تنوع بخشی بین المللی با دارایی واقعی امن و تابش خیره کننده حجاب از ریسک ارزی

This paper analyzes diversification benefits from international securitized real estate in a mixed-asset context. We apply regression-based mean-variance efficiency tests, conditional on currency-unhedged and fully hedged portfolios to account for systematic foreign exchange movements. From the perspective of a US investor, it is shown that, first, international diversification is superior to a US mixed-asset portfolio, second, adding international real estate to an already internationally diversified stock and bond portfolio results in a further significant improvement of the risk-return trade-off and, third, considering unhedged international assets could lead to biased asset allocation decisions not realizing the true diversification benefits from international assets

Do investments in international securitized real estate markets make a statistically significant contribution to an internationally diversified mixed-asset portfolio, and does currency risk exposure have an impact on these results? These questions have become more and more popular for both private and institutional investors for several reasons. First, (securitized) real estate has been a fast-growing asset class around the world during the last decades. In many countries, REIT legislation has been introduced, improving the institutional framework and legal setting of real estate companies; both the number of listed real estate companies and their market capitalization have increased tremendously, coverage by analysts and investors has augmented, and therefore, securitized real estate now cumulatively offers a suitable opportunity to overcome the drawbacks of investments in direct real estate. Second, many studies show that cross-country diversification benefits for pure stock portfolios have been decreasing over time due to increasing financial integration of global stock markets. Thus, investors are looking for other assets, such as real estate, to provide diversification benefits. Third, recent contributions by Lustig and Verdelhan (2007), Lustig et al. (2011), and Verdelhan (2011) provide evidence for systematic risk in exchange rates.2 If currency risk is not independent and random, buying bonds, stocks, and real estate in different currencies does not lead currency exposure to diversify away in international portfolios (see Verdelhan (2011), p. 6). Consequently, the portfolio diversification benefits from an asset class like international real estate may be masked by common movements in currency risk premia, and thus may lead to an underestimation of the true benefits. Therefore, given the mentioned facts and the increased relevance of real estate investments in the recent past, the central questions raised above present the guideline and motivation for our analysis. To our knowledge, the paper at hand is the first study applying mean-variance efficiency tests to such a broad range of markets and assets, covering a time period of more than 25 years, while simultaneously considering currency risk, which we regard as an important contribution to the existing literature of international portfolio diversification where currency risk is often neglected even it is a substantial part of risk. Compared to previous studies applying spanning tests as discussed below, our analysis extends the investment universe from international bond and stock markets to international securitized real estate markets and explicitly considers the impact of currency risk. For the empirical analysis, we use monthly data on bond, stock, and real estate markets from ten countries covering the period from 1984 to 2010. The markets considered, which are located in Asia, Australia, Europe, and North America, cover a large portion of market capitalization in global stock and real estate markets and can therefore be considered representative. The statistical significance of the diversification benefits is analyzed by regression-based spanning tests for the complete efficient frontier and intersection tests for the global minimum variance portfolio as well as the tangency portfolio as suggested by Huberman and Kandel (1987), de Roon and Nijman (2001), and Kan and Zhou (2012). The mean-variance frontier can by definition only shift outwards when a set of assets is added to the investment universe. However, mean-variance spanning tests can be used to check whether a shift of the mean-variance frontier is too large to be attributed to chance. Performing statistical tests allows us to measure and compare diversification benefits from international bonds, stocks, and real estate by means of statistical significance. Taking the perspective of a US investor, we start with a mixed-asset portfolio based on US bonds, US stocks, and US real estate. In three steps, we successively add international bonds, international stocks, and international real estate to this portfolio. We test the contribution of the different assets by two settings – first, by using currency-unhedged returns and, second, by using fully (unitary) currency-hedged returns – in order to account for additional systematic movements in currency markets. If a specific currency loads on a common currency factor, this may introduce unintended co-movements between asset classes within a foreign country, measured in USD. If the currencies of several countries simultaneously load in the same direction on a common currency factor, unintended co-movements between investments in these countries might follow. Thus, if investors consider international assets on an unhedged basis to derive an asset allocation decision, the core asset price risk and the exchange rate risk are being taken as inseparable, and the “full diversification benefits of international investing” can be masked (see Dales and Meese (2001), p. 10). To evaluate the impact of currency exposure in international portfolios, we use fully (unitary) hedged international asset returns to contrast our results based on unhedged returns. For the hedged returns, we use forward foreign exchange market data and account for hedging costs in this regard. The use of hedged returns allows to (almost) separate asset price risk from exchange rate risk in a realistic and quite applicable manner, since currency markets are highly liquid, and it is thus relatively easy to unwind an unintended exposure in the forward foreign exchange market.3 Accordingly, our results based on hedged returns should provide a more accurate picture on the diversification benefits of core assets, i.e. international bonds, stocks, and real estate.4 The findings of our analysis related to international bond and stock markets show that mean-variance efficiency is only weakly rejected for currency-unhedged returns while the intersection hypothesis for the tangency portfolio is not rejected at all, which is in line with the results from previous studies like Eun et al. (2010) and Glabadanidis (2009), among others. However, the results are stronger for international real estate, indicating that the efficient frontier is shifted upwards significantly when international real estate is added to an existing portfolio consisting of international bonds and stocks. This finding is even strengthened when the currency risk exposure is fully hedged. The results from an out-of-sample analysis confirm our findings. Therefore, the contribution of international real estate to an internationally diversified bond and stock portfolio is statistically significant and for an investor also economically meaningful. Finally, we study time trends in the diversification benefits discussed. We find that diversification benefits from international real estate are, in comparison to common stocks, large and relatively stable over time. However, during the recent financial crisis, also the diversification benefits from real estate turn statistically insignificant and economically meager. Interestingly, international bonds equalize the reduced diversification benefits nearly one to one. This paper is structured as follows. Section 2 provides a brief literature review. Section 3 discusses the empirical methodology and Section 4 describes the data characteristics. In Section 5, we report our empirical results and discuss the implications of our findings. In Section 6, we apply out-of-sample tests including investment frictions in form of short selling constraints as a robustness check, and study the time trends of diversification benefits of all three asset classes. Section 7 offers a summary and provides concluding remarks.

