1# RECORD 35e9cf817db8c35d9d40580566f6516793f3cc922# TITLE: Sea Level Rise, Homeownership, and Residential Real Estate Markets in South Florida3# YEAR: 2020.0 VENUE: Professional Geographer DOI: 10.1080/00330124.2020.18185864# AUTHORS: Xin-Yu Fu; J. Nijman5# PAGES: 11 CANDIDATE TABLE PAGES (1-based): [6, 7]6# ABSTRACT: This article builds on a small but rapidly growing body of research that seeks to determine the impact of sea level rise on the pricing of residential properties. Through a spatial hedonic regression analysis of real estate markets in two Florida counties (Miami–Dade and Pinellas), we assess the influence of different exposure levels on market discounts. Our article stands out in terms of its focus on two comparative case studies and its differentiation between properties that are primary homes versus nonprimary homes. We find that generally discounts are positively associated with exposure levels and overall Miami–Dade experiences higher discounts than Pinellas due to the former’s lower average elevations. We also observe different market behaviors of primary versus nonprimary home buyers and these are partially dependent on affluence. In Miami–Dade, price discounts are less for highly priced properties purchased by nonprimary owners. We attribute this to different buying motives and risk tolerance of affluent nonprimary homeowners. We argue that nonprimary ownership, particularly in high-end waterfront residential real estate, is tempering gradual market adaptation to sea level rise exposure risk, which could have detrimental longer-term consequences in terms of market volatility.789===== PAGE 1 (score 4, numeric-paren 0, table-words 0, var-words 4) =====1011UvA-DARE is a service provided by the library of the University of Amsterdam (https://dare.uva.nl)12UvA-DARE (Digital Academic Repository)13Sea Level Rise, Homeownership, and Residential Real Estate Markets in South14Florida15Fu, X.; Nijman, J.16DOI1710.1080/00330124.2020.181858618Publication date19202120Document Version21Final published version22Published in23The Professional Geographer24License25CC BY-NC-ND26Link to publication27Citation for published version (APA):28Fu, X., & Nijman, J. (2021). Sea Level Rise, Homeownership, and Residential Real Estate29Markets in South Florida. The Professional Geographer, 73(1), 62-71.30https://doi.org/10.1080/00330124.2020.181858631General rights32It is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s)33and/or copyright holder(s), other than for strictly personal, individual use, unless the work is under an open34content license (like Creative Commons).35Disclaimer/Complaints regulations36If you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please37let the Library know, stating your reasons. In case of a legitimate complaint, the Library will make the material38inaccessible and/or remove it from the website. Please Ask the Library: https://uba.uva.nl/en/contact, or a letter39to: Library of the University of Amsterdam, Secretariat, P.O. Box 19185, 1000 GD Amsterdam, The Netherlands.40You will be contacted as soon as possible.41Download date:20 okt 2026424344===== PAGE 6 (score 35, numeric-paren 0, table-words 7, var-words 12) =====4546terms are introduced into the prior model specifica-47tion: if the owner (buyer) of a property is a primary48homeowner, the interaction term for the PHOs49equals 1 and 0 otherwise; the same applies to the50NPHOs. All computations were processed in GeoDa,51an open source spatial econometric software.52Results53Table 3 provides summary statistics of our samples in54both counties. Note again that the mean home values55are nearly twice as high in Miami–Dade, and homes56and especially lot sizes in Miami–Dade are bigger.57The presence of very expensive coastal real estate in58Miami–Dade also causes the standard deviation to be59higher than in Pinellas. The average distance to the60coast61is62about63the64same65in66both67counties,68but69Miami–Dade properties tend to be relatively more70exposed: more homes are located in FEMA flood71zones and more are located below 6-ft elevations,72despite Pinellas having a considerably longer coastline.73The overall regression results for the two counties,74without distinction between primary and non-primary75homeownership, are shown in Table 4. Note that the76price discounts are overall higher in Miami–Dade than77in Pinellas, which can be explained by Miami–Dade’s78greater overall vulnerability due to its considerably79lower elevations. This probably also explains why a80FEMA flood plain location in Miami–Dade shows a81discount, whereas in Pinellas it does not.5 But what is82the relationship between SLR risk exposure