The choice of unsmoothing coefficients for real estate appreciation indices : with notes on correction procedures, appraisal-smoothing, nonsynchronous appraisal, asset return correlation and market efficiency

Due to the smoothing problem with appraisal-based indices, there is uncertainty about the true volatility of value change returns on real estate markets. The selection of unsmoothing procedures proposed in the literature obviously cannot remedy this. The article argues that it is hardly possible to uncover historical latent market values and volatilities, as various sources of error in the compilation of indices may overlap and change over time. These so-called smoothing phenomena are mentioned in various sources in the literature and are discussed in more detail here. The correction methods proposed in the literature are also discussed critically and it is explained that many of these in essence contain a simple correction formula already presented by Blundell & Ward. The results of calculations based on this correction formula and on an innovative approach by Bond & Hwang are therefore presented in the paper. The article suggests not relying on a point estimate resulting from the application of a selected correction procedure. Instead, an attempt could be made to narrow down plausible value ranges of the time series characteristics (volatility, mean value and asset correlations) relevant to investment decisions. Using the NCREIF Appreciation Index, the paper shows how the volatility of the corrected index returns changes with the level of the unsmoothing coefficient. To find plausible value ranges, it could be considered, for example, that high transaction costs are already an obstacle to market efficiency in real estate markets. The paper therefore illustrates that volatility is already substantially reduced with a low entry of serial correlation in time series.