Innovative approaches to the assessment of the impact of the shadow economy on social development: an analysis of causation

The article investigates the causal links between the level of the shadow economy and indicators of social development of 13 European countries. Based on the analysis of research results of scientists, three hypotheses concerning relationship between the analyzed indicators: there is a one-way link between social development indicators and the level of the shadow economy; there is a one-way causal link between the level of the shadow economy and the country's social development; there is a bilateral relationship between the analyzed indicators were tested. Based on the causal analysis the main hypothesis that the increase in the level of the shadow economy is accompanied by a decrease in the level of social development of the country is proved. The purpose of the paper is to analyze the correlation between the level of the shadow economy and the indicators of social development of the country using methods of causal analysis and panel data from 2010 to 2019. Based on comparative analysis of the levels of the shadow economy, share of the population at the level of poverty and share of the population in a difficult financial situation the statistically significant indirect relationship between indicators has been proved. The analysis of the relationship between the level of the shadow economy and the share of personal income tax and social contribution to GDP was made using the Granger test. Reliability established hypotheses and the nature of the relationship between the indicators on the basis of the construction of end-to-end regression (Wald test), regression with fixed individual effects (Brush-Pegan test) and regression with random individual effects (Hausman test) was carried out. Empirical calculations show that it is appropriate to describe the relationships between the level of the shadow economy and the amount of single social contribution, personal income tax and Index of Social Development by using a model with fixed individual effects. All indices are statistically at 5% significant level.