Simulating Variation in Families' Spending across Marketplace Plans

Abstract

Objective: To examine variations in premium and cost-sharing across marketplace plans available to eligible families.

Data sources: 2011-2012 Medical Expenditure Panel Survey (MEPS), 2014 health plan data from healthcare.gov, and the 2011 Medicare Part D public formulary file.

Study design: We identified a nationally representative cohort of individuals in the MEPS who would have been eligible for marketplace coverage. For each family, we simulated the total out-of-pocket payment (premium plus cost-sharing) under each available plan in their county of residence, assuming their premarketplace use.

Data collection/extraction methods: Confidential state and county of residence identifiers were merged onto MEPS public use files and used to match MEPS families to the plans available in their county as reported in the publicly available data from healthcare.gov.

Principal findings: We found substantial variation in total family health care spending, especially premium component, across marketplace plans. This is true even within a plan tier of the same minimum actuarial value, and for families eligible for subsidies. Variation among families with income below 250 percent of the FPL is larger than variation among families with higher income.

Conclusions: Our simulations show substantial variations in net premium and out-of-pocket payments across marketplace plans, even within a plan tier.

Keywords: Health insurance exchanges; health insurance marketplace; simulation.

Figure 1

Variations in Simulated Spending across Plans Available to Each Family, by Plan Tier

1. Notes. Premiums and out‐of‐pocket spending are net of tax credits and cost‐sharing subsidies. Gray area denotes 25th–50th percentiles, and black area denotes 50th–75th percentiles. The three numbers shown on each bar are mean max, mean median, and mean min.

2. There are three reasons why the min and max values for one plan tier are not necessarily the same for all plans across all tiers. First, as also described above, what is represented on the graphs is not the absolute max (min) across all plans for all families, but the average across all people of the max (min) amount across plans available to that family. Second, not all families have access to all tiers of plans—for example, some regions did not offer platinum plans—so there is a composition issue in play as well. Third, in rare instances the min (or max) cost of plans available to a particular family in one tier might be higher than the next tier up. For example, there are cases where a particular company did not offer a bronze plan in a particular county, but its silver plan was cheaper than the plans offered by its competitors even in the bronze tier.

Figure 2

Variations in Simulated Spending across Plans Available to Each Family, Stratified by Family Income

1. Notes. (A) Families with income below 250 percent of the FPL. (B) Families with income between 250 percent and 400 percent of the FPL. (C) Families with income above 400 percent of the FPL. Premiums and out‐of‐pocket spending are net of tax credits and cost‐sharing subsidies. Gray area denotes 25th–50th percentiles, and black area denotes 50th–75th percentiles. The three numbers shown on each bar are mean max, mean median, and mean min.

2. FPL, federal poverty line.

3. There are three reasons to explain why the min and max values for one plan tier are not necessarily the same for all plans across all tiers. First, as also described above, what is represented on the graphs is not the absolute max (min) across all plans for all families, but the average across all people of the max (min) amount across plans available to that family. Second, not all families have access to all tiers of plans—for example, some regions did not offer platinum plans—so there is a composition issue in play as well. Third, in rare instances the min (or max) cost of plans available to a particular family in one tier might be higher than the next tier up. For example, there are cases where a particular company did not offer a bronze plan in a particular county, but its silver plan was cheaper than the plans offered by its competitors even in the bronze tier.

Figure 3

Variations in Simulated Premium and Cost‐Sharing across All Plans, Stratified by Premarketplace Total Family Health Care Spending

1. Notes. Premiums and out‐of‐pocket spending are net of tax credits and cost‐sharing subsidies. Gray area denotes 25th–50th percentiles, and black area denotes 50th–75th percentiles. The three numbers shown on each bar are mean max, mean median, and mean min.

Figure 4

Variations in Simulated Premium and Cost‐Sharing across Silver Plans among Families with Income below 250 percent of the FPL, Stratified by Premarketplace Total Family Health Care Spending

1. Notes. Premiums and out‐of‐pocket spending are net of tax credits and cost‐sharing subsidies. Gray area denotes 25th–50th percentiles, and black area denotes 50th–75th percentiles. The three numbers shown on each bar are mean max, mean median, and mean min.