The Economic and health consequences of lead paint abatement regulations

The Economic and Health Consequences of Lead Paint Abatement

Regulations

by

Ludovica A. Gazze

M.Sc. Economics and Social Sciences, Bocconi University (2010)

B.A. Economics and Social Sciences, Bocconi University (2008)

Submitted to the Department of Economics

in partial fulfillment of the requirements for the degree of

Doctor of Philosophy

at the

MASSACHUSETTS INSTITUTE OF TECHNOLOGY

June 2016

MASSACHUSETTS INSTITUTE

OF TECHNOLOGY

JUN 10 2016

LIBRARIES

ARCHIVES

@

2016 Ludovica A. Gazze. All rights reserved.

hereby grants to MIT permission to reproduce and distribute publicly paper

and electronic copies of this thesis document in whole or in part.

Signature redacted

The author

Author

Certified by. .

Certified by. .

Accepted by..

Signature redaci

Signature redact

.... ... ...

Department of Economics

May 15, 2016

Led

.. ... ... ...

Joshua D. Angrist

Ford Professor of Economics

ed

Thesis Supervisor

.

... .... ...

Benjamin A. Olken

Professor of Economics

Thesis Supervisor

Signature redacted---

_...

_...

J. Caballero

Ford International Professor of Economics

The Economic and Health Consequences of Lead Paint Abatement Regulations

by

Ludovica A. Gazze

Submitted to the Department of Economics on May 15, 2016, in partial fulfillment of the

requirements for the degree of Doctor of Philosophy

Abstract

This thesis consists of three chapters on the economic and health consequences of lead paint abatement regulations.

The first chapter studies the effect of state-level lead paint abatement regulations on housing markets, focusing on house prices, rents and the allocation of households across houses with different health risks. State lead abatement mandates require owners of old houses to mitigate lead hazards in the presence of small children. I estimate the effects of these mandates on the housing market using a triple differences strategy that exploits differences by state, year, and housing vintage. The estimates suggest a large fraction of the abatement costs fall on property owners, with house prices for multi-family properties declining by 6.4% and single-family homes declining 4.3%. These effects persist for at least a decade, consistent with low abatement rates. Families with small children bear part of the mandates' costs, too: after a mandate, these families are 17% less likely to live in old houses, and they pay higher rents for safer homes. These results suggest that the mandates have important real distributional consequences despite evidence of low abatement rates.

The second chapter analyzes the impact of state-level lead paint abatement regulations on children's health and educational outcomes in a difference-in-differences framework.. Lead poisoning has long-lasting conse-quences on children's health, as well as on their cognitive and non-cognitive abilities. Abatement mandates reduce the rate of elevated blood lead levels by 29%. Moreover, the mandates decrease the rate of enroll-ment in special education in exposed cohorts by 8.1%, indicating a reduction in the number of children with disabilities. A back of the envelope calculation suggests that this decrease in the rate of enrollment in special education induces savings between $17.5 and $111 million per state-cohort on average, while the increased lifetime earnings from the reduction in blood lead levels could lead to increased tax revenues in the order of $2.8 million per state-cohort on average. However, the reduction in special education enrollment does not appear to be reflected in improvements in educational outcomes, as I find no evidence that average fourth-grade test scores and disciplinary actions change with the mandates.

The third chapter analyzes both the selection of houses into compliance and individuals' valuation of lead-safe housing using address-level housing, environmental and health data. Lead paint in old homes is the major source of lead poisoning for US children despite federal and local regulations concerning the mitigation of lead hazards and the disclosure of lead status of a house whenever known. By leveraging detailed information on property owners, I find that small landlords are more likely to have non-compliant properties. Distressed landlords as indicated either by high loan to value ratios or by distressed sales have similar bad outcomes. By leveraging sales data, I estimate that the finding of a lead hazard significantly decreases the value of a house.

Thesis Supervisor: Joshua D. Angrist Title: Ford Professor of Economics

Thesis Supervisor: Benjamin A. Olken Title: Professor of Economics

Contents

Acknowledgements

List of Tables

List of Figures

1 The Price of a Safe Home: Lead Abatement Mandates and the Housing Market 1.1 Introduction . . . . 1.2 A Model of Abatement . . . . 1.2.1 Set-up . . . . 1.2.2 Abatement Mandate . . . . 1.2.3 Discrimination . . . . 1.2.4 Information . . . .

1.3 Background and Data . . . .

1.3.1 Regulatory History of Lead Paint

1.3.2 D ata . . . .

1.4 Empirical Analysis: Prices and Allocation

1.4.1 Sale Prices . . . .

1.4.2 Heterogeneous Price Effects . . .

1.4.3 Rents . . . . 1.4.4 Allocation . . . . 1.4.5 Mechanisms . . . . 7 9 12 16 . . . . 16 . . . . 19 . . . . 19 . . . . 20 . . . . 22 . . . . 24 . . . . 24 . . . . 24 . . . . 27 . . . . 29 . . . . 3 1 . . . . 33 . . . . 34 . . . . 3 5 . . . . 36 4

1.4.5.1 Discrimination . . . .

1.4.5.2 Information . . . .

1.5 Empirical Analysis: Families' Expenditures

1.6 Conclusion . . . .

Figures . . . .

Tables . . . .

2 Lead Policies, Lead Poisoning, and G Introduction Background 2.2.1 Lead 2.2.2 Lead 2.2.3 Speci 2.3 Data . . . . 2.4 Lead Poisoni 2.5 Child Disabil 2.6 Education El 2.7 Infant Health 2.8 Conclusion Figures . . . . . Tables . . . .

Poisoning, Fertility, and Abatement Mandates . a1 Education . . . . .. Eucaio... ng Effects . . . . ity Effects . . . . Tects . . . . and Fertility Effects .

. . . .

. . . .

. . . .

3 Selection into Environmental 3.1 Introduction . . . . 3.2 Background . . . . 3.3 Data . . . . 3.3.1 Data Sources . . . . 3.3.2 Data Linkage . . . . overnment Spending

. . . .

. . . .

I

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

Compliance: Evidence from Lead Abatement Mandates

. . . .

. . . .

. . . .

. . . .

