The recent collapse in property values has high-lighted the cyclicality of property returns and how they are increasingly dominated by global capital flows. Peter Hayes examines the implications
History tells us - unsurprisingly - that real estate downturns lead to changes in portfolio management styles. If performance turns out to be worse than expected, managers naturally try to determine why and what can be done to improve in future. Corrections in recent decades have led many investors to change their approach to investing. The downturn of the mid-1970s prompted direct property investors to diversify overseas, while the early 1980s correction raised awareness of the impact of macroeconomic risk on portfolio performance, and the early 1990s highlighted the need for rigorous diligence in the valuations processes. The question today, then, is what lessons might portfolio managers take in the wake of one of the worst property downturns in history?
If history is a guide, the practice of portfolio management will be altered more than the theory. That is not to say current portfolio management theory is perfect. Critics of modern portfolio theory and property pricing models point to problems with high transactional costs, the risk introduced by the lag between approvals and pricing, and the lumpy and heterogeneous nature of commercial assets. But such shortcomings are actually more technical than fundamental, and say more about the unique characteristics of property than about any shortcomings of risk-return analysis. They point out the need to adapt rather than drop modern portfolio theory.
Property portfolio theory
Portfolio decision-making in theory comes down to two separate but interrelated questions:
According to Martin Hoesli & Bryan MacGregor, professors at the University of Aberdeen in Scotland who wrote a textbook on the principles and practice of portfolio management, each of the questions can be answered in one of two ways, producing four basic strategies that they label modern, traditional, large funds and passive.
These styles are illustrated in figure 1, which highlights the choices for a portfolio manager between taking positions based on forecasts or a selected benchmark. The choices available to determine stock selection - either based on pricing models or diversification strategies - can be read from right to left. Structure and stock have to be considered together. Although stock selection has to match the portfolio's strategy, the eventual structure of the portfolio is driven in part by product availability.
Although this approach works well as a summary, it just skims the surface of the considerations made by a portfolio manager. For instance, how useful or accurate are forecasting or pricing models? What is the scope of diversity? How easy is it to replicate a benchmark? These are just some of the important questions that are raised by portfolio managers to determine how well-served they are by their strategies. Additionally, what is not immediately clear is whether the answers to more detailed questions will give rise to new lessons at the expense of forgetting lessons learned in the past. At the broadest level, these considerations include problems with real estate data, the significance of herd behaviour, and the linkages between credit cycles and property cycles.
Real estate data
The interpretation of real estate data is a well-documented problem for portfolio managers. Real estate is an inherently private industry and players are reluctant to share information with one another. This gives rise to gaps as well as apparent inconsistencies in data published within the industry.
Real estate data fails on many levels. Its failures include limited coverage, a lack of timeliness, a short history and smoothing, not to mention that the accuracy and consistency of data can change even over short periods of time. Other common problems include clarifying geographical and physical definitions of ‘prime' and ‘secondary' real estate, using take-up as a proxy for net absorption and separating temporary incentives from headline rents. Working through all these problems takes time and effort. Other issues with data include the lack of complete or accurate data, which makes it more difficult to build models, and the divergence between transaction yields and valuation yields, particularly when transaction volume is low (figure 2).
All these data problems make it difficult to build and interpret models. Poor data can lead to poor decision-making, no matter how sophisticated the processes. Knowing the limitations of data is therefore crucial for a portfolio manager's decision-making process. However, these problems are not new, as is demonstrated by the industry's backing of research initiatives such as those led by Investment Property Databank (IPD), the RICS and the Investment Property Forum (IPF).
The limited nature of real estate data creates additional problems for portfolio managers when it comes to acting on the signals produced by pricing models and forecasts. Limited commercial real estate data curbs the number of results, which means that portfolio managers often end up producing the same models, generating the same results, and reaching similar conclusions. The result is that they are likely to act in the same way while thinking that they are acting independently. In effect, the models used to guide decision-making can produce spurious results. But this is not the only cause of herd behaviour. The limits to real estate data in elements such as accuracy, timeliness and frequency will push managers to use soft indicators to guide decision-making. One reaction to uncertainty caused by limited data is to mimic other portfolio managers, which in turn may cause other managers to follow suit.
Herd behaviour helps to generate large swings in activity. It is a manifestation of an apparent coordination failure in the real estate market - a failure that is linked to inefficient pricing. As with the problem of real estate data, herd behaviour is a well-documented issue for portfolio managers. It is thus unclear whether any new lessons about the topic might be learned in the context of the most recent financial crisis.
