Relatively little is known in the academic literature about the idiosyncratic returns of individual real estate investments, though quite a few commercial properties command prices commensu- rate with the market values of small publicly traded companies. I use purchase and sale data from the National Council of Real Estate Investment Fiduciaries (NCREIF) to compute holding period price-appreciation returns for commercial properties. In stark contrast with liquid asset returns, idiosyncratic drift and volatility estimates diverge as the holding period shrinks. This puzzling phenomenon survives a variety of controls for vintage effects, systematic risk hetero- geneity, and sample selection biases. I derive an equilibrium search-based illiquid asset pricing model which, when calibrated, fits the data very well. Thus a structural model of illiquidity seems crucial to a descriptive theory of real estate investment returns. These insights can be extended to other illiquid asset classes such as private equity, mergers and acquisitions, large whole loans, and other real assets. The model can also be used to price derivatives such as debt claims.