VIEWPOINT
What Do We Really Know About the Impact
Of Price Changes on Contraceptive Use?
By Barbara Janowitz and John H. Bratt
G
rowing demand for family planning services in the developing
world continues to drive up the
total cost of providing services. The United Nations Population Fund recently projected the costs of family planning and reproductive health in the year 2000 at US
$17 billion.1 With ever-larger cohorts entering their reproductive years, it is unclear where the funds to satisfy the demand for services will come from.
Most current funding for family planning programs comes from three principal
sources: government revenues, donor contributions, and fees collected from clients.
Although government funding for family
planning programs may increase in the
near term, most of this support will be targeted to programs run by ministries of
health. Donor funding for family planning
appears to be declining in real terms; in
some countries, the U. S. Agency for International Development (USAID) is phasing
out its contributions.2 Therefore, nongovernmental organizations, which often
depend heavily on donor funds, may have
no choice but to increase fees for family
planning or other services. Even government programs may be forced to institute
or increase user charges for family planning
to offset deficits caused by stagnant tax revenues and competition between government agencies for scarce public resources.
But increased reliance on user fees is
controversial. Lewis’s review of the literature on the impacts of user fees in family planning has shown that demand is
sometimes sensitive to price changes.3
Barbara Janowitz is director and John H. Bratt is senior
research associate, Service Delivery Research Division,
Family Health International (FHI), Research Triangle Park,
N. C., USA. This article is based, in part, on B. Janowitz
and B. Gould, “Options for Financing Family Planning,”
which was funded by the United Nations Population
Fund (UNFPA) and published in Report on Family Planning Program Sustainability, UNFPA Technical Report No.
26, New York, May 1985. Additional support was provided from FHI corporate funds.
38
Fees may reduce access to services, especially among poor women, and may lead
to declines in contraceptive prevalence. In
addition, many programs measure success in terms of the number of clients
served, and so declines in demand resulting from price increases are unacceptable to program managers.
Before establishing or increasing fees for
family planning services, program managers need to know the likely impact of
price changes on demand. Several new
studies have appeared since Lewis’s review, and their authors generally have either used econometric models to analyze
cross-sectional data or attempted to carry
out experimental or quasi-experimental
studies in which demand responses to
price changes are observed over time.
The Modeling Approach
Three studies have used econometric modeling techniques to attempt to show how
the choice of a contraceptive method is affected by a range of variables, including the
method’s price, the socioeconomic characteristics of users and the interactions of
these and other variables.4 The data used
in the models are cross-sectional, and come
mainly from large-scale surveys such as the
Demographic and Health Surveys (DHS).
A general conclusion of these studies is
that the demand for contraception is not
very responsive to price changes. In two
analyses conducted in Thailand using different methodologies, Akin and Schwartz5
and Ashakul6 found inelastic demand for
most contraceptive methods. Jensen, using
an econometric model for Indonesia,
found that the elasticity of demand was
generally low, except for injectable contraceptives.7 However, elasticity of demand is not a constant; for example, Jensen
found larger demand elasticity at higher
private-sector prices than at public-sector
prices. Moreover, Jensen found that the impact of some price changes was greater for
poorer individuals, which should not be
surprising given that any price increase has
a larger relative impact on the ability of the
poor to buy goods and services.
A basic problem with the modeling approach is a lack of consensus on whether
model outputs can be used to predict the
real impact of price changes. Creese, referring to the literature on health, argues
that “professional disagreement about the
appropriate way to model demand reveals
the relative infancy of this analytical approach in the special context of predicting
health behavior.”8 The very same critique
could be made with respect to family planning, a field in which even fewer demand
studies have been conducted. Griffin, who
surveyed the literature on the pricing of
health services, cautions that the modeling
approach, with its reliance on cross-sectional data, cannot account for changes in
consumer behavior.9 “Over time, consumers will adjust their consumption patterns to the new prices, and the higher public health sector prices will elicit a supply
response from private providers.”
