| Abstract [eng] |
Relevance of the topic. Household consumption, together with investment, government expenditure and net exports, is used in calculating the most important indicator reflecting a country’s economic welfare – gross domestic product. Analysis of household consumption expenditure and its structure can be one of the indicators that helps to assess not only the country’s economic situation but also the population’s quality of life, since a higher level of consumption is considered to correspond to a higher standard of living. Object of the study – total household consumption expenditure and its structural components. Aim of the study – to assess the impact of micro-level factors on total household consumption expenditure and its structure. Methods – scientific literature review, statistical tests, machine learning models. Tasks of the study: 1. to review the factors determining changes in consumption expenditure on the basis of a scientific literature analysis; 2. describe the data used in the study, their preparation steps, and the modeling methods applied; 3. to perform feature selection in order to identify the variables affecting household consumption expenditure; 4. to build a model to determine how consumption expenditure and its structure depend on various household characteristics in Lithuania. The project consists of three main stages. A scientific literature review identified the main consumption related factors and their impact and also discussed previous research and methods applied. The methodological part described the data used, their characteristics and provided a literature analysis of the methods applied. The empirical part presented the research decisions, the best models obtained, their results and interpretations. Based on SHAP analysis applied to the extreme gradient boosting model, it was found that the growth of total consumption expenditure of Lithuanian households is most closely related to a larger household size. Age and consumption expenditure are linked by a non linear relationship. Expenditure increases if households belong to a higher income quintile, have larger usable floor space, raise children, have higher education or work in higher-skilled occupations. Similar conclusions were drawn when analysing the categories of food and non-alcoholic beverages, information and communication services, and insurance and financial services expenditure. In addition, it was established that expenditure on information and communication rises if a person does not live in the countryside, is an employee and does not live in a single-family house, while expenditure on insurance and financial services increases if the person’s main income is not a pension, does not live alone, and is a homeowner with an outstanding mortgage. |
