The National Flood Insurance Program: Is It Financially Sound?

• Published date: 01 March 2019
• Author: Perry Beider,David Wylie,Terry Dinan
• DOI: http://doi.org/10.1111/rmir.12116
• Date: 01 March 2019

Risk Management and Insurance Review

Published 2019. This article is a U.S. Government work and is in the public domain in the USA,

Vol.22, No. 1, 15-38

DOI: 10.1111/rmir.12116

FEATURE ARTICLE

THE NATIONAL FLOOD INSURANCE PROGRAM:

ISITFINANCIALLY SOUND?

Terry Dinan

Perry Beider

David Wylie

ABSTRACT

This article uses data provided by the Federal Emergency Management Agency,

which implements the NFIP, to estimate the difference between annual premi-

ums and expected costs associated for the program asa whole and for inland and

coastal regions. In addition, we examine the role of discounts, cross-subsidies,

and FEMA’s method of setting what it considers to be full-risk rates in explain-

ing the outcomes that we observe.

INTRODUCTION

In 1968, Congress established the National Flood Insurance Program (NFIP), which

currently provides the vast majority of flood insurance in the United States. The Federal

Emergency Management Agency (FEMA), which administers the NFIP, develops flood

maps and sets insurance rates, and the federal government underwrites the program’s

risk. The program is authorized to borrow funds from the U.S. Treasury (up to a limit

established by Congress) if the premium receipts from the program fall short of the

program’s costs. Since the NFIP’s inception, lawmakers have sought to balance two

competing goals: setting rates that are actuarially sound and keeping rates low enough

to ensure program participation.

For many years of the NFIP’s history, the program’s receipts and costs were roughly in

balance. In 2005, however, the NFIP experienced an unprecedented volume of claims

resulting from Hurricanes Katrina, Rita, and Wilma, causing the programto borrow $17

billion from the Treasury (Congressional Budget Office, 2009). Subsequent hurricanes,

particularly Sandy in 2012 and Harvey in 2017, added more debt and, at present, the

NFIP owes the Treasury $20.5 billion, even though lawmakers forgave $16 billion of

NFIP debt in 2017. The NFIP’s debt has raised concerns about the financial soundness

of the program and reignited discussions of the appropriate balance between actuarial

and affordability goals.

This article sheds light on the NFIP’s fiscal health. In particular, we find that the program

has an expected 1-year shortfall of $1.4 billion and we identify two primary causes of

The corresponding author, Terry Dinan, may be reached at 202-226-2927 or by email at

Terry.Dinan@cbo.gov. The authors work at U.S. Congressional Budget Office.

15

16 RISK MANAGEMENT AND INSURANCE REVIEW

that shortfall: the use of “full-risk” rates that fall below modeled estimates of expected

claims and the use of discounted rates. In addition, we demonstrate that the balance

between expected costs and premium collections varies between inland counties and

coastal counties, with coastal counties having a net shortfall of $1.5 billion and inland

counties having a net surplus of $0.2 billion. We demonstrate that the difference in the

balance between coastal and inland counties is due to both explicit and implicit subsidies

that are built into NFIP rates.

Previous research has discussed the program’s history and its current operation, in-

cluding the tension between competing goals (Knowles and Kunreuther, 2014), factors

that affect demand for flood insurance (Michel-Kerjan and Kousky, 2010; Landry and

Jahan-Parvar, 2011; Michel-Kerjan et al., 2012), policies that might boost the demand for

flood insurance (Dixon et al., 2006; Michel-Kerjan et al., 2012; Kousky and Kunreuther,

2014; National Research Council, 2015; GAO, 2017), the effects of flooding and of flood

insurance on home values (Bin et al., 2008; Kousky,2010), and the redistributional effects

of the NFIP (Bin et al., 2017).

Most closely related to our work are studies that examined data on historical claims to de-

termine the importance of various factors in determining the magnitude and frequency

of claims (Michel-Kerjan and Kousky, 2010; Kousky and Michel-Kerjan, 2015) and that

examined the rate-setting process to shed light on implicit and explicit cross-subsidies

(Congressional Budget Office, 2009; Kousky et al., 2017). Our work builds on those stud-

ies. In particular, we use forwardlooking estimates of expected costs to examine the mag-

nitude and causes of imbalances between expected costs and premiums. Because of the

infrequent and irregular occurrence of floods, particularly catastrophic ones, a compari-

son of the NFIP’s expected costs with its revenue provides a moremeaningful assessment

of the program’s fiscal health than comparisons of historical costs and revenues.1

Data provided to us by FEMA allowed us to conduct this research. Those data include

estimates of county-level expected flood losses from storm surges and precipitation

from inland storms—estimates generated by commercially available catastrophic risk

models to inform FEMA of its risk exposure prior to purchasing reinsurance for the

2017 hurricane season. In addition, FEMA and their consultants provided technical

advice that informed our efforts to construct two elements of expected flood losses that

catastrophic risk models did not have the capability to provide: expected losses due

to hurricane-related precipitation and tropical storms. Finally, FEMA provided us with

policy-level information on the 5.1 million NFIP policies in place on August 31, 2016,

which allowed us to map policies into counties and flood zones, examine the financial

health of subsets of the program, and identify factors that affect variations in financial

health.2While FEMA and their consultants provided data and technical advice, they

were not partners in this research and did not influence our findings or conclusions.

The next two sections provide background on the NFIP and a description of the data

and methodology used in our analysis. Those sections are followed by four more, each

1This article provides a technical description of the analysis underlying the findings of a 2017

Congressional Budget Office Report, The National Flood Insurance Program: Financial Soundness

and Affordability, and provides additional sensitivity analysis not contained in that report.

2FEMA does not generally provide public access to policy-level data due to privacy concerns.

Our access to those data was due to our affiliation with the Congressional Budget Office.