Risk parity
Why do risk parity portfolios hold more of a quiet asset than a wild one, and what happens when the wild one moves anyway?
In plain words
Risk parity is a way of sizing holdings so that each one contributes a similar amount of risk to the whole portfolio, rather than each one receiving the same dollar amount. A more volatile asset gets a smaller weight, and a calmer one gets a larger weight.
Why it matters
Equal dollar weights are rarely equal risk. A portfolio that spreads money evenly can be dominated by its single most volatile holding, so most of its gains and losses come from one place. Risk parity aims to make risk contributions more even, but it depends on measuring volatility and correlation, both of which can jump.
In a real week: ETH, week of 11 June 2018
In the week of 11 June 2018, ETH had an annualised realised volatility of 108.2%, a weekly range of 20.0%, and a weekly return of -5.24%, with its largest one-hour move at 8.64% on Thursday at 16:00 UTC. Its hourly correlation with BTC was 0.84, labelled close, and the ETH/BTC ratio changed -0.61%. A risk parity weighting would hold far less ETH than BTC, given ETH's volatility, while the close correlation limits how much that smaller weight actually diversifies.
A common mistake
Treating a risk parity weighting as fixed. Volatility and correlation are re-estimated regularly, so weights drift, and a calm asset can turn volatile before the weights adjust.