in the values for both indicators, as evinced from the lines positioned at the median scores. This, again, prima facie accords with the conceptual understanding of national policy styles, since this expects a high level of stability to be a hallmark of a style. Second, recalling the coding of the SGI indicators, two scores are used to denote the same level. Thus, a one-point change in the scores cannot be regarded as a substantial degree of change. Consequently, in many cases, when country scores did change, these changes corresponded to minor institutional alternations and reinforce the high overall degree or pattern of stability—albeit with a few but important exceptions, which are discussed below. Table 1 offers an even more straightforward presentation of the variation in the data over time. It reports the variance for both indicators as well as characterizes it as “no change” (variance of 0), “minor change” (variance smaller than 1), “moderate change” (variance between 1 and 2), and “major change” (variance greater than 2). The characterization of change is based on the indicator that demonstrates the greater cross-temporal variance. Figure 5 gives an overview of the countries and into which cell of a 2 × 2 matrix they would fall, which comprises strategic planning (high vs. low) and public consultation (high vs. low) as the two dimensions. This presentation is instructive since it shows that the positioning on each dimension is correlated: countries with high scores on one dimension also tend to have high scores on the other. Exceptions from this rule are Iceland, Germany, Luxembourg, and Switzerland. These countries are characterized by a broad involvement of societal actors but a low degree of strategic planning. The opposite constellation holds for Ireland, Mexico, South Korea, and Spain: these countries score low on the dimension on public consultation but high on strategic planning. These findings highlight the utility of indicators such as those contained in the SGI for identifying relatively long-term policy styles, classifying them, and highlighting their general evolution. However, the data are also useful for explaining seeming anomalies, that is, cases which either do not conform to the general picture or which proved less stable over the period examined. As these data show, the United Kingdom is the only country that experienced moderate change, as the strategic planning dimension declined after 2017 (Busch et al., 2020, 37). A reduction in the “rationalist” approach to policy-making in the United Kingdom after the Brexit referendum is also witnessed by Richardson (2018a, 2018b). He associates this with the marginalization of interest groups but also states that in those areas that are important for designing the post-Brexit era interest groups with pertinent expertise will become increasingly influential. Poland, Malta, South Korea, and the United States, however, all stand out as the countries that experienced a major change in their scores during the observation period. The United States is the country that experienced the largest degree of change on both indicators. For Poland, the variance is large for the dimension on public consultation

Welfare state support during the COVID‐19 pandemic: Change and continuity in public attitudes towards social policies in Germany

Abstract: Abstract Our analysis asks whether the pandemic situation affects welfare state support in Germany. The pandemic has increased the health and income risks calling for welfare state intervention. While increased needs, more deservingness, and higher state responsibility during such a crisis would suggest augmented support generally and among those at risk, this might be a short‐term effect and cost considerations could reverse this trend. We study public attitudes towards four key social policy areas based on the German Internet Panel (GIP). We use three waves prior and further three waves since the pandemic had been declared in March 2020. The analysis shows both continuity in the popularity of social policies, in particular health and pensions, and some short‐term increase in support for unemployment and family policies. The results after nearly 2 years suggest rather continuation with some thermostatic short‐term boosts in support instead of any long‐lasting change.

