HEALTHCARE 300-500 words NO PLAGERISM Due May 5 at 11:59 PM

National Tax Journal, December 2016,69(4), 883-904 https://doi.O rg/10 .1 7310 /n tj.2 0 1 6.4.08

W H A T C A N TA X DATA TELL US A B O U T T H E U N IN S U R E D ? E V ID E N C E F R O M 2 0 1 4

Ithai Z. Lurie and Janet McCubbin

About 14 percent o f fam ilies that file d tax returns fo r 2014 reported a spell o f uninsurance fo r at least one fam ily member. Uninsurance rates were higher fo r young adults, unmarried persons, low-income families, and fam ilies in states that did not expand eligibility fo r Medicaid. These results are generally consistent with estimates from survey data. M any fam ilies who were uninsured in 2014 appear eligible fo r M edicaid or Premium Tax Credits. Outreach to these fam ilies coidd ef­ fectively reduce the number o f uninsured. Better data fo r 2015 and later years will allow fo r more comprehensive and detailed estimates about uninsured Americans.

Keywords: ISRP, exemption, uninsured

JEL Codes: H22, H24,113

I. INTRODUCTION

he Affordable Care Act o f 2010 (ACA) intersects with the tax system in several ways. Most importantly, the ACA created subsidies, including the Premium Tax

Credit (PTC), to help low- and moderate-income people obtain affordable health insur­ ance. The AC A also requires individuals to obtain health insurance coverage, receive an exemption from the coverage requirement, or pay a penalty. In addition, large employers that do not provide affordable coverage to full-time employees may owe an assessable payment if one or more full-time employees receive the PTC.

In this paper we use tax return data to learn about the characteristics of the uninsured in 2014. Data on the insured and uninsured can help us understand the efficacy of subsidies and penalties in inducing health insurance coverage. In addition, information about the reporting of insurance status will help the Internal Revenue Service (IRS) improve tax forms and instructions. Information about the characteristics and behavior of the uninsured will also help IRS and others to improve outreach efforts aimed at

Ithai Z. Lurie: O ffice o f Tax Analysis, U.S. D e p a rtm e n t o f theTreasury, W ashington, DC, USA (lthai.Lurie@ treasury.gov)

Janet M cC ubbin: O ffice o f Tax Analysis, U.S. D e p a rtm e n t o f th e Treasury, W ashington, DC, USA (Janet. M cC ubbin@ treasury.gov)

884 National Tax Journal

encouraging coverage. Finally, tax data on the uninsured can corroborate and supple­ ment survey data about health insurance coverage — data that are widely used for tax expenditure modeling and myriad other purposes.

We find that 14.5 percent of families that filed tax returns for 2014 reported a spell of uninsurance for at least one family member. Uninsurance rates were higher for young adults, unmarried persons, low-income families, and families in states that did not expand eligibility for Medicaid. These results are generally consistent with estimates from survey data. Many families who remained uninsured in 2014 appear to be eligible for Medicaid or PTC. Outreach to these families could effectively reduce the number o f uninsured. Our findings to date are tentative, because information about insurance status on 2014 returns is limited. Better data for 2015 and later years will allow for more comprehensive and detailed estimates about uninsured Americans.

II. DATA

Taxpayers who have full-year coverage for all of the individuals on their tax returns report this by checking a box on Form 1040. Tax return filers who are uninsured any part o f the year, or who have a dependent who is uninsured for any part of the year, may claim exemptions from the health coverage requirement on Form 8965. Taxpayers who do not have coverage for themselves or for a dependent for at least one month, and who do not claim an exemption, report the individual shared responsibility payment (i.e., the penalty) on Form 1040. In this paper we use tax year 2014 data to examine the characteristics o f tax return filers in 2014. Specifically, we analyze the population o f tax year 2014 returns processed by IRS through March 2016. We define uninsured families or persons as those who claim an exemption from the health care coverage requirement or who pay a penalty for failure to have coverage or both for at least one month.

Taxpayers are responsible for providing coverage, claiming exemptions, and receiving PTCs for themselves and for any person they may claim as a dependent. Dependents generally do not report on their own health insurance status, even if they file tax returns. Thus, we exclude returns filed by dependents from our analysis and define a tax family as a non-dependent return, including the primary and secondary filers and all dependents.

