LDI Research Seminar- Is Blood Thicker Than Water? Cardiologist Peer Effects ...
Stephen Parente
1. Consumer Response to a
National Marketplace for
Individual Insurance
Stephen T Parente
University of Minnesota
October 2, 2009, University of Pennsylvania
Contributions by co-authors Roger Feldman, Jean Abraham and Yi (Wendy) Xu as
well as coordinator Ruth Taylor and Lisa Tomai of T3 Health LLC were invaluable.
Critical data, software and methods were supported by the Robert Wood Johnson
Foundation, AHRQ, DHHS/ASPE and HSI Network LLC
3. Rational for Analysis
• Since 1945, McCarran Ferguson act prohibits
the sale of any insurance across state lines.
• Since 1978, ERISA has enabled an opt-out for
employers to self insure across state lines.
• Currently 55%+ of non-Medicaid and non-
Medicare insured receive insurance enabled by
ERISA.
• For the ‘median voter’ with non-public
insurance, McCarran Ferguson applies no more.
4. Policy Proposal
Since 2005, members of Congress (e.g., John Shadegg [R-AZ]) has
proposed that individual health insurance be offered nationally
instead of in state-specific markets.
The University of Minnesota was awarded a contract to study the
likely effect of a national market on take-up of individual health
insurance coverage.
The research objective is to simulate the impact of having a national
market for individual (non-group) coverage and provide advice to
policymakers regarding the strengths and weaknesses of
such a proposal.
5. ‘ARCOLA’ Simulation Model
• ARCOLA simulates national health plan take-
up from policy proposals in the individual and
group markets
• Unique combination of attributes:
– Based on conditional logit model of health plan
choice with data from 4 large employers
– Includes HRA and HSA plans
– Choice model includes measures of chronic illness
burden at contract level
• Can simulate effects of policy changes:
– Premium modifications by tax deduction or credit
– Full or select individual mandates
– State and national market differences
6. National Market Simulation
• Background: Our model predicted take-up of
HSA plans in the individual market quite
accurately (Health Affairs: Feldman, Parente
et al., 2005)
• Population: adults in the MEPS who are aged
19-64 and are not students, not covered by
public insurance, and not eligible for coverage
under someone else’s ESI policy
• Baseline sample uninsured & turned down:
32.3 million people nationally
7. National Market Simulation Steps
1. Create a synthetic version of the MEPS that
assigns people to states based on demographics
2. Identify minimum, moderate and maximum
marginal impact of state regulations on
individual-market premiums
• Community rating
• Guaranteed issue
• Any willing provider
• Mandated insurance benefits
1. Develop initial set of scenarios for policy
• Scenario 1: Competition among 5 largest states
• Scenario 2: Competition among all 50 states
• Scenario 3: Competition within regions
8. Health Insurance Regulations
• Mandates require insurers to cover particular
services or providers
• Guaranteed issue laws require insurers to sell
insurance to all potential customers
• Community rating requires insurers to limit
premium differences across individuals
• Any willing provider (AWP) laws restrict
insurers’ ability to exclude providers from
their networks
9. Literature Review
• We reviewed studies of the individual
insurance market
• We could not find any studies that used ideal
‘dif-in-dif’ research design
• Other papers looked at the effects of
regulations on premiums only for people who
held insurance – we ruled these out
• Only 4 studies met our criteria: 3 working
papers and one peer-reviewed study by
Hadley and Reschovsky (Inquiry, 2003)
11. Simulation Step #4
• Select ‘target state’ in which person can buy
insurance
• Remove the effect of regulations in home state
from premiums and add the effect of regulations
in target state
– In general, target state will have fewer regulations
and lower premium
– Exceptions: (1) target and home state are the same;
(2) high-cost person with community rating in home
state may lose advantage of community rating in
