Do Wind Hazard Mitigation Programs Affect Home Sales Values?1
Daniel R. Petrolia (corresponding author) Professor
Department of Agricultural Economics Mississippi State University d.petrolia@msstate.edu / 662.325.2888
Shea Ishee
Director of Information Development American Cotton Shippers Association shea@acsa-cotton.org
Seong D. Yun Assistant Professor
Department of Agricultural Economics Mississippi State University seong.yun@msstate.edu
J. Reid Cummings
Associate Professor of Finance and Real Estate Department of Economics, Finance, and Real Estate University of South Alabama cummings@southalabama.edu
Joshua G. Maples Assistant Professor
Department of Agricultural Economics Mississippi State University josh.maples@msstate.edu
Acknowledgments: We thank the staff at the Insurance Institute for Business & Home Safety for providing access to the FORTIFIED dataset and Lars Powell for his assistance in obtaining it. We also thank the staff at Smart Home America and the Mississippi-Alabama Sea Grant Consortium for their assistance throughout this study. This publication was supported by the U.S. Department of Commerce’s National Oceanic and Atmospheric Administration under NOAA Award NA18OAR4170080 and the Mississippi-Alabama Sea Grant Consortium. This work was also supported by the National Institute of Food and Agriculture and the Mississippi Agricultural and Forestry Experiment Station via Multistate Project W-4133 “Costs and Benefits of Natural Resources on Public and Private Lands: Management, Economic Valuation, and Integrated
1 Data provided by Zillow through the Zillow Transaction and Assessment Dataset (ZTRAX). More information on accessing the data can be found at http://www.zillow.com/ztrax. The results and opinions are those of the authors and do not reflect the position of Zillow Group.
Decision-Making” (Hatch Project MIS-033140). The views expressed herein do not necessarily reflect the views of any of these organizations.
Do Wind Hazard Mitigation Programs Affect Home Sales Values?2
Abstract
The FORTIFIED Home Program is a voluntary construction and re-roofing program designed to strengthen residential properties to withstand damages from severe weather events. The FORTIFIED designation framework can accommodate sustainable new home construction and existing property retrofitting efforts. Applying a two-state utility framework to Zillow’s ZTRAX and IBHS’s FORTIFIED designation data, we estimate the effect of a FORTIFIED Home designation on residential property values using sales transactions for coastal Alabama from 2011 to 2021. We add to the growing body of literature examining the relationship between eco- friendly construction applications and home prices by deriving the contribution of a FORTIFIED designation to enhanced risk reduction and lowered insurance costs. The results of our full- and nearest-neighbor-matched sample design suggest a likely price premium significance on the order of 2-4% for FORTIFIED homes.
Keywords: FORTIFIED Home Program; Sustainability; Green design; Wind hazard- mitigation; Hedonic price model; Residential property values; Building certifications
2 Data provided by Zillow through the Zillow Transaction and Assessment Dataset (ZTRAX). More information on accessing the data can be found at http://www.zillow.com/ztrax. The results and opinions are those of the authors and do not reflect the position of Zillow Group.
Introduction
Hurricanes, tornadoes, hailstorms—all catastrophically severe weather events—can wreak havoc on a home’s roof and underlying support structures. However, even winds as low as 50 miles per hour can place a roof in harm’s way (Insurance Institute of Business and Home Safety, 2021a). A recent study by the U.S. Congressional Budget Office estimated annual U.S. insurance losses from wind at $14 billion and combined losses from wind and related flooding at $34 billion (U.S. Congressional Budget Office, 2017). Unfortunately, trends show no apparent letup is in sight, with one United Nations study citing a 40% increase in severe storms over the last 20 years
(Yale Environment 360, 2020).
A recent Bloomberg report states that just two recent major hurricanes alone—Harvey and Irma in 2017—inflicted $134 billion in property damage in Texas and Florida (Golle, 2017). Prices for many building materials were increasing even prior to these storms. However, they moved even higher for key lumber framing and sheathing materials, components accounting for as much as 20% of a home’s total cost (Floor Daily, 2020). Many builders and homeowners have opted to incorporate FORTIFIED construction methods and techniques into their building projects to mitigate risks and reduce the expected costs of future losses.
Because installing a FORTIFIED roof can increase costs, a logical question is whether doing so actually adds value for the homeowner and, if so, to what extent. We estimate the effect of a FORTIFIED Home designation on residential property values using sales transaction data from coastal Alabama from 2011 to 2021. We use a two-state-dependent utility framework to derive the contribution of a FORTIFIED designation on a homeowners’ willingness to pay for reducing the probability of wind damage risk and lowering potential storm damages to levels
fully covered by insurance benefits. Our results show that a FORTIFIED designation is most likely associated with a 2-4% price premium.
Literature
The Insurance Institute of Business and Home Safety (IBHS) established the FORTIFIED Home Program in 2010. It is a voluntary, sustainability-minded new construction and remodeling program designed to strengthen residential properties against specific types of severe weather. The program provides access to uniform, enhanced construction standards, a wide-ranging network of qualified FORTIFIED construction contractors and sub-contractors, and a vigorous, systematic third-party verification and designation process. A primary goal of using FORTIFIED construction methods is to increase resiliency for homes and communities against natural disasters, thereby reducing homeowners’ overall insurance costs and increasing property values (Travelers, 2020). However, because installing a FORTIFIED roof can increase costs, a logical question is whether doing so actually adds value for the homeowner and to what extent.
Over the past decade, an increasing body of literature has focused on the value impacts of newly constructed or remodeled homes that incorporate ‘green’ or ‘sustainable’ design features. These monikers and others, such as ‘eco-friendly,’ ‘renewable,’ and ‘resilient,’ are frequently used as labels in the quantification of home value impacts relative to the inclusion of one or more energy-efficient features (e.g., heating, ventilation, and air-conditioning systems, windows and doors, plumbing and lighting fixtures) or new or recycled building materials in a variety of building processes (Aroul, Hansz, and Yang, 2021; Aroul and Hansz; 2012; Aroul and Rodriguez, 2017; Best et al., 2021; Bloom, Nobe, and Nobe, 2011; Bond and Devine, 2016;
Cerin, Hassel, and Semenova, 2014; Ciochetti and McGowan, 2010; Deng and Wu, 2014; Freybote, Sun, and Yang, 2015; Goodwin, 2011; Hyland, Lyons, and Lyons, 2013; Pride, Little, and Mueller-Stoffels, 2017; Qiu, Wang, and Wang, 2017; Shen et al., 2021; Yoshida and Sugiura, 2015). Similar related studies examine the impact of energy efficiency certification programs on residential sales prices (Brounen and Kok, 2011; Kahn and Kok, 2014; Reichardt et al., 2012; Walls et al., 2017). Results trend to higher price premiums for green homes.
Another vein of the sustainable real estate literature includes studies suggesting higher home prices for structures built or renovated with hazard-resistant features or under the requirements of more stringent building codes. For example, Simmons and Kruse (2000) and Simmons et al. (2002) find that windstorm mitigation features positively affect resale values. Dumm, Sirmans, and Smersh (2011; 2012) show that homes in hurricane-prone locations built using stricter building codes have higher property values. Gatzlaff et al. (2018) indicate that visible hurricane mitigation features increase sales price, but hidden features do not unless revealed by an inspection.3 However, the adoption of stricter building standards does not always increase values. Dehring (2006) reports that land prices in coastal areas decline in the presence of stricter building standards, as costs of compliance outweigh safety benefits, and related work shows that other factors such as insurance premiums, wind zones, or flood zones may influence sales price (Atreya and Czajkowski 2019; Nyce et al., 2015).
A recent working paper (Awondo et al., 2019) examines the relationship between FORTIFIED designation status and home values. Using data on home sales near the coast in
3 A related strand of the literature focuses on consumer behavior associated with hazard mitigation choices, including Burras, Done, Simmons, and Czajkowski (2018); Dumas, and Graham (2002); Carson, McCullough, and Pooser (2013); Dehring and Halek (2013); Kleindorfer and Kunreuther (1999); Jasour et al. (2018); Peacock (2003); and Petrolia et al. (2015).
Baldwin County, Alabama, they find that a FORTIFIED designation increases average sales price by 7% (Awondo et al., 2019). We build on their efforts using a more extensive, broader, and more recent sample and a different empirical approach measuring the effect of a FORTIFIED designation on residential property values using sales transaction data from coastal Alabama from 2011-2021.4
FORTIFIED Home Program
There are three levels of FORTIFIED designations: Roof (previously known as FORTIFIED Bronze), Silver, and Gold, with each level building progressively upon the previous. The FORTIFIED Roof designation focuses on minimizing roof damage and water intrusion. Building processes include using sealed roof decks, ring-shanked nails, wider drip edges, and impact-resistant shingles on homes built in hail-prone areas. The FORTIFIED Silver designation adds stronger gable bracing, anchored attached structures (e.g., porch, carport), pressure-rated garage doors, and more robust window and door protection. FORTIFIED Silver is generally possible in new construction or retrofitting an existing home. The FORTIFIED Gold designation goes even further, requiring a continuous load path, anchored chimneys, and reinforced wall sheathing. In most cases, FORTIFIED Gold is only possible in new construction.
Nationally, 55% of FORTIFIED homes are FORTIFIED Roof, 43% are FORTIFIED Gold, and 2% are FORTIFIED Silver (Insurance Institute of Business and Home Safety, 2021b). Although FORTIFIED homes exist throughout the United States, most are in southeastern coastal communities. As of November 2021, there are 32,933 FORTIFIED homes, with 83% of
4 We use Zillow’s ZTRAX dataset, a nationwide dataset of property transactions available to researchers at no cost. Note, however, that Zillow recently announced plans to discontinue the ZTRAX program, effective September 30, 2023.
them found in the state of Alabama (Insurance Institute of Business and Home Safety, 2020b). Nearly all of Alabama’s FORTIFIED homes are in its two coastal counties, Baldwin and Mobile.
A homeowner initiates a FORTIFIED designation by working with a qualified FORTIFIED Evaluator to decide on the proper designation level. The Evaluator inspects the existing home or reviews plans for new construction, issues a “Current Condition Report” (Smart Home America, 2020), and identifies changes required to achieve the desired designation. Once the project is underway, the Evaluator works with the homeowner, contractor, and sub- contractors to inspect progress and document key designation metrics. Upon completion, the Evaluator sends the results to an IBHS engineer who issues the appropriate FORTIFIED designation (Smart Home America, 2020). The designation is valid for five years and requires re- inspection by a FORTIFIED Evaluator to maintain it. The cost of FORTIFIED components is covered much the same way as are other renovations, either as part of the overall construction cost or roof replacement.
