Mobile Edition
Print
Get the Mag
Weekly UpdatesThe correlation matrix for international property returns is illustrated in Table 4. All negative correlation coefficients have been bolded. The Finnish property market is positively correlated with all other property markets, except the German market. With eight out of thirteen countries (Canada, Denmark, France, Ireland, the Netherlands, Norway, Spain and Sweden) correlations were 0,8 or more suggesting no diversification benefits, and of the positive correlation coefficients only one, with Portugal, was low. Thus, it can be stated that the Finnish market did not provide significant diversification benefits under the studied period.
Altogether, the returns in the studied markets were strongly correlated with each other, the only negative correlations being with Germany or Portugal. Also Switzerland had some moderate correlations, for example with U.K. and U.S.
Table 5 contains the country allocations in efficient portfolios. One should pay attention to the risk-return characteristics of the efficient portfolios, especially at the return levels of 4 to 6%, where portfolio risk approaches zero. The efficient portfolios are dominated by German properties at the low to moderate risk levels and Portuguese properties in the moderate to high risk levels. The high negative correlations of these markets lead to significant diversification benefits in the portfolio.
The allocation to U.K. properties grows with the portfolio return, which is explained by the fact that the U.K. market provided high absolute returns with moderate risk. U.S. property enters the efficient portfolios at the lowest risk levels and Switzerland makes a short appearance in the return level of 8%. It is also interesting to notice that of the fourteen countries, nine (Canada, Denmark, Finland, France, Ireland, Netherlands, Norway, Spain and Sweden) do not enter the efficient sets at all.
Most optimal allocations have extreme weights on few countries (Germany, Portugal and U.K.). The extreme weights are, however, not pure corner solutions, but are largely affected by the correlation structure of the asset returns. In practice such allocations might be unrealistic, and thus the optimisation was conducted using a 20% maximum allocation constraint. The results of the optimisation are illustrated in Table 6.
The use of constraints increases the number of countries in portfolios. Each optimal portfolio must consist of at least five countries, but in practise all optimal portfolios have an allocation to six countries. For Portugal and U.K., i.e. countries that had high allocations throughout the return spectrum in unconstrained portfolios, the constraint just limits the allocation to the highest level permitted. For Germany the case is different. The new countries in the portfolios change the dynamics of the portfolio, thus reducing the optimal allocations to German properties below the constraint level on return level of 20%. When applying the allocation constraint also the Finnish real estate enters the optimal portfolio at the lowest return levels.
Unhedged returns
To see the effect of currency risk on the optimal portfolios, the analysis of international real estate portfolios was conducted using returns converted into euros. Seven of the studied countries (Finland, France, Germany, Ireland, Netherlands, Portugal, Spain) are part of the eurozone, so the adjustment for currency risk only affects the returns of the remaining seven countries (Canada, Denmark, Norway, Sweden, Switzerland, U.K. and U.S.). The summary statistics of the adjusted property data is illustrated in Table 7.
The changes in the risk-return profiles are marked, the size of effect, however, varies across countries. The effect is the largest for Switzerland and U.S., where the coefficient of variation multiplies to almost fourteen- and sevenfold, respectively. For Canada, Norway and U.K. the coefficient of variation approximately doubles. The only country where the effect is remains unsubstantial is Sweden. After the adjustments, Portugal and Netherlands provide the best risk-return trade-off in the sample.
The correlation matrix for unhedged returns is provided in Table 8. Again, negative correlations are emphasised. The correlation coefficients for Germany and Portugal remain mostly negative, and the currency adjustment also turns some of the correlations for Switzerland negative. The effect of the currency adjustment on correlation coefficients varies across countries. Surprisingly for Canada, Norway, Switzerland and U.S. the correlation coefficients predominately increase, whereas for Denmark and U.K. the correlation coefficients mostly decrease.
The structure of efficient portfolios is illustrated in Table 9. The portfolios are again dominated by German and Portuguese property; whereas the U.K. property enters optimal portfolios only in the lower return levels. As for hedged returns, Denmark, Finland, France, Netherlands, Norway, Spain, Sweden and U.S. do not enter the efficient sets. Switzerland, which had a small allocation in the optimal hedged portfolios at the return level of 8%, does not enter the portfolios, whereas Canada and Ireland do.
