This research focuses on macroeconomic risk factors pertaining to the various types of real estate exposure (direct, listed, and non-listed) and how the resulting return predictability can be used in a mixed-asset portfolio framework. Comparing sensitivities to risk factors is important to assess whether indirect (listed and non-listed) exposures react in the same way as direct investments to the macroeconomy and how well such investments replicate direct real estate behavior. The various types of real estate exposure generally respond similarly to risk factors: GDP, money supply, construction costs, expected inflation, and expected economic activity positively impact returns, while the term and credit spreads, unemployment, and unexpected inflation negatively affect returns. We then rely on the identified risk factors to predict the expected returns and volatility of real estate, stocks, and bonds. These forecasts are used to build mixed-asset portfolios for various investment horizons. The benefits of including real estate in a portfolio and the possible substitutability between the three types of exposure are analyzed. The empirical analyses are conducted using U.S. data spanning over 30 years.