3.4 Empirical Analysis: Selection into Inspection and Compliance . . . 100

5 2.1 2.2 37 37 38 39 41 45 56 56 59 59 60 61 62 65 67 70 71 72 74 79 92 92 94 96 96 99

3.5 Empirical Analysis: Value of Abatement . . . .

3.6 Empirical Analysis: Turnover . . . .

3.7 C onclusion . . . .

Figures . . . .

Tables . . . .

References

Appendices for Chapter

1

Appendix Figures

Appendix Tables

Appendices for Chapter 2

Appendix Figures

Appendix Tables

Appendices for Chapter 3

Appendix Tables 6 102 105 106 108 115 127 135 137 141 152 153 155 167 169

Acknowledgements

I am extremely grateful to the members of my committee, Josh Angrist, Ben Olken, and Jim Poterba for their invaluable advice and guidance throughout the dissertation writing process. Always generous with their time, they have taught me how to recognize valuable research questions and how to craft relevant projects through constructive discussions at each stage of my research.

Josh Angrist and Ben Olken have both shaped my research from the very beginning of my experience at MIT, while Jim Poterba has been the last addition to my committee, but by all means not the least. Josh has made sure not only that my econometrics is sound, but also that my writing is decent. Ben has patiently followed me in my exploration of several fields of economics, and his work discipline has been invaluable to me in defining a research agenda that I am passionate about. Jim has shared his public finance wisdom to this novice with great enthusiasm, questioning everyone of my arguments to make them tighter. I genuinely wish that we had cross paths earlier, but now I know that I can count on his mentorship going forward.

MIT Economics proudly advertises its open door policy. Indeed, I am truly indebted to the many faculty that were always available to share their expertise and discuss ideas. I am expecially thankful to the Labor and PF faculty for their insightful questions and even more insightful comments. During my third year of graduate school, Michael Greenstone inspired me to work in environmental issues. His departure from MIT has not stopped me from seeking his advice, and I am excited to be working with him in the future. Moreover, I am indebted to Bill Wheaton and Albert Saiz for introducing me to the wonders of Real Estate Economics. Finally, I am extremely thankful to Sloan faculty members Joe Doyle for welcoming me to discuss our common interests and Tavneet Suri for welcoming me to discuss anything, really.

The research in this dissertation would not have been possible without the help of several individuals and agencies who shared their data with me, and then were stuck with answering all my questions. Special thanks to the Taubman Center for State and Local Government for providing access to the DataQuick data repository when I was an exchange scholar at the Harvard Kennedy School; to MDPH, especially Paul Hunter, Masha Fishbein, Terry Howard, and Bob Knorr, for sharing their data; to MDE, especially Tontalia Stinney and Jonathan Klanderud for sharing their data; to NC DHHS, especially Tena Hand, for sharing their data; to NJDOH, especially Jaydeep Nanavaty, for sharing their data; to Daniel Sheeham and the staff members at the MIT GIS Lab for their help in working with the GIS data; and to Sergio Correia and Michael Stepner for sharing their codes for REGHDFE and MAPTILE, respectively.

For Chapter 1, I also thank Sally Hudson, Josh Krieger, Matt Lowe, Scott Nelson, Daan Struyven, Melanie Wasserman, and participants in the MIT PF/Labor seminar and MIT Labor lunch for their comments and suggestions. For Chapter 2, I also thank Elizabeth Setren and participants in the MIT PF Lunch and NBER Children and Education programs. Previous drafts of Chapter 2 circulated under the title "Little Lead Soldiers: Lead Poisoning and Public Health".

Writing a dissertation is a lonely endeavor, but I wouldn't have produced a single page without my peers. I am proud of our efforts to be open about our struggles and our health. Various generations of MIT students shared their experiences with me, cheered me up during a bad day, walked me through math derivations, or simply stopped to ask me how I was doing. Among these, honorable mentions go to the following people: The founding members of the MIT gang, Angela Kilby and Brendan Price, and honorary member Josiah Seale for honoring me with their friendship.

My officemates, Jie Bai, Hoai-Luu Nguyen, and Maheshwor Shrestha, who never complained about this loud woman calling her data names in Italian. Their patience is the proof of the existence of altruistic motives. The strong women with whom I could open up and discuss identification strategy and personal priorities without fear of judgement: Arianna Ornaghi, Sarah Moshari, Manisha Padi, Yufei Wu.

The Italian connection with whom I share a sense of humor and a long history: Enrico Cantoni, Tommaso Denti, Francesco Lin, Giovanni Reggiani, Annalisa Scognamiglio.

Alex Bartik for our weekly coffee meetings that helped me be on track and stimulated new ideas. One day our parallel research tracks have to intersect!

I am similarly indebted to the department's support staff who always have an answer and a smile. Thomas Dattilo, Eva Konomi, Ruth Levitsky, Beata Schuster, Aaron Sullivan, and Annie Weiss smoothed out the job market craziness for me. Emily Gallagher's kind words helped me remember that human interactions are still more important than stars in a regression table.

My life would be incomplete without my friends outside the Economics bubble. The old ones across the ocean constantly remind me why I do what I do. Grazie a Marta Arniani, Alberica Bazzoni, Carlotta Cossutta, Alessia De Stefani, Nusha Viola Ghironi, Sarah Parolin, Livia Piazza, Camilla Pietrabissa, Nura Tafeche, Francesca Tonelli. The old ones in town-ish, Haiwen Chen, Rebecca Compton, Fiona Cunningham, Luca De Angelis, and Caterina Scaramelli, are there to celebrate or cry with me or just hug me. The new ones teach me everyday to smile whenever an "obstacle" comes my way: I am fortunate to be part of the CAQ family. I am particularly thankful to Ian Gafanhoto Fein and Gabe Gigante Malseptic for teaching me to be patient with myself - _{or trying to. And I am grateful to have started this journey with Maggie Bon Bon Georgieva,}

Jaimie Cegonha Remillard, Adam Geode Tavares, Rina Vespa Thomas, and Clara Sombrinha Vandeweerdt.

