I have worked in India and several African countries mainly as a health administrator, clinician, and policy advisor and secondarily as a researcher. I am all for evidence-based thinking and have seen how this is sorely lacking in the formulation of health policy. Having suffered the frustrations of managing public systems where significant effort is spent motivating people to do what they are already paid to, I am also strongly attracted to market and performance-based mechanisms that use incentives to compel performance. With that said, I find this discussion narrow in its perspective and focused on the ‘trees’ of specific theoretical points with a need to take a step back and look at the overall ‘forest.’ When I read some of what is proposed about fees, even when seemingly supported by research studies, the conversation seems to be describing a situation at odds with the reality I see on the ground. That makes me think that either I am grossly misperceiving reality and/or the research, in composite, is not forming a coherent picture or is missing angles of analysis not readily conducive to theory or focused empirical research. With this in mind, a few thoughts: The reason why health systems underperform is multifactorial with different issues playing a greater or smaller role in different settings. Lack of political will and leadership. Weak institutions and operations. Poor planning, management, and implementation capacity. Lack of adequate resources. Absence of internal and public accountability. Inability to attract talent. Undignified salaries. Etc. While incentives created through health financing policy can go a long way in spurring improvements on many of these issues, financial incentives alone cannot resolve these problems and also come with considerable tradeoffs, especially in poor communities where market and public accountability mechanisms are weak. Fees, in particular, create financial barriers for the poor making access to care prohibitive for many. Poor people cannot protest unreliable service delivery by withdrawing money that never had in the first place. Potential mitigating strategies like targeted subsidies or even comprehensive public financing of non-public service provision entail additional challenges and implications that merit serious consideration (I examine this below). Furthermore, while financial incentives could help induce some performance improvements, they cannot wholly absolve the need for building institutions that work, strengthening supply chains, establishing appropriate monitoring and supervision, bolstering management capacity, etc. Achieving improved outcomes and making health systems work is a complex and requires many factors that are necessary, but may not be independently sufficient (and, hence, why many initiatives do not achieve changes in outcomes when done in isolation). No health financing policy may work as a ‘magic bullet’ that single-handedly corrects all health system deficiencies or adequately aligns incentives without generating unacceptable equity costs. Targeting subsidies to the poor is a great theoretical idea but ‘exemption’ programs like this have been tried and have rarely worked. This is because of the challenge of identifying who can or cannot pay in communities with ubiquitous poverty and income that is precarious, seasonal, and informal and, thus, insecure, highly variable even within the same year, and difficult to measure. This makes subsidy-targeting extremely challenging and time, effort, and cost-intensive (particularly if done at scale). I have worked on subsidy-targeting programs in Rwanda and Ghana and can attest to this. In Rwanda, such wealth ranking has had more success but that is in part due to unique preexisting community organizational structures (with households already administered in localized units of 10 to 20 households), high population density even in rural areas, and uniquely effective public administration that enjoys strong authority and popular trust. The idea of attaching subsidies to particular medicines or illnesses is also problematic. Patients will usually not know what illness they have or medicines they will need until after a healthcare worker sees them. Since many illnesses can present with similar symptoms, they will not know whether the care they are seeking will be ‘covered’ until they have already been seen, at which point, if the diagnosis is an ‘uncovered’ illness or requires ‘uncovered’ medicines, they are stuck with a bill. This would discourage care-seeking behavior generally. For this reason, any medication or diagnosis-specific subsidy would have to cover a range of conditions (i.e., the entire spectrum of basic primary care infections) or a comprehensive package of essential medications. This is what many countries are already trying to do. For boosting the quality of diagnosis and treatment, training and supervision (particularly of prescribing habits) is indispensable and cannot be substituted for by financial incentives. An approach where every person’s healthcare needs were publicly financed and service delivery provided by the private and NGO sector could theoretically work and use market mechanisms to produce competitive, quality services. However, this scenario would require strong oversight and regulation of non-public healthcare delivery to prevent inappropriate and unsafe practices (i.e., women sent for unnecessary c-sections) and guard against the emergence of a two-tiered system where wealthier people who can ‘top-up’ their public insurance get better care and the poor who are solely publicly funded receive lower quality services. In addition, governments contracting out provision in this way must ensure that a sufficient range of care is reliably provided for remotely located, poor, or otherwise marginalized populations where financial incentives may not be sufficient to attract private and NGO providers (i.e., private doctors in India who refuse to treat low-caste patients no matter how much money they are offered, a situation that I have personally encountered and is not exceptional). This type and degree of oversight and regulation requires remedying the same institutional failures, corruption, and weak public administration that undermine direct public service delivery and compel the call for non-public provision in the first place. The research showing that people will pay for private care in the face of non-performing public systems makes an obvious point. However, in my experience, a greater proportion of people go without any care at all in the face of unaffordable fees for private or public services and non-performing public services. In my experience, the brunt of undue mortality, particularly child and maternal deaths, happens at home among patients who never reach any kind of care at all. Misdiagnosis, while certainly an issue, is not the problem in these cases; no diagnosis is. Even if no barriers to access are faced and service delivery is optimized, many communities may still not adequately seek care due to a limited history of interfacing with formal health systems, cultural barriers, and low health awareness. Increasing demand – by removing fees, considering CCTs, tackling health-seeking behavior – would create the strongest incentive for service delivery to perform with underperformance met with public outcry. Fees, at this stage, would be less a weapon of accountability and, in many places, discourage already weak demand.