Improving Consumer Choices on Health Insurance Marketplaces in the United States

Jon Kingsdale
Fieldwork by:
Colorado, United States of America
15,000 individuals enrolled in an insurance plan purchased through Connect for Health Colorado
2015 - 2016
Outcome of interest:
  • Provider Performance
Intervention type:
  • Information
  • Health care delivery
  • Insurance
  • Nudges and reminders
AEA RCT registration number:

The Affordable Care Act dramatically expanded the use of regulated Marketplaces to provide individuals with health insurance coverage. Researchers studied the impact of providing personalized and generic information about potential savings via letters and e-mails on enrollee health plan choices.

Policy issue

Since the creation of the health insurance Marketplaces by the Patient Protection and Affordable Care Act (ACA) in 2010, an increasing number of Americans without employer-sponsored insurance rely upon the Marketplaces to purchase insurance. The Marketplaces seek to provide affordable, high-quality insurance by relying on consumer choice. In theory, consumers are able to enroll in plans that reflect their preferences, which places pressure on insurers that raise premiums too much or have overly limited provider networks. However, consumer choice may generate limited competitive pressure in the Marketplaces if consumers choose to stay enrolled in default plans instead of switching to lower cost plans. Policymakers have sought strategies to encourage consumers to shop for coverage and thereby strengthen insurer incentives to deliver value.

Context of the evaluation

Colorado is a medium-size state, with a diverse population of 5.5 million living in both large urban and rural communities. Connect for Health Colorado, the state-operated health insurance Marketplace, offers a more robust set of plan choices than in most other states. In 2015, consumers in Colorado could choose from over sixty different plans through ten insurance carriers, compared to the federal Marketplace, which offered only forty-six health plans in the same year. Additionally, compared to the national average, enrollees in Colorado are more likely to have bronze tier plans and less likely to receive subsidies via tax credits or cost sharing reductions.

Prior research on Medicare Advantage and Medicare Part D suggests that interventions that provide consumers with information to help them choose plans in their best interests induces shopping and plan switching. However, there have been no evaluations to assess how such informational “nudges” affect consumer behavior on the Marketplaces established by the ACA.

Researchers studied the impact of providing personalized and generic information about potential savings on enrollee health plan choices. Photo:

Details of the intervention

Researchers conducted a randomized evaluation to test whether sending generic and personalized information about the potential to save on insurance premiums could encourage enrollees to shop on the state Marketplace website and switch plans. During the 2016 open enrollment period, researchers randomly assigned approximately 15,000 households eligible for automatic reenrollment in a health plan purchased through Connect for Health Colorado to receive one of three information interventions:

  • “As-usual” control group: Households in this group received only the usual communications sent by Connect for Health Colorado.
  • Generic communications: In addition to “as-usual” communications, households in this group received two letters and two emails with generic information about the potential to save from switching plans and encouragement to visit the Marketplace website or phone its call center to shop for a plan.
  • Personalized communications: In addition, to “as-usual” communications, households in this group received the same two letters and two emails as the “generic” group with the addition of personalized information about the premium savings the households could achieve if they switched plans, including the subsides they could expect to receive, relative to what they would pay if they chose to stay enrolled in their current plan.

To measure the impact of the letters and emails, researchers used administrative data from the Marketplace to track household shopping activity and plan selection.

Results and policy lessons

Impact on Shopping:
Both personalized and generic communications increased the rate of shopping for health insurance plans on the Marketplace website by about 6 percentage points from a base of 25 percent (an increase of about 23 percent). Personalized information about potential household savings had no more effect than a generic reminder.

Impact on Plan Switching and Premium Paid:
Neither the personalized nor the generic communications caused any detectable impact on health insurance plan selection. These interventions did not prompt consumers to switch plans or save on premiums. One potential explanation is that the intervention did not provide a specific recommended lower-cost plan for consumers to switch to, but rather directed consumers to the Marketplace website and call center to research this themselves. A prior randomized evaluation of an intervention targeting enrollees of a Medicare Part D prescription drug plan, which specifically recommended the lowest cost plan, increased switching rates by 11.5 percentage points over the control group’s level of 17 percent. Additional explanations of the absence of impact on plan switching and premiums include exclusion of some of Colorado’s most price-sensitive consumers, potentially lower levels of motivation to save money amongst study participants, and choice overload due to the high number of plan options on the marketplace.

Policy Implications:
Simple “nudges” with even generic information can encourage consumers to actively shop for a health insurance plan. Why consumers renew more expensive plans and how to best facilitate plan-switching remain open questions.