1) State policies expanding dependent health insurance coverage to young adults up to age 26-29 have become increasingly popular, with over 25 states enacting such policies as of 2008. 2) Preliminary analysis found these policies increased young adults' likelihood of being covered as dependents on a private health plan by about 2 to 3 percentage points, offsetting a similar drop in young adults obtaining their own private coverage. 3) The policies did not significantly impact the uninsured rate, suggesting the expanded dependent coverage substituted for other private insurance rather than reducing the number of uninsured. However, the analysis was limited by only a few years of post-policy data in most states.

1. State Policies Expanding Dependent
Coverage to Young Adults in Private
Health Insurance Plans
June 27, 2009
Academy Health State Health Research
and Policy Interest Group Meeting
Chicago, IL
Alan Monheit*,^, Joel Cantor*, Derek DeLia*,
Dina Belloff*, Margaret Koller*, Dorothy Gaboda*
*Rutgers University Center for State Health Policy
^UMDNJ School of Public Health
Funded by the Robert Wood Johnson Foundation SHARE Program

2. Outline
• Young adult coverage
• State dependent coverage expansion policies
• Preliminary impact analysis
• Conclusions and limitations
• Next steps
Center for State Health Policy
Institute for Health, Health Care Policy and Aging Research 2

3. Young Adults at High Risk of Lacking
Coverage and are Large Share of Uninsured
50%
40% 50-64
Under
15%
19
Percent Uninsured
30% 20%
30%
23% 36-49
20% 18% 23%
19-29
12% 13%
30%
10% 30-35
12%
0%
< 19 19-29 30-35 36-49 50-64 Age Distribution of Uninsured
Source: Kriss JL, SR Collins, B Mahoto, et al. “Rite of Passage? Why Young Adults Become Uninsured and How New Policies Can
Help, 2008 Update.” The Commonwealth Fund, Issue Brief, May 2008. Pub. # 1139.
Center for State Health Policy
Institute for Health, Health Care Policy and Aging Research 3

4. Source of Coverage for Young Adults (Age 19-29)
Not Full-Time Students Full-Time Students
12.5 million 7.6 million
Own-
employer
Own-
employer Uninsured
Uninsured 8%
17%
25%
39%
Non-group
or college 20%
16% plan
49%
Employer-
7% 6%
13% dependent
Other
Non-group Employer-
or college Other
dependent
plan
Source: Kriss JL, SR Collins, B Mahoto, et al…The Commonwealth Fund.
Center for State Health Policy
Institute for Health, Health Care Policy and Aging Research 4

5. Implications of High Uninsured Rate
• Critical developmental period to address risks of obesity,
smoking, sexually transmitted infections, etc.
• Uninsured young adults are two to four times…
– more likely than peers to delay/forgo care or an Rx due to costs
– less likely to see a medical provider or have a usual source of
care
• Uninsured young adults 20% more likely to report trouble
paying medical bills or carrying medical debt
• Absence from risk pools has consequences for others
Sources: Kriss JL, SR Collins, B Mahoto, et al…The Commonwealth Fund.
Callahan ST and WO Cooper. 2006. “Access to health care for young adults with disabling chronic conditions.” Archives of Pediatric and
Adolescent Medicine. 160:178-182.
Merluzzi TV and RC Nairn. 1999. “Adulthood and aging: Transitions in health and health cognition.” In Whitman TL, TV Whitman, and RD
White (eds). Life-Span Perspectives on Health and Illness. (pp. 189-206). Mahwan, NJ: Lawrence Erlbaum.
Center for State Health Policy
Institute for Health, Health Care Policy and Aging Research 5

6. State Dependent Coverage Expansion
Enactment Timeline
25 states as of 2008
CT
FL
ID
IN
Original enactments shown in black ME
Expansions shown in blue italics MD
MO
MT
MN
NH
SD FL
TX SD MA VA IA
IL CO RI WA KT
UT NM NJ DE WV NJ
1994 2003 2005 2006 2007 2008
Center for State Health Policy
Institute for Health, Health Care Policy and Aging Research 6

7. Change in Age of Dependent Eligibility (as of 2008)
STUDENTS NON-STUDENTS
Number with Reform (25 total) 19* 23
Greatest Increase in Age Limit No limit 12 years
Mean Increase in Age Limit
3.5 years** 5.7 years
(among reform states)
Notes
Based on date of enactment.
*Includes one state (RI) that increased age limit for part-time students only.
** Excludes two states (TX, IA) that eliminated the upper age limit for full-time students.
Center for State Health Policy
Institute for Health, Health Care Policy and Aging Research 7

8. Change in Age of Dependent Eligibility (as of 2008)
Student Non-Student
No Limit
14
12
Increase in Age Limit
10
8
6
4
2
0
RI* DE IN NM VA MN WA CO WV MA IL ME MD MT CT KY UT FL MO SD NH NJ TX IA
Based reforms enacted as of December 2008.
Center for State Health Policy
*RI raisedHealth,limit for part-time students from 18 to 24 (i.e., treating PT as FT students).
Institute for age Health Care Policy and Aging Research 8

9. Other Provisions
• Unmarried – 22 states
• No dependents – 4 states
• Other limits
– Most states – residency for non-students, but not FT students
– 9 states – financial dependence or living with parents
– 6 states – continuous or creditable coverage
• Included markets
– Most states – all regulated markets and public employee plans
• Premium rules
– 12 states – cost averaged into group premium
– 8 states – establish premiums for new dependent enrollees
Center for State Health Policy
Institute for Health, Health Care Policy and Aging Research 9

