The term structure of interest rates describes the relationship between short- and long-term rates and embeds the market's expectation of future interest rates. This has led to a large literature concerned with modelling the term structure and hence attempting to extract this information. This thesis is concerned with both modelling and forecasting the UK term structure, with a focus on the application of density forecasting and decision-based forecast evaluation. We test the Expectations Hypothesis of the term structure and more generally, examine if the term structure is best described by a statistical or theory informed model. Interest rate forecasts are essential for policymakers and practitioners alike. Since density forecasts provide the entire distribution about the forecast, we argue that they are appropriate for an investor concerned with the uncertainties about future asset returns. We find economic theory to have explanatory power for the term structure and the UK money market to be consistent with the Expectations Hypothesis. Further, we demonstrate how density forecasting techniques can be applied to forecast asset returns and inform portfolio allocation decisions; and how these optimal allocations are sensitive to the forecast uncertainties about the expected future returns and the assumptions made regarding return predictability. Furthermore, given the importance of forecast evaluation, our results highlight the need to judge forecasts in the decision making context for which they are ultimately intended. That is, our findings advocate the use of decision-based criteria that assess forecasts from the user's perspective, i.e. in terms of economic value, rather than conventional statistical measures. Under decision-based methods, we find that the investor may gain in terms of wealth by assuming returns are predictable and using a theory informed model to forecast. In short, we find economic theory to be significant for both modelling and forecasting the term structure.