Policy Brief #18
Social and Economic Policies as Health Policy: Moving Toward a New Approach to Improving Health in America
Based on a chapter in Making Americans Healthier: Social and Economic Policy as Health Policy, Robert F. Schoeni, James S. House, George A. Kaplan, and Harold Pollack, eds. The volume can be ordered from the Russell Sage Foundation by clicking here.
• Ignoring all other benefits that might arise to society from increased schooling, educational grants appear strikingly cost-effective when evaluated solely as investments in population health.
• Large economic and social surveys (such as the March CPS) would increase their value to the policy community tremendously if even a modest set of health status data were collected.
• The accelerating dominance in the U.S. of medical expenditures exemplifies an unwise approach to reducing death and illness. If the public knew that social and economic policies and programs could be equally - perhaps even more - effective in improving health than some medical-care interventions, they might be more willing to embrace them.
Though not explicitly designed to do so, social and economic policies impact health at both the individual and the population level. Education, income support, civil rights, macroeconomic, employment, housing, environmental, law-enforcement, agriculture, and transportation policies are just some of the domains in which this occurs. Human health is too multifaceted, its determinants too varied, and our current state of knowledge too limited however, to easily understand the full scope and complexity of the links between social and economic policy and health. If we are to realize the full promise of economic and social policies as also health policies, we must grapple with these issues. We must try to firmly establish the causal impact of social and economic policies on health. Where causality can be established, we must attempt to better understand the pathways through which these effects occur. This is imperative not just for scientific understanding, but in order to develop more cost-effective policies. Finally, in order to do this, we must generate better data and research designs.
The Difficulty of Determining Causality
Establishing causal links between particular social and economic policies and health outcomes is not easy. There is deep disagreement among policy makers and researchers alike around the issue of inference – particularly with regard to the necessity and limitations of prospective randomized experiments. True experiments are few and far between. In certain areas, for example civil rights or macroeconomic policy, randomized assignment of individuals to different policies may be either unethical or impossible. In other areas, such as income maintenance or housing, in which policy experiments are feasible, the logistical and financial obstacles can be substantial. For example, though providing invaluable knowledge, some of the best known randomized experiments (e.g. the RAND Health Insurance Experiment, Movement to Opportunity) cost many millions of dollars and took years to implement and analyze (Burtless 1995; Heckman and Smith 1995).
In other domains, such as pilot welfare-to-work interventions where randomized experiments are more easily performed, generalizability is often open to question – despite strong internal validity. Conditions in one locale may vary from the nation or state as a whole, and a small temporary program may have a different impact than a large permanent one. Even when generalizable randomized experiments are performed, they often turn out to answer different questions from those that we who are interested in health outcomes want to answer. Existing experimental studies are often limited in focus and timeframe.
For all of these reasons, health researchers often turn to other study designs to address their policy questions. Cross-state differences, or changes in policies within states over time – sometimes referred to as “natural” or “policy” experiments – can provide alternative (and not infrequently employed) identification strategies. This approach offers valuable insights while at the same time, raises questions: why did some states choose more stringent policies or implement new policies more quickly than others?; can we draw conclusions about current or future policies based on the results of interventions and policies implemented in different eras and contexts?; will marginal program expansion have the same impact as initial implementation?
The Pathway Problem
Given that a particular social or economic policy has causally affected an outcome, it has almost always done so via multiple pathways. Untangling the relative size and direction of these pathways is itself a difficult task. Yet untangle we must if we are to address two issues of practicality, namely, can we do it?, and cost-effectiveness. These issues are well illustrated in the area of education policy. David Cutler and Adriana Lleras-Muney (chapter 2, Making Americans Healthier: Social and Economic Policy as Health Policy) describe the large and well-documented educational gradients in health outcomes and behaviors. But why less-educated people display greater behavioral risk is less clear. Is it due to their being less informed? Perhaps it is to relieve job or other life stress? Maybe their economic circumstances lead them to choices those more well-off can avoid? The best path to improving health outcomes will depend on the answer.
As discussed by Daniel Keating and Sharon Simonton (chapter 3, Making Americans Healthier: Social and Economic Policy as Health Policy), the example of early childhood education programs is illustrative. Such programs might improve health by raising cognitive performance or by providing a venue for nutrition services or by improving parenting skills. Each of these pathways, as well as a number of others, is plausible, and offers distinct implications for program design and implementation. Current research indicates that early childhood education is valuable; it does not yet clearly show which of these distinct pathways is most promising for improving health. The stakes are high, however, because existing research raises the prospect that public educational investments may provide large, but usually neglected, benefits for health.
Back-of-the-envelope calculations by Cutler and Lleras-Muney suggest that each $1000 increase in grant aid to extend educational attainment of low-income children is associated with between .03 and .10 years of life. To those outside the domain of health policy, these benefits might seem small, but in fact, such an improvement in longevity is markedly larger than those obtained from widely accepted health interventions such as colorectal cancer screening (Frazier et al., 2000). Put another way, education grants appear highly cost-effective relative to standard thresholds used to evaluate medical and public health interventions. Cutler and Lleras-Muney’s calculations imply that education grants cost between $10,000 and $33,333 per year of life gained. By current convention, medical and public health interventions are considered cost-effective when they cost less than $100,000 per “quality-adjusted-life-year” (Hirth et al., 2000). Ignoring all the other benefits that might arise to society from increased schooling, educational grants appear strikingly cost-effective when evaluated solely as investments in population health. To maximize the cost-effectiveness however, we need to better understand which aspects of education, operating through which pathways, have the strongest effects on health. Education is, of course, just one example of this phenomenon. Other social and economic policies (civil rights, housing, employment, agriculture, etc.) pose similar questions.