There are hardly words to express my gratitude for my family. My mother's devotion to her work as an ob-gyn has taught me the meaning of service and has been an inspiration for putting human beings at the center of my research. I can only start to imagine the strength she needed to let her only child find her place in the world. My extended family has been a source of support for both of us throughout my studies at MIT: I owe them the peace of mind necessary to pursue my Ph.D. My final thoughts go to my father, who opened the doors of the world to me.

List of Tables

State-Level Abatement Mandates . . . .

Price Effects . . . .

Price Effects, Mandates Targeting Only Rental Properties . .

Price Effects by Tracts' Characteristics . . . .

Rental Market Effects, Extensive and Intensive Margins . . .

Allocation Effects . . . .

Allocation Effects, DataQuick-HMDA Sample . . . .

Rent Effects by Number of Bedrooms and Children's Presence

Price Effects Before and After Title X . . . .

Tenancy Effects . . . . . . . . 46 . . . . 47 . . . . 48 . . . . .. ... 49 . . . . 50 at Baseline 51 52 53 54 55 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 1.10 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 2.10 79 80 81 82 83 84 85 86 87 88

9

State-Level Abatement Mandates . . . .

Links between Disabilities and Lead Poisoning . . . .

Lead Poisoning Effects, Rates of Elevated Blood Lead Levels . . . . .

Effects of the Mandates on Special Education Needs . . . .

Effects of the Mandates on Special Education Needs, by Disability

Effects of the Mandates on Special Education Needs, Steady State

Effects of the Mandates on Special Education Needs, By Race . . . . .

Effects of the Mandates on Test Scores, Mathematics . . . .

Effects of the Mandates on Test Scores, Reading . . . .

2.11

2.12

2.13

Effects of the Mandates on Infant Health . . . .

Effects of the Mandates on Fertility . . . .

Effects of the Mandates on Maternal Characteristics . . . .

3.1 Characteristics of Units in the Massachusetts Linked Sample . . . .

3.2 Characteristics of Units with an Inspection or BLL Record . . . .

3.3 Characteristics of Inspected and Abated Units, MA . . . .

3.4 Characteristics of Units with Children with EBLL and Repeat-Offender Units . . . .

3.5 Summary Statistics: Lead Status . . . . 3.6 Predictors of BLL Record or Inspection . . . . 3.7 Predictors of BLL Record, Inspection, or Abatement . . . . 3.8 Predictors of Bad Inspection Outcomes or EBLL, and of Repeat Offense . . . . 3.9 Effect of Identifying a Lead Hazard on Property Values, IV . . . . 3.10 Effect of Identifying a Lead Hazard on Property Values, RD . . . . 3.11 Effect of Identifying a Lead Hazard on Property Values, First Stage RD . . . . 3.12 Effect of Identifying a Lead Hazard on Households' Mobility . . . . 3.13 Effect of Identifying a Lead Hazard on Households' Mobility by Tenancy Status . . . . A 1.1 Sum m ary Statistics . . . . A1.2 Price Effects for Multi-Family Properties, Alternative Specifications . . . . A1.3 Price Effects for Multi-Family Properties, Alternative Sets of Fixed Effects . . . . A 1.4 Sale E ffects . . . .

A1.5 Price Effects, by Occupancy . . . . A1.6 Price Effects, 5- and 10-Year Windows around Mandates . . . . A1.7 Price Effects by Year of Construction . . . . A1.8 Allocation Summary Statistics . . . .. . . . . A1.9 Allocation Effects, Robustness Checks . . . . A1.1(Allocation Effects by Housing Structure . . . .

A1.lfTenancy Effects by Income Quartiles . . . .

10 . . . . 89 . . . . 90 . . . . 91 115 116 117 118 119 120 121 122 123 124 124 125 126 141 142 143 144 145 146 147 148 149 150 151

A2.1 State Financing Systems for Special Education Services . . . .1

A2.2 Lead Screening Effects, Rates . . 159

A2.3 Lead Screening Effects, Counts . . . 160

A2.4 Lead Poisoning Effects, Counts of Elevated Blood Lead Levels . . . 161

A2.5 Costs to Society associated with EBLLs . . . 162

A2.6 Effects of the Mandates on Special Education Needs, Counts . . . 163

A2.7 Effects of the Mandates on Infant Health, by Mother's Characteristics . . . 164

A2.8 Effects of the Mandates on Fertility, by Mother's Characteristics . . . 165

A2.9 Effects of the Mandates on Fertility, Older Housing Stock . . . 166

A3.1 Predictors of BLL Record or Inspection . . . 169

A3.2 Predictors of BLL Record, Inspection, or Abatement . . . 170

A3.3 Predictors of Bad Inspection Outcomes or EBLL, and of Repeat Offense . . . 171

A3.4 Effect of Identifying a Lead Hazard on Property Values, BLL=5 Threshold... . . . 172

11 . . . .156

List of Figures

Equilibrium with Abatement . . . .

Equilibrium with Discrimination, Market for Leaded Houses . . . .

Equilibrium with Information on Lead Risks . . . .

Transaction Data Coverage . . . .

Price Effects . . . .

Price Effects, By Year of Construction . . . .

Sorting into Old Houses, By Income and Family Status . . . .

Allocation Effects: Child Under Six . . . .

Allocation Effects: Buyer's Income . . . .

Rent Expenditure Effects, by Family Status . . . .

2-1 Distribution of Children Served under IDEA by Age Group and Disability

2-2 Distribution of Children Served under IDEA by Age Group and Disability

Lead Poisoning Effects, Rates of BLLs above 10pg/dL, Alternative

Effect of the Mandates on Special Education, by Age at Mandate

Effects of the Mandates on Infant Health . . . .

Effect of the Mandates on Fertility . . . .

The Black-White Test Score Gap over Time, Mathematics . . . . .

The Black-White Test Score Gap over Time, Reading . . . .

Sample Lead Inspection Form, MD . . . .

Sample Letter of BLL above Level of Concern, MA . . . .

Specifications . ... ... 21 . . . . 23 . . . . 24 . . . . 41 . . . . 41 . . . . 42 . . . . 42 . . . . 43 . . . . 43 . . . . 44 74 . . . . 74 75 75 76 . . . . 77 . . . . 77 78 . . . 109 . . . 1 1 0 12 1-1 1-2 1-3 1-4 1-5 1-6 1-7 1-8 1-9 1-10 2-3 2-4 2-5 2-6 2-7 2-8 3-1 3-2

. . . . .

