—The Economic and Statistical Significance of Stock Returns on Customer Satisfaction
According to Jacobson and Mizik [Jacobson, R., N. Mizik. 2009. The financial markets and customer satisfaction: Reexamining possible financial market mispricing of customer satisfaction. (5) 810–819], excess stock portfolio returns for firms with strong customer satisfaction are small and statistically insignificant, and if there is any above-market performance at all, it is due to a small set of firms in the computer and Internet industries. But their data seem to suggest the opposite. The returns are actually both exceptionally large and significant. Using monthly data, their portfolio consisting of strong American Customer Satisfaction Index (ACSI) firms outperformed the market by 0.0053, corresponding to 6.4% cumulative risk-adjusted above-market returns on an annual basis over a 10-year period—a performance that would beat at least 99% of all large-cap U.S. stock funds tracked by Morningstar. Using a different treatment of risk, their annualized risk-adjusted return is a whopping 8.4% better than market. After eliminating computer, Internet, and utility companies, they find that the monthly risk-adjusted abnormal returns drop to 0.0045, which corresponds to an annual above-market return of 5.4%. This too is better than 99% of all actively managed stock funds in the population. Yet Jacobson and Mizik conclude that these returns are not statistically significant and that there is no evidence that stock returns from firms with strong customer satisfaction outperform the market over the long run. The failure to reject the null hypothesis is probably due to a lack of statistical power in Jacobson and Mizik's analysis. We discuss why this is likely the case and then present new data updating the results from our original article [Fornell, C., S. Mithas, F. Morgeson III, M. S. Krishnan. 2006. Customer satisfaction and stock prices: High returns, low risk. (1) 3–14]. The above-market returns persist and are both economically and statistically significant.
