Welfare Spending and Poverty Rates
Chart 1: Poverty Rates and Spending
What is Poverty?
Agreement on defining the term is difficult. Stark (2009) says that poverty should be addressed by looking at several different theories. One concept views poverty as, “a moral obligation that is grounded in compassion.” Stark defines the rights of the poor, and explains poverty through utility and liberal theories. He notes that income measures do not adequately capture the term because, “Most poor Americans have access to clean drinking water and electricity, unlike most of the global poor.”
O’Boyle (1990) questions whether the term should be defined in absolute or relative terms. He notes that it is a concept that is, “both absolute and relative because human beings are at once individual and social.” He ultimately defines it as, “a lack of economic resources to meet physical need” (p. 2).
Poverty for U.S. citizens is defined as described by the guidelines set forth by the Federal Government. For example, the calculation of the Supplemental Poverty Measure (SPM) thresholds is to be based on expenditures for food, clothing, shelter, and utilities (FCSU), as well as the value of in-kind government benefits for FCSU that are accounted for in resources by the Census Bureau (Garner, 2011).
There is considerable debate on the issue. In reviewing the relationship between welfare spending per capita and poverty, some studies show evidence to support that increased welfare spending decreases poverty rates. Brady, Fullerton & Cross (2009) found that, “For each standard deviation increase in welfare generosity, the odds of poverty decline by a factor of 2.3.” Schram (1990) showed evidence that more spending results in less poverty. Fording & Berry (2007) showed a decrease in the poverty rates with increased cash benefit.
Yet these same authors that suggested that welfare spending decreased poverty rates acknowledged the opposite finding. Authors had to concede that they found cases of increased poverty with increased spending. Brady, Fullerton & Cross (2009) acknowledged that, “The welfare state is not a panacea, as there is mixed evidence that generous welfare states might worsen the standing of young households.” Schram (1990) noted that poverty rates have increased while welfare spending has, “remained at high levels” (p. 135). Schram addressed the issue of welfare dependency and the “personal responsibility of the poor.” He noted families who would rather take a job once they become self-sufficient. He states that the findings are, “not meant to suggest that welfare spending produces less poverty.” Fording & Berry (2007) showed that if the cash benefit rate is fixed, poverty rates increase. They attribute this to a “work disincentive effect” (p. 49). These previous studies suggest that justification for increased welfare spending is at least questionable. Other approaches should be examined.
Methodology and Research Question
This analysis uses secondary data on poverty rates and welfare spending from the U.S. Census Bureau. The Census Bureau is the most comprehensive source for poverty rates in the nation. Data is collected from the entire U.S. population. All ages and socio-economic conditions are sampled.
A poverty study by the North Dakota State Data Center supports the validity and value of Census Bureau data (Rathge & Olson, 2006). The reliability of the data is expected to be high. However, the Census Bureau notes that it is subject to non-sampling errors (U.S. Census Bureau, 2007). This report accepts the data as reliable and valid.
The period of the study is from 1959 through 2009. Fifty one observations for both variables were analyzed. A second, smaller, time period of the last ten years of data was also analyzed. The study did not measure the influence of market conditions, unemployment rates, educational levels and unskilled wages of those living in poverty.
The study assumed that welfare spending is increased during times of economic decline and increased unemployment. The study also assumed that the educational levels and unskilled wages of those in poverty were relatively equal but acknowledge that future studies need to control for these possible influences.
The data is analyzed quantitatively using linear regression to answer the research question, “Does increased welfare spending reduce poverty rates?”
Welfare spending per capita and poverty rates are displayed above in Chart 1. The first ten years of the display show a rise in welfare spending and a sharp decline in the poverty rate. The last ten years of the display show increased spending and an increase in the poverty rate. An overall observation is that the line for welfare spending rises over time and the poverty rate did not deviate by more than 3% from the mean since 1970.
Descriptive statistics are found in Table 1 below. The mean poverty rate is 14.11%. The mean number of dollars spent per person on welfare is $365.01. The standard error of the mean for welfare spending is large as the amount of spending increased significantly each year. The standard error for poverty rates is low revealing little fluctuation.
Table 1: Descriptive Statistics
Linear regression and correlation analysis was conducted on the entire 51 year period as well as for the last 10 years of the study. Results are shown in Table 2. A review of the bivariate linear regression and correlation data did not reveal any outliers.
The standard error of the mean for this model is 2.73, suggesting the sample represents the study. Welfare spending and the poverty rate are significantly correlated at 0.33 (p <.02). This correlation indicates that 33% of the influence in welfare spending is explained by the poverty rate. Likewise 33% of the influence of the poverty rate is shown by welfare spending.
The correlation analysis shows a negative coefficient. A negative value indicates a relationship such that as welfare spending increases, poverty rates decrease. The influence or r square value between these variables is 0.11 meaning that 11% of the variance of poverty rates is influenced by welfare spending per capita.
The correlation between the relationship of welfare spending per capita and the poverty rate support the research question. However, this research is limited to one time period and only one independent variable. Other factors, such as unemployment rate, educational levels, unskilled wages and co-participation by charitable organizations should be investigated in future analysis.
What Do You Think?
Do you think evidence suggests that increasing welfare spending reduces poverty?
