Abstract
In this paper we first discuss the consumption behaviour of Norwegian farm households. Then, possible consumption models are outlined and our data sample is described. In the next section we discuss the use of panel data methods to estimate our consumption function. Finally, the results from estimating a consumption function with the DPD computer program are discussed. The preliminary results indicate that the GMM estimation using the system estimator of Blundell and Bond (1998) may be superior to the other methods. The parameter of the lagged dependent variable is inside the limits indicated by the OLS and the within estimator. The second lag of the dependent variable is rejected as an instrument and there are strong indications of serial correlation.
Later on, we have to transform the model to make the serial correlation disappear. In a later version of this paper we will test if the MPC is changing over time. We will also test if there are different effects from price- and production dependent agricultural income and other income.
Later on, we have to transform the model to make the serial correlation disappear. In a later version of this paper we will test if the MPC is changing over time. We will also test if there are different effects from price- and production dependent agricultural income and other income.