Relying on tax return data creates analytical challenges and important limitations to our findings. First, because insurance status was not reported on tax returns before 2014, we cannot show the effect o f the AC A on coverage.1 Second, the 2014 data are largely self-reported, and taxpayers face incentives to over-report coverage. However, we find several indications that the self-reported data are largely accurate. Third, our data are limited to families that file tax returns, and exclude the approximately 10 percent

1 Data from the Current Population Survey and other sources find that rates o f uninsurance fell significantly between 2013 and 2014, in a manner that is consistent with the intent of the ACA (Smith and Medalia, 2015).

What Can Tax Data Tell Us about the Uninsured? Evidence from 2014 885

of the U S. population (including 7 percent o f the non-elderly population) that does not appear on an income tax return. Lastly, the tax return data do not provide insur­ ance status for each person or the duration of any spell of uninsurance; in many cases we know only that at least one family member was uninsured for at least part of the year.2

III. EXEMPTIONS FROM THE COVERAGE REQUIREMENT

Exemptions from the coverage requirement are available for several specific cir­ cumstances provided under the statute and for general hardship, as determined by the Marketplaces (under the purview o f the Department o f Health and Human Services).3 Some exemptions are granted only by the Marketplaces, some are claimed only on the tax return, and some can be taken in either way.

Exemptions granted by the Marketplace are reported to IRS by the Marketplace and reported by tax return filers on Part I of Form 8965. The data reported to IRS by the Marketplace include the specific individual and months for which each exemption applies and the type o f exemption. In 2014, nearly all Marketplace exemptions reported to IRS were applicable for the full year. Exemptions that may be granted by the Marketplace include those for: general hardship (e.g., bankruptcy, fire, homelessness, death of a family member); unaffordable coverage based on projected income;4 membership in a religious sect that objects to accepting insurance benefits; individuals unable to renew 2013 coverage; and certain other exceptions available only for 2014.

Exemptions that may be claimed only on the tax return are reported in Parts II or III of Form 8965. These include exemptions for. income below the filing requirement (but nevertheless filing a tax return);5 unaffordable coverage, based on actual income or

2 More complete information will be available for future years. To facilitate IR S’s administration o f the ACA tax provisions and to help taxpayers correctly report on their coverage status, insurers and employ­ ers provide information about health coverage to enrollees and to IRS. This information includes the type o f coverage and months o f coverage for each person enrolled in insurance that meets the ACA coverage requirement. Coverage through a health insurance Marketplace (which is required in order to obtain the PTC) is reported by the Marketplaces on Form 1095-A. Large employers report offers o f coverage and enrollment in self-insured plans on Form 1095-C. Other insurers, government health programs, and small employers providing self-insured coverage report monthly enrollment for each person on Form 1095-B. While Marketplace data are available for 2014, employer and insurer reporting on Forms 1095-B and C did not begin until this year, for coverage in tax year 2015.

3 See Internal Revenue Code §5000A, 45 CFR 155.605 and 26 CFR 1,5000Afor the health insurance cover­ age requirement, penalty calculation, and exemptions.

4 The exemption for coverage unaffordable based on projected income is allowed: (1) if an individual is eligible for employer-sponsored coverage but the required employee contribution for the coverage exceeds 8 percent o f projected household income for the year; or (2) if the required contribution (after accounting for PTC) for the lowest-cost bronze Marketplace plan that would cover all family members who are not eligible to purchase employer-sponsored coverage or have not received another exemption exceeds 8 percent of projected household income.

5 Families who do not file tax returns do not need to claim exemptions.

886 N a t io n a l T a x J o u r n a l

based on aggregate self-only coverage;6 citizens living abroad or certain non-citizens; short gaps in coverage o f three months or less; and a few other exceptions that apply in 2014 only.

Some exemptions may be granted by the Marketplace (in which case they may be reported in Part I of Form 8965) or claimed on the tax return (in which case they are reported in Part III of Form 8965). These include exemptions for: income below 138 percent of the Federal poverty level (FPL) and resident of a state that did not expand Medicaid; member of a health care sharing ministry; member of an Indian Tribe or recipient o f Indian Flealth Services care; and incarceration.

Exemptions are generally reported for each person and each month, and we provide results for both families and individuals below. However, the exemption for having income below the filing requirement applies to the entire tax family for the entire year, and uninsurance for specific family members or months is not reported. Therefore, we categorize every person on a tax return with this exemption as exempt and uninsured for the entire year when analyzing the data at a person-level. Interpreting every person who qualifies for this exemption as uninsured will overstate the number o f uninsured persons.

