FOOD INSECURITY AND FOOD RESOURCE UTILIZATION AMONG LOW-INCOME HOUSEHOLDS IN NORTH ALABAMA
Abstract
The objective in this study is to determine the participation rate of low-income households in food assistance programs, and to examine the factors that predict their participation in public-funded and private/community-based food assistance programs in North Alabama. The analysis draws on primary data collected using a food security and socio-economic survey that was administered to 700 households between August 27 and September 30, 2016. The main tools of analysis include descriptive statistics and logistic regression model. The logistic model results reveal that household income, gender of household head, household head education, household size and ethnicity are among the significant predictors of food security resource utilization in the study area.
Keywords: Food insecurity, urban households, food assistance programs, participation rate, logit model
Introduction
Food security according to the Life Sciences Research Office definition means access to enough food for an active, healthy life (Wunderlich and Norwood, 2006). It includes at a minimum (a) the ready availability of nutritionally adequate and safe foods and (b) 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). Thus, food insecurity exists whenever the availability of nutritionally adequate and safe foods or the ability to acquire acceptable foods in socially acceptable ways is limited or uncertain (Wunderlich and Norwood, 2006). Based on this definition, 12.3% of U.S. households were food insecure in 2016, with food insecurity rates of greater than 38% for households with incomes below the federal poverty level (Coleman-Jensen et al., 2017). Food insecurity has been even greater in the U.S. southern region where 39% of households were food insecure in 2016. Alabama is one of the most affected states in the nation (just behind Mississippi and Louisiana), with 18.1% of households being food insecure, over a 3-year period, 2014-2016 (USDA ERS, 2017). Furthermore, Alabama’s high food hardship rate (20% of the population in 2016; FRAC, 2017) and poverty rate (17% of the population in 2016) have meant that a large numbers of low-income households are struggling to cover their basic food needs. According to the Center on Budget and Policy Priorities, 851,000 Alabamians— or roughly 17% of the state’s population—received Supplemental Nutrition Assistance Program (SNAP) benefits in 2016. Of those 71 percent are families with children; 32 percent are in families with members who are elderly or have disabilities; and 40 percent are in working families (Gore, 2017).
Ideally, food security could be improved and sustained through a wide range of policies and programs designed to address household food inequalities. While these programs have improved the security of many communities, food insecurity prevalence has continued, especially among low-income households and those with children. Anderson et al. (2015) have attributed the continued prevalence to the low-participation in the different food assistance programs by many food-insecure households. Other studies have explored the potential reasons for the low-participation rates (Martin et al., 2003; USDA Food and Nutrition Service, 2007; Fuller-Thomson and Redmond, 2008), particularly for SNAP. This study builds on previous studies that have addressed food insecurity and hunger in urban areas in the U.S. The study seeks to examine the participation rate in food assistance programs among low-income households, and to determine the factors that predict households’ participation in public-funded and private/community-based food assistance programs.
Study Area
The study was conducted in the Huntsville Metropolitan Statistical Area (MSA), Alabama (Figure 1). Located in two north Alabama counties (Madison and Limestone), the Huntsville MSA is the fastest growing and second largest MSA in Alabama. The MSA is comprised of nineteen cities (Huntsville, Ardmore, Athens, Brownsboro, East Limestone, Elkmont, Gurley, Harvest, Hazel Green, Madison, Meridianville, Monrovia, Moores Mill, New Hope, New Market, Owens Cross Roads, Redstone Arsenal (U.S. Army post), Toney and Triana) of which Huntsville is the largest. The MSA has experienced tremendous growth in the past ten years, expanding from a population of about 342,376 in 2000 to 444,752 in 2015; over a 27% increase in population. Such growth is a testament to the metropolitan area’s robust economy.
Figure 1. Map of Huntsville Metropolitan Statistical Area, Alabama
Food insecurity within the Huntsville MSA is a continuing problem. As the economy fluctuates, more people are seeking assistance from SNAP demonstrating that food insecurity is growing. According to a recent review conducted by the Alabama Public Health, North Alabama counties have some of the highest food insecurity rates in the state. Nearly one in four children in North Alabama are food insecure—recent statistics suggest that there are 59,610 children living under these conditions across North Alabama. The data analyzed in the current study are drawn from selected neighborhoods in the city of Huntsville, which is the fourth-largest city in Alabama. The 2010 census estimated Huntsville’s population at 180,105. By 2016 the estimated population grew to 194,057.
