Category: Public policy

MY ANNOTATED NOTES (work in progress)

Glied, S. “Public Health and Economics: Externalities, Rivalries, Excludability, and Politics,” In The Contested Boundaries of American Public Health, eds. Colgrove, J., Markowitz, G., Rosner, D. New Brunswick: Rutgers University Press, 15-31, 2008.  Source:

===

The scope of activities under the jurisdiction of public health departments has broadened to include many that seem far removed from control of the list of communicable disease that the U.S. Public Health Service published in 1921 (U.S. Public Health Service 1921). The term “public health” itself has been conflated with the expression “population health” or even the idea of preven­tion more generally. Addressing population health or prevention, in turn, is as readily a responsibility of large private insurers or of individual physicians as of any public entity. Public health departments today face increased responsi­bility for a growing share of the endless threats to the population’s health and, at the same time, a loss of a distinctive mission in a world populated by HMOs, insurers, and a host of other private organizations.

This predicament means that the budgets of city, state, and federal public health agencies are perpetually under attack, even while the agencies’ duties expand. In the wake of 9/11, health departments around the country complained that new antiterrorism responsibilities—even where supported by new funding—were siphoning already inadequate funds away from real public health problems. Is there a distinction between public health and population health or prevention? What part of those pieces are the core functions of public health?

Perhaps surprisingly, one place to look for an answer to the question of the boundaries of public health is in neoclassical economics. Economics, as a dis­cipline, has displayed no more than passing interest in the subject of public health. Yet, it is possible to construct a fairly clear definition of the core activities of public health through some simple extensions of the basic economic model of health.

The Production of Health

The basic economic model of health, which I will use as the basis for a theory of public health, is the theory of health production that was developed by Michael Grossman in 1972.¹ In Grossman’s model, health is a flow of benefits generated by a nontradable capital good, health capital. Individuals determine how much health capital to produce (and how to produce it) by making subjec­tively rational tradeoffs between this health capital and goods not related to health. These tradeoffs depend on the prices of the inputs to health capital, on individual preferences, and on each individual’s initial endowments of income. time, genes, and so on.

In this model, the flow of health generated by health capital is valuable both in terms of investment and in terms of consumption. As an investment, having better health yields fewer sick days and a longer work life. These benefits, in turn, generate higher income, which can be used to purchase other nonhealth goods and services. As a consumption good, better health is a benefit for its own sake. The consumption value of health means that health would have substan­tial economic value even if had no effect on productivity or economic output.

The Grossman model is unreservedly individualistic. Health, because it is not tradable, is less “public” than almost any other good or service. Within this formal model, the health of the population is simply the sum of the health pro­duced by all the individuals within that population.

The value of the population’s health as an investment good, in terms of the output that health produces, has been quantified. Burton Weisbrod’s 1961 book The Economics of Public Health, for example, emphasizes the substantial value of health in terms of economic productivity.² The value of the population’s health as a consumption good has also been measured. Recent estimates by Nordhaus (2005), for example, suggest that the consumption value of improve­ments in health since 1900 has grown about as rapidly as the growth in real income. The cost of producing health can also, in principle, be measured. It includes the cost of all actions taken to improve health—including the costs of medical care consumed, time spent in all activities associated with the produc­tion of health, whether intended to prevent disease or to treat it, and the lost utility associated with treatment and preventive actions (including, for example, the enjoyment lost from not smoking).

Population health, as it arises in this model, is in no sense “public health.” Grossman’s original model offers no obvious role for the public sector. Of course, as Arrow (1963) points out, there are many market failures around medical care and medical insurance. Efficiency may best be served by having government intervene in these markets to make them function better. But medical care is just one among many inputs into the production of health in the Grossman model, and health can certainly be produced without any medical care at all.³ Intervention in the medical care market may improve the efficiency of the economy, but it doesn’t naturally comprise “public health.”

