By Leighton Ku, Douglas J. Besharov and Peter Germanis

This article originally appeared in The Public Interest, Spring 1999.

In their critical review of the Women, Infants and Children (WIC) nutrition program ("Is WIC As Good As They Say," The Public Interest, No. 134, Winter 1999), Douglas J. Besharov and Peter Germanis conclude that WIC is less effective than commonly believed and that the research evidence is flawed. They suggest various ways to restructure the program in fashions more to their liking, although they offer little evidence that their proposals are feasible or would be more effective. I appreciate the spirit of constructive criticism brought by Besharov and Germanis, but I disagree with many of their conclusions.

Let me begin by observing that Besharov and Germanis's article is incomplete and stilted. Public-health research about the WIC program has been amazingly fertile. A brief search of scientific papers published in health journals collected by the National Library of Medicine found more than 200 published studies concerning WIC, and there are other excellent papers in public-policy or economics journals, as well as unpublished reports. Despite its relatively small size, WIC has been one of the most thoroughly studied federal programs. Researchers have investigated WIC's effects on: birth weights, infant mortality, reductions in Medicaid expenditures, nutrient intakes, children's cognitive development, immunization rates and use of health services, childhood anemia, and breastfeeding rates. Many of the most interesting papers are not rigorous evaluations but reports of field studies or descriptive analyses that help us look at questions just like those that Besharov and Germanis raise: how to improve breastfeeding rates and nutritional counseling, and how to use WIC to improve immunization rates or to coordinate WIC with home-visit programs.

Much of their paper discusses flaws in studies about whether WIC improves birth outcomes. They note methodological problems, like selection bias in many of the studies, but fail to note that Jack Metcoff of the University of Oklahoma and his associates conducted a randomized experiment to determine how WIC affects pregnant women's birth outcomes. This study, published in 1985, found that WIC increased birth weights by an average of 91 grams and meets most of their methodological objections.

Growth in the program since then has made it ethically impossible to replicate this randomized study, but most subsequent studies, conducted using a plethora of data sources, statistical methods, and outcome measures, generally reach the same conclusions-that WIC helps low-income pregnant women bear heavier, healthier babies. Come on guys, how long do we have to study something to conclude reasonably that it works? Are you acting like the fringe groups who wonder whether smoking tobacco really causes lung cancer?

BESHAROV and Germanis also understate the strength of the evidence. For example, they note that Ray Yip and his colleagues at the Center for Disease Control (CDC) found that decreasing anemia among WIC participants could be caused by a secular decline in childhood anemia. What they do not mention is that the CDC also found that fewer WIC children were anemic between the time they first entered the program and recertification, usually six months later. These specific and rapid improvements could not be caused by general secular changes and almost certainly are due to the iron-fortified foods in the WIC food package. These clinical findings are consistent with results from the National WIC Evaluation and those from Donald Rose of the Economic Research Service of USDA and his associates, indicating that WIC children consumed more key nutrients, including iron. While there have not been definitive, randomized studies for children, the research findings are consistent and compelling.

Equally important, WIC can improve children's health by boosting immunization rates. Because WIC is typically located in health clinics and serves millions of low-income children, it is well-suited to promote immunization. Mothers and their young children must regularly come to WIC clinics to be certified or to get food vouchers. The National WIC Evaluation found that children participating in WIC were more likely to be immunized than nonparticipating children. Several recent studies, including two published in the Journal of the American Medical Association, showed how changes in WIC program operations can further increase immunization rates. State and local WIC clinics have also formed partnerships with Vaccines for Children, Medicaid, and local maternal and child health programs to reinforce preventive health care for children. By increasing immunizations, WIC can indirectly reduce the incidence of preventable diseases like measles.

Other aspects of the WIC program have been less thoroughly studied. But contrary to Besharov and Germanis's assertion that they are unveiling a secret hidden from the American public, these gaps have been well-known for a long time and publicly discussed by the General Accounting Office, the National Academy of Sciences, myself, and fellow researchers like Barbara Devaney of Mathematica Policy Research. And we have made steady gains in our understanding of the program; the U.S. Department of Agriculture, which administers WIC, has long funded research about WIC. Do I wish for more and better evaluations? Of course. I would love to see more research about the effects of participation in WIC during childhood and whether the current program design provides the best package of benefits. However, I realize that rigorous studies are difficult to design, cost a lot, and take many years to complete. The fiscal reality is this: Research budgets are limited, and evaluation studies must compete for resources with other priorities.

