Bertil Hatt

Since November 2005, I am working on a PhD under the supervision of Professor Éric Brousseau --- under a joint contract between ÉconomiX & Orange Lab Sense (Sociology & Economics of Networked Services).

(For insiders: a recent internal presentation, on April 2007)

Research topic

Brian Arthur and Paul David showed that positive externalities led to a natural monopoly. This result considers that each user values equally every previous usage: for phone or mail, only the relatives matter. How can we introduce this relation constraint?

Since 1999 properties of large networks have been studied thanks to digitalization:

some agents are far more connected, though the majority remains hardly active;
agents are homophile: they preferably link by groups, but paradoxically any pair can be tied with a chain of hardly six steps on average.

These apparent contradictions are resolved by a unique structure, both resilient, robust and legible, locally efficient ofr individal users.

In 2005, distinctions appeared between large networks: in social, conscious networks, the better connected agents tend to be connected to other well connected agents ; this degree-homophily or “assortativity” is decisive in the network efficiency --- like the very uneven degree distribution.

Along what criterion can we structure social network?
What are the impacts on the diffusion of an innovative service ?
How does this structure impact competition model?

Three perspectives for one project

A first paper tries to isolate elementary social insitutions, link them to the social network structure, and link them to database statistics, such as:

comments on blogs and blogroll;
semantic and content association through group and tagging;
anonymised communication database (most internal, confidential so far)

A second paper considers the impact of assortativity on the adoption of a technology: local clusters and cores encourage viral adoption; without, a reasonnable service that demands massive adoption to be relevant might remain ignored. Product recommendations could emerge from this modeling.

Finally, a last paper considers competition: two coherent services, by being imperfectly compatible are not benefitting from network compatibility; they are loosing the undecided and, by trying to secure their base, they can block a service from appearing. An open competition can therefre by beneficial for users, but also providers who might neglect the impact of their decision.

(Translation in progress, please come back tomorrow)

Curriculum Vitae

Academia

2005-2008 Paris-X Nanterre & FranceTelecom R&D

PhD on the diffusion of a social service

2003-2004 Paris-X Nanterre

MPhil Insitutional Economics

Main thesis on Comparing models of Market intermediaries
Other work on Semantic content of Job offers, Social Network Services, Market setup process

1999-2003 Énsaé (National School for Statistics and Economic Administration -- ParisTech)

Majoring micro-economics
Final thesis on word-of-mounth in the job market
Other thesis on: Cumulative patents, Magazine readership, Poland political transition and Student life conditions

1997-1999 Competitive exam preparation

Work Experience

2005-2008 Orange Labs

PhD student (business model, content & intermediation)

2004-2005 PA Consulting

Consultant-analyst (energy, Europe)

2003-2004 EDF R&D

Market analyst

(European market watch)

2001-2003 Air France

Junior Enterprise Consultant

(setup a monitoring tool for marketting costs)

été 2000 Fitch Ratings

Quantitative Analyst

(feasability of company basket rating)

été 1999 Sinopia AM HSBC

Assistant-manager for structured funds

If you are interested in my research, do not hesitate to look at my blog.