John Gibney for The 2004 Moffatt Prize in Economics

The study finds that food bank patrons are predominantly female, unemployed, and have lower education and household income levels than the general population.  Although Williamsburg VA experiences the same or similar patterns, the Williamsburg area has more African-American and female food bank patrons than both the general Williamsburg population and a nationwide sample of food bank patrons. 

As part of the portrait of food bank users, the study looks at their coping mechanisms to deal with lack of food.  Food bank patrons nationwide and especially in Williamsburg VA not only frequently visit food banks but also borrow money for food or borrow actual food from friends and relatives in their efforts to achieve food security.

This study also reviews the literature on food banks.  While Williamsburg roughly parallels national trends, there are significant differences, particularly in the race and educational levels of the users.

Regression analysis shows that low-income households headed by persons who have completed no more than a high school degree are particularly susceptible to food insecurity.  In addition, these types of households are more likely to use food stamps.  A range of robustness tests support these basic findings. 

 I. Introduction – National and Local Responses to Hunger:

The United States, a ‘food rich’ country by almost all definitions, still experiences hunger and food insecurity in its most vulnerable populations.  In 2001 the United States Department of Agriculture (USDA) found that 11.1% of households (12.1 million) had experienced food insecurity, “the limited or uncertain availability of nutritionally adequate and safe foods,” in the preceding 12 months.  Furthermore, the USDA found that 3.5% of households (3.8 million) experienced conditions leading to hunger, “the uneasy or painful sensation caused by a lack of food.”  Food insecure Americans often utilize food banks, defined by the USDA as “distribution centers that provide groceries and other basic supplies,” in order to achieve food security.[2]

Food banks serve about 12.5 million people monthly.  Utilization rates of food assistance services have increased since 2000, raising issues concerning the adequacy of the country’s response to food insecurity and hunger.  Surprisingly little empirical research however, focuses on the economic and social impacts of the Emergency Food Assistance System.  The Emergency Food Assistance System (EFAS) is the network of food banks and soup kitchens run largely by private organizations but with some local government support.  In contrast, a much more extensive field of research exists on food stamp participation rates, WIC utilization rates, and welfare efficiency versus access studies.[3]  Food banks deserve more attention because, in addition to putting food on dinner tables, food bank handouts serve as a form of budget relief for households.  This free food often helps low-income households offset the rising costs of prescription drugs and housing and it helps these families make up for their relatively low participation rates in food stamps.[4]

Across the United States, although housing costs are high and prescription drug prices are rising, welfare participation rates remain surprisingly low.  These trends signal that low-income families use coping mechanisms, like food banks, in addition to or independent of government sponsored programs, such as food stamps and WIC (Women, Infants, and Children program).  Since the late 1970s low-income households have struggled to find affordable housing.  Only 34% of the households in the bottom fifth of the income distribution receive housing assistance, in the form of Section 8 vouchers or public housing.[5]  These households often can barely afford to cover utilities, taxes, and maintenance costs.  Furthermore, low-income households also face trouble meeting rising prescription drug prices.  Nationwide spending on prescription drugs rose 16% from 2000 to 2001, and has more than tripled since 1990.[6]  Despite these trends, low-income households still exhibit poor food stamp participation rates; the USDA found that only 57% of those eligible for food stamps participated in 2002.[7]

These two trends—a higher cost for basic necessities and a lower participation in public assistance—indicate that disadvantaged people must have additional, informal resources to cope with hunger.  A primary coping mechanism for many of these households is to get food at food banks.

Personal volunteering and observations at Williamsburg food banks led the author to choose the area as a case study in order to test hypotheses on food bank usage on a local level.  According to local food bank operators, causes of increased food bank usage include the area’s high housing costs, seasonality of local employment, the low wages of many retail and service jobs, and welfare reform.[8]  Local demographic research confirms the food bank operator’s hypotheses, and consequently makes Williamsburg food bank patrons a unique group for comparison with national food bank users.  The Williamsburg data analysis in this study builds upon the area’s first report of the local emergency food assistance system in 2003, performed by the William and Mary Sharpe Program, in partnership with the Greater Williamsburg United Way.[9] 

Williamsburg’s high housing costs, rising prescription drug prices, and low food stamp participation rates push many local households to use food banks.  According to 2000 Census data, Williamsburg possesses only 0.21 low-income housing units available for every low-income renter household, compared to the state average of 0.46.[10]  From 1992 to 2002 the average number of prescriptions per person increased from 7.3 to 11.6.  This paralleled an increase in the total number or prescriptions from 1.9 billion to 3.3 billion over the same time period, and during this time prices have risen dramatically.[11]  Despite these handicaps to indigent families, food stamp usage for James City County, Williamsburg City, and York County has declined steadily since 1993, falling from a high of 4526 users to 2502 users in 2000.[12]  

Williamsburg possesses a visible food bank population, with the local United Way referring about 350 local residents to food banks a month in 2003.[13]  The four major food banks in Williamsburg VA, Operation Hope, Helping Hands, FISH, and Grove Christian Outreach, serve on average 300-400 families a month.  The largest of these food banks, Helping Hands, donated over 425,000 pounds of food alone in 2003.[14]  With over ten local food banks, Williamsburg possesses a substantial local Emergency Food Assistance System (EFAS), attracting not only local patrons but also people from neighboring counties and cities including Newport News, Gloucester, and Charles City County.[15]

Williamsburg’s food bank patrons consist of a higher percentage of African-Americans than the national experience.  This study hypothesizes that this difference stems from local labor market inequality.  African-American families earn disproportionately less than White families in Williamsburg; with 44% of African American families earning less than $25,000 a year compared to 12% of White families, and in James City County, the figures are 32% to 8% respectively.[16]

In light of these statistics and themes, several demographic and policy questions arise:

·        Who utilizes food banks?  Do national trends carry over to Williamsburg VA?

·        How do food bank patrons make ends meet? 

·        How can the community increase food security? 

In an attempt to answer these questions this study provides a snapshot of the scope and breadth of the challenges faced by many families both nationwide and in the Williamsburg area.

More specifically, this paper uses data from the Current Population Survey (CPS), census statistics, and survey data on Williamsburg food bank patrons.  These data sets provided the national data used in this study.  The CPS survey on food security, collected by the U.S. Bureau of the Census in partnership with the U.S. Bureau of Labor Statistics, originated from the Community Childhood Hunger Identification Project Survey (CCHIP).[17]  Each December, the CPS presents information on food security, referred to as the Federal Food Security Supplement

The local data on Williamsburg food bank patrons comes from a local provider and client survey modeled after the CPS Food Security Supplement.  The local provider survey measures the size of Williamsburg food banks, their coverage range and rates, the percentage of food banks that are faith based, and the various food banks’ abilities to meet increased needs.  The provider research also highlights both the pounds and dollar amounts of food salvaged from local stores and the pounds and dollar amounts of food distributed to clients.  With the provider data the study includes information collected on local food stamp program participation numbers, local TANF (Temporary Assistance to Needy Families) participation rates, and local poverty rates.[18]

The client survey of Williamsburg food bank patrons utilized the Federal Food Security Supplement and other demographic questions in order to survey 101 local area food bank users.  By interviewing a sample of food bank patrons from three of Williamsburg food banks, the client survey strove to fulfill four main functions:

1.      Describe Williamsburg food bank patrons

2.      Determine the food-security status (as outlined by the USDA) of local food bank patrons

3.      Understand the coping mechanisms that local food bank patrons use to deal with economic insecurity

The study results are summarized as follows:

·        Foods bank patrons, both nationwide and in Williamsburg, are predominantly female, unemployed, and possess lower education and household income levels than the general population.

·        Compared to the U.S. food bank patron population, Williamsburg possesses higher rates of African-American and female patrons.  Williamsburg’s patrons tend to have higher levels of educational attainment.  Williamsburg food bank patrons display a higher tendency to use food banks and to borrow money or food from friends or relatives as coping mechanisms than the general food bank population.

·        As educational attainment and income rise, food security increases in both the local and national sample.  The increases are steeper for Williamsburg respondents as those with household incomes in the $10,000-$20,000 range are 13% more food secure than those households with less than $10,000 income.

