This paper explores the effect of learning by doing for smuggling activity in Taiwan. A theoretical model for empirical testing takes explicit account of the learning by doing and risk associated with smuggling. The period considered is from 1966 to 1990, and the effect of learning by doing for police expenditure is also examined. Police expenditure is treated as a public good.
The Almon polynomial distributed lag estimate and the minimum final prediction error are employed to determine the effect of learning by doing. The empirical results show that both smuggling activity and police expenditure do have an effect of learning by doing, with a one year lag. In the short run, a more vigorous policy towards the enforcement of laws against smuggling will raise the risk of smuggling and which in turn may reduce illegal imports.
