Sea Level Rise, Homeownership, and Residential Real Estate Markets in South Florida

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.

本文研究海平面上升对住宅房产的影响，目前这类研究比较少但在逐渐增加。通过美国佛罗里达州Miami-Dade 县和Pinellas县房地产市场的空间Hedonic 回归分析，评价了海平面上升的暴露水平对市场降价的影响。通过两个案例，讨论了主要住宅（用于居住）和非主要住宅的差异。我们发现，降价与暴露水平呈正相关性，处于低海拔的Miami-Dade 县的降价超过Pinellas 县。我们还关注了主要住宅和非主要住宅的购买者的不同市场行为，这些行为部分依赖于富裕程度。在Miami-Dade 县，非主要住宅拥有者所购买的高价房产的降价较少，这是因为富裕的非主要住宅拥有者的不同购买动机和风险承受能力。我们认为，非主要住宅的所有权，特别是高端、临海的住宅房地产市场，正在逐渐适应海平面上升的暴露风险，这可能带来市场波动的长期不利后果。

Este artículo contribuye su parte a un cuerpo de investigación pequeño, pero de crecimiento rápido, que busca determinar el impacto de la elevación del nivel del mar sobre el precio de las propiedades residenciales. Por medio de un análisis de regresión hedonista espacial de los mercados de finca raíz en dos condados de la Florida (Miami-Dade y Pinellas), evaluamos la influencia de diferentes niveles de exposición sobre los descuentos del mercado. Nuestro artículo sobresale en términos de su foco en dos estudios de caso comparativos y su diferenciación entre propiedades que se destinan primariamente para residencia contra las que primariamente no son para residencia. Hallamos que generalmente los descuentos están asociados positivamente con los niveles de exposición y que, en general, Miami–Dade experimenta descuentos más altos que Pinellas debido a los promedios de elevación más bajos que se registran en el primero. Observamos también diferentes comportamientos del mercado en los compradores primarios de residencias frente a los no primarios y estos parcialmente dependen del factor riqueza. En Miami–Dade los descuentos en el precio son menores para propiedades de muy alto precio adquiridas por propietarios de intención no primariamente residencial. Esto lo atribuimos a diferentes motivos de la compra y a la tolerancia al riesgo de los propietarios ricos con propósito no primariamente residencial. Sostenemos que la propiedad no primariamente residencial, en particular en las unidades residenciales de lujo del litoral, está atemperando una adaptación gradual del mercado al riesgo de exposición a la elevación del nivel del mar, que podría tener consecuencias perjudiciales a largo plazo en términos de volubilidad del mercado.

Key words:

• Florida
• homeownership
• real estate
• risk
• sea level rise

关键词:

• 佛罗里达州
• 房屋所有权
• 房地产
• 风险
• 海平面上升

Palabras clave:

• elevación del nivel del mar
• finca raíz
• Florida
• propiedad de la vivienda
• riesgo

Acknowledgments

We are grateful to Risa Palm and Dan Immergluck for helpful comments on an earlier draft of this article. The usual disclaimer applies.

Notes

1 Bernstein, Gustafson, and Lewis (2019) define non-owner occupancy as the buyer not being registered at the property after the sale; their data are from the real estate assessor and transaction data sets in the Zillow Transaction and Assessment Dataset. Their usage of the term “non-owner occupiers” can be confusing because they mean to refer to the buyers of subsequently non-owner-occupied properties, not the occupiers who are typically renters. Our definition of nonprimary homeownership pertains to purchases by nonprimary homeowners. In our data, we cannot tell whether these buyers subsequently occupy the home or not. Clearly, some second home buyers make the purchase as an investment and do not live there; if so, and in case they rent out the property, it would be non-owner occupied (cf. Bernstein, Gustafson, and Lewis 2019), and there is likely to be some overlap among these categories. However, because our analysis is confined to single-family homes (and less likely to involve large multi-property investors that operate in the rental market), we assume that most nonprimary home buyers in our database do occupy the property, though typically only part of the time.

