Dynamic Linkages between MVA and Internal Performance Measures: A Panel Cointegration Analysis of the U.S. Insurance Industry

• Published in: Assurances et Gestion des Risques 2011, Vol.79 (3-4), p.223-250
• Main Authors: Bernier, Gilles, Mouelhi, Chaouki
• Format: Article
• Language: English
• Subjects: Earnings per share; Econometrics; Economic value added; Insurance industry; Internal controls; Market value; Return on assets; Value added
• ISSN: 1705-7299; 2371-4913
• DOI: 10.7202/1091876ar

This article relies on recent developments in econometrics of non-stationary dynamic panel data in order to examine the linkages between a firm’s market value added (MVA) per share, a measure of its external performance, and well-established internal performance measures. Indeed, several panel unit root tests and two panel cointegration tests [namely Kao (1999) and Pedroni (2004)] are applied on a sample of 24 firms from the U.S. insurance industry over the period 1991-2004 to test for the existence of a long-term equilibrium relationship between MVA and the following five internal performance measures: earnings per share (EPS), free cash flow per share (FCF), economic value added per share (EVA), return on assets (ROA) and return on equity (ROE). Our main results show that cointegration between MVA and EVA as well as between MVA and ROA are the two more statistically powerful relationships. Several explanations are provided for the above findings supported by robustness analyses using panel error-correction models (PECM) and a comparative analysis of cointegration versus correlation results.