Professor Emmanuel Mourlon-Druol is the co-author of the article ‘Economic union in the debates on the creation of the euro: new evidence from the tapes of the Delors Committee meetings’, together with Enrico Bergamini from Collegio Carlo Alberto, in Turin. The article, published in the Journal of Digital History, uncovers the topics discussed within the Delors Committee – the group of central bankers that set out in detail how to create the euro. This analytical work benefitted enormously from the use of digital tools supporting historical research.
Before we discuss your article, could you briefly tell us about your teaching and research activities at the EUI?
I am Professor of History of European Cooperation and Integration at the EUI Department of History. This year I ran a research seminar on the study of elites with Dr Pierre Alayrac, as well as a seminar on digital methods, organised with Professor Giancarlo Casale. I have also been the principal investigator of the ERC-funded research project EURECON, which ended last year. Finally, I am currently finishing a book on the making of Europe’s Economic and Monetary Union (EMU) from the 1950s to the Maastricht Treaty, which will come out with Cornell University Press next year. In the book, I look at the origins of issues that are still bedevilling today’s EU, and, we think, were not tackled extensively in the past. I investigate how European policymakers envisaged an economic union – that is, fiscal governance, the role of the common European budget, banking regulation and supervision – that could support a possible monetary union, and why and how this rich and vivid debate led to the Maastricht construct. The work of the Delors Committee is an important step in that story.
Could you tell us why is the Delors Committee important? Why did you decide to write about the discussions of that Committee?
The process that led to the creation of the euro is based a lot on one report: the Delors Report, published in April 1989, which was extensively incorporated in the Treaty of Maastricht. The Delors Report is, thus, well-known for having set out in detail how to create a monetary union in Europe. But it was also much vaguer on 'economic union'.
The Delors Report was drafted by the Delors Committee, and that is why that Committee is so important in the history of the creation of the EMU. The Delors Committee worked in the second half of the 1980s and included the twelve governors of the central banks of the European Economic Community, plus three experts. Unusually, the Committee members agreed to record on tape all of their meetings. A treasure trove that the archives of the European Central Bank recently made available, and includes the paperwork of the Committee, the tapes, a detailed inventory of them and of the Committees’ proceedings. Historians knew that there existed word-for-word transcripts of the meetings. But what Enrico Bergamini and I discovered is that these transcripts only represented a small part of the entire discussions. The full recording on tape is 47 hours long, and we calculated that only approximately one-third was transcribed. This led us to some questions: Why was only some content transcribed? Did they write down only what they considered important? To answer those questions, we had to dig into the handwritten notes of the two Committees’ rapporteurs and from historians we somehow also became detectives.
How did you proceed with your work? Which methods did you use?
We wanted to analyse what the Delors Committee members actually discussed, more than what the Committee published at the end (the so-called Delors Report). To do so, we used mixed methods, that is, a qualitative text analysis, based on a close reading of the transcripts, together with a quantitative analysis of the Committee’s discussions, both transcribed and in the recorded tapes. We analysed the handwritten notes of the rapporteurs, Gunter Baer and Tommaso Padoa-Schioppa. We realised that, whenever the rapporteurs made an asterisk next to a name, this corresponded to the parts that were transcribed. We can assume that the transcription corresponded to what the rapporteurs perceived were the most important parts of the discussion.
The full transcript is a substantial piece of text, it consists of about 90,000 word text – roughly the size of a PhD or of an academic monograph. As a historian, the main question at this stage for me was how to provide evidence of what was discussed. It would have been very different to analyse 47 hours of recording plus a 90,000 words text without the support of some digital methods, and some questions we could not have answered. We first tried to quantify what was transcribed and what was left out, and used some very basic metrics to understand the words most used for instance.
More importantly, we used Natural Language Processing (NLP), that is, a computational method to analyse text corpora. Here we wanted to uncover 'latent topics' in the text, namely, subjects of discussion that the computation and our interpretation will allow to uncover. From the Delors Report, it was unclear to what extent and how members had discussed the topic of 'economic union', given that this part was vaguer in the report than the 'monetary union' part.
The title of your article mentions 'new evidence'. Please tell us what you discovered.
The new evidence is the tapes of the meetings. Combined to the transcripts, this allowed us to unpack what the Delors Committee discussed. 'Economic union' – that is, fiscal governance, the role of the EEC budget in the monetary union, banking regulation/supervision – is often said to have been overlooked in the negotiations on the creation of the euro. And the Delors Committee too, is often said to have focused predominantly on “monetary union.” However, our analysis found out that the Delors Committee discussed extensively the topic of the economic union. In fact, thanks to NPL, we can show that 'economic union' was the second most debated topic, after the design and the functioning of the future European Central Bank. However, this ‘economic union’ does not feature prominently in the final report. This is due both to the difficulties of defining what ‘economic union’ entails and of agreeing on how to proceed towards an economic union. We also discovered that the theme of banking regulation and supervision was tackled only marginally in the Committee's discussions.
We also discovered some interesting internal dynamics in the Committee. We looked at who spoke the most within the Committee. The first three names are highly predictable: The chair of the Committee, Jacques Delors, comes first, followed by the president of the German central bank, Karl-Otto Pöhl, and of the French central bank, Jacques de Larosière. However, the fourth most active member is not the representative of another big member state or one of the experts, as we could have imagined (e.g. Italy, UK), but instead Maurice Doyle, the governor of the Irish central bank. This is due to Doyle’s keen interest in regional policy, a topic he often raised in his interventions, thereby confirming that ‘economic union’ related topics were regularly tackled.
Please tell us how you divided this vast work.
Our analysis is the results of a real collaborative effort. I would like to highlight first of all how much the work of historians depends on the work of archivists. In our case, one of the archivists of the European Central Bank, Matthias Weber, listened to the entire 47 hours of recording to then establish the inventory we worked on. This is an enormous preparatory work without which we could simply not have proceeded. I wrote the article together with Enrico Bergamini, a PhD researcher in economics who has a strong interest in data science. The way we divided our work was straightforward: Enrico focused on the quantitative analysis and the development of the NLP models, while I concentrated on the contextual background of the meetings and the interpretation of the results. Our work process was a continuous back-and-forth between Enrico and I about methodology and interpretation of results.
I think that one take away from that experience is that historians cannot shy away from those many digital methods that exist because they can be very helpful in our work. The most important is to identify which is the right method to use at the right time to answer our research questions. Sometimes, digital methods are not needed to answer a research question, and that is absolutely fine. But historians cannot pretend some digital tools do not exist and just carry on ignoring them. What matters is to use the right tool at the right time.
And if some of these digital methods require specific technical competencies, there is no need for historians to shy away from co-authorship either. Digital history tools are recent, they are constantly evolving, and they require specific technical competencies. Of course dedicated trainings can help. But why not also combine your strengths and weaknesses with the ones of a colleague, for example, when you realise that you do not possess the technical skills needed to use some of the many digital history methods available? A data scientist and a historian will see the same object through different lenses, and that makes the collaboration only more productive – and, dare I say, fun!
Read the full article by Emmanuel Mourlon-Druol and Enrico Bergamini, ‘Economic union in the debates on the creation of the euro: new evidence from the tapes of the Delors Committee meetings’, published by the Journal of Digital History.
Each academic year, the EUI Department of History chooses a research theme that receives particular attention. In the academic year 2023-24, the selected theme is “Digital History”, and the focus is on how digital sources and methods can be used in historical research. To know more about events or projects linked to the research theme of the year, visit the Department of History webpage.