Underestimated or overestimated: matching function elasticities biased due to worker inflows and outflows

Matching function elasticities are biased, that is, underestimated or overestimated, if inflows and outflows of job seekers are omitted. I analyse mathematically the direction of this bias and estimate parameters for random, stock-flow and job queuing models, adjusted econometrically for the temporal aggregation bias in the data. I analyse various sets of endogenous and exogenous variables from registered unemployment data in Poland. Estimates show that unemployment affects matching more than discouragement, and that unemployment stock elasticity is underestimated if discouraged workers are neglected.