Equity, then Social Insurance

1
Equity, then Social Insurance
2
What is equity?

• Horizontal equity -- Equals should be treated
  equally. Usually thought of with respect to
  taxes.
• Vertical equity - Unequals should be treated
  unequally. For example, rich should pay more
  than poor, but should rich pay higher
  percentage than poor?
• General sense of equity in health care has to do
  with everyone getting some.

3
Access v. Utilization

• Possibly unique to the health field, we break
  these up.
• Do people have access to care, even if they dont
  use it?
• This is different from utilization, which refers
  to visits, days, etc.

4
Equity in the Delivery of Health Care Some
International Comparisons

• van Doorslaer and Wagstaff argue that most accept
  the premise that health care should be financed
  according to ability to pay, but should be
  delivered according to need. How do we address
  equity in the delivery of health care.
• They note that there is some contention between
  those who want to use utilization as a measure,
  and others who want to use access as a measure.

5
Equal treatment for equal need

• Link between provider incentives and patient
  income is likely to be important in determining
  how close a health care system comes to achieving
  equal treatment for equal need.
• The most obvious way that the income of a patient
  might influence provider behavior is if the
  provider is paid differently depending on whether
  the patient is being treated publicly or
  privately.
• General proposal Assessment of equity in the
  delivery of health care ought to be based on
  treatment differentials adjusted for differences
  in need.

6
Tests for inequity

• Tests for inequity
• Medical Expenditures
• ?p ?phi upi Poor
• mi
• ?r ?rhi uri Rich
• ?p, ?r are expected expenditures if healthy ?p,
  ?r are increments if you get sick.
• Test
• Equity if ?r ?p, ?r ?p.
• You want, of course, to adjust this stuff for
  demographics, health status, etc.

7
Equity and Inequity
Inequity
Equity
expenditures
expenditures
?r
?r
?p
?p
h
h
1.0
1.0
health
health
8
2 part models

• You would want to measure these with 2 part
  models
• Do people use care (or have access to care)
• If they use (have access), how much do they use,
  spend, etc.
• Also propose a type of Gini coefficient that
  basically works across income classes.
• They ask How much would each income group have
  received on average if its age structure and
  morbidity levels had been the same as those of
  the population at large?
• Use the expenditure shares based on the equations
  above to estimate shares for groups of the
  population ordered by income levels.

9
Gini Coefficients

• Not quite a standard Gini. Suppose we have this
• Income quartile Expenditures Cumulative
• 0-25 15 15
• 25-50 35 50
• 50-75 35 85
• 75-100 15 100

• Standard Gini would be 15, 15, 35, 35 Comes
  out to be 0.2. You should calculate it.

Cumulative
Lowest Expend.
10
Gini Coefficients

• Not quite a standard Gini. Suppose we have this
• Income quartile Expenditures Cumulative
• 0-25 15 15
• 25-50 35 50
• 50-75 35 85
• 75-100 15 100

• This Gini turns out to be 0. If spending favors
  rich, curve will be below diagonal -- Gini gt 0
  if vice versa, Gini lt 0.

Cumulative
Income Quartile
11
Table 20.1

• Country Measure of Inequality Statistical Test
• Denmark -0.055 Yes
• Ireland -0.076 Yes
• Italy -0.036 Yes
• Netherlands 0.025 No
• Spain 0.146 Yes
• Switzerland -0.043 No
• UK 0.013 Yes
• US 0.028 Yes
• Positive favors rich Negative favors poor.

12
Conclusions

• Persons in equal need in U.S. are not treated the
  same, although this finding depends on whether
  one examines only the under-65 (in which case the
  inequity seems to favor the worse off), or the
  entire adult population (in which case the
  inequity favors, on balance, the better off).
• Appears to be inequity favoring the well-off even
  in some countries where there is universal and
  comprehensive public coverage, notably Spain and
  U.K.
• Health care systems which do not have universal
  and comprehensive public coverage are not
  necessarily those with the highest degree of
  inequity.

13
Calculating Lorenz Curves

• Cumulative Percentage of Population F
• Cumulative Percentage of Care F1.
• Define ? (F - F1)/ ? 2
• ? (F F1)/ ? 2
• Define Lorenz curve function as
• ? a?? (? 2 - ?)?
• In logs, estimate regression
• log (?) log (a) ? log ? ? log (? 2 - ?)
• Gini G is
• G 2 (3 ? ?)/2 a B(1?, 1?),
• Where B is the statistical Beta function.