Some states incentivize homeowners to obtain a FORTIFIED designation. Wind damage mitigation is somewhat of a public good. It benefits the homeowner and surrounding homes damaged by debris during a storm. In 2009, Alabama was the first state to pass legislation authorizing FORTIFIED property insurance discounts, expanding it over the next decade. The Alabama Insurance Underwriting Association is the state-sponsored entity charged with wind insurance policy authority. It offers FORTIFIED premium discounts ranging between 20% and 50%. In 2012, Mississippi did likewise, authorizing the Mississippi Windstorm Underwriting Association to issue FORTIFIED premium discounts ranging between 20% and 30%. Other states also offer insurance premium discounts for FORTIFIED designations, ranging between 13% and 48% in South Carolina, between 6% and 19% in North Carolina, and between 5% and
10% in Georgia. States also encourage homeowners to take part in the program through other incentives. For example, Alabama offers a FORTIFIED retrofit tax deduction. It recently offered
$10,000 grants to coastal residents to undertake wind hazard mitigation efforts through the Strengthen Alabama Homes Program (Alabama Department of Insurance, 2020).
There is an extreme lack of publicly available information on the added cost of FORTIFIED construction. Awondo et al. (2019) report that a FORTIFIED Roof designation costs roughly $4.74 per square foot more than a home without the designation. Malik (2017) reports extra cost estimates for FORTIFIED roof components, including $600-$900 for added roof deck nailing and sealing, $600-$1,200 for higher-rated impact-resistant shingles, and $450 or more for FORTIFIED designation evaluation and certification fees. Breitenback (2017) states that meeting “minimum FORTIFIED standards” (which we assume means FORTIFIED Roof) requires an additional $10,000. Farris (2017) reports a “$12,000 retrofit [cost] on [a FORTIFIED] roof.” The Multihazard Mitigation Council (2017) estimates extra FORTIFIED Silver costs ranging between $3,000 and $4,000 for a 2,000 square foot home, depending on wind speed risk and distance from the coast. FORTIFIED Gold designation can add 1% to 3% to overall construction cost (personal communication with Mr. Graham Green, Smart Home America, 2/28/2020). Actuarial consulting firm Merlinos and Associates (2012) states that “FORTIFIED construction methods add 5% to 10% to building costs.” Additionally, the FORTIFIED Evaluator inspection fee, required every five years, can range between $175 and
$750 (Malik, 2017; Rogers, 2019).
Pursuing any level of FORTIFIED designation requires added capital. It is also clear that a homeowner may reduce annual insurance costs. However, whether incurring the added expense
of achieving a FORTIFIED designation adds any realizable monetary value for the homeowner at the point of sale may influence a homeowner’s decision to pursue the designation.
Analytical Model
Hedonic modeling is a long-accepted real estate valuation regression technique that considers property and related neighborhood characteristics.5 Previous FORTIFIED homes research expects a positive price impact of FORTIFIED homes in the Hedonic Price Function (HPF) (e.g., Awondo et al., 2019). However, we follow Smith (1974) and use a two-state- dependent utility model with supply uncertainty. We derive two cases in which the features of FORTIFIED homes would lead to a favorable hedonic price. We expect that FORTIFIED home features likely contribute to homeowners’ willingness to pay for reduced risk and minor overall damage. Our benchmark model is the flood hazard with insurance model with supply uncertainty developed by MacDonald et al. (1987), which we follow using the same mathematical notation.
We suppose a rational buyer will choose where to live by maximizing utility (𝑈) subject to budget constraints (𝑌) if the budget consists of the numeraire goods (x) and the housing price depends on n attributes of house, community, and environment (𝑎𝑖 for i = 1, ⋯ , n). With housing price as a function of attributes r(a), the buyer will maximize utility 𝑈(𝑥, 𝑎) subject to budget constraints (Y = x + r(a)) by choosing the combination of 𝑥 and 𝑎. From the optimal
𝑖
𝑖solution, the hedonic price of the i-th attribute (𝑎 ) is 𝜕𝑟
𝜕𝑎𝑖
= 𝑟𝑎𝑖 , which is the willingness to pay
(WTP) for alternative amounts of attributes holding the indirect utility (V(Y, a)) constant (Rosen,
5 See Affuso et al., 2018; Bao and Wan, 2007; Dahal et al., 2019; Farmer and Lipscomb, 2004; Lipscomb, 2004; Peterson and Flanagan, 2009; Osland, 2010; Sirmans et al., 2005; Shultz, 2018; Shultz and Schmitz, 2009; Wolverton and Senteza, 2000; Wyman and Worzala, 2020.
1974). Assuming the 𝑛-th attribute represents fortified home features, we expect 𝑟𝑎𝑛 ≥ 0 as a normal good.
With tropical storm hazards, the buyer maximizes utility under uncertainty. Smith (1974) first suggested the option price model of supply uncertainty. He used the HPF to analyze how WTP reduces hazard risk in the HPF. Extending Smith (1974), MacDonald et al. (1987) explained how insurance and risk interact in the HPF theoretically and empirically. Following MacDonald et al. (1987, pp. 364-365), we suppose two states: State 1 for a desirable state (e.g., no storm damage) and State 2 for an undesirable state (e.g., storm damage). With the probability of the desirable status (𝑝) and insurance payment depending on the probability (𝐼(𝑝)), we can assume two budgets as:
State 1: 𝑌 = 𝑥 + 𝑟(𝑎, 𝑝) + 𝐼(𝑝), and (1)
State 2: 𝑌 = 𝑥 + 𝑟(𝑎, 𝑝) + 𝐼(𝑝) + 𝐿(𝑎𝑛) − 𝑃(𝑎𝑛).
In Equation (1), L(∙) is the loss amount by storms, P(∙) is the known payment received by the insured, 𝑎 presents a vector of the 𝑛 attributes, and 𝑎𝑛 captures FORTIFIED home features.6 As previously described, the FORTIFIED home features increase the resistance to storm damage, and thus, we define storm losses as a function of 𝑎𝑛, i.e., 𝐿(𝑎𝑛) in Equation (1). In addition, because a benefit of participation in the FORTIFIED Home Program is the resulting insurance discount, the received payment is also 𝑃(𝑎𝑛).7
6 Fortified home features could represent multiple attributes. In this case, we can present L and P as a function of multiple attributes; inclusion of multiple attributes, however, does not change or affect our final model.
7 In MacDonald et al. (1987), both L and P are assumed as known constants. Awondo et al. (2019) also derive a similar extension to ours, but their model includes FORTIFIED home costs and assumes full insurance coverage only. In our model, we conceptualize the ex-post attributes to buyers only and the role of the FORTIFIED home features in L and P.
Let 𝑌1 = 𝑌 − 𝑟(𝑎, 𝑝) and 𝑌2 = 𝑌 − 𝑟(𝑎, 𝑝) − 𝐿(𝑎𝑛) + 𝑃(𝑎𝑛). From Equation (1), the option price (𝑂𝑃) can be defined by:
(𝑝 + 𝜎)𝑉1(𝑌1 − 𝐼(𝑝) − 𝑂𝑃, 𝑎) + (1 − 𝑝 − 𝜎)𝑉2(𝑌2 − 𝐼(𝑝) − 𝑂𝑃, 𝑎) (2)
= 𝑝𝑉1(𝑌1 − 𝐼(𝑝), 𝑎) + (1 − 𝑝)𝑉2(𝑌2 − 𝐼(𝑝), 𝑎), where 𝜎 is the incremental probability of obtaining the desirable state, through the implicit
differentiation, we can find the hedonic price of the probability to the desired state and option price differential as:
𝑝 𝑝
𝑝 𝑝𝑟 = 𝑉1(∙)−𝑉2(∙) − 𝐼 , and (3)
𝑝𝑑𝑉1+(1−𝑝)𝑑𝑉2
𝑑𝑥 𝑑𝑥
𝜎
𝜎𝑂𝑃 = 𝑉1(∙)−𝑉2(∙) . (4)
(𝑝+𝜎)𝑑𝑉1+(1−𝑝−𝜎)𝑑𝑉2
𝑑𝑥 𝑑𝑥
at the beginning increment to 𝑝 (i.e., 𝜎 = 0), we have 𝑟𝑝 = 𝑂𝑃𝜎 − 𝐼𝑝 (from Equations (3) and (4)).
Not surprisingly, the notational expression of Equations (3) and (4) are analogous to
MacDonald et al. (1987) because L and P are still exogenous to 𝑝. As expected, storm losses are higher than received insurance payments. If 𝐿 > 𝑃, then 𝑉1 > 𝑉2 and thus, 𝑂𝑃𝜎 > 0 and 𝑟𝑝 >
−𝐼𝑝. This finding means that an individual would be willing to pay to increase the probability of the desired state (i.e., a component of incremental 𝜎) regardless of the change in insurance cost8, 1987 ). Because the FORTIFIED features contribute to increasing the probability of the desired status, the FORTIFIED home attributes affect the buyer’s utility and the expected hedonic price of 𝑟𝑎𝑛 > 0.
We can also consider a special case when 𝐿 = 𝑃 is more likely for FORTIFIED
designated homes to receive insurance benefits. We reason that if the FORTIFIED features are
8 (MacDonald et al.
sufficiently effective against storm damages, then the loss amount (𝐿(𝑎𝑛)) gets smaller, and
𝐿(𝑎𝑛) is likely to be fully covered by the received payment (𝑃(𝑎𝑛)). In this case, 𝐿 = 𝑃 and thus,
𝑉1 = 𝑉2 and 𝑟𝑝 = −𝐼𝑝, which means the sales price differential will be determined by the change in insurance cost from the change of probability of storm hazard. In addition, the buyer’s best choice is determined mainly by the FORTIFIED home features that reduce storm damages to the amount fully covered by insurance. Again, the buyer is likely to prefer FORTIFIED homes
(𝑟𝑎𝑛 ≥ 0). It is noteworthy that MacDonald et al. (1987) also demonstrate that 𝐿 = 𝑃 is feasible in their model. However, their case is only feasible when the exogenous storm damage (even though it is significant) is fully covered by insurance. Making L and P endogenous to 𝑎𝑛, the model we employ demonstrates that buyers are likely to enjoy the benefits of FORTIFIED features with a smaller loss amount and possible full insurance coverage. In an empirical analysis, this propensity is evident if the price differential of the added costs of the FORTIFIED designation contribute to a property’s sale price (i.e., the hedonic value of the FORTIFIED home is positive and statistically significant (𝑟𝑎𝑛 ≥ 0)).