at various83levels and pricing within the two counties?84In Pinellas, we observe a progressively higher dis-85count at increased SLR exposure levels, as hypothe-86sized. The discount is by far the highest, 12 percent,87for the most exposed properties below 2 feet. The dis-88count is 2 percent at 2–3 ft exposure. Above 3 ft, there89is actually already a price markup (negative discount):903 percent at 3–4 ft, 6 percent at 4–5 ft, and 4 percent at915–6 ft. In Miami–Dade, however, we observe no signif-92icant discount for the most exposed properties (<2 ft),93but we do see 3 to 7 percent discounts between 2 and 594Table 495Spatial hedonic regression results for Pinellas County and Miami–Dade County96Variables97Pinellas County98Miami–Dade County99Coefficient100Std. error101Coefficient102Std. error103Intercept1042.501050.021062.701070.03108Number of bedrooms1090.031107.2 10–41110.041120.002113Lot size, sq. ftþ1140.071150.0031160.0011178.4 10–5118Living area, sq. ftþ1190.181200.0041210.231220.002123Year property built1241.20 10–4.1256.8 10–51261.40 10–41274.2 10–5128Euclidean distance to coastline, ftþþ1290.011309.4 10–41310.031320.001133Euclidean distance to major road, ftþþ1340.011356.0 10–41360.011370.002138FEMA 100-year floodplain1390.051400.0041410.031420.004143SLR exposure less than 2 ftþþþ1440.121450.021460.021470.03148SLR exposure 2 to 3 ftþþþ1490.021500.011510.071520.01153SLR exposure 3 to 4 ftþþþ1540.031550.011560.031570.01158SLR exposure 4 to 5 ftþþþ1590.061600.011610.011620.01163SLR exposure 5 to 6 ftþþþ1640.041650.011660.011670.005168q1690.451700.0031710.471720.004173R21740.791750.81176Log-likelihood17714,156.117825,291.4179Moran’s I180308.68 (p < 0.001)181228.15 (p < 0.001)182Sample size18377,45418458,264185p < 0.001;186p < 0.01;187p < 0.05;188.p < 0.1.189Notes: Year and city/town dummies were included in the model but omitted here for simplicity;190þ1 sq. ft ¼ 0.0929 sq. m; þþ1 ft ¼ 0.3048 m; þþþ2 ft 0.6 m, 3 ft 0.9 m, 4 ft 1.2 m, 5 ft 1.5 m, 6 ft 1.8 m.191FEMA ¼ Federal Emergency Management Agency; SLR ¼ sea level rise.192Table 3193Summary statistics for regression variables for Pinellas County and Miami–Dade County194Pinellas County195(N ¼ 77,454)196Miami–Dade County197(N ¼ 58,264)198Mean199Std. dev.200Mean201Std. dev.202Housing price (103 US$, unadjusted)203292204264205555206849207Number of bedrooms2081.932091.122103.262110.96212Lot size (103 sq. ftþ)2132.472141.262159.2921619.49217Living area (103 sq. ftþ)2181.762190.852202.042211.20222Year property built (as compared to year 2019)22342.6722419.922254622640227Distance to coast (103 ftþþ)2281.912291.432302.092311.45232Distance to major road (103 ftþþ)2331.992342.062350.942360.80237FEMA floodplain2380.332390.472400.482410.50242SLR exposure (6 ft or 1.8 m)2430.212440.402450.322460.47247Notes: þ1 sq. ft ¼ 0.0929 sq. m; þþ1 ft ¼ 0.3048 m. FEMA ¼ Federal Emergency Management Agency; SLR ¼ sea level rise.24866249Volume 73, Number 1, 2021250251252===== PAGE 7 (score 31, numeric-paren 0, table-words 6, var-words 11) =====253254ft; dropping to 1 percent above 5 ft. Miami–Dade does255not see any substantial discounts above 4 ft. Below, we256return to the apparent anomaly in Miami–Dade of no257discounts for the most exposed properties.258In the next step of the analysis, we differentiate259the impact of SLR exposure between PHOs and260NPHOs. Table 5 shows the regression results with261quite opposite patterns in the two counties. In262Miami–Dade County, discounts are mostly higher263for PHOs than for NPHOs, confirming our hypoth-264eses. However, in Pinellas County, it is NPHOs’265properties that exhibit higher discounts. For the266most exposed properties in Pinellas, the discounts267for PHOs and NPHOs are, respectively, 10 percent268versus 13 percent; at 2–3 ft exposure, they are 1 per-269cent versus 3 percent; and at 3–4 ft exposure the dif-270ferences are a 5 percent price markup versus a 1271percent price discount.272In273Miami–Dade,274the275findings276for277the278most279exposed280properties281are282especially283revealing.284Properties below 2 ft show a substantial 15 percent285discount for PHOs but no significant discount for286NPHOs. Thus, the lack of an aggregate discount for287the most exposed properties in Miami–Dade (see288Table 4) can be attributed to the large presence of289NPHOs. Our data show that 58 percent of all290buyers in this segment are NPHOs, with an average291value of NPHO properties of $1.7 m versus $1.2 m292for PHOs. These prices are much higher than in293Pinellas (NPHO properties below 2 ft in Miami294Dade are 4.6 times as expensive as in Pinellas).295Moreover, in Pinellas, NPHOs are a minority and296NPHO homes are priced