Aside from this more general issue, specific methodological problems in the three
studies raise questions about the applicability of study conclusions. For example,
all three models require data not only on
the price of the contraceptive method used
by the client, but also on prices of methods
not used. This information was generated
by asking respondents what they thought
they would pay for methods they were not
using, or by calculating mean prices for
narrowly defined subsamples.10 However, it may not be reasonable to expect that
users of one contraceptive method would
be able to give accurate information on current prices of other methods. For example,
younger women who use family planning
to delay their next birth probably will not
have priced permanent methods such as
tubal ligation and vasectomy or even longacting methods such as the IUD and the
International Family Planning Perspectives
implant. Also, many women who are currently using a long-acting method obtained it several years ago, and therefore
may lack up-to-date information on current prices of reversible methods such as
the pill and the condom.
A second methodological problem is the
practice of converting actual prices paid
by users into prices per couple-month or
couple-year of protection. For the pill, the
price per couple-month of protection is the
same as the price paid each month by the
user for a new supply. But in the case of
long-acting methods such as sterilization,
the user pays the price for the entire period of protection at the time of acceptance.
Thus, although the cost per couple-month
of protection may be very low, the initial
expense to the user is high. The conversion
of actual prices into prices per month effectively obscures the difference between
the immediate cost to the user of accepting sterilization and the cost of adopting
the pill. Therefore, the elasticity estimates
produced by the model may be distorted.
The Experimental Approach
The common characteristic of studies in the
“experimental” category is that they attempt to measure demand responses to actual changes in prices. Methodologies used
range from formal operations research designs with experimental and control groups
to less rigorous pretest-posttest designs to
simple observation of demand response in
a study conducted for another purpose.11
The results of these studies suggest that
demand for family planning services is
sensitive to changes in price. León and
Cuesta used a nonequivalent control
group design to evaluate the impact of
price increases at clinics run by an Ecuadoran nongovernmental organization.12 In
the three months following the price increase, a larger decline in attendance was
documented in clinics that raised their
prices by 48–61% than in those that raised
their prices by 7–16%; in clinics that did
not change their prices, attendance declined but to a lesser degree. Similar findings were reported in a study conducted
in Colombia with PROFAMILIA, which
compared changes in demand for the implant under three different pricing
schemes.13 Clinics in which implant prices
decreased the most experienced the
largest increases in new implant acceptors.
In another study carried out in Ecuador,
a nongovernmental organization increased
prices for all of its services by 20%, and the
demand response was analyzed at a convenience sample of eight clinics.14 The
number of IUD acceptors declined during
Volume 22, Number 1, March 1996
the period immediately following the price
increase, but the same pattern of decline
had occurred in the previous year and
therefore could be attributable to seasonal trends rather than to the price increase.
Finally, a study examined the impact of
funding decreases on access to sterilization
services at nongovernmental family planning clinics in Brazil, Mexico and the Dominican Republic. The investigators reported substantial declines in the demand
for sterilization at 14 of 17 nongovernmental clinics where prices had been increased.15
Although the findings reported in these
four studies are reasonably consistent, certain methodological shortcomings raise
concerns about the validity of their conclusions. In the case of the quasi-experiment carried out by León and Cuesta, each
clinic director decided how much to increase prices; as a consequence, it is not
possible to determine whether changes in
demand occurred because of price changes
or because of the characteristics of the clinics.16 In the PROFAMILIA study, clinic directors may have disagreed about how
vigorously to promote the implant; if promotion strategies were not uniform across
clinics, then the changes in demand could
have been related to variations in marketing efforts as well as to price differentials.
A second issue that receives little attention in these studies is the question of
method and source substitution. Opponents of price increases often argue that
higher prices will lead to lower contraceptive prevalence, especially among poor
women. But when a single provider of
family planning services increases its
prices, clients have three basic options:
They can pay the higher prices, they can
stop practicing contraception or they can
switch to lower-priced methods or
providers. Therefore, although price increases at one clinic may lead to steep declines in the number of clients served by
that particular clinic, the overall impact
on use of family planning may be negligible because clients can substitute other
methods or service providers.
