Regulations in Nursing Home Markets: Estimation of a Competitive Model

1

• Regulations in Nursing Home Markets Estimation
  of a Competitive Model
• Andrew Ching
• University of Toronto
• Fumiko Hayashi
• Federal Reserve Bank of Kansas City
• January 6, 2007
• The views expressed in this paper are those of
  the authors and do not necessarily reflect those
  of the Federal Reserve Bank of Kansas City or the
  Federal Reserve System.

2

• Issues
• - Access to care, especially Medicaid consumers
• - Quality
• - Price
• Regulations
• - Certificate of Need (CON) restricts supply of
  beds
• - Medicaid reimbursement rates are set by
  government
• - Quality should be common to all patients in a
  given nursing home

3

• Objective
• Construct a structural model
• - Takes into account key features of the industry
• - endogenous quality
• - rationing
• - Enables us to
• - quantify the extent of rationing
• - estimate price/quality elasticity of demand
• - investigate how market outcomes changes with
  bed availability
• - (simulate policy changes)

4

• Model (Key assumptions)
• Two demands Private-pay (P,Q,D) Medicaid (Q,D)
• A NH sets private-pay price and quality
• Medicaid price is exogenous (typically it is
  lower than private-pay price)
• A NH cannot change capacitysome NHs face a
  binding bed constraint
• A NH admits private-pay patients first, then
  Medicaid patients

5

• Model (Demand-side)
• - Private-pay consumer is utility of entering NH
  j
• - Medicaid consumer is utility of entering NH j
• - Quality
• - There is an outside alternative
• - Private-pay consumers choose NHs first
• - Medicaid consumers are the residual claimants

6

• Model (Medicaid demand)
• Medicaid consumers are randomly ordered to choose
  an alternative to maximize their utilities
• Due to the binding bed constraints, some Medicaid
  consumers may not be able to choose their most
  preferred nursing homesthose consumers choose
  the best nursing home from nursing homes with
  available beds
• Since private-pay consumers choose nursing homes
  first and the number of private-pay patients is
  small, they do not face this rationing problem

7

• Model (Supply-side)
• - Set price and partial quality (q)
• - Maximize profit
• w/BBC
• w/o BBC
• - Cost function

8

• Data
• 1) 1998 Wisconsin Annual Survey of Nursing Homes
• - NH characteristics
• - grouped patient characteristics for each NH
• by age-gender by payment-gender county
• 2) 1990 Census of Population and Housing Special
  Tabulation on Aging
• - population by age, gender, income level, and
  county

9

• Estimation
• - Estimate demand- and supply-side separately
• Demand use GMM3 sets of moments, 2 enter GMM
• 1) match the observed private-pay demand with
  the predicted demand to obtain Q
• 2) choose parameters in utility fun. to match
  the micro moments ( of patients by age-gender,
  by payment-gender, by county)
• 3) assuming ? is mean independent of X, regress
  Q on X to obtain ? and ?
• Supply utilize FOCs and demand-side equations
• and estimate cj and ?cj/ ?qj
  jointly

10

• Results
• - Most parameter estimates are as expected
• - utility ? as price ?, quality ?, distance ?
• - quality ? as staff intensity ?
• - cost ? as partial quality (staff intensity) ?
• - Interesting implications
• - 109 NHs are estimated to have a BBC
• - 11 of Medicaid consumers cannot enter
  their first-choice NHs (9 enter other NHs)
• - relationship b/w BBC and quality/price
• - elasticity and substitution patterns

11
Demand-side Equations
12
Supply-side Equations
13
Effects of Price and Quality Changes on Demand
14

• Conclusion
• - A structural model is useful in
• - quantifying the extent of rationing
• - measuring price/quality elasticity of demand
• - examining how firms compete as the extent of
• rationing increases
• - Future Research possible policy experiments
• - Change in CON
• - Change in Medicaid reimbursement rates