Adding a Public Element to the Production of Health

The Grossman model focuses on the production of individual health. Within the framework of this model, we can imagine a host of situations where there are interactions among the individuals within the population. Consider some simple ones. I get immunized against measles, hence the prevalence of measles—and your risk of measles infection—declines. [Sanjeev: IF THE VACCINE ACTUALLY WORKED, everyone would take the vaccine in their own private interest. It is far better to have 100% protection than to have <100% protection from others taking the vaccine. This is a key issue that economics ignores – cooking up casles in the air.] I smoke cigarettes near you. and your risk of lung cancer increases. I dump sewage into the local river that feeds the municipal water supply, increasing the incidence of water-borne disease in the city population. I examine New York City mortality statistics and publish a paper showing that many children ages three to eight die from win­dow falls, and you install window guards to protect your children. Taking this logic further, in an extension along lines suggested by Lester Thurow (1971), I make a contribution to build a medical facility for homeless people, and you are no longer saddened by the fact that people are going without care. In eco­nomic terms, we call these situations problems of “public goods” (more specif­ically, “nonrival goods” and “nonexcludable goods”) and “externalities.”

Economists often use the terms “public goods” and “externalities” inter­changeably, but they are not quite the same. Externalities are all situations where my actions have a positive (positive externality) or negative (negative externality) effect on your well-being or your ability to produce other goods and services, and where you have not been compensated for these effects of my actions. The reduction in risk that you, a stranger, obtain from my choosing to be vaccinated against measles is an externality. Economic theory suggests that the market is likely to produce too much of goods that cause negative external­ities and too few of goods that cause positive externalities.

Public goods (or bads) are goods (or bads) that are nonexcludable and non-rivalrous in consumption. Goods are nonexcludable if it is not feasible to exclude some people from consuming a good that others are consuming. If you stop someone from dumping sewage into our collective water supply, it will be impossible to exclude me from the benefits of your action. If a good is nonexcludable, there is an inherent “free rider” problem. [Sanjeev: It is close to impossible to find any instance of the free rider problem in pubilc health] hSince even nonpayers can consume the good, no one wants to pay for it. Everyone has an incentive to have someone else incur the costs of the good while benefiting from the outcomes. In the case of nonexcludable bads, any action that I take to eliminate the bad will also redound to everyone else’s benefit. For this reason, goods that are nonexcludable will tend to be underproduced. Likewise, bads that are nonexcludable will tend to be overproduced.

Goods are nonrival if one person’s consumption of a good does not dimin­ish the amount available to another. Providing the good to one additional con­sumer would provide that consumer with some gain but would cost nothing at all. If goods are nonrivalrous, it may be possible to exclude someone from using them, but it would be highly inefficient to do so. Once a researcher has con­ducted a study showing that window guards reduce mortality, others can obtain that information at very little additional cost. The economy will not pro­duce the most efficient possible outcomes if externalities and public goods exist. These situations generate a need for the use of the coercive power of the state in some way. In many, even most, potential cases of externalities, the externality problem can be, and usually is, resolved efficiently if the govern­ment develops an enforceable system of property rights.⁴ [Sanjeev: that’s not an economics-based solution: the failure to define property rights is not a market inefficiency] Thus, there is gener­ally no externality problem associated with visitors smoking in your private home. You have an enforceable property right to constrain the behavior of visitors within your home. Should you wish to prohibit smoking within your home, the police power of the state stands ready to enforce your prohibition.

In cases of public goods—externalities where the good (or bad) is nonexcludable and nonrivalrous—it is generally not possible to achieve economically efficient outcomes simply by allocating and enforcing individual property rights. For example, it would be difficult to assign property rights in the air we breathe. Moreover, while we may all have a legal right to breathe clean air, no single one of us has much incentive to devote much time and effort to the enforcement of this right. If one of us did take action to keep the air clean, everyone else would benefit without incurring any costs at all. Similarly, we would all benefit if infor­mation on the causes of death were collected and made available, but no indi­vidual would rationally make the investment to collect these data themselves.

Lester Thurow’s argument about charity follows along these same lines. We all benefit (psychically) if poor people receive adequate medical care. But, on the whole, we’d be just as happy if someone else spent their money doing it as if we did it ourselves (barring anticipation of rewards in some world to come). Thus, Thurow argues, redistribution is also a public good. Society inevitably does too little of it.

In the context of Grossman’s model of the production of health, these exter­nalities and public goods related to health offer a role for the public sector. In practice, governments divide up the responsibilities among departments and agencies. Governments develop systems of allocating individual property rights, including health-related property rights, and enforce these through their legal system. Purely redistributive activities fall most naturally into the purview of the U.S. Treasury Department.⁵ The remaining “public goods” situ­ations related to health comprise the traditional focus of health departments.