Individuals may be eligible for more than one type of exemption for a given month, and in some cases are instructed to report all of the exemptions for which they qualify. However, taxpayers are not required to determine or report all of the exemptions for which they qualify. Therefore, we use an ordering rule to place individuals in mutually exclusive categories o f the type o f exemption used. We first identify who claims the exemption for having income below the filing requirement, followed by those who have low incomes and reside in a state that did not expand Medicaid, citizens abroad and certain non-citizens, persons with a short-coverage gap, persons with unaffordable coverage, and other exemptions. Placing persons in mutually exclusive categories sim­ plifies the analysis. But as a result, the exemption for having income below the filing requirement will appear more frequently and the exemption for unaffordable coverage less frequently than would be the case if we used a different ordering rule or none at all.

In addition, some codes used on Form 8965 may refer to more than one type of exemption. In particular, code “G” may indicate that coverage is unaffordable based on aggregate self-only coverage that the individual has low income and resides in a state that did not expand eligibility for Medicaid, or that other special cases for 2014 apply. We assume that a code G on a return with income below 138 percent of the poverty level filed from a state that did not expand Medicaid is explicitly for that exemption and that the remaining code G entries indicate that coverage was unaffordable.

6 The exemption allowed because coverage is unaffordable based on actual income is similar to that de­ scribed in footnote 4 above, but is based on actual household income rather than projected income. The exemption for unaffordable aggregate self-only coverage is allowed for the year if for any month: (1) the cost of employer-sponsored self-only coverage for two or more members is affordable when tested individually, (2) the cost of employer-sponsored family coverage is unaffordable, and (3) the combined cost of employer-sponsored individual coverage is unaffordable.

W h a t Can Tax D ata Tell Us a b o u t th e U n in s u re d ? E v id e n c e fr o m 2 0 1 4 887

IV. CALCULATION OF THE INDIVIDUAL PAYMENT RESPONSIBILITY

Taxpayers who are uninsured for all or part of the year (or who have an uninsured dependent) and who do not qualify for an exemption from the coverage requirement must make an individual responsibility payment. The amount of this payment, or pen­ alty, depends on the number o f persons without qualifying health insurance coverage per month, income, and the family’s filing threshold. Specifically, for 2014 the penalty for each month is

(1) (l/12)xm ax|m in[95x(A ?<i +0.5A?J,95x3],

0.01 x( household income - filing threshold} } ,

where AMs the number o f uninsured adults for the month, JV is the number of uninsured children under age 18, and household income is adjusted gross income (AGI) plus tax exempt interest and amounts o f foreign earned income or foreign housing excluded from AGI.

The penalty for the year is the sum of the monthly amounts, limited to the cost of the national average bronze plan that could cover the uninsured, non-exempt family members for all months of uninsurance, up to a maximum of five persons. For 2014, the national average bronze plan is $204 per person per month, for a maximum of $1,020 per family per month.7 In 2014, the average annual penalty per return with a penalty was about $210 (Internal Revenue Service, 2016).

Only the total penalty for the year is reported by the taxpayer on the tax return. In some cases, the number o f uninsured persons by month can be inferred from the pen­ alty amount. In other cases, we know only that at least one person was uninsured for at least part of the year. For the analysis in this paper, we present information about tax families who pay any amount o f penalty. In addition, we present information about families who pay the maximum amount of penalty conditional on their income and family size. If a family is paying the maximum amount, we know that the family paid some penalty every month, meaning that at least one person was uninsured each month. For a single person, this amounts to being fully uninsured. In addition, for most families who are paying a maximum penalty that exactly equals $95x(Na + 0.5A ), we know that the family is fully uninsured. Flowever, for families o f two or more people who are paying 1 percent o f income over the filing threshold, we cannot discern the number of uninsured persons. Thus, paying the maximum amount might best be thought of as correlated with fully uninsured status or as indicating a greater intensity o f uninsurance.

7 The $95 amount is increased to $325 for 2015 and $695 for 2016, and indexed to inflation thereafter. The 1 percent of income is increased to 2 percent for 2015 and 2.5 percent for 2016 and thereafter. The national average bronze plan premium is based on the average o f the median premium for a 21-year old non-tobacco user in each county, weighted by county populations. Internal Revenue Service (2014) provides more details.

