The U.S. food stamp programtoday known as the Supplemental Nutrition Assistance Program (SNAP)was developed during the presidencies of John F. Kennedy and Lyndon B. Johnson. Its purpose i

The Review of Economics and Statistics VOL. XCIII MAY 2011 NUMBER 2

INSIDE THE WAR ON POVERTY: THE IMPACT OF FOOD STAMPS

ON BIRTH OUTCOMES

Douglas Almond, Hilary W. Hoynes, and Diane Whitmore Schanzenbach*

Abstract—This paper evaluates the health impacts of a signature initiative of the War on Poverty: the introduction of the modern Food Stamp Pro- gram (FSP). Using variation in the month FSP began operating in each U.S. county, we find that pregnancies exposed to FSP three months prior to birth yielded deliveries with increased birth weight, with the largest gains at the lowest birth weights. We also find small but statistically insig- nificant improvements in neonatal mortality. We conclude that the sizable increase in income from FSP improved birth outcomes for both whites and African Americans, with larger impacts for African American mothers.

I. Introduction

IN this paper, we evaluate the health consequences of a sizable improvement in the resources available to Ameri-

ca’s poorest. In particular, we examine the impact of the Food Stamp Program (FSP), which in 2007 provided $34 billion in payments to about 13 million households, on infant health. Our paper makes two distinct contributions. First, although the goal of the FSP is to increase the nutri- tion of the poor, few papers have examined its impact on health outcomes. Second, building on work by Hoynes and Schanzenbach (2009), we argue that the FSP treatment represents an exogenous increase in income for the poor. Our analysis therefore represents a causal estimate of the impact of income on health, an important topic with little convincing evidence due to concerns about endogeneity and reverse causality (Currie, 2009).

We use the natural experiment afforded by the nation- wide rollout of the modern FSP during the 1960s and early 1970s. Our identification strategy uses the sharp timing of the county-by-county rollout of the FSP, which was initially constrained by congressional funding authorizations (and ultimately became available in all counties by 1975). Speci- fically, we use information on the month the FSP began operating in each of the roughly 3,100 U.S. counties and examine the impact of the FSP rollout on mean birth weight, low birth weight, gestation, and neonatal mortality.

Throughout the history of the FSP, the program para- meters have been set by the U.S. Department of Agriculture (USDA) and are uniform across states. In the absence of the state-level variation often leveraged by economists to eval- uate transfer programs, previous FSP research has typically resorted to strong assumptions as to the comparability of FSP participants and eligible nonparticipants (Currie, 2003). Not surprisingly, the literature is far from settled as to what casual impact (if any) the FSP has on nutrition and health.

Hoynes and Schanzenbach (2009) use this county rollout to examine the impact of the FSP on food consumption using the PSID. They found that the introduction of the FSP increased total food spending and decreased out-of-pocket food spending. Importantly, consistent with the predictions of canonical microeconomic theory, the magnitude of the increase in food expenditures was similar to an equivalent- sized income transfer, implying that most recipient house- holds were inframarginal (that is, they would spend more on the subsidized good than the face value of the in-kind transfer). As one of the largest antipoverty programs in the United States—comparable in cost to the earned income tax credit (EITC) and substantially larger than Temporary Assistance to Needy Families (TANF)—understanding FSP effects is valuable both in its own right and for what it reveals about the relationship between income and health.1

We focus on birth outcomes for several reasons. First, families represent an important subgroup of the food stamp caseload. Over 60% of food stamp households include chil-

Received for publication February 4, 2009. Revision accepted for publi- cation December 9, 2009.

* Almond: Columbia University and NBER; Hoynes: University of California, Davis and NBER; Schanzenbach: Northwestern University and NBER.