3-3 Sample Letter of BLL above Level of Concern, MA . . . .

3-4 Units with Identified Lead Hazards and Abated Units over Time, MA . . . .

3-5 Trends in Sale Price Relative to the Identification of a Lead Hazard . . . .

3-6 Abatement Costs and Number of Abated Units over Time, MA . . . . 3-7 BLLs Frequencies, MA . . . . 3-8 Probability of Lead Hazard and House Price by Blood Lead Level . . . . A1.1 Enforcement, M A . . . . A1.2 Enforcement, OH . . . . A1.3 Share of Old Houses in Wayne County, Michigan . . . . A1.4 Median Age of the Housing Stock by Tract as of 2010 . . . .

A1.5 Correlation between Age of the Housing Stock and Demographics at the Tract Le A1.6 Price Gap between Old and New Houses, By State . . . .

. . . 111 . . . 112 . . . 112 . . . 113 .. ...113 . . . 114 . . . 137 . . . 137 . . . 138 .139 ,el . . . 139 . . . 140 A1.7 Allocation Effects: Occupancy . . . . A2.1 Number of Children below 72 Months of Age with Elevated Blood Lead Levels . . . . A2.2 Effects of the Mandates on Infant Health . . . .

13

140

153 154

Introduction

This thesis characterizes the effects of state-level lead laws on the housing market, public health, and govern-ment spending. Environgovern-mental scholars deem lead poisoning the number one (preventable) environgovern-mental health threat to American children.1 Lead hazards caused by deteriorated paint in homes built prior to 1978, and especially prior to 1950, are the main source of child lead poisoning in the US (Brown and Jacobs, 2006). As such, local, state and federal government agencies face the the following optimization problem: eradicate lead poisoning at minimum cost. HUD (2011) estimates that nationwide, lead paint still lingers in 37 million houses; therefore, deleading the entire housing stock appears to be prohibitively expensive. More-over, it is unclear that it would be efficient to delead the 31 million houses with lead paint but no children residents: in the short run, these houses do not present a threat to public health, although market turnover and idiosynchratic preferences for location-specific amenities might drive families with small children into these homes in the future. As a result, 19 states have enacted regulations that mandate the mitigation or elimination of lead hazards in old houses in the presence of small children. In addition, HUD and state-level agencies provide subsidized loans and grants for deleading.

These abatement mandates attempt to regulate a complex market with highly heterogeneous goods which constitute a large share of households' expenditures. As any other regulation concerning the private pro-vision of a good that has potential ripercussions on public health, these mandates neetd to be incentive-compatible. In other words, enforcement and penalties for non-compliance need to be strict enough to make non-compliance a dominated option. In addition, the government needs to ensure that agents are able to pay for the costs of abatement.

By using a triple differences methodology and house-level data on transactions and the housing market allocation, Chapter 1 finds that old houses depreciate substantially after the introduction of an abatement mandate. Moreover, families with small children appear to move out of old homes after a mandate, and suggestive evidence points to discrimination on the part of landlords as one factor in the reallocation of families with small children into new homes. Neither of these facts is consistent with the mandates inducing high abatement rates, because abated houses would see their value increase and would attract families with small children. By exploiting house-level data on lead inspections and children's blood lead level, Chapter

'See, e.g., https://blog.epa.gov/blog/212 10/ 10587/

3 confirms that only 28% of non-compliant houses eventually get abated in Massachusetts, the first state to implement an abatement mandate in 1971. In particular, it appears that landlords who own five or fewer properties and owners in distress as measured by their loan to value ration are less compliant than the average property owner, which suggests that liquidity constraint might play a large role in the decision to delead a house.

As a result of the reallocation in the housing market, Chapter 1 shows that rental expenditure for families with small children increases after the introduction of a mandate. This finding raises the question of who should bear the cost of rehabilitating the US housing stock to reduce lead hazards. In this context, it becomes clear that issues of affordable housing are inevitably linked with environmental and health justice. Pollution levels are a significant factor determining house values in hedonic models. Hence, families of low socioeconomic status disproportionally live in neighborhoods with dilapidated housing and environmental hazards, as they cannot afford to live in safe and adequate conditions. To the extent that Medicaid, special education services, the Social Security Administration and other government agencies cover part of the medical, educational, and professional consequences of living in unsafe housing, unsafe housing generates large fiscal externalities on the whole population.

By using a difference-in-differences methodology and state-level health and special education data, Chapter 2 presents evidence on the health effects of the mandates as well as on some of the fiscal effects of these regulations. Although Chapters 1 and 3 suggest the mandates do not spur much additional abatement, Chapter 2 finds large reductions in lead poisoning rates after the introduction of these regulations. A back of the envelope calculation cannot rule out that the reduction in the number of children with elevated blood lead levels could be entirely due to the reallocation of families with small children into new homes. The improvements in children's health appear to translate into substantially lower rates of child disability as measured by enrollment in special education. These findings indicate that the mandates have large fiscal benefits and free up resources that could be employed to subsidize further deleading. In particular, given the findings in Chapter 3, a promising intervention would include expanding access to subsidized funds for deleading to small investor owners.