888 N ational Tax Journal

In the next section, we examine rates o f uninsurance among different types o f families and the distribution of uninsured families by filing status, age, income relative to the poverty level, and other characteristics. Specifically, we tabulate the number o f families paying a penalty, claiming an exemption, or both. Then we look more closely at the types of exemptions claimed, at the individual (rather than family) level.

V. RESULTS

We find that of tax year 2014 returns processed through the end of March 2016,11.7 million claimed an exemption only, 7.2 million paid a penalty only, and 0.8 million reported both for tax year 2014 (Figure 1). Thus, in total, 19.7 million families reported at least one month o f uninsurance for at least one family member. This amounts to 14.5 percent of the 135.5 million families who filed 2014 returns.8

It is difficult to interpret this rate o f uninsurance, because it may reflect spells as short as one month or as long as a year, and it may reflect uninsurance for only some family members or for the whole family. Thus, our measure is expected to overstate the share of individuals who are uninsured all year or at any point in time. On the other hand, because our data exclude very low-income families who do not file tax returns,

F ig u re 1

N u m b e r o f R eturns w ith P enalty, E x e m p tio n , o r Both

Penalty Only Exemption Only Category

Both

8 This is nearly identical to tabulations reported by IRS for tax year 2014 returns processed through October 2015 (Internal Revenue Service, 2016).

W h a t Can Tax D ata Tell Us a b o u t t h e U n in s u re d ? E v id e n c e fr o m 2 0 1 4 889

our measure could understate the true rate of uninsurance, at least among the non- elderly. The Bureau of the Census reports that for 2014 an estimated 10.4 percent o f individuals were uninsured all year and nearly 12 percent were uninsured at the time they were surveyed.5 * * * 9 While our measure is conceptually different and excludes very low-income persons, it is o f the same order of magnitude and suggests that taxpayers accurately reported their status on their tax returns, in a manner similar to what they report in survey data.

Figure 2A shows that unmarried taxpayers are more likely to be uninsured. This is consistent with Census tabulations. Somewhat surprising to us, families with children are more likely than those without children to report a spell of uninsurance. About 20 percent o f head of household filers, that is, unmarried persons with children, report a spell of uninsurance, compared to 17 percent of single filers without children, 12 percent of married couples filing jointly with children and only 7 percent o f joint filers without children. The higher rate of uninsurance among filers with children may be due to the age composition o f the family. It does not necessarily mean the children were uninsured, and may indicate that a parent was uninsured. Figure 2B shows that single taxpayers without children account for about half o f all tax returns reporting a spell of uninsurance.

The rate of uninsurance was highest for families in which the primary taxpayer was under age 25 (22 percent) or age 25 to 29 (22.5 percent), and declined steadily after age 29. Families in the under-30 age group account for about one-third o f uninsured families. As expected, the rate of uninsurance drops markedly, tojust 3 percent, when the primary taxpayer reaches age 65 and becomes eligible for Medicare (Figures 3A and 3B).

Lower-income families are more likely to be uninsured, and are more likely to claim an exemption than to pay a penalty (Figure 4A).10 About 28 percent of families with incomes below 100 percent of the federal poverty level (FPL) claimed an exemption only, 4 percent paid a penalty only, and 1 percent did both. Among those with incomes between 100 and 138 percent of the poverty level, about 18 percent claimed an exemp­ tion only, 11 percent reported a penalty only, and 2 percent did both. While taxpayers with income below the filing threshold are exempt from the coverage requirement, it is nonetheless possible for poor families to be subject to the penalty. This can occur because the filing threshold is below the FPL for single and head of household filers, and for married couples with two or more children." Thus, some poor families may indeed be subject to a penalty. However, it also appears that some families paid a penalty when they could have claimed an exemption. In fact, IRS has contacted about 400,000 families to notify them that they might have paid too much for tax year 2014.

Families with incomes below the poverty level account for 37 percent of uninsured families, and those between 100 and 13 8 percent of the poverty level account for another 16 percent (Figure 4B).

5 See Smith and Medalia (2015) for the Census data cited in this paper. 10 For all figures showing income as a percentage o f FPL, income is defined as AGI plus tax-exempt interest,

foreign earned income and housing, and Social Security income that is otherwise excluded from AGI. This is the definition used to determine eligibility for Medicaid and PTC.

