Is it OK to flash money to snip the goods?


This page in:

I was circumcised in the hospital as a very young infant. Most children do get circumcised in Turkey, although I suspect that many are not as lucky as I was, including my younger brother, who went through the ordeal when he was around six years-old. I remember him in some pain and discomfort for what seemed like a long period of time to me at the time, even though it was probably no longer than a few weeks if not days…

Until recently, the decision to circumcise a child (or to get circumcised as an adult) was a cultural decision, perhaps with a dash of health considerations thrown into the mix. However, with three recent RCTs confirming that the likelihood of contracting HIV is significantly lower for males who undergo medically-performed circumcision (MC), it has now become a public health issue.

My understanding of the issue is that high take-up of MC is necessary for communities/societies to reap the full benefits. This is no different than the take-up of insecticide treated bed nets or deworming pills, each of which has large externalities and works much better if most people in the community have also adopted its use. However, it seems that after an initial surge in take-up, the percentage of the adult male population that underwent MC has reached a plateau below optimal levels in many countries. In some cases, this is true even with campaigns that make the information on the benefits and the availability of safe MC widely available.

Of course, information and access constitute only two of the many possible barriers to higher take-up rates. First, for many people, it may be individually rational to remain uncircumcised given the perceived benefits conditional on their subjective probabilities of HIV infection with or without MC, compared to its costs (psychic and pecuniary). Second, there may be cultural barriers to being circumcised: individuals may be less likely to opt in where they’re likely to be in a small minority of circumcised men (Hence, general equilibrium effects from a ‘big push’ could be very different than the case of a marginal increase in take-up). There may be perceived costs with respect to pleasure derived from sexual activity, sexual prowess, or reception from potential sexual partners. Finally, many people are rightly scared of medical procedures, especially in developing countries. Costs are immediate and the benefits are in the future… So, for every uncircumcised individual, this is a cost-benefit analysis that is not at all straightforward – even with full information.

Now suppose that calculations of social rates of return suggest that it would benefit the government to subsidize MC to such an extent that the tax rate on MC would be negative. Could any government actually implement this? (Let’s assume, to simplify matters, that there is no donor push behind this and the funding will fully come from the national budget, as CCT programs work in many countries) The problem is that even the suggestion of paying people a small amount to increase take-up seems to be immediately stirring memories of the proposal to give a transistor radio to every person who agreed to undergo sterilization in India back in the sixties. In fact, we don’t have to go back that far: a small program that provides IUDs to HIV-positive pregnant women in Kenya is causing justified outrage over issues of ethics and legality (see also here and here). While this case indeed seems to be an egregious one (the same NGO apparently also offers cash to drug addicts in the US money for sterilization), the policy question remains and brings to the fore one of the most straightforward applications of public finance: the use of taxes and subsidies to discourage/encourage the consumption of certain goods.

Let’s think of this another way: in the vast continuum ranging from offering poor families money to keep their children in school, vaccinate their children, or get prenatal checkups, all the way to offering HIV-positive women money to adopt long-term contraceptives, where does MC fall? Where do we draw the line on which goods can have negative tax rates? What is our guiding principle in drawing this line?

I suspect that the key here is ‘informed consent’ or ‘informed choice.’ Each of these terms deserves some discussion. For any incentive to not be considered “coercive,” the consumer has to first be able to make his choice under full information. Given that information asymmetries are larger for poorer, less educated individuals, there is a large burden that the government has to clear in terms of making sure that the costs and the benefits of MC are understood by everyone equally. Second, inequities in society make the notion of choice problematic as well: if a short-term need is urgent enough, weighed against any (and possibly heavily discounted) future costs of MC, individuals will take the money and undergo the procedure. However, this is inherent in any subsidy: the consumption patterns of the poor will be more sensitive to changes in prices than the non-poor – whether it is cigarette taxes or incentives to keep children in school, especially at low levels of subsidy.

In any society, there are conditions that the government should not impose on its citizens to in return for assistance to those who are down on their luck. However, we’re not talking about MC as a condition to benefit from safety net programs here: we are instead talking about a specific behavior change intervention in health with no explicit redistributive component. It must surely help, then, if everyone had access to a basic safety net, meaning that we could be a bit more certain that the people who opt into MC are not doing it in order to avoid skipping their two meals a day. Furthermore, I am not certain that remaining uncircumcised is an unalienable right: should governments care about the right of men to enjoy sex with more skin (or foreskin) to such an extent that they should ignore a three-fold increase in the risk of infection for every act of unprotected sex with an HIV-infected person?

Right now, governments and donors alike are trying to devise ways to increase the take-up of MC in countries where the HIV incidence remains high and the scale-up efforts have stagnated. It is possible that there are untried campaigns that are potentially effective and do not require the use of subsidies (like better information campaigns, working with couples or women, etc.). In fact, for designing such interventions, we could simply turn to the three questions Jed Friedman posed at the end of his post last week: we’d have to figure out ways to (a) identify the constraints to higher take-up, (b) effectively impart the benefits of MC, and (c) finance encouragements to adoption. Allow me to go one step further in thinking of possible intervention designs.

