Citation
Socioeconomic Determinants of Veteran Homelessness

Material Information

Title:
Socioeconomic Determinants of Veteran Homelessness
Creator:
Ramgoolie, Adita Mychaela
Publication Date:
Language:
English

Notes

Abstract:
For many, veterans are those individuals who have selflessly sacrificed for the freedom and benefit of our country. Officially, a veteran is classified as "a person who served in the active military, naval, or air service and who was discharged or released under conditions other than dishonorable." This paper delves into several socioeconomic factors that influence the homelessness rates amongst veterans, by state, in the United States. Analyzing three models that utilized multiple regression, six variables were explored between these models to ascertain their degrees of statistical significance on veteran homelessness rates, if any. One particular variable was held constant amongst each model, Veteran Suicide Rate by State, and was found to be statistically significant in each, by at least the ninety percent confidence level. The findings of this study shine a spotlight on several determinants that may influence the future of veteran homelessness through spurring the notion of further research of related factors that may affect future policy decisions directed at combating and preventing homelessness amongst veterans in the United States. ( en )
General Note:
Awarded Bachelor of Arts, magna cum laude, on May 8, 2018. Major: Economics
General Note:
College or School: College of Liberal Arts and Sciences
General Note:
Advisor: Michelle Andrea Phillips, Ph.D.. Advisor Department or School: Economics

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Adita Mychaela Ramgoolie. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.

UFDC Membership

Aggregations:
UF Undergraduate Honors Theses

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

Socioeconomic Determinants of Veteran Homelessness Adita Ramgoolie April 9, 2018 Thesis Advisor: Dr. Phillips ECO 4935

PAGE 2

Ramgoolie 1 I. INTRODUCTION For many, veterans are those individuals who have selflessly sacrificed for the freedom and benefit of our country. While many may agree, a more official definition by Title 38 of the he active military, naval, or air service and who was discharged or released under conditions other than Moreover, homelessness is defined as the state of not having a permanent dwelling, or by which one is living in an area designated as an emergency shelter, an area not meant for human habitation, or transitional housing. Homelessness may be an indicator for the overall economic health of society, often being tied to socioeconomic variables such as poverty levels, unemployment rates, inco me inequality, and/or substance abuse Research shows that Homeless Families Climbing Due t Approxi mately 11% of the adult homeless population are veterans, according to the National Coalition for Homeless Veterans (NCHV). This is interesting to note, as data from the U.S. Census B ureau illustrates that only 6% of the United States population is made u p of veterans. The NCHV details War, Cold War, Vietnam War, Grenada, Panama, Lebanon, Persian Gulf War, Afghanistan and drug cultivation effort s in South America. As stated by thirds served our country for at least three years, and one With this in mind, this study will delve into several socioeconomic factors and how they correlate with the differences in the percentages of homeless veterans between states in the United States.

PAGE 3

Ramgoolie 2 II. SAMPLE There are forty nine observations in my sample, each corresponding to the fifty states of the United States of America excluding Hawaii Hawaii was found to be an outlier, and was thus excluded. The years that correspond with the time th e data were recorded is between the ye ars of 2015 2017 and is explained explicitly for each variable below. III. DEPENDENT VARIABLE Veteran Homelessness rate : (Vet_HR) The dependent variable for this study is the percentage of homeless veterans, including both sheltered and unsheltered, amongst states. This is calculated by taking the number of a for this variable is obtained from 2016 Point in time (PIT) Estimate of Homeless Veterans by state from the U.S. Department of Housing and Urban Development (HUD). IV. INDEPENDENT VARIABLES 1 Unemployment Rate of Veterans by State : (Vet_UR) To understa nd the differences in veteran homelessness amongst states, this study will be utilizing the unemployment rate of veterans (18 years and over), by state, from 2016 annual averages by the Bureau of Labor Statistics from the United States Department of Labor. Unemployment rate reflects the health of the economy to a certain degree. With higher levels of unemployment, it is expected that consumer consumption decreases while prolonged 1 Other variables considered but ultimately left out of this study were Gini Index of Income Inequality, Religious Affiliation Rate of Veterans by State, and Estimated Expenditures for Public Schools 2015 2016. The data associated with these variables were not properly representative and/or accurate for the sample.

