Affordable Care Act Enrollment in Texas After Rating Area Adjustments

ABSTRACT

Objective: To evaluate the association between Texas Affordable Care Act rating area change and health plan enrollment, plan selection, and premiums from 2022 to 2024 for urban and rural counties.

Study Design: Texas integrated a rating area consisting of all 177 rural counties into nearby metropolitan rating areas in 2023. We analyzed this policy using enrollment data from the Marketplace Open Enrollment County-Level Public Use Files from 2022-2024.

Methods: We calculated the growth in enrollment across rural and urban counties and estimated linear regression models to understand whether enrollment grew faster in rural counties than in urban counties after the policy change.

Results: Total marketplace plan enrollment increased by 80% (95% CI, 70%-90%) in urban counties and 76% (95% CI, 68%-84%) in rural counties. Urban and rural counties experienced the largest growth among enrollees aged 35 to 44 years (urban: 107%; 95% CI, 94%-119%; rural: 103%; 95% CI, 95%-112%) and enrollees with incomes between 100% and 150% of the federal poverty level (urban: 124%; 95% CI, 106%-142%; rural: 116%; 95% CI, 106%-127%). The share of counties reporting gold plan enrollment increased in urban and rural counties from 70% to 95% and 51% to 93%, respectively. Rating area changes were not associated with differential enrollment changes across rural and urban counties.

Conclusions: We found similar growth rates in enrollment for rural and urban counties. Marketplace enrollees were more likely to choose a gold plan, suggesting that they shifted away from less-expensive bronze plans.

Am J Manag Care. 2025;31(6):In Press

Takeaway Points

• Geography is an important aspect of insurance market design. Evidence suggests that individual marketplace rating area geographic definitions affect access to affordable insurance.
• Analyzing data from Texas, we found similar growth rates in enrollment for rural and urban counties.
• Integrating rural counties into metropolitan rating areas improved the plan selection for rural marketplace enrollees.

The Affordable Care Act (ACA) established regulated marketplaces to provide affordable access to health insurance.1-3 However, access to affordable health insurance depends on effective risk pooling as adverse selection can drive up premiums and, at its worst, limit insurer participation altogether.4-7 One way to ensure proper risk pooling is to select balanced geographic areas that make up an insurance rating area with many healthy individuals eligible for coverage. Different approaches in setting up ACA rating areas have led to varied performances of the individual marketplace in terms of premiums, plan choice, and insurer participation.8-12 Historically, the default approach to defining ACA rating areas in Texas was the number of metropolitan statistical areas (MSAs) plus 1 catchment area capturing all non-MSA rural counties, called MSA+1. This approach has been used in states such as Alabama, New Mexico, and Virginia.13

One important dimension affecting individual marketplace performance is the division of rural and urban areas (and thus consumers) into separate rating areas and, as a result, insurance markets. In rural markets, access to insurance has historically been more limited relative to urban markets due to both limited insurer participation and health plan options, and sometimes higher premiums relative to urban markets, especially in the absence of subsidies.14-20 Although some of these differences can be explained by differential health needs, lower population densities of rural counties may impede creation of a well-distributed risk pool of insured individuals, and provider shortages may make the formation of suitable networks more costly.20-23 This is especially important for states with large rural populations, such as Texas, where 27 rural counties were served by only 1 insurer in 2022.

Prior to 2023, Texas utilized an MSA+1 approach that established 25 rating areas solely based on county MSAs and an additional catchall category for all remaining rural counties outside of MSAs, for a total of 26 rating areas. To address insurance access concerns in rural Texas counties, state officials revamped the allocation of rating areas from a catchment area (MSA+1) approach that created 1 rating area for all 177 rural counties to a more homogenous approach in which adjacent rural counties were integrated into nearby urban rating areas starting in 2023. The change in rating areas for rural counties shifted all 177 rural counties and their 184,000 enrollees in 2022 from the single rural rating area into different adjacent urban rating areas to form 27 new mixed urban-rural rating areas. The stated goals of adjusting the rating areas were realignment of rural counties with major tertiary networks, lower premiums for rural residents that accurately reflect health care costs in each county, and expansion of coverage and choice in rural counties.12 However, depending on the composition of the health risk profile in these counties, the integration of the higher-risk rural population with urban residents may lead to more adverse selection in the newly formed rating areas, which in turn may lead to less plan choice and higher premiums, especially when the rural population represents a large share of the market.

