Standard consumption-based models typically fail in pricing asset returns. In a famous seminal paper, Mehra and Prescott (1985), using a standard consumption model, prove the presence of a puzzle (i.e. equity premium puzzle). The recent financial literature still has to provide a convincing resolution to the well known puzzle. In contrast to this literature, which mainly focuses on the United States data, our paper simply replicates the closed form solution estimation, as in Mehra (2003), for a bunch of developed and emerging markets. On one side, our estimations confirm the existence of the puzzle and lead to bizarre values of the coefficient of relative risk aversion. On the other side, we claim that the key consumption model assumptions, the choice of a proper riskless asset and the lack of data, generate obstacles in finding robustness in the estimations of the CRRA coefficients, both in developed and emerging markets.