In this paper, we mined the financial data of several enterprises in CSMAR database and constructed an asset pricing model based on level, slope, and curvature (LSC). By combing the characteristics of data assets and their value drivers, the key factors affecting asset pricing are extracted with the help of principal component analysis. Then the relationship between corporate credit risk and asset excess return is analyzed by using Fama-Macbeth one-factor regression method. And combined with GRS and other tests, the explanatory power of asset pricing on returns under different models is comparatively assessed. The credit risk coefficients of the models in the FamaMacbeth test range from -0.473 to -0.115, and there is a significant negative correlation between them and the portfolio excess returns. There are monotonically increasing or decreasing trends in indicators such as expected excess return and average excess return. And the predicted return of the asset pricing method in this paper is closer to the real return, with the same upward trend. The first three components obtained by principal component analysis explain 90.3%, 3.7%, and 1.4% of the portfolio, respectively. The GRS statistics of the LSC pricing model in this paper are lower than the baseline model by 0.286 to 0.930, which has stronger pricing explanatory power. This study expands the theoretical framework of data asset pricing, which is of value in the marketized allocation of enterprise data assets in the era of digital economy.
