Transcript An Institutional and Econometric Study of Health Care in China

An Economic Analysis of
Health Care in China
Gregory C Chow
Princeton University
June 8, 2006
Outline
1. Introduction
2. Changes in Health Care System
3. Demand Functions for Health Care
4. Government’s Program for Health Care
5. Supply of health care and Future Development
6. Evaluation of the Current Health Care System
7. Conclusion
2. Changes in Health Care System
§Institutions before 1980’s

A cost-effective three-tear health care system
improved the health of the Chinese people:
.reduction of diseases
.decline in the annual death rate
17 per 1000 population in 1952→6.34 per 1000 in 1980
.increase in life expectancy
early 1950s: 40.8 years→ early 1960s: 49.5 years → late 1970s:
65.3 years
§ Institutions since 1980’s

Rural:
.Privatization of farming led essentially to the
abandonment of public health provided by the government.

Urban:
.Privatization of state-owned enterprises was a very slow
process that took over two decades.
.The government tried to provide a substitute for the public
provision of health care through the state-owned
enterprises.
§ Health Care Expenditures and
Funding Resources
Year
Total
(100
million)
Government
Budgetary
Social
Expenditure
Resident
Individual
Percent
Government
Percent
Individual
1995
2257.8
383.1
739.7
1135.0
17.0
50.3
1996
2857.2
461.0
844.4
1551.8
16.1
54.3
1997
3384.9
522.1
937.7
1925.1
16.4
52.8
1998
3776.5
587.2
1006.0
2183.3
16.0
54.8
1999
4178.6
640.9
1064.6
2473.1
15.8
55.9
2000
4586.6
709.5
1171.9
2705.2
15.5
59.0
2001
5025.9
800.6
1211.4
3013.9
15.9
60.0
2002
5790.0
908.5
1539.4
3342.1
15.7
57.7
2003
6584.1
1116.9
1788.5
3678.7
17.0
55.8
§Health Care Expenditures and
Funding Resources
Health Care Expenditure
100 million
7000
6000
5000
4000
3000
2000
1000
0
1995
1996
1997
1998
total expenditure
social expenditure
1999
2000
2001
2002
2003
government budget
resident individual
year
3.Demand Functions for Health Care
§ Estimation Using Time Series Data

The amount of health care services measured in
1995 prices q = health care expenditure /relative
price index of health care service table

Regression of lnq on lny and lnp based on the
9 annual observations from 1995 to 2003 yields:
lnq =1.194(.382) lny–0.730(.241) lnp–4.831(4.027)
R2/s = 0.620/.0447
----- (1)
next
Time-Series Data on Aggregate Demand for Health Care
Year
Consumer GDP
Price
Index
Price
index of
healthcare
Government
revenue
Total
consumption
expenditure
Quantity
of health
services
1995
3.028
58478.1
1.000
6242.20
33635.0
2257.8
1996
3.279
67884.6
1.124
7407.99
40003.9
2542.0
1997
3.371
74462.6
1.381
8651.14
43579.4
2451.0
1998
3.344
78345.2
1.619
9875.95
46405.9
2085.5
1999
3.297
82067.5
1.808
11444.08
49722.7
2311.2
2000
3.310
89468.1
2.009
13395.23
54600.9
2283.0
2001
3.333
97314.8
2.220
16386.04
58927.4
2263.9
2002
3.306
105172.3
2.402
18903.64
62798.5
2410.5
2003
3.346
117390.2
2.616
21715.25
67493.5
2516.9
§

Estimating Income Elasticity with
Cross-section Data
Regressing the log of medical expenditure per
capita on the log of total expenditure per capita
yields table:
total expenditure elasticity se
Urban
1.080
0.023
Rural
1.003
0.023

Adj-R2
0.9981
0.9980
Corresponding data for 2003 yield similar total
expenditure elasticities.
next
Cross-section data on per capita health expenditure and total expenditure in 2002
Low income
households
Lower
Middle
income
households
Middle
income
households
Upper
middle
income
households
High
income
households
Urban:
Total
expenditu
res
3259.59
4205.97
5452.94
6939.95
8919.94
Medicine
and
medical
services
225.67
286.56
382.83
510.15
657.33
Rural:
Total
expenditu
res
1006.35
1310.33
1645.04
2086.61
3500.08
Medicine
and
medical
services
57.57
74.88
90.73
116.49
201.72
§ Price Elasticity by Combining
Cross-section and Time Series Data