Chapter 1

The Price of a Safe Home: Lead

Abatement Mandates and the Housing

Market

1.1

Introduction

The Centers for Disease Control and Prevention (CDC) estimates that 535,000 children born in the US in the 2000s suffered from lead poisoning (Wheeler and Brown, 2013), a condition that is associated with reduced IQ (Ferrie et al., 2015) and educational attainment (Currie et al., 2014; Reyes, 2015b) and an increased risk of criminal activity (Reyes, 2007, 2015a; Nilsson, 2009; Feigenbaum and Muller, 2015).1 In the first half of the last century, however, lead paint was extensively used for residential purposes: in fact, HUD (2011) estimates that nationwide, lead paint still lingers in 5.5 million houses inhabited by small children, the population most at risk for lead poisoning, resulting in lead hazards in 3.7 million homes inhabited by small children, or 21% of houses with small children (Dewalt et al., 2015). Beginning in 1971, a growing recognition of lead hazards motivated an increasing number of states to mandate abatement, i.e., control, and, in certain cases, elimination of lead hazards in older units inhabited by children. However, abatement is expensive: Koppel and Koppel (1994) estimate that it can cost between $500 and $40,000 per unit, depending on the extent of the lead hazard. Unsurprisingly, not all owners comply with the mandates: 1.5 million houses were

'This figure refers to children with blood lead levels (BLL) above 5p.g/dL. In 1991, the CDC defined BLLs> 10pg/dL as the "level of concern" for children aged 1-5 years. However, in May 2012, the CDC accepted the recommendations of its Advisory Committee on Childhood Lead Poisoning Prevention (ACCLPP) that the term "level of concern" be replaced with an upper reference interval value defined as the 97.5th percentile of BLLs in US children aged 1-5 years from two consecutive cycles of Natonal Health and Nutrition Examination Survey (NHANES). In general, the definition of elevated blood lead levels (EBLL) for regulation purposes changes across jurisdictions and over time.

abated between 1999 and 2006 (HUD, 2001, 2011),2 and families with small children complain that landlords discriminate against them to avoid abatement (Berman et al., 2013; Williams, 2010). A such, the mandates, and environmental regulations more generally, might be void, or even have counterproductive effects, if they are not incentive compatible (Duflo et al., 2013).

This paper analyzes the effect of state abatement mandates on the housing market equilibrium, providing a novel and extensive characterization of the incidence of these policies. Prior to the mandates, rich families with small children appear to sort into new houses, suggesting that households know about lead hazards and trade off consumption and health.3 After the introduction of an abatement mandate, we expect families with children to move into old houses as these houses become lead-safe and appreciate. If, however, compliance rates are low, and/or if owners discriminate against families with small children as a consequence of the mandate, old houses depreciate, and families with small children move into new ones. In this paper, I study which equilibrium prevails in st-ates that implement mandates by asking two questions: Who bears the costs of the mandates, owners or renters? Who lives in houses that are likely to contain lead?

To answer these questions, I combine a rich set of data on housing market outcomes and use a triple differences approach to compare outcomes for old and new houses within a state before and after a mandate's introduction. This comparison is informative because lead regulations specifically target old houses, which are more likely to have lead hazards. My empirical analysis proceeds in two steps. First, I use sales data, collected by DataQuick from public deeds, to estimate the effect of the mandates on house values; using building structure as a proxy for tenancy, a choice variable for owners, I investigate the effect of the mandates on units in multi- and single-family buildings separately to allow the mandates to affect rental and owner-occupied properties differently. Second, using data from the American Housing Survey (AHS), I estimate the effects of the' mandates on the probability that a family with a small child lives in an old house and on housing expenditures.

I find two sets of results. First, the costs imposed by the mandates are capitalized into lower home values: a unit in an old multi-family building depreciates by 6.4%, i.e., by $4.80 per square foot, or 60% of the average abatement cost, and this depreciation persists up to ten years.4 Old single-family properties persistently lose 4.3% of their value, and fewer of these units appear to enter the rental market after a mandate. Second, after a mandate, families with small children are 17% less likely to live in old houses than before. Together, these results suggest that owners do not immediately comply with the mandates: if they did, abated houses would appreciate, and families with children would move into these safer homes. While landlords bear part of the mandates' costs in terms of lower house prices, they pass along a portion of these costs to tenants with small children: rents increase by 6.4-7.4% for old, family-friendly units, a result that is not consistent with

2

The US Department of Housing and Urban Development estimates that their funding and enforcement efforts contributed

to making 420,000 houses lead-safe.

3I1 the paper, I use the terms "families" and "households" interchangeably, although I have observations at the household

level.

4

Because nationwide data on abatement projects are not available, I compute the average cost of abatement on data referring to projects conducted in Massachusetts in 2014 and funded by the US Department of Housing and Urban Development (HUD).

lower willingness to pay for old houses by families after a mandate. These changes in the housing market imply that after a mandate, some families with children face higher rents in old units, while others live in new and more expensive houses: in total, I calculate that families with children spend $400 (or 6.4%) more per year on rent for several years after the introduction of a mandate.

In related work, I find that the mandates decrease the probability of lead poisoning Gazze (2016a). However, the higher housing expenditures, spread over several years, appear to be of the same order of magnitude as the mandates' benefits on average for the families these regulations are intended to protect. Accordingly, considering the impact of the policy on the housing market changes the assessment of the mandates from a beneficial policy into a neutral one, on average, for families with small children. The costs borne by these families seem to indicate a failure of the Coase theorem (Coase, 1960): although the mandates supposedly empower families with small children with rights to abatement, lax enforcement and discrimination are such that owners mantain de facto control, and tenants pay them transfers to have houses abated or pay higher rents to move out of harm's way.

My findings suggest that it is important to characterize how abatement mandates change the housing market equilibrium in order to compute the net impact of these policies, in line with the vast literature on government mandates and their unintended consequences (Summers, 1989; Gruber, 1994; MaCurdy and McIntyre, 2001).5 By analyzing the incidence of the mandates, this paper provides some caveats to the work by Aizer et al. (2015), who show that Rhode Island's abatement mandate successfully decreased lead poisoning among African Americans, thus halving the black-white test score gap in the state. Moreover, I contribute to a broad literature that explores the health effects of pollutants and neurotoxins commonly found in homes (Currie et al., 2011; Cohen-Cole, 2006; Evans, 2006). Leventhal and Newman (2010) argue that research "on housing and children's development is still in its relative infancy" in terms of both methodology and theory; indeed, the evidence of the effectiveness of lead abatement is mixed, and its effectiveness depends heavily on the techniques used to abate.

The paper proceeds as follows. Section 1.2 outlines a model to show that the impact of a mandate on prices and allocation depends on the strength of enforcement and on the extent of owners' discriminatory behavior. Section 1.3 provides background on lead poisoning as well as on the regulations studied in this paper, describing the data I use. Section 1.4 estimates the impact of the mandates on hquse prices and the allocation of households across houses. Sections 1.5 discusses the impact of the mandates on families' expenditures. Section 3.7 concludes with policy implications.