In part due to herd behaviour, the performance of commercial real estate in the modern era has been relatively volatile. Figure 3 highlights the evolution of UK property values and income return since the 1920s. Notably, while income return has been remarkably stable, capital growth has been markedly volatile. Even though capital values repeatedly endure boom-and-bust cycles, the peak-to-trough capital value decline of 44% in the most recent downturn makes it among the most severe.
Predicting the exact timing and nature of a downturn remains almost impossible for market participants. Property is a capital-intensive industry and therefore has an intrinsic vulnerability to wider financial pressures. Debt availability is the lifeblood for investors, so bank lending conditions play an important role in the cyclicality of real estate markets. In times of easy credit, such as those experienced in the mid-1980s or 2003-07 periods, seemingly well-capitalised banks had surplus funds and were able to lend on increasingly aggressive terms. The flow of credit to property managers allowed them to enhance returns though gearing, but also led to an increasingly large pool of money available to spend and, crucially, pressure on fund managers to spend it.
In boom times characterised by easy credit, the real estate market becomes increasingly dominated by ‘forced buyers', or those motivated by the need to deploy cash even if the ‘right' product is not available. This depresses yields on prime properties, as buyers chase the best assets, and reduces the pricing gap between prime and secondary assets due to so-called ‘mission creep' or ‘style drift'. In a downturn, the reverse is true. Forced buyers turn into forced sellers as banks withdraw financing and falling values require investors to post additional capital to preserve lending ratios. This motivates additional sales, widening the gap between supply and demand in the market, and potentially resulting in sharp price corrections. Figure 4 illustrates the changing spread between low- and high-yield properties, as reported by the IPD Quarterly Index for UK central and inner London office markets. The gap between high-quality and transitional properties fluctuates during the course of a cycle.
Again, credit cycles are an inherent part of real estate. Previous booms and busts in real estate are closely aligned with cycles in credit markets. Thus there are not necessarily any new lessons in this regard for portfolio managers to learn from the most recent financial crisis.
Up to this point, we have dealt with issues that are not specific to the current market. To that end, the latest boom-and-bust episode can be considered as just another, albeit severe, example of a property cycle. The most recent downturn has been characterised by some standard factors, including the retreat of lenders, falling prices and rising spreads on secondary assets. As with all cycles, however, there are differences. Today, for example, low interest rates and strong policy responses by authorities have helped prevent a sharp rise in distressed sales. Even so, market participants have just had a brutal reminder that property continues to be a highly cyclical investment class. Nothing else has changed to suggest that such cycles won't occur again.
So what lessons can be drawn by portfolio managers to prepare for future cycles? One lesson could well be about how synchronised the downturn turned out to be, not just across sectors and geographies but globally. In the same way that there was a global boom in property returns off the back of excess liquidity, there has been a global collapse in values (with some lags). This synchronisation of global capital markets is a reflection of the growth of cross-border global investing. While such globalisation has been occurring for a long time in the listed real estate sector, in recent years it has become a key factor in driving returns in the private market.
This gives rise to two concerns for portfolio managers. The first is the apparent limits to diversification strategies when property returns appear to be increasingly driven by synchronisation. The second is the significance of global capital flows in driving property performance. The performance of the market during the boom years between 2003 and 2007, when asset values shot up as a result of excess global liquidity, demonstrates how much returns can be delinked from fundamentals. Since the financial crisis, investors have given greater significance to income, linked as it is to the fortunes of the occupier. But there is no telling when global capital flows will come to dominate portfolio performance, both upside and downside.
In conclusion, portfolio managers need to be wary of overreacting to the financial crisis. After all, although the recent downturn was particularly bad, many elements of the market remain unchanged. Managers continue to face problems with data, herd behaviour remains a possibility and the cyclicality of credit could bring about a repeat of the real estate cycle.
Nevertheless, we know that shocks to the property market can bring about significant change. The current downturn has revealed the extent to which global capital movements can dominate the market, first during the boom and then in the subsequent bust. This may reverse temporarily, as lenders and investors retreat back to favoured home markets or mature countries, but the secular forces of globalisation look unlikely to be held back for too long.
For the portfolio manager, such an environment highlights the growing importance of tactical portfolio allocation: the choice between prime and secondary stock and locations. Yet we must be careful not to dismiss the importance of diligence in selecting individual assets. For investors with longer-term horizons, short-term capital movements can be a distraction, and there is no substitute for judicious asset selection and asset management.
Peter Hayes is European director of research at Pramerica Real Estate Investors. This article is based on a presentation he gave at the recent IPD Conference in Brighton, UK