An investigation of the Norwegian consumption function. Income distribution and wealth effects

Abstract

Since the financial crisis, Norwegian private consumption has fallen as a share of household disposable income. This weak development in consumption was not predicted by the contemporaneous consumption models and led to a “structural breakdown” of these models.

This thesis will attempt to build a new model for aggregate consumption that is better able to explain the developments since the financial crisis. This is done by using cointegration analysis to estimate a long run relationship and then include this in an error correction model for private consumption. With a basis in the current consumption function in Statistic Norway’s KVARTS model, the paper demonstrates the breakdown of the incumbent consumption function and conducts two separate analyses into possible explanations for the breakdown.

A first finding is that the income distribution, measured by a Gini coefficient or the wage share, does not seem to affect household consumption on the aggregate level. In another exercise the wealth variable present in the current model is split into different components. In the long run, including net housing wealth and net financial wealth separately seems to improve the model. Financial wealth is a larger determinant of household consumption in the long run than housing wealth. In the short run, the degree of liquidity affects the effect of financial wealth on consumption, while controlling for short run dynamics of debt does not improve the model.