5

See also Kuminoff et al. (2013) for a review of the literature on equilibrium sorting models.

1.2

A Model of Abatement

I use a simple infinite-horizon model of an urban rental housing market to derive two sets of predictions regarding the introduction of abatement mandates. First, the mandates always hurt owners of leaded homes. However, the effect of the mandates on the value of old properties overall depends on the strength of enforcement: the more houses are abated, the more old houses will appreciate. Second, families with children move into old units as they are abated, trading off consumption and health. However, if owners discriminate against families with small children, a mandate lowers the share of families with children in old houses and increases their housing expenditure. Similarly, a mandate lowers the share of families with children in old houses if it increases knowledge about lead hazards.

1.2.1

Set-up

Every period, a set of households of measure one optimize their consumption of a composite good, c, produced with a perfectly elastic supply at price pc = 1, and of housing services, h. Households do not save or borrow and have no other assets; therefore, their consumption is equal to their income net of the housing expenditure. Houses differ only in the presence of lead paint, and each household rents one housing unit at cost rh, where

h E {L, N,

0}.

Households vary across two dimensions: per-period income, yj

E

si = 1 indicates that the household has a small child. The household per-period maximization problem can

be written as follows:

maxhU(h; yi, si, a, rTL, rN) = io9(Yi - rh ) -

W

0)Ho

where

WI(h

small children, Ho is the disutility from the outside option, and Ho > a, > ao > 0. Although no one chooses the outside option, it pins down the rent levels in equilibrium. Hence, we can define

r

of leaded houses, L, relative to safe ones, N, and perform comparative static analysis on r.

By the concavity of utility of consumption, poor households sort into leaded houses: although everyone dislikes lead, poor households derive a higher marginal utility from the additional consumption they get by living in a leaded house and saving on rent. Moreover, for each level of income, families with small children are willing to pay more to live in a safe home. Hence, the demand for leaded houses is decreasing in r.

On the supply side, foreign landlords maximize the net present value of rental income. Assuming that

without the mandates, rents are constant over time as supply and demand of houses are constant, the price of a house can be written as follows:

Ph = NPV(rh) = ( = . (1.2)

t=O

where i is the market interest rate and without loss of generality, the maintenance costs are negligible. For simplicity, I assume a fixed supply of houses of measure one, a fraction of which have lead paint initially.6 An abatement technology turns leaded houses into safe homes at cost A, homogeneous across owners without loss of generality. Abatement is profitable if A is lower than the present value of the markup charged for safe houses, in which case all landlords want to abate. Hence, in an equilibrium with both leaded and non-leaded houses, the two values have to be equal.

1.2.2

Abatement Mandate

Unexpectedly, the government introduces an abatement mandate: with some probability r > 0, the owner of a leaded house receives a lead order requiring her to abate at cost A

M

legal expenses. After she abates, the owner will be able to reap higher rent for a safe home starting with the next period. The parameter 7 measures enforcement, i.e., the probability of a lead order, which can be smaller than the probability that a child lives in a leaded house. Indicating the equilibrium objects under the mandate with a superscript M, the values of non-leaded and leaded houses can be written as follows:

-(1+i)

NPVNM' -rN (1-3)

NPVt' = I + - NPV + _NPV'm - 1rAm (1.4)

1+z 1+i

In other words, the value of a leaded house equals its rent today plus its expected future value. With probability 1 - 7, the house won't be abated; with probability 7r, the owner receives a lead order and abates at cost AAI. The assumption that households are perfectly informed about lead hazards prevents the mandate from shifting demand for leaded houses. Normalizing rN without loss of generality and solving for the value of a leaded house recursively, I obtain:

NP VIV - (1 + 0)(iTM+ 7r) - A (1 + i)7r (1.5) L _{i(i + 7)}

(i + X)

where the first term in equation (1.5) is the expected stream of rents from a currently leaded house and the second term is the present value of abatement cost.

By a revealed preference argument, NPV-' < NPV: the mandate lowers the value of a leaded house by

6

The predictions in this section hold if we relax this assumption to allow for an elastic supply of non-leaded houses. By definition, developers cannot add old houses, and I assume that no demolition or renovation takes place. Below, I discuss how the model's intuition carries through if we allow owners to sell rental properties to owner-occupiers.

introducing an additional cost with positive probability 7r. If the drop in the value of leaded houses is big enough that NPVt" < NPVN - A, abatement becomes the best solution; if instead the drop in value is

small relative to the cost of abatement, owners will wait for law enforcement to require abatement, which yields qualitatively similar, albeit slower, changes. Which equilibrium prevails depends on the elasticity of demand, i.e., on the relative size of each group, p, and the joint distribution of income and family status. A high share of families with children makes enforcement more likely, incentivizing abatement.

Figure 1-1: Equilibrium with Abatement

Rent for Leaded Houses Rent for Non - Leaded Houses

SN/

S'

SL

SN

6V 0 Leaded Houses 1 -0 1 - 0' Non - Leaded Houses

The figure shows the equilibrium in the housing market after a mandate induces abatement. The left panel depicts supply and demand of leaded houses. Abatement reduces supply of leaded houses to S'. As leaded houses become more scarce,

their rent increases from rs to

T'.

Figure 1-1 shows how the mandate changes the housing market equilibrium. For ease of illustration, I plot supply for leaded and non-leaded houses as vertical, but the they are actually elastic depending on the abatement cost. As abatement reduces the number of leaded houses from 0 to 0', fewer people live in such houses. In particular, households with children move into abated houses, unambiguously increasing the share of children in old houses. In contrast, the effect of the mandates on relative rents is theoretically ambiguous. When landlords abate, the supply of non-leaded houses increases, and their relative rent decreases. Moreover, a market for owner-occupied units would allow landlords to sell their property to a homeowner, potentially at a lower price. Such sales decrease the total supply of rental houses, putting upward pressure on rents. However, outside the scope of this model, a mandate could also decrease rents for leaded houses: for instance, if the enforcement probability depends on tenants' complaints, then owners can "buy" silence by offering discounts.