In this study, we assessed whether the integration of rural counties with nearby urban counties in Texas into new ACA rating areas was associated with differential changes in overall enrollment, the income and demographic composition of enrollees, and plan choice and premiums in rural and urban counties before and after the rating area change. Directly comparing changes in enrollment post rating area change between urban and rural counties allowed us to partially account for a concurrent introduction of the temporary plan subsidies that drove a large increase in marketplace enrollment across Texas.

Taken together, our findings shed light on associations between ACA rating areas and sociodemographic characteristics of enrollees by county rural-urban status, in addition to their metal tier selection and premiums. This study builds on previous literature by adding enrollment data and highlighting differential changes across rural and urban counties.12 We hypothesize that the integration generally led to increased access to affordable health plans for rural residents, although it may have also led to overall higher premiums in the new rating areas for urban residents.

METHODS

The primary source of data for our analyses was the Marketplace Open Enrollment County-Level Public Use Files from 2022-2024, which were published by CMS. These files include information on county-level enrollment; plan selection relative to the previous year (conditional on selecting a plan in the prior year); plan type (metal level); mean monthly premiums (before and after advanced premium tax credits [APTCs], which quantify tax subsidies received by eligible enrollees); and sociodemographic characteristics of enrollees, including age, sex, self-reported race and ethnicity, and household income as a percent of the federal poverty level (FPL).

We classified Texas counties into the 26 rating areas in 2022 and the newly created 27 rating areas in 2023-2024 (post period) using publicly available information from the Texas Department of Insurance. A limitation of the data was that some elements were not reported if enrollment in a county was below 11 in a given year. For this study, we described enrollment characteristics of only counties reporting data every year between 2022 and 2024. Generally, we observed enrollment information for all 77 urban counties and 85% to 90% of the rural counties, depending on the sociodemographic subsample enrollment count. In the eAppendix (available at ajmc.com), we display county characteristics for all counties that reported data in all years or just in 1 of the years. Rural counties were defined as counties that were part of rating area 26 (the catchall rating area) in 2022.

When we evaluated whether enrollees shifted toward higher actuarial value plans in rural counties, we utilized information from all counties to evaluate a shift in the distribution of enrollees’ plan selection. Specifically, we created binary variables equal to 1 if the county reported enrollment counts for a plan type (bronze, silver, gold) and 0 if the count was missing due to CMS suppression rules. We then compared trends from before with after the rating area change for each plan type. Additional information on CMS suppression rules is provided in the eAppendix.

Analysis

We used descriptive analyses comparing enrollment, plan type, and enrollee characteristics for the rural and urban counties before and after the change in rating area. All enrollment outcomes of interest are ratios, where the numerator is the county-level count of marketplace enrollees of a particular type and the denominator is total county population of residents aged 18 to 64 years. We normalized rates per 1000 residents. We created mean growth rates from the period before the rating area change (hereafter, pre-period) to the period after the rating area change (hereafter, post period) for each county and presented a mean of county-level growth rates. We then utilized standard t tests to evaluate whether the mean growth rate was different from 0. When reporting the enrollee shares of sociodemographic groups, we used standard t tests to determine whether the mean enrollee shares changed from the pre-period to the post period.

To compare the effect of the change in the rating areas on rural counties relative to urban counties, we estimated the following regression model:

yct = βRural × Postct + γInsurerct + αt + αc + εct,

where yct is the outcome of interest for county c in year t. The main independent variable of interest, Rural × Postct, is a time-varying indicator variable that is equal to 1 for all rural counties after the rating area change (2023-2024) and 0 otherwise. β represents the mean associational effect of changes in the rating areas for rural vs urban counties after the policy change; αt are year fixed effects that account for year effects; αc are county fixed effects, which account for time-invariant unobservable differences across counties in our sample; Insurerct is the number of insurers offering marketplace plans in county c and year t, which serves as a proxy for market competition and could affect plan offerings, pricing, and enrollment; and εct is the error term.