We thank Justin McCrary for providing the Chay-Greenstone-McCrary geography crosswalk and Karen Norberg for advice on cause-of-death codes. This work was supported by a USDA Food Assistance Research Grant (awarded by the Joint Center for Poverty Research at Northwestern University and University of Chicago), the Population Research Center at the University of Chicago, and USDA FANRP Project 235, ‘‘Impact of Food Stamps and WIC on Health and Long Run Economic Outcomes.’’ We also thank Ken Chay, Janet Currie, Ted Joyce, Bob LaLonde, Doug Miller, Bob Whitaker, and seminar participants at the Harris School, Dart- mouth, MIT, LSE, the California Center of Population Research (UCLA), Duke, Cornell, UC Irvine, IIES (Stockholm University), the NBER Sum- mer Institute, and the SF Fed Summer Institute for helpful comments. Alan Barreca, Rachel Henry Currans-Sheehan, Elizabeth Munnich, Ankur Patel, and Charles Stoecker provided excellent research assistance, and Usha Patel entered the regionally aggregated vital statistics data for 1960 through 1975.

The online appendix referred to throughout the article is available at http://www.mitpressjournals.org/doi/suppl/10.1162/REST_a_00089.

1 The cost of the FSP was $33 billion in 2006 (compared to $24 billion for TANF, $33 billion for the EITC, and $5.4 billion for WIC, the Special Supplemental Nutrition Program for Women, Infants and Children).

The Review of Economics and Statistics, May 2011, 93(2): 387–403

� 2011 by the President and Fellows of Harvard College and the Massachusetts Institute of Technology

dren, and one-third have at least one preschool-age child. Second, birth outcomes improved substantially during the late 1960s and early 1970s. Third, to the extent that the FSP improved birth outcomes, later-life health outcomes of these cohorts may have also benefited (Barker, 1992; Black, Devereux, & Salvanes, 2007). Finally, the vital statistics data used in this project are ideally suited for analyzing FSP rollout: the birth (death) microdata contain the county of birth (death) and the month of birth (death). This, com- bined with the large sample sizes (for example, more than 1 million birth records per year in the data set), allows us to use the discrete nature of the FSP rollout with significant statistical power.

We find that infant outcomes improve with FSP introduc- tion. Changes in mean birth weight are small, increasing roughly half a percent for blacks and whites who partici- pated in the program (effect of the treatment on the treated). Impacts were larger at the bottom of the birth weight distri- bution, reducing the incidence of low birth weight among the treated by 7% for whites and between 5% and 11% for blacks. Changes in this part of the birth weight distribution are important because they are closely linked to other new- born health measures. Although not all treatment effects are statistically significant, they point consistently to improve- ments in birth weight following the introduction of the FSP. We also find that the FSP introduction leads to a reduction in neonatal mortality, although these results rarely reach statistical significance. We find very small (but precisely estimated) impacts of the FSP on fertility, suggesting that the results are not biased by endogenous sample selection. All results are robust to various sets of controls, such as county fixed effects, state-by-year fixed effects, and county- specific linear trends. Moreover, FSP impact estimates are robust to and little changed by county-by-year controls for federal spending on other social programs, suggesting our basic identification strategy is clean. Finally, we present an event study analysis that further supports the validity of the identification strategy.

Food stamps are the fundamental safety net in the United States. Unlike other means-tested programs, there is no additional targeting to specific subpopulations. Current benefits average about $200 per recipient household per month. Our analysis constitutes the first evidence that despite the fact that if did not target pregnant mothers (or even women), introduction of the FSP improved newborn health.

II. Introduction of the Food Stamp Program

The modern FSP began with President Kennedy’s 1961 announcement of a pilot food stamp program that was to be established in 8 impoverished counties. The pilot programs were expanded to 43 counties in 1962 and 1963. The suc- cess with these pilot programs led to the Food Stamp Act of 1964 (FSA), which gave local areas the authority to start up the FSP in their county. As with the current FSP, the pro-

gram was federally funded, and benefits were redeemable at approved retail food stores. In the period following the pas- sage of the FSA, a steady stream of counties initiated such programs, and federal spending on the FSP more than doubled between 1967 and 1969 (from $115 million to $250 million). Support for requiring counties to participate in FSP grew due to a national spotlight on hunger (Berry, 1984). This interest culminated in passage of 1973 amend- ments to the Food Stamp Act, which mandated that all counties offer FSP by 1975.