In summary, this section highlights two facts. First, the mandates always hurt owners of leaded homes. The

assumption of a fixed supply of houses makes landlords more inelastic than tenants: even if rents increase due to the increased costs, this rise does not fully compensate owners. Indeed, it is possible for rents and prices to move in opposite directions, as the mandates introduce a wedge between the stream of future rents and the value of a property. Second, because compliance with the mandates and additional abatement procedures make older homes safe, families with small children are attracted to older houses.

1.2.3

Discrimination

In this section, I derive the effect of a mandate when owners discriminate against families with small children by charging them higher rents to account for the mandate's costs. Under discrimination, a mandate lowers the share of families with children in old houses. Technically, price discrimination only refers to markets for homogeneous goods, and houses are hardly homogeneous, but I use this term in its legal interpretation. For simplicity, I allow discrimination only under the mandate and only in the leaded segment of the market.7 Under this scenario, subscripted with D, the value of a leaded house is a weighted average of the value of renting to families with small children, NPVT5D(1), and without small children, NPVE(0), with weights given by the fraction of families with children in the population, p:

NPV = pNPVj (1) + (1- p)NPV (0) - OT (1.6)

where OT is the expected fine for discriminating. Given that abatement can only be triggered by a child living in the house, landlords face a null probability of abatement by discriminating. Letting c = be the probability of a lead order conditional on a child living in a leaded house, I obtain:

NPVL(1) = i+l NPV?+ .NPV -

eAA

1_N~

NPVL(0) =

0

1+i

where T₁ and ro are rents paid by families with and without small children, respectively. Plugging (1.7) and (1.8) into (1.6) and solving for the value of a leaded house, I have:

(1 + i) {i [tri + (1 - t)o] +} (1 + i)r(

NPVLD = .- AAuT -

OT

(1.9)

(i

The first term in equation (1.9) is the net present value of rents, a weighted average of rents paid by families with and without small children. The second term is the expected abatement cost, which depends on

7

The results in this section hold in the more general case in which discrimination is possible at all times and in all markets. The mandate increases the cost of providing leaded houses to families with small children, which will be reflected in rents. Landlords in the non-leaded sector take advantage of the increased demand for safe homes by families with children and raise rents for these households as well. Hence, the total change in the relative rent of leaded houses will be danpened, but the

direction of the change is the same.

Figure 1-2: Equilibrium with Discrimination, Market for Leaded Houses

Relative Rent for Leaded Houses T

S/ S1 SoSo, D1 Do Ti ... T ... .... ... ... To... ... ... Leaded Houses 0/ 01

0-010-01

The figure shows the leaded segment of the housing market equilibrium with an abatement mandate and price discrimination.

D1 and Si represent demand for and supply of leaded houses for families with small children, while Do and So represent demand for and supply of leaded houses for childless families. r is the relative price of leaded houses that would prevail without discrimination, given by the intersection of the demand curves and the solid supply lines. Dashed supply lines S' and S' illustrate the equilibrium with price discrimination, where rent for families with and without children are given by r, and ro, respectively.

enforcement. By a revealed preference argument, the mandate still lowers the value of leaded houses.8

If NPVE > NPVM, the mandate induces discrimination and lowers the share of families with children in

leaded houses. Moreover, under discrimination, the mandate increases the housing expenditures of families with children because they either move to safer and more expensive houses or pay higher rents for the same homes. Figure 1-2 illustrates the market for leaded houses under this scenario. Let D, and Do be the demand functions for leaded houses of households with and without small children, respectively. The solid lines Si and So are the quantities supplied to families with and without children when mandates are in place but price discrimination is not possible: in this case,

r

D₁ (r) + Do(r) = 0 and S3 =

Oj

j

E

VFor discrimination to be valuable, i + 7r - pt > 0 is required. A standard value for the interest rate, i = 0.02, and the population share of household with children, t = 0.15, yield c > 0.87. Such a high enforcement probability is unusual, but it is conditional on the presence of lead hazards in the house.

Figure 1-3: Equilibrium with Information on Lead Risks

Relative Rent for Leaded Houses T

t

SL

Ii~[,...

0

The figure shows the leaded segment of the market when information changes the demand for leaded houses. Information lowers demand for leaded houses to D' , decreasing their rent rL to I'

1.2.4

Information

In addition to shifting the supply of leaded houses, the mandates might provide information regarding the risks of lead poisoning for small children, which in turn might decrease families' willingness to pay for these houses. Figure 1-3 depicts the extreme scenario in which the mandates only act through this information channel, focusing on the leaded segment of the market for ease of illustration: DL represents the demand for leaded houses before the mandate. The mandate changes the perceived cost of lead poisoning for families with children to a, > ceo, making D' steeper. As a result, families with children move out of old houses, causing excess supply, and rent for old houses decreases until the market clears. As no abatement happens, there is no wedge between rents and home values, and old houses depreciate.9

1.3

Background and Data

1.3.1

Regulatory History of Lead Paint

Lead paint is commonly found in old houses in the US. Starting in the late 19th century, paint manufacturers typically used lead as an additive in residential paint to increase durability of the paint coat, with paint manufactured prior to 1950 containing up to 50% lead by weight (Reissman et al., 2001). In response to the

9

1t is possible that the change in demand and the resulting change in relative prices spurs voluntary abatement. For brevity,

I do not discuss this case.

24 1 'rL

\%QL

DL

growing body of evidence of the harm associated with lead, in the late 1950s, some manufacturers voluntarily reduced the lead content of paint to 1%, a level that can still induce severe lead poisoning (Hammitt et al., 1999). Finally, in June 1977, the Consumer Product Safety Commission (CPSC) lowered the allowed level of lead in paint to 0.06%, effectively banning lead paint altogether from 1978 on.1 0 Notably, the ban covers new paint, and not the pre-existing housing stock (Mushak and Crocetti, 1990). Moreover, unless the paint coat containing lead is removed, lead remains in a house indefinitely. As a result, the incidence of lead paint in the current housing stock increases with structures' age, from 8% for houses built in the 1970s, to 86% for homes built before 1940 (HUD, 2011).