Data
Eighty-three percent of all FORTIFIED homes are in Alabama. Over 99% of them are situated in its two coastal counties, Baldwin and Mobile, which are the focus of our study (See Figure 1). Mobile County, found on the west side of Mobile Bay, covers 1,644 square miles, has a population of 413,210, and contains the cities of Mobile at the northwest head of the Bay and Dauphin Island off of the Bay’s southwest coast. Baldwin County, found on the east side of Mobile Bay, covers 2,027 square miles, has a population of 223,234, and contains the coastal beach resort cities of Gulf Shores and Orange Beach.
We obtain residential property transactions and characteristics data from the Zillow Transaction and Assessment Dataset (ZTRAX). Zillow describes ZTRAX as “the country’s largest real estate database: It is available free of charge to U.S. academic, nonprofit, and government researchers. Zillow updates ZTRAX quarterly and contains approximately 150 million records of property transactions across the United States” (Zillow 2020). We use observations for Baldwin and Mobile counties in Alabama between 2011 (the year prior to FORTIFIED’s beginning) and Q1 2021 (see Figure 1).
The raw ZTRAX data consist of two distinct datasets: characteristic and transactional.
We match property characteristics to individual property sales to obtain a single dataset, dropping any property found in one dataset but missing in the other. Each sale becomes a unique observation for a specific property having multiple sales. We drop non-residential property observations, non-single-family home observations, not deed transfers, and those missing street addresses. At this point, we have 76,511 ZTRAX observations.
IBHS’s FORTIFIED Homes dataset holds a complete list of all FORTIFIED-designated homes, including address, designation level, date of designation, and date of expiration, between 2011 and 2021. Some homes have multiple observations because a home may be re-designated multiple times. We follow Wasi and Flaaen (2015) to correct address discrepancies (e.g., “123 East A Street” versus “123 A Street East”). We also correct typographical errors in addresses.
Several observations were exact duplicates, which we dropped. Observations with duplicate addresses but different FORTIFIED designations indicate that the home was re-designated. There are two possibilities: overlap re-designation and lapse time re-designation.
Overlap re-designation refers to a home re-designated before the earlier designation expires. In this case, the home never experiences a gap in the designation. We merge these cases
into a single observation, using the earliest designation date and the latest expiration date. The second case is lapse re-designation. Lapse re-designation means there is a period between first designation expiration and re-designation. We decided to keep the most recent designation with lapse observations only due to difficulties matching multiple such observations to single sales observations.
Additionally, some homes had different FORTIFIED designations over time (e.g., a change from FORTIFIED Silver to FORTIFIED Gold). We decided to keep the most recent designation. This process results in a single observation for a given FORTIFIED home. At this point, we have 25,192 FORTIFIED observations.
We then merge ZTRAX and FORTIFIED datasets by street address, city, and zip code. We identify 9,560 ZTRAX observations that correspond to FORTIFIED observations, i.e., we match sales of homes that had FORTIFIED designations at some point. After comparing sales dates to FORTIFIED designation dates, we determined that 4,028 of them held the FORTIFIED designation at the time of sale, and 5,532 received the FORTIFIED designation after the sale. We classify these latter observations as “future FORTIFIED.” In some sense, we can consider these as pre-treatment observations, that is, homes that will eventually receive the FORTIFIED designation.
We then drop observations with missing data, primarily missing year built, number of bedrooms, and number of bathrooms, reducing the sample size from 59,242 to 31,876. We then drop observations with erroneous information and outliers: those with sale dates that precede construction dates, those with greater than five floors, and those with square footage less than
400. Additionally, following Walls et al. (2017), we drop observations with sales prices below the 1st percentile and above the 99th percentile. Our final sample contains 30,286 properties. The
2,908 with a FORTIFIED designation (see Table 2) include 2,608 (90%) FORTIFIED Gold, 32
(1%) FORTIFIED Silver, and 275 (9%) FORTIFIED Roof. There are 2,788 FORTIFIED homes in Baldwin County and 120 in Mobile County. Thirty percent of FORTIFIED designated homes sold as new construction, and nearly all (99%) were FORTIFIED Gold. There are 3,495 sales classified as Future FORTIFIED. Of these, 59% were designated FORTIFIED Roof after the observed sale, 41% were FORTIFIED Gold, and less than 1% were FORTIFIED Silver. The population of FORTIFIED designations is roughly equally divided between FORTIFIED Gold and Roof designations. In contrast, our sample is heavily tilted toward FORTIFIED Gold designations, primarily because those with FORTIFIED Roof designations did not sell during our observed time frame.
Matched Samples and Estimation Method
The ideal framework for our analysis would use the HPF to evaluate sales premiums of homes. The most rigorous way would be to implement a modern causal inference technique by setting the FORTIFIED home features as treatment. Unfortunately, our dataset structure poses some crucially unmanageable challenges. First, the level of FORTIFIED home features is heterogeneous and has no single index definition. Since the data do not hold the cost amounts necessary to achieve FORTIFIED designation, we can only track the three FORTIFIED designation levels. Second, although sales year is known, we have no neighborhood characteristics across years as before-and-after variables are not available for the significant attributes. For these reasons, we adopt the standard HPF regression analysis and incorporate extensive estimation robustness checks. To avoid the dominance of non-FORTIFIED designation effects, we test several specifications with different subsamples using a nearest neighbor
approach. While we do not claim any causal effects of the FORTIFIED homes, we contend that our literature-based, robust hybrid matching and regression testing of FORTIFIED homes is appropriate (Daw and Hatfield, 2018).
In addition to estimating effects using the whole sample, we use nearest-neighbor matching to construct matched samples. A matched sample contains all treatment observations (in our case, FORTIFIED sales) plus one or more non-treatment observations for each treatment observation that most closely resembles it based on observable conditions. We construct three nearest-neighbor samples for each sales price cutoff used in the analysis (no cutoff, $25,000,
$50,000, $75,000, and $100,000). We match using hedonic features of square footage, square footage squared, age, age squared, year built, bedrooms, bedrooms squared, full bathrooms, half bathrooms, pool, floors, garage, and fireplace. We also control for latitude, longitude, flood zone, census tract (dummies), sale year (dummies), and sale quarter (dummies). We rely on Stata’s “teffects nnmatch” routine to find the matched observations (StataCorp, 2020). Our matched sample with no price cutoff has 11,632 observations (three non-treatment sales for each one of the 2,908 treatment and pre-treatment sales).
We estimate a log-linear hedonic price regression of the following form:
3
3ln pijt 1FORTIFIEDijt 2 FutureFORTIFIEDijt β AgeCategoriesijt
SqFeet
SqFeet
2 Bedrooms
Pool
Floors
Garage
4 ijt 5
ijt 6
ijt 7
ijt 8
ijt 9
ijt
10
Fireplaceijt
11
FullBathijt
12
FullBathijt
13 HalfBathijt
14
MilestoCoastijt
2
215Coastalijt 16 FloodAijt 17 FloodVijt 18 RepeatSalesijt + jt eijt
where
pijt is the price of house i in census tract j sold in year-quarter t.
1 is the estimated
coefficient on our crucial policy variable, FORTIFIED. Because nearly all FORTIFIED observations in our sample are FORTIFIED Gold, we do not differentiate between Gold, Silver,
and Roof designations. The effect of Future FORTIFIED homes is captured by 2 . The remaining betas capture the effects of the various house characteristics. We include a fixed effect
jt for each census tract by year and by quarter to control for unobserved neighborhood characteristics that vary over time and an idiosyncratic error term eijt . We use cluster-robust standard errors clustered on census tract x sale year x quarter.
Results
We estimate 25 regressions on various subsamples: the full sample, a subsample excluding sales of homes greater than 39 years old, a subsample excluding new home sales, a subsample excluding pre-2015 sales, and a matched sample. For each of these subsamples, we then estimate regressions for 5 sales price cutoffs: no cutoff, $25,000, $50,000, $75,000, and
$100,000. To simplify the presentation, we report the estimates of just four of these regressions in the main text: (1) the full sample of all home sale observations (N = 30,286); (2) the subsample that excludes sales less than $100,000 (N = 26,063); (3) the matched sample with no price cutoff (N = 11,632); and (4) the matched sample that excludes sales less than $100,000 (N
= 11,212) (see Table 3). We summarize the variation in the FORTIFIED effect across all 25 regressions in Figure 2; the full results of all regressions are reported in the Appendix (see Tables A1-A5).
We find that a FORTIFIED designation significantly and positively impacts sales price, but the magnitude depends on the sample used. We find the most significant estimated effect of a FORTIFIED designation in the all-sales regression: a 7.7% increase in sales price. However, the increase is less when the sample excludes sales below $100,000, at 3.6%. The estimated impact of a FORTIFIED designation declines as the sales price cutoff increases from no cutoff to
$100,000 (Figure 2). Excluding older home sales or sales earlier in the study period have only marginal impacts relative to the all-sales results. The matched samples result in more modest impacts. The estimated impact is a 4.9% increase with no price cutoff. The impact declines as the price cutoff increases, down to a 2.2% increase when sales below $100,000 are excluded.
However, when the sample excludes new home sales, we find a significant impact only when sales below $100,000 are excluded. Here, the estimated impact is a 2.4% increase. Overall, Figure 2 demonstrates that results across the different subsamples are more variable at low price cutoffs but tend to converge as the price cutoff increases.
The coefficient on future-FORTIFIED is significant and positive in the full-sample models but not in the matched sample models. This result indicates that homes destined for a FORTIFIED designation in the future may sell above that of non-FORTIFIED homes, to begin with, suggesting that pre-treatment homes may be predisposed to sell at higher prices. However, with matching, this effect is mitigated.
Most other coefficient estimates are as we expect. Age, square footage, floor, fireplace, half-bathroom, and flood zone effects tend to be significant across the board. The garage coefficient is insignificant, and the repeat sales coefficient is marginally significant in just one model. Unexpectedly, bedroom and pool effects switch signs from positive in the whole sample to negative in the matched sample. Bathroom, miles-to-the-coast, and coastal effects switch from significant in the whole sample to not significant in the matched sample.
Discussion
We find that a FORTIFIED designation is associated with higher sales prices, with the magnitude ranging between 2% and just under 8%. To the extent that our matched sample provides a better estimate, the magnitude is more likely between 2% and 4%. Our models converge to this same range when sales below $100,000 are excluded. Because most of our FORTIFIED observations are Gold designations, our results most likely reflect the impact of this particular designation. In reality, FORTIFIED designations are almost equally split between Gold and Roof designations. However, because most of our Roof-designated homes did not sell during our study period, we cannot estimate the impact of particular designations.