Figure 1 plots the percentage of counties with an FSP from 1960 to 1975.2 During the pilot phase (1961–1964), FSP coverage increased slowly. Beginning in 1964, pro- gram growth accelerated, and coverage expanded at a steady pace until all counties were covered in 1974. Furthermore, there was substantial heterogeneity in the tim- ing of adoption of the FSP, both within and across states. The map in figure 2 shades counties according to the date of FSP adoption (darker shading denotes a later start-up date). Our basic identification strategy considers the month of FSP adoption for each county the FSP ‘‘treatment.’’3

For our identification strategy to yield causal estimates of the program, it is key to establish that the timing of FSP adoption appears to be exogenous. Prior to the FSP, some counties provided food aid through the Commodity Distri- bution Program (CDP), which took surplus food purchased by the federal government as part of an agricultural price support policy and distributed those goods to the poor. The 1964 Food Stamp Act allowed counties to voluntarily set up an FSP, but the act also stated that no county could run both the FSP and the CDP. Thus, for counties that pre- viously ran a CDP, adoption of the FSP implies termination of the CDP.4 The political accounts of the time suggest that debates about adopting the FSP pitted powerful agricultural interests (which favored the CDP) against advocates for the poor (who favored the FSP; see MacDonald, 1977; Berry, 1984).5 In particular, counties with strong support for farm-

2 Counties are weighted by their 1970 population. Note this is not the food stamp caseload, but represents the percentage of the U.S. population that lived in a county with an FSP.

3 This timing lines up exceptionally well with county-level FSP spend- ing as measured in the Regional Economic Information System data. See online appendix table 3.

4 This transition in nutritional assistance would tend to bias FSP impact estimates downward, but we do not think this bias is substantial because of the limited scope of the CDP. The CDP was not available in all coun- ties, and recipients often had to travel long distances to pick up the items. Further, the commodities were distributed infrequently and inconsistently, and provided a narrow set of commodities. The most frequently available were flour, cornmeal, rice, dried milk, peanut butter, and rolled wheat (Citizens’ Board of Inquiry 1968). In contrast, food stamp benefits can be used to purchase all food items (except hot foods for immediate consump- tion, alcoholic beverages, and vitamins).

5 In fact, as Berry (1984) and Ripley (1969) noted, passage of the 1964 Food Stamp Act was achieved through classic legislative logrolling. The farm interest coalition (southern Democrats, Republicans) wanted to pass an important cotton-wheat subsidy bill while advocates for the poor (northern Democrats) wanted to pass the FSA. Neither had majorities, yet they made an arrangement, supported each others’ bills, and both bills passed.

388 THE REVIEW OF ECONOMICS AND STATISTICS

ing interests (such as southern or rural counties) may be late adopters of the FSP. Counties with strong support for the low-income population (such as northern, urban counties

with large populations of poor) may adopt FSP earlier in the period. This systematic variation in food stamp adoption could lead to spurious estimates of the program impact if

FIGURE 2.—FOOD STAMP PROGRAM START DATE BY COUNTY (1961–1975)

Authors’ tabulations of food stamp administrative data (U.S. Department of Agriculture, various years). The shading corresponds to the county FSP start date, where darker shading indicates later county imple- mentation.

FIGURE 1.—WEIGHTED PERCENTAGE OF COUNTIES WITH A FOOD STAMP PROGRAM, 1960–1975

Authors’ tabulations of food stamp administrative data (U.S. Department of Agriculture, various years). Counties are weighted by their 1960 population.

389INSIDE THE WAR ON POVERTY

those same county characteristics are associated with differ- ential trends in the outcome variables.

In earlier work (Hoynes & Schanzenbach, 2009), we documented that larger counties with a greater fraction of the population that was urban, black, or low income indeed implemented the FSP earlier, consistent with the historical accounts. We sought to predict FSP adoption date with 1960 county characteristics—those recorded immediately prior to the pilot FSP phase. That analysis showed that larger coun- ties and those with a higher share of black, elderly, young, or low income implemented earlier and those where more of the land was used in farming implement later.6 Neverthe- less, the county characteristics explain very little of the var- iation in adoption dates (see online appendix figure 1). This is consistent with the characterization of funding limits con- trolling the movement of counties off the waiting list to start up their FSP: ‘‘The program was quite in demand, as con- gressmen wanted to reap the good will and publicity that accompanied the opening of a new project. At this time there was always a long waiting list of counties that wanted to join the program. Only funding controlled the growth of the pro- gram as it expanded’’ (Berry, 1984, pp. 36–37).