·        Williamsburg food bank patrons possess lower food stamp participation rates than the national and state averages.

II. Literature Review:

In 1995 the CPS implemented the Federal Food Security Supplement survey.  Since its inception, numerous studies have both tested the Food Security Supplement survey for validity and used the survey as a tool to analyze policy concerning hunger and food insecurity.  As discussed at the Food Research and Action Committee Conference (2004), concerns with the Food Security Supplement arise over the method of using an 18 question survey to assign households to food security categories.[19]  Some researchers feel that the CPS underreports hunger because households tend to shy away from questions concerning child food security.  In an effort to investigate these hypotheses, numerous studies have tested the validity of the CPS food security survey in measuring hunger (Hamilton 1997, Bjoras 2001).[20]

Some of these studies use alternative criteria to predetermine the food security status of households and then compare the results to actual classification by the CPS supplement.  Other studies examine the consistency of response patterns for different demographic groups over time.  Overall these studies show a strong correlation between the food security status determined by the CPS and the actual food condition in the household as measured by weekly household food expenditures, actual nutrient intake, and participation in food assistance programs.[21]

Relatively few studies, however, focus on this project’s main concern, the coping mechanisms that food bank patrons use in order to achieve food security.  Most studies on food bank patrons remain descriptive in nature.  These studies, however, do provide valuable statistics with which to characterize the food bank population.

Nord et al (2000) analyzed CPS food security statistics and computed food security and food spending for households that received food assistance.[22]  His findings show that households that use food assistance services experience hunger about twice as much as nonparticipating households in the same income ranges and with similar household composition.  Food insecurity and hunger rates were even higher for patrons of emergency kitchens.  In addition to research on food bank patrons, Nord analyzed the whole CPS food security sample and found that African-American households, Hispanic households, households in poverty, and single-parent households exhibit lower food security rates than the general population.

Reschovsky (1991) performed one of the first empirical studies on Emergency Food Assistance Systems.[23]  Reschovsky estimated food supply equations for patrons in upstate New York food pantries.  The study models food supply as interdependent between recipient demand (food bank users) and donor demand (beneficial giving).  Areas with high poverty rates, especially rural places, can experience static or lowered food supply levels in the face of growing demand.  Conversely, relatively affluent areas with minimal poverty rates can have fluctuating supply dependent on food bank user’s demands.

This author’s study relies heavily on Biggerstaff, Morris, and Casebolt (2002).  Biggerstaff et al. analyzed 1500 interviews of food bank users in Virginia from 1997 to 1998.  The researchers used a multistage cluster sampling method to select a random sample of Virginia food bank users.  Like this study, Biggerstaff et al. administered the surveys in November and December.  The Biggerstaff survey asked questions regarding food stamp participation, TANF participation, Medicaid enrollment, and “recent hardships”--setbacks experienced by the household that led them to become economically insecure.  Demographic analysis by the authors found that “unemployment as a result of disability or recent job loss and low wages” were major contributing factors to relying on food pantries.  The authors performed logistic regression on food stamp participation and found that being a single parent, possessing lower education, and being female made respondents more likely to use food stamps.

America’s Second Harvest and the USDA have both published detailed reports on food pantries and their users.[24]  America’s Second Harvest serves as the national headquarters for approximately 80% of all U.S. food banks.  The USDA runs the United States’ food assistance programs, helps coordinate various conferences on the emergency food system, and offers free promotional material on food stamps, WIC (Women and Infants with Children), and the School Lunch/Breakfast program.[25]  In 2001 both organizations published large nationwide reports on clientele of food banks.[26]  Second Harvest surveyed 11,329 households compared to the USDA which surveyed 2,397.  The reports contain similar results and serve as the generally accepted background research on food bank patrons and their characteristics.

According to the USDA report, 29% of pantry users are single-person households while 45% of pantry client households include children under the age of 18.  Roughly 70% of pantry client households receive food assistance in metropolitan areas.  America’s Second Harvest returns similar results except for a slightly higher percentage of households receiving food assistance in metropolitan areas, 80%.

The residential status of pantry households shows their vulnerability to high housing costs.  The USDA reports that 62% of food bank patrons rent their residence and about 8% of the patrons are homeless.  Furthermore, 50% of food bank patrons live in either an apartment or a mobile home.[27]  Second Harvest asked food bank patrons about spending choices involving food and other needs.  The survey showed that 45% of recipients had to choose between buying food and paying for utilities or heat, 36% of recipients had to choose between food and paying for housing, and 30% choose between buying food and paying for medical care.

The USDA and Second Harvest find similar trends in low-income levels for food bank patron households.  The USDA found that 93% of food bank patrons have incomes at or below 130% of the poverty level, and therefore meet the gross income requirement for the Food Stamp program.  Both studies reported similar household income averages: $10,776 per year for the USDA report and $10,960 from Second Harvest.  Both reports show low employment rates for food bank patrons: Second Harvest found that only 24.7% of food bank patrons are currently employed compared to the USDA’s employment rate of 26.4%.

Food pantry households usually possess the durable goods needed to process meals.  Only 3.4% lacked access to a stove, oven or microwave and only 4.5% of households lacked access to a refrigerator.  Transportation remains a bigger issue as 48.5% of respondents lacked access to a working car, truck, or motorcycle.[28]

Both the USDA and Second Harvest report difficulty in calculating the turnover rates of the food bank population.  The USDA reports a monthly turnover rate of between 4 – 8%, which translates into annual number of pantry clients in the 18.0 to 23.5 million people range.  Second Harvest reports a higher number of around 21.1 to 26.1 million people a year.

Another difference emerges between the reports’ food security statistics.  Both surveys implement the six-item short form of the 18-item standard core questions of household food security.  Second Harvest, however, found that 28.6% of food bank households were food secure compared to 20.8% for the USDA survey.[29]  The USDA also found significantly more food bank patron households are hungry.  The USDA classifies 41.5% of households as food insecure with hunger, compared to 36.0% for Second Harvest.  Researchers suggest that different data collection methods, statistical sampling error, and seasonal differences spur the difference in figures.

While not focusing directly on food bank patrons, Kathryn Edin and Laura Lein (1997) show that neither welfare nor low-wage work provide enough income to cover the basic needs of single mothers and their children.[30]  Edin and Lein collected data on 214 single mothers on public assistance and compared those mother’s situations to 165 of their working counterparts.[31]  Single mothers used creative solutions to make ends meet including utilizing food banks, taking under-the-table handouts from relatives and friends, visiting various charities, shopping at thrift stores, taking work not subject to formal accounting procedures, buying stolen goods, and providing sexual favors.  The mothers at low wage work, less than $8 per hour, experienced more instability than those on welfare due to costs associated with work including childcare, transportation, health care, and irregular working hours.  Edin and Lein highlight the tradeoffs that members of the low-income population must make in order to make ends meet.

While the previously mentioned reports give substantial information on the characteristics of food bank patrons nationwide, relatively little research exists on local Williamsburg food bank patrons.  The College of William and Mary’s Sharpe Program report, Food Banks in the Williamsburg Area 2002-2003, represents the first attempt to research and document all the local food banks.  The report surveyed 121 local food bank patrons and analyzed their demographic characteristics, the items they received at food banks, and their food security.  The Sharpe program found that 91% of all recipients possessed the durables needed to convert food bank handouts into meals including ovens, stoves, refrigerators, and microwaves.  Fifty percent of local pantry clients have children, and only 12% of local clients work full time.  Females disproportionately utilize local area food banks, comprising 74% of respondents.  Thirty percent of respondents experienced gaps between when they received food and when they needed food.  Forty-eight percent of clients would like to see more financial aid services in the community, and 38% would like to see more shelter options in the community.

Local government agencies--Human Services in Williamsburg and Social Services in James City County--keep records of food banks and sometimes refer low-income households to them.  The social services organizations also give caseload statistics of food stamps in their annual reports, but these reports do not emphasize the community food bank system.[32] 

Local food banks gain some publicity from the Virginia Gazette, Williamsburg’s local newspaper, which sometimes features food drives for local food banks and regularly lists Helping Hands’ operating hours in its daily events section.  In 2003, the United Way recognized Thumper Newman with the Cruickshank Award for his contributions to the community with Helping Hands Food Bank and A Gift From Ben Food Delivery.[33]  This prestigious recognition has helped to raise community awareness of food banks.

