by Fiona Sweeney
The recent takeover bid for Hammerson by Rodamco has once again focused property minds on the merits of global investment. The fact that Hammerson held an internationally spread portfolio with property in the States, Canada, Australia and Europe was one of its major attractions.
In a world where countries are becoming significantly interrelated, many fund managers are now faced with the question — should they look to global investment? This has become particularly significant when considered against the back-drop of the recent dramatic increases in overseas activity in the UK and the keen competition which many domestic investors are now experiencing even on their own patch. Overseas investors are taking the lead in a market which is fast becoming a property market without frontiers.
While the theory and practice of internationally diversifying investment funds into real and financial assets such as stocks and bonds is well known in investment circles, not so well researched are the potential international diversification opportunities that are available to investors who focus on real estate. Indeed, the idea of international diversification is often forgotten or discarded when it comes to addressing the investment strategy of the property portfolio.
This article argues that international real estate diversification can be a viable investment strategy. The findings and implications are relevant to any institution or individual who has the resources to expand globally or establish a real estate investment portfolio.
Two goals are usually sought by investors when broadening their portfolios to include foreign assets. The first is the reduction of any risk associated with the volatility of investment returns on any one single market. The second is an improvement in the level of return by moving funds around to take advantage of the different cyclical stages characterising business conditions, financial markets and exchange rate fluctuations. Even though world financial markets are functioning in an increasingly integrated fashion, there is still significant variation in economic climate, business environment, individual investment market characteristics and investor preferences to generate varying market returns and risk profiles.
The adage “do not put all your eggs in one basket” underlines the belief that spreading funds across different investments is risk-reducing and that concentrating them enhances risks. For some years now modern portfolio theory and techniques have been employed in the management of investment risk and return for institutional stock and bond portfolios. By contrast there has been nearly a complete neglect of such theory and techniques in the management of real estate portfolios, domestic or overseas. To manage risk and return, modern portfolio theory splits the total risk on any investment into two parts. That part of investment return, volatility or risk which is associated with market influences is referred to as systematic risk; and that which is specific to the investment itself, unsystematic risk.
“Decomposition” of risk
In terms of real estate assets, unsystematic risk is associated with an individual property’s characteristics such as location, regional and local economic conditions affecting demand for the property and the competitive supply of similar properties, its physical design and construction, the tenants’ roster and their credit-worthiness, the structure of tenant leases and the level of property management together with planning and zoning controls. All these factors contribute to the performance of the asset and to the potential volatility of that performance.
These risk factors are perceived as being different from property to property, so that investors by careful and diligent selection of real estate investments can construct portfolios which by virtue of diversification tend to cancel out the unsystematic risk element. Location and property type are among the principal factors selected as determinants of diversification, the rationale being that by spreading funds across property types the entire portfolio is protected from events or factors which affect a particular type of property. In addition, spreading property investments across a country can reduce portfolio exposure to any one regional or local economic event. Consequently, unsystematic risk will largely disappear as an influence on the return of a well-diversified portfolio. To the extent that the return on an individual property is influenced by purely local events, it is unsystematic and washes out in a large diversified portfolio.
Systematic risks, on the other hand, are perceived as being external factors which are not specifically related to the property but affect all investments in the market, although not necessarily to the same degree. Factors such as national economic policy, domestic monetary policy, budgetary and other financial uncertainties, inflationary pressure, election cycles and demographic characteristics will influence the overall markets risk and return profile. Such volatility represents a source of risk which is extremely difficult, if not impossible, to avoid investing in, although local real estate markets may exhibit differing sensitivities to these sources of systematic risks. Consequently with sufficient real estate diversification, the risk of a portfolio is reduced to systematic or market-related risk alone. Some investment advisers suggest that we must put up with this market risk, but the diversification opportunities of one other market should be considered — the international real estate market.
Comparative international performance
Return measures
While international economies in the past have partly influenced each other, world real estate markets have tended to be much more isolated and have not always moved in tandem with one another. This is illustrated in the graphs which trace the performance of prime office property rental values over the period 1976 to 1988 in local currency terms. Individual real estate markets have peaked and troughed at different times, while the sizes of the peaks and troughs have varied substantially.
