Relationship between Shareholder Value Added and the Accounting-based Performance Measures: An Econometric Analysis

Abstract

This empirical study aimed to investigate the most important accounting-based performance measures that contribute to shareholder value added (SVA). Our research employed econometric modelling, with a specific emphasis on panel data analysis. We carefully selected variables by analysing the quarterly earnings announcements from companies in the Dow Jones. The chosen variables were then subjected to fixed- and random-effects regressions, with several specification tests used to determine the best estimator. To avoid spurious regression in our panel data analysis, we applied a causality framework that involved testing for stationarity, cointegration, and causality. Additionally, we estimated an error-correction model and a cointegration regression to identify the long-run relationship between the dependent and independent variables. Our sample data covered the 30 companies in the Dow Jones over 20 years (2001-2020), and Stata 17.0 and EViews 12 were used to perform the econometric analysis.

The key findings of the research were as follows: First, the analysis provides evidence that diluted earnings per share, the change in free cash flow, return on assets, the change in total debt, and investment have a positive and significant impact on shareholder value added. Second, the variables are cointegrated, and the cointegration regression was successfully identified. Third, the error-correction term was significant in the error-correction model, confirming its role in adjusting to deviations from the long-run equilibrium. Finally, after thorough testing, a fixed-effects panel-data model was estimated.

Our empirical study focused on accounting-based measures and examined a larger number of variables than prior research. This differs from previous studies that compared the theoretical explanatory power of value-based and traditional accounting-based measurements using only a few variables of each. Our selection of variables was based on semantic analysis of quarterly earnings announcements, ensuring the consistency of our findings. In summary, this study aims to identify the accounting measures to develop an econometric model that predicts annual shareholder value added (SVA).