European Property Stocks Trade In Reversal Patterns; Country Indices Follow Momentum

Often investors try to increase their portfolio performance by playing specific stock patterns, which has triggered us to do some in depth research in this area. We have constructed a (short-term) trading model that explores trading patterns in European property stocks and real estate country benchmarks. We find that, historically, the worst performing companies of the last month tend to be the best performing companies of the coming months. And indeed, the opposite is also often true: the best performers of the last month are often the worst performers of the coming months. We believe the results shown by our model are highly significant, as they strongly pass the z-test (for outperformance) and t-test (for consistent outperformance over time). Our model indicates that, depending on transaction costs, the optimal risk-return portfolio consists of one to five stocks, which will be held for one to four months. Without transaction costs the strategy to buy each month the worst performing stock would have resulted in the highest outperformance of 2456% over the ten year period 1995-2005 (CAGR 38%). The optimal strategy with 1% transaction costs is to hold one stock for three months (834% outperformance or CAGR 25.0%). In all cases the stocks should not be held longer than six months, as the outperformance diminishes. These findings clearly point to reversal patterns in European property stocks and indicate, we believe, that one should adopt a contrarian strategy to take advantage. We find opposite results if we exchange the companies in our model for the FTSE/EPRA country indices, ie real estate country benchmarks. Now we find that the contrarian strategy does not work, but that one should opt for a momentum strategy. Our model indicates that the worst performing country of last month also significantly underperforms in the following months. On the other hand, the best performing country of last month is still significantly outperforming one month later. After that, the country outperformance diminishes during the following month, but tends to return one month later. The model indicates that the optimal strategy for country allocation is to buy the three best performing countries and hold those for seven months. We believe these findings could be of use to European fund managers. Taking into account these stock price reverting and country momentum patterns an investor in European real estate stocks could benefit in several ways by: 1. buying the underperformers and selling the outperformers of last month; 2. delaying a purchase in an outperformer and sale in an underperformer; 3. constructing a portfolio at zero cost by shorting the outperformers and buying the underperformers; 4. (on a country level) ignoring the worst performing countries, while holding on to the three best performing countries for seven months. // Our research contributes to the existing literature. Current research on momentum strategies is focused on the US REIT market, and to our knowledge there has been no research published on the listed European property market. While some US articles also find the existence of reversal patterns in the US REIT market, we believe our findings to be more compelling, as US-REIT share price patterns are found by applying filters: for example, buy the stocks with negative return or large share price drops (>5%), or that are not profitable if transaction costs and risk are taken into account. The model we studied always selects a pre-determined number of companies. We have also found one paper that concludes that momentum, not contrarian, strategies work in the US-REIT market. Furthermore, a recently published article on the UK direct property market finds that momentum strategies should be applied to achieve outperformance. We come to a different conclusion for listed companies, but more or less the same conclusion for country benchmarks based on listed securities. We constructed a model that, each month, ranks 28 European property companies in terms of performance over the ten year period 1996 to 2005. We believe that these 28 stocks are a fair reflection of the listed European property market. As all of our stocks currently still exist, we have not included takeovers in our sample. The user of the model can select the number of stocks in the portfolio, the holding period, the expected transaction costs and whether he/she wants to buy the best or the worst performers of last month. The performance of the portfolio is measured against a market capitalisation weighted. We have used total return data from Datastream. The portfolio returns are tested against the benchmark return for outperformance with the z-test and for consistent outperformance by regressing the monthly benchmark returns on the portfolio returns. We tested with the t-test whether the constant of the OLS regressing was significant or not. In order to get a feeling for the portfolio turnover we calculated the average of the monthly number of portfolio changes divided by the total number of possible changes. Next we calculated the standard deviation (risk) of the portfolio and benchmark return and divided the compounded annual return by the risk to determine the risk-return profile of the strategy. We believe that a portfolio is ëbeatingí the benchmark if it not only outperforms, but also passes both statistical tests and has a better risk-return profile. We included transaction costs (0.5% and 1%) for both buying and selling the security in order to assess whether one benefits from these trading strategies. We used the same approach for the country strategies as for companies. We included 13 countries over ten years in our sample.