Measuring Policy Performance, Democracy, and Governance Capacities: A conceptual and methodological assessment of the Sustainable Governance Indicators (SGI)

Abstract: This paper provides a critical assessment of the Sustainable Governance Indicators (SGI) and compares it with other prominent indices that address specific components of governance: V-Dem, WGI, and BTI. We offer a comparative assessment of content validity of these governance measures, their data generation processes, and their convergent validity. We conclude that the SGI’s most important contribution is the conceptualization of policy performance as a discrete index. Other relative strengths are the theoretical embeddedness and the exclusion of irrelevant meanings of governance, and the conceptualization of three governance components (Governance, Policy Performance, and Democracy). However, in terms of geographic and temporal coverage, the SGI is clearly inferior to WGI and V-Dem. The handling of third-party statistical data, the absence of uncertainty scores, and the (a-theoretical) aggregation of different indicators are additional shortcomings of the SGI. Finally, the SGI’s iterative process of expert deliberation has merits but is prone to biases. 本文批判评估了可持续治理指标（SGI）并将其与应对具体的治理组成部分的其他重要指数相比较，它们分别是：民主多样性（V-Dem）、全球治理指标（WGI）以及贝塔斯曼转型指数（BTI）。我们比较评估了这些治理措施的内容正当性、数据产生过程以及其收敛效度。我们的结论认为，SGI最重要的贡献是将政策表现概念化为一个离散指数。其他相对优势则是理论嵌入和排除不相关的治理含义，以及对三个治理组成部分（治理、政策表现和民主）的概念化。不过，就地理覆盖范围和时间范围而言，SGI明显次于WGI 和V-Dem。SGI的缺陷还包括对第三方统计数据的处理、缺乏不确定性分数、以及对不同指标进行（非理论性的）聚集。最后，SGI在专家商讨方面的迭代过程具有意义但却容易出现偏见。 Este documento proporciona una evaluación crítica de los Indicadores de Gobernanza Sostenible (SGI) y los compara con otros índices destacados que abordan componentes específicos de la gobernanza: V-Dem, WGI y BTI. Ofrecemos una evaluación comparativa de la validez de contenido de estas medidas de gobernanza, sus procesos de generación de datos y su validez convergente. Concluimos que la contribución más importante del SGI es la conceptualización del desempeño de las políticas como un índice discreto. Otras fortalezas relativas son el arraigo teórico y la exclusión de significados irrelevantes de gobernabilidad, y la conceptualización de tres componentes de gobernabilidad (Gobernanza, Desempeño de Políticas y Democracia). Sin embargo, en términos de cobertura geográfica y temporal, el SGI es claramente inferior a WGI y V-Dem. El manejo de datos estadísticos de terceros, la ausencia de puntajes de incertidumbre y la agregación (a-teórica) de diferentes indicadores son deficiencias adicionales del SGI. Finalmente, el proceso iterativo de deliberación de expertos del SGI tiene méritos pero es propenso a sesgos. The last three decades have seen a boom in the development of social science indicators and indices. The supply of new indicators is as much a result of improved data availability, measurement procedures, and estimation techniques as it reflects the growing demand for such information on the side of policy-makers and researchers (Anheier, 2018; Rotberg, 2015). The menu of methodologically advanced indicators available now spans almost the entire social sciences, and the coverage of the information made available encompasses most societies in the world and long historical periods. Two subfields in which the development of indicators has made particularly great strides are governance and democracy, which are both abstract, contested, and multidimensional concepts that encompass closely related phenomena. In fact, some governance indicator projects conceptualize democracy as a core dimension of (good) governance. For example, composite indices like the World Bank's Worldwide Governance Indicators (WGI), the Bertelsmann Transformation Index (BTI), and the Sustainable Governance Indicators (SGI) provide numerical measures on different dimensions of governance that capture the quality of democracy, the capacity of governments to provide public goods and services to their citizens, and their policy performance. Other indices provide sector-specific governance measures (Haber & Kononykhina, 2018). In contrast, prominent democracy measurements such as the ones published by Freedom House and the Varieties of Democracy (V-Dem) project differentiate between more comprehensive conceptualizations of governance and democracy in the narrow sense. Therefore, some reviews of the state of art in governance indicators treat the two strands of measurements synonymous, arguing that “narrow” democracy indicators constitute a subset of a “new generation of governance indicators” (Anheier, 2018). Many attempts have been made to provide comprehensive reviews of the different attempts to quantify governance and its individual components. Already in 2006, the authors of an OECD report estimated “that there are now some 140 user-accessible sets of governance indicators, comprising literally thousands of individual indicators” (Oman & Arndt, 2006, p. 30). In 2015, Rotberg and Bhushan introduced their Governance Indexes in Comparative Perspective Database, which contains summaries and classifications for 94 governance indices (Rotberg & Bhushan, 2015). In 2018, Haber and Kononykhina (2018) critically assessed a selection