An analysis of the Solvency II regulatory framework’s Smith-Wilson model for the term structure of risk-free interest rates
研究了欧盟偿付能力II监管中强制使用的Smith-Wilson利率模型，指出其参数设定（尤其是高终极远期利率）可能人为抬高利率曲线，导致寿险和养老金负债被严重低估，与监管目标相矛盾。
In the European Union financial regulation requires that life and pension (L&P) companies use the Smith and Wilson (2000) model for the term structure of risk-free interest rates when valuing their liabilities and long term guarantees. Some key features of this model are that it allows for a perfect fit to market observed bond prices, and that its extrapolated long rates converge towards a constant level, the Ultimate Forward Rate (UFR). Both this level and the rate at which convergence towards it takes place are directly specified via parameters of the model. Since the Smith–Wilson model is not one of finance theory's standard term structure models, we introduce the model and summarize its most important mathematical properties. We also describe how the European Solvency II regulation came to embrace this particular model. The paper moves on to document how the regulation also imposes quite detailed and tight restrictions on how the Smith–Wilson model should be parameterized and applied. We argue that many of these implementation instructions – one of which is the regulator's specification of a very high UFR – seem biased in the same direction and that this could indicate a systematic attempt to “lift” the term structure curve up and away from its true location whereby artificially high discount rates are induced. The result of the bias is not only significant undervaluation of L&P liabilities but also a peculiar contradiction of the Solvency II overall objective of enhancing financial stability and of protecting policyholders via the promotion of economic valuation in accordance with market consistent principles. The paper's analysis is accompanied by valuation illustrations based on data on the liability composition of an actual medium-sized Danish pension fund. The results suggest that the undervaluation of liabilities resulting from use of the regulation compliant Smith–Wilson model can be massive compared to results obtained from some alternative and more freely calibrated models.