Literature review
Arcuri et al. (2016) provided a detailed review of food security programs in the U.S., and other developed economies. The general theme that emerges from previous food security studies is that the response has been both public and private food assistance programs. The majority of nutrition assistance to the low-income population in the U.S. is provided by the public sector, through the SNAP (former Food Stamps); School Lunch and Breakfast Programs; and the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC). Daponte and Bade (2006) and Hoefer and Curry (2012) provide an overview of federal food assistance in the U.S., describing both SNAP and WIC. SNAP, administered by the states through local welfare offices, is the biggest program of cash assistance targeted to food purchases, addressing more than 45 million individuals, with a total expenditure of almost $74 million 2015. WIC is administered by the Food and Nutrition Service of the USDA through 90 state agencies, working with local organizations and clinics which directly provide the service. Both programs are entirely funded through federal resources and each year, Congress assigns a certain amount of money to the programs (Cohen, Ilieva, 2015; Arcuri et al., 2016).
Private organizations—food banks, food pantries, soup kitchens, and shelters—have helped to prevent even greater rates of hunger in America’s low-income population by playing a valuable, complementary role to federal assistance programs (Robaina and Martin, 2013). Daponte and Bade (2006) show the interdependence between the evolution of public food assistance and the origin of private food assistance. They maintain that the Emergency Food Assistance Program (TEFAP) “institutionalized the private food distribution network” in the U.S., by purchasing commodities for distribution through private charities (Daponte and Bade, 2006:679). Popielarski and Cotugna (2010) report on a social enterprise venture in the form of an agency-run grocery store developed by the Food Bank of Delaware with the goal of improving the availability, accessibility, and affordability of food for the citizens of an inner-city neighborhood. Vitiello et al. (2014) investigated food banks’ involvement in local agriculture to understand if these programs are influencing the traditional role of food banks in community food systems.
Reentry, the role of community interventions in addressing food insecurity has revived a long-standing debate. Some researchers argue that food banks exacerbate rather than alleviate food insecurity by masking it, undermining social justice and relieving governments of their duties (Anderson, 2013; Riches and Tarasuk, 2014; Booth and Whelan, 2014; Lambie-Mumford and Dowler, 2014; Tarasuk, 2014; Caraher, 2015). In contrast, others have underline the importance of food banks, affirming the importance of their role in addressing hunger and health issues (Wakefield, 2013; Webb, 2013). In this respect, it is argued that the strategic position community organizations have in changing food insecurity intervention strategies may be strengthened.
In addressing the different aspects of food insecurity problem, previous studies have revealed an interesting ethnic/racial difference in response to food security support. For instance, Yu, Lombe, and Nebbitt (2010) found that food insecurity was much higher in African American households (48%) that participated in a food stamp program, when compared to Caucasian households (29%) participating in the same program. Further, Caucasian households were found to score significantly lower in food insecurity when receiving supplemental informal food supports than African Americans who also received supplemental food supports. Similarly, the USDA Economic Research Service (Coleman-Jensen et al., 2017) collects data on food security and hunger in the annual report, Food Security Status of U.S. Households, with the most recent report being released in 2017 on data collected from 2016. The report revealed that rates of food insecurity were higher than the national average for African-American and Hispanic households under or near the poverty line with children that were headed by single parents. Specifically, African-American households had a 22.5% rate of food insecurity, while Hispanic households had 18.5% rate. In addition, single mother households had a rate of 13.9% food insecurity.
Studies have also documented important differences in food security resource utilization across racial, demographic, and social dimensions (Martin et al., 2003; Greenwald, 2017). Using data from the 1999 Current Population Survey, Bhattarai et al. (2005) examined food stamp and food pantries participation for low-income households. They revealed that household income, the level of food insecurity, household structure, and metro versus nonmetro residence affected participation decisions in both programs. Within the general population, Anderson et al. (2015) contend that a family’s personal resources relative to its size is key to determining its food security. Other determinants have been reported to include access to and use of food security programs, such as SNAP, WIC, food pantries and community gardens (Ratcliffe et al., 2011). Studies have also found elevated risk of food insecurity among low income individuals, families with children in the household, residents of ”food deserts,” migrants, and other socially and economically disadvantaged people (Patton-Lopez et al., 2014; Mayer et al., 2014; Greenwald and Zajfen, 2017).
Researchers have also explored potential reasons for the low participation rates in food assistance programs (Martin et al., 2003; USDA Food and Nutrition Service, 2007; Fuller-Thomson and Redmond, 2008). Particularly, Fuller-Thomson and Redmond (2008) have argued that, barriers to information and confusion regarding eligibility, level of benefits, and program policies discourage potential participants from applying for SNAP. Other obstacles that may further discourage participation include a lack of transportation or difficulty reaching program administration offices, as well as, language barriers (Fuller-Thomson and Redmond, 2008). Also, a fear of being stigmatized for seeking social assistance often discourages individuals from seeking food assistance (Fuller-Thomson and Redmond, 2008). The Food and Nutrition Service has also indicated that the likelihood of participation in food assistance programs increases as the level of benefit increases (USDA Food and Nutrition Service, 2007). As Martin et al. (2003) have noted, it is important to identify why some choose to utilize food assistance in order to develop appropriate programs and policies to better address the needs of those eligible.
