European elections are just around the corner. Like regional elections, they have been described as second-order elections, strongly influenced by and subordinate to national politics. With multi-level governance becoming increasingly commonplace in Europe and the world, does the assumption of second-orderness still hold true? Or are we witnessing the emergence of increasingly independent political arenas?
Are European and regional elections really subordinate to national politics?
Low turnout and little interest of voters in European Parliament (EP) elections have led political scientists to describe them as second-order elections. This influential model on the subordination of EP elections compared to first-order national politics, developed by Reif and Schmitt in 1980, has since been widely employed and expanded to other seemingly subordinate political arenas, such as the regional and local level. Put simply, it argues that voters do not abstract between different levels of government and their respective competencies, but merely use European (or other second-order) elections to signal approval or disapproval of national politics. Because less is at stake, more voters decide to stay home, particularly those that are generally content with national politics. Besides just turnout rates, do voters really view other political arenas as inherently inferior to national politics?
Today, more and more scholars take issue with this rather simplistic conception of multi-level governance (MLG).  Two interconnected developments motivate the assumption that the concept of second-order elections is not really up-to-date.
Multi-level governance has become commonplace
First, several layers of government – local, regional, national – have become commonplace in many countries throughout the world since the concept was first developed. For example, the United Kingdom initiated devolution – the conferral of powers from Westminster to parliaments and assemblies in Scotland, Wales and Northern Ireland – in the late 1990s. At least partly to appease separatists in Catalonia and the Basque Country, Spain has similarly devolved more powers to its regions. Many Eastern European countries, as part of a process coined Europeanization, built up administrative structures at local and regional level upon joining the European Union (EU). Today, albeit to different degrees, such systems of MLG in EU Member States are widespread and advanced. Unique in the world is the directly elected supranational level – the European Parliament – that complements MLG in the EU.
Second, particularly regarding the work of the EP, the last decades brought along an ever-increasing amount of competences and policy areas covered by the Parliament as well as the politicisation of European politics. Nowadays, the EP is co-legislator in many domains that directly affect EU citizens and the times of the permissive consensus are long gone: EU citizens care a lot more about European politics than they used to – whether that is affirmative interest or outright rejection of the EU.
What about regional elections?
So how does the concept of second-order elections fare in regards to regional elections? Where do they stand in comparison to both national and European elections? Do voters abstract these different levels of government or do they use all elections to judge the performance of their national governments?
Let’s take a look at some exemplary countries: When analysing voter behaviour in Spain, a country that has experienced far-reaching decentralisation over a short period of time, results quite clearly indicate that a majority of voters tends to base their electoral decisions on regional issues (about 50-60 %). This number is even higher in Spain’s most autonomous regions with an elaborate regional identity (f.ex. about 70% in Catalonia) (Cabeza, 2018). While European elections in Spain remain largely nationally driven (only 10-20 % make their decision thinking on European issues), regional elections are quite clearly first-order: The electorate makes its decision not to evaluate national politics but to advance regional issues they think of strongly. 
In Germany, elections in the Länder are still frequently classified as second-order elections, but the extent of this second-orderness depends strongly on the timing of the election and its potential consequences – on what is at stake. Mid-term regional elections are most strongly influenced by federal politics. If regional elections can directly impact the national government or cause major shifts in national parties – for example when a national party leader would make his future dependant on the outcome of a regional election – the degree of second-orderness is reinforced. One reason often noted for the inferiority of Länder elections is the uniformity of political parties in Germany. Unlike in Spain, parties with a strong regional identification, the CSU being the only exception, are weak or do not exist and as such, regional and federal party competition is strongly intertwined. However, the German electorate tends to be more courageous at the Länder level, approving of party coalitions that seem unlikely or undesired at the federal level. Länder are often experimenting grounds for federal politics: The first coalition between Social Democrats and the Greens was formed in Hesse in 1980 and became the preferred option of Social Democrats at the federal level thereafter. Likewise, at Länder level, both the Greens and the Left currently lead a government (in Baden-Wuerttemberg and Thuringia respectively). This allows for the assumption that the electorate does view the political arena as distinct, at least in part, and that regional electoral decisions do differ from those in national elections.
If we broaden the analysis, a more nuanced perspective on elections below the national level emerges: In a wide-ranging study based on a dataset that includes the results of 2933 regional elections in Europe, Australia, Canada, and Japan, Schakel and Jeffrey identified strong second-order effects for only 18% of these elections!  Assuming that all regional elections are somewhat second-order is hence an extremely flawed starting point for analysis. The truth is certainly more complex, and although there will always be interdependencies between different levels of government, the regional level seems to be increasingly independent. Although less pronounced, this development can also be observed for the European elections. Some conditions of the European political arena still favour their second-order character though, most notably the lack of clearly identifiable European parties or the non-existence of a usual government-opposition confrontation.
Not all sub- or supranational elections are second-order
As competencies are devolved from the nation-state level to sub-national and supranational bodies, voters increasingly accustom with this new form of multi-level governance and make qualified voting decisions dependant on the level of government. Although the national arena remains the most important in many cases, elections at different levels are no longer categorically inferior to national politics, a development that is likely to continue in the EU and beyond.
 Karlheinz Reif & Hermann Schmitt, “Nine Second-Order National Elections – A Conceptual Framework fort he Analysis of European Elections Results”, European Journal of Political Research, 1980.
 See for example Kopeke and Ringe, 2006, Brug et. al., 2016 & Nielsen and Franklin, 2017.
 Laura Cabeza, “First order thinking in second order contests: A comparison of local, regional and European elections in Spain”, Electoral Studies, 2018.
 Jochen Müller, “German regional elections: Patterns of second-order voting”, Regional & Federal Studies, 2018.
 Arjan Schakel & Charlie Jeffrey, “Are Regional Elections really “Second-Order” Elections?”, Regional Studies, 2013.
Julie Hassing Nielsen & Mark N. Franklin, The Eurosceptic 2014 European Parliament Elections, Palgrave Macmillan, 2017.
Hermann Schmitt, “The European Parliament Elections of June 2004: Still Second-Order?”, West European Politics, 2005.