PAGE 4

Ramgoolie 3 unemployment may impact an erosion of useful talents or skills. More so, pr olonged unemployment leads toward lower overall income and the lack of ability to pay for an education or training required to re enter the workforce. This only exacerbates the difficulties of affording a home or shelter, which segues into higher homelessness rates. VA Expenditures : (VA_Expend.) Veterans' benefits refer to a wide range of remunerations, including monthly life insurance, disability checks, health care, home loans, and education through the GI bill, among others. Total 2016 state level Veterans Affairs (VA) expenditures will be included in this study, and have been collected by the National Center for Veterans Analysis and Statistics, Department of Veterans Af fairs. The expenditures for each state will be divided by the number of veterans in that state to account for differences in Studies show that higher awareness of a particular negative issue, helps to reduce its presence. While there are many programs and charities aimed toward helping veterans, whether it be with physical and mental issues or economic difficulties, there is not any single organization that is as inclusive to aiding all types of veterans in some way as gov ernment expenditures are. For instance, the Wounded Warrior Project is an organization that provides free programs and services to address the mental and physical health needs of wounded warriors, along with career and benefits counseling. However, the pro veterans and service members who incurred a physical or mental injury, illness, or wound, co incident to their military service on or would not be representative of the total care given to all veterans, specifically those who served before September 11, 2001. Hence, in this study awareness is reflected in the ratio between total eran population. The data includes total expenditure by each state, breaking each figure down between allocation towards compensation

PAGE 5

Ramgoolie 4 and pension, construction, education and vocational employment, loan guaranty, general operating expenses, insurance, medi cal care, and unique patients. Thus, this variable helps to homelessness levels. It is hence theorized that, the higher the expenditures, the lower the rate of homelessness within that state. Percent of Veteran Households with C hildren, per S tate : (Vet_Child) This data was collected by the National Center for Veterans Analysis and Statistics from the U.S. Department of Veterans Affairs and details the 2017 perce ntage of veteran households that have children in each state. i neurologically altered once they have children. The article says that areas such as the hypothalamus, thalamus, and prefrontal cortex are altered in ways that: increase the production of them. Hence, this variable can be sai d to be a motivator for individuals to not only care for themselves, but for their offspring, which reduces their chances of becoming homeless, due to the added responsibility and socio biological change. Veteran Suicide Rate by State : (Vet_SuicideR) This data was collected from the Department of Veterans Affairs, from a report released in 2016 on the rate of veteran suicides in each state. It is used as a proxy for substance abuse and mental illness, which are closely correlated to suicide rates. Based up on information form the

PAGE 6

Ramgoolie 5 urban areas; and suffer from mental illness, alcohol and/or substance abuse, or co occurring Thus, I hypothesize that as veteran suicide rates increase, so too would their homelessness rates. Overall State Unemployment Rate : (State_UR) The overall unemployment rate for a given state may influence the ability of a veteran to find work. Therefore, this study uses data collected from the Bureau of Labor Statistics from the United States Department of Labor for the year 2016. The unemploymen t rate is reflective of each with the veteran unemployment rate, I predict that higher levels of state unemployment will correlate with higher homelessness rates as the more difficult it is to be employed, the more difficult it is to have a home, or place of reliable, permanent shelter. State Real GDP per Capita : (RGDP_PerC) This study utilizes data collected from the United States Bureau of Economic Analysis for FocusEconomics, a leading provider of economic analysis and forecasts an important indicator of economic performance and a useful unit t o make cross county comparisons of average living standards and economic wellbeing." Henceforth, I surmise that as the real GDP per capita of each state increases, homelessness rates of veterans would decrease, as higher standards of living, as represented by the GDP per capita, would indicate a lower chance of homelessness rates in that state.

