It is common practice to employ returns, price differences or log returns for financial risk estimation and time series … transformations. We forecasted risk (volatility) and price value and compared the results of all models using original, unmodified … prices. From the results, models showed that, on average, a logarithmic transformation achieved better volatility predictions …

estimation windows. It develops theoretical results for random walks when their drift and/or volatility are subject to one or …This paper considers forecast averaging when the same model is used but estimation is carried out over different … more structural breaks. It is shown that compared to using forecasts based on a single estimation window, averaging over …