So far this contribution has summarised the evolution of popular support for the euro. Next, we try to identify the major economic determinants of support for the single currency. Here we adopt the approach embedded in the literature on popularity functions (Nannestad & Paldam, 1994). The popularity function is a key concept in public choice or political economy. The basic idea behind the popularity function is that the public/the voters hold the government in office responsible for the performance of the macro-economy and/or economic events in general. Usually the popularity of a government as revealed by public opinion polls is postulated as a function of a set of macroeconomic variables like inflation, unemployment, and growth of GDP per capita. The intuition is that inflation and unemployment are negatively related to the popularity of the government while GDP growth is positively related. In short, the government is rewarded/punished for the behaviour of the macro-economy.
There are obvious similarities between the popularity of a government and the popularity of a currency like the euro. The popularity of a government is influenced by its ability to manage the economy at large, in short by its economic governance. Likewise, we expect the governance of the euro – that is, the monetary governance of the ECB – to be a function of the outcome of monetary policy, first of all as it is reflected in the rate of inflation in the euro area as measured by official statistics. Much of the research on public support behind the euro and European political integration is focused on economic determinants – although political and historical factors also have an impact.Footnote 5
The empirical tests of popularity functions generally confirm the significance of macroeconomic variables, in particular the rate of unemployment and inflation. However, the lack of stability of econometric results is a problem in this field of work (see e.g. Kirchgässner, 2009, who shows for the case of Germany that the popularity function has disappeared in recent times). Endogeneity is also a serious concern in the econometric analysis.
When estimating popularity functions for the euro, we contribute to the empirical literature in two respects. First, we explicitly take feedback effects between support for the euro and the macroeconomic environment into account. Tackling this issue (endogeneity, in technical terms), we are able to produce unbiased estimates. To this end, we adopt the Dynamic Ordinary Least Squares approach (Saikkonen, 1991; Stock & Watson, 1993; Wooldridge, 2009) in the aggregate analysis utilising a panel dataset with 228 aggregated observations from 2002 to 2011 (EB57–75). The results obtained are complemented at the individual level.
Second, in addition to the standard macroeconomic variables – namely, inflation, unemployment and growth – we explore the role played by country-level and personal perceptions and attitudes exploiting information from a cross-sectional database with a maximum of 157,899 individual observations from 2002 to 2011 (EB 57–74).
Table 9.2 captures the support for the euro over time, controlling for country characteristics that are time-invariant, for swings in the error term and for feedback between the macro-determinants of support and support itself. By applying panel time series, we can be sure not to have omitted important variables.Footnote 6 The finding that the error terms are stationary (without systematic influence on support for the euro) implies that our coefficient estimates are unbiased and remain unbiased if further controls are added to the regression.
Table 9.2 Relationship between inflation, unemployment, GDP per capita growth and net support for the euro. Panel analysis (aggregated level), 2002–11 Table 9.2 shows the results for the determinants of support for the euro in a panel analysis covering 12 countries i (I = 12) and at different points of time t for the period from spring 2002 to spring 2011. We view this time period as the theoretically most appropriate one as from January 2002 onwards European citizens could actually use the euro as a real money in daily business. As we have argued above, a distinction of four different time periods seems to be appropriate, our analysis will include the third and fourth time period. Similar to other studies in the field (Gärtner, 1997, Kaltenthaler & Anderson, 2001; Banducci et al., 2009, Kaelberer, 2007 and Hobolt & Leblond, 2009), we expect that the price stability (inflation) should be a key influence on support for the euro. However, based on the popularity function literature we would also expect unemployment and growth of GDP per capita to exert a significant influence (see here e.g. Kaltenthaler & Anderson, 2001; Banducci et al., 2009 and Hobolt & Leblond, 2009).
Equation 1 in Table 9.2 includes inflation, unemployment and growth of GDP per capita as the explanatory variables and analyses our full sample. Inflation is negatively and significantly related to support for the euro at the 95% level. Growth of GDP per capita is not significantly related to support for the euro when estimating the observations in our full sample. Unemployment is negatively related with net support for the euro at the 90% level. As we have argued, the pre-crisis period (2002–08) should be kept distinct from the crisis period (2008–11). Eqs. 2 and 3 split the full sample into a pre-crisis period and a crisis period to explore the impact of the financial crisis on popular support for the euro.Footnote 7 Splitting the full sample into a pre-crisis period and a crisis period reveals that the significant negative effect of inflation on support for the euro is driven by the pre-crisis period in which inflation exhibits a strongly negative and significant (99% level) effect on support for the euro. As expected from the literature on popularity functions, GDP growth is strongly and positively associated with support for the euro in the pre-crisis sample. In the crisis sample, column 3, we detect no impact of the three macroeconomic variables on support for the euro. We thus conclude that inflation is the most robust driver of support for the euro.
So far we have examined the macroeconomic determinants of popular support for the euro over time (within effects). In Table 9.3 we move to a micro-level approach to analyse the impact of inflation – our main determinant of support for the euro – based on a cross-section analysis of respondents in all EA-12 countries. Controlling for the socio-economic variables of age, gender, education, left/right placement and the macroeconomic variables of unemployment and GDP growth per capita as well as for country and time fixed-effects (including 157,899 individual observations), inflation has the expected effect in all three samples. Respondents in a country with higher inflation tend to support the euro less than respondents who live in a country with more moderate inflation. Banducci et al. (2003) and Hobolt and Leblond (2009) argue that support is determined by both socio-tropic and egocentric motives. Citizens are on the one hand concerned about the situation in their country, while at the same time they also care about their personal situation. Moreover, Banducci et al. (2009) posit that the actual economic reality – as summarised in official economic statistics – does not necessarily agree with the perceived economic situation. Therefore, to account for citizens’ perceptions towards inflation, columns 2 and 3 include citizens’ perceptions towards their country’s situation (socio-tropic view) and their personal situation (ego-centric view).Footnote 8 Columns 2 and 3 analyse whether perceptions of inflation also have an impact on net support. Our results confirm that this is the case. All coefficients have the expected signs. In a cross-section design, however, the official inflation rate also affects euro support negatively and significantly in the crisis sample.
Table 9.3 Inflation and support for the euro – Probit analysis (individual level), 2002–10 2.1 What Is Novel about Our Findings?
The literature has not yet analysed the impact of the financial crisis on citizens’ support for the euro. We find that the financial crisis – at least so far – has had no impact on public support for the euro when analysing aggregated data with a fixed-effects DFGLS panel analysis. Our findings from the pre-crisis period confirm a negative and significant relationship between inflation and support for the euro. For a pre-crisis sample, a similar result has been established, inter alia, by Banducci et al. (2009). In addition, in accordance with Banducci et al. (2009), perceived inflation also has the expected highly negative signs.
Beyond this consensus, we are able to show that on the aggregate level other variables, such as the budget deficit, the exchange rate, attitude towards the EU, euros in circulation, etc. are unable to influence the support for the euro in any consistent way. This result is in contrast to the findings of Banducci et al., 2003, 2009 and Hobolt & Leblond, 2009, who identify the exchange rate and the attitude towards EU membership as significant drivers of support for the euro.
Our analysis confirms that in contrast to a dramatic fall in citizens’ trust in the ECB (Roth et al., 2011b) driven by the financial crisis, public support for the euro has remained stable so far.