We view the weakness of this model fit as a strength when it comes to our identification approachin that much of the variation in the implementation of FSP appears to be idiosyncratic. Nonetheless, in order to control for possible differences in trends across counties that are spuriously cor- related with the county treatment effect, all of our regres- sions include interactions of these 1960 pretreatment county characteristics with time trends as in Acemoglu, Autor, and Lyle (2004) and Hoynes and Schanzenbach (2009).

FSP introduction took place during a period of tremen- dous expansion in cash and noncash transfer programs as the War on Poverty and Great Society programs were expanding. To disentangle the FSP from these other pro- grams, the county-by-month variation in FSP rollout is key. Further, given that virtually all means-tested programs are administered at the state level, our controls for state-by-year fixed effects should absorb these program impacts. To be sure, however, our models include controls for per capita real county government (non–food stamp) transfers.7

III. Background Literature

The goal of the FSP is to improve nutrition among the low-income population. As such, many studies have exam- ined the impact of the FSP on nutritional availability and intake, food consumption, food expenditures, and food inse-

curity (see Currie, 2003, and Fraker, 1990, for reviews of the literature).

Almost all existing studies of the impact of the FSP use research designs that rely on comparisons of program parti- cipants to nonparticipants at the individual level. This approach is subject to the usual criticisms regarding selec- tion into the program. For example, a number of researchers (Currie, 2003; Currie & Moretti, 2008; Fraker, 1990) have pointed out that if food stamp recipients are healthier, are more motivated, or have better access to health care than other eligible women, then comparisons between partici- pants and nonparticipants could produce positive program estimates even if the true effect is 0. Conversely, if food stamp participants are more disadvantaged than other families, such comparisons may understate the program’s impact. In fact, as Currie (2003) reported, several studies, including Basiotis, Cramer-LeBlanc, and Kennedy (1998) and Butler and Raymond (1996), find that food stamp parti- cipation leads to a reduction in nutritional intake. These unexpected results are almost certainly driven by negative selection in participation.

Many researchers who evaluate the impact of other gov- ernment programs avoid these selection problems by com- paring outcomes across individuals living in states with different levels of benefit generosity or other program parameters. A long literature on the effects of cash assis- tance programs is based on this type of identification strat- egy (Moffitt, 1992; Blank, 2002). Unfortunately, the FSP is a federal program for which there is very little geo- graphic variation (aside from the variation we use in this paper) or variation in eligibility criteria or benefit levels, so prior researchers have had to employ alternative approaches.

Identification issues aside, it is noteworthy that few FSP studies examine the impact on health outcomes. We are aware of two studies. Currie and Cole (1991) examine the impact of the FSP on birth weight using sibling comparisons and instrumental variable methods and find no significant impacts of the FSP. Our work is closer to that of Currie and Moretti (2008), who use the county rollout of FSP in Califor- nia to analyze birth outcomes. They find that FSP introduc- tion was associated with a reduction in birth weight, which was driven particularly by first births among teens and by changes for Los Angeles County. As discussed below, this negative effect is possible if the FSP led to fertility changes or increases in the survival of low-birth-weight fetuses. The timing of FSP assignment in Currie and Moretti (2008) differs from ours in that they consider FSP availability at the begin- ning of pregnancy and its impact on birth weight, whereas we focus on availability toward the end of pregnancy.8

The literature (see the review in Currie, 2009) provides few estimates of the causal impact of income on birth

6 For more detail, see table 1 in Hoynes and Schanzenbach (2009). 7 The Special Supplemental Food Program for Women, Infants and

Children (WIC), available to low-income pregnant women and children up to age 5 in families, was introduced in 1974. Given the timing of WIC implementation relative to FSP, there is little concern that the introduc- tion of WIC biases our estimates of the introduction of FSP, and results limited to pre-1974 are qualitatively similar.