Take an average of 1.080 and 1.003 or 1.042 as our
estimate of income elasticity of demand for health care,
which is close to the estimate based on time series data
alone as reported in equation (1)

Use time series data to estimate the price elasticity :
(lnq -1.042 lny) = -0.636 (.047) lnp - 3.228 (.033)
R2/s = 0.9637/.04192

Price elasticity is 0.636
----(2)
§ Income Eelasticity by Provincial Data
for Urban and Rural Residents

Adding lnp to both sides of equation (1) yields
ln(pq) = c + a lny + (1- b) ln p + e

---- (3)
If the lnp on the right-hand side of (3) is uncorrelated with
lny , using provincial data on health care expenditure from
CSY 2005, we have:
Urban: ln(pq) = -2.237(1.415) + 0.919(0.154) lny R2 =0.5501
Rural: ln(pq) = -4.434(1.299) + 1.162(0.163) lny R2 =0.6379

The average of the above two income elasticities is (0.919
+ 1.162)/2=1.041.
§ Inequality in Health Care Spending
from Regression Analysis

s(lnpq) = (a/R)s(lny)

For urban residents across provinces, the factor a/R
equals 0.919/0.742 or 1.239. For rural residents it is
1.162/0.799 or 1.454.(in 2004)

Inequality in medical expenditure is larger than inequality
in income across provinces for both urban and rural
residents.

The ratio of inequality for rural residents is higher partly
because the rural residents have a higher income
elasticity of demand for medical expenditure.
4.Government’s Program for Health Care
§ On Demand Side

"Decision on Health Reform and Development by the
Central Party Committee and State Council." (January
15, 1997)

Basic objective : to insure that every Chinese will have
access to basic health protection.

Rural : to develop and improve CMS through education,
by mobilizing more farmers to participate and gradually
expanding its coverage; 40 yuan subsidy per account.

Urban：a basic medical insurance system was
established in 1998, financed by 6％of the wage bill of
employing units and 2％ of the personal wages.
§

On Supply Side
In 2004 the government is in the process
of allowing some hospitals in urban and
rural areas to be run privately to reduce
the burden to the government.
5. Supply of Health care and
Prospects for Future Development
§ Constant Supply

The amount of health care supplied remained
approximately constant between 1989 and
2003(as with the quantity q in Table 2).
1989
1997
2002
2003
# of Hospital Beds
per 10 000 Population
22.8
23.5
23.2
23.4
# of Doctors
per 10 000 Population
15.2
16.1
14.7
14.8
Change
of No. of
Doctors
and No.
Graduat
es
Year
Number
of
Doctors
1000’s
Number of
Graduates
1000’s
Retirees
(1/35 No.
in year
before)
Estimated
Increase in
No.
Doctors
Actual
Increase
in No.
Doctors
Implie
d % of
Retire
ment
1997
1985
61.239
1998
1999
61.379
56.714
4.665
14
.02387
1999
2045
61.545
57.114
4.431
46
.00778
2000
2076
59.857
58.429
1.428
31
.01411
2001
2100
62.638
59.314
10.738
24
.01861
2002
1844
79.500
60.000
3.324
-256
.15976
2003
1868
111.356
52.686
58.67
24
.04737
2004
1905
154.187
53.371
100.816
37
.06273
§ Shift of Health Resources from Rural
to Urban Population

In 2001 the number of health clinics in villages and
townships was reduced by 1139; the number of doctors
and health care personnel was reduced by 30,000.

From 1990 to 2000, government spending in total health
care spending in rural areas was reduced from 12.5
percent to 6.6 percent.

The shifts in relative demand in favor of urban residents
who could afford to pay and received more government
funding for medical care resulted in the shifts of supply to
the urban residents at the expense of rural residents.
§ Forecast of Rate of Increase in the
Supply of Doctors

Assuming the number of doctors in the next few years to be 2400
thousand (with 160 thousand graduates per year, and number of
graduates to be 200 thousand per year.

The number retired will be 2400/35 = 68.57 thousand, resulting in a
net increase of 200 – 68.57 = 131.43 thousand, or a rate of increase of
131.43/2400 = 0.05476.

