Panel estimation of state dependentadjustment when the target is unobserved

Understanding adjustment processes has become central in economics. Empirical analysisis fraught with the problem that the target is usually unobserved. This paper develops,simulates and applies GMM methods for estimating dynamic adjustment models ina panel data context with partially unobserved targets and endogenous, time-varyingpersistence. In this setup, the standard first difference GMM procedure fails. I proposethree estimation strategies. One is based on quasi-differencing, and it leads to two different,but related sets of moment conditions. The second is characterised by a statedependentfilter, while the third is an adaptation of the GMM level estimator.[...]