Better integration of real estate with other asset classes will increase ongoing moves towards global exposure, as Jean-Martin Aussant and Peter Hobbs reveal
The traditional, home-biased focus of real estate investing is starting to change. The globalisation of real estate is being driven by the largest sovereign wealth and pension funds, many of which have explicitly global mandates. But there is a broader trend, driven by the recognition of the diversification benefits of international real estate exposure. Most of these investors have started to understand the role of real estate in a multi-asset-class context, and this perspective tends to increase the demand for international real estate, further eroding the real estate home bias.
The appetite for international exposure across asset classes is nicely demonstrated by work conducted by Towers Watson (for equities and bonds) and MSCI (for real estate). This work shows that the bias toward domestic investment is lowest for equities but far higher for fixed-income and for real estate. There is a logic for fixed-income having a relatively strong home bias, given its role in hedging domestic liabilities, but this would appear to be somewhat less compelling for real estate.
Despite this home bias, a range of recent studies point to an increasing appetite for foreign real estate, driven by concerns over the pricing of domestic markets, particularly in the US, Canada and Australia, as well as the diversification benefits. These benefits have been complemented by the increasing options for investing internationally, with a series of more robust and better governed investment platforms covering most of the world’s real estate markets.
These moves to international exposure raise questions related to risk, particularly in terms of the amount of overseas real estate, its location and the leverage that should be employed. The global financial crisis (GFC) proved that international diversification can be used to mitigate the risks of a strong downturn in an individual country. The GFC also revealed the extent to which inter and intra-country correlations can increase in a crisis, emphasising the need for truly global investment strategies to mitigate risk.
The significant benefits of international diversification are illustrated in figure 2, which shows the return implications of different global exposures. The chart compares the performance of the IPD Global index with that of the UK and the global series ex-UK. At a glance, it is possible to see the greater volatility of the UK, particularly during the global financial crisis. The chart also suggests that global ex-UK generates higher return for far lower risk than a pure UK exposure. This is an illustration of the benefits of international diversification, which are compounded when correlations and diversification benefits are taken into account.
This example demonstrates that a reduction of home bias can improve portfolio diversification, but it is based on naïve comparisons of the risk and returns of global markets. In particular, it uses ‘appraised’ performance series that understates the true volatility of markets.
Although they are beyond the scope of this article, a series of techniques have been developed in order to de-smooth real estate indexes, the most recent of which has been conducted by MSCI in order to integrate a private real estate risk model (commonly known as PRE2) into the Barra Integrated Model (BIM). The full description of the method is in Shepard, Liu, and Dai (2014). In summary, the Barra risk model uses many data inputs – appraisal-based data, listed real estate returns, and real estate transaction prices – within a Bayesian de-smoothing framework. The use of numerous sources of real estate data successfully improves the risk estimates and reduces the noise associated with appraised series. The result is a set of risk estimates that portray real estate as having higher volatility and greater correlations, between countries and with other asset classes, than previously thought.
MSCI’s real estate risk research raises a series of questions relating to home bias in real estate investment. First, is the argument for global diversification undermined if covariance between national markets is higher than first thought?
Second, does the increase in risk caused by the use of debt negate the benefits of overseas investment, given that the model estimates real estate to have higher variance than first thought?
Third, how do the risk implications of international exposure vary from country to country?
The MSCI research team has used the Barra model to create a number of case studies, which help answer these series of questions. A starting point in this analysis is the risk implications of a real estate allocation by, for example, a UK-based investor. In this example, the investor has allocated 10% of its capital to UK real estate and the standalone risk of this allocation averages 9.8%, excluding leverage. In this case, the correlation of real estate with the other asset classes shown is high, but its overall risk contribution is still low. Fixed-income also contributes little risk to the overall portfolio, despite its large allocation, with the largest risk contribution coming from the equity exposure.
One of the central benefits of international real estate derives from the significant differences that persist between countries. These differences are captured in the chart showing the historic return against the volatility of the main markets covered by the PRE2 model, with the size of the bubbles representing the size of the real estate markets. Figure 4 shows that the UK has tended to generate slightly below average performance with high volatility, which contrasts, for instance, with France and Sweden, which have tended to generate higher returns for lower volatility. Although the chart excludes the issue of correlations between markets, it suggests that a UK investor with overseas real estate exposure might benefit in particular from risk reduction, while for a German investor it might come from return enhancement. The relatively high-risk and low-return Japan-based investor might, in contrast, benefit from both return enhancement and risk reduction from international exposure.
