Recent Treacherous US Treasury Borrowing by National Flood Insurance Program

The National Flood Insurance Program (NFIP) experienced “only one catastrophic loss year in its 40-year history, impairing the program’s ability to pay current obligations, administrative expenses, as well as interest on the debt to the Treasury,” writes Rawle O. King, analyst in financial economics and risk assessment, government and finance division, Congressional Research Service (August 11, 2008). (1) That catastrophic loss year was, of course, Fiscal Year 2006, the (fiscal) year in which Hurricanes Katrina, Rita and Wilma struck (note the hurricanes struck in the fall of 2005, which was after the 2006 fiscal year began.) The NFIP is one of the many, many programs operated by the Federal Emergency Management Agency (FEMA).

"Tombstones rise above muddy waters at a Galveston cemetery flooded by the storm surge. Photo: Matt Slocum/Associated Press." Hurricane Ike, September 2008. Source: http://www.nytimes.com/slideshow/2008/09/13/us/0913-IKE_9.html; accessed September 18, 2008.

The Table below from King’s report shows that the NFIP, created in 1968 by the US Congress in response to rising flood losses and escalating costs to the general taxpayers for disaster relief, borrowed $16.66 billion from the US Treasury in Fiscal Year 2006 to cover outlays to NFIP insureds impacted by Hurricanes Katrina and Rita. None of that money has been repaid to the US Treasury and there is serious talk of forgiving the full debt owed by the NFIP to the Treasury (more below). (2) Indeed, the NFIP borrowed $50 million from the US Treasury in early April 2008 just to meet its semi-annual interest payments. “The NFIP’s outstanding borrowing is the combined result of paying insurance claims and servicing the debt on the borrowing,” explains King. (3)

One could grow anxious over the potential insolvency of the NFIP. However, the problem is much greater than the woes of the NFIP. Consider the recent 2008 Midwest flood, which does not even rank in the top fifteen significant flood events in the US in terms of NFIP payouts. Of the total $20 billion in economic loss attributed to the 2008 Midwest flood, only $273 million were in NFIP claims! In a similar manner, Hurricanes Katrina and Rita resulted in more than $200 billion in economic losses of which NFIP covered $21.9 billion. (4) Who is paying for the uninsured losses in these two disasters? The answer is the American taxpayer.

Private insurers have lone excluded flood damage from homeowner’s policies because they consider flood risk to be an uninsurable peril, notes King. “Insurers reportedly cannot accurately estimate losses and most lack the ability to pool and spread flood risks over a large and diverse group of (uncorrelated) insureds in order to minimize the possibility of multiple claims for the same event. “ (5)

The US Congress in 1968, however, believed that it could fill the void created by the private insurance sector by selling subsidized flood insurance to owners of residential and commercial properties in NFIP participating communities. However, the market penetration of flood insurance remains pathetically low, in spite of legislation after legislation to close loopholes. The result is billions of dollars in uninsured losses, such as those associated with the Midwest flood of 2008 and Hurricanes Katrina and Rita in 2005.

What are the root causes of the federal government outlay gap between NFIP insureds and non-insureds? The latter consistently enjoy disaster relief through the various disaster relief programs provided by FEMA. If the owners of residential and commercial properties do not purchase NFIP flood insurance (for whatever reason) and continue to live in at-risk locations, no problem! When the 100-year flood comes along, disaster assistance flows quickly from FEMA to the stressed property owners. The behavior of uninsureds is not irrational; by contrast, it is perfectly rational, given the system in which the uninsureds are operating.
For example, various government entities build levees, which offer property owners a sense of protection but which periodically fail, as they did in 1993, 2005, and 2008. The property owners do not purchase flood insurance—why should they, they reason, when the levees have been built specifically to prevent flooding?

Another example is inaccurate flood maps that do not reflect residual flood risk behind levees, dams and other structural flood control systems.

However, the most astonishing example of incentivizing the irresponsible behavior of uninsureds is that their needs will be provided for whether they buy the flood insurance or not. It is irrational to think that providing disaster relief to uninsured individuals will increase their desire to purchase flood insurance. The low flood insurance market penetration in flood-prone areas is proof of this fallacious way of thinking. (6)

Midwest flood 2008: hog on a levee. Source: http://farm4.static.flickr.com/3141/2613716593_5750b6e7fe.jpg?v=0; accessed September 18, 2008.

• Dramatically reduce the financial cost of multiple flood insurance payments (i.e., repetitive loss properties) that account for a disproportionate share of the NFIP’s total claims payouts. This would require eliminating the premium subsidies available to repetitive loss properties.

• Strengthen floodplain management regulations designed to restrict development in high risk areas and require new construction to be elevated three feet above the base flood elevation.

• Improve flood risk assessment and mapping of the nation’s floodplains and include 500-year floodplains and areas behind levees.

• Strengthen and enforce mandatory purchase requirements.

• Forgive the full debt owed by the NFIP to the Treasury.

• Require actuarially-based premiums in the NFIP. (7)

How has the US Congress responded to the Midwest flood of 2008? Simple. It enacted emergency supplemental appropriations legislation to compensate flood “victims.” On June 30, 2008, President Bush signed into law Public Law 110-252 to, among other things, appropriate $8.48 billion for natural disaster relief and recovery, including $5.64 billion for construction of flood prevention and protection structures in Louisiana and $2.84 billion for flood assistance in Midwestern states. (8)

Hurricane Katrina flooding, 2005. Source: http://serc.carleton.edu/images/research_education/katrina/katrina_flood_400.jpg; accessed September 18, 2008.

Meanwhile the House (H.R. 3121) and Senate (S. 2284) have passed reform and modernization acts for the NFIP, which have not yet (at the time of this writing) been reconciled. H.R. 3121 actually expands the tottering NFIP to cover damage caused by wind (in addition to flooding—the two often occur together). The Senate version lacks this largess. S. 2284 would forgive the $17.5 billion debt owed by the NFIP to the US Treasury and create a “new bipartisan Commission on Natural Catastrophe Risk management and insurance” to study ways to address the availability of catastrophe insurance coverage in high-risk areas. (8)

Summary
The high degree of uninsured flood losses during the 2005 hurricanes and the 2008 Midwest flood raises the policy question of “who should appropriately bear the cost of the decision to live in potentially high-risk areas, including areas behind levees, dams and other flood control structures. In the absence of flood insurance, the cost of repairing the flood damage will be born either by the property owner from [his or her] own financial resources or through federal disaster assistance—not flood insurance payments.” (9) The problem—and it is a big problem—is that since 2005 the NFIP has been forced to borrow heavily to provide this disaster assistance to uninsureds. In the short term, the NFIP is becoming an insolvent insurer of last resort. The only way out of this financial conundrum is a span of disaster-free years that would allow the Treasury debt to be paid off with incoming NFIP premiums.
Notes:

1. Rawle O. King: “Midwest flooding disaster: Rethinking federal flood insurance?” CRS Report for Congress RL 34610, August 11, 2008. Available at http://opencrs.cdt.org/document/RL34610; accessed September 17, 2008.
2. Ibid, CRS-15.
3. Ibid, CRS-17.
4. Ibid, CRS-9.
5. Ibid, CRS-4.
6. Ibid, CRS-7.
7. Ibid, CRS-15.
8. Ibid, CRS-21.
9. Ibid, CRS-12.