Assets and neighboring: an exploration into household assets and efforts to be a good neighbor.

A non-random, cross-sectional sampling procedure was utilized in this study to explore the relationship between the level of household assets and their corresponding level of neighboring. Surveys from 111 households were collected from September 1997, to April 1999, in seven Ohio counties to elucidate the relationship between the level of assets in working poor households and selected household demographic variables, and their propensity to provide various forms of community assistance over the previous month from the time of survey. Findings revealed that households 1) with more than one adult and 2) with lower levels of monthly earned income were more likely to provide community assistance to their neighbors. The implications of an asset-based social welfare policy strategy will be discussed.

Introduction

The theoretical model utilized in this study is Sherraden's (1991) asset-based theory of economic and social development. Sherraden is the current director of the Center for Social Development (CSD) at Washington University, and the survey tool (obtained by permission from CSD) used for data collection is similar to the one the CSD developed for their current national evaluation of asset-based programs or Individual Development Accounts, referred to as IDAs throughout this paper. Sherraden promotes a synthesis of economic development and social welfare via long-term asset accounts aimed at accumulating savings for life goals via the establishment of matched savings accounts or IDAs for working poor individuals and households. Accumulated savings can then be converted into long-term assets in the form of home ownership, continuing education, and/or a small business. While postulating a long list of potential benefits due to asset accumulation via IDA's, this paper will focus primarily on Sherraden's hypothesis that increased levels of material and human assets are associated with higher levels of community assistance/neighboring.

Numerous researchers have elucidated the relationship between the economic survival strategies of working poor households and their reliance on neighboring, support networks, and the informal economy (Edin and Lein, 1997; Parcel and Menaghan, 1997; Duncan and Brooks-Gunn, 1997; Wilson, 1996; Mingione, 1991). The plight of the working poor is of tremendous importance to our society due to the fact that nearly one in five children in America today lives in poverty and the working poor population is rapidly expanding (U.S. Census, 2000; Children's Defense Fund, 2000; Zaslow and Emig, 1997; Caputo, 1991). Social welfare researchers have challenged the efficacy of a laizze-fare capitalistic economic model which results in even greater social and economic disparity between social classes, communities, and geographic regions, even when the poor are working (Midgley, 1995; Kondrat, 1994; Estes, 1993; Smith and Wallerstein, 1992).

What Sherraden (1991) proposes is a theory of asset-based welfare policy in which the poor are provided the opportunities/life chances (Darendorf, 1979), just like middle and upper-class Americans, to establish a financial stake in the American economic system via the holding of actual material/human capital assets. Individual Development Accounts (IDAs) are the primary social welfare policy tool advocated by Sherraden to provide a financial stakehold for the working poor. However, little research has been completed to date as to the effectiveness of IDAs as an anti-poverty strategy for working poor households. The primary research question concerning IDA implementation addressed in this paper is whether or not higher levels of asset accumulation by working poor households are associated with higher levels of community assistance/neighboring? Or, as some social researchers have observed, will IDAs (asset accumulation) actually lead to less neighboring and sense of community (Lasch, 1995; Putnam, 1994; Rohe and Stegman, 1994), leaving behind those households with the lowest levels of material capital and human assets?

Specifically, the research question addressed in this paper is: what is the relationship between the level of assets in working poor households at the time of this study, including key household demographic characteristics, and their corresponding level of neighboring?

Methodololgy

This study is a secondary analysis of data collected for the purpose of evaluating the effectiveness of various forms of Resident Development Fund (RDF) Projects financed by Ohio Capital Corporation for Housing (OCCH). OCCH is a not-for-profit housing corporation providing more than 4,000 units of subsidized housing in 68 project locations across Ohio. OCCH's Resident Development Fund is $300,000 set aside for the "purpose of supporting initiatives to create opportunities for residents of affordable housing to improve their economic situation and achieve greater self-sufficiency" (OCCH Annual Report, p. 6, 1996). Of the numerous OCCH Resident Development Fund projects funded during the study, seven were structured around the implementation of Individual Development Accounts (IDAs) as conceptualized by Sherraden (1991). Monthly savings by IDA participants were matched by OCCH funds by a ratio of 2-to-1. That is, for each dollar a participant saved, OCCH deposited two dollars in their savings account. Matching funds could only be utilized by participants for three "legitimate" uses: 1) to purchase a home, 2) to continue education, or 3) to start a small business.

