Was forwarded an interesting article from Bloomberg the other day about some of the struggles in risk parity strategies, particularly Bridgewater's All Weather.
This highlights the one glaring flaw in this approach---you are using past returns, correlations, and volatility to set allocations. Interestingly, this is the same reason why Modern Portfolio Theory does not work.
To us risk parity can be improved in three ways:
1. Use maximum drawdown instead of standard deviation as a risk measure. Nothing is perfect but clients tend to view risk more as drawdowns from a high water mark than day to day and month to month volatility.
2. Use methodologies instead of asset classes. You can get a better idea of what to expect from a methodology, say momentum or counter trend, than you can from an asset class.
3. Using the above two methods can avoid having large fixed income exposure reducing or eliminating the need for leverage.