Previous applications of the Lucas (1978) representative agent framework, particularly Mehra and Prescott (1985), assume that the agent's consumption, in each period, is solely composed of income received from an initial endowment of an equity share. This paper examines the impact of incorporating an additional assumption of initial, non-equity endowment, aimed at improving the model's representation of the underlying valuation process The new formulation shows that the excess volatility puzzle is caused by the failure to take into account the recent change in corporate payout policy, from single channel (dividends) to multi-channel (dividends and repurchases). By resolving this issue it also shows, given sufficiently long and stable periods, that the representative agent framework can, using reasonable economic assumptions, accurately model the expected return and volatility of risky assets – in this case the equity market. However, the modified framework fails to resolve the equity premium puzzle. This failure is due to its inability to adequately model the risk-free asset and thus suggests that when market imperfections, such as borrowing costs, limited participation and incomplete markets - shown by previous research to explain the behaviour of the risk-free asset (Kocherlakota, 1996), are incorporated into this model, that the puzzle 'will be resolved. Thus this paper provides considerable support for the previously threatened foundations of modern financial economics, the Lucas-Breeden class of consumption-based asset pricing models.