of 37 governance indices. In addition, there are numerous attempts to evaluate governance indices, particularly focusing on issues related to theory and measurement (Arndt, 2009; Andrews et al., 2010; Davis et al., 2012; Rothstein & Teorell, 2012; Gisselquist, 2015) as well as evaluations of the various democracy indices in particular (Lauth et al., 2000; Munck & Verkuilen, 2002; Lauth, 2004; Pickel & Pickel, 2006; Coppedge et al., 2017; Coppedge et al., 2020; Schlenkrich, 2021). On top of that, there is extensive data quality assessment literature (for an overview, see McMann et al., 2021). The existing literature reviews and indicator evaluations highlight three common insights. First, governance is a multidimensional phenomenon and there is no convergence regarding its conceptual understanding. Second, some problems of conceptualization and measurement of an abstract and contested concept such as governance may never be solved definitely. Third, the most widely used governance measurement approach has been composite indices. The goal of this paper is to not create another inventory of governance indices but rather to critically assess the SGI and compare it with a selection of prominent indices that address governance more broadly or address specific components of what is generally considered to be governance. The SGI project, first published in 2009, is curated and funded by the German Bertelsmann Foundation. It aims to provide systematic measurements and indicators of the reform capacity in 41 OECD and EU countries. Its approach combines three different “pillars” of sustainable governance and includes Policy Performance, Democracy, and Governance as distinct dimensions of sustainability. By doing so, the SGI project offers a comprehensive approach of measuring sustainable governance, while it also competes with social science data collection efforts measuring either democracy, such as V-Dem, policy outcomes, such as the BTI, or governance quality, such as the WGI. Our contribution is to systematically apply a step-by-step guide for assessing a measure's quality recently introduced by Kelly McMann and her co-authors (McMann et al., 2021) to the SGI. We use McMann et al.’s approach to systematically assess the SGI because it is a comprehensive data quality assessment approach that integrates three components: (1) content validity; (2) the validity and reliability of the data generation process; and (3) convergent validity (McMann et al., 2021, p. 1).1 This comprehensive assessment of the SGI advances our understanding of how (not) to measure (sustainable) governance and thus also provides a user-friendly introduction to the SGI and related approaches and indices. In the following section, we discuss the extent to which the SGI pillars capture the higher-level theoretical concepts the project intends to measure (content validity assessment). Next, we examine the validity and reliability of the data generation process of the SGI. A unique characteristic of the SGI is the fact that it combines expert-coded items and statistical “objective” data to operationalize and measure the reform capacity of OECD and EU countries.2 We discuss the dataset management structure, data source, and respondent coding procedures as well as the aggregation models the SGI executes and present the temporal and spatial coverage. In the subsequent section, we assess the convergent validity by comparing the SGI measures to alternative comparable measures, to discuss the theoretical and conceptual strengths and limitations of the SGI. Finally, we summarize the SGI’s advantages and disadvantages relative to other existing governance indicators as well as its value for comparative policy analysis.3 The SGI seeks to provide data that allows researchers and policy-makers to understand how OECD and EU countries can “achieve sustainable policy outcomes and imbue political decision making with a longer-term focus” (Schraad-Tischler & Seelkopf, 2016, p. 2). Thus, the SGI mainly focuses on the capacity for policy reforms (Brusis, 2010; Jäckle & Bauschke, 2009) but also includes the democratic quality of political institutions and processes as well as the governance capacities as prerequisites for sustainable governance. Its methodology changed in 2014 and thereafter. The SGI also competes with data curation and collections projects that measure the quality of governance, democracy, or both such as the World Bank's WGI (Kaufmann et al., 2007), the V-Dem project (Coppedge et al., 2021), or the Bertelsmann Transformation Index (BTI) (Brusis et al., 2012; see Table A1). The SGI defines “sustainable governance” as “the political management of public affairs that adopts a long-term view of societal development, takes into account the interests of future generations, and retains capacities for social change” (Brusis & Siegmund, 2011, p. 3). Building on this definition, the SGI differentiates between three “pillars” of sustainable governance (Schraad-Tischler & Seelkopf, 2016, p. 3). According to Munck and Verkuilen's vertical ordering of the different components of indices according to their degree of abstraction (“concept tree”; Munck & Verkuilen, 2002), these “pillars” are