This paper details a method for estimating the unknown parameters of a regression model when the estimates of the dependent variable should be embedded in an input-output table with accounting constraints. Since in regression modelling the dependent variable is usually transformed either to achieve homoscedasticity of the residuals or for a better interpretation of the model, the estimating procedure becomes an optimization problem of an opportunely defined Lagrangian function with non-linear constraints. After detailing the algorithm and deriving the asymptotic distribution of the restricted estimator, the methodology is applied to estimate the flows of tourism within and between Italian regions with a gravity model. The procedure can be seen as an extension of Byron's (1978) balancing method.
Regression modelling of the flows in an input-output table with accounting constraints
STANGHELLINI, Elena
2009
Abstract
This paper details a method for estimating the unknown parameters of a regression model when the estimates of the dependent variable should be embedded in an input-output table with accounting constraints. Since in regression modelling the dependent variable is usually transformed either to achieve homoscedasticity of the residuals or for a better interpretation of the model, the estimating procedure becomes an optimization problem of an opportunely defined Lagrangian function with non-linear constraints. After detailing the algorithm and deriving the asymptotic distribution of the restricted estimator, the methodology is applied to estimate the flows of tourism within and between Italian regions with a gravity model. The procedure can be seen as an extension of Byron's (1978) balancing method.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