11 For example, for a single person in 2014, the filing threshold was $10,150 and the poverty level was $11,490.

890 National Tax Journal

Figure 2A

■ Penalty Only Exemption Only "B oth

Figure 2B

Distribution of Uninsured Returns by Family Type

C/)c CDO'

■ O0) 3

c ’ c 3 o 0l_ 03

SZ i f )

0.60

0.50

0.40

0.30

0.20

0.10

0

Family Type

W hat Can Tax Data Tell Us a b o u t th e U ninsured? Evidence fro m 2014 891

Figure 3A Share o f Families that is Uninsured by Age

Age

■ Penalty Only Exemption Only ■ Both

Figure 3B D istribution o f Uninsured Families by Age

Age

892 NationalTax Journal

Figure 4A

Share o f Families that is Uninsured by FPL

0.35 ■

« 0.25

1 0.20 - LL

O <u 0.15

Income as a Share of FPL (l/FPL)

■ Penalty Only Exemption Only a* Both

Figure 4B

D istribution of Uninsured Families by FPL

Income as a Share of FPL (l/FPL)

W h a t Can Tax D ata Tell Us a b o u t t h e U n in s u re d ? E v id e n c e fr o m 2 0 1 4 893

Prior to 2014, most states did not provide Medicaid to non-disabled adults without children. The ACA provides incentives for states to extend eligibility for Medicaid to all persons with incomes below 138 percent o f the poverty level, but 23 states decided not to do so for 2014.12 Taxpayers in these states with incomes below 100 percent of the poverty level are generally not eligible for Medicaid or for subsidized Marketplace coverage. Taxpayers in these states with incomes between 100 and 138 percent o f the poverty level are not eligible for Medicaid but may receive subsidized Marketplace coverage, including both PTC and help with out-of-pocket costs.13 All persons with incomes below 138 percent of the poverty level and residing in a state that did not expand Medicaid are exempt from the coverage requirement. Many received this exemption when they went to the Marketplace seeking coverage. In addition, this exemption may be claimed on the tax return.

Other researchers (Council of Economic Advisers, 2016) have found that rates of uninsurance are lower in states that expanded Medicaid, both before and after 2014. Similarly, we find that uninsurance rates are 12 percent in states that expanded Medicaid eligibility and 18 percent in those that did not. Families in states that did not expand Medicaid were more likely than others to claim an exemption from the coverage require­ ment (11 percent versus 7 percent). Families filing from a state that did not expand Medicaid eligibility account for 44 percent of all families and 54 percent o f uninsured families (Figures 5A and 5B).

As expected, much of this difference in uninsurance rates is attributable to poor families who could receive Medicaid if they resided in a state that expanded eligibility (Figures 6A and 6B). For example, the rate o f uninsurance among families with incomes below 100 percent o f the poverty level is 42 percent in states that did not expand Med­ icaid and 26 percent in states that did expand Medicaid. The rate o f uninsurance among families with incomes between 100 and 138 percent of the poverty level is 37 percent in non-expansion states and 25 percent in expansion states. However, even at higher income levels, rates o f uninsurance are slightly higher (1 to 5 percentage points higher) in non-expansion states. This suggests that the lower rates of uninsurance are partly due to the Medicaid expansion directly, partly to spillover effects such as increased

12 The 23 states that did not expand Medicaid eligibility for 2014 include: Alabama, Alaska, Florida, Georgia, Idaho, Indiana, Kansas, Louisiana, Maine, Mississippi, Missouri, Montana, Nebraska, North Carolina, Oklahoma, Pennsylvania, South Carolina, South Dakota, Tennessee, Texas, Utah, Virginia, and Wyoming. Wisconsin expanded eligibility to include adults with incomes up to 100 percent o f the poverty level and New Hampshire expanded eligibility effective August 15, 2014. These are included as expansion states, although low-income taxpayers residing in these states could claim an exemption from the coverage re­ quirement for 2014. Categorizing Wisconsin and New Hampshire as non-expansion states does not make a noticeable change in the results. Alaska, Indiana, and Pennsylvania subsequently expanded Medicaid eligibility beginning with all or part of 2015; Montana for 2016; and Louisiana is expected to expand in June 2016. Thus, by the latter h alf o f 2016 there will be 18 states that have not extended Medicaid eligibil­ ity to all poor adults (those below 100 percent o f the poverty level).