We performed subsample analyses by rating areas that experienced a large influx of enrollees from rural counties relative to their enrollment in the urban county based on 2022 enrollment counts. To do so, we explored whether rural counties with larger growth in marketplace enrollment experienced differential effects of the change in rural rating areas on premium. We re-estimated the earlier equation where the independent variable of interest is a binary variable equal to 1 for rural counties with above-median growth in ACA marketplace enrollment in the post period and 0 for rural counties with below-median growth in ACA marketplace enrollment in the post period. Statistical significance was determined at the α = .05 level; all analyses were conducted via Stata 16 (StataCorp LLC).

RESULTS

Table 1 shows the mean enrollment characteristics per 1000 residents in urban and rural counties before and after the rating area change. eAppendix Tables 1 and 2 report results for all counties reporting data in 1 or both years. The average urban county had 95 enrollees with a marketplace plan in 2022 per 1000 residents, whereas rural counties had 105 enrollees with a marketplace plan per 1000 residents. Total marketplace plan enrollment increased from the pre-period to the post period by 80% (95% CI, 70%-90%; P < .001) in urban counties and 76% (95% CI, 68%-84%; P < .001) in rural counties. Most of the enrollees were returning customers: In urban counties, the return rate was 72% in the pre-period and 72% in the post period, whereas the return rate in rural counties was 74% in the pre-period and 73% in the post period.

In urban and rural counties, most enrollees were older than 45 years, enrollees were predominantly women, and the majority had incomes between 100% and 150% of the FPL. Urban counties experienced the largest growth among enrollees aged 35 to 44 years (107%; 95% CI, 94%-119%; P < .001), followed by those aged 26 to 34 years (91%; 95% CI, 78%-105%; P < .001) and 45 to 54 years (83%; 95% CI, 72%-95%; P < .001). Enrollment in rural counties also increased the most among individuals aged 35 to 44 years (103%; 95% CI, 95%-112%; P < .001), followed by individuals aged 26 to 34 years (84%; 95% CI, 76%-91%; P = .001) and 45 to 54 years (81%; 95% CI, 73%-89%; P < .001). Larger growth in enrollment among younger groups of enrollees (aged 35-44 and 26-34 years) in rural and urban counties may point toward improvements in the risk pool for marketplace plans.

The growth of enrollees in urban counties was also concentrated among individuals who self-identified as Hispanic (39%; 95% CI, 33%-46%; P < .001), with smaller but statistically meaningful growth among non-Hispanic White enrollees (16%; 95% CI, 12%-19%; P < .001). In rural counties, the growth rate was also largest among Hispanic enrollees (38%; 95% CI, 33%-43%; P < .001) followed by non-Hispanic White enrollees (17%; 95% CI, 14%-20%; P < .001). Enrollment growth from the pre-period to the post period was the largest for individuals with incomes between 100% and 150% of the FPL in both urban (124%; 95% CI, 106%-142%; P < .001) and rural (116%; 95% CI, 106%-127%; P < .001) counties.

In terms of plan selection, the share of enrollees in gold plans increased in both urban and rural counties from 14% to 32% (P < .001) and 19% to 41% (P < .001), respectively. The Figure shows that the share of counties reporting enrollment counts for bronze plans decreased for both urban (97% to 66%, or 31 percentage points; P < .001) and rural (75% to 41%, or 34 percentage points; P < .001) counties. However, the number of counties reporting gold plan enrollment increased substantially among urban counties, from 70% in 2022 to 95% in 2023-2024 (P < .001), whereas the number of rural counties reporting gold plan enrollment increased from 51% to 93% (P < .001). eAppendix Table 3 reports the underlying data. In terms of plan cost, we observed no increase in the mean total premium for enrollees in urban or rural counties after the rating area change (P = .135 and P = .340, respectively). The mean premium after subsidy (APTCs) declined somewhat for enrollees in rural counties by –1% (95% CI, –13% to –3%; P < .001) but decreased for enrollees residing in urban counties by 29% (95% CI, –34% to –23%; P < .001).

