Transcript Expanding the Role

National Health Accounts
Joseph P. Newhouse
Harvard University
Main Points
 Should account for non-market inputs,
especially time
 Comparisons of spending across time and
space can yield useful inferences
 Decomposing change in medical spending
into price and quantity requires
measurement of output by episode
A Caveat
 My experience is with the US accounts, and
my examples reflect a developed country
bias
 But I think the conclusions apply generally
Non-Market Transactions
 The accounts measure goods and services
traded in the market
True of both health accounts and national
income and product accounts (NIPA)
 Latter often used to measure changes in well
being

Well Being and Non-Market
Transactions
 Time is an important input into health care,
but time has an opportunity cost that is not
captured in the accounts
Time as a Complement
 Time is sometimes a complement to market
inputs
Own time spent traveling to and receiving care
 Time of family members assisting others

–

Mother taking child to physician
Time spent recovering from illness (“Take 2
aspirin and go to bed”)
Time Making Production of
Health More Efficient
 This is a role usually assigned to education
 But people spend time trying to get more
health out of a given set of market inputs
For example, time spent talking with others
about providers of care or otherwise seeking
information
 Time spent gathering information on health
effects of lifestyles; health sections in the press

Time as a Substitute for Market
Inputs
 Informal care of frail elderly
 Health promotion; wellness (e.g., exercise)

Difficult boundary lines here (e.g., sleep)
Measuring Time Used in
Production of Health
 Suppose one wanted to add time to a
satellite account; this would require separate
time use survey
 Issues of valuation; persons not working
 Issues of boundaries
 Joint production

Exercise might have other benefits
Conclusion on Time
 The accounts understate by an unknown,
but probably non-trivial amount the
resources devoted to health care
 Recent NAS publication on satellite
accounts including time inputs; see next
slide (book also covers medical price
indices)
A Recommended Book
Beyond the Market: Designing Non-Market Accounts for the United States; Washington: National Academy Press, 2005.
Usefulness of Accounts
 Some would cite comparing levels of
spending across countries

Sometimes such comparisons have arguably
had an effect; e.g., UK decision to increase
spending to OECD average
Rates of Change
 Within country one can not only calculate
share of GDP (already available from
NIPA), but how rate of change varies
among health care sectors
For example, share of spending going to
pharmaceuticals
 But public sector spending known from budgets

Comparative Rates of Change
 I have found comparative rates of change
useful

I am struck by the similarity of rates of change
both across countries and over time
Annual Real % Cost Increase per
Capita, G-7*, 1960-2002*
8
7.1
Real % per Year
7
6
Average=4.9%
5.3
5.1
5
4
4.2
4.0
3.4
3
2
1
0
Can
Fra
Ger
Jap
UK
US
% Annual
Increase in Real
Personal Health
Care Spending
per Person,
1960-2002
Country
*Italy missing data before 1990. Germany 1970-2002, Japan 1960-2001. Source: OECD Health Data 2004 and US GDP deflator.
Similar Increase in Real US
Annual $/Person by decade
Medicare and Medicaid enacted
%/person/yr (real)
7
6.2
6
5
4
5.3
5.2
Average = 4.4%
4.5
3.7 3.7
3.0
3
2
1
0
% Annual
Increase in Real
per capita
Personal Health
Care Spending
40s 50s 60s 70s 80s 90s 0003
Decade
Managed care
Sources: CMS National Health Accounts. Newhouse, JEP 1992(3), Stat Abst, Ec Rpt Pres. GDP Deflator.
What Do These Data Tell Us?
 Any explanation of the cost increase in
medical care needs to hold across countries
and decades
Differences among countries in financing
institutions are not the explanation
 Costly advances in medicine explain much of
the increase and probably will continue

Costly advances: Newhouse, Jnl Econ Perspectives, 1992.
The Increase Was Probably
Worth It
 The roughly similar rates of increase
everywhere are a crude market test
 In US case confirmed by Cutler: CVD and
neonatal mortality advances alone can
justify the entire US $ increase post 1950
 Nordhaus: Value of US Δlife expectancy
1900-95  Value of ΔNational Income
Cutler, Your Money or Your Life, Oxford, 2004; Nordhaus: The Health of Nations; NBER, 2002, W8818.
A Question to Ponder
 Would you rather have 2005 health levels
and 1955 incomes or 1955 health levels and
2005 incomes?

No formal survey, but Nordhaus’ informal
survey suggest many opt for the former,
consistent with his finding
–
Choice of former goes up with age
Defects of Current Price Indices
 Current medical price indices suggest much
of expenditure increase is a price increase
Implies falling productivity in medical care
 Sometimes used to justify expenditure caps

 But official price indices are badly biased
upward for many reasons, including the
omission of health gains
Price index bias: Berndt et al., Handbook of Health Econ; Newhouse, NBER W8168, Academia Ec Rev March 2001.
Toward Better Price Indices
 Need to construct price indices from Δcost
of episode and Δoutcomes

Price indices based on medical inputs such as
MD visit cannot account for Δquality of care
–

For example, better scanner looks like Δprice
Heart attack work suggests falling price of heart
attack treatment; need to carry out similar work
for other conditions
Heart attack price: Cutler et al., QJE, November 1998.
Conclusions
 Useful expansion of National Health
Accounts to measure time used in the
production of health
 Comparative measures across countries at a
point in time and within countries across
time can yield useful inferences
 Need to base price indices on episodes, not
prices of medical care inputs