How far does the pricing of a market reflect the risk involved? A new model developed by Jenny Buck, Mark Callender and Alex Krystalogianni sheds light and identifies substantial variances across global markets
As domestic returns soften, interest in global real estate investment is increasing, and countries that have not previously been considered established real estate investment markets are now being considered. Not only do the markets in different countries operate in different ways, but geopolitical and legislation issues also differ. A key part of any global investment strategy must therefore be to assess these risks and compare returns on a risk-adjusted basis.
We have built a model that assesses the risk of 30 different countries around the world, and this paper summarises the methodology and conclusions of this work. We have made the following assumptions in developing our model:
real estate risk premium = Transparency + Liquidity risk + Volatility.
This has been adopted as we believe that its key components drive the depth, liquidity and maturity of the real estate market and will therefore affect the extent to which the risk premium associated with the asset class should be adjusted.
These constituents have been determined as follows.
Transparency: We have adopted the established and recognised Jones Lang LaSalle transparency index for global real estate investment markets for our transparency constituent. The index provides for five attributes of real estate transparency: availability of investment performance indices; availability of market fundamentals data; listed vehicle financial disclosure and governance; regulatory and legal factors; professional and ethical standards.
Liquidity risk: We believe that liquidity risk is associated with the uncertainty of exiting an investment, both in terms of timeliness and cost. The ability to exit an investment quickly and with minimal cost greatly depends on the type of asset being held. The greater the time it takes to exit a position and/or the higher the cost of selling out of the position, the more compensation investors should require.
Our liquidity indicator is expressed as the ratio of the investment transactions (average of the last two years) relative to the size of the investment market. We used CBRE data for the European markets and DTZ data for the Asian markets. This liquidity indicator has been further adjusted to allow for unusually high rates of turnover in certain countries over the past two years and for the size of the local investment market versus the size of the global investment market.
Volatility of returns: This should reflect the uncertainty of future cash flows, the ex-ante volatility of returns. Since the ex-ante volatility of returns cannot be accurately calculated, we used historical data: the historical volatility of prime office rental growth (1990-2006). Although prime rents are an imperfect guide to property cycles and we would prefer to use IPD-type total returns series which cover all sectors, prime office rents are the only time series which are available in all countries.
Risk-free rate: The risk free rate is the yield on 10-year government bonds at the end of June 2007. Variations in the risk-free rate implicitly reflect investors' expectations of future movements in exchange rates. For example, the lower level of bonds yields in Japan relative to the US reflects an assumption that the yen will appreciate relative to the dollar.
A second assumption is that the risk-free rate accurately reflects the market's perception of the country. However, if the bond market in a country is thin and dominated by one or two local players such as a large state pension fund, then the risk-free rate may not be a true reflection of the risks associated with that country. We therefore have increased the risk free rates for Malaysia and Thailand by 100bps, following discussions with our global bond colleagues.
Having established the assumptions and constituents of our model, we then built the model. This involved standardising the real estate transparency, liquidity risk and volatility numbers to create the real estate risk premium.
While volatility creates risk, investors can to some extent mitigate the impact of property cycles, if they are able to buy at the market troughs and sell at the peaks. Many investors have an ambivalent attitude to transparency, because they believe that they can exploit opaque markets to their own advantage.
We acknowledge that the upper and lower limits which we have set for the three property risk premia are based on subjective judgement. We are not aware of any surveys of investor attitudes which quantify the three risk premia for real estate. The one reference point has been to look at the actual long-term difference in annual total returns between property and government bonds which has been around 175bps in the US, 200bps in the UK and 275bps in continental Europe. The model has been calibrated so that the overall risk premia, combining the three elements, are consistent with these observed differences in returns.
Not surprisingly, given that market pricing is dependent on transaction volumes, there is a strong correlation between the risk premia for market transparency and for liquidity - see table. However, the two premia are discrete variables and there are certain markets (eg, Denmark, Switzerland) which are relatively transparent, but have low liquidity. Neither market transparency nor liquidity are highly correlated with volatility. For example, certain Asian markets, such as Hong Kong and Singapore, are highly liquid, but relatively volatile.
The risk premia for the 30 countries in the analysis are illustrated in Figure 1, along with their constituent parts. The data show that:
Although the research is interesting, it is of no use unless it is applied. We use this work in several ways: helping us to understand further the risks associated with the particular countries that we are investing in, identifying over- and/or under-priced markets and determining regional allocations.
In attributing the risk premium associated with a particular country in to its three constituent parts, we can understand the generic risk associated with a country in more depth and are therefore able to make more informed investment judgements.
When our risk premium model is combined with the risk-free rate, we can determine the required rate of return for each property market. If these required rates are then compared against forecast returns, we can begin to identify over- and under-priced markets.
Figure 2 compares our current risk adjusted forecasts for the period end-2007 to end-2011 with the required rate of return in the 30 countries. The relatively short data series for Asian property markets means that we have less confidence in the forecasts for the region than in the forecasts for Europe and North America, which carry a much higher level of conviction.
In conclusion, assessing the risk associated in investing in overseas real estate markets is critical. The results of the model intuitively seem right and are consistent with academic studies and with the observed long-term risk premium for real estate in continental Europe, the UK and the US.
It should be noted, however, that countries with high risk premium will not necessarily be the countries that provide the highest risk adjusted returns, as their prevailing risk-free rates may be very low, as is the case in Japan at the moment.
Furthermore, all models are only as good as the data that go into them and as the international real estate markets evolve and mature the accuracy of our model and our forecasts is expected to improve. In the meantime, fund and or property due diligence must remain a fundamental part of the process for determining the risk associated with an investment.
Jenny Buck is head of multi-manager at Schroder PIM
Mark Callender is head of property research at Schroder PIM
Alex Krystalogianni is head of international property forecasting at Schroder PIM