Effects of misspecification of lag structure in certain two-variable distributed lag models

摘要: Abstract Distributed lag models are a type of dynamic econometric model often used in demand analysis. Such take account the fact that consumers do not respond instantaneously to changes their economic situation. The Koyck model, qt = δqt−1 + βppt βyyt ϵt, is particularly simple form fitted data. If logQt, pt logPt, yt logYt where Qt, Pt, Yt quantity, price and income data, respectively, then βp, βy represent one-period elasticities demand, while β p (1−δ) , y long-term elasticities, respectively. This paper considers data generated by which has different geometrically declining responses yt, specified parameters δ1, δ2, It determines large sample biases elasticity estimates arising from incorrectly fitting either static or such using ordinary least squares.