Diversification of portfolio risk: reconciling theory and observed weightings

During the 1990s and early 2000s a body of literature emerged that explores and seeks to improve property portfolio risk diversification strategies. The studies were motivated by the view that strategies commonly used are suboptimum because of the reliance, due to data availability, on political administrative regions to represent asset classes. Asset classes should comprise homogeneous assets, with comparable (if not identical) risk and return profiles. This enables investors to follow the diversification principles set out by Markowitz (1952). However, political administrative regions are not based on property market fundamentals, nor, indeed, economic fundamentals in most cases, but on historic and governmental factors. Individual regions may contain, for example, both urban and rural areas; prospering and declining areas; and manufacturing and service employment centres. Looking to the fundamentals of property market performance, these contrasting characteristics underpin differential investment outcomes. These studies have consistently challenged the use of administrative regions by investors for the diversification of risk, finding local area characteristics and property market fundamentals as important in portfolio strategy. The focus of the paper is to re-evaluate these studies, now that sufficient time has passed for data to become available to assess what would have happened if investors had followed the diversification strategies suggested. A sample of 73 local markets is analysed using total returns data from IPDís UK Key Centres 2008 report, reflecting institutional property investment trends and market performance over the ten year period 1998-2007. Efficient frontiers are developed that enable evaluation of the opportunities presented by the classifications of local markets and, in addition, sectoral and regional market segregation. The optimal weightings suggested by the classifications are compared to actual aggregate institutional investment weightings. The differences in allocation are explored. The analysis is then extended to cover different phases of market performance.