Tax Αvoidance and Transfer Pricing: A VECM Regression Model
Purpose: The study examines the tax incentives related to pricing decisions between affiliated companies which are tax residents of Greece and focuses on the pricing behavior of intra-group transactions among related parties. Design/methodology/approach: In the context of the empirical analysis, a panel data regression analysis was performed, using a vector error correction model (VECM) with two lags. The data used in the analysis were retrieved from the AMADEUS (Analyze MAjor Databases from EUropean Sources) Tp-Catalyst (TP 96, March 2018 edition) database and consists of 2,131 companies from most Greek economy sectors, of which 971 are independent and 1,160 are affiliated with other companies. The total sample consists of 17,048 observations, of which 7,768 refer to independent companies and 9,280 to affiliated companies. The research span is from 2010 to 2017 and a two-lag VECM regression model was implemented. Findings: The results of the study are generally in compliance with the international literature. According to the study, affiliated companies appear to have lower profit margins, lower tax burden and a lower Berry ratio than independent companies. Practical implications: The study elaborates on the positive effects of the new Income Tax Code and the new Tax Procedure Code, according to which companies operating in Greece and conducting intra-group transactions, are obliged to document such transactions in the context of a price documentation file. Originality value: The empirical research of the present paper is unprecedented in Greece. The size of the sample in terms of number of companies, number of available ratios and time depth, allowed the analysis through a VECM model to examine the possible use of transfer pricing by affiliated companies to decrease their tax burden. A similar lagged model of autoregressive vectors has not been identified in the relevant literature. The results provide the authorities a benchmark for future audits, since indicate that the manipulation is significant both in short, as well as in a long-term time frame. The use of a VECM model with two lags revealed that the manipulation is not limited within one financial year and the authorities should inspect data from a wider time range when conducting their audits.