The desperation threshold: a model to explain decisions in poverty

Abstract

In this dissertation, I propose the desperation threshold model, which seeks to explain decisions in situations of poverty as rational ones. The model assumes that people in poverty aim at staying above a "desperation threshold" -- a critical amount of resources that represents their basic needs. Throughout the dissertation, I show that this assumption can explain several behavioral patterns associated with poverty -- and with economic inequality at the population level. I then try to assess its empirical validity. I explore this idea in two modelling chapters, one empirical chapter using secondary data, and an integrative literature review chapter. In the first modeling chapter (Chapter 2), I show that the desperation threshold can account not only for the high proportion of property crime in deprived or unequal populations, but also for the frequency of violent crime, as a toughness signal to deter exploitation. In the second modeling chapter (Chapter 3), I show that the desperation threshold generates elevated time discounting when basic needs are on the line. Chapter 4 aims at testing the model's prediction empirically. I show that the empirical literature is currently split, with studies predicting and finding either an increase or a decrease of risk taking in situations of poverty. I propose that the desperation threshold model can reconcile these seemingly contradictory findings. I derive testable predictions from the model, and test them on secondary survey data from France and the UK. I obtain partial support for a U-shape between resources and risk taking, but a clear polarizing effect: both risk taking and risk avoidance are more frequent among the poorer participants. In the integrative review chapter (Chapter 5), I give more perspective on the desperation threshold. I first state the model in its purest form, highlight and justify its core assumptions. I then review the empirical evidence from diverse fields and methods, relevant to test the validity desperation threshold model. I find that similar models -- under different names -- have been proposed and tested across several social and biological science disciplines, though their full implications and explanatory potential have rarely been recognized. I conclude that the desperation threshold model, along its intellectual antecedents, has promising empirical support, and is already a fruitful paradigm for the study of poverty.