Last week, I had two Health Affairs blogs published on the ways that the ACA choice environment is complex and confusing and potential administrative solutions. I got quite a bit of good feedback and I want to respond to some of it.
What about agents?
Insurers don’t like putting people in plans they leave, but I’m skeptical about a CMS algorithm, and the time it takes to print and mail consumer-level notices might be a couple of weeks depending on time, complexity, and size.
— Gabriel McGlamery (@jgmcglamery) June 4, 2021
YEP! Agents are important. A good agent is an expert who can sort through a lot of confusing and conflicting features quickly and effectively. A lot of research has found that attention is the constrained commodity in the decision process. Individuals who make active choices with attention are more likely to make better choices than people who are inattentive for whatever reason. Agents can be a form of outsourced attention. However, given the way that agents are paid, either a flat Per Member Per Month or as a percentage of premium, the one-off non-iterated incentive is to sell something instead of sell the best thing. We probably need to think about how agent compensation interacts with dominant plan placement. More importantly, not everyone has an agent or a navigator or anyone else who can pay expert attention to the choice problem, so improving defaults to get rid of absolutely horrendous options is likely welfare improving.
FIX THE GLITCH. While I agree that dominated plan choice is a consumer concern that needs to be addressed…it’s pretty low on my priorities. Let’s get more working people subsidies then deal with the outliers
— Jenny Chumbley Hogue (@kgmom219) June 4, 2021
I disagree. Improving defaults and fixing the glitch are not mutually exclusive options. They are not in competition for constrained Congressional attention and they are in at most modest competition for CMS/regulator attention. These can be done at the same time. Both should be fixed and if for one reason or another the family glitch is not fixed, there is still a significant welfare improvement of improving defaults.
Nice piece. Will/should health plan quality ratings be part of a logic to reduce the complexity burden?
— Lori Suzanne (@lori__suzanne) June 5, 2021
I am extremely reluctant to even think about including quality ratings with any default changing options. I am comfortable about changing peoples’ plan defaults when the changes are within the context of both revealed preference (there is a reason a person chose Insurer X and Network A, so we keep them in Insurer X and Network A) and fully observable characteristics of plans. We can see that Silver (X,A) has a $5500 deductible/MOOP and a $250/month premium, while Gold(X,A) has a $2,000 deductible/MOOP and a $225/month premium and say that if we hold previous year choices to choose X,A constant, then the switch is welfare enhancing while if a person discovers that they hate insurer X and like Insurer Z, they’ll be making an active choice and revealing new preferences anyway. I feel confident about this logic.
Quality ratings in ACA plans have a couple of big problems. First, they are composites that are measured at the insurer/state level. An insurer can be amazing at behavioral health and horrendous at preventive care measures and get a decent rating. Another insurer can be horrendous at behavioral health and really good at preventive care measures and pharmacy management and get a similar rating. For some people, they want the behavioral health to be awesome, and other people want the preventive care stuff to be awesome while being completely indifferent to other measures. We can not extract a clear statement of preference and then premium/quality trade-off at the individual level if we’re switching people between insurers.
Secondly, we don’t really know if the quality measures are a strong signal (after removing selection from the equation) with actual improvements in morbidity and mortality. Abaluck et al have found that Medicare Advantage quality ratings are not tied to causal changes in mortality. More importantly, they have found that some plans are much better than other plans at reducing death. We don’t have that framework for Exchange plans.
I think that improving the choice universe and choice experience of the ACA individual marketplace is important. I know both HC.Gov and the SBMs are trying hard and trying creatively to make decision-making easier, more transparent and nudged towards improved outcomes, but this is tough if our expectation is that this is an individual level problem.