An equilibrium analysis of low income property development ... · An equilibrium analysis of low...

23
Who wants affordable housing in their backyard? An equilibrium analysis of low income property development Rebecca Diamond & Tim McQuade Stanford GSB & NBER April 6, 2017 Diamond & McQuade (Stanford GSB & NBER) LIHTC Effects April 6, 2017 1 / 20

Transcript of An equilibrium analysis of low income property development ... · An equilibrium analysis of low...

Who wants affordable housing in their backyard?An equilibrium analysis of low income property development

Rebecca Diamond & Tim McQuade

Stanford GSB & NBER

April 6, 2017

Diamond & McQuade (Stanford GSB & NBER) LIHTC Effects April 6, 2017 1 / 20

Affordable Housing as Place-Based Policy

Increasing geographic income segregation & housing costs puts gov’s rolein promoting affordable housing at forefront of policy debate

Policy question of where to build affordable housing

Large Diversity in Geographic Targeting of Affordable Housing

SJ, SF, NYC incentivize 15-20% of new development affordable in same buildingLIHTC: Larger subsidy to poorest neighborhoods

Low-income housing may have equilibrium effects on neighborhood

Benefits to tenants in affordable housing must be considered along with costs andbenefits to surrounding neighborhood residents

Diamond & McQuade (Stanford GSB & NBER) LIHTC Effects April 6, 2017 2 / 20

Welfare effects of LIHTC on neighborhood residents

Quantify costs & benefits of affordable housing to surrounding neighborhoodresidents

Analyze affordable housing construction funded by Low-Income HousingTax Credit (LIHTC)

Develop housing choice model:

Identify preferences for LIHTC proximity by measuring LIHTC impacts onhouse prices and households’ housing purchases

Analyze neighborhood welfare impacts of LIHTC

Compare impacts across neighborhoods of different income and racialcomposition

Diamond & McQuade (Stanford GSB & NBER) LIHTC Effects April 6, 2017 3 / 20

Welfare effects of LIHTC on neighborhood residents

Quantify costs & benefits of affordable housing to surrounding neighborhoodresidents

Analyze affordable housing construction funded by Low-Income HousingTax Credit (LIHTC)

Develop housing choice model:

Identify preferences for LIHTC proximity by measuring LIHTC impacts onhouse prices and households’ housing purchases

Analyze neighborhood welfare impacts of LIHTC

Compare impacts across neighborhoods of different income and racialcomposition

Diamond & McQuade (Stanford GSB & NBER) LIHTC Effects April 6, 2017 3 / 20

Preview of Results

Reduced Form Prices Effects of LIHTC construction after 10 years:

6.5% price appreciation in lowest income areas

2.5% price depreciation higher income, low minority areas

No depreciation in high minority areas

Welfare Impacts

LIHTC in lowest income, low minority area:

Avg HH benefits $8,843Aggregate benefit of $116 Million

LIHTC in higher income, low minority area:

Avg HH benefits -$3,295Aggregate loss of $12 Million

Diamond & McQuade (Stanford GSB & NBER) LIHTC Effects April 6, 2017 4 / 20

Outline

1 Data

2 Model & Estimation Method

3 Reduced Form Estimates

4 Preference Estimates & Welfare

5 Conclusion/Next Steps

Diamond & McQuade (Stanford GSB & NBER) LIHTC Effects April 6, 2017 5 / 20

Data

DataQuick: 1988-2012

Transaction history & property characteristics collected from countyrecorders’ and assessors’ offices

Home Mortgage Disclosure Act (HMDA): 1991-2012

Merged with DataQuick to observe income of homebuyers

HUD Public Use LIHTC Database: 1987-2012

Contains information on 39,094 LIHTC projects

Project address, year funds allocated, number of units

Diamond & McQuade (Stanford GSB & NBER) LIHTC Effects April 6, 2017 6 / 20

Counties in Sample

Diamond & McQuade (Stanford GSB & NBER) LIHTC Effects April 6, 2017 7 / 20

Identifying LIHTC Treatment Effect on House Prices

Exploit highly local, micro aspect of our data:

Very local difference-in-differences strategy

Compare price changes close to LIHTC vs. price changes slightly further away

Key Assumption:

LIHTC construction uncorrelated with highly local price trends

Prices can’t trend differentially 0.1 miles vs 0.2 miles from LIHTC site other thanb/c of LIHTC developmentWe test this: No trends in of this nature before LIHTC construction

Diamond & McQuade (Stanford GSB & NBER) LIHTC Effects April 6, 2017 8 / 20

Impact of LIHTC on House PricesProjects located in Census Block Groups with Median Income < $26k

-0.01

1.2

0

0.01

1

0.02

10

0.03

Log

Hou

sing

Pric

e

0.8

0.04

Miles to LIHTC Site

0.6 5

0.05

Years Since Funding

0.06

0.40

0.2

0 -5

Diamond & McQuade (Stanford GSB & NBER) LIHTC Effects April 6, 2017 9 / 20

Impact of LIHTC on House PricesProjects located in Census Block Groups with Median Income $26k-$38k

-0.03

1.2

-0.02

1

-0.01

10

0

Log

Hou

sing

Pric

e

0.8

0.01

Miles to LIHTC Site

0.02

0.6 5

Years Since Funding

0.03

0.40

0.2

0 -5

Diamond & McQuade (Stanford GSB & NBER) LIHTC Effects April 6, 2017 10 / 20

Impact of LIHTC on House PricesProjects located in Census Block Groups with Median Income $38k-$55k

