An examination of investors' use of nonfinancial measures
MetadataShow full item record
In this dissertation, I examine investors’ use of nonfinancial measures (e.g., customer satisfaction) on two dimensions. First, I consider the relevance of nonfinancial measures for investors and introduce a prescriptive four-step model for using such measures in conducting fundamental valuation. A model that articulates how investors can use nonfinancial measures is important because such measures can provide forwardlooking information that is not available from traditional accounting measures. However, nonfinancial measures only benefit investors if they can interpret and integrate the measures with other available information. Performing these tasks with nonfinancial measures is difficult for investors because they observe nonfinancial measures presented in a variety of ways, as no standards currently exist regarding such disclosures. The model provides a framework for investors to realize the benefits of using nonfinanical measures by articulating a process that facilitates interpreting and integrating these measures to make investment-related judgments in the current financial reporting environment. My dissertation also considers investors’ judgment problems associated with observing nonfinancial measures presented in various ways. Specifically, I conduct an experiment to investigate judgment problems that investors encounter when they observe nonfinancial measures presented using different scales (i.e., nominal, ordinal, interval, or ratio). My results provide evidence that investors rely more on a nonfinancial measure when its scale matches the scale the investor uses for his judgment, at the expense of an equally relevant measure whose scale does not match his judgment scale (termed a scale compatibility effect). Importantly, the effect exists in investors’ valuation-related judgments (e.g., assessments of price-earnings multiples). My results also show that the scale compatibility effect is reduced when investors evaluate several values for nonfinancial measures simultaneously. Implications of these results are discussed.