Several empirical studies have examined the volatility of either the components or the whole of real estate returns of different sectors, using data sets from a variety of sources. A number of studies looked in detail into the unconditional distributions of returns, such as Young et al (2006), other studies attempted to employ more sophisticated quantitative techniques attempting to support stochastic conditional distributions, such as Patel and Sing (2000), and Fourt et al (2006) looked into asset specific distributions using a large data set. This paper examines separately the stochastic distribution of real estate assets from the distribution of market aggregates, proposing a theoretical explanation for the link between the two. The conditional empirical distribution of each kind of capital returns is then derived by analysing the residuals of a simple first order autoregressive deterministic structure, which may or may not be complemented by higher moment conditionality GARCH family of models. The study examines capital returns of a range of real estate sectors, using both established market indices and transaction based indices that are developed for the needs of the study, and discusses the derived conditional distributions.