Methodology
This study uses econometric approach to examine the factors that predict food-insecure household’s participation in public and community-based food security programs in North Alabama. The variables explained in the model are dichotomous, taking a value of one when the household utilizes public/private food security programs and the value of zero otherwise. In such cases where Y is a dummy variable, binary choice models should be applied. The main idea behind the model is to find the relationship between the probability (Pi) that Y will take a 1 value and the characteristics of considered individuals. A general class of binary choice models assumes that
(1)
Where
probability, ,
cumulative distribution function (CDF),
parameters, ,
value of explanatory variables, for i-th household,
number of explanatory variables,
sample size.
The two common binary choice models are the binomial logit model and the binomial probit model. The current analysis employs the former, which can be represented as:
(2)
Where and denotes the logistic cumulative distribution function. The explanatory variables can have the form of both dichotomy and quantitative variables. For a quantitative variable , Greene (2000) shows that:
(3)
where f(.) is the density function that corresponds to the cumulative distribution function F(.). Because the CDF is monotonically increasing in its argument, the second term in the chain rule derivative given in (3) is always positive. As a result, the sign of the parameters always equals the sign of the partial derivative of interest. The partial derivatives are given by:
(4)
where ?(.) is the logistic CDF.
The advantage of the logit model over the probit model is that effects of changes in explanatory variables can be interpreted in terms of odds ratios. Odds are defined as the ratio of two probabilities Pi and (1- Pi), i.e. the ratio of the probability of occurrence of an event to that of non-occurrence. The logit model odds equal
, because
(5)
The exponential relationship provides an interpretation of odds ratio. For a unit change in, the odds are expected to change by a factor of , holding all other variables constant. The parameters of logit model are estimated by maximum likelihood method (ML).
The estimated models use household response to the question (Have you received food assistance from the listed food security program?: Yes, No responses) as the dependent variables. The actual realization of the dependent variables yi are assumed to follow,
(6)
Using the key food assistance programs/food security programs (snap, wic, school, fbank, fpantry and cgarden) as dependent variables, six logistic regressions are estimated using the general specification:
(7)
The dependent variables are binary, representing respondents’ utilization of the various food assistance programs. Because of the nature of the data i.e., categorical question with two choices: yes (yi = 1) or no (yi = 0), a binary logit regression model was adopted to assess the factors that are expected to predict utilization of food assistance programs. The independent variables are hypothesized to predict respondents’ utilization of the various food assistance programs. In the case of the explanatory variables, the estimated results are interpreted using the odds ratio.
Data and Sampling
Primary data were collected through a household food security and socio-economic telephone survey conducted on August 27 through September 17, 2016. The questionnaire was administered in 14 low-income neighborhoods within the city of Huntsville (Chelsea, West Huntsville, Huntsville Park, Brandontown, Oakwood, Rutledge Heights, Lakewood, Vaughn Corners, Rideout Village, Terry Heights, Brookhurst, Meadow Hills, Cavalry Hill and Edmonton Heights). The selected neighborhoods are located in a cluster of census tracts that are defined as food desert (USDA ERS, 2016). These neighborhoods were chosen because they are typical in many ways of inner city communities in the southern United States. Their populations include a large proportion of minorities, female headed households, with incomes below the poverty line, high unemployment rates, high crimes, among other disparities. Within these neighborhoods, three stage cluster random sampling with probability proportion to size sampling technique was used to select a sample of 700 respondents. After cleaning the data for incomplete responses, a usable sample of 679 households was compiled.
The food security section of the survey was based on the shorter version of Household Food Security Survey Module (HFSSM) administered annually in the Current Population Survey (CPS), the Food Security Supplement ((Bickel et al., 2000). The standard 18-item module, or core module, is able to capture the various combinations of experiences, behaviors, and food conditions for each level of food security observed currently in the U.S. using a variety of indicators (Bickel et al., 2000). A shorter version of this module is frequently used when survey time and space is limited (Bickel et al., 2000). This version, consisting of a six question subset, was adopted in the current study. The advantage of adopting the 6-item subset is that the survey findings can be compared directly with national and state-level standard benchmark statistics published annually by USDA Economic Research Services (ERS) and with many national- or regional-level tabulations of population subgroups available in the USDA ERS reports. The questionnaire also corrected information on food security programs (SNAP, WIC School-operated, food banks, pantries and community gardens) utilized by the respondents, opinions about services received from selected programs, and socio-economic background variables such as age, employment status, marital status, ethnicity, education, presence of children in the household, income, household size.
