Campbell and Shiller analyze the relationship between stock prices, earnings, and expected dividends. Using U.S. stock market data from 1871–1986, they find that long-term earnings averages help predict the present value of future dividends. Their vector-autoregressive (VAR) model shows that stock returns can be forecasted using moving averages of earnings and current stock prices. The study highlights that stock returns and dividend-price ratios are too volatile to be explained by news about future dividends, and that this volatility is closely related to the predictability of long-term returns. They argue that excess volatility and predictability of long-term returns are not separate phenomena but are fundamentally connected. The study also shows that long-term returns are more predictable than short-term ones, and that dividend-price ratios and earnings-price ratios are strong predictors of stock returns. The results confirm that a long moving average of earnings can forecast future dividends, and that stock prices and returns are too volatile to fit a simple present-value model. The study concludes that earnings data, when averaged over many years, are a powerful predictor of stock returns, and that the volatility of stock prices is not consistent with the present-value model.Campbell and Shiller analyze the relationship between stock prices, earnings, and expected dividends. Using U.S. stock market data from 1871–1986, they find that long-term earnings averages help predict the present value of future dividends. Their vector-autoregressive (VAR) model shows that stock returns can be forecasted using moving averages of earnings and current stock prices. The study highlights that stock returns and dividend-price ratios are too volatile to be explained by news about future dividends, and that this volatility is closely related to the predictability of long-term returns. They argue that excess volatility and predictability of long-term returns are not separate phenomena but are fundamentally connected. The study also shows that long-term returns are more predictable than short-term ones, and that dividend-price ratios and earnings-price ratios are strong predictors of stock returns. The results confirm that a long moving average of earnings can forecast future dividends, and that stock prices and returns are too volatile to fit a simple present-value model. The study concludes that earnings data, when averaged over many years, are a powerful predictor of stock returns, and that the volatility of stock prices is not consistent with the present-value model.