The actions of health departments in response to the existence of these public goods feed back into the framework of individual health production. Health departments may provide goods and services directly, often at minimal cost to the user. Health departments may use tax dollars to subsidize services provided by other providers, again effectively lowering prices. They may exert a direct coercive effect on individual behavior, through regulation of activities generally (smoking regulations, for example) or restricting the behavior of indi­viduals (isolation of disease carriers). These coercive actions effectively raise the price (whether financial or nonfinancial) of the regulated activity substan­tially. More recently, health departments have joined with other areas of gov­ernment to tax public bads. The taxes raise the prices of these bads.

Once the new set of prices induced by health department action has been introduced, the prices enter the individual production function for health, and individuals once again make individually rational decisions. If the public health activities are indeed optimal, individual decisions will now lead to what are both individually and socially efficient choices (though these may include choices to engage in unhealthy activities).

Public health activities interact with private decisions and may even increase the return to private investments in health. This potential increase in the payoffs to individual investments in health that follows public investments means that addressing the insufficiency of public goods can have multiplier effects on population health. Dow, Philipson, and Sala-i-Martin (1999) and Cutler and Miller (2005a) demonstrate such spillovers, improvements in health outcomes following public health interventions that exceed those directly gen­erated by the interventions themselves.

The new set of prices induced by the health department may also affect dif­ferent people differently (depending on how responsive they are to prices). Those who are very responsive to changes in financial costs—generally those of lower incomes—will tend to be more affected by these interventions. Thus, the actions of public health departments often have a redistributive component, whether intentionally or unintentionally.

Pure(r) Public Goods

The economic model suggests a definition of public health as that subset of gov­ernmental activities that addresses health-related public goods. This definition potentially encompasses a great deal, since many activities have some public good components, although the public good component is often trivial. To see this, consider a ridiculous example: If I have my teeth whitened, all other riders on the subway may receive slightly more utility from my smile. It is hard to believe that this external effect leads to an important efficiency loss from too lit­tle tooth whitening. Traditional health department activities tend to be those that address those public goods problems where the free market leads to great­est underproduction.

Nonrivalrous Goods

Much of the traditional work of public health consists of the collection, tabula­tion, and analysis of routine epidemiological data. The information that results from these disease surveillance activities is not naturally embodied in a drug or treatment or product. In consequence, it is entirely nonrivalrous in consump­tion, and, in practice, nonexcludable as well. Once the information has been collected and made available to one user, it can be made available to all other users at no additional cost. While it might be possible, with considerable effort, to exclude nonpayers from gaining access to these data, such a prohibition would be inefficient, because the incremental cost of access is zero. Moreover, exclusion is likely to prove difficult. Anyone who did purchase the information could make it available to others at no personal cost, because my health has no direct effect on the health of others. If a good is nonrival. society is better off if we do not limit access to it.

These properties of epidemiological information suggest that, in many cases, no private market would likely arise to collect and tabulate it for the popu­lation as a whole. An insurer might wish to monitor the health status of its enrollees and might survey their health conditions (although the evidence sug­gests that few insurers do). Even this engaged insurer, however, would be unlikely to have an interest in the general health status of the overall popula­tion, including those not now or ever likely to be covered by its policies.

Without any incentive to promote a private market, information of this sort will tend to be underproduced without public intervention. The production of this information by a health department can improve the technology available to each individual to produce health. Knowledge of disease risks allows people to make better choices with the endowments they have available to them. As Glied (2001) shows in the case of childhood injury epidemiological data, the gains from public health information can be very large.

The gains from public health production of epidemiological surveillance information are greatest in the cases where that information cannot be tied to some known private good. If surveillance information can be embodied in an excludable, nonrival good or service—for example, where the results of epi­demiological surveys can be used to design and market products—the gains from public production of the information will be correspondingly small. The efficiency benefits of the public production of surveillance information are greatest in situations where the outcome of the investigation and the potential value of that outcome are not clearly linked to any private good.

Like the production of surveillance information, tracking infectious dis­ease carriers and subsidizing or coercing them to cease infecting others consti­tutes a service that is nonrivalrous and nonexcludable. People who spread disease confer negative externalities on others. Unless treatment is available, disease carriers may not wish to know that they are infected. Once informed of their status, some infected people will behave altruistically and avoid infecting others. But others will have little incentive to reduce their own disease-spread­ing activities.