13 In addition to being eligible for PTC, families who obtain Marketplace coverage and who have incomes below 250 percent o f the poverty level may be eligible for cost-sharing reductions, which reduce their out-of-pocket copayments and deductibles.

894 National Tax Journal

Figure 5A

Share o f Families that is Uninsured by Expansion Status

No Expansion Expansion

Expansion Status

■ Penalty Only Exemption Only * Both

Figure 5B

D istribution o f Families by Expansion Status

No Expansion Expansion

Expansion Status

S h

a re

o f

F a

m ili

e s

W h a t Can Tax Data Tell Us a b o u t th e U ninsured? Evidence fro m 2014 895

Figure 6A

Share that is Uninsured by FPL Non-Expansion States

Income as a Share o f FPL (l/FPL)

■ Penalty Only Exemption Only B B o th

Figure 6B

Share that is Uninsured by FPL Expansion States

Income as a Share o f FPL (l/FPL)

■ Penalty Only Exemption Only ■ Both

896 N a tio n a l T a x J o u r n a l

outreach generating more coverage among families with slightly higher incomes, and partly to underlying state characteristics that are correlated with the choice to expand.

Reported insurance status does not vary much by tax preparation mode. About 55 percent of families use a paid preparer to file their tax returns and 38 percent prepare their own returns with the help o f software. This is true for both insured families and uninsured tax return filers. Families that prepare their own returns using software are less likely to report a penalty (4 percent) and more likely to claim an exemption (10 percent) than are other groups, but the difference is not large. Those who prepare their own returns without software are less likely to report that they are uninsured (5 percent), and those using volunteer assistance that primarily serves low-income families are more likely to be uninsured (16 percent), but both of these groups are very small, accounting for less than 7 percent of families.

We focus next on families who paid a penalty. Many families who pay a penalty are likely to be eligible for subsidized Marketplace coverage. We find that 11 percent o f primary taxpayers who reported a penalty for anyone in their family in 2014 were Marketplace policyholders at some point in 2015, indicating that someone in the family had Marketplace coverage.14 Primary taxpayers who paid a penalty and had incomes between 138 and 200 percent of the poverty level in 2014 were most likely to appear as Marketplace policyholders in 2015, at 14 percent. But lower-income and higher-income families subject to penalties in 2014 also appear to have moved into the Marketplace in 2015 (at least as measured by observing the primary taxpayer as a policyholder), including 8 percent below poverty in 2014 and 9 percent o f the uninsured with incomes over 400 percent of the poverty level (Figure 7).

As noted earlier, we cannot infer the duration of uninsurance or the number of unin­ sured family members from the amount of the penalty in many cases. However, we know that a family that paid the maximum amount conditional on the family’s income and family size had at least one uninsured family member in every month. About 56 percent of families paid the maximum amount. This includes 68 percent of single filers with a penalty, who we know were uninsured all year

We next analyze the types o f exemptions that are claimed at the individual level. A total of 22.2 million people were represented on Forms 8965 filed for tax year 2014; that is, for each of these 22.2 million people, an exemption was claimed for them at the individual level or the exemption for income below the filing threshold that applies to the whole family was claimed. About 46 percent of people with exemptions are associ­ ated with a family claiming the exemption for income below the filing threshold. This result is driven in part by the fact that we count every person on returns claiming the filing requirement exemption as exempt, and because we count this exemption first. Other relatively common exemptions are those for residents o f Medicaid non-expansion

Some of these primary taxpayers were also policyholders in 2014, since part of a family may be insured and another part uninsured in any month, or the family may be uninsured for only part o f the year.

W hat Can Tax Data Tell Us about the Uninsured? Evidence from 2014 897

F ig u re 7

Percent of Penalty-paying Families with Primary Found as Policyholder in the 2015 Marketplace, by FPL

0.20

0 .1 8

M 0 .1 6 .CD

I 0 .1 4 li­ en

. £ 0.12 > x 0 3Q_

0.10 co | 0 .0 8 o

0 .0 6 co

w 0 .0 4 0.02

0.00 llillll

Income as a Share of FPL

states, citizens abroad and certain non-citizens, and lack o f affordable coverage options, each accounting for about 15 percent o f exemptions claimed (Figure 8).