target state
• Simulate the net effect of removing regulations
on health insurance take-up
12. Details & Assumptions
• Premium data:
– HSA from ehealthinsurance.com for HSAs
– HRA from composite of 3 of our empl0yers
– Kaiser/Commonwealth for all other plan designs
• State-specific premium inflators/deflators derived
from Musco et al. AHIP report on individual health
insurance
• Individual market premiums were experience rated
for age and gender (except community rated states)
• Small group market (<250 employees) premiums
were adjusted by state-specific regulatory effects
• Employee premiums in large firms were tax-adjusted
• HSA premiums include a $1K/$2K investment in
accounts
13. Scenario 1: Competition among 5 largest States
4,688,254
Status
Quo Mininum Moderate Maximum
Indiviudal
HSA 4,655,291 4,493 0% 806,865 17% 1,282,626 28%
PPO High 7,515,552 33,396 0% 2,486,440 33% 4,456,992 59%
PPO Low 180,379 263 0% (22,243) -12% (30,380) -17%
PPO Medium 1,534,799 3,886 0% 20,139 1% 12,174 1%
Uninsured 28,848,310 (42,038) 0% (3,291,201) -11% (5,721,413) -20%
Group Market
HMO 5,505,466 (0) 0% (179) 0% (1,487) 0%
HRA 6,166,134 (4) 0% (791) 0% (2,711) 0%
HSA Offered 307,298 (0) 0% (37) 0% (165) 0%
HSA No-offer 11,088 69 1% 27,301 246% 135,973 1226%
PPO High 16,535,831 (2) 0% (578) 0% (3,229) 0%
PPO Low 665,950 (0) 0% (72) 0% (796) 0%
PPO Medium 53,470,814 (62) 0% (25,093) 0% (119,262) 0%
Turned Down 3,530,681 (0) 0% (552) 0% (8,323) 0%
Within Sample National
Mininum Insurance Estimate: 42,038 59,873
Moderate Insurance Estimate: 3,291,753 4,688,254
Maximum Insurance Estimate: 5,729,735 8,160,532
Scenario 1
Least Regulated Top 5 State - Texas
14. Scenario 2: Competition among States
8,490,592
Status
Quo Mininum Moderate Maximum
Indiviudal
HSA 4,655,291 337,126 7% 1,380,706 30% 1,679,969 36%
PPO High 7,515,552 982,018 13% 4,570,144 61% 7,423,340 99%
PPO Low 180,379 (10,102) -6% (37,231) -21% (52,021) -29%
PPO Medium 1,534,799 39,324 3% 45,805 3% 32,344 2%
Uninsured 28,848,310 (1,348,366) -5% (5,959,423) -21% (9,083,632) -31%
Group Market
HMO 5,505,466 (16) 0% (508) 0% (4,985) 0%
HRA 6,166,134 (157) 0% (1,711) 0% (5,990) 0%
HSA Offered 307,298 (6) 0% (86) 0% (428) 0%
HSA No-offer 11,088 3,780 34% 64,982 586% 353,446 3188%
PPO High 16,535,831 (79) 0% (1,424) 0% (9,120) 0%
PPO Low 665,950 (3) 0% (231) 0% (2,841) 0%
PPO Medium 53,470,814 (3,511) 0% (58,965) 0% (297,398) -1%
Turned Down 3,530,681 (8) 0% (2,057) 0% (32,684) -1%
Within Sample National
Mininum Insurance Estimate: 1,348,374 1,920,411
Moderate Insurance Estimate: 5,961,480 8,490,592
Maximum Insurance Estimate: 9,116,316 12,983,844
Scenario 2
Least Regulated State - Alabama
15. Scenario 3: Competition among States in 4 Regions
7,772,544
Status
Quo Mininum Moderate Maximum
Indiviudal
HSA 4,655,291 264,970 6% 1,220,825 26% 1,546,262 33%
PPO High 7,515,552 815,292 11% 4,230,546 56% 6,879,526 92%
PPO Low 180,379 (8,763) -5% (35,444) -20% (49,259) -27%
PPO Medium 1,534,799 36,709 2% 40,486 3% 26,151 2%
Uninsured 28,848,310 (1,108,208) -4% (5,456,413) -19% (8,402,679) -29%
Group Market
HMO 5,505,466 (12) 0% (301) 0% (2,402) 0%
HRA 6,166,134 (125) 0% (1,467) 0% (4,667) 0%
HSA Offered 307,298 (5) 0% (69) 0% (285) 0%
HSA No-offer 11,088 2,894 26% 48,592 438% 224,457 2024%
PPO High 16,535,831 (60) 0% (996) 0% (5,184) 0%
PPO Low 665,950 (2) 0% (116) 0% (1,264) 0%
PPO Medium 53,470,814 (2,685) 0% (44,738) 0% (196,852) 0%
Turned Down 3,530,681 (4) 0% (905) 0% (13,803) 0%
Within Sample National
Mininum Insurance Estimate: 1,108,213 1,578,364
Moderate Insurance Estimate: 5,457,318 7,772,544
Maximum Insurance Estimate: 8,416,482 11,987,111
Scenario 3
Least Regulated State in 4 Regions - AL,AZ,NE,NH
16. Implications
• Largest insurance take-up is competition
among all 50 states with one winner
• Most pragmatic scenario, with good impact, is
one winner in each regional market
• No way to assess impact of such a migration on
provider access or quality of care
• Significant opportunity to reduce the number
of uninsured in each scenario
18. Responses to Arguments
Against Change in Status Quo
• Insurance is local. Insurers can’t navigate this.