Identifying and isolating a disease carrier to protect one person benefits all others who might have been exposed at no additional cost. The widespread bene­fits of tracking disease carriers and subsidizing their treatment mean that no single susceptible individual has much incentive to do it. Instead, faced with the possibility that others are disease carriers, uninfected individuals have an incentive to change their own behavior to avoid contact with potential suscepti-bles. This risk-avoidance behavior is personally costly and may, in certain cir­cumstances, increase the general prevalence of disease, by reducing the number of potential uninfected contacts (see Michael Kremer’s analysis of HIV, 1996).

Public Goods Monopoly Infrastructure

Another traditional role for health departments has been the development and maintenance of large-scale public health physical infrastructures, particularly municipal water supply and sewage treatment systems. The careless disposal of sewage is a clear example of a public bad. Dumping generally occurs in pub­lic property or rivers and lakes that serve large populations. Thus, poor private sewage disposal generates a collective negative externality.  [Sanjeev: our body produces waste which is not an inefficiency. The fact that we need to dispose waste safely IN CITIES is NOT because it waste is inefficient but because human waste is harmful when allowed to accumulate in streets of cities. When such waste was disposed in nature – in the jungle or agricultural fields – it caused NO harm, being decomposed in a few days]. 

Health departments address this negative externality through the develop­ment and subsidization of sewage treatment systems. In this situation, the role of public health goes beyond the construction of the facility. The operation of the facility must continue to be subsidized over time so that people do not have an incentive to revert to less costly, externality-causing alternatives. Often, tax-based subsidization of such facilities is coupled with systems of penalties designed to coerce people to use only these disposal options.

Water filtration and clean water systems perform a public health function that parallels that of sewage treatment, although their economic characteristics are rather different. Use of untreated water is primarily a private bad; there is no externality to correct in this case. Moreover, the benefits of water treatment are excludable. Households can be (and generally are) charged for their local water utilization.

Water treatment systems, however, have a low (though not zero) degree of rivalry. A network of water pipes can provide service to additional subscribers at very low cost. Municipal water systems are natural monopolies. A single water supply and treatment facility can produce water much more efficiently than can multiple competing systems. The economic role of health depart­ments in this case is to own or regulate a region’s monopoly water treatment system. The early twentieth-century debates over whether municipalities ought to develop their own water treatment facilities or contract for the provision of water reflects the economic ambiguity of the public role in the provision of clean water. In their analysis of the genesis of municipal water filtration, Cutler and Miller (2005b) conclude that the observed public role in water treatment reflects, primarily, the advantage of local governments in financing such large-scale natural monopoly public investments.

Benefits Conferred on Both Current and Future Generations

Many health risks, including those associated with sewage and unclean water, can be avoided by taking precautions (avoiding crowded places, washing hands often, staying out of restaurants, cooking all eggs to hard-boiled). Taking such precautions is costly. The expense of taking such precautions means that an entrepreneur who develops a less costly avoidance strategy—think of antibac­terial hand cleansers—can reap benefits. The existence of a market for avoid­ance strategies often obviates the need for public intervention.

In some cases, however, an intervention that reduces avoidance costs has “permanent” payoffs—costs must be incurred today but returns are experi­enced both by the current and by all future generations. The payoffs from these public health investments are not “excludable” to these future generations. In effect, future generations free ride on current investments. The existence of benefits to future generations, who do not participate in either private or pub­lic decision making today, implies that there will be too little private invest­ment in such activities today.

Many public health investments in the development of clean water and sanitary sewage disposal systems have this characteristic. For example, during the 1890s, the municipality of Chicago reversed the flow of the Chicago River so that the city’s water supply would no longer be contaminated by sewage (Blake 1956). This action continues to benefit the residents of Chicago today, a century later.

The advancement of public health knowledge offers another example of such payoffs to future generations. Research that identifies disease risk and pro­tective factors today will continue to be useful into the indefinite future. The development of the germ theory of infectious disease, for example, provided considerable benefit to the generation alive at its introduction but also con­tinues to provide benefit today (Deaton 2005).

Another case of future benefit concerns the complete eradication of dis­ease. As Tomas Philipson points out, the eradication of smallpox provides immensely more benefit than would control of the disease. Future generations are entirely spared the need to take precautions against developing smallpox (Philipson 1995, 2000). At the same time, the eradication of the disease means that there is no private good (not even a vaccination) that can be sold to these future generations.

