The aim of the paper is an investigation on the reliability of historical returns for the Italian property market, where the quality of information seems not standardized. In Italy, such as for many other countries, the returnsí indices for direct markets are provided by several sources that differ among them in terms of methodology adopted (appraisal-based vs transaction-based approaches) and in term of indexís composition. These differences produce a lack of informative standardization that could negatively affects the predictability of market and that can be explained by a strong real estate marketís fragmentation, as well as informative and marketís organizational inefficiency. The implications arising from the existence of not homogeneous indices involve the main market operators such as: real estate funds (referring to their IRR forecasting ability); banks (referring to the estimation of LGD for mortgage loans); investors (referring to their investmentís choices). By using the historical returns provided by four index provider (the period is 2002-2008, returns are quarterly), we submit data to some quantitative tests, designed to investigate the following topics: 1) the smoothing degree of returns caused by appraisal based methodologies 2) the lack of a sufficient degree of homogeneity among the returns of all the property indices. The smoothing effect is investigated through the implementation of an ARIMA process, while the reliability of the dataset is analyzed by comparing the distributions of returns following a stationarity approach. Our results seem to underline a significant level of smoothing of the returns for the main sectors of market (residential, commercial, office) and a divergence of the quality among the sources of data that appear not consistent with an efficient market. Finally, we give an example of how the lack of standardization of real estate database affects the LGDís banks forecast process and the expected IRR of real estate funds.