located at the concept level. This first pillar examines the Policy Performance in terms of three different policies which are located at the attribute level: economic policies, social policies, and environmental policies. Each attribute has different components (SGI refers to them as categories) as shown in Figure 1. Second, Quality of Democracy is conceptualized by four attributes: electoral processes, access to information, civil rights and political liberties, and rule of law. The third pillar examines governance capacities in terms of executive capacity and accountability of a country. In contrast to the other two pillars, the governance concept uses two sub-concepts, namely executive capacity and executive accountability. The sub-concept executive capacity is operationalized by steering capacity, policy implementation, and institutional learning as attributes that each have a number of components. The sub-concept executive accountability has four attributes: citizens, legislature, intermediary organizations, and independent supervisory bodies. The SGI conceptualizes all three main dimensions (or “pillars”) with a special focus on longer-term sustainable developments. The theoretical embeddedness of the SGI concept with three pillars of sustainable governance deserves credit but lacks theoretical consistency and more detailed discussion. The SGI sees governance capacity as well as democratic quality as prerequisites for sustainable governance. However, it is at least questionable whether the quality of democratic governance or rather the stability of democratic decision making is a key factor for sustainability. In addition, the discrimination between executive accountability as a sub-concept of governance instead of an element of (liberal) democracy lacks a theoretical justification. Another conceptual issue is that the SGI does not differentiate between dimensions of accountability, for example, vertical, horizontal, and diagonal (Lührmann et al., 2020), but mixes these dimensions of accountability into different components, which results in conceptual inconsistency. As shown in Figure 1, the SGI uses 71 qualitative expert-coded indicators and 85 quantitative statistical indicators. Each component is measured by at least one expert-based indicator and additional statistical indicators for the Policy Performance index and executive accountability sub-concept. The Democratic Quality index, as well as the executive capacity sub-concept, relies exclusively on expert-coded indicators, which is not surprising due to the nature of both concepts. Similar to the BTI, the SGI also offers narrative descriptions of expert-rated indicators, so-called country reports of several thousand words, which serve a dual purpose. First, they provide a source of detailed case-based information for users who also seek more detailed descriptions of relevant institutions, processes, and actors in a particular county. Second, the reports make numeral ratings more transparent, and the process of iterative text revisions (see below) through peer review improves the validity of the numeral ratings and increases the inter-coder reliability. The Policy Performance index of the SGI is the most innovative contribution of the SGI in the empirical measurement of sustainable governance. There is no shortage of policy indicators, provided by OECD, World Bank, and other national or international organizations. However, to our knowledge, the SGI is the first systematic and cross-national attempt to conceptualize and measure the performance of public policies in different policy areas in the form of a composite index as one aspect of sustainable reform capacity of OECD member countries. Other existing governance indicators that aim at measuring performance are either not policy-specific or limited to sector-specific policies. Therefore, in this section, we assess the content validity of the Policy Performance index without comparing it to other measurement approaches. The theoretical construct of interest is public policy outcomes in different policy areas, which are highly contested aspects of sustainable governance. Neither political actors, “who are prone to conduct and use evaluations strategically” (Stephenson et al., 2019, p. 664), nor researchers have come to a common understanding of evaluating public policy outcomes. However, the SGI approach in measuring policy performance in a comprehensive and cross-national way is a systematic attempt to do so. Thus, we cannot refer to a standard definition of public policy performance when evaluating the content validity of the respective index. The SGI Policy Performance index captures the performance of states in terms of economic policies, social policies, and environmental policies as distinct attributes of public policy. Each of the three attributes refers to a particular policy dimension and includes two to eight components (see Figure 1). These components refer to specific policy fields and are measured by different statistical and expert-coded indicators. For example, the economy component captures one expert-coded indicator, asking “How successful has economic policy been in providing a reliable economic