International Diversification Strategies for Direct Real Estate

Abstract

This paper will disentangle the performance of international real estate into property type performance and region selection. This helps to create an international diversification strategy for direct real estate. We use constrained cross-section regression with dummy variables for regions and property types to measure the best risk reducer. We analyze the impact of currency changes on total returns by looking at a hedged and un-hedged portfolio, both stock and equally weighted. The findings show that geographic factors have the largest influence on the volatility of international real estate returns. The average variance of the regional effects is higher than the property type effects and therefore the regional effects have a higher influence on the variation of the total portfolio. However, the regional effects are less stable through time, compared with the variance and correlation of the property type effects. Also the property type effect seems to become a more important factor for the return over time, especially when the return is expressed in local currency.

Appendices

Appendix A. Average total return analysis by city and property type.

Region/Country	City/MSA	Industrial			Office			Residential			Retail
			σ			σ			σ			σ
		Total	Total		Total	Total		Total	Total		Total	Total
		Return	Return		Return	Return		Return	Return		Return	Return
Asia
China	Beijing				22.1%	20.1%	[48]	18.0%	17.9%	[56]	-1.2%	14.9%	[32]
	Shanghai				13.3%	19.9%	[41]	19.8%	19.9%	[40]	15.4%	17.0%	[40]
	Hong Kong				9.9%	19.9%	[64]	17.6%	19.6%	[64]	1.6%	13.9%	[24]
Indonesia	Jakarta				24.3%	27.3%	[64]	25.1%	27.2%	[28]	21.6%	29.4%	[44]
Japan	Tokyo				-6.8%	7.7%	[64]
Malaysia	Kuala Lumpur				7.1%	4.1%	[48]	11.7%	7.1%	[48]	4.0%	11.6%	[24]
Philippines	Manila				3.3%	5.7%	[31]	0.4%	7.9%	[32]	4.5%	5.7%	[24]
Singapore	Singapore				1.0%	12.8%	[49]	10.2%	10.2%	[60]	1.6%	13.5%	[49]
Thailand	Bangkok				6.2%	10.4%	[40]	6.5%	5.5%	[44]	8.2%	13.7%	[36]
Australia
Australia	Brisbane	12.3%	2.9%	[40]	6.5%	3.6%	[64]				10.7%	2.4%	[40]
	Canberra				8.0%	5.2%	[64]
	Melbourne	12.1%	2.4%	[35]	5.2%	6.1%	[64]				11.0%	3.4%	[64]
	Perth				5.3%	7.3%	[64]				13.0%	4.0%	[64]
	Sydney	10.5%	4.6%	[64]	5.9%	6.8%	[64]				12.0%	3.3%	[64]
Continental Europe
Belgium	Antwerp	12.1%	9.6%	[58]	10.7%	6.5%	[58]				11.2%	8.0%	[64]
	Brussels	11.5%	10.5%	[58]	10.4%	8.2%	[64]				11.7%	8.3%	[64]
Czech Republic	Prague	15.0%	6.5%	[28]	8.4%	9.4%	[46]				22.5%	12.9%	[16]
Denmark	Copenhagen	12.3%	6.6%	[12]	9.6%	3.7%	[16]				6.6%	8.1%	[16]
Finland	Helsinki	12.1%	9.7%	[16]	7.3%	7.4%	[16]				7.2%	6.8%	[16]
France	Lyon	16.0%	8.8%	[36]	10.3%	8.8%	[52]				30.7%	15.5%	[38]
	Paris	8.7%	10.5%	[58]	3.9%	10.0%	[54]				20.2%	13.1%	[38]
Germany	Berlin	5.6%	10.1%	[29]	2.0%	13.4%	[53]				7.2%	9.5%	[64]
	Frankfurt	7.5%	7.0%	[43]	7.5%	10.6%	[64]				7.3%	10.0%	[64]
	Munich	9.3%	6.7%	[16]	7.2%	9.6%	[64]				8.1%	10.5%	[64]
Hungary	Budapest	11.1%	8.9%	[26]	9.9%	8.5%	[46]				15.9%	6.5%	[ 8]
Ireland	Dublin	14.1%	10.2%	[56]	11.7%	12.0%	[57]				19.0%	12.4%	[16]
Italy	Milan	14.5%	10.6%	[36]	7.5%	11.4%	[52]				14.1%	9.2%	[32]
Luxembourg	Luxembourg	13.9%	7.2%	[50]	6.9%	8.5%	[52]				11.7%	6.9%	[16]
