This study suggests a way to estimate the aggregate consumption function by estimating and then aggregating individual household consumption functions with micro household data rather than using macro aggregate variables. The primary purpose of this paper lies in setting up a model that uses micro data to estimate and predict aggregate consumption, which is one of the pillars of a national economy. This approach approximates individual household consumption functions to an aggregate consumption function when they exhibit non-linearity with regard to particular household characteristics. This method is considered more adequate for mid-/long-term analyses than for short-term predictions because the analysis time period of the estimated aggregate consumption function needs to be linked with the time period when the micro data are acquired. Also, micro data mainly reflect volatile socio-economic distributions of population.
When individual household characteristics including earnings influence consumptions in a non-linear way, the cross-section distribution of the explaining variables determines aggregate consumption. Along with this, this study adopts the methodology proposed by Hildenbrand and Kneip (2005), which describes with statistical figures such as averages and co-variances the distributional characteristics of socio-economic variables including the age of the head of household, level of education, wealth, family composition and income, and captures changes in the aggregate consumption through variations of the statistically described variables. This approach can heighten the model’s explanatory power especially by taking into account co-variances among variables in estimating the aggregate consumption, while most previous methods depend only on the averages of the variables.
An analysis of data on 3,087 households (from the Korea Labor Income Panel Study in its 1-6th survey sessions) entered into the suggested model, however, failed to show the full advantages of this approach, mainly due to the short time period of the surveys. Despite its limitations, the aggregate consumption function estimation with micro data assists in understanding and predicting more precisely how the changes in the population’s distributional characteristics, for example, income distribution, affect aggregate consumption under stable economic circumstances, which makes it necessary to refine and improve the methodology.
미시자료를 이용한 총소비함수의 추정(Estimating aggregate consumption with household panel data)
|Series Title; No||정책포럼 / 2005-10|
|Subject Country||South Korea(Asia and Pacific)|
|Subject||Economy < Macroeconomics|
|Holding||KDI; KDI School|