framework and in fostering international competitiveness?” (SGI, 2020, p. 20), and six statistical indicators including GDP per capita, GDP growth, and inflation. In addition, the SGI uses components in regard to global public goods, more precisely “Global Financial System,” “Global Social Inequalities,” and “Global Environmental Protection.” By doing so, the SGI uses domestic policy-making components and components that measure “the extent to which governments actively contribute to the provision of global public goods” (SGI, 2020, p. 8), and thereby mixes two dimensions of policy-making that are mutually interdependent. Furthermore, within the P9 component “Health,” the SGI combines health policy as a policy output measure with an infant mortality measure, which is an outcome variable. Moreover, infant mortality is closely and causally interrelated with economic prosperity and also democratic quality (Gerring et al., 2021). The combination of policy outputs and outcomes in the policy performance concept is—from your view—undertheorized, but seems reasonable in light of the goals of the SGI. Researchers typically construct measures of governance through aggregation of some factual data, or aggregation of the responses of a set of expert respondents to survey questionnaires, or some combination thereof (Hollyer, 2018). The SGI relies on such a combined approach. In the following, we use Bayesian factor analysis to assess the content validity of the Policy Performance index.3 The findings provide moderate empirical support regarding the construct validity of this first SGI pillar. As shown in Table 1, all six components of the economic policies attribute load on a single dimension. However, the fit for taxes, budgets, and global financial system is somewhat weaker, while the uniqueness of different components is relatively high. Table A5 (Supporting Information Appendix) indicates that a two-dimensional factor solution decreases the uniqueness score clearly, but does not indicate a clear pattern of two distinct dimensions of economic policies. The social policies attribute also captures a wide range of components from education to global inequalities. All components strongly load on a single dimension, although the fit for integration is substantially weaker. Table A5 shows that the two-factor solution only moderately decreases the uniqueness score indicating that the one-dimensional factor solution is adequately operationalized. The integration component measures how effectively policies support the integration of migrants into society, and thus is the only component in the attribute that distinguishes between migrants and natives. Finally, the two components concerning environmental policies also show strong factor loadings and thus provide support for constructing one attribute dimension. In addition, we also tested for some of the components how the indicators at the lowest level load on the respective components. Tables A10–A12 clearly indicate that the indicators load slightly different on the components. In sum, the theoretical constructs seem reasonable, but the content validity at the lowest level of the SGI is in need of improvement. The results in Table 2 show that all three attributes load onto a single index and reflect one underlying concept, though the fit for environmental policies is substantially weaker and its uniqueness score indicates that a large portion is unexplained. However, Table A6 indicates that the two-factor solution in fact decreases the high uniqueness score of the environmental policies attributes, but does not lead to a clear pattern regarding two different dimensions. Thus, the aggregation into the policy performance pillar seems to be supported by the data. With the global diffusion of democracy in the late twentieth century, there has been a profusion of measures of democracy and democratic quality. Although there is no consensus about underlying models of democracy, concepts, variables, yardsticks, and methods (Geissel et al., 2016), the scholarly debate has increasingly moved away from questions about what democracy is or is not to questions of reliability and validity of democracy measures (Munck & Verkuilen, 2002). Among the more than a dozen different measurements that claim to evaluate the quality of democracy, the V-Dem project is the most ambitious and methodologically advanced one (cf. Coppedge et al., 2020). It has become the de facto gold standard in empirical democracy research. Therefore, V-Dem is our benchmark to discuss and evaluate the conceptualization of the democracy dimension in the SGI data collection effort. The SGI deviates in important regards from the V-Dem approach to conceptualize and measure democracy. First, it is accepted among students of democracy that there is not a single universal definition of democracy, but a variety of different models, each building on specific core principles (Held, 2006). For example, the V-Dem project distinguishes between five high-level principles of democracy (electoral, liberal, participatory, deliberative, and egalitarian), operationalizes these principles, and collects data to measure those (Coppedge et al., 2020). Of these principles