Pierre Hollegien

Bienvenue à Pierre Hollegien, qui nous a rejoint début octobre pour un stage de recherche jusqu’à fin décembre.

Pierre Hollegien est étudiant à l’Université Paris Dauphine dans le Master 2 Economie internationale et développement, mention recherche. Il a effectué toutes ses études à Paris Dauphine et a effectué sa troisième année en échange à la Singapore Management University.

Il a intégré le projet ANR TOFLIT 18 dans le cadre d’un stage d’assistant de recherche auprès de Guillaume Daudin. Il effectuera diverses missions telles que le traitement et l’analyse de données, du travail bibliographique, de l’analyse de documents, et les diverses autres tâches d’un assistant de recherche, avec pour objectif l’extension à la France d’une étude menée par Patrick Wallis sur le commerce des médicaments au XVIIIème siècle en Angleterre (cf http://eprints.lse.ac.uk/28577/1/WP143.pdf).

« Ce stage de recherche mené en parallèle de mon master 2 doit me permettre de me préparer au mieux à une thèse portant sur l’histoire économique en France ou à l’étranger. Après avoir eu plusieurs expériences dans des entreprises privées, j’ai souhaité me rediriger vers la recherche universitaire où j’espère pouvoir exercer dans le futur. Adepte d’histoire et de philosophie, je souhaiterais travailler sur le développement économique au cours de l’histoire, et particulièrement sur le lien existant entre ce-dernier et le développement des Idées en Europe occidentales. Ainsi travailler avec l’équipe ANR TOFLIT18 est l’opportunité pour moi de concilier une expérience concrète en recherche économique et un sujet ambitieux qui m’intéresse tout particulièrement.


CV, Pierre Hollegien, 2014

October 20th meeting

October 20th internal meeting

Present : Hugo Bruel, Loïc Charles, Guillaume Daudin, Béatrice Dédinger, Jérôme Héricourt, Emmanuelle Lavalée, Ivan Ledezma, Guillaume Plique

1) What can the economists do

Ivan showed us a very interesting document summing up the interpretation program planned in the initial proposal (see the attached documents).
Summing up, here are the suggested directions :
A) French trade & trade theory
1) French specialisation.
2) Sources of international specialisation of trade. Was it classical (techno ?) factor endowments (neoclassical) ? Other
3) Study of political events / natural experiments
4) Competition structure in each port

B) Economic development and structural change
1) Sources of growth
2) Market size and economic development.
3) Dutch disease effect of colonial trade
4) Evolution of specific sectors
4) French economic geography
5) Consumption

C) Financial and balance of payment issues (testing mercantilist ideas?)

There is a discussion about what is feasible and who wants to do what.

Pierre Hollegien and Alexandre are working on B-5
Pierre Gervais seems to be interested in A-4

Hugo Bruel seems willing to work on trade policy and trade treaties

Ivan and Emmanuelle are more interested in A-1 and A-2 (especially comparing the « more-autarkic» situation of the revolution to the one before). Jérôme is more interested in A-1. However, because of the difficulties linked with finding international comparative data, these studies might be done at the level of the French regions, linking it to B-4.

Some general principles are discussed :
– Start with the easy and the concrete
– It is easier to do when one looks at specific sectors or political events
– Jérôme believes natural experiments have a potential

We decide that the next economist-focused Toflit meeting will be on 08/01/15.

2) September 19th meeting in Dublin

Béatrice went to a meeting about an Irish foreign trade statistics project (http://www.duanaire.ie/dbases/trade_ireland) organized by Patrick Walsh.
They were interested in our work (and Ricardo). Their project is interesting, but just beginning. They have nice graphs of fiscal data which structure could inspire us.

They seemed to be quite interested in some longer-term collaboration. We discuss it. The obvious outlet would be an European project. Loïc warns against trying for the creation of a large database : as we have seen with the Sound, the discrepancies are just too large and too difficult to explain. Working on frameworks and classifications might be more realistic.

