Is there an 'unmet need' for birth control


This page in:

Update: Lant Pritchett has kindly responded to my invitation and posted his thoughts: "No need for unmet need." Check out the comments section.

A water cooler conversation (yes, geeky conversations about development are more common than VCU-Butler matchup in the Final Four around here, sigh…) I was having with a couple of colleagues last week made me wonder if the answer to the question in the title of this post is not clear. One of my colleagues, an economist, started by saying that she came into this thinking that ‘unmet need’ was a hyped-up notion with no good evidence behind it – a position she (and I) suspected is held by many economists: we think some demographers made this stuff up!

She then added that she has now come to think that there may be something to the notion that women in some developing countries would like to use contraceptives but that they simply cannot find or access them. She specifically said, if my memory serves me right, that if we were to increase the supply of birth control methods in these places, the take-up rates would go up. I cannot remember whether she then drew the conclusion that unwanted pregnancies would go down as a result: that would be the natural conclusion one might draw.

Well, a recent paper by Ashraf, Field, and Lee suggests otherwise in the case of Zambia. In a clever experiment that they implemented in a health clinic in Lusaka, they had two treatment arms (in addition to a control group who just received information): individual counseling, where women were counseled alone and given vouchers for free access to long-term and reliable methods of birth control (Depo-Provera and Jadelle); and couples counseling where the vouchers were physically handed to the husband.

Most of the paper is about the comparison of the two treatment arms (more on that below), but the answer to the question of increasing supply is answered by examining the outcomes in the couples counseling arm vs. the control group: while the take-up of different birth control methods were up 15 percentage points in the couples treatment arm, the prevalence of unwanted births was identical in the two groups. This suggests that the women who tried the new methods (mainly injectables and pills) were already managing to avoid unwanted pregnancies with the methods available to them before the intervention. The authors conclude:

“The policy implications of this comparison are straightforward. Increasing access while requiring spousal consent will not reduce excess fertility in settings like urban Zambia where modern contraceptives are already reasonably though by no means freely available. Though doing so is likely to change patterns of utilization towards more convenient and reliable long-acting methods, those positioned to take advantage of better access will be couples already in control of fertility through existing - and perhaps even traditional - methods. In sum, it appears that excess fertility in these settings is not driven by the high cost of birth control given that reducing direct and indirect costs had no impact on unwanted births. In contrast, evidence from our experiment indicates that technologies or policies that shift control of fertility from men to women are likely to reduce excess fertility and unwanted births, though with a welfare cost to men.”

The last sentence draws from the comparison of the individual vs. the couples treatments. The authors found that women who received counseling alone were 28% more likely to use a concealable form of contraception that led to a 57% reduction in unwanted births (defined as births to women who reported, at baseline, not wanting to give birth within the next two years). The authors suggest that most of this effect was due to women hiding the use of the voucher from disapproving husbands.

The paper is worth your read, not the least because it pulls off a difficult feat that has caused some tension in the IE field with the recent emergence of RCTs. Some papers reveal a previously unknown behavioral parameter using a clever experiment. Other papers try to address very policy-relevant questions while not necessarily providing any new insights into human behavior. But, only a small number of studies manage to do both and I feel that this paper is one of them. Not only do the results imply that there are substantial inefficiencies in the intra-household bargaining between husbands and wives who disagree on the number of total children they want (causing sub-optimal strategies for both), but also make the straightforward point that simply increasing the supply of reliable and long-acting birth control methods may not reduce unwanted pregnancies unless we also deal with the difficult issue of spousal consent.

Of course, this is only one study from Zambia. My colleague might still be right for many other settings. Do you know of any other studies that would suggest otherwise?

