Observability and Social Interactions in Consumption

Using a randomized survey experiment in the Italian Survey of Consumer Expectations, we study how observability and commonality shape responses to a €1,000 employer bonus. Making an idiosyncratic bonus observable to coworkers lowers the marginal propensity to consume by 3 percentage points. Conditional on observability, extending the bonus to all coworkers raises it by 6 percentage points. Under a social-interactions model, the latter effect implies a social-interaction parameter of 0.16 and a multiplier of 1.2. Thus, consumption responses depend not only on resources but also on whether income shocks are private, public, or shared within a salient peer group.