qotd: New evidence that individuals do not make sensible health insurance decisions

National Bureau of Economic Research
May 2015
NBER Working Paper No. 21160
Do Individuals Make Sensible Health Insurance Decisions? Evidence from a
Menu with Dominated Options
By Saurabh Bhargava, George Loewenstein, and Justin Sydnor

Abstract

The recent expansion of health-plan choice has been touted as increasing
competition and enabling people to choose plans that fit their needs.
This study provides new evidence challenging these proposed benefits of
expanded health-insurance choice. We examine health-insurance decisions
of employees at a large U.S. firm where a new plan menu included a large
share of financially dominated options. This menu offers a unique litmus
test for evaluating choice quality since standard risk preferences and
beliefs about one's health cannot rationalize enrollment into the
dominated plans. We find that a majority of employees – and in
particular, older workers, women, and low earners – chose dominated
options, resulting in substantial excess spending. Most employees would
have fared better had they instead been enrolled in the single
actuarially-best plan. In follow-up hypothetical-choice experiments, we
observe similar choices despite far simpler menus. We find these choices
reflect a severe deficit in health insurance literacy and naïve
considerations of health risk and price, rather than a sensible
comparison of plan value. Our results challenge the standard practice of
inferring risk attitudes and assessing welfare from insurance choices,
and raise doubts whether recent health reforms will deliver their
promised benefits.

Conclusion

Our principle contribution is to document widespread and costly
violations of dominance in the health-insurance decisions of a large and
diverse sample of U.S. employees. Unlike most prior work, our paper is
able to transparently assess choice quality without making assumptions
about risk preferences or beliefs, by analyzing employee choice from a
standardized plan menu with a large share of financially dominated
options. For the majority of employees who chose dominated plans—a
disproportionate share of whom were of low income—the adverse financial
consequences of poor choice outweigh the estimated benefits of recent
policy measures (e.g., the effect of the exchange on insurance premiums,
or the penalty associated with the individual mandate) that have
attracted far more policy attention.

Our initial results prompted us to address the question of why employees
make such disadvantageous decisions. Through a series of online
experiments with hypothetical choices modeled on the firm's plan menu,
we find evidence for a modest role of search/menu complexity and much
larger role of low health literacy. Confirming other research, we find
widespread deficits in employee understanding of the insurance choices
they face. Even with simple menus with few options, we observed
significant errors in plan choice, suggesting that plan choice is
difficult to improve. The online studies suggest that poor choice may
partly reflect a heuristic understanding of health insurance such that
consumers sort into plans based on perceived health risk and inferences
about plan generosity from the rank-ordering of plan deductibles rather
than a careful assessment of financial plan value.

The promise of recent reforms that expand choice and aim to increase
provider competition is premised on the assumption—challenged by our
research—that enrollees will make sensible plan choices. While efforts
to improve choice through simplification, education, and other
modifications to the choice environment may improve the quality of
decisions, our results suggest value in shifting focus from helping
consumers navigate complicated insurance options to simplifying and
standardizing the options themselves. At the extreme, if all firms
offered identical products, then competition would be far more likely to
focus exclusively on price and quality. Beyond implications for policy,
our study asserts that traditional attempts to infer risk preferences
from health insurance choices may be misguided. Rather than reflecting
rational deliberations involving cost, need, and risk, many health plan
choices likely reflect heuristic choice strategies grounded in a
fundamental deficit of health plan literacy.

http://www.nber.org/papers/w21160

****


Comment by Don McCanne

Regarding whether or not individuals make sensible decisions when
purchasing health insurance, these authors conclude, "many health plan
choices likely reflect heuristic choice strategies grounded in a
fundamental deficit of health plan literacy." That is, individuals do
not understand health plan options well enough to make sensible choices
on their own.

So much for the political campaign that led to the Affordable Care Act -
you remember, the campaign that called for CHOICE, when by choice they
meant choosing exactly the health insurance you want. As the authors
state, "We find these choices reflect a severe deficit in health
insurance literacy and naïve considerations of health risk and price,
rather than a sensible comparison of plan value." This raises "doubts
whether recent health reforms will deliver their promised benefits."

The authors further state, "At the extreme, if all firms offered
identical products, then competition would be far more likely to focus
exclusively on price and quality."

But we can go them one better. Suppose we had one identical
comprehensive product for everyone, and removed price as an issue by
funding care through an equitably-funded, universal risk pool. Then we
would all get the care we need, and select our care based on quality.

It may be that the perception of quality would be based on "heuristic
choice strategies" with a "deficit of health (quality) literacy," but
limiting decision options to only that of perceived quality, whether or
not the perception is valid, would result in a vastly superior concept
of choice than that which applies to the health plan marketplace -
whether the individual insurance market, the ACA exchanges,
employer-sponsored plans, or Medicare Advantage and Part D plans.

Obviously, under a single-payer improved Medicare for all, the only
choice that we would have to make would be our health care professionals
and institutions, and that choice would be based on perceived quality -
a much better deal for us all.