Artificial Intelligence and Its Application in the Study of the Legal Complexity of the Value Added Tax Act in Mexico

Abstract

AbstractThe text is a raw material, researchers need to extract information and patterns of value. Through the use of AI tools in conjunction with the hard sciences, it is now possible to access significant sources of knowledge that previously remained hidden in the form of patterns of ideas and feelings stored in large volumes of text. The analysis of the raw text of the Law of Value-Added Tax (VAT) considered the three elements: structure, language, and interdependence. With these three elements, a legal complexity index was constructed, and the results of the model’s parameters show the following: the value for the legal complexity variable was negative (−1.39), which means that when the legal complexity index per unit increases, tax collection will decrease 1.39%. It is helpful to remember that interdependence is the component that outweighs the rest within the legal complexity index. The GDP estimator showed a positive sign, and its magnitude was 4.51; this means that when this estimator increases 1%, VAT collection could increase a 4.5%.KeywordsArtificial intelligenceText miningValue added taxTax collectionLegal complexityJEL ClassificationE62C01H30K34O33