The Capitalization of Insurance Premiums in House Prices

• Author: G. Stacy Sirmans,Greg Smersh,Randy E. Dumm,Charles Nyce
• Published date: 01 December 2015
• Date: 01 December 2015
• DOI: http://doi.org/10.1111/jori.12041

THE CAPITALIZATION OF INSURANCE PREMIUMS IN

HOUSE PRICES

Charles Nyce

Randy E. Dumm

G. Stacy Sirmans

Greg Smersh

ABSTRACT

This study uses Miami-Dade County, Florida home sales, and Citizens

Property Insurance Corporation data for the period 2004 through 2009 to

measure the capitalization effect of increases in premiums on house prices.

Using hedonic pricing models, spatial autocorrelation models, and differ-

ence-in-differences models, we find that new information was conveyed to

homeowners in the higher risk areas by the 2004/2005 storms and that

consumers appear to use the insurance premium as a “risk signal.” We also

find some support for the hypothesisthat the risk of potential hurricane losses

is communicated to potential homebuyers through windzone maps.

INTRODUCTION

Using an expected utility approach, we examine the effect of natural disasters on

insurance premiums and, in turn, the effect on local competition relative to the

demand for housing. The macroeconomic model developed in Nyce (1999) and the

microeconomic models developed in MacDonald, Murdoch, and White (1987) and

Bin, Kruse, and Landry (2008) provide the theoretical foundation for how natural

hazard risk and insurance repricing of that risk can affect house prices. The study uses

hedonic pricing models, spatial autocorrelation models, and difference-in-differences

(DND) models to measure the effect of changes in homeowners’ insurance premiums

on house prices in the Miami-Dade County area. We show how the reestimation of

future insurance losses following a natural disaster increases insurance premiums

and thereby lowers property values. As such, this article adds to the literature in a

meaningful way by examining how and to what extent insurance costs affect broader

Charles Nyce is an Assistant Professor of Risk Management and Insurance at the Florida State

University. Randy E. Dumm is the William T. Hold Professor of Risk Management and

Insurance at the Florida State University. G. Stacy Sirmans is the Kenneth G. Bacheller Professor

of Real Estate at the Florida State University. Greg Smersh is an Assistant Professor of Real

Estate at the University of South Florida. Charles Nyce can be contacted via e-mail: cnyce@cob.

fsu.edu. The authors wish to thank the Florida Catastrophic Storm Risk Management Center for

financial support of this article.

© 2014 The Journal of Risk and Insurance. 82, No. 4, 891–919 (2015).

DOI: 10.1111/jori.12041

891

market economic activity. In addition, we believe that a better understanding of why

and how insurance supply and demand factors affect markets in general is critical for

providing more effective public policy decisions regarding the regulation and pricing

of insurance. This article informs those decisions by providing key insights into how

insurance costs impact on real estate prices.

Previous research, using primarily floodplain data, has shown that increased

insurance costs are negatively capitalized into house prices. While flood and

windstorm risks share some similarities (i.e., both are catastrophic and have areas of

high exposure), there are significant differences between the two as flood coverage

has greater coverage availability, higher subsidy levels, and lower potential premium

volatility.

1

Another distinction between these two catastrophic exposures relates to the type of

information or signal that consumers receive when purchasing homes. Mortgage

lenders require a survey that includes flood mapping as a condition of sale; therefore,

buyers are made aware of the flood risk without the flood insurance premium acting

as the primary risk signal. In comparison, mortgage lenders do not require that

borrowers specify that the house location is within a specific windzone.

2

The result

may be that consumers are less well informed about windstorm risk than flooding

risk as the windzone maps are not as well known. Therefore, the windstorm premium

may convey more information to consumers than flood insurance premiums. In other

words, accurate insurance pricing can better objectify catastrophic windstorm risk,

which has historically been subjectively misperceived (Kunreuther, Novemsky, and

Kahneman, 2001; Daniel, Florax, and Rietveld, 2009; Kunreuther and Michel-

Kerjan, 2009).

As a recurring cost of owning a home, the premium for homeowners’ insurance has

an effect on the homeownership decision, but in areas not exposed to catastrophic risk

one would expect this impact to be minimal. However, in areas with high insurance

costs or rapidly increasing premiums, these costs would likely have an adverse effect

on housing demand where the resulting capitalization of these higher costs into house

prices would create a downward pressure on prices. The supply of housing could also

be affected if the cost of homeownership becomes prohibitive and homeowners

1

These differences are driven in part by who provides coverage as the National Flood Insurance

Program is the primary writer of flood coverage in the United States while catastrophic wind

risk coverage is provided by private insurers or residual insurance markets at the state level.

2

For flood risk, we would argue that it is the maps that are the risk signals, not necessarily the

insurance premium. Flood risk has risk maps (floodplain maps) and the insurance premiums

for flood risk are based on these maps. While there are risk maps (windborne debris zones) for

catastrophic wind risk, these maps are more opaque to potential homebuyers and it is unlikely

that this information is revealed to buyers at the time of sale like the flood maps. There is some

mapping of windstorm risk through windborne debris zones. These zones highlight

maximum wind speeds for engineering and building codes purposes, but the windzones

contain no information about the meaning of the windzone category or about the frequency

those wind speeds occur.

892 THE JOURNAL OF RISK AND INSURANCE