Next week, I plan to ruminate about speed dating as an instrument for marriage. After that, I plan to ponder whether running too many experiments will cause many development economists to burn out sooner than ideal. Stay tuned…


P.S. My friend and fellow blogger Jed Friedman, while commenting on a draft of this post, wrote to me that many studies find the demand for contraceptives to be highly price inelastic, suggesting that "unmet need" is not empirically supported. He pointed me to this paper by McElvey, Thomas, and Frankenberg and the literature cited therein.



Berk Özler

Lead Economist, Development Research Group, World Bank

Join the Conversation

Karen Grepin
April 07, 2011

Grant Miller has a paper that has looked at a related question in Colombia using the roll-out of family planning programs in the country during the 1960s. He finds a small significant effect. However, the effect is small relative to overall fertility declines. Of course the intervention was a broader package of activities, not just providing contraception itself and it was 50 years ago but it leads to a similar conclusion: just making services more readily available may not lead to big changes in fertility.…

William Savedoff
April 07, 2011

The overall point may be true, and the McElvey paper clearly supports it. But I found it somewhat odd that you use the Ashref et al paper as if it were further support for the general critique of unmet need. You focus on the demonstration that availability isn't really the issue. Fine. But the fact that women counseled by themselves and given vouchers for concealable contraceptives had fewer births is very strong evidence of "unmet need" for birth control in this particular society. The study not only reveals the unmet demand but also the conditions under which women can exercise their preferences - and the conditions (influenced or controlled by partners) when they can't.

So, you're right in the sense that it doesn't support the strategy of reducing prices and increasing availability. However, it does demonstrate the value of finding strategies to address womens' social subordination directly or, as a second best, to provide access to concealable contraceptives.

Berk Özler
April 07, 2011


Thanks for the comment. I don't think we have a difference in interpretation. The conclusions you've drawn are clear from my post. Perhaps you disagree with the amount of space devoted to the comparison of the two treatment arms vs. the comparison of couples treatment and the control group. As you rightly pointed above, these comparisons lead to two different and equally important conclusions. The paper is mostly about the former, while my post mostly about the latter.

However, I should say that I personally have never heard of 'unmet need' being discussed to mean "women need contraceptives but their husbands won't allow them." It usually refers to availability and price, changes in which would not reduce unwanted pregnancies -- as the authors clearly state in the paper.

Let's hope that the term 'unmet need' for contraceptives indeed gets replaced by 'unsatisfied demand' -- whatever the barriers are -- and that we can find better strategies to help women space births and reduce unwanted pregnancies.

Thanks again,


Dominic Montagu
April 07, 2011

Hi Berk, Thanks for raising a challenging topic.

Back in 1994 Lant Prichett wrote a brilliant article in Population and Development Review arguing that because there is not a causal link between contraceptive prevalence and and fertility, after adjusting for 'desired fertility', expanding contraceptive supply is not going to effect fertility in LMICs. Essentially, he argued that women are able to attain close to their desired fertility with or without modern contraceptives.

The fallacy in Prichett's argument was the unaddressed endogeneity in his analysis: desired fertility is driven by actual fertility. If you live in a society where everyone has seven children, you might want to have only five, but having a goal of one or two children would be unthinkable. If average fertility in your society is at two children per woman, then it's quite likely that you could imagine happily having only one or two children.

The problem with simplifying supply (contraceptives) and eventual outcomes (fewer births) is that there are many intermediary factors that are difficult to account for, and that ultimately must be examined at either a macro level (Prichett's global analysis, even though I disagree with his conclusion), or a micro level (husband control of decision making in one hospital in Zambia), but not both. Maybe we need the phrase "All contraception decisions are local?" I don't know the literature well, but I've heard lots of anecdotes about the popularity of injectables in Bangladesh and other countries where wives seek family planning without their husband's approval. It seems the same may be happening in Zambia. This is a sad state of affairs, but my guess is it's only one manifestation of 'unmet need'.

I do like the idea of a shift from unmet need to unsatisfied demand. A need with no demand might make sense for political activism, but not for programs or policies.

Thanks for the interesting discussion.