When paint surfaces deteriorate, residents, and especially children, are exposed to health hazards from lead-contaminated dust. Lead dust enters the human system through ingestion or inhalation. Once in the bloodstream, lead impairs cognitive and non-cognitive ability at levels as low as 1 - 2ptg/dL, 80 times lower

than the level of concern for iron (DNTP, 2012): Lanphear et al. (2005) estimate an IQ loss of 3.9 points when BLLs increase from 2.4 to 10 pg/dL, with lower IQ decrements associated with further BLLs increases. These effects are irreversible, and treatment can only help prevent further accumulation of the toxin (Rogan et al., 2001). Small children are especially exposed to lead-contaminated dust from paint and windowsills due to normal hand-to-mouth activity, and they might grow accustomed to the sweet taste of lead paint (Fee, 1990).11 Moreover, lead is most damaging to small children: they absorb and retain more lead than adults and their neurological development is particularly susceptible to neurotoxins (see, e.g., McCabe, 1979).

As of today, 19 states have enacted abatement mandates, as summarized in Table 2.1.2 Although physicians had warned against the hazards of lead paint since the early 1900s (Ruddock, 1924), the first regulation banning the use of lead pigment for interior use in the US was only adopted in 1951 by the Baltimore Commissioner of Health (Fee, 1990). In 1970, the US Surgeon General released a policy statement calling for "adequate and speedy removal of lead hazards" from the homes of lead-poisoned children below six years of age (Steinfeld, 1971). Massachusetts was the first state to follow suit, introducing in 1971 one of the strictest lead paint regulations in the country, requiring property owners to permanently control lead paint hazards in the home of any child under the age of six.

In my analysis, I treat all mandates as homogeneous to increase statistical power. In results not reported in the paper, I find little evidence that the impact of the mandates depends on their characteristics. Nonetheless, the mandates differ in terms of their coverage, what triggers a lead order, and type of abatement required. For instance, some regulations cover all properties, whereas others focus on rentals; similarly, some states

1

Seven years earlier, in 1971, the Lead-Based Poisoning Prevention Act (LBPPA) directed the Secretaxy of Housing and Urban Development to "prohibit the use of lead-based paint in residential structures constructed or rehabilitated by the Federal Government, or with Federal assistance in any form after January 13, 1971." (LBPPA, 1971)

"For years, it was argued that only children affected from pica disorder, i.e., the persistent and compulsive craving to eat nonfood items, were subject to eating lead paint, and that careless parents were to blame for their children's lead poisoning. Markowitz and Rosner (2000) quote a letter, dated December 16, 1952, by an official of the Lead Industries Association asserting that childhood "lead poisoning is essentially a problem of slum dwellings and relatively ignorant parents" and that "until we can find means to (a) get rid of our slums and (b) educate the relatively ineducable parent, the problem will continue to plague us."

12

Regulations were identified with a search through LexisNexis and Westlaw.

require abatement as soon as a small child lives in the house, whereas others need a triggering elevated blood lead level (EBLL) case to issue a lead order, and the definition of EBLL changes across states and over time.13 Finally, the regulations differ in the extent of mitigation that they require and the protection of liability granted to owners who abate. 14

Anecdotally, enforcement of these mandates is lax, and abatement is slow. Unfortunately, there is little data on inspections, lead orders, and lead-safe certificates, and the existing figures are plagued by misreporting, as off-the-books voluntary lead inspections at sale are the norm in regulated states.1 5

Data from Maryland indicates that 200,000 units, i.e., only a third of rental properties, have been inspected and certified under the state law that requires all rental properties to be registered.'6 In addition, Appendix Figure A1.1 shows that even in states with strict regulations, like Massachusetts, inspections are rare. And deleading projects are even more infrequent: in Massachusetts, the number of abatement projects reported to the Department of Labor since June 2010 totals 7,500 properties, on a total of 2.8 million units, but not all of these projects have been finalized. Finally, in Michigan, which implemented a mandate in 2005, 584 abatement projects were reported to the Department of Health and Human Services in fiscal year 2015, while there have been an average of over 1,000 projects in the fiscal years 2009-2013 in Ohio, which implemented a mandate in 2003. Appendix Section 3.7 further explores the weak link between the mandates and inspections and abatement.

Aside from the state-level mandates I study in this project, both the federal and local governments have regulated lead paint. At the federal level, the Lead-Based Paint Poisoning Prevention Act of 1971 prohibits lead paint in federal homes and provides funds for deleading. In June 1977, the CPSC effectively banned lead paint for residential use in private homes from 1978 on. However, the first federal measure to deal with the existing stock of lead paint in older US homes was the Residential Lead-Based Paint Hazard Reduction Act of 1992 (Title X), which became effective on December 12, 1996. The act mandates disclosure of known information on lead hazards before the sale or lease of houses built prior to 1978. At a more localized level, city governments also deal with issues related to lead paint and may enact regulations that are stricter than their state's requirements.'7 To the extent that the timing of these city-level regulations is not correlated with the introduction of the state-level mandates, the lack of systematic information on local laws does not

affect the validity of my findings.

Various abatement techniques are available, and abatement costs vary wildly depending on the technique employed and the extent of the lead hazards present in a property. The US Department of Housing and

'3_{Furthermore, only a few states, such as Massachusetts, mandate universal blood lead screenings for children. In states} where lead inspections are triggered only by EBLLs, the inspection and abatement rates will depend on screening.

14

For instance, Vermont requires owners only to perform Essential Maintenance Practices, i.e., interim control practices that ensure the paint does not deteriorate. Furthermore, over time the states have changed the definition of abatement and of admissible interim control practices.

15

See any online forum for home-buyers for posts like the following, accessed on 07/16/2015 at http://www.city-data.com/foruin/massachusetts/1875347-ma-lead-law.html: "We hired a lead inspector anyway at the time of the home in-spection, and even though I was verbally given results nothing was disclosed to anyone."₁₆

Source: Author's calculation on data from the Maryland Department of the Environment.

1 7

See, e.g., Philadelphia Lead Paint Disclosure and Certification Law, effective on December 21, 2012, requiring landlords to ensure that property rented to families with children six years and younger is lead-safe. Notably, Pennsylvania has no state-level mandate.