Let's speak more? How the ECB responds to public contestation

ABSTRACT

Although the post-crisis politicisation of the ECB is widely acknowledged, little empirical evidence exists about how this important non-majoritarian institution has responded to public contestation. This article starts filling this gap by investigating whether and how public opinion affects ECB communication. Based on automated text analysis of the speeches delivered by Executive Board members (2001–2017), the article shows that negative public opinion is associated with an expansion of the scope of ECB communication and a reduction in the salience attributed to monetary policy issues. These results challenge the view according to which the ECB conceives of its sources of legitimation based almost exclusive on the achievement of its mandate. In particular, our findings suggest that increased politicisation leads the ECB to reassess the sources of its legitimation strategy from a strategy based on output achievement towards one based on participation to broader policy debates.

KEYWORDS:

• ECB
• legitimacy
• politicisation
• communication
• central banking
• public opinion

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes on contributors

Manuela Moschella is associate professor of International Political Economy at the Scuola Normale Superiore.

Luca Pinto is assistant professor of Political Science at the University of Bologna.

Nicola Martocchia Diodati is a postdoc fellow at the Scuola Normale Superiore.

ORCID

Manuela Moschella http://orcid.org/0000-0002-3159-9533

Luca Pinto http://orcid.org/0000-0003-1165-4475

Nicola Martocchia Diodati http://orcid.org/0000-0003-4743-8918

Notes

1 Financial Times, Anti-capitalist protesters target ECB in Frankfurt, March 18, 2015; Financial Times, ECB adjusts to life amid the protests, April 23, 2015.

2 Benoit Coeuré, ECB Executive Board Member, as reported in The Financial Times, ECB hits the road to pitch its policies, 8 October 2018.

3 We also investigated whether reverse causality exists between public support and communication expansion. Our empirical tests (Clausen, Kraay, & Nyiri, 2011) suggest that the presence of reverse causation does not undermine our conclusions (see Appendix A).

4 For instance, as compared to other major central banks in the advanced economies, the ECB has not published account of monetary policy meetings until 2015 and these accounts do not provide information on the individual preferences of ECB policymakers.

5 Our analysis starts in 2001 in order to discount potential problems related to data availability and public awareness of ECB activity that might have been limited in the first years of its operations.

6 All speeches were automatically retrieved from the Bank for International Settlements website. The speeches of Sirkka Hämäläinen (1998–2003) and Eugenio Solans (1998–2004) were not available.

7 Information is regularly provided also through research publications and, since 2015, through the publication of the accounts of monetary policy meetings.

8 More than 80 per cent of the speeches in our corpus are not part of the ECB legal and operational obligations such hearings and press conferences.

9 We estimate topic models using the stm package in R (Roberts, Stewart, & Tingley, 2018).

10 We aggregated topics to help the reader to better visualise the results of the topic model analysis. The only relevant macro-category for the purpose of the statistical analysis is the one on monetary policy, which we discuss in what follows.

11 Measures of issue or attention diversity have been extensively used in several studies related to public policy, agenda formation and party competition (see Greene, 2016).

12 Respondents were asked to choose among three categories: “Tend to trust it”, “Tend not to trust it” and “Do not know”. For each year, data were taken from the Eurobarometer of May and October. Data on net trust in the ECB were matched to the speeches according to the date of the document. Population-weighted trust values are employed. We also compute a measure of net trust limited to Eurozone countries. The results show that there is no significant difference between the impact of the former compared to net trust at the European level (see Appendix E). Given the role played by the ECB also in non-eurozone countries, we thus used data on EU trust. We also included two variables measuring the difference between net trust toward the ECB and net trust toward the EU Parliament and the EU Commission. We found that only trust in the ECB is significant for ECB communication (data are available upon request).

13 Data are taken from the annual consolidated balance sheet of the Eurosystem at the end of each month.

14 We run the analysis using different model specifications: a speaker-level fixed effects; random intercepts for speaker and time; a random slope for time-counter between speakers; a model including all the ECB representatives who delivered at least ten speeches; and, finally, a model including only observations starting from 2002 (the year of Euro introduction). The results are similar to those reported. Data are available upon request.

15 Both unemployment and public trust show a very high variance inflation factor, indicating that knowing the former is possible to predict the latter quite well.

Additional information

Funding

This work was supported by Italian Ministry of Education, SIR Grant 2014 ‘Unconventional Central Banks: Making Monetary Policy in Hard Times’ [RBSI14KCWY].