8 Table 3 shows the sensitivity of our impact estimates to the timing of FSP assignment.

390 THE REVIEW OF ECONOMICS AND STATISTICS

weight. Cramer (1995) finds that mothers with more income have higher-birth-weight babies, although income is identi- fied cross-sectionally. Kehrer and Wolin (1979) find evi- dence that the Gary Income Maintenance Experiment may have improved birth weight. However sample sizes are small (N ¼ 404 births), and although positive effects were found for woman as being and high risk for low birth weight (young, smokers, short birth interval), perverse effects were found for woman classified as being of low risk low birth weight. Currie and Cole (1993), using IV and mother-fixed effects estimators, find that AFDC income leads to improvements in birth weight. Baker (2008) uses the 1993 expansion in the EITC, which disproportionately benefited families with two or more children, finding a 7 gram increase in the birth weight of subsequent children. In general, the literature has been plagued by imprecise esti- mates due to small sample sizes as well as a lack of well- identified sources of variation in income. As a result, we argue that our paper provides some of the best evidence to date on the impact of income on birth outcomes.

IV. Food Stamps and Infant Health

The FSP introduction represents an exogenous and siz- able increase in income for the poor. Canonical microeco- nomic theory predicts that in-kind transfers like food stamps will have the same impact on spending as an equiva- lent cash transfer for consumers who are inframarginal. Hoynes and Schanzenbach (2009) use the same FSP rollout identification approach and data from the PSID to examine the impacts of food stamps on food expenditures; they find that recipients of food stamps behave as if the benefits were paid in cash. Therefore, not only can we think of the FSP introduction as a large income transfer, we can think of it as for the most part the equivalent of a cash income transfer.

With this framing, an increase in income could lead to changes in infant health through many channels. We would expect that spending on all normal goods would increase, therefore leading to increases in food consumption regard- less of whether the benefits are paid in cash or in kind. We have little information on how particular subcategories of food demand change with FSP availability: Hoynes and Schanzenbach (2009) are able to measure impacts on total food expenditures, but cannot provide information on the quantity or quality of food consumed (or other goods).

The medical literature on the determinants of birth weight provides a useful structure for thinking about the possible channels for the health effects of the FSP. As Kra- mer (1987a, 1987b) suggested, birth weight is usefully decomposed into that related to the gestation length (prema- turity, or GL) and growth conditional on gestation length (intrauterine growth, or IUG). Of the two, GL is thought to be more difficult to manipulate, though empirically more important than IUG in affecting birth weight in developed countries (Kramer, 1987a, 1987b). Maternal nutrition and cigarette smoking are the two most important determinants

of IUG that are potentially modifiable (Kramer, 1987a, 1987b). Finally, there is evidence that birth weight is gener- ally most responsive to nutritional changes affecting the third trimester of pregnancy.9 Kramer (1987a) writes, ‘‘It is important to analyze additional health measures in addition to birth weight: A final reminder concerns the need for future research to keep sight of the truly important out- comes of infant and child mortality, morbidity, and func- tional performance. After all, birth weight and gestational age are important only insofar as they affect these out- comes’’ (p. 510).

We examine impacts on neonatal mortality because it is commonly linked to the health environment during preg- nancy; it is therefore plausible that FSP transfers may have been a factor. Estimates from Almond, Chay, and Lee (2005) indicate that a 1 pound increase in birth weight causes neonatal mortality to fall by 7 deaths per 1,000 births, or 24%. Postneonatal mortality, by contrast, is viewed as being more determined by postbirth factors.10

This discussion suggests that we would expect FSP to affect birth weight and neonatal mortality but not necessarily gestational length. One obvious channel for food stamp impacts is through improvements in nutrition. The introduc- tion of the FSP transfer increases total family resources and is predicted to increase the quality and quantity of food con- sumed, thereby leading to improvements in infant health. The increased transfer income could also encourage behaviors that could harm infant health, such as smoking or drinking.11

Health improvements may work through other channels as well, for instance, reducing stress (such as financial stress) experienced by the mother, which itself may have a direct impact on birth weight. We explore these issues by separately testing for FSP impacts on length of gestation and birth weight and by exploring the sensitivity of our impact estimates to the timing of FSP assignment by pregnancy trimester.

Overall, we expect that access to the FSP should improve infant health. The same forces that improve infant health, however, could also lead to a change in the composition of births. In particular, if improvements in fetal health lead to fewer fetal deaths, there could be a negative compositional effect on birth weight from the improved survivability of marginal fetuses. This could bias downward the estimated

9 See the literature review of Rush et al. (1980). For example, the cohort exposed to the Dutch famine in the third trimester had lower average birth weight than cohorts exposed earlier in pregnancy (Painter, Rosebooma, & Bleker, 2005).

