Does the information environment affect the value relevance of financial statement data?

Abstract

Recent studies demonstrate that the usefulness of financial statement data for valuation of stocks varies depending on specific economy- and firm-level conditions. This empirical study identifies a novel firm-level influential condition. It hypothesizes and finds that for firms that trade at a premium to book value the value-relevance of two fundamental financial statement value drivers (i.e. earnings and book value), is negatively related to the level of sophistication of the firm's information environment. However, for firms that trade at a discount to book value, the level of sophistication of information environment does not affect the value-relevance of these financial statement value drivers. The level of complexity of the firm's information environment is proxied by the firm's capitalized value. The empirical analysis is based on a sample of nonfinancial firms listed on the London Stock Exchange.

Acknowledgements

The author is grateful to Jo Danbolt, John Holland and Antonios Siganos for helpful comments and suggestions.

Notes

¹Collins and Kothari (1989) and Atiase (1987) note that driven by high-information demand, the amount of production and dissemination of nonaccounting information is an increasing function of the capitalized value of the firm, as professional investors devote a majority of their resources to gathering and dissemination information on larger, wider traded stocks. Hence, one may suggest that for larger firms the firm's accounting information is likely to account for a smaller proportion of the entire volume of value-relevant competing information, which is available to investors.

²The residual income valuation model states that the value of equity should equal its book value plus the present value of future expected abnormal earnings. The option-style model, on the other hand, holds that equity market value should equal (i) the value of the firm's resources in place, when the firm is not a going concern, or (ii) the discounted stream of future earnings, when the firm is expected to continue its business as usual.

³Our scale proxy is the arithmetic mean of three of the most commonly used single-variable deflators: market value of equity, total assets and sales. In contrast with the highly asymmetric and skewed distributions of regression variables that are deflated by alternative single-proxy variables, our scale proxy and deflated variables have close-to-normal frequency distributions.

⁴Consistent with Collins et al. (1999), the information content of an individual variable of interest is measured as the difference between the R ² of the full model and the R ² of the model when it omits the variable of interest.

⁵Note that the first quintile firms match the nonfinancial constituents of the FTSE 250 index. Thus, firms in this quintile are the largest in the UK and operate in the most sophisticated information environment. The actual number of observations (in Table 1) varies across quintiles and between the mv  >  bv and mv  <  bv panels because the eliminated outliers and negative book value cases are not equally spread across the quintiles.

Table 1 Changes in value relevance of earnings and book value across quintiles representing different information environments