– Self-insurance market success proves otherwise.
• Physician panels and credentialing are local.
– Self-insurance and FEHBP prove otherwise.
• Claims information systems are incompatible.
– Medicare & 1990s vintage IS for pharmaceutical
benefit management firms prove otherwise.
• States’ reflect local cultural/ethical standards.
– Human anatomy, medical science and risk do not.
21. Estimating the Impact Policies to Expand
Private Coverage for New York’s Non-Poor
Uninsured
Sponsored by the New York State Health Foundation
Albany, New York
September 22, 2009
22. Scenarios Modeled
• Removing restrictions on underwriting
– community rating
– guaranteed issue
• Allowing Health Savings Accounts into the market
– Currently, these high-deductible savings plans may not be
sold in the New York State individual market.
• Allowing the purchase of policies issued by insurers
based in and regulated by neighboring states.
• Allow the sale of “mandate lite” plans
23. ARCOLA’s strengths &
weaknesses for task
Strengths
• Peer-reviewed in Health Affairs
• Can be used for federal &
state estimates
• Is based on a microeconomic
model of health insurance
demand published in three
journals
• Is supported by consumer
driven health plan choice, cost
& use
Weaknesses
• Needs survey data from a
state to make estimates –
Zogby provided data for this
analysis
• Has not been bench-tested
with Urban or Columbia
University models with state
data
• Works only through price
effects, but that is the
dominant factor affecting
insurance choice
24. Plan Choices in the Simulation
• Direct Pay Low PPO
– restrictive network
– high co-pay
– 15 percent coinsurance
• Direct Pay Medium PPO
– Lower co-pay and coinsurance than the Low PPO
• Direct Pay High PPO
– lowest co-pay
– no coinsurance
• HSA
– High deductible , low account contribution
25. What is the Impact of Eliminating Community
Rating (CR) and Guaranteed Issue (GI) and
Introducing Health Savings Accounts?
New York Health Insurance Reform Options
2009 Estimates
Baseline Rx New York % Rx New York % Rx New York %
Individual Market Population No GI Change No CR & GI Change No CR & GI Change
& HSAs
Direct Pay - HSA 0 0 N/A 0 N/A 35,383 N/A
Direct Pay - PPO High 16,939 365,817 2060% 766,953 4428% 741,572 4278%
Direct Pay - PPO Low 9,658 8,903 -8% 5,914 -39% 5,648 -42%
Direct Pay - PPO Medium 7,649 31,172 308% 35,786 368% 34,259 348%
Uninsured 2,107,530 1,735,884 -18% 1,333,122 -37% 1,324,915 -37%
Total Direct Pay 34,246 405,891 808,653 816,861
Total Population 2,141,776 2,141,776 2,141,776 2,141,776
The combined effect of No CR & GI is a 37%
reduction in the Number of uninsured in NYS.
26. What is the Impact of Interstate Market
Competition?
If everyone took advantage of lower premiums, there
would be a 26% reduction. A 17% reduction if ¼ buy CT,PA
27. What is the Impact of Reducing the Number of
Mandates in New York?
If 20 mandates were removed, the impact would be a 3% reduction
in the uninsured, 9% reduction if 40 mandates removed.
28. Summary of Simulation Results
• Removing Community Rating & Guaranteed Issue has the
greatest impact on reducing the number of uninsured.
• Introducing HSAs into the market reduces the uninsured,
but does not have nearly the impact of removing CR & GI.
• Letting New Yorkers purchase insurance across state lines
can lead to up a 26% reduction in the uninsured.
• Reducing the number of mandates will have an impact, but
not as great as interstate competition or the removal of CR
& GI.