The R2 Puzzle

Previous research has argued that the degree of co-movement of stock returns (the R² of a market regression) at country-level can be explained by the interaction of firmspecific and market-wide information. The R² measure has been used to investigate a number of issues of potentially great importance to accounting, such as whether countries with poor corporate governance regimes and weak legal protection of private property rights are more likely to have poor information environments or to assess the informativeness of prices. To date, only limited research has been carried out to assess the reliability of an information interpretation of the R² measure at a firm-level within a country rather than at an aggregate country level. In this paper we now examine the properties of stock returns co-movement at the firm-level within two countries, UK and USA, thereby being able to filter out certain extraneous factors that could arise in cross-country settings. We analyse the performance of this overall measure by triangulating it with other information-related measures which previous research has suggested capture partial aspects of the information environment. We find some serious flaws in the methodology and our findings suggest that when using it at firm-level, it may be being driven by other factors related to uninformed trading.