Testing the Capital Asset Pricing Model on Aggregate Data: with Special Reference to the United Kingdom

This paper utilizes the Sharpe-Lintner-Mossin-Black (SLMB) Capital Asset Pricing Model (CAPM) as a framework for studying the relationship between aggregate private sector portfolios and asset returns in the United Kingdom. Previous studies in the US have provided only partial support for the SLMB CAPM in this context; UK studies have generally rejected the model. Utilizing a comprehensive new monthly dataset covering 1972-85, we investigate possible reasons for the failure of the SLMB CAPM in the UK. We seek econometric evidence for possible violation of each of the three main assumptions of the model: one-period optimization; constant first and second moments of returns; and perfect markets. We concentrate particularly on the assumption of perfect markets as this has not previously been investigated in empirical work. We utilize a wide range of nested and non-nested tests for this purpose. We find that UK data are not consistent with a very simple model of certain market imperfections, but there is still ample evidence that market imperfections of several kinds have played an important role in determining UK asset returns. The assumptions of one-period optimization and constant first and second moments also appear unjustified. We indicate the direction that future research should take to improve our understanding of asset price formation in the UK.