Furthermore, because most of our Gold designations are associated with new home sales, our results most likely reflect the impact of new home sales. This argument is supported because the estimated magnitude of impact is not significantly different from zero in four of the five models that excluded new home sales. When sales below $100,000 are excluded, the impact is significant among non-new-home sales. This finding could mean that builders most likely reap the benefits of a FORTIFIED designation in the form of a higher sales price, with a lesser impact accruing to sellers of existing homes.
Considering why the price premium is not higher than we estimate, our model assumes full and open information of FORTIFIED home attributes and their benefits to buyers. A violation of this assumption could produce a lower estimated magnitude. Considering asymmetric information possibilities, a lack of buyer awareness may provide insight. Although a nationwide program, almost all FORTIFIED homes are located in coastal Alabama. Because much of the area is vacation and retirement destinations, many buyers may not know about the program, especially those moving in from elsewhere. An anecdotal search of real estate listings in the study area reveals that although some homes have a FORTIFIED designation, the listings do not mention it. If buyer incognizance is the reason, this supports the findings of Gatzlaff et al. (2018) that visible hurricane mitigation features increase sales price but that hidden features
generally do not do so. To the extent that FORTIFIED features are not evident to prospective buyers and not touted by sellers or agents, it is reasonable to posit that they will not affect the sales price. If this is the case, IBHS, real estate agents, and local officials may wish to increase efforts to make the benefits of the FORTIFIED Home program more widely known.
Another possibility may be a lack of concern for natural hazards by buyers. Bin and Polasky (2004) find that price discounts associated with being located in a flood zone are more pronounced after a significant storm event than they are before it. Moreover, Michel-Kerjan, Lemoyne de Forges, and Kunreuther (2012) find that in response to the landfall of Hurricanes Katrina and Rita, the number of flood policies in Louisiana and Mississippi increased in 2006 at three to four times the growth rate observed in previous years. Because a major hurricane had not made landfall in Alabama since Hurricane Ivan in 2004 (although Alabama saw some impacts from Katrina in 2005), homeowners may not have viewed storm mitigation as a priority during our study period. It will be interesting to see if results change as more data become available after the very active 2020 hurricane season, including Hurricane Sally, which made landfall in southwest coastal Alabama and caused severe, widespread damage. Although we find that sales prices increased after Sally’s landfall, we do not detect an increased FORTIFIED premium. A recent Business Alabama article quotes IBHS CEO Roy Wright as saying that it is easy to tell which homes have FORTIFIED roofs in the wake of Hurricane Sally: “The ones that don’t are color-coded with blue tarps” (Bloom, 2020).
Conclusions
We wish to emphasize that the benefits associated with FORTIFIED are much broader. In some cases, their value could far exceed the costs of obtaining the designation. With hazard
mitigation measures in place, FORTIFIED homes are less likely to suffer damages as frequently as homes that do not include such measures. As a result, homeowners have lower damage expenses after a severe wind event and avoid less quantifiable costs associated with the stress and logistical challenges they face in the aftermath of a severe wind event (e.g., finding someone to tarp a roof as a temporary fix until a contractor can make permanent repairs or replacement). Homeowners may also avoid costs associated with displacement until home repairs are complete.
Additionally, some states now mandate insurance discounts for homeowners that obtain a FORTIFIED designation. In the absence of statutory requirements, some insurance companies also offer them. This distinction is important because insurance discounts are immediate and ongoing, unlike the benefits that accrue only if and when a storm occurs.
The 2020 Atlantic Hurricane Season was the most active ever recorded (CBS News, 2020), with 31 named storms—so many in fact that the list of names for 2020 generated by the World Meteorological Organization fell short, which required the last ten storm names to derive from the Greek alphabet (Almanac, 2020). Final damage totals are not yet available, but one estimate pegs the staggering total losses above $41 billion (Center for Disaster Philanthropy, 2020). We mentioned earlier that many questions remain about the connections between climate change and extreme weather events. There appears to be a growing body of experts who point to links between rising Atlantic Ocean temperatures and the upward frequency trends of extreme weather events. Although this paper is not about the merits of the climate change debate, trends point to likely increasing occurrences and severity of extreme weather events. We hope that our findings will provide helpful information and insights that policymakers can use as they consider decisions that affect the lives of so many who choose to live in areas more prone to catastrophic storm damages and the insurance companies they regulate that provide vital property coverage.
We believe a critical need exists for further research into the costs and benefits of mitigation associated with natural hazards.
References
Affuso, Ermanno, J. Reid Cummings, and Hubinnh Le. 2018. “Wireless Towers and Home Values: An Alternative Valuation Approach Using a Hedonic Spatial Error Model,” The Journal of Real Estate Finance and Economics 56(4): 653-76.
Alabama Department of Insurance. 2020. https://strengthenalabamahomes.com/, last accessed December 4, 2020.
Almanac. 2020. “Hurricane Names for the 2020 Hurricane Season.” https://www.almanac.com/content/hurricane-names-hurricane- season#:~:text=Thelistsofhurricanenames,thelistsofnames
repeat, last accessed December 9, 2020.
Aroul, Ramya R., Hansz, J. Andrew, and Jinsuk Yang. 2021. “Fix it with Green:” The Valuation Impact of Green Retrofits on Residential Transaction Price. Journal of Housing Research, 30(2), 142-162.
Aroul, Ramya and J. Andrew Hansz. 2012. “The Value of ‘Green:’ Evidence from the First Mandatory Residential Green Building Program.” Journal of Real Estate Research 34(1):
27-50.
Aroul, Ramya R., and Mauricio Rodriguez. 2017. “The increasing value of green for residential real estate.” Journal of Sustainable Real Estate, 9(1), 112-130.
Atreya, Ajita and Jeffrey Czajkowski. 2019. “Graduated Flood Risks and Property Prices in Galveston County.” Real Estate Economics 47(3): 807–44.
Awondo, Sebastain, Harris Hollans, Lawrence Powell, and Chip Wade. (2019). “Estimating Effects of Wind Loss Mitigation on Home Value.” Unpublished manuscript, University of Alabama.
Bao, Helen, and Alan Wan. 2007. “Improved Estimators of Hedonic Housing Price Models.” Journal of Real Estate Research 29(3): 267-302.
Best, Rohan, Paul J. Burke, Rabindra Nepal, and Zac Reynolds. 2021. “Effects of rooftop solar on housing prices in Australia.” Australian Journal of Agricultural & Resource Economics 65(3): 493-511.
Bin, Okmyung and Stephen Polasky. 2004. “Effects of Flood Hazards on Property Values: Evidence before and after Hurricane Floyd.” Land Economics 80(4, Nov.): 490-500.
Bloom, Bryan, Nobe, Mary Ellen, and Michael Nobe. 2011. “Valuing green home designs: A study of ENERGY STAR® homes.” Journal of Sustainable Real Estate, 3(1), 109-126.
Bloom, Nedra. 2020. “Fortified Buildings Weather Sally.” Business Alabama, September 25, https://businessalabama.com/fortified-buildings-weather-sally/, last accessed November 19, 2020.
Bond, Shaun A. and Avis Devine. 2016. “Certification matters: Is green talk cheap talk?” The Journal of Real Estate Finance and Economics, 52(2), 117-140.
Breitenbach, Sarah. 2017. “Using Special Nails to Save Roofs—and Dollars.”
Stateline, The Pew Charitable Trusts, June 21. https://www.pewtrusts.org/en/research- and-analysis/blogs/stateline/2017/06/21/using-special-nails-to-save-roofs-and-dollars, last accessed October 27, 2020.
Brounen, Dirk and Nils Kok. 2011. “On the economics of energy labels in the housing market.”
Journal of Environmental Economics and Management 62: 166-79.
Burrus, Jr., Robert T., Christopher F. Dumas, and J. Edward Graham, Jr. 2002. “Catastrophic Risk, Homeowner Response, and Wealth-Maximizing Wind Damage Mitigation.” Financial Services Review 11: 327-40.
Carson, James M., Kathleen A. McCullough, and David M. Pooser. 2013. “Deciding Whether to Invest in Mitigation Measures: Evidence from Florida.” The Journal of Risk and Insurance 80(2): 309-27.
CBS News. 2020. “The Record-Shattering 2020 Hurricane Season, Explained.” https://www.cbsnews.com/news/atlantic-hurricane-season-2020-record-breaking/, last accessed December 9, 2020.
Cerin, Pontus, Lars G. Hassel, and Natalia Semenova. 2014. “Energy performance and housing prices.” Sustainable Development, 22(6), 404-419.
Center for Disaster Philanthropy. 2020. “2020 Atlantic Hurricane Season.” https://disasterphilanthropy.org/disaster/2020-atlantic-hurricane- season/#:~:text=Attheendofthe,the2020Atlantichurricane
season, last accessed December 9, 2020.
Ciochetti, Brian, and Mark McGowan. 2010. “Energy efficiency improvements: do they pay?” Journal of Sustainable Real Estate, 2(1), 305-333.
Coastal Trim & Accessories. 2020. “Fortified Evaluations.” https://coastaltrimaccessories.com/fortified-evaulations/, last accessed October 27, 2020.
Dahal, Ram P., Daniel R. Petrolia, Robert K. Grala, Jason S. Gordon, Ian A. Munn, and J. Reid Cummings (2019). “A Hedonic Pricing Method to Estimate Value of Waterfront in the Gulf of Mexico,” Urban Forestry and Urban Greening, 41(1), 185-194.
Daw, Jamie R. and Laura A. Hatfield. 2018. “Matching and Regression to the Mean in Difference-in-Differences Analysis.” Health Services Research 53(5): 4138-4156.
Dehring, Carolyn A. 2006. “Building Codes and Land Values in High Hazard Areas.” Land Economics 82(4): 513-28.
Dehring, Carolyn A. and Martin Halek. 2013. “Coastal Building Codes and Hurricane Damage.”
Land Economics 89(4): 597-613.
Deng, Yongheng and Jing Wu. 2014. “Economic returns to residential green building investment: The developers’ perspective.” Regional Science and Urban Economics 47: 35-44.
Done, James M., Kevin M. Simmons, and Jeffrey Czajkowski. 2018. “Relationship between Residential Losses and Hurricane Winds: Role of the Florida Building Code.” Journal of Risk and Uncertainty in Engineering Systems, Part A: Civil Engineering 4(1, March): 04018001.
Dumm, Randy E., G. Stacy Sirmans, Greg T. Smersh. 2011. “The Capitalization of Building Codes in House Prices.” The Journal of Real Estate Finance and Economics 42(1): 30- 50.
