Download to read offline
![
[1] What geographic patterns of NEI (e.g., investment
income vs. government transfers) are apparent in the
upper Great Plains region during the early Great
Recession (2007-2008)?
[2] How influential is NEI in this region‟s economy?; and
[3] What factors (e.g., socio-demographics, economic,
etc.) explain the geographic variations of NEI? Very
specifically, are these variations indicating strong
relationships to different industrial sector patterns and
are they shaped by the urban system such as urban,
suburban, or exurban?](https://image.slidesharecdn.com/beaverthesisdefense-131103212037-phpapp01/85/BTD2013-6-320.jpg)
More Related Content
BTD2013
- 1. By Edward Beaver MA ThesisDefense Department of Geography, UNC-Greensboro 4 November, 2013
- 2. Intro Research Questions Theoretical/Conceptual Background StudyArea, Data and Methods Results Conclusions & Summary
- 3. NEI: monies notfrom labor Includes two types of non-labor income o Government Transfers (15% of U.S. income) • Social Security, Medicare, Food Stamps (SNAP) o Investment Income (25% of U.S. income) • Capital Gains, Savings Interest, Stock Dividends NEI “...can be seen as bringing „new money‟ into a local economy...” as exports or higher wages do, even in countercyclical fashion (Nelson, 2008)
- 4. Minimal attention fromleaders Uneven NEI geography across U.S. Government budgets in dire straits promise changes o “Grand Bargain” Potential on Taxes/Revenues Demographics in „Age of Austerity‟ o “Grey” economies require more government transfers in support o Lower birth rates= less new taxpayers o Disputes over „legitimate‟ government transfers increasingly age-centric (Brownstein, 2010)
- 5. Stage 1: SocialSafety Net (1913-1945) Stage 2: Growth & Demographics (1946-1975) NEI History Stage 3: Revolution & Deregulation (1976-2007) Stage 4: New Normal (2008-Present)
- 6. [1] What geographicpatterns of NEI (e.g., investment income vs. government transfers) are apparent in the upper Great Plains region during the early Great Recession (2007-2008)? [2] How influential is NEI in this region‟s economy?; and [3] What factors (e.g., socio-demographics, economic, etc.) explain the geographic variations of NEI? Very specifically, are these variations indicating strong relationships to different industrial sector patterns and are they shaped by the urban system such as urban, suburban, or exurban?
- 7. Geographers and othersocial scientists considered NEI‟s importance in many other types of places and conditions o Rural counties (Nelson and Beyers, 1998) o Natural resource dependent communities (Petigaraa et al., 2012) (Nelson, 2005:2008) Regional income convergence (Austin and Schmidt, 1998) Boom and bust economic cycles (Smith and Harris, 1993) o Life course migration o o (Wenzl, 2008) o States (Kendall and Pigozzi, 1994; Campbell, 2003) o Megapolitan areas (Debbage and Beaver, 2012) o Micropolitan areas (Mulligan and Vias, 2006) o Core-based statistical areas
- 8. None except Forward(1982:1990) and Wenzl (2008) incorporated capital gains and private pensions as investment income But Wenzl left out multiple NEI types, esp. transfers None ever considered NEI in metropolitan counties, the population and economic centers of the nation, especially not in the UGP
