Instrumental Variable Model: Using Lagged Sample Clauses

Instrumental Variable Model: Using Lagged. Import Impact as Instrument Xxxxxxxx & Xxxxx'x (1982) method of using the lags as IV was employed. The lag of the Chinese import impact was used as an IV for the first difference in Chinese import impact and assume that errors are sequentially exogenous. The model used is as follows: : ∆ = −1 + ∆ + ∆ + ∆ + ∆ : ∆ ln = ∆ ̂ + ∆ + ∆ + ∆ + ∆ : ∆ ln = −1 +∆ + ∆ + ∆ + ∆ ∆ is the first difference operator and −1 represents the one year lagged Chinese import share. ∆, ∆ and ∆ are time, country-time and country-industry fixed effects. Errors are clustered at country-industry pair level. For productivity variables, the following model was used: : ∆ = −1 + ∆ + ∆ + ∆ : ∆ ln = ∆ ̂ + ∆ + ∆ + ∆ : ∆ ln = −1 + ∆ + ∆ + ∆ Where ∆ is the first difference of the natural log-transformed productivity variables. ∆ and ∆ are the time and industry dummies respectively. Errors are clustered at industry level. For both models, ∆ ̂ is the first differenced Chinese import share estimate based on the first stage regression. γ < 0 is expected as the first difference is constructed such that the lag of the variable is negatively related to the first difference.