Feeling Is Believing? Evidence From Earthquake Shaking Experience and Insurance Demand

• Author: Xiao Lin
• Date: 01 June 2020
• DOI: http://doi.org/10.1111/jori.12291
• Published date: 01 June 2020

© 2019 American Risk and Insurance Association. Vol. 87, No. 2, 351–380 (2020).

DOI: 10.1111/jori.12291

FEELING ISBELIEVING?EVIDENCE FROM EARTHQUAKE

SHAKING EXPERIENCE AND INSURANCE DEMAND

Xiao Lin

ABSTRACT

This article investigates how a particular type of personal experience—

“no‐loss”experience with minor earthquakes—affects financial decisions

such as insurance purchases. We find a small temporary increase in in-

surance demand in areas that experience a shaking with moderate in-

tensity, or multiple shakings with light intensity. An analysis of Google

Trends data confirms an immediate increase in interest in insurance though

not in seismic retrofit. These findings extend the applicability of the

availability bias and hot‐hand fallacy to a broader context: financial deci-

sions may be motivated by not only loss experience, but also recent no‐loss

experience, as people may extrapolate their “feeling”to something worse.

However, such experience does not motivate long‐term behavioral change.

INTRODUCTION

Personal experience influences behavior through new information or through

the experience itself (Simonsohn et al., 2008). Accordingly, empirical evidence on

such can be explained by Bayesian learning model (Viscusi, 1991; Gallagher, 2014)

or by heuristics such as availability bias (Tversky and Kahneman, 1973)or hot‐hand

fallacy (Gilovich, Vallone, and Tversky, 1985).

1

The importance of heuristics in

explaining financial decisions due to experience has been widely recognized in the

Xiao (Joyce)Lin is at the School of Risk Management, Insurance, and Actuarial Science, the

Peter J. Tobin College of Business, St. John’s University, 101 Astor Place, New York, NY

10003. Lin can be contacted via e‐mail: linx@stjohns.edu. I thank Justin Sydnor, Mark Browne,

Jed Frees, Jennifer Alix‐Garcia, and Margie Rosenberg for guidance and advice throughout

different phrases of this project. I also thank Joan Schmit, Martin Halek, Patty Born, Ilyana

Kuziemko, two anonymous referees, participants at the Wisconsin School of Business semi-

nars, participants at the World Risk and Insurance Economics Congress, and participants at

the St. John’s research seminars for helpful comments. Data support from the California

Department of Insurance is gratefully acknowledged.

1

Both availability bias and hot‐hand fallacy are first documented in the psychology literature.

Direct evidence found that people evaluate the probability of an outcome to be higher if that

outcome is more readily available in memory or has just been observed repeatedly.

351

economics literature (Johnson, Tellis, and Macinnis, 2005; Haselhuhn et al., 2012;

Aseervatham, Born, and Richter, 2013). Prior studies focus on examining the effect

of experience of financial loss (or gain)on subsequent financial decisions. Important

questions remain, however, about the influence of “no‐loss”experience. Not all

experiences are equal. For instance, a “feeling”of shaking from a Magnitude 5.0

earthquake should be different from a loss experience from a damaging Magnitude

8.0 earthquake, yet both are similar in the sense that one can easily be related to

another. To test if financial decisions respond to a broader set of personal experience

than what has been widely documented, we turn to a residential earthquake in-

surance market, and investigate homeowners’insurance purchase decisions, among

others, after experiencing earthquake shaking that typically does not involve a fi-

nancial loss.

In the current study, we investigate the role of “feeling”an earthquake in inducing

behavioral response, as well as the degree and type of response induced. We expand

the definition of personal experience by focusing on shaking experience with a set of

relatively minor earthquakes that happened in California from 2003 to 2013. During

this time period, California experienced a handful of earthquakes, yet none are

particularly damaging.

2

Around the same time period, the U.S. Geological Survey

(USGS)published seismic hazard maps every 6 years (2002, 2008, and 2014), and a

visual examination of these maps show little changes in seismic risk in California

over time.

3

A rational economic agent therefore should not change her risk per-

ception or insurance purchase decision. Even a behavioral agent, due to a lack of

loss experience, may not be motivated to purchase more insurance considering that

insurance would not help with mitigating minor earthquake shaking. Any finding

to the contrary suggests possible influence from an extrapolated version of personal

experience.

We find that experiencing shaking from a relatively minor earthquake results in an

increase in the earthquake insurance take‐up immediately, controlling for fixed

effects at the zip‐code level and time‐trend. The intensity of shaking has to reach a

moderate level to have an impact; but a decrease in the shaking intensity can be

compensated by an increase in frequency. However, any effect is short‐lived as we

do not observe any sustained higher level of insurance take‐up beyond 1 year. This

could be due to homeowners who may eventually get around to purchase earth-

quake insurance next year being “pulled forward”to buy insurance immediately, as

a shaking experience makes earthquake risk more salient; it could also be due to

2

The loss ratios for the dominant earthquake insurance underwriter in California have never

been larger than 0.02 percent except in 2003 (the loss ratio was 1.19 percent for that year, and

then was adjusted down in the next 2 years). The direct losses on this insurer’s book were

typically in the tens of thousands of dollars per year, which did not amount to a total loss of

one single house per year. Source: https://www.insurance.ca.gov/0400‐news/0200‐studies‐

reports/upload/EQ_PML_RPT_2002_2010.pdf

3

A place where seismic risk has changed significantly is the Central‐Eastern US, where

fracking activities have induced earthquakes and changed the predicted seismicity. Source:

https://earthquake.usgs.gov/hazards/hazmaps/

352 THE JOURNAL OF RISK AND INSURANCE