My parents' retirement fund consists largely of an apartment building, a substantial portion of which is rented by section 8 (ie low income) tenants. The variance in cash flow is pretty dramatic, from both an expenses and a non-payment of rent perspective. Tenant protections are robust enough that, more than once, they've ended up in a situation where it takes four months of missed rent + legal fees before a non-paying tenant could be evicted, which is a cool $10k chopped right off of their income. To say nothing of the occasional expenses you wouldn't see in higher-income properties: the same tenant protections mean that units can become almost hazardously filthy before the landlord can intervene, and when they leave, they obviously don't tend to clean up.

Higher rents effectively function as an insurance policy for these expenses, and your insurance policy is obviously going to overcover rather than undercover. Higher risk (ie variance) investments mean higher returns, for the simple fact that the value of a high-variance investment will be lower than that of a low-variance one with a given expected return, for obvious reasons.

The only reason to be ignoring variance If your goal is shitty advocacy instead of intellectual honesty, which, as others have pointed out, is obviously the case with the authors of this study.