The previously mentioned reports give useful information on the characteristics of food bank patrons.  These reports, however, often do not focus on the coping mechanisms used by food bank patrons to achieve food security.  The reports remain descriptive in nature and generally do not analyze the relationships between certain demographic characteristics and household choices to use food stamps, food banks, and other coping mechanisms to achieve food security.[34]  The Second Harvest Survey begins to tackle some of these issues, by analyzing the tradeoffs that food bank patrons must make. 

The literature notes the various limitations that arise in surveying the food bank population.  As noted by Nord et al (2000) in Guide to Measuring Household Food Security, one cannot generalize findings for the food bank population to any larger group, even the group of all low-income members of the same community.[35]  Nord shows that one difficulty in studying food bank patrons lies in their “essentially self-selected nature and indeterminate, often fluid, boundaries of the group.”  Despite these concerns, studies on the food bank population of a community are valuable in that they formally and authoritatively document the existence of food insecurity and hunger in a needy part of the population.  Furthermore, food bank studies shed light on the effects of welfare reform and recessions on the vulnerable population of food bank patrons.  Study of the Williamsburg food bank population and their coping mechanisms can help develop an understanding of the makeup of the local economy.

III. The Williamsburg Area Economy and Its Impact on Food Security:

This study selects the Greater Williamsburg area as a case study due to its unique economic and social makeup. 

In 2001, Williamsburg City possessed a population of 12,102 and James City County had 50,249 people.  The intertwining of districting lines and the sharing of social and governmental services between the city and county make it impossible to study either Williamsburg or James City County alone; therefore, the study takes aggregate census numbers and computes statistics for Williamsburg/James City County.[36] 

Williamsburg’s labor market conditions, high housing costs, rising prescription drug prices and low welfare participation rates drive many low-income families to utilize local food banks.[37]  Households use Williamsburg food banks in an attempt to avoid hunger and in order to enhance their earnings by having their food expenditures subsidized. 

Williamsburg’s service-based economy, with its tourism focus, leaves a substantial portion of the workforce with low earnings.  Furthermore, the locality’s industry and occupational distribution contributes to a substantial difference in earnings between males and females and between races.  Lower paying service jobs in the arts, recreation, accommodation and food services comprise 23% of Williamsburg employment and 16% for James City County.[38]  Little advancement often exists for these service workers as the hotel and food industry regularly imports in competing workers from foreign countries to meet the tighter job market in the summer months.[39] 

Median earnings figures from the 2000 census show that women earn disproportionately less than men in Williamsburg, with women earning $28,822 annually compared to $30,738 for men.  A much larger divide exists in James City County with women earning $29,011 and men $46,539.[40]  Gender inequality in earnings most likely comes from industry differences rather than education level differences.  Women and men in Williamsburg and James City County possess relatively equal education rates, but women are disproportionately employed in education, health, social services, recreation, accommodation, and food services. 

African-American families earn disproportionately less than White families in both James City County and Williamsburg City.  Only 5% of African-American families earn above $100,000 in James City County compared to 29% of White families.  The pattern holds true for Williamsburg with 0% of African-American families earning over $100,000 compared to 28% of White families.  In a set of combined statistics, encompassing both Williamsburg and James City County, African-Americans possess an 18.43% poverty rate compared to a 6.33% poverty rate for Whites.  These differences in income most likely come from different education levels for African-Americans and Whites.  In James City County 61% of Whites hold a bachelor’s degree compared to 11% of African-Americans.  Williamsburg follows the pattern with 39% of Whites holding bachelor degrees compared to 10% of African-Americans.

In a focus group conducted at a local Williamsburg food bank, Operation Hope, 75% of participants noted difficulties in paying for housing, prescription drug prices, and difficulties in enrolling in welfare.  For these patrons the monthly challenges to pay rent, electricity, childcare, and prescription drugs often leave them one paycheck away from financial ruin.  Many of these households turn to food banks in an effort to achieve food security and provide budget relief.  Participants voiced similar concerns at round table talks at the Williamsburg Virginia Organizing Project meetings, citing high housing costs and barriers to welfare use as reasons to utilize food banks. 

Williamsburg has relatively high housing costs, with 53.2% of owner-occupied houses valued above $200,000 in Williamsburg and 38.9% of houses in James City County.  Expensive housing makes it difficult for low-income workers in the service industries to afford housing in the area.  The majority of these workers rent apartments in the area or live elsewhere.  These workers often spend the largest part of their income on rent with 35.12% of Williamsburg renters paying more than 35% of their income on gross rent.  In contrast, in Virginia as a whole 26.3% of households pay 35% or more of their household income on rent.[41] 

In 2003 the Williamsburg Housing and Redevelopment Authority (WHRA) possessed 104 public housing units in the locality and had very few vacancies.  With over two years worth of applications for public housing sitting on their shelves, WHRA managers see no immediate way to meet the demand for low-income housing.[42]  James City County does not have public housing; rather, it participates in the Section 8 Housing Vouchers program.  The James City County Office of Housing and Community Development stopped accepting applications for the vouchers several years ago, citing an already impossibly long waiting list (in 2003 the county administered 118 vouchers).  A grant received in February 2004 will provide $2.9 million dollars for 28 more affordable homes, but housing experts still predict an excess demand for low-income housing.[43]  These recent low-income housing gains have triggered protests by James City County residents who fear their property values will fall and crime will rise due to the construction of low-income housing.  These residents cite the problems with Burnt Ordinary, a low-income apartment and housing complex, in the western Williamsburg that suffers higher than average crime rates.[44]  Proponents of low-income housing, like Ripley Heatwole Company which has built successful low-income housing in Hampton VA, emphasize the importance of strict applicant screening and the mixing of affordable housing with higher priced housing.

In addition to the increased demand for low-income housing, results from the 2004 Sharpe Program Study, Homelessness in Williamsburg, show a homeless population in the Williamsburg, numbering between 30-100 people.  Many homeless individuals work, but most earn below subsistence wages.  The census cannot count these individuals, due to their lack of an address, which might bias local poverty statistics upwards.  Eighty percent of the heads of local assistance agencies cite high housing costs as a major contributing factor to hunger and homelessness in Williamsburg. 

Many clients of food banks frequently cite rising prescription drug prices as a reason they seek food assistance.  While it is hard to come by specific Williamsburg data on these phenomena, local clinics for low-income families, like Olde Towne Medical Center, claim to have higher utilization rates than years past.  A survey done by America’s Second Harvest found that 30% of food bank clients report “having to choose between paying for food and paying for medicine or medical care.”  The same survey finds that 85% of food bank patrons lack any private health insurance and 20% of the food bank population lacks any health insurance at all.  A lack of health insurance makes many low-income households vulnerable to rising prescription drug prices.  Nationwide prescription drug spending has more than tripled since 1990 reaching $140.6 billion in 2001.  Currently, prescription drugs comprise 11% of personal health care spending and this figure is projected to rise further.[45] 

Low welfare participation rates, especially in food stamps, should drive low-income families to food banks in an effort to supplement their budget.  Food stamp usage for James City County, Williamsburg City, and York County has declined steadily since 1993, falling from a high of 4526 users to 2502 users in 2000.[46]  While some of the fall may be attributed to a rising economy, research by the Food Research Action Committee finds that the 1996 Welfare Reform Laws significantly reduced food stamp participation.  Welfare reform imposed tougher eligibility restrictions including new earnings limits, immigrant status restrictions, and required work programs, which dramatically lowered FSP participation rates nationwide.  Williamsburg mirrors this downward national trend in FSP participation.

