This work examines equity risk premium predictability in periods of market efficiency and market inefficiency. Efficiency is measured by the return's degree of multifractality, calculated from the multifractal detrended fluctuation analysis method. For the S&P 500 index during the 1951–2022 period, the results show that market efficiency varies over time, with recurrent periods of statistically significant inefficiency (multifractality). Moments of inefficiency are associated with (i) a higher level of financial uncertainty—financial uncertainty Granger causes the degree of multifractality (inefficiency), (ii) a greater variability in the patterns of dependence of returns and also (iii) with periods of more relevant volatility clusters. In times of market inefficiency (efficiency), the use of financial (technical) indicators shows statistically significant in‐sample and out‐of‐sample accuracy for equity risk premium prediction. When the market is efficient (inefficient), the use of financial (technical) indicators should be avoided due to the degradation of their predictive capacity. To build accurate and statistically significant predictions of the risk premium, thus enhancing decision‐making processes, investors should monitor the informational efficiency status of the market before selecting financial and technical indicators as predictive variables. Finally, during market inefficiency periods, there is a greater polarisation of investors' opinions, with increased attention to fundamental variables for risk premium prediction, leading to a breakdown in price trend patterns, explaining the worst predictive capacity of technical indicators.