Notable low points were London in 1976, Amsterdam and Hong Kong during 1983-84. Notable peaks have included London in 1988, Paris in 1985, Amsterdam in 1980 and Brussels in 1987. Looking to world centres, Hong Kong, Sydney and New York showed exceptional rates of rental value growth in the early 1980s and more recently Hong Kong and Madrid have shown very high rates of growth. In general, most centres experienced rising rental growth during the period. This general trend is not too surprising considering the high inflation rates that were experienced by most economies during this period.
The growth in rental values in other major cities has generally been significantly higher than that shown by European centres. However, the range of returns indicates a much higher volatility and hence risk in these markets. There appeared to be a strong similarity in trends between Hong Kong, New York and Sydney, particularly in the 1980s. The European cities appear to demonstrate trends of performance within a similar band, although there are leading and lagging markets. All rose from a low point in 1976 to peak in the early 1980s, dropping away in the mid-1980s and rising again most recently. The figures provided for London relate to the City — and it is worth noting that performances in other major districts of central London are still rising.
Risk measures
Two types of summary measures, the total return and the standard deviation, are shown in Table 1. The total return figures are the annualised compound growth figures and the standard deviation is a basic statistical measure of the volatility or riskiness of the market’s return performance. There is considerable divergence in both return and risk among the markets. For instance, both Hong Kong and Madrid on average have performed well during the period but they are also by far the most risky, with standard deviations of 9.7% and 5.1% respectively. On the other hand, the German market has a standard deviation of only 2.6% but a return of 6.3%, reflecting the commonly held view that high potential gains imply high potential risk.
When comparing the standard deviation of returns on the various international markets, it is more useful to normalise the standard deviations to their respective averages. This measure, known as the coefficient of variation, provides a better comparison of relative volatility of returns than comparing standard deviations among markets. In simplistic terms it shows the pay-off in units of return for every unit of risk taken on.
The German and French real estate markets show comparable standard deviations of 2.57% and 2.56% respectively, but when these returns are normalised it becomes clearer that Paris, with a co-efficient of variation of nearly 5.0, has a higher pay-off than Germany. In other words, for every unit of risk taken on in this market, there is the reward of receiving 5 units of return. The ranking column orders the cities, based on this risk/return pay-off. It is apparent that the centres fall into three groups: Toronto, Sydney, London and Paris show high pay-offs; Amsterdam and Hong Kong show low pay-offs; the other cities show broadly similar pay-offs.
Correlation analysis
In addition to assessing the risk and return characteristics of the individual international markets, the correlation (that is, the degree to which two markets move together) must be measured. The highest possible correlation is 1 (perfect positive correlation) and the lowest is – 1 (perfect negative correlation). If two markets have a correlation of 1 then they move together in lock step; at the other end of the spectrum a co-efficient of -1 means that they move in exact opposition. In the middle, a correlation of 0 means that the two markets are unrelated. The higher the correlation, the lesser the extent to which an asset can be used to diversify another and vice versa. This implies that if a pair of real estate markets have a negative correlation of returns, then in circumstances where one of the markets is performing badly the other is likely to do well and vice versa. The average return of holding a real estate investment in more than one market is likely to be much safer than investing in one market alone. Table 2 shows the correlations between pairs of markets set out in the format of a correlation matrix in local currency terms.
It is apparent that no pair of cities have perfectly correlated rental growth performance. However, it is obvious that some rental values move in tandem more than others. For example, the real estate returns between New York and Sydney and New York and Hong Kong are highly correlated: this is not too surprising in the light of the international trade flows between these countries. In an earlier study it was shown that returns between cities in the same country are highly correlated, reinforcing the case for international real estate diversification. More interestingly, the matrix shows that the UK market is either negatively correlated or has low correlations with all of the cities (in particular the European) implying that when the rental values are increasing in the UK, they tend to be decreasing in these centres.
It therefore follows that more efficient real estate investment could be achieved by spreading funds among cities whose performance is not so highly correlated — and preferably negatively correlated. The ideal solution is to select a group of cities or countries whose correlations are consistently low. This would reduce the likelihood that the performance of all real estate holdings would simultaneously generate low/high returns. To assist in the identification of markets that can be combined to assist international diversification, blocks of markets that tend to move together can be identified, hence finding groups of correlations that fall within each block and have relatively high correlation relationships, and have lower correlations outside each block.
It should be recognised that while correlation analysis is useful if not essential for international real estate investment decision makers, correlation statistics do not remain static through time and constant monitoring is advised.
