STRUCTURAL EQUATION MODELING AS A TOLL FOR ECONOMIC RESEARCH

Abstract

In this paper, the benefits of applying the Structural equation modelling (SEM) technique to economic research are methodologically analyzed. SEM is a second-generation multivariate analysis technique for simultaneously measuring and analyzing complex cause-effect relationships. Interest in this technique has grown since a large amount of software has been developed for applied SEM, where the theoretical model is created using simple drawing tools (Arbuckle, 2017). SEM allows simultaneous standard and non-standard modeling of relationships between latent and observed variables with the ability to process longitudinal data, time series, elimination of autocorrelation and analysis of non-normally distributed variables. Techniques such as multiple and logistic regression and analysis of variance, have certain initial disadvantages that can affect the accuracy of the measurement: 1. simplification of the model structure, 2. they require that all variables be considered observable from the start, and 3. they assume that all variables are measured without error. With the SEM technique, often unobservable constructs, that are indirectly measured by several indicators, can be modelled simultaneously, taking into account measurement errors, so that the results are more accurate compared to standard techniques. To illustrate the methodological features of SEM, this paper uses a theoretical background, a small graphical manual and an own applied empirical analysis from practice. The study presents the results of an applied empirical research on corporate social responsibility conducted on a sample of 105 companies in the Italian region of Campania. The data were collected using a questionnaire, and the methodological technique applied was an adapted SEM that involved a wider range of factors. PCA was used to identify preliminary factors, which were rotated to reduce the number of variables. The generated principal components were integrated into SEM, where the testing of various inter-associations and the measurement of the impact on the company's profit were simultaneously carried out. Company profit was measured on a Likert scale, for the period covering the last two years. All three principal components, GEWR, APG and EFO had a significant and positive impact on the growth of company profits. An increase in GEWR by one point leads to an increase in profit by 0.57. A jump in APG by one point leads to an increase in profit by 0.45. The growth of EFO by one point, causes an increase in profit by 0.3. Following the size of the company, the type of activity, the size of the budget and the ecological situation in the local community, this study gives recommendations to managers for an effective choice of strategy.