Netherlands	Amsterdam	13.5%	9.6%	[58]	11.8%	9.5%	[64]				13.0%	7.5%	[64]
	The Hague	10.7%	10.0%	[58]	9.6%	6.1%	[64]				9.8%	8.1%	[58]
Norway	Oslo	12.6%	10.3%	[13]	4.8%	17.5%	[16]				6.4%	6.2%	[16]
Poland	Warsaw	13.5%	13.4%	[31]	7.3%	9.5%	[31]				19.5%	10.6%	[16]
Spain	Barcelona	15.0%	14.3%	[58]	8.3%	14.0%	[56]				14.8%	16.2%	[57]
	Madrid	12.4%	10.1%	[58]	10.7%	18.1%	[64]				17.6%	14.9%	[64]
Sweden	Stockholm	12.8%	14.5%	[36]	14.8%	13.1%	[40]				-0.7%	8.9%	[15]
United Kingdom	Birmingham	10.2%	4.3%	[30]	10.4%	5.6%	[64]				12.1%	4.6%	[41]
	Bristol	12.4%	6.2%	[64]	11.4%	8.6%	[64]				9.5%	4.8%	[64]
	Edinburgh				10.0%	9.7%	[64]				9.6%	6.0%	[64]
	Leeds	14.7%	8.5%	[64]	11.1%	7.1%	[64]				11.9%	5.6%	[44]
	London	12.4%	5.5%	[64]	7.9%	6.6%	[64]				9.7%	4.4%	[64]
	Manchester	10.8%	3.3%	[37]	13.6%	7.7%	[64]				11.8%	5.1%	[44]
	Reading	7.9%	4.5%	[41]	6.3%	6.0%	[64]				11.2%	4.9%	[44]
United States
United States	Atlanta	6.5%	4.1%	[64]	4.2%	5.7%	[64]	7.7%	4.5%	[64]	8.1%	3.7%	[64]
	Austin							8.0%	3.3%	[35]
	Baltimore	9.7%	3.5%	[64]							5.7%	4.7%	[56]
	Boston	8.3%	6.0%	[64]	5.9%	8.6%	[64]	11.8%	5.7%	[53]	9.9%	3.6%	[38]
	Chicago	7.7%	3.3%	[64]	4.6%	6.5%	[64]	11.3%	3.2%	[35]	7.3%	3.5%	[64]
	Dallas	6.1%	3.9%	[64]	5.3%	6.2%	[64]	6.6%	4.2%	[64]	6.3%	4.7%	[64]
	Denver				4.2%	6.7%	[64]	13.0%	3.7%	[46]	6.0%	5.2%	[64]
	Fort Lauderdale							8.0%	3.3%	[59]
	Houston				4.1%	6.8%	[64]	6.9%	5.0%	[64]	4.5%	8.1%	[64]
	Indianapolis	6.7%	3.1%	[53]
	Las Vegas							7.6%	2.8%	[59]
	Los Angeles	8.7%	5.0%	[64]	4.3%	6.4%	[64]	13.3%	4.4%	[23]	9.7%	5.8%	[64]
	Memphis	7.1%	4.5%	[64]
	Miami				10.5%	4.1%	[33]				5.5%	3.8%	[21]
	Middlesex	6.2%	4.5%	[52]	8.7%	2.3%	[20]
	Minneapolis	7.3%	4.1%	[64]	1.4%	6.3%	[64]				9.0%	2.7%	[33]
	New Haven				9.0%	16.5%	[54]
	New York				6.4%	7.1%	[64]	8.0%	3.6%	[15]
	Oakland	8.4%	6.0%	[64]	6.9%	6.8%	[64]				7.6%	3.6%	[64]
	Orange County	8.6%	5.5%	[64]	6.8%	8.9%	[64]				6.4%	5.3%	[59]
	Orlando	7.4%	6.4%	[64]				7.6%	3.7%	[58]
	Philadelphia							11.0%	7.5%	[60]	7.5%	5.3%	[44]
	Phoenix	6.3%	5.6%	[64]	8.8%	4.3%	[29]	9.0%	3.7%	[61]	8.4%	7.0%	[64]
	Portland	12.4%	3.5%	[34]
	Riverside	8.7%	5.1%	[63]				17.0%	3.6%	[20]
	San Antonio										11.3%	5.2%	[15]
	San Diego	8.5%	4.9%	[64]	6.4%	8.3%	[64]	15.0%	3.4%	[19]	6.8%	5.6%	[64]
	San Francisco				5.7%	8.2%	[64]	5.0%	1.5%	[ 6]	8.4%	5.4%	[54]
	San Jose	9.6%	8.8%	[64]	8.6%	11.9%	[55]				8.6%	6.5%	[64]
	Seattle	8.9%	3.2%	[64]	8.1%	8.1%	[56]	9.5%	4.8%	[61]	9.4%	3.7%	[57]
	Washington	8.0%	5.4%	[64]	7.1%	3.6%	[64]	10.1%	3.7%	[64]	7.1%	5.2%	[64]
	West Palm Beach							8.7%	3.3%	[50]

This appendix shows the average annual total return in local currency and standard deviation for every city/MSA included in our database. If there is a value, it means that there is a time-series available, no value means no time-series included in the database. The number of available quarters is given in brackets.

Appendix B. Relative weight of regions and property type

This figure shows the relative weight of regions in the dataset between 1Q1988-4Q2003. The weight depends on the number of cities and the market value in US dollars.

This figure shows the relative weight of property types in the dataset between 1Q1988-4Q2003. The weight depends on the number of cities and the market value in US dollars.