Surveys from 111 OCCH households are used in this study to contribute to our understanding of the relationship between the level of household assets at the time of survey and the resultant level of neighboring/community assistance given by the household over the last month. Table one is a compilation of the descriptive statistics measured in the 111 households. Gender was dropped from the analysis due to 86% of the households surveyed being female-headed, resulting in an insufficient sample size for a statistically valid analysis.

Sampling

A non-random, cross-sectional sampling procedure was utilized in which all potential IDA program participants (heads of households) were asked to fill out the IDA program evaluation survey before entry into the IDA program. The study period began with the implementation of OCCH's Resident Development Fund in September 1997. The last data was collected in April 1999. Only those households with income from labor market participation were eligible to participate in an IDA. Steps were taken to insure the participant confidentiality of both groups by only using the first and last initial of the head of household's name along with the last four digits of their Social Security number to identify their survey responses. A verbal informed consent was also obtained from each participant who completed the questionnaire, informing them of the research purposes of the study and insuring each participant of their confidentiality. Persons were not required to complete the questionnaire to receive services.

Sample Demographics

From table one, we see that 60 percent of households surveyed reported African American as their race, with 95 percent of the sample population ranging in age from 21 to 44, indicating that variation in age among the majority of respondents was fairly narrow. For the total number of adults (18 years of age or older) in each household, values reported ranged from one to six with 61 percent of all households reporting only one adult. Two-adult households represented an additional 34 percent of all households reporting, with a cumulative percent between these two categories of 95 percent. Concerning the total number of children (17 and younger) reported in each household, values ranged from a low of zero (16 households) to a high of seven (one household). The mean number of children reported is 1.7. Based on these findings, the "typical" respondent is a female, African-American single-parent, 33 years old, with two children. Forty-seven percent of respondents reported an education level of four (attended some college) with 26 percent reporting a high school education as their highest level of education completed. Therefore, to the above "typical" 33 year-old, female, African-American single-parent survey respondent could be added an educational level of having attended some college. Finally, the average monthly earned income (before taxes) for the "typical" household surveyed was $1200.

Data Analyses

The independent variables entered in the multiple regression analysis are: age and race of head of household, number of adults, and number of children in household, total household monthly earned income, total household material capital assets, total household debt, and the education level of head of household. Material capital assets are considered to be "long-term", providing households with financial stability and a "stakehold" in their lives and communities (Sherraden, 1991). Examples of material capital assets in this study include automobile ownership, as well as savings account balances. Because more than half of the households surveyed had a zero balance in their savings accounts and the remainder had negligible balances, a decision was made to include checking account balances as material capital assets.

The dependent variable entered in the multiple regression analysis is community assistance given (or level of neighboring). This concept refers to a participant's giving of community supports/services during the last month. This dependent variable was indicated by the literature review in the form of the level of "neighboring" that occurs in a community and its hypothesized relationship to a household's level of assets. As stated earlier, Sherraden (1991) hypothesizes that increases in assets will result in increases in community assistance/neighboring. However, the empirical evidence is mixed, with one study noting that levels of neighboring actually declined when comparing home owners to tenants, but that women who perceived positive changes occurring in the neighborhood were also more apt to be a good neighbor (Rohe and Stegman, 1994).

In this study, the level of survey participants' community assistance given over the last month is measured by their responses to the following nine survey questions regarding various forms of help: 1) helped with baby-sitting or child care, 2) cared for or stayed with an older or disabled adult, 3) given someone a ride, 4) helped with repairs to someone's home or car, 5) made phone calls or written/interpreted letters, 6) given someone food or loaned someone a tool, 7) helped with other kinds of work around the house, 8) watched someone's home or helped care for a pet, 9) given advice, encouragement, or emotional support? The range of possible scores for a study participant's level of community assistance given over the last month varies from a possible low score of 0 to a high score of 9. A Cronbach's alpha was calculated for the community assistance given scale, yielding a value of .6350, indicating a moderate degree of internal reliability. All study participants responded to this measure, with an overall sample score of 5.65 and a median score of 6.00, indicating a slightly negatively skewed sample. The standard deviation for the sample is 2.03.