WHILE the research evidence about WIC's effectiveness is not perfect, it is hard to think of any public program with so consistent a body of positive research findings. Further, the federal government continues to invest in research to better understand and strengthen the program. No doubt this strong research record accounts for the program's popularity. But the program also receives high marks for its design. Steering a middle course between the problem of entitlements and the inherent vagueness of block grants, WIC is a federal program that allows for local and state administration. This has satisfied WIC's state and local managers, allowing them to be creative within a framework of broadly shared goals and a time-tested program structure.

One good example of state or local innovation is WIC's infant formula rebates, in which competitive bidding reduces the price of WIC infant formula. Typically, the rebates amount to 80 percent or more of the price of formula, collectively saving more than $ 1 billion annually. The rebates began as state initiatives and have become one of the best examples of the power of competitive bidding within government. Linking WIC to immunization efforts is another example of creative local initiatives.

WIC incorporates principles shared by conservatives and liberals. It seeks to prevent problems, not just cure them. It strives for efficiency and cost-containment in operations. It fosters family responsibility by teaching parents how to improve their children's diets, as well as by providing healthy food directly. It is not an unlimited handout but provides specific benefits during a critical stage in a child's development. Finally, it permits creative collaboration between program directors at federal, state, and local levels.

BESHAROV and Germanis suggest that WIC should have even more flexibility and that paying for additional services by providing fewer people with food benefits would be the right thing to do. In particular, they propose that WIC should try home visits for pregnant women, based on a model developed and tested by David Olds at the University of Colorado. There are indeed many proven merits to home-visit programs, but I am not sure that paying for the entire range of home-visit services is WIC's mission. Many other programs are better suited to financing non-nutrition-related services for pregnant women, including Healthy Start, Title V Maternal and Child Health Block Grants, Title XX Social Services Block Grants, and Medicaid. Providing WIC nutrition services through home visits, in coordination with other programs, is an attractive idea, but this is already permitted under current program rules.

Flexibility for state and local programs is important but does not, by itself, guarantee success. The Healthy Start program gives communities great autonomy to design and implement projects to improve birth outcomes, including home-visit programs, but preliminary, interim evaluations have been disappointing. Does it make sense to divert resources from WIC's relatively well-proven model toward one yielding inconclusive results to date?

Besharov and Germanis apparently believe that WIC provides food to too many mothers and children. For example, they complain that almost half of the infants in the country get WIC benefits. What they do not mention is that there is an alarmingly high level of poverty and near poverty among American families with young children. Mothers with infants are often unable to work full time and, even when they work, often have low-wage jobs. To qualify for WIC, a family of four must generally have an annual income of about $ 30,000 or less, which most of us would agree is a relatively low income. The main reason that almost half the nation's infants receive WIC benefits is that, sadly, about half the babies are in poor and low-income families.

Without WIC, a low-income family would need to spend more than $ 1,000 a year to feed its baby and purchase the infant formula and other fortified foods provided by WIC. Few low-income families can afford to spend this much on their own. WIC ensures that these babies have nutritious foods available. In addition, competitive purchasing allows the government to purchase these foods at a fraction of the market cost. Rather than revealing a program run amuck, the high participation of infants in WIC signals a successful government policy that helps ensure the nutritional status of millions of low-income infants.

Overall, however, it has been estimated that only about 80 percent of those eligible for WIC (including women and preschool children) are able to get benefits. More health and social services for the mothers and children most at risk are urgently needed, but many other programs can provide these services, along with WIC. Let's be sure that we have utilized all the resources to their utmost and coordinated programs' efforts, before we propose taking food away from low-income women and children.

Besharov and Germanis have raised many provocative issues, and I hope that this stimulates a constructive discussion of ways to improve services for low-income mothers and children. But I have read the research literature, and I remain confident that WIC is effective and well run. We can and should refine the WIC program to improve the health of lowincome mothers and children, but let's not forget program fundamentals.