Future benefits may also arise through the avoidance today of public bads. Control of indiscriminate use of antibiotics today, for example, primarily bene­fits future generations by stemming the rise of new antibiotic-resistant strains of disease. Future generations, however, cannot compensate today’s antibiotic users for these potential benefits.

Broadening the Definition of Public Health: Impure Public Goods

Traditional public health activities—surveillance, research, sanitation, clean water—tend to have a very substantial “public goods” component. The unfet­tered private market would be unlikely to provide these goods, so in the absence of public involvement, they would tend to be underproduced. [Sanjeev: keeping track of births, deaths, etc. is a government function: the others are not. These functions were started in the private sector and any shortcomings can be readily addressed through regulation] The welfare of all of society can be improved through adequate provision of these goods and services.

But many of the activities that comprise the broader swath of public health today have a much smaller pure public goods component. Some—such as the delivery of personal health care services related to noninfectious disease (pre­ventive or otherwise)—don’t have much of a public goods component at all. The only “public good” present in this component of the Medicaid program, for example, is redistribution.⁶

In cases of “impure” public goods—goods where there is a substantial elem­ent of “privateness” or excludability. present—the consequences of public intervention are not clear-cut. Public health intervention may have unanticipated consequences that mitigate the health and efficiency effects of the intervention.

These unanticipated consequences are a natural outcome of the processes under­lying the health production function model. Changing a price in this model will generate a reoptimization of individual behavior that may undo the public action. Since prices affect different people differently, these public health actions will also have a redistributive function. Redistribution, rather than efficiency enhancement, is often the main economically beneficial effect of these activities.

When Is a Public Good Not Public?

Not all public goods (or bads) require government intervention to ensure adequate provision. In many situations, activities with a “public good” or “public bad” component are complementary to activities where excludability is not a problem. Consider restaurant inspection.

There is a substantial demand for information about the (overall) quality of restaurants and, recent evidence suggests, about their health characteristics as well (Jin and Leslie 2003). In 1998, the Los Angeles Health Department began requiring restaurants to post the results of annual inspections in their windows. Jin and Leslie (2003) show that, after the results of the scores were made pub­lic, customers became significantly less likely to patronize restaurants with a grade of B or lower (a B restaurant has a hygiene score of 80-89 on a scale where an A grade is 90-99 and a failing grade is below 60).

Information about the health quality of restaurants is a nonrivalrous public good—once I invest in learning about the quality of a restaurant, the informa­tion can be provided at no additional cost to everyone else. Thus, no individual would spend much time inspecting restaurant kitchens. This pattern would appear to suggest that information on the health quality of restaurants must be provided publicly. But this is less clear once the impact of the decline in res­taurant health quality is incorporated in the full health production model.

Without restaurant inspections, food poisoning would rise and fewer people would eat out. A private company could profit by developing a credible restau­rant inspection system and selling it to restaurants. Restaurants that paid the price of inspection would be permitted to post their grades in the window and to advertise that their quality was high. The value of the restaurant inspection company would depend on the extent to which restaurant customers viewed the reports as credible.

In the world of product safety, many analogous private quality assurance systems exist. Underwriters Laboratory, Good Housekeeping, and Best’s Insurance Reports are paid by product manufacturers to develop information that is valuable to consumers. In these cases, quality-monitoring activities, although they have public goods elements, are produced at high levels because they enable more of a complementary private product to be sold.

Public health provision of these pure quality-monitoring services and coer­cive actions to close down failures are likely to have relatively little impact on the overall health of the public. Public provision substitutes for—or crowds out—the private producer and so provides little new information. In fact, the quality of municipal restaurant inspection systems may be lower than would be that of the corresponding private system. Municipal restaurant inspection sys­tems have a monopoly on quality monitoring, and restaurant owners cannot respond to a decline in the credibility of these systems by refusing to purchase them.’