- 11. BEA Census/ACS IRS SOI ZCTA BEA capital gains refusal Multi-County Zip Codes IRS 20 months late
- 12. %Single Parent Household Median HomeValue % Non-Caucasian % Elderly % Services Employment % Uninsured % Low Birth Weight % Construction Employment % Office Employment % Prod./Transportation % MGMT/Professional Employment Employment % Employment Growth 2000-07/08 % Movers 1 Year Before % Poverty % Workforce Participation % Bachelor Degree % Unemployment % Renters %High School Dropout % Married % Diabetes % Population Growth 2000-07/08
- 13. Highest NEI % Counties Polk, MO Macoupin , IL Carlton , MN Jasper , MO St. Louis , MN Douglas , WI Pennington , SD Callaway , MO Washington , MO Greene , MO Dubuque , IA Jones , IA Cape Girardeau , MO St. Louis , MO Henry , IL Franklin , MO Blue Earth , MN Shawnee , KS Rock Island , IL Polk , MN 48.11% 45.82% 44.82% 43.78% 43.75% 43.60% 43.32% 43.27% 42.75% 42.72% 42.55% 42.46% 41.92% 41.65% 41.32% 40.83% 40.43% 40.34% 40.32% 40.24% Lowest NEI % Counties Riley , KS Scott , MN Sarpy , NE Geary , KS Carver , MN Warren , IA Platte , MO St. Croix , WI Wright , MN Dallas , IA Sherburne , MN St. Charles , MO Dakota , MN Isanti , MN Clay , MO Polk , IA Anoka , MN Jefferson , MO Pottawattamie , IA Sumner , KS 25.98% 26.65% 27.40% 27.69% 27.89% 28.98% 29.01% 29.68% 29.81% 30.00% 30.02% 30.20% 30.48% 30.55% 30.63% 31.24% 31.27% 31.51% 31.82% 32.36%
- 14. Highest Transfer % Washington, MO 29.80% Polk , MO 26.04% Wyandotte , KS 22.29% St. Louis city, MO 21.90% Carlton , MN 21.86% Webster , MO 21.81% Douglas , WI 21.48% Jasper , MO 21.23% Buchanan , MO 20.47% Lafayette , MO 19.90% St. Louis , MN 19.85% Polk , MN 19.78% Franklin , KS 18.92% Warren , MO 18.29% Jones , IA 17.99% St. Clair , IL 17.88% Macoupin , IL 17.78% Dakota , NE 17.73% Callaway , MO 17.66% Pottawattamie , IA 17.27% Lowest Transfer % Lincoln , SD 4.95% Carver , MN 6.03% Johnson , KS 6.68% Dallas , IA 7.48% Scott , MN 7.71% Washington , MN 7.93% Riley , KS 8.02% Dakota , MN 8.67% Johnson , IA 8.76% Sarpy , NE 9.06% Platte , MO 9.08% St. Croix , WI 9.18% Cass , ND 9.36% Hennepin , MN 9.61% Geary , KS 9.89% St. Louis , MO 10.15% St. Charles , MO 10.23% Douglas , NE 10.60% Monroe , IL 10.69% Douglas , KS 10.73%
- 15. Highest Investment % St.Louis , MO 31.50% Johnson , KS 29.27% Pennington , SD 28.65% Meade , SD 28.57% Douglas , NE 28.17% Macoupin , IL 28.04% Lincoln , SD 27.52% Dubuque , IA 27.39% Minnehaha , SD 27.34% Greene , MO 27.24% Douglas , KS 26.75% Hennepin , MN 26.70% Franklin , MO 26.58% Blue Earth , MN 26.51% Monroe , IL 26.31% Cape Girardeau , MO 26.19% Henry , IL 25.76% Callaway , MO 25.60% Rock Island , IL 25.55% Ramsey , MN 25.30% Lowest Investment % Washington , MO 12.95% Wyandotte , KS 12.99% Pottawattamie , IA 14.54% Dakota , NE 15.47% Ray , MO 15.70% Isanti , MN 15.99% Sumner , KS 16.14% Webster , MO 16.32% Warren , MO 16.50% Clinton , MO 16.54% Jefferson , MO 16.87% Warren , IA 16.91% Lafayette , MO 17.32% Geary , KS 17.79% Buchanan , MO 17.80% Riley , KS 17.96% St. Louis city, MO 18.01% Wright , MN 18.34% Sarpy , NE 18.34% Miami , KS 18.59%
