Abstract
A stylized fact in the real estate diversification literature is that sector (property type) effects are relatively more important than regional (geographical) factors in determining property returns. Thus, for those portfolio managers who follow a top‐down approach to portfolio management, they should first choose which sectors to invest in and then select the best properties in each market. However, the question arises as to whether the dominance of sector effects over regional effects is constant. If not, property fund managers will need to take account of regional effects in developing their portfolio strategy.
Using monthly returns data for individual properties over the period 1987:1 to 2002:12, this paper investigates the influence of sector and regional factors on commercial real estate performance, first by adopting the dummy variable approach of Heston and Rouwenhorst (Citation1994, Citation1995) and then by using the notion of cross‐sectional dispersion introduced by Solnik and Roulet (Citation2000). The results show that sector‐specific dominate region‐specific factors for the majority of the time and, in particular, during volatile periods of the real estate cycle. However, during calmer periods, sector and regional effects appear to be of equal importance. Overall, sector effects are still the most important aspect in the development of an active portfolio strategy.
Acknowledgements
The authors acknowledge the contribution of Mark Andrew of the University of Reading, whose programming work for an earlier paper assisted some of the analyses presented here. We also thank the anonymous referees for their helpful comments.
Notes
1. Hess and Liang (Citation2000) have used a similar approach to measure changes in regional market selection in the investment process in the US real estate market. The authors arguing that when the regional dispersion index is high it indicates that portfolio managers have greater opportunity to select investments that display superior performance than they do when regional dispersion is low, although the potential for selecting inferior regions also increases with increases in the dispersion index. In other words, the regional dispersion index indicates the importance of regional selection strategies in the investment process.
2. This increase was largely due to more funds joining the index rather than rapid growth of the founder funds.
3. This is partly due to portfolio structure as well as fewer properties, with the constituent funds being only lightly represented during the earlier years in sub‐segments such as Shopping Centres and City of London Offices. Nonetheless, some alternative segmentations were tested using the dispersion methodology, with little modification to the overall findings (results available from the authors on request).
4. See Nelson and Plosser (Citation1982) for a more detailed discussion of the HP‐filter, as well as alternative procedures.