The main function of public provision of this information and mandatory participation in these systems is redistributive. Some (low-income) consumers will not be willing to pay a premium for eating in a restaurant that participates in a private quality monitoring system. These consumers may eat in unmoni­tored restaurants and suffer unfortunate health consequences. After public pro­vision of restaurant inspection information, the low-quality producers will no longer be available. Whether low-income consumers are better off or not depends on who pays the cost of the quality monitoring system and on what happens to the underlying price of the goods themselves. If, as is typically the case, the price of restaurant inspections and of maintaining the quality of restaurant hygiene at standard levels is borne by restaurants, the price of eating out will rise. Low-income consumers may now be priced out of the market, improving their health but, perhaps, diminishing their overall well-being. Alternatively, they may turn to lower-cost, less-regulated producers (street vend­ors, for example) and experience more, rather than fewer, health problems.

Many health department activities have the property that they are comple­mentary to private activities. There exists a thriving private market in the dis­semination of information. News media survive because consumers want information (a public good) and this information is complementary to advertis­ing markets. Public health messages brought to the attention of the news media, and likely to be of interest to the public, will tend to be disseminated broadly (often too broadly) by the media (Philipson and Posner 1994). While the public health impact of the development and initial dissemination of this information is often dramatic (see, for example, the impact of the surgeon general’s report on smoking), the empirical evidence suggests that later public dissemination efforts have somewhat less impact. Dissemination activities, again, tend to have a redistributive focus, rather than an efficiency focus.

Substitution between Public Goods and Individual Actions

Most standard economics textbooks (and virtually all health economics texts) use immunization as the classic example of a positive externality. Each of us benefits from the immunization decisions of others, and we would all like to free ride on others’ immunization choices. [Sanjeev: This is the most absurd claim; why would someone choose LESS THAN 100% immunity? There is NO free riding in the case of vaccines. Those who don’t take it believe that harms> benefits. They are ABSOLUTELY RATIONAL, unlike econmists who cook up a “free rider problem” for vaccines. ].  The textbook theory suggests that in the presence of such externalities, market equilibrium may generate inadequate levels of immunization. The arguments around immunization also apply to preventive actions taken to prevent the spread of sexually transmitted diseases or HIV. Use of condoms by one (uninfected) person reduces the prevalence of disease and hence the risk faced by other potential (uninfected) sex partners.

The textbook model suggests that these positive externalities should be off­set by subsidies for socially responsible behavior. Immunizations—or condoms—should be offered at less than market price to increase utilization. In the context of the health production model, however, the consequences of such subsidies are not quite so straightforward.

Recent economics work points out that this textbook analysis misses a further feedback loop (Philipson 1995, 1996; Ahituv, Hotz, and Philipson 1996). In the health production model, the decision to be immunized depends on both the price of immunization (or condom use) and on the benefits of this behavior. The benefits of preventive behavior, in turn, depend on the underlying prevalence of disease. Reducing the price of prevention generates a lower prevalence of disease and this, in turn, reduces the benefits of prevention, off­setting the initial price reduction.

This feedback loop makes it very difficult for individual prevention of infectious disease to drive prevalence to zero. [Sanjeev: This is the kind of worthless analysis which is part of economic theory. The concept of eradicating a disease is well beyond the capacty of eocnomics to understsand] As disease prevalence falls, whether in the case of measles or HIV, a growing literature shows that the level of precautions taken falls as well. Depending on the responsiveness of behavior to the price of prevention and to the benefits of prevention, subsidies may or may not reduce risk of disease. In either case, however, subsidies redistribute the risk of disease. Subsidies will tend to shift the prevalence of illness away from those who were centrally concerned with the price of prevention (poor people) toward those who were centrally concerned with disease prevalence. [Sanjeev: IF VACCINES WORK, THERE WOULD BE NO NEED TO SUBSIDISE. THE GENERAL SOCIAL INSURANCE SYSTEM WHICH TRANSFERS CASH TO THE POOR WOULD BE SUFFICIENT]

In the case of diseases with high prevalence, this feedback loop is likely to he very important. Private benefits from reducing risk will be very large and the social benefit from intervention correspondingly smaller. Conversely, in the case of diseases with low prevalence that are nearly eradicated, private benefits from preventive action will tend to be small and the social benefits of interven­tion very large, particularly if they lead to disease eradication.

Public Goods Generated by Social Insurance

As infectious disease prevalence has fallen, health departments’ focus has increasingly turned toward the prevention of chronic disease. Chronic disease clearly affects the health of the population. A chronic disease—or a chronic dis­ease risk factor—is only a public health problem, however, to the extent that it generates externalities.