Pritchett: Desired Fertility and the Impact of Population Policies

Berk Özler
April 08, 2011

Thanks Dominic,

Maybe we can get Lant to guest blog on this (or any other topic of his choosing) in the near future. Stay tuned...

Lant Pritchett
April 08, 2011

There are may distinct issues that are conceptually separable, so let me not talk about fertility impacts f interventions but just address "unmet need." My claim is that the usual numbers bandied about for estimates of "unmet need" do not correspond to any definition of "unmet need" that any economist (or just common sense) could agree to. They are a advocacy construct that has been successfully used in the overall political agenda for promoting family planning. But ultimately I have also argued the notion of "unmet meed" has been counter-productive even within the movement.

First, the usual use of the word "need" implies stronger intensity that "want" or "wish" and we usually, for competent adults, don't say people "need" things they don't "want." Yet the usual "unmet need" numbers include every woman of a certain age who says they do not want a child now who is not using contraception was having an "unmet need" for contraception. This is in spite of the fact that the same DHS surveys have responses from women who do not want a child and are not using contraception about why they are not using, which includes answers like that they dislike the side effects, that they are no longer fecund, they are sexually inactive, that they have religious objections, that their husband is out of the country for a year. That is, many women give reasons suggesting they do not want contraception and only a few cite access or price as reasons for their "unmet need" status attributed to them.

(moreover, the numbers usually include women who are currently pregnant (if their current pregnancy was unwanted or mistimed plus women who haven't had a period since their most recent birth. I have shown that "unmet need" is 13 percent in France (just as a benchmark) and the numbers commonly reported suggest it is 35 percent in Ghana. But I exclude from the "unmet need" women are are pregnant, amenorhheic, or sexually inactive (all pretty good reasons for not using contraception) the number in Ghana falls to 15 percent (with similar reductions for other countries with high "unmet need").

So the notion of "unmet need" might sound like "how many women would use if contraception were available at low cost (in time, trouble, money)"--that is, some point on a demand curve and hence there might seem to be a puzzle between large "unmet need" and low price elasticity of demand (or income elasticity, after all, what economists mean by "need" is something with super high marginal utility at low levels of consumption (perhaps rapidly declining) like water so that the marginal propensity to spend on "needs" should be very high at very low levels of consumption).

But there isn't any contradiction because the numbers for "unmet need" consisted (when I last looked at them) predominantly of women who don't express any current desire to use contraception.

I am convinced the "unmet need" numbers were created to counter the objection many countries had to expanding family planning programs that women didn't really want it. But, while appearing to counter that objection, it doesn't because it doesn't measure women's expressed "want" it measures some completely arbitrary attribution of "need."

Second, I have argued the concept of "unmet need" has actually been counter-productive for the movement, in two senses.

One, it is symptomatic of the of the deep disrespect for women and their agency that the demographically driven family planning programs have often displayed. As Matthew Connelly has argued the coercion in the family planning programs in India and China was not a "mistake" it was a logical consequence of people (men mostly) who believed that women needed to use contraception to reduce population growth --whether they wanted to or not. The fact that the movement has consistently attributed "need" for contraception to women who have articulated reasons why they don't want it reveals the paternalistic approach inherent in demographically driven family planning programs--we population bomb advocates can override what you want with what you need.

Second, suppose you actually believed your own advocacy and thought that 25 to 35 percent of women having "unmet need" actually meant they had "need" in the usual sense of the word. Then meeting "need" is just logistics--all I have to do is slap the stuff out there and it will fly out the door. I don't have to worry about the client, don't have to be nice, don't have to worry about side effects, don't have to worry about arrays of methods. There is a huge difference between famine relief (delivering food to people who need it) and selling macaroni and cheese (where people might not want it and need to be sold on it).