10. Factors Potentially Limiting Impact
• ERISA preemption
– e.g., In NJ, ~33% of state population subject to state regulation
(25% in state-regulated plans; 8.6% in state health benefit plan)
• Possible burdens on insurers or employers
– Taxable as income for those over 23 years
• Possible impact on premiums and costs
– Risk selection
– Premium rules
• Unanticipated consequences
– Non-group or other risk pools
– Young adult behavior (e.g., marriage, child bearing)
Center for State Health Policy
Institute for Health, Health Care Policy and Aging Research 10

11. Impact Analysis Strategy
• CPS March Supplements (2000-2008)
– Utah and Massachusetts excluded
– 15 states implementing by 2007, ~23 state-years of experience
• Young adults (ages 19-29)
– Restricted: Single adults living with a parent (n=66,654)
– Full: All young adults (n=227,002)
• Five linear probability models predicting “COVERAGE”
– Covered by employer-sponsored insurance (ESI) as dependent
(on parent’s policy, in restricted model)
– Covered as ESI policyholder
– Non-group coverage
– Public coverage
– Uninsured
• Adjusted for complex sample design (Davern, et al.)
Center for State Health Policy
Institute for Health, Health Care Policy and Aging Research 11

12. Model Specification
COVERAGEi = a1 + a2TARGETi + a3(TARGETi*POLICYs,t) +
a4Xi + a5Zs,t + a6ADOPTs + a7STATEs + a8YEARt + ei
Where:
TARGET = expansion population dummy (regardless of year)
POLICY = state policy in effect dummy
TARGET*POLICY = interaction of being in target population and
living in a state post-policy implementation (a3 is DD estimator)
X = vector of individual characteristics
Z = vector of time-varying state characteristics
ADOPT = predictors of state policy adoption
STATE = state fixed effects
YEAR = time fixed effects
Center for State Health Policy
Institute for Health, Health Care Policy and Aging Research 12

13. “TARGET” define by…
• State of residence
• Age
• Marital status
• Student status
• Other state-specific eligibility criteria
Center for State Health Policy
Institute for Health, Health Care Policy and Aging Research 13

14. Other Variables…
• Individual characteristics (X vector)
– Demographics (age, sex, race/ethnicity)
– Fair/poor health
– Student status
– Educational attainment
– Family income (% FPL)
– Marital status (unrestricted model)
– Live with parent (unrestricted model)
• Time varying state characteristics (Z vector)
– Unemployment rate
– Percent college graduate
• Policy adoption predictors (ADOPT vector)
– Number of benefit/provider coverage mandates
– Party of governor and legislature
– Number of insurance department staff
– Elected insurance commissioner
– State net budget revenues
Center for State Health Policy
Institute for Health, Health Care Policy and Aging Research 14

15. Hypotheses
• Policy impact as intended
– Positive and significant DD estimate for ESI as dependent
– Negative and significant DD estimate for Uninsured
• Unintended substitution effect
– Positive and significant DD estimate for ESI as dependent
– Negative and significant DD estimate for ESI policyholder, non-
group coverage, and/or public coverage
Center for State Health Policy
Institute for Health, Health Care Policy and Aging Research 15

16. Policy Impact Estimates
Change in Probability of Coverage (t-statistic)
Coverage Outcome DD estimates DD estimates
Single, Live w/Parent All Young Adults
ESI as dependent* 0.0267 0.0193
(2.02) (2.96)
Uninsured 0.0007 0.0086
(0.05) (0.89)
ESI as policyholder -0.0202 -0.0201
(-1.64) (-2.31)
Non-Group Coverage -0.0094 -0.0067
(-1.23) (-1.17)
Public Coverage -0.0022 -0.0011
(-0.24) (-0.18)
*Dependent on parent’s ESI plan in restricted model, any dependent ESI in unrestricted model
Bold indicates significant at p<.10 level
Center for State Health Policy
Institute for Health, Health Care Policy and Aging Research 16

17. Predicted Coverage Status
Standard Population of Young Adults (ages 19-29)
Based on Unrestricted Model (n=227,002)
No Expansion Expansion Policy
35%
30%
25%
20%
15%
10%
5%
0%
ESI dependent Uninsured ESI policyholder Non-group Public
Center for State Health Policy
Institute for Health, Health Care Policy and Aging Research 17

18. Conclusions So Far
• Very popular strategy, policy details vary
• Expanded dependent coverage appears to substitute for
other private insurance
– ESI dependent coverage increase of about 2 to nearly 3
percentage points in the target population
– Offset by drop in own-name ESI
– No impact on uninsured rate
Center for State Health Policy
Institute for Health, Health Care Policy and Aging Research 18

19. Limitations
• Early experience
– 23 state-years experience as of 2007
– Nearly half (11 state-years) in first year of implementation,
including 4.7 state-years in 9 states that implemented in 2007
• Some eligibility characteristics unmeasured
– Parental coverage status and state of residence (eligibility
assigned by young adults’ state of residence)
– Financial dependence of young adults on parents
– Parent’s plan ERISA status
– Assumed to be random with respect to adoption
Center for State Health Policy
Institute for Health, Health Care Policy and Aging Research 19

20. Next Steps
• Update analysis with 2009 CPS
– Add 19 more state-years (including 5 states implementing in 2008)
• Additional modeling
– Confirm linear probability models with Logit or Probit
– Refine policy variable (e.g., # years post-implementation, examine
specific state policy features)
– Consider DDD approach comparing to middle aged adults
• Implementation case studies
– Stakeholder interviews in several states TBD
• NJ Family Health Survey analyses, 2001 and 2009
– Pre-post impact analysis
– Estimates of eligible population
– Risk selection
Center for State Health Policy
Institute for Health, Health Care Policy and Aging Research 20