In this paper, we investigate investors' benefits from international diversification for a sample of ten countries in a mixed-asset portfolio context including international bonds, stocks, and international securitized real estate, in particular from the perspective of a US investor. Furthermore, we have considered the impact of currency risk exposure, which is often neglected in the analysis of international diversification benefits, but could substantially influence the results since exchange rates exhibit systematic variation. We find significant diversification benefits for global bond and stock portfolios from investing in international real estate. These benefits are robust in an out-of-sample analysis and available for most subperiods covered by our sample. The only exception are the subperiods covering the recent financial crisis, during which there are no diversification benefits achievable from international real estate investments. Overall, diversification benefits seem to be larger from international real estate than from common stocks. These findings support the view that real estate returns exhibit lower international exposure than common stocks and are less globally integrated due to the local nature of real estate markets (see Eichholtz (1996)). We also document that accounting for currency risk exposure of international assets is helpful to separate out the true diversification benefits from global bonds, stocks, and real estate. Since the conducted analysis is based on historical data and applies a static or myopic framework, further research could contribute to answering the question of which variables are good predictors for investment decisions and whether the diversification benefits are still significant in a dynamic, time-varying framework. Regarding hedging currency risk and the presented evidence that currency hedging can matter for portfolio allocations, an analysis on optimal currency hedging should provide interesting insight into this topic. Against the background of the recent financial crisis in particular, constructing and conducting spanning tests using conditioning variables in an international context also presents an interesting topic for future research (see e.g. Ferson and Siegel (2010)).