Multiple Linear Regression Analysis

A multiple linear regression (MLR) analysis with hierarchical--stepwise entry was utilized to test the following hypothesis, while controlling for race and age of head of household, and the number of adults and the number of children in household under age 17. Independent variables are entered into the regression equation in two hierarchical steps. In step one, four independent variables (age and race of head of household, number of adults and number of children in household) are entered simultaneously as control variables. Therefore, the variance in each respective dependent variable explained by the above four control variables can be partialed out (controlled). Then in step two, the remaining independent variables (education level of head of household, total household monthly earned income, total household debt, and total household capital assets) are entered in a stepwise fashion into the regression. In this way, the additional variance (after controlling for age and race of head of household, and number of adults and children in household) explained by the education level of head of household (human capital), total household monthly earned income, total household debt, and total household material capital assets can be determined. For the multiple linear regression equation constructed, the full model is reported.

Hypothesis Testing Utilizing a Multiple Linear Regression Model

Hypothesis: The level of participant material/human assets is not significantly correlated with the level of participant community assistance given. The full model is statistically significant at the alpha = .05 level, with the calculated test statistic F = 2.283, p = .053 (see table 2; while the p value is slightly greater than alpha in this case, the results are reported at the alpha .05 level). The test hypothesis is therefore rejected and we conclude that the level of participant material/human assets at the time of survey is statistically significant in predicting their level of community assistance given. The total variance in the level of participant community assistance given (dependent variable) explained by the regression equation is .113 (R-square).

While holding all the other independent variables constant, the partial regression coefficients for number of adults in household (b = .841, t = 2.902, p = .005) and total household monthly earned income (b = -.001, t = -2.182, p = .032) were found to be statistically significant (alpha < .05) in predicting the head of household's level of community assistance given at the time of survey. Note that the relationship between total household monthly earned income and level of participant community assistance given is a negative one, indicating the higher the level of household earned income the lower their level of community assistance given. The regression equation for the full model is:

Com. Assistance Given = 5.925 -.015 (Age) - .531 (Race) + .841 (#Adults**) + .091 (#Children)-.001(Income*)

The number of adults in household is the most important independent variable in predicting level of community assistance given, with a standardized regression coefficient (Beta) of .312. Total household monthly earned income is the next relatively important independent variable with a Beta = -.233. The additional variance in participant level of community assistance given at the time of survey explained by the linear combination of education level of head of household, total household monthly earned income, total household debt and total household capital assets (step two of the regression) is .047 (R-square change). R-square change is also statistically significant at the alpha = .05 level (p = .032). Both the Tolerance and VIF statistics indicate that multicollinearity is not a problem in this regression analysis. Finally, to test for the assumption that the residuals are independent, the Durbin-Watson statistic was calculated, yielding a value of 1.597, indicating the independence of residuals in this regression analysis. Also, the assumption that residuals for the full model have a mean of zero is also supported by the calculated value of the residual mean = -.007.

Study Findings

The level of household monthly earned income was found to be significant in predicting the level of the head of household's giving community assistance/neighboring, with higher levels of household earned income being associated with lower levels of giving assistance (an inverse relationship). This finding is not congruent with Sherraden's hypothesis of higher income leading to greater community assistance. However, upon closer examination of the community assistance given scale, one might interpret the findings as higher income households not requiring the kinds of assistance measured in the study. Perhaps these households could simply afford to pay for these "neighborly" services rather than rely on the reciprocity of their neighbors?

Another interpretation of this finding is supported by Rohe and Stegman's research (1994) with a similar study population (125 low-income, predominately African American homeowners and 101 Section 8 renters with similar demographic characteristics) which found that homeowners were less likely to provide neighborly services than the control group of renters. In their study, Rohe and Stegman (1994, p. 170) measured a concept they referred to as "neighboring," with an index consisting of five questions: 1) how many people on your block do you know by name, 2) how many people on your block would you recognize if you saw them outside your neighborhood, 3) how many people on your block do you have a neighborly relationship with, 4) how many people on your block do you see socially at least three times a year, and 5) how many people on your block do you consider as close friends? Utilizing Sherraden's survey, the concept of a participant's level of community assistance is measured by their giving of assistance in their neighborhood over the last month. While focused on the giving of neighborhood assistance, parallels to Rohe and Stegman's "neighboring" measure can be seen, providing some degree of face and content validity to Sherraden's measure of community assistance. Therefore, according to Rohe and Stegman's findings, and this study's findings, higher asset levels do not necessarily lead to higher levels of neighboring/community assistance.