PAGE 7

Ramgoolie 6 V. SUMMARY STATISTICS AND CORRELATION MATRIX Vet_HR Vet_UR Vet_SuicideR VA_Expend Vet_Child State_UR RGDP_PerC Mean 0.1669 4.07 41.81 802.52 30.7 4.68 48519.5 Standard Error 0.0126 0.158 1.54 18.36 0.436 0.139 1266.5 Median 0.1427 4 39.2 806.86 30.878 4.8 47633 Mode #N/A 3.57 37.8 #N/A #N/A 4.8 #N/A Standard Deviation 0.0887 1.104 10.77 128.5 3.052 0.9696 8865.2 Sample Variance 0.0079 1.218 116.02 16522.5 9.31 0.9401 78592626.01 Kurtosis 5.27 0.143 0.328 2.44 0.05 0.468 0.778 Skewness 2.052 0.024 0.265 0.817 0.4595 0.0396 0.33 Range 0.457 5.039 47 685.24 13.64 4 33179 Minimum 0.0659 1.59 21.6 588.6 25.3 2.9 32102 Maximum 0.523 6.63 68.6 1273.88 38.94 6.9 65281 Sum 8.18 199.4 2048.6 39323.3 1506.5 229.2 2377455 Count 49 49 49 49 49 49 49 Variable Definitions : Vet_HR: Percent Veteran Homelessness Rate: (#Homeless veterans/ State veteran population)*100 Vet_UR : Percent Veteran Unemployment Rate Vet_SuicideR : Percent Veteran Suicide Rates by state VA_Expend: VA Expenditures per 100 veterans per state. Vet_Child: Percent Veteran households with children: (Veteran households with ch ildren/State veteran population )*100 State_UR: Percent State Unemployment Rate RGDP_PerC: Real GDP per Capita by State _____________________________________________________________________________________ Correlation Matrix: Vet_HR Vet_UR Vet_SuicideR Vet_Child VA_Expend State_UR RGDP_PerC Vet_HR 1 Vet_U R 0.101 1 Vet_SuicideR 0.228 0.252 1 Vet_Child 0.196 0.051 0.006 1 VA_Expend 0.038 0.089 0.335 0.327 1 State_UR 0.042 0.254 0.015 0.026 0.293 1 RGDP_PerC 0.245 0.09 0.408 0.089 0.389 0.111 1

PAGE 8

Ramgoolie 7 The correlation matrix shows that each of the variables generally have a low correlation with one another. Several variables show a slight, but not significant, negative correlation with one another, like the veteran unemployment rate and veteran suicide rate. Overall, the matrix shows low correlation values between the variables V I REGRESSION AND RESULTS I decided to run three models, with three independent variables, instead of one with six independent variables, to directly see the influence different independent variables would have in each scenario, if any In all three models Veteran Suicide Rate is h eld constant. Model 1: Model one includes the independent variables: Veteran Unemployment Rate, Veteran Suicide Rate, and VA Expenditures. It excludes the variable s : Percent of Veteran Hous eholds with Children, per state Overall State Unemployment Rate a nd State Real GDP per Capita Regression Statistics Multiple R 0.29 R Square 0.084 Adjusted R Square 0.023 Standard Error 0.088 Observations 49 ANOVA df SS MS F Significance F Regression 3 0.032 0.010609 1.381 0.261 Residual 45 0.35 0.007681 Total 48 0.38

PAGE 9

Ramgoolie 8 Coefficients Standard Error t Stat P value Lower 95% Upper 95% Intercept 0.047 0.097 0.486 0.629 0.148 0.24 Vet_UR 0.0148 0.0121 1.226 0.226 0.0095 0.039 Vet_SuicideR 0.00248 0.0013 1.897 0.064 0.00015 0.00511 VA_Expend 5.5E 05 0.0001 0.516 0.609 0.00027 0.00016 RESULTS: Unemployment Rate of Veterans by State With a p value of 0.2 26 which is greater than 0.05 and .1 the null hypothesis that there is no linear relationship between veteran unemployment rate and the veteran homelessness rate cannot be rejected. Thus, this variable can be considered not statistically significant Veteran Suicide Rate by State At the 90 % confidence level, the null hypothesis that there is no linear relationship between veteran suicide rate by state and veteran homelessness can be reje cted, as the p value is 0.064 which is less than 0.1. Thus, this variable can be considered statisticall y significant. The coefficient is 0.00248 therefore a one percentage point increase in the suicide rate is associated with an increase in the homelessness rate by about 0.0025. This supports the initial hypothesis that an increase in suicide rates will correlate with an increase in veteran homelessness rates. This may be due to the fact that the suicide rate amongst veterans acts as a proxy for mental illness and/or substance abuse, which are two leading causes of homelessness amongst veterans, according to the National Coalition for Homeless Veterans. VA Expenditures This va riable can also be considered not statistically sig nifi cant. With a p value of 0.609 which is much greater than 0.05 or 0.1 the null hypothesis that there is no linear relationship between VA expenditures and the veteran homelessness rate cannot be rejected.