of democracy, the electoral democracy and the liberal democracy are most widely used in empirical democracy research. The electoral democracy concept is captured by Robert Dahl's concept of Polyarchy and represents a minimum standard for democracy. It focuses on elections as the core mechanism for making rulers responsive and accountable to citizens and their demands (see Coppedge et al., 2020, p. 30; Dahl, 1971). The liberal democracy concept “stresses the value of protecting individual rights against state oppression and unrestricted majoritarian rule” (Coppedge et al., 2020, p. 33). Somewhat surprisingly, the SGI does not specify the principle of democracy that is its benchmark and, hence, its Democratic Quality index is inadequately located in these different principles of democracy (see also Pickel et al., 2015, p. 509). Furthermore, the SGI’s Democratic Quality index is conceptually extensive at the attribute level but thin at the component level and it does not discriminate between necessary, non-necessary, and sufficient conditions. At the attribute level, it includes the rule of law, resulting in a demanding liberal concept of democracy. However, compared to V-Dem's Liberal Democracy Index, which emphasizes the importance of protecting individual and minority rights against the state and the majority, SGI’s Democratic Quality index does not include essential indicators, such as effective checks and balances, which is one of the core features and often conceptualized as a necessary condition of the liberal principle. While the SGI Governance index includes a number of indicators for executive accountability, including indicators of legislative oversight, these indicators are not conceptualized as measures for the quality of democracy. Another conceptual shortcoming is that the SGI conflates some features of the liberal principle, for example, civil liberties, judicial independence, and legislative independence, in the same questions. For example, the question on civil rights asks, “To what extent does the state respect and protect civil rights and how effectively are citizens protected by courts against infringements of their rights?” Thus, at the indicator level, the SGI conceptualization manifests shortcomings regarding conflation and coherence. Similarly, there are some conceptual shortcomings regarding the measurement of the electoral principle of democracy in the SGI. The SGI covers the electoral principle of democracy—a necessary condition for liberal democracy—by two attributes: electoral process and access to information. The former includes five expert-coded indicators that are relevant but miss important additional indicators for the measurement of electoral democracy, such as whether the chief executive is (in-)directly elected, and whether elections affect the composition of the chief executives. In addition, the SGI’s electoral process component does not adequately capture whether elections are free, fair, and regular, as it only asks whether “citizens have the opportunity to take binding political decisions when they want to do so” and “to what extent do all citizens have the opportunity to exercise their right of participation in national elections” (voting and registration rights) (SGI, 2020, p. 39). Although the electoral processes component includes the fairness of candidate and party registration, thereby capturing one important dimension of clean and fair elections, there is no attribute which would explicitly capture the autonomy and capacity of the election management body, and the absence of vote and turnout buying attempts, which are frequent defects in young democracies (e.g., Gonzalez-Ocantos et al., 2012). Overall, while the concept logic is not theorized coherently, the SGI’s Democratic Quality index lacks also from conflation problems concerning different questions, as these often multi-dimension and inherently difficult to rate. The Bayesian factor analysis shows that most of the indicators are strongly loaded onto their respective attribute (Table 3). However, the loadings of indicators related to three of the four attributes—party financing, popular decision making, access to government information, and appointment of justices—only fit weakly to their respective attributes, which confirms our concerns regarding the conceptualization of some democracy attributes in the SGI. The two-factor solution presented in Table A7 (Supporting Information Appendix) also demonstrates that at least two attributes—rule of law and electoral processes—are better covered by a two-factor solution. Finally, the factor loadings of the four democracy attributes reported in Table 4 suggest that they are in fact appropriate for the underlying single attribute of democratic quality.4 In sum, it is fair to conclude that the construction of the SGI Democratic Quality index is problematic for two reasons. First, it misses some important electoral principles of democracy and conceptually conflates the liberal principle with the electoral principle of democracy.5 Second, the construction of the attributes that are supposed to reflect the underlying attributes is erroneous. Similar to democracy, the concept of governance is