The Effect of SCHIP Expansions on Health Insurance Decisions by Employers

1
The Effect of SCHIP Expansions on Health
Insurance Decisions by Employers

• Thomas Buchmueller (UC Irvine and NBER)
• Phil Cooper (AHRQ)
• Kosali Simon (Cornell University and NBER)
• and Jessica Vistnes (AHRQ)

2
SCHIP Expansions

• In 2002, SCHIP covered 5 million children
• (Almost) all children in families under 200 of
  the Federal Poverty Level are now eligible
• Compared to earlier federal minimums of 185 for
  infants, 133 for children 1-6, 100 for children
  born after 1983
• 32 states changed eligibility for older children
  (17-18) from lt100 to at least 200 FPL
• Expansions occurred late 1998/early 1999
• Variation in the timing and extent of SCHIP
  adoption across states

3
Did SCHIP affect employers health insurance
decisions?

• Did employers cut back on offers of health
  insurance in low and middle wage firms as a
  result of SCHIP?
• Did employers encourage workers to drop family
  benefits?
• By increasing employee contributions to family
  health insurance plans?
• Did workers drop family benefits?

4
Anecdotal Evidence

• RWJ outreach programs
• Evidence from state benefit managers
• Roll-backs

5
Hypotheses

• Reasons to expect an effect
• Family health insurance premiums were 7500 in
  2001, with employees contributing 1740 of the
  cost
• In 1996 , many children who would later become
  eligible for SCHIP were covered by private
  insurance and had working parents
• Administrative data show several million children
  served by SCHIP

6
Hypotheses

• Reasons to expect no effect
• SCHIP has measures designed to prevent crowd out
• Competitive labor markets of the late 1990s, even
  for low-wage employers
• Take-up of SCHIP has been low, stigma costs may
  exist
• Employers might not know employees total family
  income and employee eligibility for SCHIP

7
Importance of using employer data to answer these
questions

• Employer decisions plausibly driven by the median
  worker, and individual level data do not identify
  employer workforce characteristics
• Problems in identifying affected group in
  individual data
• Employer decisions such as premium contributions
  are not captured well in individual level surveys

8
Findings from previous studies

• Household-based studies
• Cutler and Gruber (1996)
• Dubay and Kenney (1997)
• Shore-Sheppard (1997)
• Yazici and Kaestner (1998)
• Thorpe and Florence (1998/1999)
• Blumberg, Dubay and Norton (1999)
• Cunningham, Hadley, Reschovsky (2002)
• LoSasso and Buchmueller (forthcomming)
• Employer-based studies
• Shore-Sheppard, Buchmueller and Jensen (2000)

9
Methods

• Dependent variables employer and employee level
  health insurance decisions and outcomes
• offer insurance
• offer family coverage
• cost of family
• Also, cost relative to single coverage
• Enrollment rate
• Family enrollment rate
• Affected firms those with substantial share of
  workers newly eligible for SCHIP (other firms
  serve as controls)

10
Econometric Specification

• Identifying sources of variation
• Exploiting time and state variation in
  eligibility expansions
• Model Y B1 Eligibility B2 Percent LOW
  B3 (Percent Low Eligibility) B5 Other
  Characteristics
• Where Y different health insurance outcomes

11
Data

• MEPS-Insurance Component (MEPS-IC)
• List Sample of Private Establishments
• 1997-2001
• Roughly 25,000 private sector establishments and
  their plans in each year
• Employers asked
• Whether they offer insurance
• Fraction of eligible active workers that enroll
• At all
• In single/non-single coverage
• Premiums and contributions for single and family
  coverage
• Workforce characteristics, including information
  on the percent of workers in three different wage
  ranges

12
Data (continued)

• Wages 1997-1999 lt6.50, 6.50-15.00, gt15
• Wages 2000-2001 lt9.50, 9.50-21.00, gt21
• Industry
• Age of firm
• Establishment size
• Percent of workforce that are women, union
  members, over age 50, part time
• For profit status
• Organization type (i.e. corporation)
• State dummies
• Year dummies

13
Data (continued)

• We merge in data on
• SCHIP/Medicaid Eligibility
• Computed from running detailed simulation
  programs (created by Gruber and Simon) on March
  CPS respondents as in Cutler and Gruber (1996)
• We take all March CPS children in 1996, and
  calculate the weighted fraction of them that
  would be eligible for Medicaid or SCHIP in a
  particular state in a particular year. That
  fraction is used as our measure of
  Medicaid/SCHIP generosity. The measure varies by
  state and year.

14
Data (continued)

• Workers most likely to affected by SCHIP
  expansions.
• Using PUMS we construct a proxy measure of the
  percent of families with children under 200 FPL
  by industry and state (LOW).

15
Mean of Eligibility Measure from CPS
16
Data (continued)

• Other data merged onto the MEPS-IC
• Area Resource File
• Unemployment rate
• Per capita income
• Number of active doctors in area
• Number of HMOs in county
• County Business Pattern Data
• Percent of establishments with lt 10 employees, gt
  1000 employees, in manufacturing
• Herfindahl index

17
Results

• Offers and Family Offers
• No significant results on eligibility in either
  equation
• Employers are not reacting to SCHIP through
  offers of coverage or family coverage

18
Results

• Take-up
• For hypothetical establishments with 50 percent
  of the workforce under 200 percent of the poverty
  line, a 0.188 increase in ELIG results in a
• 4 percentage point decrease in overall take-up
  rates.
• 5 percentage point decrease in family take-up.

19
Results

• Marginal Cost of Family Coverage (family
  single).
• For hypothetical establishments with 50 percent
  of the workforce under 200 percent of the poverty
  line, a 0.188 increase in ELIG results in a
• 378 increase in the marginal cost of family
  coverage
• All workers under 200 FPL - 800 increase in
  marginal cost of family coverage

20
Implications

• If employers are reacting to SCHIP expansions,
  what will happen with rollbacks?
• What is the effect of a sluggish economy?

21
Future plans

• Include other measures of SCHIP outreach,
  waiting periods, paperwork simplification
• Use more worker eligibility indices (lt100 FPL,
  200 300 FPL, etc.)
• Additional years