Table 2: Linear Regression and Correlation Statistics
The negative correlation between the relationship of welfare spending per capita and the poverty rate using fifty-one observations support the research question that increased welfare spending decreases poverty rates. However, the correlation coefficient is 0.33. This is a weak correlation. A correlation value greater than 0.8 is generally described as strong, whereas a correlation less than 0.5 is generally described as weak (Roberts, 2011). Correlation values are intended to reveal what variance or influence the two variables have on each other. It is difficult to know whether welfare spending was initiated as a reaction to the poverty rate or if the poverty rate was influenced by welfare spending.
The analysis examines fifty-one years of welfare spending. Because significant government spending on welfare did not begin until the late 1950s, the analysis might be described as a snapshot of the data. Earlier Census Bureau data on spending is reported every ten years rather than annually.
Future analysis could also consider investigating the time period from the year 2000 through 2029 so as to analyze at least thirty observations. This should allow us to make generalizations and inferences more applicable to the time frame. Further, future testing should consider the alternative hypothesis: As welfare spending increases, poverty rates increase.
Serious deliberation and consideration needs to be given to U.S. Government welfare spending. In particular, lawmakers need to make informed decisions based on the research data. Does increased welfare spending per capita decrease poverty rates?
Bivariate linear regression of U.S. welfare spending per capita and poverty rates since 1959 showed a negative correlation. This suggests that as welfare spending increases, poverty decreases.
However, linear regression using ten observations showed a positive correlation suggesting that poverty rates increased as welfare spending increased. This data shows that current poverty rates are at one of the highest levels since 1959 and that welfare spending per capita is at an all time high. Yet, the overall trend line of the graph supports the research question that as welfare spending is increased, poverty rates decrease.
Data was analyzed for the entire United States. Therefore the findings are more applicable at the Federal level. Narrowing the findings to a particular state or municipality might be less applicable. The analysis fails to account for unique circumstances such as affluence or unemployment rate of one city compared to another city located in the same or another state of a similar population. Therefore it would be difficult to generalize the national findings to a specific city. We would anticipate that the findings become more applicable as we move from the state level to regional and finally to the national level. Nonetheless, State and Federal lawmakers should consider the findings when addressing welfare legislation.
This study only looked at one independent variable. Other variables such as the unemployment rate, educational levels and unskilled wages should be examined. Another limitation is being able to make generalizations from data that is more than ten years old. Since the analysis included data that is over fifty years old, extrapolating the findings into the current time period becomes less applicable and therefore less desirable.
Public Administrators will have to look at other root causes of poverty and make a decision about how to address it. Evidence suggests that simply addressing poverty by increased spending is only part of the solution. Other factors such as the educational and skill levels of the impoverished should be examined. Studies have suggested that increasing educational levels of those with lower income can reduce poverty. It has also been suggested that increasing the skill levels of individuals is a factor in reducing poverty. Future studies should include education and employment rates as additional independent variables for analysis. The influence of faith-based institutions participating in welfare aid should also be examined. Recalling Brady, Fullerton & Cross (2009) this report found “mixed evidence” about the relationship between welfare spending and the poverty rates. The findings suggest that increased welfare spending reduced poverty rates more significantly in the first half of the time period studied as compared to the last ten years of the analysis.
Brady, D., Fullerton, A., & Cross, J.(2009). Putting poverty in political context: a multi-level analysis of adult poverty across 18 affluent democracies. Social Forces. 88(1): 271-300. University of North Carolina Press.
Chantrill, C. (2011). U.S. welfare spending chart. Retrieved from: http://www.usgovernmentspending.com/
Fording, R.C., & Berry, W.D. (2007). The historical impact of welfare programs on poverty: Evidence from the American states. The Policies Study Journal. 35(1): 37-60.
Garner, T. I. (2011). Supplemental poverty measure thresholds: Laying the foundation. Bureau of Labor Statistics. 0-33. Retrieved from: http://www.census.gov/hhes/povmeas/methodology/supplemental/research/ASSAGarner%20Poverty%20Thresholds%20paper%2012-29-10.pdf
O'Boyle, E. J.(1990). Poverty: A concept that is both absolute and relative because human beings are at once individual and social. Review of Social Economy. 48(1): 2-17.
Rathge, R. & Olson, K. (2006). North Dakota case study: Perceived income and poverty contradiction. North Dakota State University Data Center. Retrieved from: http://www.ndsu.nodak.edu/sdc/toolbox/NDSDC_CaseStudyOne.pdf
Roberts, F. & Roberts, D. (2011). Statistics 2: How well does the regression equation truly represent the set of data? Retrieved from: http://mathbits.com/MathBits/TISection/Statistics2/correlation.htm
Schram, S. F. (1991). Welfare spending and poverty: Cutting back produces more poverty, not less. The American Journal for Economics and Sociology. 50(2): 129-141.
Stark, Barbara. (2009). Theories of poverty/the poverty of theory. Brigham Young University Law Review. No. 381-425, 2009; Hofstra Univ. Legal Studies Research Paper No. 10-20.
U.S. Census Bureau (2007). 2007 economic census: Reliability of data. Retrieved from: http://www.census.gov/econ/census07/www/methodology/reliability_of_data.html
U.S. Census Bureau. (2011). Small area income and poverty estimates: state and county interactive tables. Retrieved from: http://www.census.gov//did/www/saipe/county.html
United States Federal State and Local Government Spending (2011). Welfare. Retrieved from: http://www.usgovernmentspending.com/numbers