On Friday, December 5, 2025, Chloé Beaudet will defend her thesis entitled “Towards sustainable lighting: Socio-economic analyses of light pollution reduction policies,” under the supervision of Maïa David (AgroParisTech, PSAE) and Léa Tardieu (TETIS, INRAE), at AgroParisTech (Campus Agro Paris-Saclay, 22 place de l’Agronomie, Palaiseau), amphitheater C2.0.37.
This dissertation focuses on the policies implemented by French municipalities regarding public lighting to reduce light pollution, and their socio-economic consequences. It aims to provide new insights for integrating the societal dimension into decision-making on public lighting, a dimension often overshadowed by ecological concerns alone.
In the first chapter, we focus on the social acceptability of light pollution reduction policies. Using the case study of the Montpellier metropolitan area, we rely on a discrete choice experiment to evaluate citizens’ willingness to pay for three types of policies: reducing light intensity, switching off public lighting, and changing the color of light from white to orange.
A latent class model identifies two groups of preferences: one generally favorable to the proposed policies, the other rather unfavorable, particularly toward switching off lighting between 11 p.m. and 6 a.m.
The second chapter compares two methods for mapping preferences from a discrete choice experiment at a fine spatial scale. We evaluate them in two
ways: first through a theoretical case based on Monte Carlo simulations, and then by applying them to the data from Chapter 1 to map preferences at the infra-municipal level. The results show that one of the methods is the most effective, and better captures spatial heterogeneity in preferences.
The third chapter develops a decision-support tool for lighting policies in the Montpellier Metropolitan Area, integrating both ecological and societal needs. Two spatial indicators are constructed: an ecological indicator, based on light pollution data and modeling of the needs of six species, and
a social acceptability indicator derived from Chapter 2. The combination of these indicators, integrated into an user-friendly application, provides policymakers with a tool to prioritize actions and highlights the importance of adapting lighting policies to local contexts (down to the streetlight) rather than applying a uniform approach.
The fourth chapter introduces the construction of a novel database on public lighting switch-off policies in French municipalities with more than 1,500
inhabitants. Using radiance time series from nighttime satellite data, we apply a break detection model, followed by a random forest classifier to distinguish switch-offs from other types of changes (renovation, reduced intensity). We first show that 64.4% of French municipalities adopted
a switch-off policy between 2012 and 2023, including 53.3% after July 2022, and then identify the profiles of municipalities associated with the adoption of these policies.
Finally, the fifth chapter evaluates the causal effect of public lighting switch-off policies on five types of crime between 2016 and 2023, using a staggered difference-in-differences approach. The results indicate that switching off public lighting has no impact on the types of crime under study, except for burglaries, where we observe a slight increase, mainly driven by high-density municipalities.