The lesson that actual implementation of family planning programs has consistently found is that getting uptake is hard, not just slapping it out there, which is not at all surprising, it is true of every consumer good. It is only the contradiction between the advocacy needs of the movement (convincing governments that people wanted it) and the implementation needs (actually needing to create demand--and meet much women's actually desires with more specific demands of methods, convenience, side effects, respectful non-coercive treatment, etc) that created the confusion.

So I think independently of one's sophisticated views on topics around fertility and contraception (e.g. how much of women's expressed reproductive intentions represent "just" social norms), everyone should be able to agree that the usual numbers on "unmet need" are an advocacy tool, not particularly relevant to conceptually or empirically informed discussions.

(attached is a link to the text of a presentation I gave many years ago, which I just never bothered turning into a real "paper)…

John Bongaarts
April 18, 2011

Response to Pritchett “No need for unmet need”

The most persuasive evidence that unmet need is real and substantial comes from family planning experiments such as the one undertaken in the Matlab district of Bangladesh. Matlab’s population of 173 thousand in 1977 was divided into roughly equal experimental and control areas. Starting in 1977 the quality of family planning services (including home visits, access to an array of contraceptive methods, and follow-up care) were greatly improved in the experimental half of the district while no additional services (other than much less intensive country wide services) were provided in the control half of the district.
The impact of the new services was large and immediate (Cleland et al 1994): Contraceptive use jumped from 5 to 33 % among women in the experimental area in the first 18 months and it remained about 25% higher than in the control area in subsequent years. As a result, fertility declined more rapidly in the experimental than in the control area and a difference between the areas of more than one birth per woman was maintained over time.
The Matlab experiment demonstrated that family planning programs can succeed even in very traditional societies. When this experiment began in the 1970s, Bangladesh was one of the world’s poorest and least developed countries, and there was considerable skepticism among economists that couples would be interested in limiting their fertility. This skepticism was proven groundless by the experiment and by the subsequent fertility decline that occurred throughout Bangladesh after the lessons from Matlab were incorporated in the national family planning program.
In addition, the fertility decline in the intervention area yielded substantial development benefits. A study by Paul Schultz concluded “Households in the program villages realized health and productivity gains that were concentrated among women, survival and schooling increased among children, and after 19years household physical assets were 25 percent greater per adult than in the control villages.

Shareen Joshi
June 16, 2011

This is a rather fascinating exchange. At the very outset, I should be clear that I am one of the researchers who has worked on Matlab (with Paul Schultz) and have also written some review articles on the rationale for family-planning programs lately.

I agree that we should stop emphasizing "unmet need" as a rationale for family-planning (FP) programs. I agree that it does not correspond to what any economist would call demand. I also agree that framing the rationale for contraception on the basis of "unmet need" has been counter-productive. The rationale for such programs should be much simpler: Easy and low-cost access to modern contraception give couples the opportunity to regulate their fertility and control when they have children. They make it possible for parents to have children when they most want them and choose to have them. Isn't this type of expansion of peoples freedoms and choices the whole point of development?

A secondary rationale for family-planning programs should be that they have significant spillover effects on a variety of variables that policy-makers care about: female employment, female health, children's education, etc. We have only learned this recently, from the papers by Schultz, Miller, etc. but the evidence appears to be quite convincing. Since it appears that FP programs (voluntary of course) can reinforce these other programs, that should be a great reason for policy-makers to take them seriously. FP should be viewed as just one more investment in female human capital.... And needless to say, there is tremendous evidence about the importance of female human capital.

Finally, going back to this issue of "unmet need".... I agree with Lant Pritchett that we should stop trying to quantify the need for contraception this way. But by that same logic, we should also stop trying to quantify "desired fertility" (a variable featured prominently in writings by economists, including Pritchett's hugely famous 1994 paper)! Asking a woman her "ideal" number of children is just as complicated as asking her about her need for contraception. Neither question is answered objectively. Responses to both are colored by social norms, preferences of a spouse, past fertility history, and the broader socio-economic environment. So just like you can't use "unmet need" to justify family-planning programs, we shouldn't use "desired fertility" to justify their omission from the policy agenda. Again, the rationale for FP programs should simply be that that they (a) expand people's choices and give them greater control of their fertility; and (b) they are one more investment in female human capital. Lets not worry about unmet need, desired fertility, or ideal fertility!


