Analyzing Home Lending Patterns for Discrimination in Worcester, MA with Linear Regression Curtis Wiemann Presented with assistance by Kathryn Madden,

Slides:



Advertisements
Similar presentations
Exploring the Dimensions of Foreclosure in the Context of Neighborhood Housing Markets. Michael Barndt Todd Clausen Jeff Arp Nonprofit Center of Milwaukee.
Advertisements

ANALYSIS OF IMPEDIMENTS TO FAIR HOUSING CHOICE (AI) City of Missoula.
Significance Testing.  A statistical method that uses sample data to evaluate a hypothesis about a population  1. State a hypothesis  2. Use the hypothesis.
Impacts and Responses to ‘Spatial Concentrations’ of Foreclosures in Massachusetts Communities of Color Presentation William Monroe Trotter Institute University.
Targeting NSP Funds Todd Richardson HUD Office of Policy Development and Research.
1. Spoiler Alert!!! Some movies have such a twist, you walk out of the theater in total shock! 2 “Fight Club” “The Others” “The Sixth Sense”
1 Multiple Regression A single numerical response variable, Y. Multiple numerical explanatory variables, X 1, X 2,…, X k.
Fair Lending 2001 Why are you here? Everyone has contact with customers You may be the first to be approached regarding a loan Know who to refer the.
1 Data on Housing Vacancy and Housing Cost from the Census Bureau Dr. Arthur R Cresce Assistant Division Chief for Housing Characteristics Housing and.
Millennial Housing Commission Housing Program Tutorial, May 2001 Additional Baseline Information Housing Program Tutorial May 14, 2001 Millennial Housing.
The Rural Housing Data Portal Information for Rural America Housing Assistance Council.
Home Mortgage Disclosure Act (HMDA) Data Raphael W. Bostic University of Southern California Housing Statistics User Group West meeting at the University.
Demographics 14,583 people. 6,137 housing units The racial makeup 97.31% White, 0.23% African American, 2.03% Native American, 0.76% Asian,
Looking at data: relationships Scatterplots IPS chapter 2.1 © 2006 W. H. Freeman and Company.
Chapter 7 Scatterplots, Association, Correlation Scatterplots and correlation Fitting a straight line to bivariate data © 2006 W. H. Freeman.
Baltimore City African American Middle Class Analysis and Metrics Matthew Kachura Program Manager BNIA-JFI, University of Baltimore January 10, 2008.
LECTURE 2 Understanding Relationships Between 2 Numerical Variables
Requests for permission to make copies of any part of the work should be mailed to: Thomson/South-Western 5191 Natorp Blvd. Mason, OH Chapter 17.
Copyright © 2011 Pearson Education, Inc. Multiple Regression Chapter 23.
WOODSTOCK INSTITUTE | October 2012 October 4-5 | Birmingham, UK US Financial Data Disclosure Policy Tom Feltner | Vice President Woodstock Institute |
GSE REFORM MADE SIMPLE: (CASH-OUT) REFINANCES VS. HOMEOWNERSHIP LENDING November 13,2013 Joel S Singer, CEO California Association of Realtors.
CHAPTER SEVENTEEN Consumer Loans, Credit Cards, And Real Estate Lending
Home Mortgage Refinance and Wealth Accumulation American Housing Survey User Conference Washington, DC March 8, 2011 Frank E. Nothaft and Yan Chang Freddie.
LECTURE UNIT 7 Understanding Relationships Among Variables Scatterplots and correlation Fitting a straight line to bivariate data.
Discrimination in Loan Servicing David Berenbaum Chief Program Officer National Community Reinvestment Coalition Fair Housing 2010: Time to Act 2010 National.
1 Mortgage Defaults and Foreclosures: Recent Trends and Associated Economic and Market Developments Randy Fasnacht U.S. Government Accountability Office.
HUD Consolidated Plan Cuyahoga County Planning Commission Cuyahoga Urban County September 16, 2009.
1 The High Cost of Segregation Exploring Racial Disparities in High Cost Lending Vicki Been, Ingrid Ellen, Josiah Madar, Johanna Lacoe Urban Affairs Association.
Census 2000 Supplementary Survey: An Operational Feasibility Test Nancy M. Gordon Associate Director for Demographic Programs U.S. Census Bureau July 2001.
Accessing Census Data through the American FactFinder Arthur Bakis Information Services Specialist Boston Regional Census Center US Census Bureau
1 Fair Lending Pricing Analytics Katherine Samolyk, Senior Economist, Division of Insurance and Research, Federal Deposit Insurance Corporation September.
Property Rights Protection and Bank Loan Pricing Kee-Hong Bae Korea University Vidhan K. Goyal Hong Kong University of Science and Technology.
Lecture 3 – Sep 3. Normal quantile plots are complex to do by hand, but they are standard features in most statistical software. Good fit to a straight.
Presented by Mollie Fitzpatrick, Heidi Aggeler, Jen Garner 1999 Broadway, Suite 2200 Denver, Colorado (303)
Group 3 Members:Dan Sun Hongliang Wu Hui Lai Hui Wang Ling-Ching Hsu Seok-Rahn Lee Shin-Hao Lee Yuanbo Mao Analysis of House Price in California Econ 240.
Emerging Markets Homeownership Initiative Committee Meeting: Homeownership Barriers Part II September 29, 2004.
AGGREGATE DEMAND. Aggregate Demand (AD) Shows the amount of Real GDP that the private, public and foreign sector collectively desire to purchase at each.
Chapter Ten: Inequality in Housing and Wealth By Tanya Maria Golash-Boza.
Presentation at MassHousing Prabal Chakrabarti, AVP and Director of Community Development March 3, Any views expressed are not necessarily those.
Visualizing Displacement and Gentrification: Los Angeles
The Hispanic Market United States California Southern California
CHAPTER SEVENTEEN Consumer Loans, Credit Cards, And Real Estate Lending
Global Investments, Inc. Student Coaching Slides
Implementing ACED Designations for HFA Homeownership Programs
Series 13 Regional Growth Forecast
The Impact of Credit Scoring on Under-Served Markets:
Jim Park Chairman Emeritus Asian Real Estate Association of America &
Habitat for Humanity Illinois’ Building Impact Program
Assessment of Fair Housing (AFH)
Home Mortgage Disclosure Act Lending Patterns (HMDA): Cuyahoga County
F Chapter 17 FUNDAMENTAL ANALYSIS vs TECHNICAL ANALYSIS 7/30/2018
Mortgage Discrimination
The History of the ? Economic Crisis
Affordable/Workforce Housing Issues and Programs
Shared Equity Housing and the DC Region
CHAPTER SEVENTEEN Consumer Loans, Credit Cards, And Real Estate Lending
Mortgage Ready Millennials FALHFA Annual Conference
Extreme Poverty, Poverty, and Near Poverty Rates for Children Under Age 5, by Living Arrangement: 2011 The data for Extreme Poverty, Poverty, and Near.
ANALYSIS SPATIAL DATA University of Pennsylvania
Global Investments, Inc. Student Coaching Slides
Transnational Investments, Inc. Student Coaching Slides
1 Causal Inference Counterfactuals False Counterfactuals
2University of Virginia
Brown County Financial Decision and Support Model
Current conditions.
3.3 Cautions Correlation and Regression Wisdom Correlation and regression describe ONLY LINEAR relationships Extrapolations (using data to.
Is Mortgage Credit Too Tight? What the Data Tells Us
Building Michigan Communities Conference Lansing, Michigan
Discussion of Baugh (2015) “What happens when payday borrowers are cut off from payday lending? A natural experiment” Brian T. Melzer Kellogg School of.
Future Borrowers: Challenges and Opportunities
Presentation transcript:

Analyzing Home Lending Patterns for Discrimination in Worcester, MA with Linear Regression Curtis Wiemann Presented with assistance by Kathryn Madden, AICP, Faculty at Clark University

Methodology Utilize linear regression to depict and analyze trends in home lending in Worcester, MA to determine if lending exhibited racial bias Begin with univariate analysis of mortgage lending (measured in dollars per capita of conventional/FHA loans originated) and race/ethnicity (measured in % Hispanic [all races] and African-American) Analyzed at Census Tract level (smallest unit of measurement possible for HMDA data publications) Null hypothesis: no correlation between borrower race/ethnicity and home lending Home lending influenced by many legitimate factors – to model risk and stock by Census Tract, rates of foreclosure (HUD, % of total), median income (Census, Table DP03 [Household Median Income]), and tenure (Census, Housing Characteristics [% Owner-Occupied]) were added in multivariate linear regression

Results Univariate regression returned statistically-significant linear relationships between race and home lending, but these relationships were ultimately found to have limited explanatory value in comparison to foreclosure rates and median income rates by census tract. Conjecture based on result: Race/ethnicity shares a linear relationship with median income and foreclosure rates, which indirectly shapes its linear relationship with mortgage lending. Market conditions allow for inequitable lending that, while not directly correlated with race, nevertheless leads to fewer home lending dollars in mortgages originated to borrowers of color Missing from this analysis: other racial/ethnic groups, analysis of home prices (which in turn affects mortgage origination values)

Limitations Inappropriateness of linear regression Linear regression best for controlled experiments; complexity of human factors influencing mortgage lending leads to low P-values throughout = low predictive value (~ 0.3, up to ~ 0.5) Linear regression more useful the more points added; with 54 Census Tracts analyzed, outliers more influential Data Relatively small number of home loans covered under HMDA Relatively small number of loans per tract (6 tracts with no conventional lending whatsoever – outliers) Missing information: average home valuation by tract, other economic/social characteristics that could legitimately shape access or desire to conventional home borrowing Approach Aggregate view by tract imperfect method to evaluate existence/impact of discrimination

Recommendations Expand the mandate of the CRA “The Performance and Profitability of CRA-Related Lending,” 2000 Report to Congress: 30% of home lending activities are subject to the CRA CRA mandates a slate of community reinvestment actions that may no longer reflect realities of geographical discrimination Support lending in weak market tracts If foreclosure and median income correlate with low lending, provide fiscal support to underwrite inherent risk of lending Already mission of CDCs in Worcester; broadly mission of federal HUD Possible to create municipal/state funding mechanisms to augment HUD/CDC mandate in Worcester and elsewhere

Questions and Comments?