-0.04

1.2

-0.03

1

-0.02

10

-0.01

Log

Hou

sing

Pric

e

0.8

0

Miles to LIHTC Site

0.01

0.6 5

Years Since Funding

0.02

0.40

0.2

0 -5

Diamond & McQuade (Stanford GSB & NBER) LIHTC Effects April 6, 2017 11 / 20

Impact of LIHTC on House PricesProjects located in Census Block Groups with Median Income > $55k

-0.04

1.2

-0.03

1

-0.02

10

-0.01

Log

Hou

sing

Pric

e

0.8

0

Miles to LIHTC Site

0.01

0.6 5

Years Since Funding

0.02

0.40

0.2

0 -5

Diamond & McQuade (Stanford GSB & NBER) LIHTC Effects April 6, 2017 12 / 20

Impact of LIHTC on House PricesProjects located in Census Block Groups with Income < $38k, % Minority >50

-0.021.2

-0.01

0

1

0.01

10

Log

Hou

sing

Pric

e

0.8

0.02

0.03

Miles to LIHTC Site

0.6 5

0.04

Years Since Funding

0.05

0.40

0.2

0 -5

Diamond & McQuade (Stanford GSB & NBER) LIHTC Effects April 6, 2017 13 / 20

Impact of LIHTC on House PricesProjects located in Census Block Groups with Income > $38k, % Minority >50

-0.03

1.2

-0.02

1

-0.01

10

0

Log

Hou

sing

Pric

e

0.8

0.01

Miles to LIHTC Site

0.02

0.6 5

Years Since Funding

0.03

0.40

0.2

0 -5

Diamond & McQuade (Stanford GSB & NBER) LIHTC Effects April 6, 2017 14 / 20

Indirect Neighborhood Changes

Lower income LIHTC development attracts 5% higher income homebuyers

Higher income, low minority LIHTC development attracts 1.2% lowerincome homebuyers

Development in high minority areas decreases black share of homebuyersby 3 percentage points → less racial segregation

Property and violent crime fall in low income areas

Diamond & McQuade (Stanford GSB & NBER) LIHTC Effects April 6, 2017 15 / 20

Welfare Analysis

Quantify welfare benefits of local neighborhood change

Decompose into impacts on:HomeownersRentersAbsentee landlords

Assumptions for renter welfareHouse prices equal to present value of rentsRenter preferences same as homeowners given race and income

Assume no moving costs

Diamond & McQuade (Stanford GSB & NBER) LIHTC Effects April 6, 2017 16 / 20

Mean Utility Benefit per Household: Low Minority Areas

Med Income <26K Med Income >55k

Average Benefit perHome Owner

23,402 -3,972

Average Benefit perRenter

6,502 67

Average Benefit perLandlord

6,011 -2,416

Mean welfare benefit from LIHTC construction to households who choose to live nearby. Utility is

measured in 2012 dollars.

Diamond & McQuade (Stanford GSB & NBER) LIHTC Effects April 6, 2017 17 / 20

Aggregate Utility Benefit: Low Minority Areas

Med Income <26K Med Income >55k

Aggregate Benefitto Home Owners

57.9 -9.0

Aggregate Benefitto Renters

29.2 .78

Aggregate Benefitto Landlords

29.0 -3.2

Aggregate Benefitto Society

116.2 -12.1

Aggregate welfare benefit from LIHTC construction to society. Utility is measured in million of 2012

dollars.

Diamond & McQuade (Stanford GSB & NBER) LIHTC Effects April 6, 2017 18 / 20

Incorporating Tenant Welfare

Our analysis considers only place based effects of LIHTC development

Incorporate the welfare effects on affordable housing tenantsUse MTO(Chetty, Hendren, & Katz, 2015): $99,000 PV per childWelfare benefit to tenants is $26.7 million.

Opportunity cost of not building in low income neighborhood: $116million.

Place-based effects lead to larger benefits for low-incomepopulation

Diamond & McQuade (Stanford GSB & NBER) LIHTC Effects April 6, 2017 19 / 20

Conclusion

Neighborhood impacts of affordable housing vary substantially byneighborhood income levels

LIHTC in poor areas improves neighborhoodLIHTC in richer areas lowers local neighborhood quality

Affordable housing programs often criticized when housing located inalready poor or high minority neighborhoods

80/20 mixing of market rate and affordable housing (SF & NYC)Large opportunity costs of not improving low income communities

Diamond & McQuade (Stanford GSB & NBER) LIHTC Effects April 6, 2017 20 / 20

Mean Utility Benefit per Household: High Minority Areas

Income Q1 Income Q2 Income Q3 Income Q4

Average Benefit perHome Owner

16857 2414 996 3255

Average Benefit perRenter

6475 190 342 971

Average Benefit perLandlord

6099 1288 375 1090

Mean welfare benefit from LIHTC construction to households who choose to live nearby. Utility is

measured in 2012 dollars.

Diamond & McQuade (Stanford GSB & NBER) LIHTC Effects April 6, 2017 20 / 20

Aggregate Utility Benefit: High Minority Areas

Income Q1 Income Q2 Income Q3 Income Q4

Aggregate Benefitto Home Owners

63,460 1,446 3,615 14,508

Aggregate Benefitto Renters

73,417 314.7 1,903 4,861

Aggregate Benefitto Landlords

74,236 2,672 2,262 5,907

Aggregate Benefitto Society

211,113 4,433 7,780 25,277

Aggregate welfare benefit from LIHTC construction to society. Utility is measured in thousands of 2012

dollars.

Diamond & McQuade (Stanford GSB & NBER) LIHTC Effects April 6, 2017 20 / 20