Asset Pricing Based On Investor Attention

By comparing information asymmetry theory and behavioral finance theory in the area of asset pricing, we find investor attention is overlooked by both theories. In real financial markets, however, investor attention plays vital role in information absorbing and price formation. Therefore, attention based asset pricing is worthy studying both on financial economic theoretical work and investment pratice.To test the general relationship between attention and expected return, we ultilize Internet media coverage as the stock attention proxy and examine the effect of stock attention on its trading and price behavior. We find evidence that high attention stocks are traded heavily, with greater trading volume and higher turnvoer than low attnetion stocks. Low attention stocks outperform high attention stocks in the subsequent months. If we construct a hedge strategy by buying low attenton stocks and selling high attention stocks, rebalancing monthly, we expect to capture1.98%monthly abnormal return, which is still significant after risky factor model adjustment.Sincer stock attention is autocorrelated on time series, we decompose stock attention into expected attention and unexpected attention. The former variable succeeded in predicting future stock return in negative direction while the latter could not. To the detail, we find the predictability of the attention proxy without decompostion is mainly originated from unexpected attention which is time-varing, other than expected attention defined similar to the attentive investor proportion in the model. This finding approximatly rejects the general relationship referred by our model. However, the quadratic linear relationship between expected attenton and expected return is verified by econometric models, which demostrating the U-shaped relationship between the two variables. Additionally, the sign of the relationship between expected attention and expected return is sensitive to the level of noisy trading, which is in line with our "category attention hypothesis". The fact that unexpected attention changes almost simulataneously with investor heterogeneous belief make us to control the investor disagreement effect, and then find the abnormal return caused by unexpected attention decrease sharply even to a insiginficant level, which supports" short selling constraints and investor disagreement hypothesis" by Miller (1977).