"Our research wants to establish links and patterns over time between the constituent components of the cap rates, and between real interest and risk premium in particular. We developed a statistical model to predict and explain the risk premium asked for by investors buying properties throughout Europe, with the aim to having reliable risk premium forecasts in the years ahead. The research reveals that both an univariate and a multiple linear regression model appear to be significant (satisfactory) models, predicting the risk premium from one explanatory variable (the real interest rate) or from two explanatory variables (the nominal interest rate and inflation). Moreover, dummy variables were introduced to deepen the results and to highlight the sensitivity of the risk premium to the kind of properties asked for (offices or retail), and to the various geographical regions properties are located (Holland, Belgium, Frankfurt, London). A main conclusion is that the risk premium seems to be used by investors to ëadjust' the initial yields at which they buy or sell properties. The risk premium appears to have a corrective effect, keeping the cap rates at a rather stable level. This is in contrast to the more common proposition that the risk premium would be a compensation for the increased level of risk of real property. Furthermore, risk premium may highly vary over time: from 0 to over 600 basis points, well below or above the ""assumed"" 200 basis points. "