Although we did observe significant growth in enrollment among different groups within urban and rural counties after the rating area change, regression results showed limited differential effects for rural counties relative to urban counties. Regression results in panels A, B, and C of Table 2 show no significant change in enrollment after the rating area change in rural counties relative to urban counties. We also did not find significant changes in enrollment across demographic and income characteristics. We found that enrollment in bronze plans increased in rural counties relative to urban counties after the rating area change (8.13; 95% CI, 2.39-13.87; P = .006) (Table 2, panel D). In addition, premium after subsidy increased by approximately $16 in rural relative to urban counties (95% CI, $11.27-$20.72; P < .001) (Table 2, panel D). To validate the robustness of our regression results and account for the multiple comparison problem, we also applied Bonferroni correction to all significant coefficients in Table 2. We found that the coefficient on premium after APTC is still significant.

Finally, results of subsample analysis evaluating whether rural counties with larger vs smaller growth in marketplace enrollment experienced differential effects of the change in rating areas on premiums before and after APTCs are presented in Table 3. We limited our analysis to premiums because counties with above-median enrollment growth will mechanically have higher enrollment. Specifically, mean premiums decreased by approximately $23 (95% CI, –$37.12 to –$8.90; P = .002), whereas premiums after subsidy decreased by almost $5 (95% CI, –$10.05 to $0.81; P = .095) for rural counties with above-median growth in enrollment compared with rural counties with below-median enrollment growth. eAppendix Table 4 describes the outcomes for rural and urban counties before and after the rating area change.

DISCUSSION

In our descriptive analysis of changes in the design of Texas rating areas, we found growth in urban and rural county enrollment, similar shifts in the distribution of demographic enrollee profiles, and plan selection shifting toward higher actuarial value plans, especially in rural counties. Taking enrollment growth rates in urban counties as the gold standard and using a regression model, we found no evidence that changing rural counties’ rating areas was associated with greater growth rates relative to urban counties. Instead, growth rates in rural counties were on par with urban counties. This also suggests that the integration of rural counties into nearby urban rating areas did not have negative effects on overall enrollee mix in urban counties, which is the main concern for insurance market operations.

Of note is that although rural and urban enrollees were more likely to choose a gold plan, a larger share of rural enrollees chose a gold plan in the post period (41% vs 32%), suggesting that they benefited from improved health coverage and shifted away from previously cheaper, and less protective, bronze plans that do not maximize federal subsidies. This occurred despite more unique gold plans being offered to consumers in urban areas after the rating area change (median, 28 vs 22). From a policy perspective, this change is highly desirable because it limits the financial exposure of these residents and likely increases utilization of needed health services. However, this potentially came at a price, as after-subsidy premiums increased not only for rural enrollees relative to urban enrollees but also for rural enrollees residing in counties with lower vs higher growth in marketplace enrollment. A potential increase in access to care due to enrollment into a more generous plan for some consumers has been reported elsewhere; however, the impact on enrollment and plan selection has not been described.12

Previous work has shown that access to health insurance plans improved after the Texas rating area change.12 Access to insurance plans could have been improved even more for rural residents if insurers were required to offer plans in all counties within a rating area. For example, based on 2022 plan-level data from CMS public use files, insurers offered plans in only a subset of counties in 11 rating areas. The absence of negative effects of the change in Texas rating areas on enrollment mix in 2023 may incentivize insurers to expand their plan offerings to rural counties in future years, further expanding plan choice for rural consumers.

Future research should explore the impact on health care utilization, health status, and other pertinent outcomes that could have changed after the large increases in enrollment in high actuarial value plans, such as financial security. Other avenues of research should explore the insurance market from a perspective of health plan benefits depth, including the composition of provider networks, provider choices, and consumer travel burdens.

Although our findings offer a glimpse into potential benefits of changing marketplace areas, the question remains: Are these findings generalizable, especially as other states struggle to find ways to improve access to affordable plans in the individual marketplace? Even states that do not currently utilize the MSA+1 approach should consider whether it is worth expanding metropolitan and urban rating areas outward to incorporate more rural areas.

