Risk-return convergence in international public property markets

Abstract

The main contribution of this study is to assess the risk-return convergence, as well as its relationship with the realised correlation, relative to the global public real estate, of 12 international developed public property markets during 1990–2011. Based on the Euclidean distance method we find that average risk-return distance of the sample markets thus computed has increased over time, implying a mean-variance divergence, albeit statistically insignificant. Most of the markets are more ‘divergent’, as well as being more volatile during the Asian financial crisis and Global financial crisis periods. There is some evidence that risk-return convergence is positively linked to increasing correlation with the global developed public real estate over the full study period. Finally, exchange rate variable has relatively little effect on the variation of the three distance measures. We conclude that the risk and return characteristics of the developed public property markets have not become less different from each other over time, implying that the idiosyncratic ‘real estate factor’ and ‘country factor’ of individual markets might have become more important in affecting the market integration over time. This analysis and evidence contributes to our understanding of the dynamics of international developed public property market integration in global investing.

Keywords:

• risk-return distance
• risk-return convergence
• realised correlation
• global real estate
• public property markets
• market integration

Notes

1. In general term, convergence denotes an approach towards a definite value, a definite point, a common view or opinion, or towards a fixed or equilibrium state in the long-term or short-run.

2. Henceforth, this term ‘public property markets’ is used interchangeably with the term ‘real estate securities markets’.

3. Integration of financial market is broadly classified into domestic financial market integration (e.g. among money, credit, bond, stock and real estate markets), as well as international financial market integration (e.g. across national or regional stock markets, across national or regional bond markets, as well as across national or regional securitised real estate markets, etc.). Financial markets are integrated when the law of one price holds (Adam et al., 2002). This indicates convergence of returns on assets that are issued in different countries and generate identical cash flows (Baele, Ferrando, Hordahl, Krylova, & Monnet, 2004).

4. Accordingly to Dhar and Goetzmann (2006), due to the strong growth and remarkable risk-adjusted performance over time, the securitised real estate sector has now been recognised as an ‘essential’ asset class in mixed-asset portfolios, with industry sources predicting the global real estate securities market capitalisation to increase significantly from $500 billion in 2004 to 1 trillion by 2010 (Newell, Liow, Ooi, & Zhu, 2005). However, relatively less is known regarding the degree of inter-linkages among international real estate securities markets compared to those of world equity markets and bond markets (Yunus, 2009).

5. As will be explained in the methodology section, our risk-return convergence approach develops Euclidean distance (similar to cluster analysis) to measure the degree of risk-return differences among the markets. The computation of the Euclidean distance is model free as it does not assume any asset-pricing models used in computing the distance measure.

6. In the economic growth literature, σ-convergence happens when the cross-sectional standard deviation of per capita income among regions diminishes over time. In applying this concept to detect financial market integration, diminishing cross-sectional standard deviation for standard deviations of returns can be interpreted as evidence of risk-convergence or return-convergence. Solnik and Roulet (2000) use this concept of dispersion as cross-sectional correlation to estimate the global correlation level of stock markets.

7. This global developed real estate index has 25 constituents: US, Japan, UK, Singapore, Netherlands, Sweden, Germany, Finland, New Zealand, Norway, Portugal, Hong Kong, Australia, France, Canada, Switzerland, Belgium, Austria, Italy, Greece, Spain, South Korea, Denmark, Israel and Ireland.

8. The average convergence ratio for 17 developed stock markets in Eun and Lee’s (2010) study is worked out to be 4.64.

9. Figure 2(c) was included in response to this observation provided by a reviewer.

10. The local dollar results are also largely similar. Around the GFC period, all 12 real estate securities markets report an increase in RRD of between 4.9% (Singapore) and 160.8% (US). The average RRD increase for the six European markets around the GFC period is 72.9% (UK), 49.2% (France), 40.6% (Germany), 45% (Netherlands), 29.8% (Sweden) and 39.8% (Switzerland).

11. In addition to those reported observed cross-market convergence differences, it appears that France and Germany (two of the most developed European real estate securities markets)’s RRD divergence could be affected by different factors: for example, France’s RRD divergence results (yearly divergence rate of about .02%, second highest in the Euro zone – Table 6) might be caused by changes in the idiosyncratic risk between France and the ‘World’ real estate market. Alternatively, the results could also partially be caused by changes in the relationship between France and other European real estate securities markets. Additional analysis reveals that the UK, France and the Netherlands (which together accounts for over 75% of the European real estate securities markets) all displayed risk-return divergence of different degrees. For Germany, the difference in the RRD results could possibly be due to the consequence of the different types of securities/real estate companies traded: open-ended funds and closed-ended funds. We wish to thank one reviewer for raising specific points concerning the two markets.

12. Liow et al. (2009) reported there is an asymmetry in the correlation of international securitized real estate markets under different market conditions: the correlation is higher under ‘crisis’ (bearish) market conditions than under ‘non-crisis’ (or bullish) conditions.

13. For example, the UK average RRD correlation estimate is obtained by taking the average RRD correlation of (UK/France, UK/Germany … UK/Hong Kong and UK/Australia) and so on for 11 other market-average RRD correlation estimates. In the interest of space, the RRD correlation estimates for the individual pairs are not reported.

14. This is an alternative approach to the use of parametric models such as the GARCH or the multivariate GARCH models. See Andersen, Bollerslev, Diebold, and Labys (2003), Kim, and Doucouliagos (2005), Cappiello, Engle, and Sheppard (2006) and Beine and Candelon (2011).

15. The two models are: Model I: RRD = f (correlation, GFC, AFC); Model II: RRD = f (correlation, stock market risk, AFC, GFC).

16. We are grateful to one reviewer for raising this point.

17. We are grateful to one reviewer for this suggestion.