The current sentiment I detect is that many policymakers are OK with making MC available freely. Perhaps, they might agree to the offer of a voucher to cover transport costs. I suggest adding foregone earnings to this package as well. That would be a nice start if the cost of the procedure (including the opportunity cost) was the main barrier to higher take-up. Suppose, however, that this package was not effective in increasing take-up rates sufficiently. Suppose further that it seems likely that a small additional cash incentive would do the job – even if this might be a second-best solution (we can think of a pilot experiment where randomized amounts offered showed a positive and significant gradient in MC take-up). Why would this not be acceptable? If people were worried about informed consent, we could have a two-step procedure where people are first given incentives to attend a one-hour information session. Then, they would be sent home with a second voucher for MC that they cannot redeem before one week (think of it as a cooling period to avoid impulse purchases). During this week, they’d have a chance to discuss the issue with spouses, family, and friends, have access to community health workers for any questions, as well as the ability to ask questions to health professionals via text message and receive instant answers. If there is proper qual-quant work beforehand to identify the constraints, there are ways to design pilot programs that would not only provide information on cost-effective ways of increasing the take-up of MC but are also acceptable and feasible for scale-up.

Before we let politicians or other policy-makers take away incentives off the table in a knee-jerk fashion, can’t we at least request that they clearly make their case as to why MC cannot be subsidized like schooling or vaccines are in many countries? I’d love to hear your (and their) thoughts…

Postscript: I was just made aware of this paper by Chomitz and Birdsall (1990) that treads similar ground wrt family planning. Definitely worth the read if you're interested in the topic, not the least because it makes several points that I stumbled onto much more elegantly and in a systematic manner, including the issue of the ethics of cash incentives (see pages 330-331).



Berk Ozler

Lead Economist, Development Research Group, World Bank

Join the Conversation

May 19, 2011

This is a highly controversial issue for many cultures as reflected in the article itself. In addition the actual benefits of MC remain debatable. A recent article of Robert S. Van Howe, Michelle R. Storms in The Journal of Public Health in Africa (April 2011) has the following abstract "The World Health Organization and UNAIDS have supported circumcision as a preventive for HIV infections in regions with high rates of heterosexually transmitted HIV; however, the circumcision solution has several fundamental flaws that undermine its potential for success. This article explores, in detail, the data on which this recommendation is based, the difficulty in translating results from high risk adults in a research setting to the general public, the impact of risk compensation, and how circumcision compares to existing alternatives. Based on our analysis it is concluded that the circumcision solution is a wasteful distraction that takes resources away from more effective, less expensive, less invasive alternatives. By diverting attention away from more effective interventions, circumcision programs will likely increase the number of HIV infections."

There are other studies just as critical of the practice. Given the limited resources of most governments facing the crisis, using more conventional and proven measures might well be economically more effective.

Berk Özler
May 19, 2011

Very fair point - thanks. I have to admit that when the initial news of the RCTs on MC came out, my priors were exactly the same. I was particularly worried about "disinhibition effects" (as well as potential negative externalities associated with procedures performed in less than ideal conditions in a scale-up scenario). However, I have to say that, longer-term studies (the credibility of which I have not yet investigated thoroughly) on MC have not yet found evidence of disinhibition to reverse the original findings. I will download and read the paper you refer to.

However, it remains true that, for the purposes of this post (i.e. to discuss the issue of incentives for behavior change), I abstracted from the analysis of whether MC is more cost-effective than other prevention methods. I am sure that the debate has passionate advocates on both sides, and with the news effective technologies -- whether that of a partically effective vaccine or a microbicidal gel -- the set of alternatives is getting larger. And, that is always good news.

David McKenzie
May 19, 2011

Ethical issues aside, as economists we are reasonably comfortable with subsidizing people to take certain actions when there are externalities. But then the subsidy should not be more than the externality. So it might be useful to think what the externality is here to get a sense of what level of subsidy we might be talking about.

Suppose then MC lowers the probability of getting HIV from 0.12 to 0.04,

We then need to know what the expected number of people that will get infected by a person who gets HIV will be who would not have otherwise got infected. Suppose this is 0.05 (i.e. each infected person has a 1 in 20 chance of infecting someone else who wouldn't have otherwise got infected).

Then value human life at GDP per capita of US$1000 for 30 years, discounted at 10% per year = $9500.

Then rough guess at externality is 0.08*0.05*9500=$38.
All these numbers are made up, and experts will know better than me whether they are too high or too low, but I think it would be useful to also have a reference of what sort of subsidy might be justified by economics alone.

Berk Özler
May 20, 2011

A few hours after I posted the above, Freakonomics blogged about a ballot proposal in San Francisco to ban circumcision of minors:…