Real estate portfolio diversification takes many forms, most of which can be associated with size (value). Larger portfolios are assumed to have greater diversification potential than small portfolios. In addition, since greater diversification is generally associated with lower risk it is assumed that larger portfolios will also have reduced return variability compared to smaller portfolios. If large real estate portfolios can simply be regarded as scaled-up, better-diversified, versions of small real estate portfolios, then the greater a portfolio's size, the lower the risk. This suggests a negative relationship between size and risk. If however large real estate portfolios are not just scaled-up versions of small portfolios, these relationships may not hold. This paper explores the empirical relationship between real estate portfolio size, diversification and risk using the returns from 136 UK real estate portfolios over the period 1989 to 1999. Using a conceptually sound measure of overall portfolio diversification (R²) it is shown that a significantly positive correlation between size and diversification does not necessarily translate into the expected negative correlation between size and risk. The analysis shows that increasing portfolio size may lead to a larger reduction in specific risk than previous studies have identified, with the proviso that this increase is not accompanied by other risk- enhancing activities. In the context of portfolio management this leads to a view that performance evaluation and benchmarking should seek to control for the style and specialisation of the fund (managers) because it seems clear that systematic risk does vary to an appreciable extent because of these features.