lower than PHO homes.297In Miami–Dade, properties between 2 and 3 ft show298discounts of 9 percent for PHOs vs. 5 percent for299NPHOs, and properties between 3 and 4 ft show300discounts of 6 percent for PHOs vs. no significant301discount for NPHOs.302Discussion and Conclusions303Our findings are generally consistent with recent304writings insofar as they indicate a price differential305based on SLR exposure (e.g., Keenan, Hill, and306Gumber 2018).6 At the same time, our study contrib-307utes important new insights in the significance of geo-308graphic variations of real estate markets (Miami–Dade309vs. Pinellas) and in the roles of nonprimary homeown-310ership in market responses to SLR exposure.311The results indicate that SLR exposure risk is312currently impacting the residential real estate mar-313kets in Pinellas County and Miami–Dade County,314with generally higher discounts for more exposed315properties. We find significant heterogeneity in the316degree and extent of the discounting effect between317the two counties: Miami–Dade shows a stronger318overall market response in price discounts than319Pinellas,320which321we322attribute323to324Miami–Dade’s325greater326county-wide327exposure328risk,329due330to331its332Table 5333Spatial hedonic regression results for PHOs and NPHOs in Pinellas County and Miami–Dade County334Variables335Pinellas County336Miami–Dade County337Coefficient338Std. error339Coefficient340Std. error341Intercept3422.503430.023442.703450.03346Number of bedrooms3470.033480.0013490.043500.002351Lot size, sq. ftþ3520.073530.0033540.0013558.4 10–5356Living area, sq. ftþ3570.183580.0043590.233600.002361Year property built3621.2 10–4.3636.8 10–53641.4 10–43654.2 10–5366Euclidean distance to coastline, ftþþ3670.013680.0013690.033700.001371Euclidean distance to major road, ftþþ3720.013730.0013740.013750.002376FEMA 100-year floodplain3770.053780.0043790.033800.004381SLR exposure less than 2 ft & PHOsþþþ3820.103830.033840.153850.05386SLR exposure less than 2 ft & NPHOsþþþ3870.133880.033890.143900.04391SLR exposure 2 to 3 ft & PHOsþþþ3920.013930.013940.093950.02396SLR exposure 2 to 3 ft & NPHOsþþþ3970.033980.013990.054000.02401SLR exposure 3 to 4 ft & PHOsþþþ4020.054030.014040.064050.01406SLR exposure 3 to 4 ft & NPHOsþþþ4070.014080.014090.0034100.01411SLR exposure 4 to 5 ft & PHOsþþþ4120.084130.014140.0024150.01416SLR exposure 4 to 5 ft & NPHOsþþþ4170.0024180.014190.014200.01421SLR exposure 5 to 6 ft & PHOsþþþ4220.054230.014240.0034250.01426SLR exposure 5 to 6 ft & NPHOsþþþ4270.01.4280.014290.024300.01431q4320.454330.0034340.474350.004436R24370.794380.81439Log-likelihood44014,083.744125,265.1442Moran’s I443308.55 (p < 0.001)444228.05 (p < 0.001)445Sample size44677,45444758,264448p < 0.001;449p < 0.01;450p < 0.05;451.p < 0.1.452Notes: Year and city/town dummies were included in the model but omitted here for simplicity;453þ1 sq. ft ¼ 0.0929 sq. m; þþ1 ft ¼ 0.3048 m; þþþ2 ft 0.6 m, 3 ft 0.9 m, 4 ft 1.2 m, 5 ft 1.5 m, 6 ft 1.8 m.454PHO ¼ primary homeowner; NPHO ¼ nonprimary homeowner; FEMA ¼ Federal Emergency Management Agency; SLR ¼ sea455level rise.456Sea Level Rise and Real Estate in South Florida45767458459460===== PAGE 8 (score 6, numeric-paren 0, table-words 0, var-words 6) =====461462considerably463lower464county-wide465elevations.466Miami–Dade’s average elevation is 6 ft (1.8 m), com-467pared to 30 ft (9 m) in Pinellas. Pinellas’s main cities,468Saint Petersburg and Clearwater, have average ele-469vations of 42 ft (13 m) and 32 ft (10 m), whereas the470average elevation of the City of Miami is less than 7471ft (2.1 m). Buyers in Pinellas have more options than472in Miami–Dade to purchase homes at higher eleva-473tions. Concern about SLR exposure is likely to be474more widely shared in Miami–Dade than in Pinellas,475and Miami–Dade has figured more prominently in476media reports and academic debates about sea477level rise.478But market responses to SLR are also limited: In479Pinellas, discounts are limited to properties below 3480ft of exposure and properties between 3 and 6 ft ele-481vation actually see a price markup. In Miami–Dade,482we observe price discounts across the different expo-483sure levels except for the most exposed properties,484below 2 ft. The risk of SLR seems to be inconse-485quential in Miami–Dade’s most exposed market seg-486ments. We attribute this to the different buyer487motives and risk tolerance in this segment: these488buyers are disproportionately affluent and nonpri-489mary homeowners.490Interestingly, we find that the role of second491homeowners492(NPHOs)493varies494between495the496two497counties: in Pinellas, second homes have higher dis-498counts