Several studies advocating safety first as a major concern to investors propose downside beta risk as an alternative to the traditional systematic risk-beta. Downside measures are concerned with a subset of the data and therefore the results in the studies that consider the downside beta only may be biased. This study addresses this issue by including downside co-skewness risk in addition to the downside beta risk in the pricing model. In a sample of 27 emerging markets two-stage rolling regression analysis fails to support pricing models with downside risk measures. In a cross-sectional analysis inclusion of downside co-skewness improves model fit. When considered together, downside beta is potential and downside co-skewness is a risk to the rational investor. Even though our results are inconclusive the evidence strongly suggests a need for further investigation of co-skewness risk in pricing models that adopt a downside risk framework.