10 The initial health at birth is generally much better among infants who die in the postneonatal period than among infants dying in the first month of life. For example, while 72% of all neonatal deaths had a low birth weight (below 2,500 grams), only 20% of all postneonatal deaths were low-birth-weight infants (Starfield, 1985). Postneonatal deaths tend to be caused by negative events after birth, most often by infectious diseases and accidents (Grossman & Jacobowitz, 1981). Further, postneonatal deaths may be more responsive to hospital access than neonatal deaths (see Almond, Chay, & Greenstone, 2007).

11 Although recipients cannot purchase cigarettes directly with FSP benefits, the increase in resources to the household may increase cigarette consumption, which would work to reduce birth weight.

391INSIDE THE WAR ON POVERTY

effects of the FSP on birth weight and infant mortality.12 In addition, if FSP introduction leads to increases in fertility for disadvantaged women, this could also lead to negative compositional effect and a subsequent downward bias on the estimates.13 To evaluate such channels, we test for impacts of the FSP on total births (finding no effect).

V. Data

The data for our analysis are combined from several sources. The key treatment or policy variable is the month and year that each county implemented a food stamp pro- gram, which comes from USDA annual reports on county food stamp caseloads (USDA, various years). These adminis- trative FSP data are combined with two microdata sets on births and deaths from the National Center for Health Statis- tics. In some cases, we augment the core microdata with digi- tized print vital statistics documents to extend analysis to the years preceding the beginning of the microdata. These data are merged with other county-level data from several sources.

A. Vital Statistics Natality Data

These data are coded from birth certificates and are avail- able beginning in 1968. Depending on the state-year, these data are either a 100% or 50% sample of births, and there are about 2 million observations per year. Reported birth outcomes include birth weight, gender, plurality, and (in some state-years) gestational length. Data on the month and county of birth permit linkage of natality outcomes to the month the FSP was introduced in a given county. There are also (limited) demographic variables, including age and race of the mother and (in some states and years) mother’s educa- tion and marital status. Online appendix table 1 provides information on the availability of these variables over time.

We use the natality data and collapse the data to county- race-quarter cells covering the years 1968 to 1977. We use quarters (rather than months) to keep the sample size man- ageable. The results are unchanged if we instead use county- race-month cells. We end the sample in 1977, two years after all counties have implemented the FSP and before the pro- gram changes enacted in 1978 led to increases in take-up.

Unfortunately, natality microdata are available only beginning in 1968. By 1968, half of the population lived in counties with on FSP in place. In the interest of examining the full FSP rollout, we obtained annual print vital statistics documents and digitized the available data. With these print documents, we augment the microdata with counts of the total number of births by county and year (not available by race) for 1959 to 1967 and counts of births by birth weight ranges by state, race and year (not available by county) for 1959 to 1967.14

B. Vital Statistics Death Data

These data are coded from death certificates and are available beginning in 1959. The data encompass the uni- verse of death certificates (except in 1972, when they are a 50% sample) and report the age and race of the decedent, the cause of death, and the month and county of death. We collapse the data to county-race-quarter cells covering the years 1959 to 1977.

Our mortality measure is the neonatal mortality rate, defined as deaths in the first 28 days of life per 1,000 live births. We focus on deaths from all causes, as this gives us the most power (further cutting of the county-quarter-race cells by detailed cause of death leads to many very thin cells) and is unaffected by changes in the coding of cause of death (conversion from ICD-7 to ICD-8) in 1968. We have attempted to identify causes of death that could be affected by nutritional deficiencies and also present results for these and other deaths.15 We consider nutritional causes both because the FSP was targeted at those in nutritional risk and widespread concerns about nutritional status among the poor during this period. Online appendix table 2 lists the broad categories for cause of death.

Our main neonatal results use the natality microdata to form the denominator (live births in the same county-race- quarter). This limits the sample to the years 1968 to 1977. In an extension, we use the digitized vital statistics docu- ments and county-year counts of births to construct the denominator for live births and therefore neonatal death rates (for all races) for 1959 to 1977.16

C. County Population Data

The SEER population data are used to construct e