This paper examines the role of detrended wealth in predicting stock returns. Using U.S. quarterly data, the authors find that trend deviations in wealth are strong predictors of both real stock returns and excess returns over a Treasury bill rate. These deviations outperform other popular forecasting variables like the dividend yield, earnings yield, and dividend payout ratio, especially at short and intermediate horizons. The authors argue that detrended wealth, defined as deviations from its shared trend with consumption and labor income, forecasts asset returns because it captures agents' expectations of future returns on the market portfolio. This is supported by a wide class of optimal consumer behavior models, where the log consumption-aggregate wealth ratio predicts expected returns on aggregate wealth. The paper introduces a framework where consumption, labor income, and nonhuman wealth share a common stochastic trend (they are cointegrated). Deviations from this trend summarize agents' expectations of future returns on the market portfolio. The authors show that these deviations can be expressed in terms of observable variables: consumption, nonhuman wealth, and labor income. The study uses a dynamic least squares (DLS) method to estimate the cointegrating parameters of consumption, labor income, and nonhuman wealth. The results show that the estimated trend deviation, denoted as $ \widehat{cay}_{t} $, is a strong predictor of future asset returns. The paper also compares $ \widehat{cay}_{t} $ to other forecasting variables like the log dividend-price ratio and finds that $ \widehat{cay}_{t} $ provides independent information for forecasting returns at different horizons. The authors find that $ \widehat{cay}_{t} $ has a significant predictive power for stock returns, especially at short and intermediate horizons. The variable is also less persistent than other popular forecasting variables, which makes it a better predictor of future returns. The paper concludes that detrended wealth, measured as deviations from its shared trend with consumption and labor income, is a strong and reliable predictor of future stock returns.This paper examines the role of detrended wealth in predicting stock returns. Using U.S. quarterly data, the authors find that trend deviations in wealth are strong predictors of both real stock returns and excess returns over a Treasury bill rate. These deviations outperform other popular forecasting variables like the dividend yield, earnings yield, and dividend payout ratio, especially at short and intermediate horizons. The authors argue that detrended wealth, defined as deviations from its shared trend with consumption and labor income, forecasts asset returns because it captures agents' expectations of future returns on the market portfolio. This is supported by a wide class of optimal consumer behavior models, where the log consumption-aggregate wealth ratio predicts expected returns on aggregate wealth. The paper introduces a framework where consumption, labor income, and nonhuman wealth share a common stochastic trend (they are cointegrated). Deviations from this trend summarize agents' expectations of future returns on the market portfolio. The authors show that these deviations can be expressed in terms of observable variables: consumption, nonhuman wealth, and labor income. The study uses a dynamic least squares (DLS) method to estimate the cointegrating parameters of consumption, labor income, and nonhuman wealth. The results show that the estimated trend deviation, denoted as $ \widehat{cay}_{t} $, is a strong predictor of future asset returns. The paper also compares $ \widehat{cay}_{t} $ to other forecasting variables like the log dividend-price ratio and finds that $ \widehat{cay}_{t} $ provides independent information for forecasting returns at different horizons. The authors find that $ \widehat{cay}_{t} $ has a significant predictive power for stock returns, especially at short and intermediate horizons. The variable is also less persistent than other popular forecasting variables, which makes it a better predictor of future returns. The paper concludes that detrended wealth, measured as deviations from its shared trend with consumption and labor income, is a strong and reliable predictor of future stock returns.