IN his response to our article, Leighton Ku, a careful researcher and a self-described advocate of the Women, Infants and Children (WIC) program, demonstrates what is so sadly wrong about the public-policy debate these days: Eager to defend or expand a social program, advocates and politicians trumpet favorable research findings to claim that the program "works." Often, they assert that the program actually saves public money by preventing other social problems or expenditures. Never mind that the assertion is questioned by unbiased experts in the field and that, even if true, the impact is insufficient to make a real dent in the underlying problem.

As evaluation studies have proliferated in recent years, so have such advocacy-oriented research claims. Probably the best example concerns Head Start. Research on the Perry Preschool Program dating back to the 1960s suggested that early-childhood education programs can substantially affect children's later lives, in areas such as school drop-out rates, criminal behavior, out-of-wedlock teen births, welfare dependency, and unemployment. Advocates claimed that the program saved from $ 3 to $ 7 for each dollar spent. To deal with the fact that the Perry Preschool Program only vaguely resembled Head Start, they often loosely attributed this cost-benefit to "Head Start-like" programs. Since most of us believe that a child's upbringing matters, such evaluations were widely accepted. But no other rigorously evaluated, early-childhood program has had even remotely similar results (sharply undermining even the Perry Preschool results). Even those programs that report initial, impressive improvements in children's performance conclude that these gains "fadeout" over time.

Most recently, the federal government mounted the Comprehensive Child Development Program (CCDP), a five-year, $ 125 million early-intervention program at 24 sites. Despite an average annual investment of $ 15,768 per family (above and beyond welfare payments, food stamps, Medicaid, housing, etc.), the children who received the services had outcomes that were little different from those who did not.

Family-preservation services are another example of an oversold social program. Again, the idea makes sense: Use intensive rehabilitative and supportive services to improve parenting so that children do not have to be removed from home and placed in foster care. And again, out-sized results from one or two early studies were used to justify a massive federal program-in this case, the Family Preservation and Support Program, funded at almost $ 1 billion for its first five years. The ink had not yet dried on the new law before other, more carefully designed studies deflated claims that each dollar spent on family preservation services would save $ 5 to $ 6 in foster care and other expenditures.

ONE could provide many other examples and, in effect, our article simply added WIC to the list of programs whose exaggerated impacts should be brought down to earth. We made several broad points: There is a widespread and bipartisan belief that WIC is a tremendously cost-effective program. The most common claim, to quote a 1997 speech by Agriculture Under Secretary Shirley Watkins, is that "for every dollar spent in WIC benefits, three dollars are saved in Medicaid dollars." And in 1998, an editorial in the Washington Post asserted that "repeated studies have shown that the program saves far more in health care costs than it spends." Such claims are simply untrue. The relevant research applies to only a small subset of the WIC caseload-pregnant women. Moreover, most of the research upon which the costs and benefits are calculated are plagued by unresolved problems of selfselection and simultaneity bias. The widely proclaimed results are, in other words, at the high end of possible impacts. Nevertheless, some parts of the WIC program undoubtedly make a difference in the lives of low-income women and children. Here is exactly what we said:

WIC for pregnant mothers has perhaps zero to substantial impacts on infant mortality, prematurity, and birth weight. But the rest of the program (for infants, children, and postpartum and breastfeeding mothers) has small to modest impacts on anemia and nutrient intake. There is also a glimmer of evidence that WIC's beneficial effects are concentrated among the most needful recipients.

The priority should be to learn more about what in WIC improves the health of low-income mothers and their young children-and what doesn't. Thus, and most important, we did not argue that WIC should be defunded, or even cut back. Instead, we proposed that: (1) more WIC resources should be targeted on the most needful families; (2) counseling services should be intensified-and made more directive; (3) states should be allowed to exceed federal limits on spending for WIC services; and (4) alternative service configurations should be tested. Moreover, we cautioned that these ideas "are untried and, therefore, should be carefully evaluated." Hardly a radical set of proposals.

Yet Ku says that our review is "incomplete and stilted." He says we ignored hundreds of studies, but, as he notes, most are of highly questionable validity ("not rigorous evaluations," are his words)-and are ignored in most research syntheses.

