Time-dependent or state-dependent price setting? Micro-evidence from German metal-working industries

Price setting in German metal-working industries is analysed using a monthly panel of individual price data for more than 2,000 plants covering the period from 1980 to 2001. Motivated by several models in the literature, a duration model is estimated. Price changes can be explained by a combination of state-dependence and time-dependence but time-dependence clearly dominates. Time-dependence is strongest if a price increase follows a price increase. This is typically the case during the observed period. A price increase is most likely to follow a price increase after 1, 4, 5, 8, 9, – quarters. This time-dependent effect is so strong and cost and price increases are so weak in the observed period that adjustment occurs before the sticky price sufficiently deviates from the flexible price, as traditional menu cost models assume. State-dependence seems to be most relevant in periods with decreasing demand. Then prices are reduced and the time between two price reductions only rarely exceeds four months.