Assessment of Household Food Security
Responses to the six questions on the short form of the HFSSM were scored for each respondent and summed up to generate each respondent’s raw score (Figure 2). The raw scores (ranging from 0 to 6) were used to group the respondents into food secure and food insecure households. Out of the 679 low-income households who responded, 144 were characterized as being food insecure (Table 1). The 144 food insecure household make-up the sub-sample analyzed in this study, to identify the factors that predict participation in food assistance programs. Household responses are analyzed based on responses to questions related to their participation in each of the six food assistance programs: SNAP, WIC, School-operated, Food Bank, Pantries and Community Gardens. The breakdown of the self-reported participation is presented in Table 1. As presented in the table, the overall participation rate across the sample (food secure and food insecure households) was 21%. Similar low participation rates have been reported in previous studies (Anderson et al., 2015). In terms of public-funded versus private/community-based food assistance programs, respondents utilized more of the private/community-based programs (28%) compared to public-funded programs (17%). SNAP was the most utilized public food assistance program while food banks and food pantries were the most used among the private/community-based food assistance programs in North Alabama.
Figure 2. U.S. Household Food Security Questions (USDA/ERS 6-item Module)
Source: Generated by author using information from Bickel et al 2000.
Table 1. Participation in Food Assistance Programs
Low-income
(n = 679) Food secure
(n = 535) Food insecure
(n = 144)
No participation in food insecurity programs 79% 95% 19%
Participation in food insecurity programs 21% 5% 81%
Participation in government programs 17% SNAP 12% 0% 57%
WIC 3% 0% 14%
SCHOOL 5% 1.5% 18%
Participation in non-government programs 28% FOOD BANK 12% 0.5% 55%
COMMUNITY GARDEN 6% 3% 17%
FOOD PANTRY 10% 0% 47%
Econometric Analysis
The analysis utilizes the logit regression to model the six food assistance programs variables (snap, wic, school, cbank, cpantry and cgarden), which were recoded into dichotomous variables. Detailed description and descriptive statistics of the dependent and independent variables included in the analysis are presented in Table 2. For instance, respondents who indicated that they had received SNAP benefits where coded 1 and 0 otherwise; those who indicated receiving assistance from food pantries in the past twelve months were coded 1 and 0 otherwise; and so on. In a logit model the dependent variable is expressed as the natural log of the odds of being categorized in one category as opposed to the other. Therefore, an odd expresses the ratio between the frequencies of those in one category against the frequency of not being in that category. This odd is dependent on the independent variables used in the particular logit model. This means the impact of the predictor variable may be explained in terms of odds ratios. Parameter estimates known as logit coefficients, which are estimators of the change in the dependent variable caused by a unit change in the independent variable are also calculated (Garson, 2009). In terms of interpretation, a positive logit coefficient suggests that the presence of the independent variable would increase the odds of the dependent variable, and conversely, a negative logit coefficient would decrease the odds.
Table 2. Variable Descriptions
Variable Description Mean Std. Dev
Dependent Variables
SNAP =1 if received SNAP benefits; 0 otherwise. 0.32 0.468
WIC =1 if received WIC benefits; 0 otherwise. 0.23 0.422
SCHOOL =1 if participation in national school lunch program, summer feeding programs or school breakfast program; 0 otherwise. 0.26 0.438
FBANK =1 if received food assistance from a food bank; 0 otherwise. 0.27 0.446
FPANTRY =1 if received food assistance from outlets typically operated by churches and other community non-profit organizations; 0 otherwise. 0.29 0.456
CGARDEN =1 if received food assistance from community gardens; 0 otherwise. 0.24 0.430
Independent Variables
ETHNICITY =1 if African American; 0 otherwise. 0.57 0.497
GENDER =1 if female headed household; 0 otherwise. 0.85 0.361
EDUCATION =1 if high school or lower education; 0 otherwise. 0.47 0.501
MARITAL =1 if divorced; 0 otherwise. 0.47 0.501
INCOME =1 if household income <$25k; 0 otherwise. 0.34 0.475
HOUSEHOLD SIZE = Number of people in the household. 2.41 1.40
Results and Discussion
Logit model was employed to identify factors that predicted participation in food security programs. Before fitting the model, it was important to check whether serious problem of multicollinearity and association exists among explanatory variables. For this purpose, Variance Inflation Factor and contingency coefficient tests were used for the continuous and discrete variables, respectively. The choice of the final variables in equation 7 were best on the aforementioned analyses. Tables 3 and 4 present the results, showing that, the models fitted the data reasonably. As specified, the model explained significant non-zero variations in factors influencing food security assistance program use among low-income households in North Alabama. The estimated coefficients of determination (count R-square) were fairly high across the six models, suggesting that between 68% and 77% of the responses to program participation were correctly predicted among government-funded programs (Table 3). Similarly, between 72% and 76% of the responses to participation in non-government food assistance programs were correctly predicted (Table 4). It must be noted here though, that in binary choice models goodness of fit measures are of secondary importance—what matters are the expected signs of the slope parameters in the models and their statistical and practical significance (Gujarati, 2011).