PAGE 10

Ramgoolie 9 Model 2: Model two includes the independent variables: Veteran Unemployment Rate, Veteran Suicide Rate, and Percent of Veteran Households with Children, per state. It excludes the variable s : Overall State Unemployment Rate State Real GDP per Capita and VA Expenditures. 2 Regression Statistics Multiple R 0.337 R Square 0.114 Adjusted R Square 0.0545 Standard Error 0.086 Observations 49 ANOVA df SS MS F Significance F Regression 3 0.0429 0.014 1.92 0.1395 Residual 45 0.335 0.0074 Total 48 0.377 Coefficients Standard Error t Stat P value Lower 95% Upper 95% Intercept 0.1897 0.151 1.26 0.215 0.114 0.493 Vet_U R 0.0128 0.0117 1.094 0.28 0.0101 0.036 Vet_SuicideR 0.002198 0.00119 1.84 0.072 0.000207 0.0046 Vet_Child 0.00542 0.00408 1.327 0.191 0.014 0.00281 RESULTS: Unemployment Rate of Veterans by State With a p value of 0. 28 which is greater than 0.05, the null hypothesis that there is no linear relationship between veteran unemployment rate and the veteran homelessness rate cannot be rejected. Thus, this variable can be considered not statistically significant 2 Model 2 exhibits a slightly higher R 2 2 of .08.

PAGE 11

Ramgoolie 10 Veteran Suicide Rate by State Just as in the first model, the variable of veteran suicide rate by state can be considered statistically significant at the 90% confidence level, as the p value is 0.07 2 which is less than 0.1. Therefore, the null hypothesis that there is no linear relationship between veteran suicide rate by state and veteran homelessness can be rejected with 90% confidence The coefficient is 0.002198 therefore a one percentage point increase in the suicide rate is associated with an increase in the h omelessness rate by about 0.0022 This again supports the initial hypothesis that an increase in suicide rates will correlat e with an increase in veteran homelessness rates The U.S. Department of Veterans Affairs has a specific section known as the National Center for PTSD: Posttraumatic Stress Disorder, which is an ailment closely associated with increases in suicide rates. T he Percent of veteran households with children, per state This variable is also not significant at either the 90% or 95% confidence level. With a p value of 0.191 which is greater than 0.1 or 0.05, the null hypothesis that there is no linear relationship between the percent of veteran households with children an d the veteran homelessness rate cannot be rejected. Model 3 : Model three includes the independent variables: Overall State Unemployment Rate, Veteran Suicide Rate, and State Real GDP per Capita It excludes the variable s : Veteran Unemployment Rate, Veteran Households with Children, per state and VA Expenditures.

PAGE 12

Ramgoolie 11 Regression Statistics Multiple R 0.44 R Square 0.198 Adjusted R Square 0.144 Standard Error 0.082 Observations 49 ANOVA df SS MS F Significance F Regression 3 0.075 0.025 3.69 0.018 Residual 45 0.303 0.007 Total 48 0.377 Coefficients Standard Error t Stat P value Lower 95% Upper 95% Intercept 0.21 0.12 1.72 0.093 0.47 0.037 State_UR 0.0086 0.012 0.699 0.488 0.016 0.033 Vet_SuicideR 0.0033 0.0012 2.73 0.009 0.00086 0.0057 RGDP_PerC 4.19E 06 0.0000015 2.84 0.0068 1.21E 06 7.16E 06 RESULTS: Overall State Unemployment Rate This variable cannot be considered significant at either t he 90% or 95% confidence level, p value of 0. 488 which is greater than 0.1 or 0.05, means that the null hypothesis that there is no linear relationship between and the veteran homelessness rate cannot be rejected. Veteran Su icide Rate At the 95 % confidence level, the null hypothesis that there is no linear relationship between veteran suicide rate by state and veteran homelessness can be reje cted, as the p value is 0.009, which is less than 0.05 This variable can once again be considered statistically significant, but is unlike models one and two, where the veteran suicide rate was only statistically significant at the 90% confidence level. The coefficient is 0.0033 therefore a one percentage point increase in