than primary homes, and this appears to499align with the findings of Bernstein, Gustafson, and500Lewis (2019). Here, it is possible that a significant501number of NPHOs are in fact investors who take a502more critical look at SLR-exposed properties and503negotiate prices down and rent out the property.504Accordingly, in Pinellas, our hypothesis does not505hold up. We should note again that across all expo-506sure levels in the Pinellas sample, NPHO prices are507lower than PHO prices, indicating that PHOs in508Pinellas tend to be the more affluent buyers.509In Miami–Dade, the pattern is reversed: there,510the discounts for second homes are consistently less511than those for primary homes. In Miami–Dade, with512large numbers of second home buyers in very expen-513sive market segments, our theoretical premise seems514validated: PHOs are generally more deeply invested515in the community than NPHOs, who typically516reside elsewhere for much of the time. PHOs are517likely to be more concerned with long-term stability,518whereas NPHOs are generally invested in these519properties in the short to medium term, with an eye520on the immediate returns in terms of coastal ameni-521ties and quality of life. NPHOs are also wealthier522buyers with a higher risk tolerance; thus, they are523more financially flexible, more mobile, and less524socially invested in the community.7525Hence, in Miami–Dade County, we find that526nonprimary homeownership tends to dampen mar-527ket adjustments to present and future SLR. The528financial flexibility and mobility of many (affluent)529NPHOs underlies their market behavior: at the pres-530ent time, it allows ignoring projected SLR exposure531and continued preferences for waterfront properties,532but at some point in the foreseeable future it is likely533to trigger an exodus from the same properties.534Hence, present-day intransigence could have detri-535mental longer-term consequences in terms of mar-536ket volatility. These findings may well apply to537other coastal residential real estate markets, many of538which have a substantial share of affluent NPHOs.539What are the planning implications of our find-540ings? First, our study joins several others that541observe market reactions to SLR risk exposure. We542can expect these market reactions to continue and543probably accelerate. This signals a challenge to544many coastal cities that rely heavily on their housing545markets for tax income. Most of these communities546have yet to implement significant adaptations such547as elevating roads, changing building codes, etc. (Fu548et al. 2017, 2019; Fu 2020). Such adaptation is costly549and will become more difficult to finance if market550values of coastal properties and the tax base decline.551This means that the proverbial clock is ticking for552these553communities554to555proceed556with557adapta-558tion measures.559Second, it is striking that property owners in560Pinellas (except at the highest exposure levels, below5612 ft) appear so much less concerned with SLR com-562pared to Miami–Dade, a discrepancy we attribute in563part to Pinellas having received far less attention in564the news media and public debates about SLR com-565pared to Miami–Dade. A Google search in July 2020566on “sea level rise Miami–Dade” returned about567twelve times as many results as “sea level rise568Pinellas.” This apparent gap in risk perception calls569for targeted information for the general public and570local communities in Pinellas and similarly affected571areas. Local workshops on SLR and public engage-572ment with adaptation planning are also desired but573not yet widespread or common.574Finally, our findings suggest that governments at575various levels consider remedying present-day mar-576ket intransigence through market intervention. This577is a major challenge in terms of balancing fairness578and579effectiveness580(Chandra-Putra581and582Andrews5832019) and one that tends to go against the grain in584terms of local governments typically seeking to stim-585ulate short-term growth and expansion of the tax586base rather than the opposite. The goal would have587to be to induce a sufficient and steady long-term588market response to SLR exposure risk, and this589means gradual price discounts in the most exposed590market segments. Nonprimary homeowners across591the United States already face fiscal disincentives by592the way of denial of homestead exemptions in prop-593erty taxes, but our findings suggest that these are594too marginal to affect market behaviors where they595are most needed, on the waterfront in the most596expensive market segments. Sooner rather than59768598Volume 73, Number 1, 2021599