Produire des statistiques : pour quoi faire ? L'échec de la statistique des faillites en France au XIXème siècle

Some statistics are developed within a consistent intellectual, political and administrative project, as was the case of criminal statistics which appeared and developed with criminology during the 19th century, particularly in France and Belgium. We examine the official statistics of civil and commercial justice, and particularly those of bankruptcy, which were created by the same administration as the criminal statistics in the same period (in the 1830s). We show that the statistics were well done, but that they were unable to attract users in spite of an early use by the justice administration as a management device. Neither members of Parliament nor social scientists used them, probably because they weren't developed in order to answer adequate questions that would have been embedded in a clear and developing scientific framework. This led to a decline of these statistics, in spite of the fact the data they contained provide interesting insights on 19th century's society and economy.