Most individuals claiming an exemption, 84 percent, claim to be exempt from the coverage requirement for the full year. Again this is driven in part by the fact that, by definition, income below the filing requirement qualifies every individual in the family for an exemption for the entire year, even if the family had coverage for part of the year or for some members of the family. However, most other exemptions are also claimed for the entire year. The exceptions include exemptions for short coverage gaps, which, by definition, should only be claimed for at most three months, and exemptions for unaffordable coverage, about 43 percent o f which were claimed for the full year.

Persons who are not lawfully present in the United States are generally not eligible for Medicaid or Marketplace health insurance coverage (though they may receive PTC on behalf of a lawfully present dependent enrolled in coverage). These non-citizens as well as citizens living abroad are exempt from the requirement to have health insur­ ance coverage. People who need to file a tax return but do not have a Social Security Number are issued an Individual Tax Identification Number, or ITIN, by IRS. Persons with ITINs may reside outside the United States or they may be in the United States

898 National Tax Journal

F ig u re 8

P e rc en t o f E x e m p t In d iv id u a ls b y T y p e o f E x e m p tio n

Type of Exemption

either lawfully or unlawfully. While we cannot observe who among ITIN-holders is not lawfully present, we know that nearly all filers who are not lawfully present use ITINs. ITIN holders, whether lawfully present or not, may have insurance through other means, such as an employer, family member, or individual policy outside of the Marketplace. However, we find that an exemption was claimed for almost half of persons with ITINs on 2014 tax returns. Nearly 20 percent of persons claiming any type of exemption and over 80 percent of individuals who claim an exemption for citizens abroad and certain non-citizens are ITIN holders (Figure 9).

Over a quarter of exemptions are claimed for children under age 18 (Figure 10). This surprised us, because subsidized coverage (namely Medicaid and the Children’s Health Insurance Program) is more readily available for children. The most common exemption claimed for children is for income below the filing requirement, and it is possible that many of these children do have coverage, but appear on a return where an adult is claiming this exemption for the entire family. We will know much more about these children later this year, when Forms 1095-B showing enrollment in public programs are filed with IRS. The second most common exemption for children is that for citizens abroad and non-citizens.

What CanTax Data Tell Us about the Uninsured? Evidence from 2014 899

Figure 9 Percent o f Exemptions Claimed by Persons w ith ITINs, by Type

Type of Exemption

Figure 10 D istribution o f Individuals w ith Exemptions by Age

Age

900 N atio na l Tax Journal

As other data show, the rate of uninsurance among adults generally falls with age. The small uptick for elderly persons again appears to be attributable to persons on a return with income below the filing requirement, where the exemption is perhaps claimed on behalf of another uninsured family member, and to persons who are abroad or are non-citizens.

Over half of people claiming exemptions have income below the poverty level, and another 18 percent have income between 100 and 13 8 percent o f the poverty level (Figure 11). Individuals with incomes below 100 percent of the poverty level most often claim the exemption for income below the filing requirement. Those with incomes between 100 and 138 percent of the poverty level most often claim the exemption for low-income persons in states that did not expand Medicaid. In contrast, people with incomes above 138 percent o f the poverty level most often claim the exemption for unaffordable cover­ age. Lastly, we find that about 7 percent o f individuals who claimed an exemption for 2014 have Marketplace coverage in 2015. This includes about 15 percent o f those with incomes between 100 and 200 percent of the poverty level (Figure 12).

Figure 11 D istrib u tio n o f Individuals w ith Exem ptions b y FPL

0.60 i

0.50

0.40

0.30 -

£ 0.20 <0

JZCO 0.10

0.0 0

✓

I I i £>

f J t A

07

Income as a Share of FPL (l/FPL)

W h a t C a n T a x D a t a T e ll U s a b o u t t h e U n in s u r e d ? E v id e n c e f r o m 2 0 1 4 901

Figure 12

Percent of Individuals with Exemption in 2014That Have Marketplace Coverage in 2015 By Fpl

_co CT3 =3 ■g >

7 3 _C

Q . E0) X111 o CD

CO

0.16

0.14

0.12

0.10

0.08

0.06

0.04

0.02

0

Income as a Share of FPL (l/FPL)

VI. CONCLUSION AND NEXT STEPS

For 2014,19.7 million families reported at least one month o f uninsurance for at least one family member. This amounts to 14.5 percent o f the 135.5 million families who filed 2014 returns. Of these 19.7 million families 11.7 million claimed an exemption only, 7.2 million paid a penalty only, and 0.8 million reported both for tax year 2014.

