Traditionally, the practice of asset management has been dominated by techniques relying on volatility as a risk measure. Although the academic community has been aware of shortcomings of such approaches, different alternatives have not gained much traction in the industry, mostly due to difficulties in their usage and lack of intuitive interpretation. In this talk we look at the problem of asset allocation in the Equal Risk Contribution (Budgeting) framework of Litterman and, building on the earlier work of Rocalli and co-authors, we investigate applications of the Expected Shortfall as a risk measure instead of volatility. We show that, when combined with tractable and intuitive Gaussian Mixture Models for asset returns, Expected Shortfall provides an effective way to take into account negative skewness present in some of the popular underlyings (in particular some Alternative Risk Premia strategies), thus addressing one of the main weaknesses of volatility-based portfolio allocation. We also discuss limitations of this approach and list some further research directions and open questions.