Dumm, Randy E., G. Stacy Sirmans, Greg T. Smersh. 2012. “Building Codes, Wind Contours, and House Prices.” Journal of Real Estate Research 34(1): 73-98.
Farmer, Michael and Clifford Lipscomb. 2010. “Using Quantile Regression in Hedonic Analysis to Reveal Submarket Competition.” Journal of Real Estate Research 32(4): 435-60.
Farris, Meg. 2017. “Group says hurricane-proof roof will pay for itself.” Eyewitness News (WWLTV), Published October 9, 2017, Updated October 9, 2017. https://www.wwltv.com/article/news/local/group-says-hurricane-proof-roof-will-pay-for- itself/289-482076172, last accessed October 27, 2020.
Floor Daily. 2020. “Lumber Prices Skyrocketing.” https://www.floordaily.net/flooring- news/lumber-prices-skyrocketing, last accessed December 7, 2020.
Freybote, Julia, Hua Sun, and Xi Yang. 2015. “The impact of LEED neighborhood certification on condo prices.” Real Estate Economics, 43(3), 586-608.
Gatzlaff, Dean, Kathleen McCullough, Lorilee Medders, Charles M. Nyce. 2018. “The Impact of Hurricane Mitigation Features and Inspection Information on House Prices.” The Journal of Real Estate Finance and Economics 57: 566–91.
Golle, Vince. 2017. “Construction Was Pricey Before the Hurricanes. It’ll Get Worse.” Bloomberg. https://www.bloomberg.com/news/articles/2017-09-15/construction-was- pricey-before-the-hurricanes-it-ll-get-worse, last accessed December 4, 2020.
Goodwin, Kim. 2011. “The demand for green housing amenities.” Journal of Sustainable Real Estate, 3(1), 127-141.
Hyland, Marie, Ronan C. Lyons, and Seán Lyons. 2013. “The value of domestic building energy efficiency – evidence from Ireland.” Energy Economics 40: 943-52.
Insurance Institute of Business and Home Safety. 2021a. “Why FORTIFIED?” https://fortifiedhome.org/why/, last accessed November 24, 2021.
Insurance Institute of Business and Home Safety. 2021b. “FORTIFIED Monthly Update.” https://ibhs.org/fortified/fortified-monthly-update/, last accessed November 24, 2021.
Jasour, Zeinab Y., Rachel A. Davidson, Joseph E. Trainor, Jamie L. Kruse, and Linda K. Nozick. 2018. “Homeowner Decisions to Retrofit to Reduce Hurricane-Induced Wind and Flood Damage.” Journal of Infrastructure Systems 24(4): 04018026.
Jumper, Kathy. 2011. “Fortified Homes Can Save Homeowners Money.” June 05, Press- Register, Updated Jan 14, 2019. https://www.al.com/press-register- business/2011/06/fortified_homes_can_save_homeo.html, last accessed October 27, 2020.
Kahn, Matthew E. and Nils Kok. 2014. “The capitalization of green labels in the California housing market.” Regional Science and Urban Economics 47: 25-34.
Kleindorfer, Paul R. and Howard Kunreuther. 1999. “The Complementary Roles of Mitigation and Insurance in Managing Catastrophic Risks.” Risk Analysis 19(4): 727-38.
Lipscomb, Clifford. 2004. “An Alternative Spatial Hedonic Estimation Approach.” Journal of Housing Research 15(2): 143-60.
MacDonald, Don N., James C. Murdoch and Harry L. White. 1987. “Uncertain Hazard, Insurance, and Consumer Choice: Evidence from Housing Markets.” Land Economics 63(4): 361-371.
Malik, Fred. 2017. “Is Building FORTIFIED Worth It?” Presentation, National Tornado Summit, February 14. https://tornadosummit.org/archive/2017/files/Presentations/BuildingFortified_FredMalik. pdf, last accessed October 27, 2020.
Merlinos and Associates – Actuarial Consulting. 2012. “The Impact of FORTIFIED Building Standards.” http://merlinosinc.com/fortified-building-standards, last accessed October 27, 2020.
Michel‐Kerjan, Erwann, Sabine Lemoyne de Forges, and Howard Kunreuther. 2012. “Policy Tenure Under the U.S. National Flood Insurance Program (NFIP).” Risk Analysis 32(4, April): 644-58.
Nyce, Charles, Randy E. Dumm, G. Stacy Sirmans, and Greg T. Smersh. 2015. “The Capitalization of Insurance Premiums in House Prices.” The Journal of Risk and Insurance 82(4): 891–919.
Osland, Liv. 2010. “An Application of Spatial Econometrics in Relation to Hedonic House Price Modeling.” Journal of Real Estate Research 32(3): 289-320.
Peacock, Walter Gillis. 2003. “Hurricane Mitigation Status and Factors Influencing Mitigation Status among ’Florida’s Single-Family Homeowners.” Natural Hazards Review 4(3): 149-58.
Peterson, Steven, and Albert Flanagan. 2009. “Neural Network Hedonic Pricing Models in Mass Real Estate Appraisal.” Journal of Real Estate Research 31(2): 147-64.
Petrolia, Daniel .R., Joonghyun Hwang, Craig E. Landry, and Keith H. Coble. 2015. “Wind Insurance and Mitigation in the Coastal Zone.” Land Economics 91(2): 272-95.
Pride, Dominique, Joseph Little, and Marc Mueller-Stoffels. 2017. “The value of energy efficiency in the anchorage residential property market.” Journal of Sustainable Real Estate, 9(1), 172-194.
Pride, Dominique, Joseph Little, and Marc Mueller-Stoffels. 2018. “The value of residential energy efficiency in interior Alaska: A hedonic pricing analysis.” Energy Policy 123: 450-60.
Qiu, Yueming, Yi David Wang, and Jianfeng Wang. 2017. “Soak up the sun: Impact of solar energy systems on residential home values in Arizona.” Energy Economics 66: 328-36.
Reichardt, Alexander, Franz Fuerst, Nico Rottke, and Joachim Zietz. 2012. “Sustainable Building Certification and the Rent Premium: A Panel Data Approach.” Journal of Real Estate Research 34(1): 99-126.
Rogers, Spencer. 2019. “How to Make Your Roof More Leak Resistant: Wind Insurance Incentives.” North Carolina Beach Inlet and Waterway Association. http://www.ncbiwa.org/files/WindIncentives-NCBIWA-2019.pdf, last accessed October 27, 2020.
Rosen, Sherwin. 1974. “Hedonic Prices and Implicit Markets: Product Differentiation in Pure Competition.” Journal of Political Economy 82(1): 34-55.
Shen, Xingchi, Pengfei Liu, Yueming (Lucy) Qiu, Anand Patwardhan, and Parth Vaishnav.
2021. “Estimation of change in house sales prices in the United States after heat pump adoption.” Nature Energy 6: 30-7.
Shultz, Steven. 2018. “Housing Depreciation Revisited: Hedonic Price Modeling Versus Assessor Estimates.” Journal of Housing Research 27(1): 45-58.
Shultz, Steven., and Nicholas Schmitz. 2009. “Augmenting Housing Sales Data to Improve Hedonic Estimates of Golf Course Frontage.” Journal of Real Estate Research 31(1): 63- 80.
Simmons, Kevin M., and Jamie B. Kruse. 2000. “Market Value of Mitigation and Perceived Risk Empirical Results.” Southern Economic Journal 26(1): 41-51.
Simmons, Kevin M, Jamie B. Kruse, Douglas A. Smith. 2002. “Valuing Mitigation: Real Estate Market Response to Hurricane Loss Reduction Measures.” Southern Economic Journal 68(3): 660-71.
Sirmans, G. Stacy. David A. Macpherson, and Emily N. Zietz. 2005. “The composition of hedonic pricing models.” Journal of Real Estate Literature 13(1): 1-44.
Smart Home America. 2020. “How do I get a FORTIFIED Designation for My Home?” https://www.smarthomeamerica.org/news/how-do-i-get-a-fortified-certification-for-my- home, last accessed November 20, 2020.
Smith, V. Kerry. 1985. “Supply Uncertainty, Option Price, and Indirect Benefit Estimation.”
Land Economics 61(3): 303-307.
Travelers. 2020. “Renovating Your Home to FORTIFIED Standards.” https://www.travelers.com/resources/home/renovation/renovating-your-home-to- fortified- standards#:~:text=Buildingcodessetaminimum,resilienttonatur aldisasterstomorrow, last accessed December 4, 2020.
U.S. Congressional Budget Office. 2020. “Expected Costs from Hurricane Winds and Storm- Related Flooding.” https://www.cbo.gov/system/files/2019-04/55019- ExpectedCostsFromWindStorm.pdf, last accessed December 4, 2020.
Walls, Margaret, Todd Gerarden, Karen Palmer, and Xian Fang Bak. 2017. “Is Energy Efficiency Capitalized into Home Prices? Evidence from Three U.S. Cities.” Journal of Environmental Economics and Management 82: 104–24.
Wasi, Nada and Aaron Flaaen. 2015. “Record linkage using Stata: Preprocessing, linking, and reviewing utilities.” The Stata Journal 15(3): 672–97.
Wolverton, Marvin and Jimmy Senteza. 2000. “Hedonic Estimates of Regional Constant Quality House Prices.” Journal of Real Estate Research 19(3): 235-53.
Wyman, David and Elaine Worzala. 2016. “’Dockin’USA—A Spatial Hedonic Valuation of Waterfront Property.” Journal of Housing Research 25(1): 65-80.
Yale Environment 360. 2020. “Extreme Weather Events Have Increased Significantly in the Last 20 Years.” https://e360.yale.edu/digest/extreme-weather-events-have-increased-
significantly-in-the-last-20-years, last accessed December 4, 2020.
Yoshida, Jiro, and Ayako Sugiura. 2015. “The effects of multiple green factors on condominium prices.” The Journal of Real Estate Finance and Economics, 50(3), 412-437.
Zillow. 2020. “’Zillow’s Assessor and Real Estate Database (ZTRAX).” https://www.zillow.com/research/ztrax/, last accessed November 20, 2020.