Recently, the downward spiral of FSP participation reversed itself and participation rates grew.  This growth coincided with the 2001 recession and the subsequent “jobless” recovery.  James City County experienced a similar trend in rising FSP caseloads, going from 666 in 2002 to 696 in 2003 for James City County.  This trend shows the interdependency between food stamp usage and the business cycle.  The 4.5% increase in local FSP caseloads parallels an even larger boom in Williamsburg food bank usage.  In response to increased demand, the major food banks in Williamsburg have begun salvaging food from more grocery stores and purchasing food from the Virginia Peninsula Food bank at discounted prices.[47]  All the local food banks report increased usage from 2001-2003.  Even traditionally smaller local food banks, such as St. Olaf’s in the western portion of Williamsburg, have experienced growing consumer demand.  St. Olaf’s rose from around 20 households a month before 2000 to a current monthly average of 100 households a month.  The increased usage of food banks is not just a local phenomenon: America’s Second Harvest reported that 86% of its affiliated food banks experienced an increase in requests for food assistance from 2001-2002.  Similarly New York City’s food pantries reported feeding 45% more people in 2002 than in 2000.[48]

Some of Williamsburg’s increased food bank usage likely comes from non-local residents.  Food banks located in the eastern section of Williamsburg, closer to Newport News VA, reported an increased incidence of non-Williamsburg residents utilizing their services.  Grove Christian Outreach Food Bank has experienced such high non-local demand that they now must refer individuals from Newport News to food banks in their home town in an effort to keep the bank’s participation numbers stable.[49]  Many out-of-town food patrons cite Williamsburg’s food banks’ efficient operations and friendly service as reasons to come into the community for food; in addition, many non-local people work in the greater Williamsburg area and find it convenient to stop at a food bank on their way home.[50]

The heavy utilization of Williamsburg food banks raises the question: what other types of food assistance, besides food banks, do Williamsburg’s low-income people use in order to achieve food security?  Nationally the USDA estimates that food pantries serve 9% of low-income Americans, while food stamps provide assistance to 47% of low-income Americans.[51]  It remains extremely difficult to determine the percentage of the Williamsburg population using food banks, because food banks use different documentation methods to record their patrons and because individuals patronize more than one food bank.  Nevertheless, information from the Williamsburg United Way HELP referral line indicates that food banks are disproportionately used to help low-income individuals compared to welfare and other forms of assistance.  The HELP referral line, a call in line where people can ask for assistance and undergo screening for possible services, provides the area’s most reliable statistic on food bank referrals.  In 2003 the HELP line received on average 369.8 requests per month for food assistance, with the winter and early summer months experiencing the highest demand.  A safe assumption is that at least 300 households use Williamsburg’s food banks.  This would indicate that over a third of the assistance given to low-income individuals in Williamsburg comes from food banks, which stands in stark contrast to the 9% utilization rate nationwide.

IV. Theory

This study bases its model of food security on the theory of household production cited in Rose et al (1998).[52]  Households combine store-bought foods, food bank handouts, and time spent shopping and preparing meals with the use of durable goods (e.g. stoves, refrigerators, and microwaves) and human capital (e.g. nutritional knowledge and cooking skills) in order to create meals. 

Meals maximize the household’s food utility, subject to the household’s home production function and the household’s constraints on income and time.  The production function represents the household’s ability to make nutrients, which depends on many of the following factors including but not limited to the opportunity cost of shopping and waiting for food bank handouts, capital goods, and demographic characteristics of the household.  The household constraints on income and time define the factors that limit food production.  These factors include prices of food, prices of other goods, wages, non-labor income, and the total time available to household members.  In equation form:

In the utility equation C represents household taste components, N represents nutrients found in meals, X denotes other goods, and l denotes leisure.  In the production function F denotes store-bought goods, Z denotes food bank handouts, denotes labor time spent waiting for handouts and shopping for food and preparing meals, K denotes capital goods (which includes human capital), and D denotes the demographic characteristics of the household.  Demographic and human capital variables include but are not limited to education, gender, race, welfare status, household structure and income.  The household constraints on income and time model gives price of food as , prices of other goods as PX, non-labor income as V, the wage rate as w, T as the total time available to household members, and l once again as leisure. 

These equations remain appropriate for the Williamsburg case study, because the majority of Williamsburg households possess the characteristics modeled in the equations.  For instance, 91% of Williamsburg food bank recipients possess the durable goods necessary to prepare meals including refrigerators, ovens, stoves, and microwaves.[53]

The amount of nutrients in consumed meals must pass some minimal threshold in order to categorize the household as food secure.  The USDA sets a societal threshold for nutrients in its Thrifty Food Plan.  One can model food security as a binary variable where food security, denoted by I_h  for the hth household, can take on a yes or no value (1,0).  In equation form:

V. Econometric Models:

This study models probits, a form of regression analysis, for each dependent variable: food security, food stamp usage, and various coping mechanisms.  Further, for each dependent variable the study estimates the models on the U.S. sample, the local Williamsburg sample, and on a pooled sample containing both the local and U.S. sample.  Probit regression analysis accounts for the binary nature of the dependent variables.[54] 

The study utilizes two probits categorized as primary and secondary regressions (A and B in the tables respectively).  Primary regressions use household structure (single parent, single alone), race, and gender as explanatory variables.  Secondary regressions include the explanatory variables from the primary regression and add in the additional explanatory (independent) variables of education, age, and household income.  The study utilizes these explanatory variables because they play a role in the household production function cited in Equation 2, and because previous research shows their importance in determining food security.[55]  The regression analysis excludes the 18-35 age group, over $20,000 income group, employment status, and males in order to correct for multicollinearity.

The Primary and Secondary Probits in equation form are as follows:

where  represents the cumulative normal distribution and represents the stochastic disturbance term in both regressions.

Initial regression analysis focuses on food security.  The study estimates the primary and secondary probit models on whether the respondent possesses food security, defined by Equation 4 in the theory section.  The food security models in equation form are as follows:

In an effort to further understand more about how food bank patrons achieve food security, the study estimates regressions on various other coping mechanisms.  These models use the same primary and secondary probit models and the equations parallel those in equation 5 except for a different dependent variable.  The various coping mechanisms analyzed include food stamp participation, getting food or borrowing money for food from relatives or friends, and food bank usage intensity (using food banks more than only one or two months a year).[56]

Respondent information on food stamp participation status, borrowing money or food, and on heavy utilization of food banks comes from questions asked in the Local Williamsburg Survey 2003 and the National Current Population Survey.  Responses to these questions possess only two possible outcomes: yes or no (0 or 1), making them binary dependent variables and therefore precipitating the use of probits.

VI. Data:

The Current Population Survey (CPS), a monthly survey of about 50,000 households, serves as the main source for estimates on employment and unemployment in the United States.  For over fifty years, the Census Bureau has conducted the survey, on behalf of the Bureau of Labor Statistics.[57]  The survey uses a scientifically selected sample in order to represent the civilian non-institutional population.  Published data shows information on employment status, earnings, hours of work, and occupational choice for each member of the household aged 15 years and older. 

Every December the CPS adds additional questions concerning food security to their survey.  The Food Security Supplement Data, collected by the Census Bureau for the United States Department of Agriculture (USDA), consists of answers by household respondents to questions concerning food expenditures, behaviors related to food security, and the use of food assistance programs.  The major sections of the federal food security survey are as follows (CPS code names follow):

·        Food Spending and Program Participation[58]

·        Food Sufficiency and Food Security[59]

·        Coping Strategies: Ways to Avoid or Ameliorate Food Insecurity[60]

·        Minimum Food Spending Needed[61]

The classification system for food security comes from the Life Sciences Research Office (LSRO) of the Federation of American Societies for Experimental Biology.  The LSRO uses the following terminology when discussing food security:

Food security – “Access by all people at all times to enough food for an active, healthy life.  Food security includes at a minimum:

(1) the ready availability of nutritionally adequate and safe foods

(2) an assured ability to acquire acceptable foods in socially acceptable ways (e.g., without resorting to emergency food supplies, scavenging, stealing, or other coping strategies).”

Food insecurity – “Limited or uncertain availability of nutritionally adequate and safe foods or limited or uncertain ability to acquire acceptable foods in socially acceptable ways.”