The allocations in the constrained portfolios are illustrated in Table 10. Again, the changes are not restricted to reducing allocations of some countries to the upper boundary (as for Portugal), adding new countries to the portfolio (e.g. the Netherlands, Finland) and leaving some out (e.g. U.K.), but the use of constraints changes the dynamics of the portfolio. In the constrained portfolio, the allocation to German property reduces to the set maximum on the risk levels of 6 and 8%, but is clearly below the set maximum on the higher risk levels. For Portugal, the optimal allocations in the return levels of and above 8% are on upper boundary of the constraint, i.e. 20%, whereas for the lowest return level, the allocation is markedly lower.
Dutch property, which did not enter the unconstrained portfolios, has an allocation of 20% on all risk levels. The case is similar for Finnish real estate, which has an optimal allocation of 20% on all other but the highest return levels. Other new countries in the constrained portfolios are Spain and France, which both have high allocations on the upper end of the return spectrum, and Sweden and Switzerland, which enter the portfolios at lower return levels. The countries that have moderate allocations in unconstrained portfolios, i.e. Canada and U.K, have decreased allocations in the constrained portfolios, whereas U.S. property enters the 4% return portfolio.
5. CONCLUSIONS
The purpose of this paper was to analyse, if the Finnish property market provided diversification benefits in its early years of internationalisation. In the Finnish mixed-asset context, both direct and indirect real estate investments had a large allocation in the set of efficient portfolios.
In the international real estate portfolio, the results varied depending on the constraints used. When there were no maximum allocation constraints, Finnish property did not enter the efficient portfolios regardless of the hedging strategy of the investor. When a maximum allocation constraint of 20% was set, the effects of hedging strategy started to play a more significant role in the results. For completely hedged returns, Finnish real estate offered diversification benefits at the 6% return level, whereas for unhedged returns, the allocation remained at the upper boundary of the allocation constraint until the return level of 10%. Thus the results indicate that international investors were able to obtain diversification benefits in the Finnish property market during its early years of internationalisation.
ACKNOWLEDGEMENT
The author wishes to acknowledge Investment Property Databank and Institute for Real Estate Economics (KTI) for the data they pro vided, as well as Academy of Finland (project 122525) and Tekes--the National Technology Agency of Finland (project 40079/06) for the financial support. The author is also grateful to the editor of this journal and the two anonymous referees whose comments considerably improved the quality of this article.
REFERENCES
Addae-Dapaah, K. and Yong, G.L. (1998) Currency risk and office investment in Asia Pacific, Real Estate Finance, 15(3), pp. 67-85.
Addae-Dapaah, K. and Yong, C.C. (2000) Diversification of real estate investment in the Asia-Pacific region, Pacific Rim Property Research Journal, 6(2), pp. 31-45.
Black, F. and Litterman, R. (1992) Global portfolio optimization, Financial Analysts Journal, 48(5), pp. 28-43.
Diaz, J. III and Wolverton, M.L. (1998) A longitudinal examination of the appraisal smoothing hypothesis, Real Estate Economics, 26(2), pp. 349-358.
Case, B., Goetzmann, W. and Rouwenhorst, K.G. (1999) Global real estate markets, cycles and fundamentals. Working Paper No. 99-03, Yale International Center for Finance.
Catella (2005) Market Review, Catella Property Oy., autumn 2005.
Catella (2008a) Market Review, Catella Property Oy., spring 2008.
Catella (2008b) Market Review, Catella Property Oy., autumn 2008.
Geltner, D. (1989) Estimating real estate's systematic risk from aggregate level appraisal-based returns, Journal of the American Real Estate and Urban Economics Association, 17(4), pp. 463-481.
Geltner, D. (1993) Estimating market values from appraised values without assuming an efficient market, Journal of Real Estate Research, 8(3), pp. 325-345.
Gordon, J. (1991) The diversification potential of international property investments, Real Estate Finance Journal, 7(2), pp. 42-48.
Hoesli, M., Lekander, J. and Witkiewicz, W. (2004) International evidence on real estate as a portfolio diversifier, Journal of Real Estate Research, 26(2), pp. 161-206.
FEED