Limitations

This study has some limitations. First, rating area changes occurred at the same time Texas was experiencing continued growth in marketplace enrollment driven by the introduction of the temporary elevated plan subsidies that will expire at the end of 2025 and a federal regulation that fixed the ACA’s “family glitch.”24 As such, the growth rate in enrollment, especially in gold plans, does not identify changes due only to changes in marketplace rating areas for rural consumers. However, previous work on silver loading in other states has shown that individuals with incomes between 100% and 200% of the FPL did not change plan selection. Approximately 75% of enrollees in Texas were below the 200% FPL, and as such, the effect of silver loading may be less important in this context.25,26 Second, our main analysis focused on counties that reported data for 2022-2024, and we could not evaluate plan selection and premium outcomes for all counties. However, the eAppendix results utilizing all counties reporting data in 1 or both years suggest similar conclusions. Third, because we evaluated data for only 2 years post rating area change, our findings can be interpreted as medium-term. Fourth, Marketplace Open Enrollment County-Level Public Use Files include only sign-ups during open enrollment and not actual enrollment counts of individuals who paid the premiums. Fifth, although most of our regression results in Table 2 are insignificant, they are not precisely estimated due to large SEs. Future work should investigate whether our findings are robust to the addition of more years of postpolicy data.

CONCLUSIONS

We expanded on earlier work displaying improved access to plan choice and lower premiums. In this repeated cross-sectional study of marketplace enrollment data in Texas, we found similar growth rates in enrollment for rural and urban counties after the inclusion of rural counties into geographically proximate rating areas. We also found that rural and urban enrollees were more likely to choose a gold plan, suggesting that they may have benefited from more generous health coverage as they shifted away from less-expensive bronze plans.

Acknowledgments

The authors thank Susmita Chakraborty for research assistance assembling the data.

Author Affiliations: Texas A&M University (EA, DM, SH, BU), College Station, TX.

Source of Funding: None.

Author Disclosures: Dr Haeder has received fees from the Washington State Office of the Attorney General for consulting on provider networks, has provided expert testimony in a lawsuit on Texas Affordable Care Act (ACA) plans and provider networks, and has received a Robert Wood Johnson Foundation and the Pennsylvania Insurance Department grant for work on ACA marketplaces. Dr Ukert was affiliated with Elevance Health. The remaining authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.

Authorship Information: Concept and design (EA, DM, SH, BU); acquisition of data (EA, SH); analysis and interpretation of data (EA, DM, SH, BU); drafting of the manuscript (EA, DM, SH, BU); critical revision of the manuscript for important intellectual content (EA, DM, SH, BU); statistical analysis (EA, DM, BU); and supervision (BU).

Address Correspondence to: Benjamin Ukert, PhD, Department of Health Policy and Management, Texas A&M University, 212 Adriance Lab Rd, College Station, TX 77845. Email: bukert@tamu.edu.

REFERENCES

1. Courtemanche C, Marton J, Ukert B, Yelowitz A, Zapata D. Early impacts of the Affordable Care Act on health insurance coverage in Medicaid expansion and non-expansion states. J Pol Anal Manage. 2017;36(1):178-210. doi:10.1002/pam.21961

2. Haeder SF, Yackee SW. A look under the hood: regulatory policy making and the Affordable Care Act. J Health Polit Policy Law. 2020;45(5):771-786. doi:10.1215/03616878-8543250

3. Haeder SF, Weimer DL. You can’t make me do it: state implementation of insurance exchanges under the Affordable Care Act. Public Adm Rev. 2013;73(suppl 1):S34-S47. doi:10.1111/puar.12065

4. Crespin D, DeLeire T. As insurers exit Affordable Care Act marketplaces, so do consumers. Health Aff (Millwood). 2019;38(11):1893-1901. doi:10.1377/hlthaff.2018.05475

5. Panhans M. Adverse selection in ACA exchange markets: evidence from Colorado. Am Econ J Appl Econ. 2019;11(2):1-36. doi:10.1257/app.20170117

6. Treasure G, Anderson DM, Hatcher L, et al. Plan selection, enrollee risk, and health spending on the Patient Protection and Affordable Care Act individual marketplaces, 2019. JAMA Netw Open. 2023;6(3):e234529. doi:10.1001/jamanetworkopen.2023.4529

7. Sen AP, DeLeire T. How does expansion of public health insurance affect risk pools and premiums in the market for private health insurance? evidence from Medicaid and the Affordable Care Act Marketplaces. Health Econ. 2018;27(12):1877-1903. doi:10.1002/hec.3809

8. Dickstein MJ, Duggan M, Orsini J, Tebaldi P. The impact of market size and composition on health insurance premiums: evidence from the first year of the Affordable Care Act. Am Econ Rev. 2015;105(5):120-125. doi:10.1257/aer.p20151083

