The seminar will be physical at FASS 2054 at 11.45 but it can also be attended at the following link https://sabanciuniv.zoom.us/j/91383976765
"Decision Making with Recommendation"
Recommendations play an undeniable role in decision-making. The empirical literature argues that recommendation can influence demand through two distinct channels: i) by enlarging awareness (attention channel), or ii) by altering preferences (utility channel). In this paper, we develop a framework to study these two channels. We illustrate when and how one can distinguish through which channel the recommendations affect choices. We offer both deterministic and probabilistic models. While deterministic models aim to identify the basic observable behavioral differences between these two channels, our probabilistic models are suitable for econometric estimation, which is crucial for studying aggregate behavior used in empirical work. Our parametric models offer unique identification under minimal data requirements. This enables us to make out-of-sample predictions for counter factual analysis for policy design purposes. In addition, we offer simple and intuitive behavioral postulates characterizing each model so that one can test our models.