We use a variance decomposition approach to examine why aggregate valuation ratios differ across countries. In a cross section of 22 developed countries from 1980 to 2009, we find that 50 percent of all cross‐country differences in the aggregate price‐to‐book ratio (P/B) can be explained by cross‐country differences in expected future five‐year profitability. In the second half of our sample period, this percentage exceeds that of the first half, rising to almost 64 percent. Although international differences in accounting standards and conventions may have made earnings from different countries more difficult to compare relative to dividends, we find that it is still cross‐country differences in expected future profitability, rather than dividend growth rates, that are more closely related to international differences in valuation ratios. Even among 25 emerging markets, we find that expected future profitability at the five‐year horizon can account for 29 percent of all cross‐country P/B variations. Our results show that international investors are able to identify substantial cross‐country differences in future‐earnings prospects and incorporate them into stock market valuations.

We find that aggregate profitability and asset investment exhibit robust joint predictive power for aggregate excess stock returns, consistent with the investment model of Hou, Xue, and Zhang (henceforth HXZ, 2015). These results provide out-of-sample empirical support for HXZ, as the same mechanisms that HXZ use to explain firm-specific variation in stock returns can also be used to explain variation that is market-wide in nature. Also consistent with the HXZ investment model, we find that the growth rate of short-term (long-term) assets exhibits a stronger predictive power for one-year-ahead (two-year-ahead) stock returns.

We show that the returns of individual stocks become more synchronous with the aggregate market during periods of high investor sentiment. We also document that the effect of sentiment on stock return synchronicity is especially pronounced for small, young, volatile, non-dividend-paying and low-priced stocks. This ‘difference in difference’ suggests that stocks with these characteristics are affected more by sentiment—consistent with previous studies. Our results support the hypothesis that greater constraints on arbitrage and the prevalence of sentiment-driven demand during periods of high sentiment lead to increased comovement among stocks.