Do decreasing hazard functions for price changes make any sense?

A common finding in empirical studies using micro data on consumer and producer prices is that hazard functions for price changes are decreasing. This means that a firm will have a lower probability of changing its price the longer it has kept it unchanged. This result is at odds with standard models of price setting. Here a simple explanation is proposed: decreasing hazards may result from aggregating heterogeneous price setters. We show analytically the form of this heterogeneity effect for the most commonly used pricing rules and find that the aggregate hazard is (nearly always decreasing. Results are illustrated using Spanish producer and consumer price data. We find that a very accurate representation of individual data is obtained by considering just 4 groups of agents: one group of flexible Calvo agents, one group of intermediate Calvo agents and one group of sticky Calvo agents plus an annual Calvo process.