- 16. Highest Gain FromIRS Counties Callaway , MO Lincoln , MO Christian , MO Macoupin , IL Benton , IA Meade , SD Franklin , MO Anoka , MN Jersey , IL Butler , KS Leavenworth , KS Geary , KS Carlton , MN Chisago , MN Benton , MN Lincoln , SD Douglas , WI Cass , MO Jefferson , MO Clay , MO 99.31% 90.89% 84.40% 81.62% 76.18% 76.16% 75.40% 69.69% 69.63% 67.77% 62.86% 61.22% 61.05% 59.89% 58.07% 57.32% 56.67% 54.67% 52.95% 52.13%
- 17. Principal components analysis(PCA): Component 1 Hi-Low SES 2 Diversity 3 Office/SPH 4 Mobility 5 Home Value 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Total 6.62 4.47 2.16 1.87 1.08 1 0.82 0.73 0.59 0.46 0.44 0.39 0.34 0.26 0.17 0.15 0.13 0.12 0.09 0.06 0.05 0 Eigenvalues Extraction Sums of Squared Loadings %Variance %Cumulative Total %Variance %Cumulative 30.09 30.09 6.62 30.09 30.09 20.33 50.42 4.47 20.33 50.42 9.82 60.24 2.16 9.82 60.24 8.48 68.72 1.87 8.48 68.72 4.93 73.65 1.08 4.93 73.65 4.53 78.18 3.71 81.89 3.33 85.22 2.67 87.89 2.1 89.99 2 91.99 1.76 93.74 1.57 95.31 1.16 96.47 0.78 97.25 0.68 97.93 0.6 98.53 0.57 99.1 0.41 99.51 0.26 99.77 0.22 99.99 0.01 100
- 19. Variables Constant PC1 Model 1 2305734 -5452488 Model 2 776526 -23574 PC2 1579810* 409943 1926048* 510557 31.88 0.39 99 1060650* 519160* 174022* 235921 --1369981* 556066* 174116* 336441 26.53 20.19 0.51 0.37 99 99 PC3 PC4 PC5 F-Statistic Adj. R-Square Total Observations Note: *Significant at 0.05 level Model 1: NEI $ Model 2: Investment Income Model 3: Transfer Income Model 4: NEI Ratio Model 3 1529208 -521674* Model 4 1.74 0.57195* 0.11949* 0.09 -0.02 0 72.68 0.59 99
- 20. Greatest NEI impactobserved in suburban & exurban counties, much more than in wealthy suburban counties. Where NEI concentrates, transfers and investment NEI trend toward high levels of both Counties with lower NEI trended suburban or exurban. Urban cores and very well educated counties had highest levels of investment income NEI is no more or less influential in UGP than nationally BEA accounts of NEI are inaccurate without capital gains. Gov‟t policy is very important in present & future NEI Socio-economic status indicators are most important indicators
- 21. Incorporate IRS dataacross all metropolitan counties in U.S. for regional comparisons Structural equation modeling instead of PCA and regressionbetter explains variable relationships Account for „accrued capital gains‟ somehow Promote NEI awareness among policymakers and academics
- 22. Thank you tomy advisor Dr. Selima Sultana! My thesis committee members Dr. Keith Debbage & Dr. Zhi-Jun Liu. The San Francisco Blue Cross/Blue Shield Research Team, Wake County Tea Party & N.C. NAACP for supportive and constructive comments over the summer and fall.
Editor's Notes
- #5 Aside from its growing share of local economies:
- #7 Explain how NEI changes due to government policy shiftsHighlight how including capital gains as investment income changes NEI geography
- #8 Tiebout(1962) acknowledged NEI in economic base theory butManson & Groop(1990) argued that only in the early 1980s was NEI included in economic base analyses for development purposes Along with Forward’s (1982:1990) Canadian city studies of NEI, Manson & Groop were early champions of NEI’s importance nationally (1988:1990:2000)
- #9 Wenzl was also more interested in household savings.
