‘Bubble episodes’ in real estate markets are a source of valuable pricing information, say Sotiris Tsolacos and Chris Brooks

Following the surge and subsequent collapse in real estate prices, research to establish whether bubbles drove valuations is in vogue. The early detection of bubbles, their duration and intensity, are topics of particular interest to real estate investors.

Bubbles are commonly defined as the element of the market price which exceeds or undershoots an asset’s ‘fundamental’ value. Research on assessing the deviations of actual from fundamental values has also gained momentum. We see greater attempts in the real estate industry to establish these so-called fundamental, ‘equilibrium’, ‘underlying’ or ‘fair’ values to approximate the ‘correct’ price of the asset.

We believe real estate research will remain occupied with these questions. Undoubtedly, systematic analysis to estimate fundamental values and identify possible episodes of bubbles can generate useful signals that will trigger speculation of a looming major price change (for example, a price collapse). This will, in turn, activate further analysis of the current market conditions and the probability of a forthcoming price adjustment. At this stage, behavioural factors and their own beliefs may take over and drive the timing of investor action. This article illustrates an application of a commonly used technique to detect bubbles and offers observations for the use of bubble-related analysis in practice.

The detection of bubbles can be based on several methodologies. These range from naive ones – for example, comparing recent price changes to a long-term average – to more sophisticated approaches requiring the use of more advanced econometric techniques. Below, we estimate deviations between actual and fundamental values using the van Norden and Schaller technique*, a widely used method in finance. It is applied to the three main property sectors at the pan-European level.

Figure 1 plots the actual capital values for prime offices in Europe as a whole. We discern discrepancies between the actual and the estimated fundamental capital values until 2005 which fade away, as we would expect, after corrective market action. But from 2004, we begin to see evidence of price exuberance and a sharp mismatch emerging between the two series, possibly forming a bubble for reasons that were well publicised in the real estate market (for example, availability of debt, liquidity, lower risk premia, beauty contest). The subsequent correction pushed actual capital values well below fundamentals.

Figure 2 shows a similar pattern for the European prime retail market. Before 2004, actual and fundamental values seemed to be more in line compared with offices. However, the picture from 2004 is similar to that for offices. The pattern for the prime warehouse market is different. Actual prices were above fundamentals for a long period of time, perhaps demonstrating confidence among investors that the market could continually trade at prices above fundamentals. But mirroring the other two sectors, the gap widens from 2005 and it is followed by a correction.

The dynamics of the swings around the fundamental asset price owes to several factors. The occupier market, which partially underpins any fundamental price model, and the investment market respond to different influences in the short run. The investment market, for example, can discount a future cash flow at rates which adjust abruptly to capture swings in the appetite for risk and required returns. On the other hand, the occupier market is slow to encompass developments in the economy and business activity. The forces that will impact positively or negatively on the asset’s net income reveal themselves slowly. 

The price exuberance in 2007 gave way to an undershooting of fundamental values across all three sectors. This methodology suggests that the undershooting has now been rectified. Recently, actual prices have again begun to turn away from fundamental values (move above fundamental prices), both in the office and the retail sectors.

In the warehouse sector, a balance is currently observed. The signal indicates that any further yield compression could leave the market in a vulnerable position in the sense that a yield adjustment could occur. Much will depend on how net operating income (which is a primary input in our application of van Norden and Schaller’s methodology) will support fundamental values. A further yield compression without a corresponding income improvement will drive both prime office and retail markets further into disequilibrium. In the present situation, the focus should be on the factors driving income, which will in turn determine fundamentals prices.

The study of bubble and signals for price correction is useful but in practice it is certainly not as simple as presented here. Users and those engaging in bubble analysis should bear in mind four points.

• Bubble tests are always based on the presumption of a correctly specified model to generate fundamental values. In our case, we observe a distinctive pattern in the estimation of fundamental prices in the warehouse sector. Perhaps a different methodology could have produced different dynamics.

• Relating to the first point, fundamental asset price models are not universal between geographies, making the process even harder.

• Once there is deviation between actual and fundamental values, there are several tests available to establish the presence of bubbles. Some of these tests may give contradictory results – that is, one test may establish a bubble with another test not allowing for bubbles for the same data series.

• When actual prices rise above fundamental values, it does not necessarily imply irrationality. Rational investors, aware of mispricing, are still trading. On their judgement, they can sell at a higher price (taking advantage of irrational or ignorant investors) at the time when it is obvious that the bubble is about to collapse. There is a view in the stock market that it is riskier to bet against the bubble than to ride it.

* Van Norden, S. and Schaller, H., (1999) ‘Speculative Behaviour, Regime-Switching and Stock Market Crashes’, in Non-Linear Time Series Analysis of Economic and Financial Data, P. Rothman (ed.), Kluwer Academic Publishers, Boston.

Sotiris Tsolacos (far left) is director at Property & Portfolio Research. Chris Brooks is professor at ICMA Centre, Henley Business School

Topics