Table 3 presents the estimated logit model coefficients predicting the participation of low-income households in public-funded food security programs (SNAP, WIC, and school-operated). Based on the results, ethnicity appears capable of influencing the decision to participate in food security programs. The variable show a significant positive sign, which supports the notion of a strong relationship between ethnicity (African American) and household participation in all the three public-funded food security programs. In Model 1 (SNAP), the expected odd ratio of ethnicity is equal to 2.2, which indicates that a low-income household headed by an African American has 2.2 times chance to participate in SNAP compared to low-income White headed household. Model 2 (WIC program) and Model 3 (School-operated programs) yield similar results, suggesting that the odds of participating in public sponsored food security programs are higher among African American headed households compared to White headed households.
Previous research has identified different groups that are more likely than others to participate in food assistance programs, particularly SNAP. Predictor variables identified by Biggerstaff et al. (2002) include being a child under the age of eighteen, being a single parent, having low education, being a minority, being a woman, and being unemployed. Similar results were found by Smith et al. (2016) who showed that those more likely to participate in government programs, such as SNAP, included non-Hispanic blacks and less educated. In line with the findings reported in these studies, the education variable is shown to be a significant predictor, but only in Model 3 (School-operated programs). The expected odd ratio of education is equal to 2.8, which implies that, less educated (high school or lower) households have 2.8 times chance to utilize national school lunch, summer feeding or school breakfast programs than households headed by individuals with more education. Although most of the other variables have the hypothesized signs, the results suggest that they are not significant predictors of households’ participation in public-funded food assistance programs in North Alabama. Thus, the results for a sample of low-income food insecure households in North Alabama reveals that being minority and having less education.
Table 3. Participation in Public Food Security Programs
Model 1: SNAP Model 2: WIC Model 3: SCHOOL
Coeff. SE Odds ratio Coeff. SE Odds ratio Coeff. SE Odds ratio
Constant -1.383** 0.645 — -1.680** 0.715 — -2.242*** 0.739 —
Ethnicity 0.797** 0.399 2.218 0.863* 0.457 2.371 1.022** 0.451 2.779
Gender 0.601 0.502 0.548 0.679 0.538 0.507 0.023 0.576 1.023
Education 0.238 0.382 1.269 0.438 0.426 1.549 1.043** 0.433 2.839
Marital status 0.047 0.405 1.048 0.092 0.447 1.096 -0.453 0.454 0.636
Income 0.212 0.388 1.236 0.474 0.422 1.607 -0.562 0.443 0.570
Household size 0.178 0.137 1.195 0.020 0.157 1.020 0.105 0.143 1.161
Pseudo R2 0.043 0.053 0.079 Count R2 68% 77% 75% Log likelihood -86.36 -73.44 -75.55 Observations 144 144 144 *,**,*** denotes significance at the 10%, 5% and 1% levels, respectively.
The results for the private/community-based food security programs (food banks, pantries and community gardens) are reported in Table 4. It is revealed that, contrary to the result observed in the public-funded food security models, ethnicity seems to play a different role in the private/community-based models. The estimated coefficients for ethnicity are statistically significant only in the food bank model (Model 4) and carries a negative sign. The expected odd ratio is equal to 0.29, suggesting that there is a 0.29 times chance that low-income households headed by African American will not utilize food banks to mitigate their food insecurity situation compared to their White counterparts. Furthermore, marital status is shown to be a significantly predictor of household participation across the three private/community-based food security programs. The expected odd ratio of marital status (being divorced) is equal to 3.2, 2.1 and 2.2 in Model 4 (Food Banks), Model 5 (Pantries) and Model 6 (Community Gardens), respectively. The results indicate that low-income households headed by someone divorced/separated have 3.2 times chance to utilize food banks, 2.1 times chance to utilize food pantries, and 2.2 times chance to utilize community gardens in addressing their food insecurity situations compared to married couple households. Also, household size appears to be a significant predictor of participation, but only in the food pantries model (Model 5). The expected odd ratio of household size is equal to 1.2, suggesting that, households in North Alabama with two or more have 1.2 times chance to utilize food pantries to address their food insecurity situation compared to households with less people. In support of this result, previous studies have indicated that most households visiting food pantries are small—with the average client household size of 2.7 persons (Cohen et al., 2006). The average household size for the current study is estimated at 2.4 persons.