- #10 Upper Great Plains (UGP) classified by the U.S. Census as the West North Central Census sub-region Seven states: Iowa, Minnesota, Missouri, Nebraska, Kansas, South Dakota, North Dakota99 metropolitan statistical area (MSA) counties with 14.5 million peopleEleven MSA counties from Wisconsin and Illinois included because they were part of UGP-based MSAsClassified as suburban, wealthy suburban, major urban center, small urban center and exurban counties
- #11 Why study the UGP?It avoided the worst of the Recession’s impact in housing and employment (comparative to the Southeast and Midwest)its lack of extreme wealth concentration and inequality (comparative to the Northeast) its relative geographic proximity & concentration of counties (comparative to the West which ranges from Washington State to New Mexico)
- #12 Most prior studies used only NEI data from the Bureau of Economic Analysis (BEA).BEA refuses to include capital gains because they regard them as asset price changes, not ‘production or income’BEA counts pensions as earnings rather than as NEI when they are issued to pensionersCapital gains & pensions combined range from 3-7% of total national income every year, data available from IRSThis is a significant ‘gap’!Only problem to incorporate IRS data: zip code onlyMatched with counties via ARCGIS & MS Excel.10% of zip codes were located in multiple counties based on the % of population from Zip Code Tabulation Areas (ZCTAs), a county with 76% of the population was assigned 76% of the IRS data and the other county assigned 24%.IRS data was added to investment income totals
- #13 Data for socio-demographic, economic, and health variables collected from 3 year American Community Survey (2007-2009), and Robert Wood Johnson Foundation’s CountyHealthRankings.orgOriginally planned to include BEA & IRS NEI data for 2007-2009, but IRS never released 2009 data (now 18 months late!)22 variables in totalPrincipal component analysis (PCA) was utilized to address multi-collinearity concerns with variables
- #14 Above-average counties- more than half were major or small urban cores, including St. Louis city (MO), St. Louis, MN (Duluth), and Greene, MO (Springfield). All but one of the counties in the below average tier (Polk, IA) were suburban or exurban. Several in this group included the wealthiest counties in the U.S., with Carver, Wright, Anoka, Dakota, and Sherburne from Minneapolis and one each from Kansas City (St. Charles, MO) and Des Moines (Dallas, IA)
- #15 The counties with below average proportions of transfers shared two primary features:Most had below average % of elderly populations (< 14.9%) and higher than average workforce participation % (> 70.1%Counties with higher proportions of transfers shared primary commonalities, including:above average % of elderly (> 14.9%), lower than average workforce participation %(< 70.1%), higher than average poverty % (> 10.9%) and lower than average bachelors or better educational attainment % (< 24.9%
- #16 Investment income is heavily represented in urban core counties of metropolitan areas, including Hennepin at 27.0%, Douglas at 28.2%, and St. Louis at 31.5%. Wealthy counties were notably absentMajority of below-average tier was suburban and exurban counties. 14 had above-average % of transfers (> 14.1%), half had average or above % of elderly residents (> 14.9%), and most had average or below % of workforce participation (< 70.1%) and below average % of bachelors degree attainment (> 24.9%).
- #17 Median increase in investment income from including capital gains data was 36.6%.Mostly suburban and exurban counties comprised the above average tier of gainersEight counties had 10% or more of their total income overlooked by the BEA
- #18 utilized as a variable-reduction technique creates a smaller group of ‘principal components’, artificial variables accounting for as much of the variance in the original variables as possible. PCA was suitable for use because the variables were all measured at the continuous level, had linear relationships, and outlier tests indicated no significant outliers among the variables. 5 variables were created by PCA, accounting for 73.7% of total variance of all variables.Component 1 (hi-lo SES) contains variables tracking both weaker and stronger socio-economic status (SES) including education, workforce participation and health condition(Leigh and Blakely, 2013). Component 2 (Diversity) is a signal of diversity, with linked variables including % of non-whites, % renters and % in poverty. Component 3 (Office) contains the linked variables of single-parent households and employment in the office sector, a lower-paying sector that has less educational requirements (Roberts et al., 2012). Component 4 (Mobility) highlights mobility trends, especially for recent migrants who may lack health insurance as they seek employment, a dilemma that can worsen economic vulnerability (EBRI, 1999). Component 5 is reserved for median home value.