Exchange rate risk
In addition to return volatilities, the investor in a foreign market is also exposed to exchange rate variability, since currency fluctuations can cause the return on an investment to differ from that expected when converted back to local currency. It would be important to calculate the correlation matrices and the returns from various currency angles, but it can be argued that exchange rate risk does not have a true role to play, as currency risks can easily be edged to separate the view on the market from the view on currency. In other words this risk can potentially be eliminated.
Optimal international portfolio
The merits of international diversification can be further illustrated by the construction of optimal real estate portfolios. An optimisation model was employed to select portfolios which fall within the efficient frontier. The efficient frontier comprises the set of portfolios derived from the selected real estate markets having the maximum expected return and risk dimensions with the characteristic that there is no portfolio that is not part of the efficient frontier that has a higher expected return for a given level of risk. If a portfolio is on the frontier it is not possible to reduce its risk without also reducing its expected return or to increase its return without also increasing its risk. Portfolios on the efficient frontier range from the portfolio with the minimum risk to the maximum expected return. The optimal portfolio chosen will depend on an investor’s risk/return profile and will be the one that most appropriately reflects the manner in which the investor trades risk for return.
A very risk-averse investor would select a portfolio that has low risk and consequently a low return. An aggressive investor would select a portfolio with a higher risk and consequently a higher return. Table 3 highlights combinations of portfolios on the efficient frontier.
The highest return portfolio is 100% in Sydney giving a return of 17.4% and a risk of 4.1%. The lowest risk portfolio is diversified into seven of the eleven markets and shows a return of 11.8% and a risk of only 0.6% — significantly lower than the risk on any single market, even the lowest, Toronto, with an SD of 2.0%. In considering optimal portfolios it is important to take account of any impediments to investment in these markets, for example, Tokyo is a particularly difficult market for overseas investors to enter, hence, optimisation must be practically approached and realistic constraints on market weightings introduced.
Implications for portfolio strategy
The relationship of return to risk for real estate assets has not been as well researched or defined as it has for other classes of assets. This deficiency has seriously handicapped property portfolio managers, particularly when developing diversification strategies and investment policies. Needless to say, inadequate research can lead to mispriced assets and counterproductive diversification strategy.
In arriving at the optimal real estate portfolio diversification strategy, traditional practice disregards the significance of possible correlations among real estate returns in favour of a blind one of each appraisal. The general approach to property portfolio structuring seems to be to buy properties when they become available if they look like good deals, with little regard for the equally important issue of how the acquisition fits with the other holdings in the portfolio and what, if any, effect it will have on the overall risk and return objectives of that portfolio. Hence, minimisation, this approach can simply result in, providing an equal opportunity for any one property to sink the portfolio.
Risk may be compounded, returns may be needlessly watered down by inefficient combination of portfolio of assets and management costs may be multiplied by having more segregated properties than needed. Assessment of market risk and the correlation/coefficient between markets can aid and reduce the need for diversification and increase efficiency and hence the return on a portfolio.
Part of the solution obviously rests in developing a two-tiered approach to portfolio management, a macro level and a micro level.
At the macro level, analysis should focus on the composition and investment characteristics of the portfolio as a whole, identifying major strategic investment options and long-run implications.
At the micro level, the analysis should focus on project analysis, employing traditional real estate project analysis and focusing on the selection of the individual properties that make up the portfolio, evaluating a property’s specific risk/reward potential against the investor’s performance targets. Macro analysis drawing on the techniques of modern portfolio theory derives the characteristics of risk and return for the portfolio as a whole from different combinations of the individual property types and geographic locations.
Conclusion
The result of this study at the macro level suggests that international real estate diversification can be a viable strategy for both the UK and the foreign investor. International property investment does provide significant diversification benefits and is a uniquely effective portfolio diversifier because the economic and fundamental variables that cause property values to fluctuate are not the same as the variables that cause fluctuations in the value of other assets.
Markets may not, however, continue in the future to offer the same risk and return performance as was shown over the period of the study, and even efficient portfolios may not all the time perform to the highest expectation. Nevertheless, it is very clear that the potential gains from international diversification have been identified as being positive and that the importance of continued international real estate allocation monitoring cannot be overstressed.
References
(1) Richard Ellis World Rental Survey 1988
(2) Sweeney F M “International Real Estate Diversification A Viable Investment Strategy” Journal of Property Management Vol 5 No 4
(3) Sweeney F M “20% in Property a Viable Strategy” Estates Gazette February 13 1988