After subtracting annual population increase of 0.006 we obtain a
rate of increase of 0.049. This is substantially less than the increase
in demand due to increase in real income.
§ Explanation of Rapid Increase of
Health Expenditure

Taking the derivative of equation (3) with respect to time
we have
dln(pq)/dt = 1.042 dlny/dt + (1-0.636)dlnp/dt
---- (4)
where, dlny/dt = (ln 35083.7- ln19312.4)/8 = 0.0746
dlnp/dt = (ln1 - ln 2.36738)/8 = 0.1077.

The sum of the income effect 1.042(0.0746) = 0.0777
and the price effect (1–0.636)(0.1077) = 0.0392, yields a
total of 0.117 for the exponential rate of increase in
medical expenditure per year.
§ Estimate Rate of Increase in Price
and in Expenditure

By assuming the exponential rate of growth for real GDP
to be 0.08 and rate of growth in the quantity supplied is
zero, solving the equation 1.082(0.08) – 0.646 x =0
yields a rate of increase in the relative price of health
care equal to 0.134 ( about as large as the average
annual rate of increase in the period 1995-2003).

If the rate of increase of medical supply is 0.049 per year,
the rate of increase in price will be 0.058, instead of
0.134.
§ Explanation of Increase in the Ratio
of pq to GDP

As a fraction of GDP health care expenditure increased
from 3.86%in 1995 to 5.61% in 2003.

Adding lnp to and subtracting lny from both sides of the
equation (2) to yield:
ln(pq/y) = 0.042 lny + 0.364 lnp – - 3.228
----(5)

The expenditure for health care as a ratio of income or
GDP increases as a result of the income term if income
elasticity is larger than unity. This effect is very small.

The ratio increases due to the second or price term in the
demand equation if price elasticity is less than unity.
6.Evaluations of the Current System
－ Unequal Treatment of Urban and Rural Population

Urban: the government has assisted the working population in the
transformation to the current system of insurance financed.
Urban: medical care is mostly publicly supplied

Rural: services privately supplied and market determined.
Government has encouraged and assisted the rural population to
organized CMS as a collective medical care system, covering about
100 million/800 million rural population, but in 2006 it has introduced a
government subsidy of 40 yuan per person with 10 yuan contributed
by rural resident; fund will be pooled to pay for medical expenses.
Why effect of government subsidy on
demand for rural healthcare limited

1. government subsidy of 40 yuan per person

2. the relative demand of rural and urban
may be substantial as compared with existing
healthcare spending per capita but it will not
increase total demand substantially because it is
a substitute for private spending.
population for healthcare will not be affected
substantially. Because of the more rapid rate of
increase in urban income per capita the relative
force of demand will shift in favor of the urban
population.
Government subsidy and provision for healthcare
to rural residents can increase welfare
substantially


if government spending is used to (1) pay for
medical insurance (a) to cover only the major
illnesses of the rural population, or (b) to insure
the rural population with below median income,
or (2) to operate clinics almost free of charge,
the effect on rural healthcare can be improved
substantially.
A government program of social insurance or
healthcare provision can have an important effect
because many rural residents may not voluntarily
buy such an insurance as they pay for medical
expenses only when they are seriously ill and
when it is often too late. In some villages clinics
are not available and farmers cannot get
healthcare even if they are willing to pay for it.
Possible improvement in the
Management of Health Care

Medical insurance only to pay for large expenses. No
insurance or high co-payment for small expenses

Incentive payment for physicians in public hospitals

Leasing of public hospitals for private management given
same subsidy

Encouragement of private hospitals
7. Conclusions





We have estimated an income elasticity of demand for health
services to be unity for urban population and slightly above unity
for rural population, and a price elasticity of about 0.6 by
combining cross-section and time-series data.
Demand analysis can explain the increase in expenditure on
healthcare and the increase in price as income increases given
limited supply. It also explains the increase in the ratio of health
expenditure to GDP.
There is large inequality in health expenditure per capita between
the urban and the rural population associated with income
inequality.
Rapid increase in income and government support account for
much better healthcare for the urban population.
A market economy in rural China fails to provide as much health
care as under the former collectively managed and collectively
paid system. The government is attempting to reintroduce
features of this system, with results yet uncertain.
THANK YOU