Managing Specific Risk In Property Portfolios

Specific risk in property, whether measured as standard deviation or as a tracking error against a benchmark, is a key problem for international investors. It arises through large average property capital values (lot sizes), through an uneven distribution of these values and through the inability of investors to match competitorsí portfolios, as each property is unique. Property funds offer a way to limit this problem, as all three issues are minimised by investing indirectly by using diversified funds. But this paper shows that specific risk varies significantly between sectors and countries, and unlisted funds may be more useful in some sectors and countries than others. This is not simply a function of lot size but also of ëdiversification powerí within sectors, defined as the efficiency of specific risk reduction through adding properties. Data from the UK is used to illustrate this issue and to show where the use of funds would be most useful. It also examines whether the funds currently available are large enough to offer the necessary risk reduction. In the paper we use market data to examine the extent of the problem of specific risk for property investors, and we couple this issue with the proposition that the growth in the unlisted fund market is explainable by reference to this problem. The propositions being tested are as follows: * an allocation to direct property will carry high levels of specific risk; this risk varies significantly between the segments, and it is harder to achieve efficient diversification in some segments; * this problem coupled with the use of benchmark-driven portfolio structures makes it exceedingly hard to control diversification by sector and well as within sectors without huge sums to invest; * there is a strong case for using indirect property vehicles to combat specific risk at the sector level. // To test these propositions, we examined how much cash would have been be needed over a specific period to limit the tracking error of a group of properties within one segment to a given risk target. We then compared the results across segments; and we draw some conclusions about those segments where diversification is relatively easy and those where it is relatively hard. We use a unique dataset describing UK property vehicles in combination with new IPD data measuring risk at the individual property level and together this data creates an original contribution in an under-researched area.