contested. At the core of the debate is the question of whether governance relates to quality of government exercise of power, the performance of administrative institutions and government's management of public affairs, or the quality of public services and goods that governments provide to them (Haber & Kononykhina, 2018). All of the three meanings have in common “the act of steering and regulating social behavior” (Fukuyama, 2016, p. 90). The SGI Governance dimension is defined as “political leadership's capacity to steer processes with success” (Schraad-Tischler & Seelkopf, 2016, p. 10). Thus, it corresponds to the second meaning of governance. The SGI conceptualization then differentiates between two dimensions: executive capacity and executive accountability. The former “emphasizes a government's capacity to deliver sustainable policies” (ibid). The latter takes into account whether actors and institutions make use of their participatory and oversight competencies. Thus, the SGI focuses on “the forms of interaction between a government and other stakeholders in the policy-making process” (ibid). The SGI Governance Index relates to the WGI, which define governance as “the traditions and institutions by which authority in a country is exercised” (Kaufmann et al., 2010, p. 4). However, compared to the SGI Governance index, which is theoretically embedded in the conceptual literature, the WGI has been criticized due to its a-theoretical character of its governance concept (Thomas, 2010) and the absence of “an explicit theory of what constitutes good governance” (Fukuyama, 2016, p. 97). While the SGI Governance index focuses on executive capacity and accountability, the WGI comprises six governance indices that cover the “process by which governments are selected, monitored and replaced, […] the capacity of the government to effectively formulate and implement sound policies, and […] the respect of citizens and the state for the institutions that govern economic and social interactions among them” (Kaufmann et al., 2010, p. 4). The WGI compiles and aggregates a wide range of third-party measures, for example, democracy indicators from Freedom House and V-Dem, government quality from BTI, or survey data from Afrobarometer and Latinobarometer, but does not theorize how the six baskets of indicators were chosen and whether and how they relate to each other (e.g., Fukuyama, 2016). In addition, the WGI assumes that voice and accountability, “which can otherwise be understand as media freedom and democracy – are considered intrinsic aspects of good governance” (Fukuyama, 2016, p. 98). The a-theoretical combination of democracy and media freedom as aspects of governance makes the empirical test how democracy relates to governance challenging. In sum, the content validity of the WGI is problematic, as it does not map how the measures capture the higher-level theoretical concept, and, additionally, it is largely unspecified what the theoretical concept of interest is. Of the six WGI indices, two (voice and accountability; political stability and the absence of violence) focus on the electoral principle of democracy, whereas rule of law and control of corruption indices approach the liberal principle of democracy. Thus, the appropriate comparison for the content validity assessment of the SGI’s Governance index is the WGI’s remaining two government indices on the capacity of the government to effectively formulate and implement policies that is government effectiveness and regulatory quality. The Bertelsmann Transformation Index can be considered as another reference point (see Møller & Skaaning, 2021; Müller & Pickel, 2007 for assessments). The BTI “analyzes and evaluates whether and how developing countries and countries in transition are steering social change toward democracy and market economy” (Bertelsmann Stiftung, 2020, p. 116). It differentiates between political transformation, economic transformation, and governance. The former two build the so-

Pandemic responses at the subnational level: Exploring politics, administration, and politicization in Swedish municipalities

Abstract: Abstract The Swedish response to the pandemic at the national level has attracted considerable international attention, but little focus has been placed on the way municipalities dealt with the crisis. Using Hay's dimensions of politicization, namely the capacity for human agency, deliberation in the public domain, and social context, we analyze the politicization of the municipal response to the pandemic in Sweden. We do this based on the analysis of the decision making process to activate (or not) an extraordinary crisis management committee. We find inter alia, that (i) only a quarter of the municipalities activated the committee while a majority of them had an alternate special organization in place; (ii) support to the existing organizational structure was more salient than creating an extraordinary organization, and (iii) a robust municipal structure was deemed to be one able to withstand shocks without resorting to extraordinary governance arrangements. We find a ‘conditioned politicization’ of the response, privileging administration over politics.