Hunger – “The uneasy or painful sensation caused by a lack of food.  The recurrent and involuntary lack of access to food.  Hunger may produce malnutrition over time.”[62]

The USDA uses 18 specific questions to determine the actual food security status for households.  These core questions, displayed in Appendix 6, ask households whether they have experienced a certain behavior with regard to food security in the last 12 months.  The questions start with less severe statements such as whether the household “worried whether our food would run out before we got money to buy more.”  The questions become more indicative of severe hunger as they progress and end with questions concerning child food security such as: “Did your child (or any of the children) ever not eat for a whole day because there wasn’t enough money for food.”[63]  Households without children, consequently, only answer 10 of the 18 questions.

The higher the number of affirmative responses the interviewee gives, the greater the food insecurity of the household.  As the level of food insecurity rises, hunger rises because the USDA models food security as a continuous measure.  Households are classified as “food secure,” “food insecure without hunger,” and “food insecure with hunger.”  Households without children are classified as “food secure” if they give fewer than three affirmative responses, as “food insecure without hunger” if they give between three and five affirmative responses, and as “food insecure with hunger” if they give six or more affirmative responses.  Households with children are classified as “food secure” if they give fewer than three affirmative responses, as “food insecure without hunger” if they give between three and seven affirmative responses, and as “food insecure with hunger” if they give eight or more affirmative responses.  All food security and summary statistics are weighted using the appropriate household and individual weights from the CPS.

The Williamsburg area cross-sectional data comes from surveys of clients from the three largest local food banks (Operation Hope, Grove Christian Outreach, Helping Hands).  The local Williamsburg survey, shown in Appendix 5, contains demographic and labor force questions modeled after the general CPS survey and the 18 question Federal Food Security Supplement.  The local survey also incorporates food coping mechanisms questions from the CPS.  Design and implementation ideas for the survey come from the USDA Community Food Security Assessment Toolkit, which provides standardized measurement tools for assessing various aspects of community food security.[64]  Each participant received an ID number, which helped prevent repeat surveying.[65]

The three food banks surveyed represent an accurate cross section of the food pantry population in Williamsburg.[66]  These food banks cater to different demographic groups, are located in different parts of the Greater Williamsburg area, possess different eligibility restrictions, and have different operating hours.  For example, Helping Hands possesses the most accessible operating hours; open Monday, Wednesday, and Friday from 4-6pm in the James City Community Church.  Operation Hope and Grove Christian Outreach operate on Wednesdays from 2pm-6pm and 11am-1pm respectively.

These differences allowed the researchers access to a greater cross section of the food pantry population than focusing on one food bank.  All three of the banks did, however, possess queues, or lines, which allow a researcher ample time to conduct an interview. 

Helping Hands lies closest to the physical center of the Greater Williamsburg area, Operation Hope is in the area near Busch Gardens, and Grove Christian Outreach is located farthest east of the food banks, almost in Newport News.[67]  Helping Hands possesses the least stringent requirements in that they have no forms to fill out (save for signing ones name in a ledger) in order to use the food bank.  Operation Hope has slightly more stringent guidelines requiring a form for each family.  The forms record each family’s name and reason for using the food bank, but the pantry does not demand pay stubs or a referral from the United Way of Greater Williamsburg.  Grove Christian Outreach requires a name, address, and proof of income in order to receive food from their USDA handouts line, but they only require a name and address in order to receive food from their regular pantry.

Different demographic groups utilize specific food banks, but a considerable amount of overlap of food bank usage occurs in the area.  Out of 37 surveys at Helping Hands, 58.3% of respondents were African-American, 38.9% White, and 2.7% Asian.  Operation Hope caters to a predominantly African-American clientele from neighborhoods near Busch Gardens.  In 27 surveys at Operation Hope 74.1% of respondents were African-American, 22.2% White, and 3.7% Hispanic.  Grove Christian Outreach reaches the White population in Grove and a substantial African-American population from throughout the county.  Out of 36 surveys at Grove Christian Outreach, 41.7% of respondents were African-American, 50% were White, 5.6% Hispanic, and 2.8% Asian. 

The majority of Williamsburg food banks, including those not surveyed, cite a disproportionately large African-American presence.  The farther western part of the Greater Williamsburg area lacks a large food bank in which to survey patrons.  Currently St. Olaf’s and the Salvation Army meet the needs of food bank users in the western part of the county, usually coming from Burnt Ordinary low-income housing and representing all races.

The local data validity comes from sampling approach, time frame of surveying, use of the CPS questionnaire, and the extensive training of researchers.  Sampling techniques included surveying every other person, surveying every third person, and surveying everyone in the line.  The author administered 73 of the local surveys and volunteers administered an additional 28 surveys.  Refusals of the local survey totaled 8 making for a 7.4% refusal rate.  Surveying took place between November 17 and December 14, 2003.  This time frame may have depressed local food insecurity numbers, due to generous food donations around the winter holidays, but surveying during the winter parallels the Current Population Survey Food Security Survey, conducted in December.[68]  In order to ensure accurate local data the volunteers underwent extensive training, modeled after Cornell University’s “Training Program for Researchers in the Use of Human Subjects.”[69]  Following the Cornell model, each survey client was given autonomy, beneficence, and fair justice.

VII. Demographic Analysis:

Table 1 reports demographic information from the Williamsburg food bank patrons and compares it to the general population of the Greater Williamsburg area (including James City County and Williamsburg) and Virginia.  The table shows that Williamsburg food banks possess significantly higher rates of female usage and African-Americans than their proportions in the general populations of Williamsburg and Virginia.  The high African-American usage of food banks should draw special concern because of Williamsburg’s relatively low African-American population.  Further, Williamsburg food bank users have lower education rates than the population at large, but they have slightly larger household sizes.  These data match trends in the literature on the food bank population.[70]  Williamsburg food bank data also shows that food pantry users have lower incomes and employment rates than the general population at large.  This logical trend finds support in the literature as Second Harvest reported that 47.9% of its food bank household population earned less than $10,000 a year in 2001.[71] 

As noted in table 1, this study uses 2003 data for the Williamsburg food bank patrons but can only compare this data to the 2000 Census figures for the general population of Williamsburg and Virginia.  Income and employment status for Williamsburg and Virginia may be inflated due to the economic expansion in 1999 (the year of the data collection).  While this presents a challenge in interpreting the evidence in the tables, the general trends in the table should stay the same over time.  Those trends point to a lower earning capability for food bank families and lower employment rates for food bank families.

Another difference between data are the tables which compare the local Williamsburg sample (2003) to the U.S., South, and South Atlantic data, which come from the 2001 CPS Food Security Supplement.[72]  The differences between 2001 and 2003 become clear by analyzing the differences in poverty rates and food insecurity rates during this time. 

The most recent poverty statistics show that conditions worsened from 2001 to 2002 with the national poverty rate rising from 11.7% in 2001 to 12.1% in 2002.  African-Americans also experienced higher poverty rates over this time period, 23.9% to 24.1% poverty rate in 2002 (depending on race definition) compared to 22.7% in 2001.  Family poverty rates grew rising from 9.2% (6.8 million families) to 9.6% (7.2 million families).  Food insecurity also rose from 10.7% in 2001 to 11.1% in 2002.  The prevalence of food insecurity with hunger rose from 3.3% to 3.5% over the same time period.  In 2002, 19.3% of food bank households obtained emergency food from a food pantry, church, or food bank during the preceding 12 months compared to an 18.6% utilization rate in 2001.[73] 

The trends in food security and poverty point to harder economic times for food bank patrons and the general low-income population in 2002 as compared to 2001.  The trend towards difficult times most likely continued throughout 2003 as the “jobless” recovery, positive growth in the economy but not in the labor market, lengthened.[74]  The weak labor market throughout 2003 continued to impose hardship on millions of American workers, their families, and their communities.  As people remained unemployed in 2003, many turned to food stamps and other forms of welfare.  Food stamp participation statistics show a 2.7 million-person increase from November 2002 to 2003 and a rise of about 4.5 million persons compared to November 2001.[75] 

These statistics and findings show that 2003 food bank patrons experienced tougher economic conditions than the 2001 sample.  The economically hard times most likely transferred to Williamsburg’s food bank population, because many of Williamsburg’s users work in seasonal and service jobs, which are highly susceptible to recessions.  The Williamsburg local food bank statistics, collected in 2003, represent households in riskier economic times than those in the regional samples of 2001.  Based on these assumptions, nationwide food security statistics for 2003 will align more closely with Williamsburg’s food security rates. 