Information about uninsured families and individuals gleaned from tax data for 2014 is generally consistent with survey and other administrative data. Census data for 2014 suggest that nearly 12 percent o f individuals in the United States were uninsured at the time they were surveyed and 10.4 percent were uninsured all year, compared to our estimate o f 14.5 percent of families filing tax returns reporting a spell o f uninsurance. We find that rates of uninsurance are higher for unmarried persons, younger adults, lower-income families, persons residing in states that did not expand eligibility for Medicaid, and certain immigrants.

At least 56 percent o f families who paid a penalty did so for all 12 months of the year (for all or part of the family). Among single filers who paid a penalty, at least 68

902 N ational Tax Journal

percent were uninsured all year. Similarly, most individuals who claimed exemptions did so for the full year. However, this result is largely driven by the fact that the most common exemption, that for income below the filing requirement, by definition applies all year and the particular months o f uninsurance or persons uninsured are not reported.

Both the consistency with survey data and some of the patterns o f responses suggest that most taxpayers reported their insurance status correctly. For example, most people claiming the exemption for citizens abroad and certain non-citizens have ITINs. Most people claiming the exemption for persons below the filing requirement have low incomes.

Our findings about the characteristics of the uninsured suggest that many of them are eligible for Medicaid or subsidized Marketplace coverage. About 37 percent of unin­ sured families have incomes below 100 percent of the poverty level, 16 percent have incomes between 100 and 138percent of the poverty level,and 18 percent have incomes between 138 and 200 percent o f the poverty level. While some o f these families are not lawfully present or are poor residents of states that did not expand Medicaid, most are not and additional outreach to these families could encourage them to obtain coverage.

Thus, the Departments of Treasury and Health and Human Service are partnering to send notices to taxpayers who paid a penalty or claimed an exemption for tax year 2015 and appear to be eligible for Medicaid or Marketplace coverage, during the next Marketplace open enrollment period. Several notices will be tested in a randomized controlled study, so that effectiveness of various messages can be measured.

Because we are relying on self-reported tax return data, some o f which is reported only at a family or annual level, we do not currently have comprehensive information on types o f coverage, the number of persons with coverage, or the duration of uninsur­ ance. This information is important for administering a number of AC A provisions. It is also important for understanding key tax subsidies for health insurance, including the premium tax credit and exclusion of employer-sponsored health benefits. Fortu­ nately, beginning with tax year 2015, we will have much better information on types of coverage by person and month. The Office of Tax Analysis will soon be incorporating person-month level information into our Individual Tax Model.

ACKN O W LED G M EN TS A N D DISCLAIMERS

This paper does not necessarily represent the views of the Department of the U S. Treasury. The authors would like to thank Nathan Born for excellent work with final edits to the paper. We would like also to thank participants in the National Tax Asso­ ciation 2016 Spring Symposium and participants in a presentation at the Office of Tax Analysis for their insightful comments.

DISCLOSURES

The authors have no financial arrangements that might give rise to conflicts o f interest with respect to the research reported in this paper.

W hat Can Tax Data Tell Us about the Uninsured? Evidence from 2014 903

REFERENCES

Council o f Economic Advisers, 2016. "The Affordable Care Act at Six: Progress on Cover­ age, Costs, and Quality.” Presentation, March 22. Council o f Economic Advisers, Washing­ ton, DC, https://www.whitehouse.gov/sites/default/files/page/files/20160322_aca_six_year_ anniversary_slides.pdf.

Internal Revenue Service, 2014. “Revenue Procedure 2014-46.” Internal Revenue Service, Washington, DC, https://www.irs.gov/pub/irs-drop/rp-14-46.pdf.

Internal Revenue Service, 2016. Letter from Commissioner John Koskinen (January 8). In­ ternal Revenue Service, Washington, DC, https://www.irs.gov/pub/newsroom/irsJetter_aca_ stats_010816.pdf.

Smith, Jessica C , and Carla Medalia, 2015. “Health Insurance Coverage in the United States: 2014.” Current Population Reports P60-253. U S. Census Bureau, Washington, DC, https://www. ccnsus.gov/content/dam/Census/library/publications/2015/demo/p60-253.pdf.

Copyright of National Tax Journal is the property of National Tax Association and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.