This rationale is also supported by historian Christopher Lasch's (1995) research and that of communitarian Amitai Etzioni (1988) who both challenge the capitalistic market place paradigm with a call to civility and social responsibility. While there can be no doubt that profits have risen dramatically in many capitalistic economies (including the U.S.), the disparity between haves and have nots has also risen (U.S. Census Bureau, 2000), findings congruent with Sherraden's analysis of wealth/assets in American society. However, the above finding in this study, while preliminary, should provide a caution in the implementation of IDAs as an anti-poverty strategy. Unless efforts are made to extend the benefits of asset ownership to the "poorest of the poor," a poverty program which is dependent upon participant savings from earned income may only "cream" those households who already are financially better off, leaving communities and neighborhoods even more divided along socio-economic lines. Households with more income may be less likely to provide commuinity assistance, potentially weakening the survival strategies of lower socio-economic communities. Or perhaps worse yet, these "successful" households could actually move out of poor neighborhoods, leading to an even greater concentration of poverty and limited economic resources in these areas as they pursue opportunity elsewhere. Sherraden's asset-based theory accepts the neo-classical economic concept of the social mobility of capital, buying into the pursuit of wealth and "happiness" via following economic opportunity wherever it may lead. While this financial/economic strategy has proven successful for many middle and upper-class Americans, working poor households and communities may resist leaving the neighborhoods and communities where their trusted financial and economic survival strategies are established, findings consistent with other studies (Task Force on Persistent Rural Poverty, 1993; Mingione, 1991; McGranahan, 1988). Inversely, closer interpretation of the community assistance given scale utilized in the study could lead one to associate greater need with those who have higher levels of community assistance given. Since wealthier households are less likely to give community assistance, it would appear as though poorer families, who are more likely to give community assistance, might be less likely to have the savings required to benefit from IDA participation. Additional research is needed to illuminate this relationship between higher levels of community assistance given (especially in the forms measured in this study), such as providing childcare, elderly care, transportation, home/auto repair etc. in one's community and the survival strategies of working poor households.

Finally, the number of adults in a household was also found to be significant in predicting the level of community assistance given, at the time of this survey. An additional paired samples (2-tailed) t-test was performed between household income and the number of adults in household to explore their relationship. No significant correlation was found between these two independent variables (t = -18.715). Perhaps those households with more than one adult have the "opportunity" to provide community assistance at a greater frequency than those with only one parent? Also, this researcher observed that in numerous study households with more than one adult, the additional adult was found to be elderly and/or disabled, making minimal contributions to monthly earned income via formal labor market participation. Additional research is also needed to elucidate this relationship between community assistance given and family form.

Summary

Unfortunately, the current economic and political consensus in the United States continues to stigmatize the poor and blame them for their poverty rather than the social, economic, and political structures which promote "production for profit" rather than investment in human potential. The same economic environment has pervaded the global economy, placing vulnerable populations, especially poor women and children, at risk. How to manage this paradox of the responsibility of society, on the one hand, and the responsibility of the individual on the other, in bringing about a "common vision" of economic and social justice is, in this researcher's mind, the synthesis of knowledge building in social work which now calls the profession to task.

Table 1
Descriptive characteristics of households surveyed (N = 111)

Variable                             Percent  Mean      S.D.

Race of Head of H.H. (n = 109)
  African American                      59.6
  Caucasian                             38.5
  Hispanic                               0.9

Age of Head of H.H. (n = 110)                    33.74      7.75
  21-25                                 14.5
  26-30                                 21.0
  31-35                                 23.6
  36-40                                 25.4
  41+                                   15.5

Number of Adults in H.H. (n = 111)                1.47      7.75
  1                                     61.3
  2                                     34.2

Number of Children in H.H. (n = 111)              1.75      1.31
  0                                     14.4
  1                                     30.6
  2                                     35.1
  3                                     13.5

Level of Education Completed by Head . (n = 110)
  Grade, middle or jr. high               .9
  Attended high school                   7.3
  H.S. graduate or GED                  25.5
  Attended some college                 47.3
  Graduated from college                12.7
  Attended graduate school               6.3

Total H.H. Earned Monthly Income (n = 109)     1200.50    669.01
  0                                      6.4
  1-500                                  7.3
  501-1000                              22.9
  1001-1500                             35.8
  1501-2000                             17.4
  2001+                                 10.1