Previous studies have showed that participation rate in food assistance programs is different by gender (Mabli et al., 2011; Martin et al., 2003; Gundersen et al., 2008). Generally, women are more likely to participate in programs than men and this finding was supported by the food bank model (Model 4). The expected odd ratio of gender is equal to 0.35, indicating that, female headed households have 0.35 times chance to utilize food banks to address household food insecurity situation compared to households headed by men. Income also significantly predicts participation, but only in the food bank model. The expected odd ratio of income is equal to 3.2, indicating that, low-income households have 3.2 times chance to utilize food banks to address household food insecurity situation compared to households with higher income. The overall results among the private/community-based food security programs highlight the role that food banks play in helping low-income households in dealing with food insecurity issues. Food banks, utilized by 27% of low-income food insecure households in north Alabama, are indeed an important food resource. Notably, many people who use public-funded programs also use food banks. Of the 12% households that received SNAP benefits, for example, 6.6% also used food banks.
Table 4. Participation in Community Food Security Programs
Model 4: FBANK Model 5: FPANTRY Model 6: CGARDEN
Coeff. SE Odds ratio Coeff. SE Odds ratio Coeff. SE Odds ratio
Constant -0.667 0.655 — -2.347*** 0.744 — -1.928*** 0.73 —
Ethnicity -1.235*** 0.442 0.291 -0.264 0.451 0.734 0.125 0.418 1.133
Gender 1.065*** 0.53 0.345 0.652 0.570 0.441 0.083 0.57 1.087
Education -0.405 0.414 0.667 0.429 0.432 1.689 -0.411 0.41 0.663
Marital status 1.161** 0.459 3.194 1.136*** 0.483 2.066 0.803* 0.435 2.233
Income 1.156*** 0.429 3.178 0.543 0.440 1.036 0.511 0.41 1.667
Household size 0.147 0.150 1.158 0.290* 0.153 1.240 0.097 0.147 1.101
Pseudo R2 0.123 0.075 0.038 Count R2 78% 72% 76% Log likelihood -73.81 -69.38 -76.82 Observations 144 144 144 *,**,*** denotes significance at the 10%, 5% and 1% levels, respectively.
Conclusion
A major aim of this paper was to identify factors that predicted low-income households’ participation in food assistance programs. Study results suggest that thus far, the primary responses to household food insecurity in North Alabama have been local-level food-based initiatives, predominantly food banks, but also community-based food pantries and programs such as community gardens aimed at enhancing food skills and food access. While it has long been recognized that such initiatives do not address the economic issues that underlie food insecurity, the perception that these programs play a valuable role in addressing the unmet food needs of food-insecure households persists and is supported by the current study. Our results highlight the need for systematic evaluations of community-based food initiatives to determine their relevance and accessibility for food-insecure households. The logit model results revealed that, participation in food assistance programs among low-income households in North Alabama was strongly related to predisposing and need factors, such as female headed households, living without spouse, low education levels and ethnicity, as well as enabling factors, such as low household income. These findings will be helpful in identifying segments of the population to be targeted for food assistance programs and thus increasing the effectiveness and target population penetration of these food assistance programs.
The study has some limitations that should be considered when drawing broader conclusions from the results. First, the sample size is relatively small for this type of analysis. Second, both predictor and outcome variables were based on self-reported conditions.
Acknowledgment
This research was supported by USDA/NIFA 1890 Capacity Building Program and by contributions from the Alabama A&M University, College of Agricultural, Life and Natural Sciences. The opinions expressed herein are those of the author.
References
Arcuri, S., G. Brunori, and F. Galli. (2016). “Insights on food assistance provision in High Income Countries: a systematic literature Review.” Paper for the IX ESPAnet Conference “Welfare models and Varieties of Capitalism. The challenges to the socio-economic development in Italy and Europe” (Macerata, 22-24 September 2016). https://www.researchgate.net/publication/308615301_Insights_on_food_assistance_provision_in_High_Income_Countries_a_systematic_literature_review
Alabama Food Bank Association. (2017). “Hunger in Alabama.” https://alfba.org/hunger-in-alabama/ Retrieved January 14, 2017.
