|Year of selection||2017|
|Institution||Institut des politiques publiques – IPP|
Type of support
130 000 €
In all complex systems, the overall impact of a change in variables is hard to predict. The French
tax and benefit system is no exception. Its non-linearity, the diversity of individuals’ characteristics,
and the interaction between different taxes and benefits make it arduous for policymakers or
analysts to properly anticipate the impact of reforms. In view of improving the policy debate in
France, Dr. Brice Fabre’s post-doctoral project aims to build a dynamic computer model
encompassing the complete French system of taxes and public benefits. This ambitious enterprise
relies on the use of French administrative data, « which has been made recently available to the
research community through secured access », the researcher explains. A prime objective of the
model is to incorporate behavioural responses to changes in tax and benefit systems, as « they are
key to discriminate between alternative policy options ». What is more, the model will be available
in open source.
The type of computerized analytical tool Fabre is working on is called microsimulation. « In simple
terms, it consists in simulating taxes and social benefits of each micro-level unit (an individual, or a
household) of a database representative of a given population, under any tax and benefit system.
Then, it is possible to draw conclusions that apply to higher levels of aggregation such as an entire
country », explains Dr. Brice Fabre. « To estimate the effects of various reform options, it is
standard for public administrations to rely on a microsimulation model which takes into account the
full diversity of individuals’ characteristics ». « But these current state-of- the-art models are based
on household surveys, with limited information on detailed tax characteristics, and a limited
sample », he pinpoints.
Data fusion : towards more accurate results
This project is expected to create an input of much improved quality by being based on a large set
of administrative data, including anonymous income tax records, benefit records, social security
records, etc. To integrate this data into the model, the research makes use of a process called data
fusion. « Data fusion is a rapidly developing technology in the context of Open Data. It is receiving
growing attention in economics, but has rarely been used in the context of microsimulation », the
researcher points out. Another important innovation of Fabre’s project is the investigation of
behavioural responses to changes in taxes and benefits and their impact on total tax revenues.
Indeed, by matching various administrative data together, the objective is also to study how
individuals are most likely to react to various reforms, taking into account labour supply, tax
avoidance, income shifting, etc. A database merging different administrative data sources is
expected to give more precise evidence and to bring new knowledge on these behavioral responses.
« The development of such a model, available in open source, is key to improve the policy
debate on changes in the tax and benefit system », says Fabre. « This is all the more true in a
context where there is a growing attention towards structural reforms of our tax systems. » By
aiming to make his model available in open source, Dr. Fabre’s project promises to considerably
change the expertise available in the public debate and bring about game-changing social