Table 1. Variable Definitions
Variable Description
Sales Price Sales price of home
FORTIFIED =1 if FORTIFIED designated at time of sale, 0 otherwise Future FORTIFIED =1 if received FORTIFIED designation at some point after
observed sale, 0 otherwise
Age (multiple categories) Year built minus sale year Square Footage (1,000)
Bedrooms Number of bedrooms
Full Bathrooms Number of rooms containing sink, shower, bathtub, and toilet Half Bathrooms Number of rooms containing sink and toilet
Pool =1 if property has a pool, =0 otherwise
Floors Number of floors
Garage =1 if reported garage description is “garage”, = 0 if “carport”, “mixed garage type”, or “NA”
Fireplace =1 if home contains a fireplace, =0 otherwise
Miles to Coast Distance in miles from home to nearest coastline using convex
hull method
Coastal = 1 if home located < 0.5 mile from nearest coastline
Flood Zone A =1 if home is located in FEMA SFHA Zone A, =0 otherwise Flood Zone V =1 if home is located in FEMA SFHA Zone V, =0 otherwise Repeat Sales Count of times home sold during 2011-2021
Fixed Effects Census Tract Sale Year Sale Quarter
Table 2. Summary Statistics of FORTIFIED and non-FORTIFIED Homes
FORTIFIED non-FORTIFIED
non-FORTIFIED
matched
Mean
Std.
Dev.
Mean
Std. Dev.
Mean
Std. Dev.
Sales Price
265,463
100,983
192,959
106,945
234,129
99,390
Future FORTIFIED
0.13
0.33
0.35
0.48
Age
2.95
6.35
15.06
12.18
6.46
7.57
Square Footage (1,000)
2.70
0.75
2.43
1.05
2.54
0.72
Bedrooms
3.66
0.66
3.30
0.62
3.50
0.63
Full Bathrooms
2.35
0.60
1.79
0.75
2.25
0.55
Half Bathrooms
0.21
0.41
0.19
0.41
0.15
0.36
Pool
0.04
0.19
0.08
0.27
0.02
0.15
Floors
1.12
0.30
1.11
0.30
1.08
0.26
Garage
0.01
0.09
0.07
0.25
0.01
0.08
Fireplace
0.55
0.50
0.57
0.50
0.53
0.50
Miles to Coast
19.02
10.20
23.83
12.06
18.93
10.05
Coastal
0.004
0.07
0.01
0.10
0.01
0.08
Flood Zone A
0.03
0.16
0.05
0.23
0.03
0.16
Flood Zone V
0.0003
0.02
0.004
0.06
0.001
0.03
Repeat Sales
1.51
0.61
1.35
0.55
1.43
0.56
Observations
2,908
27,378
8,724
Table 3. Main regression results. All models estimated using log-linear fixed-effects regressions, with census tract x
sale year x sale quarter fixed effects. Significance levels are indicated as *** (p-value < 0.01), ** (< 0.05), * (< 0.10).
No Price Cutoff
≥ $100,000
No Price Cutoff
≥ $100,000
coef se
coef
se coef
se
coef se
No Price Cutoff
≥ $100,000
No Price Cutoff
≥ $100,000
coef se
coef
se coef
se
coef se
Full Sample Matched Sample
FORTIFIED At Sale
0.077
***
0.012
0.036
***
0.007
0.049
***
0.014
0.022
***
0.006
Future FORTIFIED
0.024
**
0.011
0.016
***
0.005
0.008
0.023
0.015
0.011
Age 1
0.029
**
0.013
-0.004
0.007
-0.011
0.019
-0.017
0.012
Age 2
-0.016
0.023
-0.032
**
0.013
-0.041
0.037
-0.028
0.017
Age 3
-0.043
**
0.020
-0.042
***
0.012
-0.068
*
0.035
-0.042
***
0.015
Age 4
-0.003
0.020
-0.036
***
0.011
-0.087
*
0.049
-0.056
***
0.015
Age 5
-0.011
0.021
-0.044
***
0.012
-0.104
0.066
-0.060
***
0.015
Age 6-9
-0.042
***
0.016
-0.067
***
0.010
-0.128
***
0.031
-0.077
***
0.014
Age 10-14
-0.095
***
0.015
-0.108
***
0.009
-0.175
***
0.03
-0.140
***
0.014
Age 15-19
-0.171
***
0.016
-0.160
***
0.010
-0.207
***
0.033
-0.168
***
0.017
Age 20-24
-0.247
***
0.016
-0.212
***
0.011
-0.250
***
0.041
-0.204
***
0.020
Age 25-29
-0.334
***
0.019
-0.263
***
0.012
-0.340
***
0.043
-0.276
***
0.024
Age 30-39
-0.379
***
0.024
-0.283
***
0.017
-0.371
***
0.068
-0.282
***
0.041
Age 40-49
-0.417
***
0.026
-0.315
***
0.02
-0.478
***
0.111
-0.423
***
0.125
Age 50-74
-0.399
***
0.033
-0.200
***
0.031
-0.242
*
0.127
-0.162
0.113
Age 75+
-0.378
***
0.072
-0.086
0.065
-0.410
***
0.137
-0.302
**
0.120
Square Footage (1,000)
0.349
***
0.015
0.321
***
0.012
0.486
***
0.054
0.532
***
0.033
Square Footage²
-0.024
***
0.002
-0.022
***
0.002
-0.041
***
0.007
-0.048
***
0.005
Bedrooms
0.014
**
0.006
0.004
0.004
0.002
0.014
-0.024
***
0.007
Pool
0.025
*
0.013
0.018
**
0.009
-0.127
***
0.04
-0.091
***
0.023
Floors
0.058
***
0.012
0.066
***
0.008
0.023
0.037
0.061
***
0.022
Garage
0.004
0.011
0.003
0.008
-0.084
0.059
-0.002
0.027
Fireplace
0.079
***
0.007
0.064
***
0.004
0.068
***
0.013
0.067
***
0.008
Full Bathrooms
0.314
***
0.028
0.218
***
0.023
0.181
*
0.107
0.123
0.078
Full Bathrooms²
-0.044
***
0.006
-0.021
***
0.005
-0.024
0.021
-0.005
0.015
Half Bathrooms
0.098
***
0.008
0.094
***
0.005
0.077
***
0.019
0.052
***
0.010
Miles to Coast
-0.005
**
0.002
-0.006
***
0.001
0.004
0.004
0.000
0.002
Coastal
0.264
***
0.033
0.163
***
0.026
0.042
0.089
0.122
*
0.066
Flood Zone A
0.098
***
0.022
0.106
***
0.014
0.061
0.088
0.072
***
0.022
Flood Zone V
0.170
***
0.066
0.181
***
0.048
0.578
***
0.192
0.168
**
0.079
Repeat Sales
-0.008
0.005
0.005
*
0.003
-0.011
0.017
0.010
0.008
Observations
30,286
26,063
11,632
11,212
Figure 1. Map of Alabama’s Baldwin and Mobile Counties, showing location of FORTIFIED homes at time of sale (blue) and non-FORTIFIED homes (red). White lines are census track boundaries.
Figure 2. Plot of estimated percent change in sales price due to FORTIFIED designation by regression model. ns = not significant.
Table A1. Regression results using all sales, by price cutoff. All models estimated using log-linear fixed-effects regressions, with census tract x
sale year x sale quarter fixed effects. Significance levels are indicated as *** (p-value < 0.01), ** (< 0.05), * (< 0.10).
No Price Cutoff
≥ $25,000
≥ $50,000
≥ $75,000
≥ $100,000
coef
se
coef
se
coef
se
coef
se
coef
se
FORTIFIED
0.077
***
0.012
0.065
***
0.011
0.041
***
0.009
0.031
***
0.007
0.036
***
0.007
Future FORTIFIED
0.024
**
0.011
0.018
*
0.01
0.004
0.008
0.015
**
0.006
0.016
***
0.005
Age 1
0.029
**
0.013
0.027
**
0.011
0.003
0.009
-0.006
0.007
-0.004
0.007
Age 2
-0.016
0.023
-0.017
0.02
-0.022
0.016
-0.031
**
0.014
-0.032
**
0.013
Age 3
-0.043
**
0.02
-0.046
**
0.019
-0.063
***
0.016
-0.048
***
0.012
-0.042
***
0.012
Age 4
-0.003
0.02
-0.009
0.019
-0.037
***
0.014
-0.052
***
0.013
-0.036
***
0.011
Age 5
-0.011
0.021
-0.009
0.019
-0.044
***
0.016
-0.047
***
0.014
-0.044
***
0.012
Age 6-9
-0.042
***
0.016
-0.035
**
0.014
-0.06
***
0.012
-0.072
***
0.011
-0.067
***
0.01
Age 10-14
-0.095
***
0.015
-0.097
***
0.014
-0.114
***
0.011
-0.121
***
0.009
-0.108
***
0.009
Age 15-19
-0.171
***
0.016
-0.177
***
0.015
-0.183
***
0.012
-0.179
***
0.01
-0.16
***
0.01
Age 20-24
-0.247
***
0.016
-0.252
***
0.015
-0.258
***
0.012
-0.244
***
0.011
-0.212
***
0.011
Age 25-29
-0.334
***
0.019
-0.333
***
0.017
-0.326
***
0.014
-0.305
***
0.013
-0.263
***
0.012
Age 30-39
-0.379
***
0.024
-0.372
***
0.022
-0.357
***
0.018
-0.326
***
0.018
-0.283
***
0.017
Age 40-49
-0.417
***
0.026
-0.416
***
0.024
-0.397
***
0.021
-0.356
***
0.02
-0.315
***
0.02
Age 50-74
-0.399
***
0.033
-0.392
***
0.03
-0.352
***
0.029
-0.301
***
0.029
-0.2
***
0.031
Age 75+
-0.378
***
0.072
-0.388
***
0.07
-0.28
***
0.064
-0.183
***
0.06
-0.086
0.065
Square Footage (1,000)
0.349
***
0.015
0.353
***
0.015
0.349
***
0.014
0.335
***
0.012
0.321
***
0.012
Square Footage²
-0.024
***
0.002
-0.025
***
0.002
-0.025
***
0.002
-0.023
***
0.002
-0.022
***
0.002
Bedrooms
0.014
**
0.006
0.014
**
0.006
0.006
0.005
0.007
*
0.004
0.004
0.004
Pool
0.025
*
0.013
0.034
***
0.012
0.027
**
0.011
0.026
***
0.01
0.018
**
0.009
Floors
0.058
***
0.012
0.06
***
0.011
0.058
***
0.01
0.065
***
0.008
0.066
***
0.008
Garage
0.004
0.011
0.002
0.01
0.002
0.009
0.002
0.008
0.003
0.008
Fireplace
0.079
***
0.007
0.078
***
0.006
0.075
***
0.005
0.07
***
0.004
0.064
***
0.004
Full Bathrooms
0.314
***
0.028
0.293
***
0.027
0.281
***
0.024
0.232
***
0.023
0.218
***
0.023
Full Bathrooms²
-0.044
***
0.006
-0.041
***
0.006
-0.037
***
0.005
-0.026
***
0.005
-0.021
***
0.005
Half Bathrooms
0.098
***
0.008
0.096
***
0.007
0.103
***
0.006
0.096
***
0.006
0.094
***
0.005
Miles to Coast
-0.005
**
0.002
-0.004
**
0.002
-0.003
**
0.002
-0.004
***
0.001
-0.006
***
0.001
Coastal
0.264
***
0.033
0.236
***
0.029
0.21
***
0.029
0.173
***
0.028
0.163
***
0.026
Flood Zone A
0.098
***
0.022
0.106
***
0.018
0.095
***
0.015
0.107
***
0.014
0.106
***
0.014
Flood Zone V
0.17
***
0.066
0.198
***
0.056
0.178
***
0.049
0.155
***
0.049
0.181
***
0.048
Repeat Sales
-0.008
0.005
-0.009
*
0.005
-0.001
0.004
0.003
0.003
0.005
*
0.003
Observations
30,286
29,978
29,095
27,904
26,063
Table A2. Regression results excluding sales of homes > 39 years old, by price cutoff. All models estimated using log-linear fixed-effects regressions, with census tract x sale year x sale quarter fixed effects. Significance levels are indicated as *** (p-value < 0.01), ** (< 0.05), * (<
0.10).