The importance of externalities in motivating intervention in the case of chronic disease has generated numerous empirical economic studies assessing the external costs of tobacco use, alcohol use, physical inactivity, and most recently obesity. The seminal work in this genre (Manning et al. 1991) exam­ined the external costs of tobacco and alcohol use.

In some of these cases—most notably, the substantial external costs of alcohol-related automobile accidents—the external costs identified are classic public bads. In the case of obesity and physical activity, and tobacco as well, the exter­nal costs are largely those associated with public and private insurance pro­grams (such as Medicare and Medicaid) that cover the disability and medical costs of those who become ill in consequence of the risky behavior (and the offsetting benefits to private and public pension programs of premature mortality).

The external costs of illness associated with insurance programs can be, in principle—and often are, in practice—internalized through variable premiums. Life insurers have long charged higher premiums to smokers, as do non—group health insurers. A growing economics literature on obesity appears to indicate that the external costs of obesity are mainly borne by obese individuals them­selves, through lower wages that may offset their higher medical costs (Cawley 2004; Bhattacharya and Bundorf 2005).

Social insurance programs do not typically charge variable premiums based on risk behaviors or lifestyles. In the health production model, the fail­ure to charge variable premiums in social insurance programs reduces the costs of engaging in unhealthy behavior.⁹ In this context, taxation of the unhealthy behavior merely corrects for the existing and inefficient subsidy to unhealthy behavior generated by social insurance. Perhaps taxing the unhealthy activity has lower administrative costs than would assessing premiums appropriately.

Taxing bads to correct social insurance externalities may confer welfare benefits, but this is likely just a second-best solution to the existence of the ini­tial distortion. Moreover, the strategy of using public health taxes to undo the redistributive function of social insurance may act to erase the progressive bene­fits of social insurance itself (Remler 2004).

The Political Economy of Public Goods

Economic theory suggests a set of conditions that define when public health action is most beneficial—when these actions are most irreplaceable. Public health activities enhance efficiency most when they address issues where there is a serious “free rider” problem of nonexcludability or where a good is entirely nonrivalrous (both from the perspective of suppliers and purchasers). Not all free rider problems or problems of nonrivairy warrant public health interfer­ence, however. Public health action is not as necessary when the private mar­ket will generate the same activities. The private market may do so either because of complementarities between public health activities and private excludable goods markets or because public health actions substitute for the effects of the prevalence of disease itself. Public health does perform an eco­nomically useful function when it stems externalities generated by other social programs, but this is only a second- best activity necessitated by the distortions introduced through the social programs themselves. Public health is at its best when it is producing basic epidemiological or risk factor research, or putting into place measures that will benefit future generations.

This set of criteria, however appealing, places the actual enterprise of pub­lic health in a rather unfortunate position. Public health, economic theory says, is most useful and beneficial when nobody can observe cash savings because of the actions of public health; when public health activities don’t even try to reduce taxes; when the potential benefits of public health actions are unclear; and when the potential beneficiaries of public health activities aren’t even born yet!

The political participation that generates public health itself, however, is a public good. Mobilizing to protect public health requires costs that generate benefits—public goods—of service to the entire population. The economic def­inition of public health requires that it target areas that, by definition, have no apparent constituency to support them and where the outcomes of the activ­ities themselves are not readily predictable and countable. Often, the more observable an activity is, the less essential its public health function.

Public health, as a common good, suffers not only on the demand side, where all other public goods produced by government must also face the prob­lem of a lack of constituency. The production of public health is an activity that is inherently local, labor intensive, and fragmented. Unlike, for example, the concentrated suppliers of national defense services, the public health work­force is a diffused constituency that must overcome its own collective action problems to lobby government for increased funding. The core functions of public health have no lobby but the good government types and altruistic sanitarians.

In this context, the expansion of the boundaries of public health is, in eco­nomic terms, a survival strategy. By providing prevention services and addressing population health needs more broadly, it creates constituencies who benefit directly, immediately, and observably from the services, information, and sub­sidies provided by public health. The benefits of these activities are so readily apparent, however, mainly because they substitute for obvious private activities. The danger is that these sideshow activities will overtake the main stage. The expansion of the scope of public health—rather than creating budgetary space for core activities—may leave us with a public health system that provides services redundant to the private market, while neglecting those public health functions where public action is truly indispensable.