Hybrid methodology for sparse selection of generalized estimating equations model for the drivers of firm value

Abstract

The study proposes a two-step hybrid methodology for sparse generalizedestimation equations modeling of the drivers of shareholder value creation.Through the methodology, the validity of the Gordon constant growth modelis established and other non-dividend factors’ contribution to shareholdervalue creation is assessed. The two-step hybrid method involves pickingout the right intra-subject correlation matrix and set of regressors usingEAIC and QIC respectively (EAIC-QIC) and then obtaining the penalizedGEE estimators of the selected model. Penalization is useful in removingredundant regressors from the final model. The performance of the proposedmethod was compared to that of exclusively using QIC method in selectingboth the correlation matrix and set of regressors. The study results showedthat, whereas EAIC preferred the parsimonious order one autoaggressive{AR(1)} structure for the data, QIC preferred the unstructured matrix whichestimates the highest number of correlation parameters. Using the AR(1)structure and Algorithm 2, the GEE model chosen had higher efficiencycompared to when QIC is used to select both the correlation matrix andregressors. Based on the results, the study concludes that adopting hybridmethods enhances efficiency of GEE estimators. On firm value, the studyconcludes that besides the elements in the Gordon-Constant growth model,the financial health of a firm is a vital indicator of value creation by firms.