Table 2 presents the same information on Williamsburg’s food bank patrons from Table 1 but compares this information to other food bank populations grouped by census region, South Atlantic, South, and U.S.  The South Atlantic sample includes Delaware, District of Columbia, Florida, Georgia, Maryland, North Carolina, South Carolina, Virginia, and West Virginia.  The South sample includes all of the states in the South Atlantic region plus Alabama, Arkansas, Kentucky, Louisiana, Mississippi, Oklahoma, Tennessee, and Texas.  The U.S. sample includes all fifty states. 

Williamsburg food banks cater disproportionately to African-Americans and to females when compared to national and regional samples.  Williamsburg food bank patrons possess relatively higher education levels than their regional and national counterparts, but Williamsburg food bank patrons have similar household sizes and incomes to the census sub-samples.  Income and race emerge as leading factors in pushing households towards food banks.

Table 3 presents information on selected food security supplement variables for the Williamsburg food bank population and the national and regional sub-samples.  Williamsburg food bank patrons possess lower food stamp uptake rates than any of the national sub-samples, which mirror the data on Williamsburg and James City County’s low food stamp participation rates.  Consequently, Williamsburg food bank patrons more intensively use local food banks than patrons surveyed in the national and regional samples.  Furthermore, Williamsburg food bank patrons also have a higher tendency to borrow food or money for food from friends or relatives than the U.S. or South sample, but they do not have a higher borrowing rate than the South Atlantic.  The follow-up survey question to borrowing food or money for food from friends or family, however, shows that Williamsburg food bank patrons who answered affirmatively have a higher intense usage of borrowing than the national and regional sub-samples.

Table 3 also presents summary statistics for various coping mechanisms used by food bank patrons.  Williamsburg food bank patrons more intensively utilize food banks compared to the U.S. sample or the South or South Atlantic.  All of the sub-samples contain approximately equal numbers of food bank patrons borrowing food or money for food from friends or relatives.  Compared to the U.S. sub-sample, Williamsburg respondents who do borrow money or food from relatives, however, exhibit a greater propensity repeatedly to ask friends or relatives for help.

Questions specific to the Williamsburg food bank survey, including transportation questions and questions about Hurricane Isabel, show that 36.63% of food bank patrons had trouble getting to local food banks and that 69.07% of patrons were forced to use food banks due to Hurricane Isabel.  The Williamsburg United Way should pay attention to these findings as they evaluate and comment on Williamsburg’s disaster response system.

Tables 4 to 6 show detailed answers to the Federal Food Security Supplement for different subsets of the food bank populations.[76]  Williamsburg food bank patrons closely follow the national and regional samples in terms of answering trends for the CPS questions.  A high correlation exists between food insecurity and answering the questions affirmatively, as shown in table 5.  This makes logical sense as those who have a harder time achieving food security should find themselves in more dire straits.

The regional comparisons made in Table 4 show that the South Atlantic Census region has the highest percentage of food insecurity.  Virginia, however, does not fit the South Atlantic trend as numerous studies place the state among the most food secure states in the country.  Over a three year average, Virginia ranks 45^th out of 50 in food insecurity rates for states.[77] 

Virginia’s trends do not translate to Williamsburg, as local data shows that local residents possess a high propensity to rely on a few kinds of low-cost foods to feed children, and that, more alarmingly, children tend to skip meals or go hungry in Williamsburg at a higher rate than the U.S., South, or South Atlantic.  These high child food insecurity statistics (questions 12-16 in table 4) may stem from the smaller sample size of the Williamsburg food bank population.  Regression analysis attempts to pare through this and is presented in the results section.

  Table 6 divides the questions into the following situational categories:

·        Households perceive food budget or supply to be inadequate

·        Households perceive food eaten by adults or children to be inadequate in quality

·        Households report reduced food intake, or consequences of reduced food intake, for adults

·        Households report reduced food intake, or consequences of reduced food intake, for children

Food bank patrons, in both Williamsburg and the U.S., are more likely to report a higher frequency of feeling that their food budget or supply is inadequate than to report that they reduced children’s food or intake.  This trend finds support in the literature; families tend to cut back on food quality and adult meals before endangering children.[78]  Alarmingly, a high percentage of both U.S. and Williamsburg food bank patrons reported an inability to afford foods to make balanced meals.  While food bank handouts do provide food relief, the food does not always meet the highest nutritional standards.[79]

Tables 7 through 10 and Tables 12 and 13 present food security statistics for the Williamsburg and the U.S. food bank populations.  The Emergency Food Assistance System statistics in Table 8 come from the USDA 2001 report on food bank clients.  Table 7 further divides “food insecurity with hunger” into “food insecure with moderate hunger” and “food insecure with severe hunger.”[80]  Households with children experience higher rates of food insecurity than those without.  For the U.S. sample, however, this trend does not repeat itself in the most severe state of food insecurity “food insecure with severe hunger.”  In Williamsburg, food bank households with children disproportionately experience insecurity with severe hunger, but the small sample size might drive this.[81]  Overall security statistics show that Williamsburg families experience higher levels of food insecurity than the U.S. sample, but lower food insecurity levels when compared to the EFAS sample.

The disparity between the U.S. and Emergency Food Assistance System statistics in Table 8 makes comparing food security rates difficult.  The CPS surveys households and does not interview transients, homeless, or displaced persons.  This likely under-represents the food insecure population because many homeless individuals and displaced families utilize food pantries in order to make ends meet.  For example, in 1997 and 1998, Biggerstaff et al (2002) found 15% of Virginia’s food bank patrons were homeless.

Table 9 presents local Williamsburg food bank security statistics.  A t-test for Table 9 found no statistically significant differences between the food security levels for the local food banks.  Table 10 reports food security and food insecurity rates for different demographic groups.  For example, only 26.9% of female food bank patrons are food secure in the U.S. sample and only 13.33% in the Williamsburg sample.  Table 11 presents information on the number of household visits to local food banks.  Finally, tables 12 and 13 present food security statistics by race for both the Williamsburg and U.S. samples.  While African-Americans tend to patronize Williamsburg food banks more often than other racial groups, they do not have lower food security rates.

The statistics on Williamsburg food bank patrons show that African-American females with low household income disproportionately use food banks.  The Williamsburg food bank patron statistics differ slightly from the statistics on Virginia food bank patrons as shown by Biggerstaff et al (2002).  Biggerstaff et al (2002) interviewed 1,500 food pantry and soup kitchen users in Virginia and reported a 66.2% female participation rate and a 46% African-American participation rate, compared to an 82.2% female participation rate and 55.5% African-American participation rate in this study. 

Compared to Biggerstaff et al (2002), Williamsburg food bank patrons possess higher education levels than the Virginia sample.  In Williamsburg, 34.7% of food bank patrons have achieved a high school degree and some college or greater compared to only 17.0% for the Virginia sample.  Biggerstaff et al (2002) note that for 25.7% of food pantry patrons, the chief reason to get food assistance was because they were working poor who could not make ends meet.  Their sample also found that 16.1% of respondents cited losing food stamp benefits and that 22.0% of respondents had their phone service cut off during the year.  Future research in the Williamsburg area should incorporate some of the hardship questions that Biggerstaff et al (2002) utilize.

Demographic analysis helps to build a picture of food bank patrons.  The preceding tables show that food bank patrons, both nationwide and in Williamsburg, are predominantly female, unemployed, and possess lower education and household income levels than the general population.  Compared to the U.S. food bank patron population, Williamsburg possesses higher rates of African-American and female patrons.  Williamsburg’s patrons also tend to have higher levels of educational attainment. 

Food bank patrons utilize many coping mechanisms in order to make ends meet.  These low-income households rely on food stamps, borrowing food or money for food from relatives or friends, and visiting food pantries in order to achieve food security.  Williamsburg patrons tend to use non-welfare based coping mechanisms, as shown by their poor food stamp participation rates.