PAGE 13

Ramgoolie 12 the suicide rate is associated with an increase in the h omelessness rate by about 0.003 This coincides, once more with the initial hypothesis that an increase in suicide rates will correlate with an increase in veteran homelessness rates This variable was used as a proxy for substance abuse, which is often a cause of homelessness for all people, not just veterans. As explained by Addictive disorders disrupt relationships with family and friends and often cause people to lose their jobs. For people who are already struggling to pay their bills, the onset or exacerbation of an addiction may cause them to lose their housing. State Real GDP per Capita With a p value of 0.0068 which is less than .05, the null hypothesis that there is no linear relationship between state real GDP per capita and veteran homelessness can be reje cted with 95% confidence. Unlike my initial hypothesis that as state real GDP per capita increases, veteran homelessness would decrease, the data shows otherwise. With a coefficient of 4.19E 06 a one pe rcentage point increase in the real GDP per capita, per state, is associated with an increase in the h omelessness rate by about 0.000004. The effect is miniscule and may be explained by external factors, such as how GDP does not take into effect income distribution or policies that may influence economic standards of living and homelessness VII. CONCLUSION : All models exhibit the independent variable: veteran suicide rates per state, as being statistically significant with models one and two with 90% confidence and model three at 95% confidence This is not much of a surprise, as veteran suicide rates are high ly correlated with people who suffer from mental illnesses such as PTSD. The National Center for PTSD states that Seeing as this variable was used as a proxy f or drug and alcohol abuse, it is significant to

PAGE 14

Ramgoolie 13 point out that various research explains that drug and alcohol abuse is significantly higher amongst homeless people. According to the National Coalition for the Homeless, cities reported that one of the top items needed to combat homelessness (for the entire population, not just veterans), are additional substance abuse services However, the report continued to point out that substance abuse is both a cause and a result of homelessness, and thus both issues need to be addressed concurrently. Moreover, i t is worth noting that the sample size of this study is relatively small, and that further research that includes a larger sample size spanning many years may influence certain values w hen a regression is run. It would be interesting to delve into data from around the late 1980s that would include a different era of veterans in a time when data on substance abuse and mental illness were now being released Unfortunately, after much inves tigation, it seems that much of the data around this time is on a national level, versus on a per state level. Regardless this study may be further developed using data around the mid to late 1990s, after the first Gulf War from 1990 91. It is compelling to note th at in the first two models, the Veteran Unemployment Rate, (Vet_UR), was almost sta tistically significant at the 75% and 80% confidence levels, while the Percent of veteran households with children, per state, ( Vet_Child), was st atistically sign ificant at the 8 0% confidence level. While these findings have a great deal of variation not explained by these variables, with a larger sample, it would be intriguing to see what happens to these variables p values. The independent variables used in this study, whether or not they have been found to be statistically significant, are nonetheless important in analyzing the overall determinants of veteran homelessness, and the policies or programs that may be enacted or modified to affect these people. More so, it is encouraging that recent reports from the U.S. Department of Housing and Urban Development (HUD), U.S. Department of Veterans Affairs (VA), and the U.S. Interagency