Total H.H. Debt (n = 110)                     12514.02  16631.29
  0                                      8.2
  1-5K                                  31.8
  5K-10K                                23.6
  10K-15K                               10.0
  15K-20K                                5.5
  20K-30K                               10.9
  30K+                                  14.5

Total H.H. Material Capital Assets (n = 101)  13725.37  27870.62
  0                                      9.9
  1-5K                                  47.5
  5K-10K                                21.8
  10K-15K                                3.0
  15K-20K                                3.0
  20K-160K                              14.8

Table 2
Multiple regression analysis of IDA survey participant level of
community assistance given and preselected independent variables
(n = 111)

Variable             b      Beta     t        p

Age                  -.015  -.057   -.543  .589
Race                 -.531  -.129  -1.222  .225
No. of Adults         .841   .312   2.902  .005 **
No. of Children       .091   .060    .580  .563
Total Earned Income  -.001  -.233  -2.182  .032 *

Excluded Variables:
H.H. Debt             .098   .097    .924  .358
H.H. Capital          .013   .012    .120  .905
Education             .148   .159   1.414  .161

R = .336; R-square = .113; F = 2.283; p = .053 *; R-square
change = .047; p = .032 *; Tolerance Statistic (.854-.920);
VIF Statistic (1.087-1.171); Durbin-Watson (1.597)
Mean of Residuals (.007)

* Significant at .05 level

** Significant at .01 level

Bibliography

Caputo, R. (1991). Patterns of work and poverty: Exploratory profiles of working-poor households. Families in Society, vol. 72, Oct. 1991, p. 451-60.

Children's Defense Fund (2000). The state of America's children yearbook. Washington, D.C.: Children's Defense Fund.

Darendorf, R. (1979). Life chances: approaches to social and political theory. Chicago, IL: University of Chicago Press.

Duncan, G. and Brooks-Gunn, J., Eds., (1997). Consequences of growing up poor. New York: Russel Sage.

Edin, K. and Lein, L. (1997). Making ends meet: How single mothers survive welfare and low-wage work. New York: Russel Sage Foundation.

Estes, R. (1993). Toward sustainable development: From theory to praxis. Social Development Issues, 15(3), 1-29.

Etzioni, A. (1988). The moral dimension: Toward a new economics. New York: Free Press.

Kondrat, M. E. (1994). Culture and power in technology transfer. Social Development Issues. Vol. 16(3) 45-67.

Lasch, C. (1995). The revolt of the elites and the betrayal of democracy. New York: W.W. Norton and Co.

McGranahan, D. (1988). Rural workers in the national economy. Rural development research report no. 69. U.S. Government Printing Office: Washington, D.C.

Midgley, J. (1995). Social development: The development perspective in social welfare. Sage Publications: Thousand Oaks, CA.

Mingione, E. (1991). Fragmented societies: A sociology of economic life beyond the market paradigm. Cambridge, MA: Basil Blackwell.

Ohio Capital Corporation for Housing (1996). Annual Report. Columbus, OH: OCCH.

Parcel, T. and Menaghan, E. (1997). Effects of low-wage employment on family well-being. The Future of Children, vol.7, no.1, 116-121.

Putnam, R. (1995). Bowling alone: America's declining social capital. Journal of Democracy, Vol. 6, No. 1, 65-78.

Rohe, W. M. and Stegman, M.A. (1994). The impact of home ownership on the social and political involvement of low-income people. Urban Affairs Quarterly, 30, 152-172.

Sherraden, M. (1991). Assets and the poor New York: M. E. Sharpe.

Smith, J. and Wallerstein, I. (1992). Creativity and transforming households: The constraints of the world-economy. Great Britain: Cambridge University Press.

U.S. Bureau of Census (2000). Statistical Abstract of the United States. Washington, D.C.: Government Printing Office.

Wilson, W. J. (1996). When work disappears: The world of the new urban poor. New York: Alfred A. Knopf.

Zaslow, M. and Emig, C. (1997). When low-income mothers go to work: Implications for children. The Future of Children, vol. 7, no. 1, 110-115.

TIM REUTEBUCH
University of Wisconsin-Whitewater

Tim Reutebuch University of Wisconsin Whitewater, WI 53190-1790 [e-mail]