VIII. Regression Results

Tables 14 through 17 present the results from the probit regressions.  For each sample the primary and secondary probit are estimated with the chosen dependent variable (e.g. food security).  The columns labeled dy/dx show the difference in the probability of the selected variable occurring between two groups.  For example, the female variable in the probit results shows the percentage increase or decrease in a dependent variable if the respondent is female, therefore, not male.  Standard errors are italicized underneath the derivatives.

The study reports probit results for three different samples of food bank patrons, Williamsburg, the U.S., and a pooled sample.  A chow test, run on ordinary least squares regressions for the three samples, shows that the pooled regression is unnecessary because of the substantial difference between the residuals for the Williamsburg sample and the national sample.[82]  The study leaves the pooled regressions in the data charts to serve as a contrasting point for the regressions, but analysis should focus on comparing the Williamsburg food bank patron’s model and the U.S. food bank patron’s model.

Table 14 shows the results for the food security probit.  This study’s probit on food security finds that low education, failing to complete high school, and low-income make households more susceptible to food insecurity.  A $10,000 rise in income increases food security by 13% in the Williamsburg sample and 3% in the national sample.  In all the samples, single parent households possessed a high propensity for food insecurity.  In both the Williamsburg and national samples senior citizens, on average, are more food secure than their younger counterparts, those aged 35-64. 

Not surprisingly, the literature often shows a correlation between food insecurity and low-incomes, low education, and single parent households.  Few studies, however, actually perform regression based analysis on the subject.  The regression, shown in Table 14, supports the general economic theory regarding the determinants of food insecurity. 

The inclusion of food stamps as an explanatory (independent) variable in the probit on food security does not significantly change the coefficients of any other explanatory variables.[83]  The major differences between the Williamsburg food bank patron regression and the national sample regression are the non-White dummy variable, the single alone dummy variable, and the female dummy variable.[84]  The non-White variable and female variable contain substantial noise, as indicated in their standard deviations, which makes them an imprecise estimator.  The single alone variable posts more of a problem in that it turns up statistically significant in both the Williamsburg and US model yet takes on different signs for the models.  This most likely comes from the relatively small sample size of the Williamsburg population.  In sum, table 14 shows the factors that effect the various equations that determine food security.

Table 15 shows results from the probit models that estimate the determinants of food stamp usage.  In both the U.S. and Williamsburg sample, food stamp usage decreases with higher education, higher income, and higher age.  Raising one’s income from the less than $10,000 a year range to the $10,000-$20,000 range decreases food stamp participation by 22% in the Williamsburg sample and by 16% in the national sample.  This can be interpreted in two ways: either the household gains the ability to buy food with the extra income or the household becomes ineligible for food stamps due to the income increase.  The literature in the subject presents a muddled picture as some families opt out of food stamps when they become ineligible for Temporary Assistance for Needy Families (TANF) due to income gains while other families opt out of food stamps because they can personally achieve higher food security with the extra income.[85] 

Households headed by single females show a strong propensity to use food stamps.  These results fit with previous research and with USDA reports on the characteristics of food stamp using households.  Race emerges as a noisy estimator in that being non-white in the samples decreases food stamp participation.  This most likely comes from sample size as the literature predominantly shows that African-Americans have higher rates of food stamp usage.

This study’s food stamp regression analysis finds many of the same results as Biggerstaff et al (2002).  In both studies single parent households and households with less than a high school degree possessed a higher propensity to use food stamps.  Furthermore, both studies find that females disproportionately utilize food stamps and that higher income levels discourage food stamp usage.  Biggerstaff et al performed logistic analysis, which closely mirrors this study’s probit analysis.[86]  Interestingly, Biggerstaff et al report that “hardships” experienced by the household decrease food stamp usage.  This means that households that experience loss of employment or similar hardships do not immediately utilize food stamps.  This may point to coping mechanisms alternative to welfare that food insecure families in Virginia utilize.

Tables 16 and 17 show probit results for two different coping mechanisms that food bank patrons use in order to achieve food security.  Table 16 performs a probit on whether respondents “got food or borrowed money for food from relatives or friends.”  Over 50% of the food bank patrons in both the U.S. and Williamsburg samples reported receiving food or borrowing money for food from friends or relatives.  The regression results show that younger, less educated, and lower income patrons are more likely to borrow from friends or relatives.  Williamsburg females exhibit a higher propensity to borrow from friends or relatives than males, but the patterns for females on a national level is less clear.

Table 17 reports regression results for heavy or intense use of food banks.  Eighty-seven percent of Williamsburg food bank patrons cited using a food bank more than only one or two months a year, compared to 49.58% for the U.S.  The regression returned few statistically significant results, but general patterns emerge which show that higher amounts of income, higher educational attainment, and older age levels make one less likely to heavily rely on food banks.

The regressions on coping mechanisms, including food stamp participation, borrowing food, and using food banks, show that generally those with less income and education more intensively utilize these services.  These groups also remain the most food insecure as show in Table 14.

Table 18 shows the individual dy/dx values for probits on individual CPS food questions, for both the U.S. and Williamsburg model.  The table divides the questions by their subject matter and runs the regressions on the respondent answering affirmatively to the question.  All of the probits in Table 18 use the secondary regression model, which includes all of the explanatory variables.  Analysis of Table 18 shows that females, low income households, and less educated households exhibit greater tendencies to answer affirmatively to questions regarding food insecurity.

Regression analysis shows ways that local communities can increase food security.  The analysis shows that as educational attainment and income rise, food security increases in both the local and national sample.  Raising wages and education emerge as important societal issues in reducing food insecurity.

IX. Summary and Conclusion

Myths and stereotypes abound concerning the food bank population.  The evidence presented in this study helps to dissolve some of these preconceived notions.  This study’s data analysis helps to paint a picture of food bank patrons in the U.S. and Williamsburg and to uncover many of the challenges that these people face in their struggle to achieve food security.  Food bank patrons utilize various coping mechanisms in order to make ends meet, including food stamps, borrowing food or money for food from friends or relatives, and heavily utilizing local food banks. 

The Williamsburg Area case study uncovers interesting patterns in local food bank usage.  While lower income African-American females disproportionately use the local food banks, they do not experience higher food insecurity rates than their White counterparts.  Compared to other food bank populations, Williamsburg’s food bank patrons possess higher levels of education, but they also have lower employment rates.  The demographic makeup of Williamsburg food bank patrons parallels the local labor market structure, where African-Americans earn less on average than other racial groups.  Further, Williamsburg food bank patrons have food security rates that closely match those found in the USDA’s 2001 study of food pantries.  Compared to CPS sub-samples, however, Williamsburg food bank patrons possess lower food security rates. 

Regression analysis showed a high correlation between food insecurity, on the one hand, and low education rates and low-income levels, on the other.  Substantial gains in food security would occur from increasing household income levels for food bank patrons.  More research should focus on the hardships food bank households face, and the various reasons that households go to food banks.

Local communities can increase food security by increasing cooperation and communication among organizations in the area.  In order to achieve greater food security in the Williamsburg area, all major players in the Williamsburg economy must work together in an effort to help food bank patrons.  Central Virginia Food Bank Director Fay Lohr shares this sentiment:

"Non-profits alone can't solve the problem of hunger. Government alone can't solve the problem of hunger. The private sector alone can't solve this problem. We must all work together, and this project (increasing food stamp participation) will move us along the path to meeting the nutritional requirements of those who need assistance."[87]

Local government should recruit more local food bank patrons to participate in the federal food stamp program.  In addition to feeding hungry people, food stamps bring federal funds into the locality and help boost businesses and employment in the area, all at low cost to local government.  The federal government funds 100% of food stamp benefits and 50% of the state’s administrative costs.  Ways to increase food stamp participation numbers include shortening applications, expanding office hours, recruiting at local food banks, and allowing for applications to be made at other local sites such as Olde Towne Medical Center, the United Way, and other local agencies.[88] 

Private businesses should follow the model set forth by Food Lion, named the retail regional donor of the year in 2002 by America’s Second Harvest.[89]  The Emerson Good Samaritan Act protects private businesses from liability for donating food to food banks as long as the business had “good intentions” in the donation process.  Local businesses also should strive to pay a living wage.  Regression analysis shows a high correlation between income and food security, while data analysis shows the low paying nature of many of Williamsburg’s jobs.  A raise to an $8/hr living wage would raise food security for many food bank patrons.

