Funded by Scottish Widows Investment Partnership (SWIP)
Real Estate investment strategies of financial institutions have a major influence on pension funds and regional investment / growth. Currently most property investment strategies are London centric and follow the IPD (Investment Property Databank) benchmark. Our research for SWIP has caused them to begin to change the geographical weighting in real estate investment portfolios across the UK. The mathematical models we developed have been adopted by this major financial institution with a significant real estate portfolio, pursuing more regional investment. The consequences of portfolio rebalancing are encouraging greater investment allocation across the regions of the country.
Addressing the Challenge
This specific project built upon earlier work conducted in conjunction with colleagues at Heriot-Watt University for SWIP that identified opportunities for regional investment diversification. Developing this, and employing more advanced econometric techniques for more in-depth analysis, this study focused specifically on:
- timing of purchase and sale of real estate assets
- identifying risk and return combinations of real estate investment markets across regional cities
- evaluating whether there is too much or too little investment in particular markets
- conditional volatility to identify real risk levels.
Professor Michael White is the director of the Real Estate Economics and Investment Research Group, he also provides academic leadership in mentoring new researchers and PhD students.
Michael's research interests are in the field of real estate market analysis. Within this area he has undertaken research in regional, national, and international commercial real estate markets. Michael's research has also analysed residential markets in the UK and in EU countries. This body of work has developed theoretical and policy issues in examining market behaviour. The theoretical issues have related to market adjustment processes over time, spatial interactions across locations, and the separation of long and short run influences on market behaviour.
Making a Difference
In order to do this a number of specific research tasks were completed:
- construction and estimation of a market model that identified long term trend values and, furthermore, predicted market cyclical turning points
- development and application of a modern portfolio approach to examine risk and expected returns across different city markets
- examination of how relative city positions from (2) above can change over a cycle and factors affecting trend positions
- identifying if investment was above or below the level it should have been.
- identification of ARCH processes and subsequent extension of analysis to test for ARCH-M and TARCH/T-GARCH processes.
Investment in regional real estate markets is linked to macroeconomic and financial factors in task 1. This identifies the key market drivers important for real estate asset values and explicitly considers the econometric issues that arise in identification of long run trend paths and short run adjustment speeds. Expected returns based upon task 1 are used in the model in task 2 together with the unconditional risk measure to compare real estate markets across different cities.
This is examined over an economic cycle in task 3. We can then identify if cities have too much or too little investment in task 4 based upon the criteria in task 2. In task 5 we account for the possibility that the standard risk measure underestimates the true risk attached to investment since risks may be temporally correlated and estimate autoregressive conditionally heteroscedastic models. Combining these steps is relatively new in regional real estate market analysis. However all steps above link to methodological developments in the analysis of real estate, particularly linking theoretical market models to estimated empirical models.
SWIP have indicated that they have begun to apply it in their analysis of regional real estate markets. This pension fund holds one of the largest portfolios of real estate assets in the UK, worth approximately £10 billion at current market values. Their regional portfolio allocation had previously closely followed the IPD benchmark. Based upon the research undertaken here, a re-evaluation of regional investment is taking place within the company. The investment returns made will feed into the size of the firm’s pension funds and hence have a direct impact on the value of pensions paid from the fund.