Sales Included No Price Cutoff ≥ $25,000 ≥ $50,000 ≥ $75,000 ≥ $100,000
coef
se
coef
se
coef
se coef
se
coef
se
FORTIFIED
0.081 ***
0.012
0.069
*** 0.011
0.042
***
0.009 0.031
***
0.007
0.035 ***
0.006
Future FORTIFIED
0.024 **
0.011
0.018
* 0.01
0.005
0.008 0.016
***
0.006
0.017 ***
0.005
Age 1
0.029 **
0.013
0.027
** 0.011
0.003
0.009 -0.007
0.007
-0.004
0.006
Age 2
-0.012
0.022
-0.014
0.02
-0.019
0.016 -0.03
**
0.013
-0.031 **
0.013
Age 3
-0.04 **
0.019
-0.044
** 0.019
-0.061
***
0.015 -0.047
***
0.012
-0.041 ***
0.011
Age 4
-0.002
0.02
-0.008
0.019
-0.036
***
0.014 -0.053
***
0.012
-0.037 ***
0.011
Age 5
-0.009
0.02
-0.007
0.018
-0.043
***
0.016 -0.046
***
0.013
-0.044 ***
0.012
Age 6-9
-0.041 ***
0.016
-0.034
** 0.014
-0.06
***
0.012 -0.073
***
0.01
-0.068 ***
0.009
Age 10-14
-0.093 ***
0.015
-0.096
*** 0.013
-0.114
***
0.01 -0.122
***
0.009
-0.111 ***
0.008
Age 15-19
-0.169 ***
0.016
-0.175
*** 0.014
-0.182
***
0.011 -0.18
***
0.01
-0.163 ***
0.009
Age 20-24
-0.245 ***
0.016
-0.25
*** 0.015
-0.258
***
0.012 -0.245
***
0.01
-0.214 ***
0.01
Age 25-29
-0.331 ***
0.019
-0.33
*** 0.017
-0.324
***
0.013 -0.304
***
0.012
-0.264 ***
0.011
Age 30-39
-0.375 ***
0.024
-0.368
*** 0.022
-0.353
***
0.018 -0.323
***
0.018
-0.283 ***
0.017
Square Footage (1,000)
0.355 ***
0.016
0.359
*** 0.016
0.355
***
0.015 0.344
***
0.013
0.329 ***
0.012
Square Footage²
-0.025 ***
0.002
-0.026
*** 0.002
-0.025
***
0.002 -0.024
***
0.002
-0.022 ***
0.002
Bedrooms
0.015 **
0.007
0.015
** 0.006
0.008
0.005 0.008
*
0.004
0.004
0.004
Pool
0.023 *
0.013
0.034
*** 0.012
0.026
**
0.011 0.024
**
0.01
0.015
0.009
Floors
0.059 ***
0.012
0.061
*** 0.011
0.058
***
0.01 0.064
***
0.009
0.068 ***
0.008
Garage
0.012
0.012
0.008
0.011
0.009
0.009 0.009
0.008
0.009
0.008
Fireplace
0.079 ***
0.007
0.078
*** 0.006
0.076
***
0.005 0.072
***
0.004
0.065 ***
0.004
Full Bathrooms
0.334 ***
0.03
0.311
*** 0.028
0.309
***
0.025 0.263
***
0.023
0.248 ***
0.023
Full Bathrooms²
-0.047 ***
0.006
-0.044
*** 0.006
-0.042
***
0.005 -0.032
***
0.005
-0.026 ***
0.005
Half Bathrooms
0.098 ***
0.008
0.096
*** 0.008
0.103
***
0.006 0.097
***
0.006
0.094 ***
0.005
Miles to Coast
-0.004 *
0.002
-0.003
* 0.002
-0.003
*
0.002 -0.004
***
0.001
-0.005 ***
0.001
Coastal
0.253 ***
0.035
0.225
*** 0.031
0.204
***
0.031 0.168
***
0.03
0.157 ***
0.026
Flood Zone A
0.097 ***
0.022
0.109
*** 0.018
0.097
***
0.015 0.106
***
0.014
0.104 ***
0.014
Flood Zone V
0.17
**
0.067
0.2
***
0.056
0.18
***
0.05
0.156
***
0.05
0.185 *** 0.049
Repeat Sales
-0.007
0.005
-0.008
*
0.005
-0.001
0.004
0.003
0.003
0.005 *
0.003
Observations
29,093
28,810
28,031
26,989
25,302
Table A3. Regression results excluding new home sales, by price cutoff. All models estimated using log-linear fixed-effects regressions, with
census tract x sale year x sale quarter fixed effects. Significance levels are indicated as *** (p-value < 0.01), ** (< 0.05), * (< 0.10).
Sales Included No Price Cutoff ≥ $25,000 ≥ $50,000 ≥ $75,000 ≥ $100,000
coef
se
coef
se
coef
se coef
se
coef
se
FORTIFIED
0.019
0.012
0.014
0.011
0.01
0.01 0.012
0.009
0.024 ***
0.008
Future FORTIFIED
0.029 ***
0.011
0.028
*** 0.009
0.016
**
0.008 0.018
***
0.007
0.014 **
0.005
Age 1
Age 2
-0.05 **
0.02
-0.047
*** 0.018
-0.026
*
0.014 -0.026
**
0.012
-0.029 **
0.012
Age 3
-0.079 ***
0.017
-0.078
*** 0.017
-0.067
***
0.014 -0.043
***
0.012
-0.039 ***
0.011
Age 4
-0.048 **
0.019
-0.048
*** 0.016
-0.046
***
0.013 -0.05
***
0.012
-0.035 ***
0.01
Age 5
-0.064 ***
0.02
-0.056
*** 0.017
-0.057
***
0.015 -0.046
***
0.014
-0.044 ***
0.013
Age 6-9
-0.1 ***
0.014
-0.087
*** 0.013
-0.075
***
0.011 -0.072
***
0.01
-0.066 ***
0.009
Age 10-14
-0.154 ***
0.013
-0.151
*** 0.011
-0.131
***
0.01 -0.122
***
0.009
-0.108 ***
0.008
Age 15-19
-0.231 ***
0.014
-0.232
*** 0.012
-0.2
***
0.01 -0.181
***
0.01
-0.161 ***
0.009
Age 20-24
-0.307 ***
0.015
-0.306
*** 0.014
-0.275
***
0.012 -0.246
***
0.01
-0.212 ***
0.01
Age 25-29
-0.393 ***
0.018
-0.387
*** 0.015
-0.343
***
0.013 -0.308
***
0.012
-0.264 ***
0.011
Age 30-39
-0.44 ***
0.022
-0.427
*** 0.019
-0.374
***
0.016 -0.332
***
0.015
-0.287 ***
0.015
Age 40-49
-0.469 ***
0.025
-0.463
*** 0.023
-0.409
***
0.02 -0.357
***
0.018
-0.314 ***
0.019
Age 50-74
-0.458 ***
0.032
-0.443
*** 0.028
-0.367
***
0.026 -0.306
***
0.026
-0.206 ***
0.028
Age 75+
-0.43 ***
0.071
-0.434
*** 0.068
-0.295
***
0.061 -0.188
***
0.057
-0.092
0.061
Square Footage (1,000)
0.349 ***
0.015
0.357
*** 0.015
0.35
***
0.014 0.332
***
0.012
0.315 ***
0.012
Square Footage²
-0.024 ***
0.002
-0.025
*** 0.002
-0.024
***
0.002 -0.022
***
0.002
-0.021 ***
0.002
Bedrooms
0.009
0.007
0.009
0.006
0.007
0.005 0.009
**
0.004
0.006
0.004
Pool
0.045 ***
0.012
0.053
*** 0.011
0.043
***
0.01 0.038
***
0.009
0.028 ***
0.008
Floors
0.062 ***
0.013
0.062
*** 0.011
0.058
***
0.01 0.058
***
0.009
0.059 ***
0.009
Garage
0.01
0.011
0.006
0.01
0.006
0.009 0.004
0.008
0.003
0.008
Fireplace
0.08 ***
0.007
0.083
*** 0.006
0.076
***
0.005 0.067
***
0.005
0.059 ***
0.004
Full Bathrooms
0.304 ***
0.028
0.281
*** 0.027
0.27
***
0.024 0.235
***
0.023
0.224 ***
0.023
Full Bathrooms²
-0.04 ***
0.006
-0.036
*** 0.006
-0.034
***
0.005 -0.026
***
0.005
-0.021 ***
0.005
Half Bathrooms
0.111 ***
0.008
0.107
*** 0.007
0.114
***
0.006 0.107
***
0.006
0.104 ***
0.005
Miles to Coast
-0.009
***
0.002
-0.007
***
0.002
-0.006
***
0.002
-0.006
***
0.001
-0.008
***
0.001
Coastal
0.266
***
0.041
0.233
***
0.034
0.212
***
0.033
0.177
***
0.03
0.171
***
0.027
Flood Zone A
0.108
***
0.02
0.109
***
0.018
0.102
***
0.015
0.115
***
0.014
0.114
***
0.014
Flood Zone V
0.182
***
0.06
0.176
***
0.056
0.166
***
0.05
0.155
***
0.051
0.188
***
0.047
Repeat Sales
-0.001
0.005
-0.003
0.005
0.002
0.004
0.004
0.003
0.005
*
0.003
Observations
25,992
25,729
24,945
23,836
22,036
Table A4. Regression results excluding pre-2015 sales, by price cutoff. All models estimated using log-linear fixed-effects regressions, with
census tract x sale year x sale quarter fixed effects. Significance levels are indicated as *** (p-value < 0.01), ** (< 0.05), * (< 0.10).