Non-profits in the Williamsburg area could provide more services to the local area if they organized more cohesively.  While a strong movement towards cohesion does exist in the food pantries, other local non-profits lack coordination.  The creation of the Williamsburg Area Shelter Relief Program, a rotating winter homeless shelter provided by local churches, could serve as a good first step towards unifying the non-profits.  Non-profits should continue to look for ways to work with local government.

In technical issues, the federal government should consider changing the Food Security Supplement of the CPS by adding questions about tradeoffs food bank patrons must make.  The CPS does not capture the choices that households must make between food, paying for rent, utilities, prescription drugs, and childcare.  These behaviors often indicate a deeper problem than running short on a grocery bill each month. 

The U.S. and Williamsburg rely on the Emergency Food Assistance System to help promote food security for low-income Americans.  Nationwide communities should use food banks as rallying points in the fight against hunger and poverty.  Legislators, welfare officers, managers, and others should remember the substantial food bank population when formulating public policy.

References:

Secondary Sourcebooks:

Edin, Kathryn and Laura Lein. Making Ends Meet: How Single Mothers Survive Welfare and Low-Wage Work. New York: Russell Sage Foundation, 1997.

Ohls, James and Harold Beebout. The Food Stamp Program: Design Tradeoffs, Policy, and Impacts. Washington D.C.: Urban Institute Press, 1993.

Schwartz-Nobel, Loretta. Growing up Empty: The Hunger Epidemic in America. Harper Collins, 2002.

Secondary Articles:

America’s Second Harvest. “State Government Responses to the Food Assistance Gap 2000: Annual Report and 50 State Survey from Food Research and Action Center and America’s Second Harvest.” 2000. <http://www.secondharvest.org/policy/gap_2000/contents.html> (20 February 2003).

America’s Second Harvest. “Red Tape Divide: State-by-State Review of Food Stamp Applications.” <http://www.secondharvest.org/policy/food_stamp_study.html> (20 February 2003).

Barbara Cohen et al. “Community Food Security Assessment Toolkit.” United States Department of Agriculture: Economic Research Service. Number 02-013, July 2002.

Biggerstaff et al. “Living on the edge: examination of people attending food pantries and soup kitchens.” Social Work. vol 47 (July 2002): 267-278.

Bjoras, George. “Food Insecurity and Public Assistance.” JCPR Working PaperSeries 2001.  <http://www.jcpr.org/wpfiles/borjas_foodinsecurity.pdf> (1 March 2004).

Brian Taber and Anna Hockaday, “FY 03 James City County Social Services Annual Report.” James City County Division of Social Services. (2003).

Economic Research Service (care of Mathematica Policy Research, Inc.) “The Emergency Food Assistance System—Findings from the Client Survey.” July 2003. <http://www.ers.usda.gov/publications/fanrr32/>(2 January 2004).

Hagstrom, Paul. “The food stamp participation and labor supply of married couples: an empirical analysis of joint decisions.” Journal of Human Resources. Vol 31 (Spring 1996): 383-404.

Hamilton et al. “Household Food Security in the United States in 1995: Technical Report of the Food Security Measurement Project.” (1997). USDA.

Hanson and Golan. “Issues in Food Assistance: Effects of Changes in Food Stamp Expenditures Across the U.S. Economy.” United States Department of Agriculture: Economic Research Service. Number 26-6 (August 2002).

Kennedy, Eileen. “Public policy in nutrition: the US nutrition safety net – past, present, and future.” Food Policy Journal. Vol 24 (1999): 325-333.

Joint Center for Housing Studies of Harvard University. “The State of the Nation’s Housing.” <http://www.jchs.harvard.edu/publications/markets/son2003.pdf> John F. Kennedy School of Government, 2003.

Kim and Mergoupis. “The working poor and welfare recipiency: participation evidence and policy directions.” Journal of Economic Issues. Vol 31 (September 1997): 707-729.

O’Brien, Doug and Hailey Aldeen. “Hunger in America 2001.” America’s Second Harvest. Third National Hunger Study. (2003).

Reschovsky, James. “The emergency food relief system: an empirical study.” Journal of Consumer Affairs. Vol 25 (Winter 1991): 258-278.

Rodgers et al. “Food Security and the Federal Minimum Wage.” National Urban League Institute for Opportunity and Equality. Special Research Report (SRR-03-2002).

Rose et al. “Socio-Economic Determinants of Food Insecurity in the United States: Evidence from the SIPP and CSFII Datasets.” Washington, DC: Economic Research Service, USDA, Technical Bulletin No. 1869.

Sashin, Daphne “JCC resists low-rent housing: Residents fear high crime, lower property values.” Daily Press (Newport News – Williamsburg VA). February 16, 2004.

Tolbert, Bill. “$2.9 million for 28 more affordables.” Virginia Gazette (Williamsburg VA). February 7, 2004.

Nord, Mark et al. “Guide to measuring Household Food Security: Revised 2000.” United States Department of Agriculture <http://www.fns.usda.gov/fsec/FILES/FSGuide.pdf>  (15 November 2003).

Nord, Mark et al. “Household Food Security in the United States, 2002.” United States Department of Agriculture. < http://www.ers.usda.gov/publications/fanrr35/>  Report No. (FANRR35) (15 November 2003).

Wegener, Victoria. “Food Stamp Education and Outreach Working to Provide Nutrition Benefits to Eligible Households.” Welfare Information Network. December 1999. <http://www.welfareinfo.org/foodstampout.htm> (25 January 2003).

Primary Sources:

Board of Directors FISH Food Pantry. Interview by author, October, 2003. Williamsburg VA.

Cindy Zangardi: Director- St. Olaf’s Food Pantry. Interview by author, October, 2003. Williamsburg VA.

Food Bank Patrons. Interviews by author, November 17 – December 14, 2003. Local Williamsburg Food Surveys, Williamsburg, Va.

Joyce O’Brien: Director of Planning- United Way of Greater Williamsburg VA. Interview by author, December, 2003. Williamsburg VA.

Reverend Ellis: Director- Operation Hope Ministry. Interview by author, October 2003. Williamsburg VA.

Thumper Newman: Director- A Gift from Ben and Helping Hands Food Pantry. Interview by author, October, 2003. Williamsburg VA.

Social Worker- Salvation Army. Interview by author, January, 2004. Williamsburg VA.

Vivian Van Holten: James City County Social Services. Interview by author, November, 2003. Williamsburg VA.

General Websites:

CPS Supplements. “Food Security Data” Bureau of Labor Statistics and Bureau of the Census. 1999. <http://www.bls.census.gov/cps/foodsecu/foodsecu.htm> (21 January 2003).

Economic Research Service. “Food Security in the United States.” United States Department of Agriculture. N.d.  <http://www.ers.usda.gov/briefing/foodsecurity/> (21 January 2003).

The Finance Project. “Overview of the Food Stamps Program.” Welfare Information Network. N.d. <http://www.financeprojectinfo.org/WIN/food.asp> (1 February 2003).

United States Department of Agriculture. “Food Stamps Overview.” <http://www.fns.usda.gov/fsp/> (22 January 2003).

Conference Sources:

Lee Mercer et al. “Food Stamp Outreach: Building Your Capacity to Connect Families with Benefits.” Presentation at the 2004 National Anti-Hunger Policy Conference, Washington DC. February 22-24, 2004.

Nicole Woo et al. “Primers on Child Nutrition Programs.” Presentation at the 2004 National Anti-Hunger Policy Conference, Washington DC. February 22-24, 2004.

Stacy Dean et al. “Best Practices for Food Stamp Access: Research, Reality Checks and Recommendations.” Presentation at the 2004 National Anti-Hunger Policy Conference, Washington DC. February 22-24, 2004

Table 7: Household Food Security Status: Comparison of Williamsburg and US

Table 8: Household Food Security Status: Comparison of Williamsburg, US, and EFAS

(EFAS: Emergency Food Assistance Report compiled by USDA in 2001)