None of the four studies in this category make an effort to measure changes in
demand among substitute providers, although some do acknowledge that other
providers may have absorbed a portion
of the demand. For example, the Haws
study concludes that “the availability of
nearby services may have some impact on
the price elasticity of demand….”17 The
Ojeda study attempts to determine the impact of price changes on use of other methods provided by PROFAMILIA but does
not consider the impact of implant price
reductions on the use of other sources.18
Nevertheless, other work on the implant
has shown that an increase in its distribution causes women to switch sources to
obtain the method.19
Two studies have attempted to determine whether price increases for a particular method or source affected overall
use of contraceptives. Ciszewski and Harvey examined the impact of price increases for condoms and oral contraceptives sold by the social marketing
company in Bangladesh.20 Prices were increased in April 1990 by an average of
60%. Social marketing company sales of
condoms declined by 46% in the following 12 months, while the impact on sales
of the pill was less dramatic. The authors
concluded that the price increases negatively affected not only sales of condoms
and pills by the social marketing company, but the total distribution of condoms
and pills in Bangladesh.
Data from contraceptive prevalence surveys conducted in 1989 and 1991 contradict the study’s assertion of declines in
overall contraceptive use.* During this period, use of the pill increased from 9.1% of
women in union to 13.9% and condom use
increased from 1.9% to 2.5%.21
An important reason for increased use
of the pill and the condom is that the government of Bangladesh substantially increased its employment of field workers
who distribute these methods, thereby establishing a viable substitute to pharmacy distribution.22 The share of resupply
methods provided by pharmacies declined
from 44% in 1989 to 26% in 1991.23 Therefore, while fewer users of the pill or the
condom purchased their method at a pharmacy, the difference was apparently more
than offset by increased provision from the
government program and other sources.
In the second study that examined impacts on the total market, Harvey used data
from 24 countries to demonstrate a negative correlation between condom prices
and sales volume in social marketing programs.24 His recommendation was that
prices should be set low enough that no
more than the equivalent of 1% of per capita gross national product would be needed to purchase 100 condoms. Although he
*Researchers have long noted the discrepancy between
distribution and use figures for the condom (see, for example: G. Ahmed, W. P. Schellstede and N. E. Williamson,
“Underreporting of Contraceptive Use in Bangladesh,”
International Family Planning Perspectives, 13:136–140,
1987). The two series of data for the time period
1989–1991 appear to be especially divergent, as the prevalence figures indicate an increase in condom use, while
the distribution figures cited by Ciszewski and Harvey
indicate a decrease (see: reference 20).
39
The Impact of Price Changes on Contraceptive Use
acknowledged that condoms sold by social
marketing programs are part of a larger
market that includes other condoms and
other methods, he did not attempt any
analysis of substitution patterns. However, given that there are good substitutes for
social marketing condoms, price increases might reduce social marketing sales, but
not necessarily reduce distribution of all
condoms or of all methods. A more thorough analysis, country by country, is required to determine how condom pricing
in social marketing programs affects contraceptive use.
Conclusions
In summary, each of the studies reviewed
here has methodological problems that
undermine the validity of its conclusions.
Thus, because we do not fully understand
the impact of user fees on contraceptive
use, the potential for user fees to expand
resources for family planning in nongovernment and government programs
remains unclear.
The only solution is to conduct well-designed experimental studies that assess
method- and source-substitution patterns
resulting from price changes. An ideal
price experiment would 1) randomly allocate service delivery points for the program under study to an experimental
group that raises prices or to a control
group that keeps prices constant, 2) collect
information on the distribution of all methods and 3) collect information on the distribution of methods by other programs.
The results of studies using this design
40
could provide a basis for action on the part
of governments, nongovernmental organizations and donors as they seek to ensure sufficient funding for family planning
services in the future. Because programs
may soon have to make decisions about
whether and how much to increase prices,
such research should be given a high priority.
References
1. United Nations Population Fund (UNFPA), Report of
the International Conference on Population and Development,
A/CONF.171/13, New York, Oct. 18, 1994.
