Estimating efficiency spillovers with state level evidence for manufacturing in the US

•We extend a classic non-spatial estimator of efficiency to the spatial case.•Unit effects are used to calculate time-variant indirect (spillover) efficiency.•Direct (own) and total (direct plus indirect) efficiencies are estimated.•The estimators are applied using a spatial cost frontier for US manufacturing.•We observe high efficiencies for Northeastern states.

Unit specific effects are often used to estimate non-spatial efficiency. We extend such estimators to the case where there is spatial autoregressive dependence and introduce the concept of spillover efficiency. Intuitively, we present an approach to benchmark how successful units are at exporting and importing productive performance to and from other units.