PAGE 15

Ramgoolie 14 Council on Homelessness (USICH), show a significant decrease in homelessness amon gst and January 2016 quadruple the previous year's annual decline and a 47 percent decrease lessness for future research VII I DATA SOURCES/REFERENCES Works Cited Baylor University. "Faith Based Organizations Shoulder Majority of Crucial Services and Develop Creative Solutions for Homelessness, New Baylor University Study Says." Media Communications | Baylor University Baylor University, 01 Feb. 2017. Web. Bureau, United States Census Bureau 18 Apr. 2015, www.census.gov/search results.html? Data Access and Dissemination Systems (DADS). "2016 American Community Survey 1 Year Estimates." American FactFinder Results U.S. Depart ment of Commerce, 05 Oct. 2010. Web. Dr. Bill. "Mommy Brain & Daddy Brain: How Parenthood Changes Your Brain." Ask Dr Sears AskDrSears, 17 May 2016. Web. FocusEconomics | Economic Forecasts from the World's Leadin g Economists FocusEconomics, 2018, www.focus economics.com/economic indicator/gdp per capita. "Geographic Distribution of VA Expenditures for Fiscal Year 2016, the National Center for Veterans Analysis and Statistics, Department of Veterans Affairs" Homeless Veterans, U.S. Department of Veterans Affairs, 26 Jan. 2010, www.va.gov/homeless/chaleng.asp.

PAGE 16

Ramgoolie 15 Malinen, Tuomas. "The Economic Consequences of Income Inequality." The Huffington Po st TheHuffingtonPost.com, 17 Dec. 2015. Web. Homeless, July 2009, pp. 1 3., www.nationalhomeless.org/factsheets/addiction.pdf. National Coalition for Homeless Veterans NCHV, 2016, nchv.org/index.php/news/media/background_and_statistics/. NEA. "Rankings of the States 2015 and Estimates of School Statistics 2016." National Education Association (2016): n. pag. National Education Association National Education Association, 2002 2017. Web. Sard, Barbara. "Number of Homeless Families Climbing Due To Recession." Center on Budget and Policy Priorities Center on Budget and Policy Priorities, 10 June 2015. Web. 2017 ANNUAL AVERAGES. Bureau of Labor Statistics U.S. Department of Labor, 27 Feb. 2018, 10 AM, www.bls.gov/news.release/pdf/srgune.pdf. U.S. Bureau of Labor Statistics. "Table 6A. Employment Status of Veterans 18 Years and over by State, 2016 Annual Averages." U.S. Bureau of Labor Statistics U.S. Bureau of Labor Statistics, 22 Mar. 2017. Web. Analy U.S. Bureau of Economic Analysis (BEA) U.S. Department of Commerce, Nov. 2017, www.bea.gov/itable/iTable.cfm?ReqID=70&step=1#reqid=70&step=10&isuri=1&7003=1 000&7035= 1&7004=naics&7005=1&7006=xx&7036= 1&7001=11000&7002=1&7090=70&7007=2016&7093=levels

PAGE 17

Ramgoolie 16 Housing and Urban Development (HUD), U.S. Department of Housing and Urban Development, 1 Aug. 201 6, www.hud.gov/press/press_releases_media_advisories/2016/HUDNo_16 117. U.S. Department of Vete Veterans PTSD: National Center for PTSD, U.S. Department of Veterans Affairs, 23 De c. 2011, www.ptsd.va.gov/public/problems/ptsd_substance_abuse_veterans.asp. U.S. Department of Veterans Affairs. Veteran Households with Children. 4 Apr. 2017. Raw data. N.p. U.S. Department of Veterans Affairs. Veterans Religious Affiliation by State. 18 Aug. 2017. Raw data. N.p. Wounded Warrior Project 2017 Wounded Warrior Project, Inc., 2003, www.woundedwarriorproject.org/mission/who we serve

PAGE 18

College of Liberal Arts and Sciences 334 Matherly Hall Economics PO Box 117140 http://people.clas.ufl.edu/michellephillips/ Gainesville, FL 3261 Email: michellephillips@ufl.edu 352 392 5017 I, Michelle A. Phillips, the undergraduate honors thesis coordinator for the Economics Department (CLAS), certify that the thesis: Title of thesis: Socioeconomic Determinants of Veteran Homelessness Written b y: Adita Ramgoolie Was approved by the E conomics Undergraduate Thesis C ommittee on: Approval Date: 4 / 13 /2018 Regards, Michelle A. Phillips