- #19 Component 1’s positive scores predominate with higher proportions of elderly, greater than median employment in lower-skill & pay employment sectors, higher unemployment & sub-optimal health conditions. Negative scores predominate with higher workforce participation rates, greater than median employment in the high-skill and high-pay professional and management sector & higher educational attainment levels. Component 1 captures higher and lower dynamics of SES. e.g., the concentration of positive scores in the Kansas City and St. Louis metropolitan areas and mostly negative scores in the Minneapolis & Des Moines MSAs. Component 2’s positive scores are prevalent with higher non-Caucasian populations, higher % of poverty and % of renters Negative scores indicate those with strong levels of marriage and construction sector employment. Possible unique role of construction employment in UGP counties with lower than average levels of Hispanic immigrant workers & possibly less downward pressure on wages as a result (Thompson, 2010). Also indicate the loadings reflect diversities of household composition, economic status and ethnicity: e.g. relatively wealthy Hennepin County, the urban core of Minneapolis MSA, has a very positive score driven in part by its greater diversity and renter demographics, as do many of the young, renter-heavy counties with major universities. Meanwhile, poorer urban core counties in the inner St. Louis and Kansas City MSAs also score highly.
- #20 Linear Regression Analysis with PCA:Stepwise regression models in SAS 9.3 to determine influence of each PCA variable on NEI types while controlling for their influence.Three options evaluated for best possible model.Model 1: Total NEI $ for each countyModel 2: Two models utilizing transfer & investment income totalsModel 3: NEI Ratio (Investment/Transfer)Model 1: Total NEIThe R-squared in the final regression model suggested that 38.6% of Total NEI variation was accounted for by two PCA components: Diversity (Comp. 2) & Uncertainty (Comp. 4)Using the unstandardized b coefficients, the estimated regression equation is:NEI = 2305734 + 1579810 (DI) – 1926048 (UN)Model 2: Total TransfersThe R-squared in the final regression model suggested that 51% of Total Transfers variation was accounted for by four PCA components: Diversity (Comp. 2), Office (Comp. 3) Uncertainty (Comp. 4) & Home Value (Comp. 5)Using the unstandardized b coefficients, the estimated regression equation is:Transfer Income = 776526 + 519160 (DI) + 174022 (OFF) - 556066 (UN) + 174116 (MHV)Model 2: Total InvestmentThe R-squared in the final regression model suggested that 37% of Total Investment variation was accounted for by three PCA components: Low-High SES (Comp. 1), Diversity (Comp. 2), & Uncertainty (Comp. 4).Using the unstandardized b coefficients, the estimated regression equation is:Investment Income = 1529208 - 521674 (LOH) + 1060650 (DIV) - 1369981 (UN) Model 3: NEI Ratio (Investment/Transfer)The R-squared in the final regression model suggested that 59.5% of NEI Ratio variation was accounted for by two PCA components: Low-High SES (Comp. 1) & Diversity (Comp. 2).Using the unstandardized b coefficients, the estimated regression equation is: NEIR= 1.74 – 0.571(LOH) + 0.119 (DI)Of the three models, the best is Model 3 (NEI Ratio).It accounts for the most variation within the components, has the least detectable bias, and has the components with the most included original variables. The second model containing the investment and transfer models has more detectable bias (Mallow’s CP score) and accounts for less of the variation in the components. The first model has considerable bias and also accounts for less of the variation in the components.
- #22 at 40% of U.S. Total Income, it must be considered as often as employment in policy debates