Anderson, S.E. 1990. “Core Indicators of Nutritional State for Difficult-to-Sample Populations.” Journal of Nutrition 120: 1559-1599.
Anderson, M. D. (2013). “Beyond Food Security to Realizing Food Rights in the US.” Journal of Rural Studies 29:113-122.
Anderson, M.P., K.F. Butcher, H.W. Hoynes, and D.W. Schanzenbach. (2015). “Beyond Income: What Else Predicts Very Low Food Security among Children?” University of Kentucky Center for Poverty Research Discussion Paper Series DP2014-06. http://www.ukcpr.org/content/beyond-income-what-else-predicts-very-low-food-security-among-children. Retrieved February 12, 2018.
Bhattarai, G.R., P.A. Duffy, and J. Raymond. (2005). “Use of Food Pantries and Food Stamps in Low-Income Households in the United States.” Journal of Consumer Affairs 39 (2): 276–298.
Bickel, G., M. Nord, C. Price, W. Hamilton, and J. Cook. (2000). “Guide to Measuring Household Food Security.” http://www.fns.usda.gov/sites/default/files/FSGuide.pdf. Retrieved March 6, 2017.
Biggerstaff, M.A., P.M. Morris, and A. Nichols-Casebolt. (2002). “Living on the Edge: Examination of People Attending Food Pantries and Soup Kitchens.” Social Work 47(3): 267-277.
Booth, S., J. Whelan. (2014). “Hungry for Change: The Food Banking Industry in Australia.” British Food Journal 116(9):1392-1404.
Caraher, M. (2015). “The European Union Food Distribution program for the Most Deprived Persons of the community, 1987-2013: From agricultural policy to social inclusion policy?” Health Policy 119(7): 932-940.
Cohen, N., & Ilieva, R. T. (2015). Transitioning the food system: A strategic practice management approach for cities. Environmental Innovation and Societal Transitions, 17, 199-217.
Coleman-Jensen, A., M.P. Rabbitt, C. Gregory, and A. Singh. (2017). “Household Food Security in the United States in 2016.” ERR-215. U.S. Department of Agriculture, Economic Research Service, September 2017.6. https://www.ers.usda.gov/webdocs/publications/84973/err-237.pdf. Retrieved February 12, 2018.
Daponte, B. O., & Bade, S. L. (2006). The evolution, cost, and operation of the Private Food Assistance Network. Institute for Research on Poverty, University of Wisconsin-Madison.
Feeding America. (2017). “Food Security in Alabama.” http://map.feedingamerica.org/county/2014/overall/alabama. Retrieved February 12, 2018.
FRAC. (2017).Profile of Hunger, Poverty, and Federal Nutrition Programs http://www.frac.org/maps/sos/sos-map.html
Fuller-Thomson, E., and M. Redmond. (2008). “Falling Through the Social Safety Net: Food Stamp Use and Nonuse among Older Impoverished Americans. The Gerontologist 48(2): 235-244.
Garson, G.D. (2009). “Logistic Regression: Statnotes.” North Carolina State University. http://faculty.chass.ncsu.edu/garson/PA765/logistic.htm. Retrieved May 8, 2017.
Greene W.H. 2000. Econometric Analysis, Prentice Hall Inc., Upper Saddle River, New Jersey.
Greenwald, H.P., and V. Zajfen. (2017). “Food Insecurity and Food Resource Utilization in an Urban Immigrant Community.” Journal of Immigrant Minority Health 19: 179. https://doi.org/10.1007/s10903-015-0331-9. Retrieved January 22, 2018.
Gujarati, D. (2011). Econometrics by Example, Pelgrave Macmillan, New York.
Gundersen, C., J.P. Ziliak. (2008). “The Age Gradient in Food Stamp Program Participation: Does Income Volatility Matter?” In Income Volatility and Food Assistance in the United States (pp. 171-214). W.E. Upjohn Institute. https://experts.illinois.edu/en/publications/the-age-gradient-in-food-stamp-program-participation-does-income- Retrieved May 8, 2017.
Gundersen, C. (2013). Food insecurity is an ongoing national concern. Advances in Nutrition: An International Review Journal, 4(1), 36-41.Kiteley, R., & Stogdon, C. (2013). Literature Reviews in Social Work. Sage.
Lambie-Mumford, H., and E. Dowler. (2014). “Rising Use of ‘Food Aid’ in the United Kingdom.” British Food Journal, 116(9): 1418-1425.
