This paper examines the predictive performance of asset prices for inflation and real output growth. It reviews the existing literature on this topic and conducts an empirical analysis using quarterly data on up to 38 candidate indicators (primarily asset prices) for seven OECD countries over a 41-year period (1959–1999). The findings suggest that some asset prices have statistically significant predictive content for output growth and inflation in certain countries and periods. However, the predictive performance of these indicators is unstable over time and across different periods. For example, an indicator that performs well in one period may not perform well in another. This instability is a common feature of asset price-based predictive relations. Forecasts based on individual indicators are often unreliable, as their predictive power varies over time. However, combining forecasts from different indicators appears to improve accuracy. For instance, the median of forecasts based on individual asset prices consistently outperforms autoregressive benchmarks. Similarly, forecasts of inflation that combine information from real activity measures and output gaps are more reliable and stable. The paper also discusses the limitations of using in-sample significance tests, such as Granger causality tests, to identify predictive relations. These tests do not guarantee that the identified relations are stable over time. Moreover, the predictive content of asset prices for inflation is generally weaker than for output growth. The study highlights the importance of considering the stability of predictive relations when using asset prices for forecasting. While some asset prices have predictive content, their reliability is often limited. Combining multiple indicators appears to be a more robust approach to forecasting economic variables. The paper concludes that the predictive power of asset prices for economic variables is context-dependent and subject to change over time.This paper examines the predictive performance of asset prices for inflation and real output growth. It reviews the existing literature on this topic and conducts an empirical analysis using quarterly data on up to 38 candidate indicators (primarily asset prices) for seven OECD countries over a 41-year period (1959–1999). The findings suggest that some asset prices have statistically significant predictive content for output growth and inflation in certain countries and periods. However, the predictive performance of these indicators is unstable over time and across different periods. For example, an indicator that performs well in one period may not perform well in another. This instability is a common feature of asset price-based predictive relations. Forecasts based on individual indicators are often unreliable, as their predictive power varies over time. However, combining forecasts from different indicators appears to improve accuracy. For instance, the median of forecasts based on individual asset prices consistently outperforms autoregressive benchmarks. Similarly, forecasts of inflation that combine information from real activity measures and output gaps are more reliable and stable. The paper also discusses the limitations of using in-sample significance tests, such as Granger causality tests, to identify predictive relations. These tests do not guarantee that the identified relations are stable over time. Moreover, the predictive content of asset prices for inflation is generally weaker than for output growth. The study highlights the importance of considering the stability of predictive relations when using asset prices for forecasting. While some asset prices have predictive content, their reliability is often limited. Combining multiple indicators appears to be a more robust approach to forecasting economic variables. The paper concludes that the predictive power of asset prices for economic variables is context-dependent and subject to change over time.