David Bishai
August 03, 2011

Unmet has the small virtue of helping policy makers retain a quantitative indicator. Spatially locating FP investments in proportion to unmet need is slightly better than randomly throwing money at reproductive health goals. Surely we can do much better than this. If unmet need were a good indicator then changes in unmet need would be followed by changes in unwanted fertility. Unfortunately, there is a very weak correlation between the change in unmet need and the change in TFR. In countries that have had more than one DHS survey, I found that 1 point of unmet need reduction is worth 0.02 points of TFR reduction, not controlling for any other variables that would confound the relationship. What is frustrating is that the DHS surveys offer so many better questions about women’s ability to access FP services that would be better for policy guidance. Why are advocates so stuck on unmet need when there are better indicators that could guide policy?

The existence of unmet need counts “in the millions” never had to be proven for there to be subsidies in family planning. For murky reasons, interpersonal interest in reproduction is as old as primates. Family planning is now subsidized by private insurers and Medicaid in the US and subsidies exist internationally. In 2010 USAID’s budget for family planning was $649 million which it used to support family planning programs in more than 50 countries. These are funds signifying a revealed preference by policymakers and citizens for an equilibrium number of family planning devices that is larger than it would have been otherwise.

The size of the subsidy ought to be guided by the potential to reduce externalities—but these are not easily enumerated. The externality story told by an economist is a complex tale wherein household level FP-purchasers are not fully integrating the interests of their families in the decision-making. Harmonious family accord on family planning strategies and goals cannot be taken for granted as demonstrated by Ashraf Field and Lee. In different families, a father, a mother, or a child may be left out of having their interests represented in a family planning decision. Whoever ends up deciding may have information deficits about the impact of birth intervals on child survival, maternal health, paternal health, and human capital. The donor believes that the benefits of FP are underappreciated and/or they may also have an ideological interest in various birth rates. The economic dilemma for donors is how much to spend and where to spend the subsidies to minimize externalities. There is little reason to believe that the current spending amounts and locations are chosen with a rigorous economic rationale. It would be great to fix or replace unmet need with a measure that really counted the number of families where subsidies would improve human functioning.

There have been few estimates of the economic value of family planning subsidies as public goods. The paper by Joshi and Schultz is a fabulous contribution to progress here, but John Bongaarts incorrectly attributes all of the treatment effects on household wealth to family planning--the Matlab treatment included a large array of child and maternal health services in addition to family planning. We all realize that Matlab’s contributions to knowledge supersede a bean-counting assessment of whether Joshi and Schultz’s estimates of wealth gains justify the immense donor investment since 1974. However finite donor resources force the question of return on investment on FP everywhere else. It should not be dodged by sanctimony or blind advocacy. There is an important stewardship role for all spending on behalf of the poor. Until economists deliver a full-fledged guide to investments in family planning space the prioritization question is going to keep getting answered with “invest where there is greatest unmet need and cheap CYPs”. The drawbacks of this approach are well stated above. The research community can do better to guide policy, but we haven’t done it yet.

Dr. Dom
August 01, 2012

We have in the Catholic Church said something similar all along, but perhaps with a bit more anthropological depth than is being offered here. Contraceptives don't work for an abundance of reasons, but it is very consoling that what we have been sang for decades -- that the population control ideologues project on the third world an anti-natalist mindset that quickly becomes an anti-child mindset. Whereas this is not what the folks in these countries want or need.

Yes to development, maternal care, education initiatives, clean water and the like, but no to the contrivances of failed western ideologies that have done more then fail here.

December 16, 2019

I am very interested in this article.