Mabli, J, S. Tordella, L. Castner, T. Godfrey, and P. Foran. (2011). “Dynamics of Supplemental Nutrition Assistance Program Participation in the Mid-2000s.” U.S. Department of Agriculture, Food and Nutrition Service, Office of Research and Analysis. https://fns-prod.azureedge.net/sites/default/files/DynamicsMid2000.pdf. Retrieved May 8, 2017.
Martin, K.S., J.T. Cook, B.L. Rogers, H.M. Joseph. (2003). “Public versus Private Food Assistance: Barriers to Participation Differ by Age and Ethnicity.” Journal of Nutrition Education Behavior 35: 249–254.
Mayer, V.L., M.A. Bachhuber, and J. Long. (2014). “Food insecurity, Neighborhood Food Access, and Food Assistance in Philadelphia.” Journal of Urban Health 91(6):1087–97.
Patton-Lopez, M.M., D.F. Lopez-Cevallos, D.I. Cancel-Tirado, and L. Vasquez. (2014). “Prevalence and Correlates of Food Insecurity Among Students Attending a Midsize Rural University in Oregon.” Journal of Nutrition Education Behavior 46(3):209–14.
Popielarski, J.A., Cotugna, N. (2010), “Fighting Hunger through Innovation: Evaluation of a Food Bank’s Social Enterprise Venture, Journal of Hunger ; Environmental Nutrition, 5(1):56-69.
Poppendieck, J. (1994). “Dilemmas of Emergency Food: A Guide for the Perplexed.” Agriculture and Human Values 11(4):69–76.
Ratcliffe, C., S. McKernan, and S. Zhang. (2011). “How Much Does the Supplemental Nutrition Assistance Program Reduce Food Insecurity?” American Journal of Agriculture Economics 93 (4):1082–98.
Riches, G. (2011). “Thinking and Acting Outside the Charitable Food Box: Hunger and the Right to Food in Rich Societies.” Development in Practice 21(4–5):768–75.
Riches, G., and V. Tarasuk. (2014). “Canada: Thirty Years of Food Charity and Public Policy Neglect.” In First World Hunger Revisited (pp. 42-56). Palgrave Macmillan UK.
Robaina, K.A., and K.S. Martin. (2013). “Food Insecurity, Poor Diet Quality, and Obesity among Food Pantry Participants in Hartford, CT.” Journal of Nutrition Education Behavior 45(2):159–64.
Smith, T.M., F.M.W. Bertmann, C.A. Pinard, D.J. Schober, K. Shuval, B.T. Nguyen, H.E. Fricke, and A.L. Yaroch. (2016). “Factors Associated With Supplemental Nutrition Assistance Program Participation among the Working Poor: Findings from 2012 American Community Survey.” Journal of Hunger and Environmental Nutrition 12(2): 1-12.
Tarasuk, V., N. Dachner, and R. Loopstra. (2014). “Food banks, welfare, and food insecurity in Canada.” British Food Journal 116(9): 1405-1417.
U.S. Department of Agriculture, Food and Nutrition Service, Office of Research, Nutrition and Analysis, Dynamics of Food Stamp Program Participation, 2001-2003, by Scott Cody, et al. Project Officer: Christine Kissmer, Alexandria, VA: 2007. https://fns-prod.azureedge.net/sites/default/files/Dynamics2001-2003.pdf. Retrieved May 3, 2017.
USDA, Economic Research Service. (2017). “State-Level Prevalence of Food Insecurity.” https://www.ers.usda.gov/topics/food-nutrition-assistance/food-security-in-the-us/key-statistics-graphics.aspx#map. Retrieved January 13, 2018.
U.S. Census Bureau, Current Population Survey. (2017). Income and Poverty in the United States: 2016. https://www.census.gov/topics/income-poverty/data/tables.html. Retrieved January 13, 2018.
Wakefield, S., J. Fleming, C. Klassen, and A. Skinner. (2013). “Sweet Charity, Revisited: Organizational Responses to Food Insecurity in Hamilton and Toronto, Canada.” Critical Social Policy 33(3):427–50.
Webb, K.L. (2013). “Introduction–Food Banks of the Future: Organizations Dedicated to Improving Food Security and Protecting the Health of the People They Serve.” Journal of Hunger and Environmental Nutrition 8(3):257–60.
Wunderlich, G.S., and J.L. Norwood. (2006). Food insecurity and hunger in the United States: An Assessment of the Measure. Washington, DC: The National Academies Press.
Yu, M., M. Lombe, and V.E. Nebbitt. (2010). “Food Stamp Program Participation, Informal Supports, Household Food Security and Child Food Security: A Comparison of African American and Caucasian Households in Poverty.” Children and Youth Service Review 32: 767-773.