9. Jacobs PD, Banthin JS, Trachtman S. Insurer competition in federally run marketplaces is associated with lower premiums. Health Aff (Millwood). 2015;34(12):2027-2035. doi:10.1377/hlthaff.2015.0548

10. Abraham JM, Royalty AB, Drake C. Plan choice and affordability in the individual and small-group markets: policy and performance—past and present. Health Aff (Millwood). 2019;38(4):675-683. doi:10.1377/hlthaff.2018.05401

11. Fang H, Ko A. Partial rating area offering in the ACA marketplaces: facts, theory and evidence. National Bureau of Economic Research working paper 25154. October 2018. Accessed June 1, 2024. https://www.nber.org/papers/w25154

12. Haeder SF, Andreyeva E, Marthey D, Ukert BD. Merging rural and urban ACA rating areas improved choice, premiums in rural Texas. Health Aff (Millwood). 2023;42(11):1527-1531. doi:10.1377/hlthaff.2023.00444

13. Market rating reforms. CMS. Accessed November 6, 2023. https://www.cms.gov/cciio/programs-and-initiatives/health-insurance-market-reforms/state-gra

14. Fehr R, Cox C, Levitt L. Insurer participation on ACA marketplaces, 2014-2019. KFF. November 2018. Accessed June 1, 2024. http://resource.nlm.nih.gov/101740400

15. Dafny L, Gruber J, Ody C. More insurers lower premiums: evidence from initial pricing in the health insurance marketplaces. Am J Health Econ. 2015;1(1):53-81. doi:10.1162/ajhe_a_00003

16. Abraham JM, Drake C, McCullough JS, Simon K. What drives insurer participation and premiums in the federally-facilitated marketplace? Int J Health Econ Manag. 2017;17(4):395-412. doi:10.1007/s10754-017-9215-y

17. Van Parys J. ACA marketplace premiums grew more rapidly in areas with monopoly insurers than in areas with more competition. Health Aff (Millwood). 2018;37(8):1243-1251. doi:10.1377/hlthaff.2018.0054

18. Lerman AE, Sadin ML, Trachtman S. Policy uptake as political behavior: evidence from the Affordable Care Act. Am Polit Sci Rev. 2017;111(4):755-770. doi:10.1017/S0003055417000272

19. Wengle E. Are marketplace premiums higher in rural than in urban areas? Robert Wood Johnson Foundation. November 1, 2018. Accessed November 6, 2023. https://www.rwjf.org/en/insights/our-research/2018/11/are-marketplace-premiums-higher-in-rural-than-in-urban-areas.html

20. Anderson D, Abraham JM, Drake C. Rural-urban differences in individual-market health plan affordability after subsidy payment cuts. Health Aff (Millwood). 2019;38(12):2032-2040. doi:10.1377/hlthaff.2019.00917

21. Cox C, Semanskee A, Claxton G, Levitt L. Explaining health care reform: risk adjustment, reinsurance, and risk corridors. KFF. August 17, 2016. Accessed November 6, 2023. https://www.kff.org/affordable-care-act/issue-brief/explaining-health-care-reform-risk-adjustment-reinsurance-and-risk-corridors/

22. Hospital closings likely to increase. Health Resources & Services Administration. October 2017. Accessed June 1, 2024. https://web.archive.org/web/20211015231549/https://www.hrsa.gov/enews/past-issues/2017/october-19/hospitals-closing-increase.html

23. Barker AR. Effect of population size on rural health insurance premiums in the Federal Employees Health Benefits Program. Health Aff (Millwood). 2019;38(12):2041-2047. doi:10.1377/hlthaff.2019.00912

24. Norris L. IRS regulations fix the ACA’s ‘family glitch’ as of 2023. HealthInsurance.org. April 19, 2023. Accessed October 6, 2024. https://www.healthinsurance.org/obamacare/irs-regulations-fix-the-acas-family-glitch-as-of-2023/

25. Sprung A, Anderson DM. Mining the silver lode. Health Affairs Forefront. September 7, 2018. Accessed June 1, 2024. https://www.healthaffairs.org/content/forefront/mining-silver-lode

26. Fiedler M. The case for replacing “silver loading.” Brookings. May 20, 2021. Accessed June 15, 2024. https://www.brookings.edu/articles/the-case-for-replacing-silver-loading/