Binary decision making, with imperfect information
This paper develops a framework where an individual has to make a binary decision, but finds it difficult to assess the true state of the world. In the model, states can be described completely by finite sequences of zeros and ones. By observing partial information, the decision maker can reveal the true state of certain bits, which he uses to update his prior beliefs about the true state of the world. I describe the data such a model would generate and show that which data can be characterized by the model. In particular, the datasets need to abide by an axiom that eliminates choice reversals. Later I introduce the idea of distance to rationalizability; this measure describes how far a given dataset is from being rationalizable using our model. Finally, the computational process to generate such distance and to recover preferences is shown using a dataset from Mexican labor courts.
Autor:
Diego Javier Jiménez Hernández
Número de revista:
34
