EUROPEAN OFFICE MARKET CYCLES: AN EMPIRICAL STUDY OF LEAD-LAG RELATIONSHIPS
||EUROPEAN OFFICE MARKET CYCLES: AN EMPIRICAL STUDY OF LEAD-LAG RELATIONSHIPS
||Book of Abstracts: 13th Annual European Real Estate Society Conference in Weimar, Germany
||The trend toward the implementation of international portfolio strategies, which is already standard practice for capital market investments, has also become increasingly common for property investments over the last years. Due to the cyclical nature of real estate markets, the success of property investments depends highly on the investors’ ability to get the timing of investment and disinvestment in the individual property markets right. Therefore it is of vital importance for the investors to know about the correlation of cyclical movements of the individual real estate markets. As the cycles do not necessarily have to occur simultaneously in all markets, it is essential to account for lead-lag relationships between the markets. Due to the importance of the issues mentioned above for the development of a successful international real estate portfolio strategy, there seems to be the need for a comparative study that provides a better understanding of the interdependencies between the cycles of the different European office markets. The aim of this empirical study is not only to analyse the strength and direction of the relationships between the individual European property markets, but also to check for lags or leads in their cycles. The analysis is based on annual prime rental data (PMA) for the major European office markets for the years 1980 to 2005. At first, the prime rental series have to be decomposed into their trend and cyclical components. The decomposition is based on the time series models that NELSON AND PLOSSER (1982) developed originally for macroeconomic time series and that WANG (2001) adapted for the analysis of real estate market variables. Given the length of the time series, the Hodrick Prescott filter (HP-filter) that was developed by HODRICK AND PRESCOTT (1997) provides an appropriate technique to calculate the trend components of the logarithmized prime rental series. Afterwards the cyclical components can be derived as a residual. The cyclical components are analyzed for two overlapping periods from 1980 to 1995 and from 1990 to 2005 in order to answer the question whether the interdependence between the cyclical components of the individual markets has strengthened or weakened over time. The relationship between the different cyclical components is measured with the help of cross correlation analysis. As one major aim of the study is to detect lead-lag relationships between the European markets, the cross correlations are also calculated for lags and leads. The empirical study allows to answer a broad range of questions for the theoretical as well as for the applied analysis of European office market cycles: * As the analysis is undertaken for two periods of time, conclusions can be drawn concerning the question whether the cyclical movements of European office markets have converged over time, i. e. whether the integration of European real estate markets has grown stronger or not; * The consideration of lags and leads in calculating the correlations allows to identify markets which lead or lag the cycle or other markets which undergo the cyclical movements simultaneously. Again, the analysis for the two different periods shows whether the lead-lag structures are stable over time; * Furthermore, the correlation coefficients inform about the strength and direction of the relationship between the cyclical movements of the different markets. This is of interest with regard to risk diversification in the context of international portfolio investment. // As six German markets are part of the study, cross country and within country correlation results can be compared.
||office market cycle; convergence or divergence of market cycles; HP-filter; cyclical and trend component; cross-correlations with leads and lags; lead-lag relationships
Post discussion ...
These pages are best viewed with any standards compliant browser (e.g. Mozilla).