2. ——, Global Population Assistance Report 1982–1991,
New York, 1993.
10. J. S. Akin and J. B. Schwartz, 1988, op. cit. (see reference
4); and E. R. Jensen et al., 1993, op. cit. (see reference 4).
11. F. R. León and A. Cuesta, “The Need for Quasi-Experimental Methodology to Evaluate Pricing Effects,”
Studies in Family Planning, 24:375–381, 1993; G. Ojeda et
al., “Effects of a Credit System and Price Reductions on
Norplant Demand and Revenues,” paper presented at
the annual meeting of the American Public Health Association, San Francisco, Calif., USA, Oct. 24–28, 1993;
J. H. Bratt et al., “Impacts of a Price Increase in Eight
Ecuadorean Family Planning Clinics,” final report, Family Health International and CEMOPLAF, Research Triangle Park, N. C., USA, 1992; and J. Haws et al., “Impact
of Sustainability Policies on Sterilization Services in Latin
America,” Studies in Family Planning, 23:85–96, 1992.
12. F. R. León and A. Cuesta, 1993, op. cit. (see reference 11).
13. G. Ojeda et al., 1993, op. cit. (see reference 11).
14. J. H. Bratt et al., 1992, op. cit. (see reference 11).
3. M. A. Lewis, “Do Contraceptive Prices Affect Demand?” Studies in Family Planning, 17:126–135, 1986.
15. J. Haws et al., 1992, op. cit. (see reference 11).
4. J. S. Akin and J. B. Schwartz, “The Effect of Economic Factors on Contraceptive Choice in Jamaica and Thailand: A Comparison of Mixed Multinomial Logit Results,” Economic Development and Cultural Change,
36:503–527, 1988; T. Ashakul, “Determinants of Contraceptive Method Choice for Women in Municipal and
Non-Municipal Areas of Thailand,” Thailand Development Research Institute, Bangkok, 1990; and E. R. Jensen
et al., “Contraceptive Pricing and Prevalence: Family
Planning Self-Sufficiency in Indonesia,” paper presented at the annual meeting of the Population Association
of America, Cincinnati, Ohio, USA, Apr. 1–3, 1993.
17. J. Haws et al., 1992, op. cit. (see reference 11).
5. J. S. Akin and J. B. Schwartz, 1988, op. cit. (see reference 4).
6. T. Ashakul, 1990, op. cit. (see reference 4).
7. E. R. Jensen et al., 1993, op. cit. (see reference 4).
8. A. L. Creese, “User Charges for Health Care: A Review
of Recent Experience,” Health Policy and Planning,
6:309–319, 1991.
9. C. C. Griffin, “User Charges for Health Care in Principle and Practice,” Economic Development Institute Seminar Paper No. 37, World Bank, Washington, D. C., 1988.
16. F. R. León and A. Cuesta, 1993, op. cit. (see reference 11).
18. G. Ojeda et al., 1993, op. cit. (see reference 11).
19. B. Janowitz et al., “Introducing the Contraceptive Implant in Thailand: Impact on Method Use and Costs,” International Family Planning Perspectives, 20:131–136, 1994.
20. R. L. Ciszewski and P. D. Harvey, “Contraceptive
Price Changes: The Impact on Sales in Bangladesh,” International Family Planning Perspectives, 21:150–154, 1995.
21. S. N. Mitra, C. Lerman and S. Islam, Bangladesh Contraceptive Prevalence Survey, final report, Mitra and Associates, Dhaka, Bangladesh, 1993.
22. S. Hussain et al., “Recruiting Appropriate Female
Field Workers: Experiences From a National Recruitment
Process in Bangladesh,” Working Paper No. 59, MCHFP Extension Project, International Centre for Diarrhoeal
Disease Research, Bangladesh, Dhaka, 1991.
23. S. N. Mitra, C. Lerman and S. Islam, 1993, op. cit. (see
reference 21).
24. P. D. Harvey, “The Impact of Condom Prices on Sales
in Social Marketing Programs,” Studies in Family Planning, 25:52–58, 1994.
International Family Planning Perspectives