The Need for Better Data
To better understand the causal impacts of social and economic policies on health, as well as the pathways through which such effects are produced, adequate health data must be collected during evaluations of social and economic policies. Much more than the typical self-report of health must be employed. Fortunately, detailed health data, from specific local surveys to national epidemiological surveys (such as the National Health and Nutrition Examination Survey, and the National Health Interview Survey) and national panel surveys (such as the National Longitudinal Study of Adolescent Health), are now routinely collected, and might be linked to social and economic policy evaluations. Even more useful would be if data collection and dissemination were more carefully coordinated between data collectors and the policy research community. Some practices, such as the masking of state identifiers, hinder the study of key policies that affect health. In other cases, elaborate data sets are narrowly prepared for one purpose alone. For example, the March Current Population Survey, which is the source for official government estimates of poverty and contains high-quality income data, would increase its value to the policy community if it collected even a modest set of health status data.
Policy research would also benefit if health considerations were taken more seriously in the design and analysis of social and economic program evaluations. When large-scale evaluations of, for example, welfare-to-work, criminal-justice, or housing interventions are undertaken, modest additional investments could allow important health matters to be explored as well. Given the large economic and social burdens associated with poor health, policy evaluations would wisely include attention to health effects. As in the case of education, cost-benefit ratios of non-health policies would be greatly enhanced by consideration of the health effects as well as more typically considered outcomes (such as income or the case of education). This can also help to unravel the paradox in the United States of high health expenditures and relatively poor health outcomes. Social policies not ostensibly aimed at medical care or public health provide a key venue through which public action improves – or erodes – population health.
A Paradox and a Promise
Policy makers face great difficulties in striking the proper balance across policy options that promote health. Ironically, the very success, not to mention popularity, of medical care threatens to undermine this balance. The rapid increase in expenditures on government health programs increasingly threatens the funding of other public activities which may have equal or greater effects on health. If, for example, increased Medicare or Medicaid spending is financed by reduced expenditures on housing, education, or income maintenance, we may experience decidedly mixed results with regard to overall population health. To put this in perspective, a task force of the Center for American Progress has proposed a set of policies, not unlike those recently enacted in the UK, designed to cut U.S. poverty in half at an annual cost of $90 billion (Greenberg, Dutta-Gupta, and Minoff 2007). Though substantial, this expenditure amounts to less than 20% of current federal health care spending. Looked at another way, according to the Congressional Budget Office in January 2005, federal Medicare and Medicaid expenditures totaled $473 billion in 2004, and were projected to increase (in real terms) by nearly 50% by 2010. Combined mandatory income security expenditures, a category that includes Unemployment Compensation, SSI, the Earned Income Tax Credit, food stamps, TANF, child nutrition and foster care, totaled just $191 in 2004, and was projected to rise by less than 3% over the same six year period.
The accelerating dominance of medical expenditures exemplifies an unwise approach to reducing death and illness in the United States. At the same time, most Americans appear substantially more willing to spend public funds to address unequal health outcomes and unequal access to medical care than they are to address unequal conditions of housing, education and cash income (Schlesinger 2004). Such narrow targeting of egalitarian aspirations may prove self-defeating, or at least self-hindering. If the public knew that social and economic policies and programs could be equally – or perhaps more – effective in improving health than some medical-care interventions, they might be more willing to embrace them. Economic and social policy is health policy.
Burtless, Gary. 1995. “The Case for Randomized Field Trials in Economic and Policy Research.” Journal of Economic Perspectives 9(2): 63-84.
Frazier, A. Lindsay, Graham A. Colditz, Charles S, Fuchs, and Karen M. Kuntz. 2000. “Cost-Effectiveness of Screening for Colorectal Cancer in the General Population.” Journal of the American Medical Association 284(15): 1954-61.
Greenberg, M., I. Dutta-Gupta, and E. Minoff. 2007. From Poverty to Prosperity: A National Strategy to Cut Poverty in Half. Washington: Center for American Progress.
Heckman, James J., and Jeffrey A. Smith. 1995. “Assessing the Case for Social Experiments.” Journal of Economic Perspectives 9(2): 85-110.
Hirth, Richard A., Michael E. Chernew, Edward Miller, A. Mark Fendrick, and William G. Weissert. 2000. “Willingness to Pay for a Quality-Adjusted Life Year: In Search of a Standard.” Medical Decision Making 20(3): 332-42.
Schlesinger, Mark. 2004. “Reprivatizing the Public Household? Medical Care in the Context of American Values.” Journal of Health Politics, Policy and Law 29(4-5): 969-1004.
About the authors
Robert F. Schoeni is professor of public policy and economics, the University of Michigan.
James S. House is Angus Campbell Collegiate Professor of Sociology and Survey Research, the University of Michigan
George A. Kaplan is the Thomas Francis Collegiate Professor of Public Health, the Univesity of Michigan.
Harold Pollack is associate professor of social service administration, University of Chicago.
The National Poverty Center is charged with promoting high-quality research on the causes and consequences of poverty, evaluating and analyzing policies to alleviate poverty, and training the next generation of poverty researchers.
The NPC is directed by Sheldon H. Danziger.
The National Poverty Center’s Policy Brief series summarizes key academic research findings, highlighting implications for policy.
The NPC encourages the dissemination of this publication and grants full reproduction right to any party so long as proper credit is granted the NPC. Sample citation: “Title, National Poverty Center Policy Brief #x”.
Major funding for the National Poverty Center is provided by the Office of the Assistant Secretary for Planning and Evaluation, U.S. Department of Health and Human Services.
Any opinions, findings, conclusions, or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the National Poverty Center or any sponsoring agency.
National Poverty Center
735 S. State Street
Ann Arbor, MI 48109-3091