Sales Included No Price Cutoff ≥ $25,000 ≥ $50,000 ≥ $75,000 ≥ $100,000
coef
se
coef
se
coef
se coef
se
coef
se
FORTIFIED
0.082 ***
0.013
0.07
*** 0.011
0.048
***
0.009 0.033
***
0.007
0.037 ***
0.006
Future FORTIFIED
0.02
0.013
0.012
0.012
0.000
0.01 0.012
0.008
0.013 **
0.006
Age 1
0.005
0.015
0.006
0.012
-0.007
0.011 -0.016
*
0.009
-0.013 *
0.008
Age 2
-0.045 *
0.025
-0.043
* 0.023
-0.038
*
0.019 -0.042
***
0.016
-0.042 ***
0.015
Age 3
-0.066 ***
0.022
-0.063
*** 0.021
-0.08
***
0.018 -0.052
***
0.014
-0.048 ***
0.013
Age 4
-0.051 **
0.025
-0.052
** 0.023
-0.066
***
0.018 -0.076
***
0.016
-0.053 ***
0.013
Age 5
-0.036
0.026
-0.031
0.023
-0.055
**
0.022 -0.057
***
0.018
-0.049 ***
0.016
Age 6-9
-0.057 ***
0.02
-0.05
*** 0.017
-0.063
***
0.013 -0.076
***
0.012
-0.069 ***
0.011
Age 10-14
-0.107 ***
0.017
-0.105
*** 0.015
-0.116
***
0.012 -0.125
***
0.01
-0.114 ***
0.009
Age 15-19
-0.171 ***
0.018
-0.173
*** 0.016
-0.171
***
0.012 -0.172
***
0.011
-0.156 ***
0.01
Age 20-24
-0.245 ***
0.018
-0.244
*** 0.017
-0.243
***
0.014 -0.236
***
0.012
-0.21 ***
0.011
Age 25-29
-0.328 ***
0.02
-0.329
*** 0.019
-0.314
***
0.014 -0.301
***
0.013
-0.261 ***
0.012
Age 30-39
-0.406 ***
0.025
-0.389
*** 0.023
-0.365
***
0.018 -0.338
***
0.017
-0.29 ***
0.017
Age 40-49
-0.41 ***
0.029
-0.409
*** 0.027
-0.388
***
0.024 -0.351
***
0.021
-0.318 ***
0.022
Age 50-74
-0.4 ***
0.036
-0.388
*** 0.031
-0.357
***
0.029 -0.319
***
0.029
-0.217 ***
0.031
Age 75+
-0.412 ***
0.072
-0.418
*** 0.068
-0.307
***
0.061 -0.226
***
0.057
-0.13 **
0.063
Square Footage (1,000)
0.331 ***
0.017
0.339
*** 0.018
0.333
***
0.016 0.326
***
0.015
0.32 ***
0.014
Square Footage²
-0.023 ***
0.002
-0.024
*** 0.003
-0.024
***
0.002 -0.022
***
0.002
-0.022 ***
0.002
Bedrooms
0.012
0.008
0.013
* 0.007
0.005
0.006 0.007
0.005
0.006
0.004
Pool
0.038 **
0.015
0.052
*** 0.014
0.039
***
0.013 0.033
***
0.012
0.026 **
0.011
Floors
0.065 ***
0.015
0.063
*** 0.013
0.061
***
0.012 0.066
***
0.01
0.067 ***
0.01
Garage
-0.005
0.013
-0.003
0.013
0.006
0.011 0.007
0.009
0.007
0.009
Fireplace
0.076 ***
0.008
0.076
*** 0.007
0.074
***
0.006 0.074
***
0.005
0.066 ***
0.005
Full Bathrooms
0.329 ***
0.034
0.313
*** 0.032
0.303
***
0.028 0.242
***
0.028
0.228 ***
0.028
Full Bathrooms²
-0.049 ***
0.007
-0.047
*** 0.007
-0.044
***
0.006 -0.03
***
0.006
-0.025 ***
0.006
Half Bathrooms
0.097
***
0.01
0.094
***
0.009
0.098
***
0.007
0.091
***
0.007
0.087
***
0.006
Miles to Coast
-0.005
**
0.002
-0.005
**
0.002
-0.004
*
0.002
-0.005
***
0.002
-0.006
***
0.001
Coastal
0.275
***
0.042
0.248
***
0.037
0.213
***
0.038
0.176
***
0.037
0.17
***
0.034
Flood Zone A
0.083
***
0.026
0.091
***
0.022
0.092
***
0.018
0.107
***
0.016
0.105
***
0.016
Flood Zone V
0.093
0.076
0.138
**
0.063
0.138
***
0.051
0.101
*
0.052
0.143
***
0.047
Repeat Sales
-0.001
0.006
-0.001
0.006
0.005
0.005
0.009
**
0.004
0.01
***
0.003
Observations
22,086
21,864
21,242
20,490
19,416
Table A5. Regression results using matched sample (3 nearest neighbors), by price cutoff. All models estimated using log-linear fixed-effects regressions, with census tract x sale year x sale quarter fixed effects. Significance levels are indicated as *** (p-value < 0.01), ** (< 0.05), * (<
0.10).
Sales Included No Price Cutoff ≥ $25,000 ≥ $50,000 ≥ $75,000 ≥ $100,000
coef
se
coef
se
coef
se coef
se
coef
se
FORTIFIED
0.049 ***
0.014
0.046
*** 0.013
0.032
***
0.011 0.020
***
0.007
0.022 ***
0.006
Future FORTIFIED
0.008
0.023
-0.002
0.021
-0.005
0.019 0.013
0.012
0.015
0.011
Age 1
-0.011
0.019
-0.009
0.019
-0.014
0.02 -0.021
0.014
-0.017
0.012
Age 2
-0.041
0.037
-0.023
0.031
-0.007
0.024 -0.025
0.018
-0.028
0.017
Age 3
-0.068 *
0.035
-0.048
0.031
-0.041
0.026 -0.034
**
0.016
-0.042 ***
0.015
Age 4
-0.087 *
0.049
-0.092
* 0.047
-0.057
**
0.027 -0.072
***
0.021
-0.056 ***
0.015
Age 5
-0.104
0.066
-0.067
* 0.036
-0.047
**
0.024 -0.057
***
0.016
-0.06 ***
0.015
Age 6-9
-0.128 ***
0.031
-0.104
*** 0.029
-0.081
***
0.026 -0.083
***
0.016
-0.077 ***
0.014
Age 10-14
-0.175 ***
0.030
-0.158
*** 0.024
-0.145
***
0.020 -0.141
***
0.013
-0.140 ***
0.014
Age 15-19
-0.207 ***
0.033
-0.179
*** 0.027
-0.169
***
0.023 -0.175
***
0.019
-0.168 ***
0.017
Age 20-24
-0.250 ***
0.041
-0.21
*** 0.033
-0.198
***
0.026 -0.211
***
0.021
-0.204 ***
0.02
Age 25-29
-0.340 ***
0.043
-0.329
*** 0.042
-0.323
***
0.039 -0.281
***
0.027
-0.276 ***
0.024
Age 30-39
-0.371 ***
0.068
-0.358
*** 0.063
-0.321
***
0.053 -0.28
***
0.045
-0.282 ***
0.041
Age 40-49
-0.478 ***
0.111
-0.381
*** 0.115
-0.406
***
0.100 -0.429
***
0.100
-0.423 ***
0.125
Age 50-74
-0.242 *
0.127
-0.248
* 0.130
-0.182
0.111 -0.148
0.112
-0.162
0.113
Age 75+
-0.410 ***
0.137
-0.415
*** 0.140
-0.355
***
0.131 -0.299
**
0.121
-0.302 **
0.120
Square Footage (1,000)
0.486 ***
0.054
0.481
*** 0.052
0.46
***
0.048 0.507
***
0.034
0.532 ***
0.033
Square Footage²
-0.041 ***
0.007
-0.044
*** 0.006
-0.04
***
0.006 -0.044
***
0.005
-0.048 ***
0.005
Bedrooms
0.002
0.014
0.008
0.014
-0.007
0.013 -0.019
**
0.009
-0.024 ***
0.007
Pool
-0.127 ***
0.040
-0.112
*** 0.036
-0.093
***
0.028 -0.090
***
0.023
-0.091 ***
0.023
Floors
0.023
0.037
0.017
0.035
0.024
0.031 0.063
***
0.022
0.061 ***
0.022
Garage
-0.084
0.059
-0.075
0.059
-0.045
0.052 -0.008
0.034
-0.002
0.027
Fireplace
0.068 ***
0.013
0.068
*** 0.012
0.066
***
0.011 0.064
***
0.008
0.067 ***
0.008
Full Bathrooms
0.181 *
0.107
0.172
* 0.101
0.175
0.111 0.135
*
0.079
0.123
0.078
Full Bathrooms²
-0.024
0.021
-0.021
0.020
-0.020
0.022 -0.009
0.016
-0.005
0.015
Half Bathrooms
0.077 ***
0.019
0.071
*** 0.017
0.064
***
0.017 0.044
***
0.011
0.052 ***
0.010
Miles to Coast
0.004
0.004
0.002
0.004
0.001
0.004
0.001
0.002
0.000
0.002
Coastal
0.042
0.089
0.053
0.086
-0.017
0.099
0.011
0.131
0.122
*
0.066
Flood Zone A
0.061
0.088
0.084
0.077
0.028
0.055
0.064
**
0.03
0.072
***
0.022
Flood Zone V
0.578
***
0.192
0.534
***
0.161
0.394
***
0.099
0.220
***
0.072
0.168
**
0.079
Repeat Sales
-0.011
0.017
-0.008
0.015
-0.010
0.013
0.010
0.008
0.010
0.008
Observations
11,632
11,564
11,464
11,336
11,212
23
23
10
10
2
2
3
3
4
4
5
5
6
6
7
7
8
8
9
9
10
10
11
11
12
12
13
13
14
14
15
15
16
16
19
19
20
20
21
21
22
22
24
24
25
25
26
26
27
27
28
28
29
29
30
30
31
31
32
32
33
33
44
44
34
34
36
36
38
38
40
40
42
42