Significantly, Ku does not defend the benefit-cost estimates we criticize, nor does he assert a particular statistical impact for the program. The plain truth is that no one knows how much difference the program really makes. As recently as 1994, Ku and other researchers at the Urban Institute concluded: "More research is needed on the effect of WIC participation for children (over half of all WIC recipients), given that this is the group that will grow most under full funding."

SO far, of course, exaggerating WIC's impact has helped fuel the program's expansion. In the last 10 years, the number of WIC recipients has more than doubled. Last year, the bill reauthorizing WIC sailed through the House of Representatives by a vote of 383 to 1. But just as the public and politicians have, over time, gained more nuanced understandings of Head Start, family-preservation services, and a host of other once-favored programs, so too will they eventually acquire a more realistic view of WIC.

In the meantime, the problems that it and other socialwelfare programs are meant to address remain and fester. Hence, rather than try to counter Ku's comments, we prefer to underscore the broader sociopolitical tendencies that they reflect.

Circle the wagons, thereby preventing programmatic improvement. Ku's real complaint is not that we have misstated the evidence of WIC's effect but that, if accepted, our view of the research evidence would undercut public and political support for the program. That is a valid concern. But obscured by rhetoric that WIC "works" are potentially life-saving questions about how to improve it and thus the well-being of possibly thousands of underprivileged children. Sound policy making requires that programs be continually evaluated and, when necessary, their services reoriented or, at least, refined.

Demonize skeptics, thereby avoiding the need to address their concerns. Although we appreciate the general tone of Ku's response, at one point he accuses us of "acting like the fringe groups who wonder whether smoking tobacco really causes lung cancer." In case there is any question: Yes, we think smoking causes lung cancer in many people. But some long-time smokers escape serious harm. Similarly, WIC benefits many, but surely not all, participants. At another point, Ku says that our proposal would be "taking food away from low-income women and children." Honest, we don't want anyone to go hungry. We just want to help those who most need the WIC program.

Blame insufficient funding, thereby excusing inadequate theory and poor implementation. When forced to acknowledge a program's weaknesses, the advocate's first riposte is that not enough money is being spent on it. By many estimates, the cost of full-day, full-year Head Start now exceeds $ 10,000 per year, compared to the average of about $ 6,000 spent by middle-class families for equivalent hours of child care. Higher funding may improve services, but often the services are simply irrelevant to the family's problems or are so poorly managed that they have little impact. (Remember the $ 15,768 per family spent by the unsuccessful CCDP.)

Include the middle class, thereby creating an everlarger constituency for government spending. As we point out, about 50 percent of all newborns qualify for WIC, as do 25 percent of children aged 1 to 5. Ku attributes this to "an alarmingly high level of poverty and near poverty." But income eligibility for WIC, one of the most generous of federal public-assistance programs, reaches to families of four with incomes up to $ 30,000 a year. One wishes these families had higher incomes but to call them "poor" is simply wrong. (In 1997, the median income for all American families with children was $ 43,545.)

Don't trust the states and localities, thereby reducing programmatic responsiveness and stifling innovation. In the 1930s, the federal government helped states because they needed additional funds; but, in the aftermath of the civilrights struggles of the 1960s, federal involvement was increasingly motivated by a corrosive mistrust of the states. The welfare reform of 1996 was a major step in the other direction, but most federal programs still mandate the specifics of state implementation in excruciating detail. America, however, is a diverse, continental nation. What works for New York City may not work for Albuquerque. More importantly, centralized planning did not work for the Soviet economy, and it has not worked for American social-welfare services. At any given time, there are a multitude of promising ways to address child poverty, inadequate nutrition, and child maltreatment. Limiting the states to a single, centrally funded approach has foreclosed the marketplace of ideas that should characterize a federal system.

EACH of these points, of course, challenges conventional social policy, at least as it has been conducted since the 1960s. Recently, we have seen crime decline, probably because of improved policing, rather than because of public-jobs programs. Similarly, out-of-wedlock births have declined, probably as a result of more conservative attitudes, rather than the free distribution of contraceptives; and welfare rolls have declined, probably because of work expectations and the end of the entitlement mentality, rather than through job training. If social programs are to succeed, they must be open to reconsideration and reorientation. We hope that raising these points does not mean we are "acting like a fringe group." Indeed, we hope there is an emerging majority that favors a more open-minded and candid discussion of America's pressing social problems.