It is in the context of these different behaviours that it is possible to explore the trade-off between UK and non-UK real estate, and the implications of adding different levels of leverage to the international exposure. In this case, leverage of 20% is applied to the domestic real estate portfolio, taking the overall standalone risk up to 12.25%. Figure 5 shows, in the green highlighted area, the effect of increasing the international real estate exposure by 20% increments with no leverage being added to the international exposure. This demonstrates the significant reductions in risk, from 12.25% for full domestic exposure to less than 6.5% for full international exposure.
The table also shows the impact of increasing leverage for different levels of international exposure. In all cases, the addition of leverage increases risk. But, for international exposure up to 40%, leverage can be increased to 60% and result in a lower level of overall risk than a purely domestic portfolio. The table also shows that high levels of leverage – generally more than 60% – have a significant impact on overall risk levels.
The answers to the two original questions are now evident, at least from the perspective of a UK investor. Even once real estate’s heightened variance and covariance is incorporated into the model, there is a strong case for risk reduction through global investment. However, one should carefully assess the level of debt used to achieve this international investment. In this case, low levels of debt do not negate the benefits of foreign investment, but higher levels increase risk to the overall portfolio.
The example of the UK demonstrates the risk reduction benefits from increasing exposure to international real estate. This example is relatively crude as it assumes exposure across all global markets covered by the PRE2 model, weighted according to the size of each market. In reality, most investors have more targeted strategies reflecting a preference for, or aversion to, particular regions (such as Asia Pacific or North America) or particular countries. These more targeted strategies tend to be based on a combination of the choices related to performance and risk, and the availability of suitable investment opportunities across the various markets. By making use of the PRE2 model it is possible to draw out the risk implications of exposure to different types of market. This is captured by figure 6 showing how the risks of real estate varies when increasing levels of international exposure based on four different international portfolios: global ex-UK; US; euro-zone; Asia. Leverage is assumed to be 20% for each of these international exposures.
The results show that each of these portfolios reduce the risk of the wholly domestic portfolio, certainly when allocating up to 40% overseas. For the US exposure, the risk starts to increase once allocations rise above 40%, due to the relative volatility and high correlations with that market. For the global and, particularly, the euro-zone exposure, there is a continual reduction in risk for the addition of greater exposure to these markets. Clearly, there is a range of scenarios that can be generated, but these examples illustrate the benefits of international exposure and how they can help drive portfolio construction as well as measuring the risk of actual exposure.
A range of additional considerations needs to be incorporated to arrive at a meaningful allocation to individual markets. This includes the differential return expectations with, for instance, Asian and perhaps the US markets generally expected to generate stronger, demographically driven growth than the euro-zone. It also includes the costs of such exposure. These costs can come from a range of dimensions, including the increased need for staffing and oversight, higher fees, currency hedging and tax leakage. Although these differential costs and return assumptions do not directly impact the risk of international exposure, they can have a significant implication for risk-adjusted returns.
While the UK example shows the compelling case for international exposure for a UK-based investor, the implications vary significantly from country to country. On the one hand, domestic markets vary in their risks and the extent to which they are correlated with other markets. So, for instance, the Dutch, Swedish and German markets have less risk than the UK, so the case for international exposure might be less significant than for a UK investor.
On the other hand, the multi-asset-class context might vary and this can have implications for the risk contribution of real estate. These variations in the multi-asset-class context are summarised by Towers Watson’s survey on pension fund allocations, which shows that Australian, UK and US pension funds, for instance, tend to have far higher allocations to equities compared with bonds than, say, Dutch, Japanese and Swiss pension funds. For the former, equity-dominated investors, real estate tends to serve to reduce multi-asset-class risk, and for the latter it tends to increase it.
The implications of these differences are summarised in figure 8, which shows the risks of real estate exposure for different countries at different levels of international exposure. In all cases, the international and domestic exposures are assumed to be leveraged at 20%. This summary suggests that it is the more volatile markets of China, Japan and the US that would benefit most significantly from international exposure. For China, for instance, the standalone risk of domestic exposure of 18.45% would be reduced to 7.69% by increasing international exposure to 80%.
The real estate home bias is starting to be eroded, with asset owners in most countries already investing internationally, or actively exploring the options for building such exposure. This trend is running in parallel with greater scrutiny from risk managers that are seeking to integrate real estate risk analysis with other asset classes. The diversification benefits of investing internationally can significantly reduce the risk of real estate exposure. As always with real estate, the implications vary from country to country and investor to investor. A range of other factors also need to be considered such as return objectives and the risks associated with implementation and market pricing. But these trends, complemented as they are by the increasing availability of real estate platforms, are set to further erode the home bias that has, until recently, been a major characteristic of the real estate asset class.
Jean-Martin Aussant is executive director and Peter Hobbs is managing director at IPD