Mixed frequency data in a heterogenous sticky price model

We develop a model to estimate price-adjustment behavior when prices are observed more frequently than key explanatory variables such as wage costs. We propose a mixed-frequency stochastic (S, s)-model that accommodates infrequently observed costs and allows for plant-, product-, and season-specific heterogeneity. The model is estimated using a likelihood-based nonlinear state-space approach, enabling estimation as if all variables were observed at the same frequency. Applied to monthly price survey data matched with plant-level annual cost data for manufacturing producers, the model yields precise estimates of both cost pass-through and price-inaction thresholds, and reduces the blurring of intermittent price changes.