Someone willing and able to head the project has to be found. Paul believes that the medialab experience (50% of the fundings come from projects) could help and that no more than a year would be necessary to set it up. Anyway, this is post-ANR stuff, so we should probably start working on it in mid-2016.

3) Ricardo discussion
Paul must integrate the last version of the database : the Dublin meeting helped us to progress on the interface. There are speed issues, however, so he must do some code optimization.
The expected release date is before Christmas.
We discuss the release of the full database as plain text files. That would allow to put it in in a data archive with versioning (not CKAN, which does not support it). We must decide on a license.
We are not sure when to do it. Immediately at Christmas? A bit later? After a presentation paper in an open archive? After a peer-reviewed paper? What are the risks of someone stealing our thunder?

Béatrice must take the decision. We will discuss it with Paul-André Rosenthal.

4) Geography
Paul met people from the Larhra. They have developed a software, Sylogie, to work with historical information. We could work with them on geographical information (though their data starts in 1815 and might be better fitted to Ricardo, actually).

Next meetings : November 21st 4pm
December 18th 10 am (Big meeting : we have funding for people coming from outside Paris)
January 8th 10 am (Economists)


task 3



October 2nd meeting

Attending : Alexandre Aubourg, Guillaume Bazot, Loïc Charles, Guillaume Daudin, Pierre Gervais, Paul Girard, Pierre Hollegien, Guillaume Plique

1) Measurement unit treatment
Alexander Aubourg presents the system he has put in place to treat measure units and transform them into metric ones
The first database includes 277 measure units and treats those for which the name is enough. For comparison, one year of accounts for a merchant in Bordeaux includes 50 units in Marprof.
The second database includes units + merchandise for those that could not be treated. This includes 1197 lines.
The third base includes units + merchandize + import/export + partner. This includes 258 lines. In fine, that includes only imports, as exports can be dealt with without taking into account partners. A priori this file should get smaller and smaller as we discover that units are used only in one importing direction. Example of razière : used only for goods coming from UK. In fine, probably the French value from the North.
Some units do not exist in French metrology : « barils »

Loïc discusses of the local sources. There are local enquiries which give the value of local units. Units are often used for trade with only one foreign partner. Using the regional data (in which the quantities are given for all years) helps identifying the units used at the national level.

We should probably not do too much on that subject, as most observations are simple to deal with.

2) 1789
Special issues for 1789. We have unit values, quantities and total values treated at the national level. Things work properly for slightly more than 50% of quantities. The other do not fit. The problems are solvable, but at the price of 10 minutes per observation.
Maybe they did aggregate values and quantities and then put an approximate unit value (the middle of the range?).
In that case maybe we should just neglect the unit value.
There is a debate on this question

The issue is not retranscription, it is the verification.
Main errors : averaging of the unit value / decimal error / forgotten column / error with quantity conversion between different units

3) Retranscription
Loïc recapitulates what has been collected. The complete serie for La Rochelle. The whole (not complete) for Rouen and Caen. The selected ones for Bordeaux. The selected ones for Nantes & Rennes.
The national has been done

Marseille should be done by the end of the year

Bayonne : all from 1746. Thasectra.
Montpellier : maybe all (but not complete).

Loïc has done some checks and often the sum of the regional source indeed gives the national source.  Which indicates that the national document is indeed useful with little modifications.

Debate on how to treat the existence of overlapping sources. Paul recommends one table per type of sources. I.e. 3 tables : one for the national / one for the directions / one for the bureaux.

4) Organizations of livedrive

– Organize the jpeg by archival source and reference
– Organize the pdf by sources (national / régional / bureaux). With the name including place – year – type of flow. There is a discussion on wether it should include a «stamp» or page with the archival reference
– Organize the excel files in yet a third group, with the same organization as the pdf

5) Drug paper
We discuss the future work of Pierre Hollegien on drug consumption in France

Next meeting : October 20th at 10am, Médialab, 13 rue de l’Université