2 Palm and Bolson (2020) provide compelling evidence that there is little change in perceptions about climate change among residents in coastal areas in the United States. Our research data pertain to property sales and this is, of course, a distinctly different population of sellers and buyers (the latter often coming from elsewhere). We would suggest that residents of coastal real estate who are not selling are more likely to play down or deny notions of climate change or sea level rise, if only to protect their home value. An actual sale makes for a different occasion: it involves possibly a seller who is cognizant of increased risk and a buyer seeking to leverage notions of exposure risk to lower the price.

3 Unlike studies on the impact of flood zones (or flood insurance) on home prices, we exclude properties located in interior flood zones, along inland waterways and other bodies of water, with generally different market segments than along the coast and where we expect risk perceptions of sea level rise to be less.

4 Tests for model misspecifications found no violations. Various coastal distance bands were also tested, and the findings were generally consistent. Besides the key independent control variables, the dummy variables of month/year of the sales and cities where sales occurred were included to account for temporal and spatial heterogeneity in sale prices. Also, we created ten spatial distance matrices with different distance bands (from 100 ft to 1,000 ft in 100-ft increments) and found that the parameters of the variables were generally insensitive to the changes of the spatial weight matrix. Finally, a 5,000-ft distance band was employed to create the spatial weight matrix for both counties because it led to the best fit of the models.

5 Location in a flood zone corresponds to a price markup of 5 percent in Pinellas and a price discount of 3 percent in Miami–Dade. In theory, properties in flood zones would be expected to show a price discount, but existing research does not always bear that out (e.g., Lamond, Proverbs, and Antwi 2005; Bin and Kruse 2006; McKenzie and Levendis 2010; Posey and Rogers 2010; Indaco, Ortega, and Taspinar 2018).

6 At the same time, our findings appear to differ from some studies on the impact of flood plain locations. Indaco, Ortega, and Taspinar (2018) find no significant price discounts for properties located in flood plains in Miami–Dade County. But they include, as do most flood plain studies, interior flood zones, whereas we focus exclusively on the 5,000-ft coastal band. We took this approach because the coastal and interior zones represent very different real estate market price segments and because we surmise that risk perceptions related to sea level rise also stronger in the coastal band than in the interior. Hence, we do not contest the findings of Indaco, Ortega, and Taspinar (2018); after all, their research question is on the effects of flood zone locations that extend well beyond our coastal band. But if the question is about the impact of sea level rise risk exposure, we think our particular geographic focus is more valid.

7 It may be that in our samples Pinellas has a larger share of non-owner occupants than Miami–Dade; that is, NPHOs as investors who rent out the property (our data set does not have that information). It seems a plausible, if partial, explanation, given that the Pinellas market segments are more suitable to a broad renter market than the much more expensive segments in Miami. A comparison of overall owner occupancy rates in selected municipalities with predominantly single-family homes (our analysis focuses solely on single-family homes) does seem to bear this out: in Pinellas, the waterfront municipalities of Madeira Beach and Belleair Bluffs have owner occupancy rates of 66 percent and 67 percent, respectively; in Miami–Dade, this compares to 85 percent and 93 percent in the waterfront municipalities of Palmetto Bay and Golden Beach.

Additional information

Notes on contributors

Xinyu Fu

XINYU FU is Lecturer of Environmental Planning in the School of Social Sciences at the University of Waikato, Hamilton 3240, New Zealand. E-mail: [email protected]. His research interests include urban resilience, climate change adaptation, and planning.

Jan Nijman

JAN NIJMAN is Distinguished University Professor and Founding Director of the Urban Studies Institute, Andrew Young School of Policy Studies, Georgia State University, and Professor of Geography, University of Amsterdam. E-mail: [email protected]. His research interests are in urban and regional development, urban-global linkages, and comparative urbanism.