The behaviour of REITs in the presence of bubbles in other asset classes have important implications for investors. Sotiris Tsolacos and Ogonna Nneji explain
Academic and industry research have both investigated the impact of REIT inclusion on real estate portfolios. The main conclusion to be drawn is that price developments diverge in the short run, making REITs less ‘real estate’, but at the same time diversification benefits can accrue.
In the short run, it is widely documented that REITs are prone to stock market influences and volatility despite the different characteristics, such as the distribution of dividends, whereas the direct market is characterised by lagged responses. In the long run, though, REIT performance will resemble that of the direct market, since the income underpinning values in the REIT market is generated in the underlying market. Research has gone further to show that even in the short run the behaviour of private and public real estate may not differ notably if we adjust for leverage and consider portfolios with similar asset and geographical composition. Recent research has also found that REITs and the general stock market are synchronised both in the bull (high return, low volatility) market and in bear periods.
These findings are useful to understand the impact of REITs on investment portfolios; however, the role of REITs can be studied further. Assets exhibit recurrent price rallies. Observed episodes of price rallies and enduring abnormal pricing can lead to bubble building. Speculative bubbles are likely to collapse and result in notable portfolio value losses. REIT investors will be interested in the behaviour of REITs in the presence of bubbles in other asset classes, given the linkages argued in the previous section. Are they synchronised and contagious? The deviation of actual prices from the levels supported by fundamentals, a key source underpinning the realisation of bubbles, can sustain for several years and be contagious. During this period investors have sufficient time to assess the bubble size, the possibility of it collapsing and to adjust their strategy to mitigate possible adverse price developments. As always, action will be dependent on risk aversion, but at least relevant signals are being received and evaluated.
The intuition is simple. Once bubbles are established they may survive or collapse in the next period. This is conditional on the bubble size. Larger and persistent over-valuations (possible larger bubbles) are likely to collapse. At times when speculative bubbles have occurred, the likelihood of price collapsing is high with apparent risks to portfolio valuations.
Bubble building research has the following implications for REIT investors:
• It raises awareness among REIT investors about the contingency of speculative bubbles bursting that can prove damaging to portfolios, hence advocating continuous real-time monitoring of bubble occurrence, especially when prices are assessed to have moved away from fundamentals;
• The REIT investor receives evidence as to whether speculative bubbles in the REIT market are exacerbated by bubbles in the direct market and/or from bubbles in the general stock market. The degree to which bubbles are transmitting and influencing each other market is assessed;
• Bubble analysis presents investors with another tool to extract signals for the possible future trajectory of REIT prices and shocks from the direct market and the stock market.
We provide evidence on these issues from an examination of REITs, direct real estate and general stock markets in five European countries: the UK, France, Germany, Belgium and Sweden. The geographical dimension is important since pricing, even for the same asset class, is likely to follow different fundamentals and behavioural factors.
In price bubble analysis we need to provide evidence to two key questions:
• Can we ascertain the presence of speculative price bubbles across assets? This requires an estimation of fundamental or sustained (long run or equilibrium) values and comparisons with actual values through time;
• Are bubbles contagious? A statistical investigation of bubble spillovers between asset classes is conducted. A REIT investor will be particularly interested in the spillover of bubbles from the private market and the stock market to public real estate.
Our results show that the three asset classes in the selected countries have displayed signs of significant under- and over-valuation over the years, with some being more severe and perhaps more synchronised than others. For example, in the UK, the under/over-valuation trend in all three markets appears highly correlated, unlike Belgium. These findings also support the thoughts of numerous industry specialists that there was a growing bubble in many markets in the 2000s before the US sub-prime mortgage meltdown which led to a global financial crisis.
Although the investigation clearly shows that there have been prolonged periods of under/over-valuation in the countries considered, we need to statistically test for whether these deviations of prices from their fundamental values can be classified as speculative bubbles. A number of statistical approaches are available to do this. The van Norden-Schaller approach we use in this paper is an appropriate direct measure of speculative asset price bubbles.
Table 1 shows that bubble formation is established in most markets. Exceptions are REITs in the UK, direct property in France and the stock market in Germany. The absence of a statistically proven bubble in the UK does not mean that prices do not move away from fundamentals and that over-valuations are non-existent. This would simply imply that the over-valuation in the selected market may not necessarily be due to speculative investments as identified by a number of conditions that need to be satisfied under the van Norden-Schaller methodology.
Table 2 gives the probabilities of a price correction when the size of the bubble is measured to be 30% (actual values 30% away from the estimated fundamental values). The probabilities differ markedly by asset class and geography. All three asset classes considered, the largest probabilities are registered in the UK and the smallest in Belgium and Germany. Focusing on the REIT market, when prices have deviated by 30% from fundamental values, the highest probability of a price correction is estimated for the UK followed by France, with the lowest probability computed for Germany. This implies that investors move quicker in the UK and France and do not wait for greater deviations between observed and estimated fundamental values.
Contagion and spillover of price bubbles and over-valuations are a concern for investors. Bubble spillover to the REIT market from the stock and direct property markets is defined as a situation whereby under/over-valuation in the stock and direct property markets jointly influence the returns in the REIT market as well as the probability of its bubble to collapse. Even though some asset classes do not statistically exhibit signs of speculative bubbles, it is worth studying whether under/over-valuations in these markets (specifically in the stock and the property markets) are likely to influence the formation or the collapse of speculative bubbles in REIT markets.
There is clear evidence of a spillover from the stock and direct property markets to the REIT market in some countries – there is evidence that under/over-valuations in the stock and property markets transmit and jointly influence REIT prices as well as the probabilities of REIT bubbles collapsing. In the UK market, for example, there is strong evidence of a spillover from both the stock and direct property markets to the REIT market. In France, however, speculation in the commercial real estate market is unlikely to influence the dynamics of the REIT market, while its stock market does. This is similar for Belgium. In the German and Swedish markets, however, speculation in the commercial real estate markets has an effect on the dynamics of their REIT markets. Table 3 summarises the bubble spillover results.
Episodes of recurrent over-valuations that could underpin price bubble formation and subsequent price collapses and corrections require continuous bubble monitoring. In particular, ongoing measurement of the bubble size will provide better assessments of forthcoming risks to values. Our research identifies bubble formation both in the direct and REIT markets with the exception of UK REITs and the French direct market. However, over-valuations do occur in the latter two.
We find strong evidence of bubble transmission between asset classes, but the pattern is not uniform in the cross section of countries we investigate. The REIT market will not offer protection from bubbles both in the stock market and the direct property market in the UK. REITs will offer protection from over-valuations in the direct market in France and Belgium but not in Sweden and Germany. Interestingly, we have found that REIT prices offer protection from general stock market speculative price bubbles in Germany and Sweden, a result relevant to investors with no real estate securities in their portfolios.
Taking the results as a whole, the paper does recommend that the relationship between the direct market and the REIT market with regard to price bubble transmission is likely to be country-specific. We argue that the study of asset-bubble formation and transmission provides significant insight to mitigate forthcoming severe downside risks originating in the presence of speculative bubbles and should be part of the investor’s toolkit when investing in REITs.
Sotiris Tsolacos is chair in real estate finance and Ogonna Nneji is lecturer in finance at Henley Business School, University of Reading