THE IMPACT OF NATURAL DISASTER RISK ON US 
MUNICIPAL BONDS 
 
 
 
 
 
A Thesis 
Presented to the Faculty of the Graduate School 
of Cornell University 
In Partial Fulfillment of the Requirements for the Degree of 
Master of Science 
 
 
 
 
 
by 
Ming Ju 
August 2020 
 
 
 
 
 
 
 
 
 
 
© 2020  Ming Ju 
 
  
ABSTRACT 
 
Hurricane Katrina has raised awareness of the potential impacts of hurricanes on municipalities, 
however, it remains unclear how natural disaster risk is perceived in the municipal bond market 
nationwide. I attempt to fill this gap by conducting an analysis to determine if natural disaster 
risk and disaster damage affect interest costs for municipal bond issuers in US during 2000-2010. 
Using county level natural hazard data, I find both disaster risk and damage matter in 
determining the interest costs for municipalities issuing debt, but only shortly after severe 
disasters. In the long run, there is no significant impact of natural disaster risk or disaster damage 
on municipal bond interest rate. 
 
Key Words: Municipal Bond, Natural Disaster 
  
BIOGRAPHICAL SKETCH 
 
Ming Ju was born in Beijing, China. After completing her schoolwork at Beijing Experimental 
High School Attached to Beijing Normal University in 2014, she entered Peking University in 
Beijing, China. She received a Bachelor of Science with a major in economics from Peking 
University in July 2018. In August 2018, she entered the applied economics and management 
master program in Cornell University.  
ACKNOWLEDGEMENT 
 
Foremost, I would like to express my sincere gratitude to my advisor Prof. David Ng for the 
continuous support of my Master study and research, for his patience, motivation, enthusiasm, 
and immense knowledge. His guidance helped me in all the time of research and writing of this 
thesis. I could not have imagined having a better advisor and mentor for my study.  
 
Besides my advisor, I would like to thank the rest of my thesis committee: Prof. Byoung-Hyoun 
Hwang, for his encouragement and insightful comments.  
 
I thank my fellow students in Applied Economics and Management apartment for the stimulating 
discussions and for all the fun we have had in the last two years. I also thank my friends Anna 
Giesmanm and Autumn Pratt for supporting me throughout my program. 
 
Last but not the least, I would like to thank my parents for giving birth to me at the first place 
and supporting me spiritually throughout my life. 
 
  
TABLES 
 
THE IMPACT OF NATURAL DISASTER RISK ON US MUNICIPAL BONDS ............................1 
ABSTRACT......................................................................................................................................3 
INTRODUCTION ............................................................................................................................7 
LITERATURE REVIEW .................................................................................................................9 
DATA ............................................................................................................................................. 12 
DATA SOURCE ................................................................................................................................. 12 
SUMMARY STATISTICS .................................................................................................................... 13 
METHODOLOGY ......................................................................................................................... 16 
EMPIRICAL ESTIMATION ......................................................................................................... 20 
CONCLUSION ............................................................................................................................... 23 
REFERENCE ................................................................................................................................. 24 
 
  
INTRODUCTION 
 
The municipal bond market has long suffered the lack of efficiency. Studies have shown that 
information inefficiencies of municipal bonds imply that the efficiency of the municipal bond 
market is between a weak and semi-strong market. In contrast, scholars of the corporate bond 
market detect a semi-strong efficiency that anticipates changes in issuer risk prior to a bond 
rating downgrade. This could be attributed to several factors. First, because municipal bonds are 
tax-exempt, a large segment of municipal bonds is held by households and insurance companies 
for long term investment purposes. Second, municipal bonds trade in decentralized markets, 
therefore lacking centralized information exchange channels. Some states only allow exemption 
of their own municipalities’ bonds from state income tax, which further contributes to 
segmentation of the market. Third, many municipal bonds are issued by small municipalities 
with a limited history in the bond market. Fourth, the majority of municipal bond investors 
follow a buy and hold strategy, which result in a relatively illiquid market. 
 
Empirical evidence suggests that the bond market does not always react efficiently to the release 
of new information. Denison (2000) and Halstead, Hegde, and Klein (2004) investigate the 
market response to the announcement of the Orange County bankruptcy. Denison (2000) finds 
that the average share prices of California bond funds experienced a statistically significant 
decline on the day of the bankruptcy announcement. Although the financial losses of the Orange 
County bankruptcy and Hurricane Katrina are both unusually large in magnitude, and both 
events had concentrated geographical impacts, investors’ reactions to the two events were 
different. As for the Orange County bankruptcy, markets reacted days before the official 
announcement. On the contrary, Hurricane Katrina was identified days before landfall and yet 
the market did not react until after the storm passed and the magnitude of the damage was 
realized. 
 
Recent studies have shown an increasing trend in extreme damages from natural disasters, which 
is consistent with a climate-change signal. Coronese, Lamperti, Keller, Chiaromonte and 
Roventini (2019) document increases in aggregated or mean damages have been modest, but 
evidence for a rightward skewing and tail fattening of the distributions is statistically significant 
and robust. This pattern is strongest in temperate regions, suggesting that the prevalence of 
devastating natural disasters has broadened beyond tropical regions. Their results indicate that 
while the effect of time on averages is hard to detect, effects on extreme damages are large, 
statistically significant, and growing with increasing percentiles. It is highly likely that natural 
disasters would have a bigger impact on the US municipal market in the future.  
 
This paper attempts to provide some new insights into the efficiency of the municipal market 
through an empirical analysis of the impact of natural risk and damage on US municipal bond 
pricing both before and after Hurricane Katrina. This paper proceeds as follows. Section 2 
reviews literature on natural disasters and municipal bonds. Section 3 describes the data used in 
this paper. Section 4 introduces the empirical methodology. Section 5 explains the empirical 
results. Section 6 concludes. 
 
 
LITERATURE REVIEW 
 
Before Hurricane Katrina, little theoretical and empirical attention had been paid to impact of 
natural disaster risks on local economies and the municipal bond market. Settle (1985) suggests 
that the financial consequences of a disaster are consisted of four factors: loss of tax base; loss of 
business affecting the source of sales taxes; amount of money the local government has 
borrowed in relationship to taxable property (debt ratio); and the number of income sources such 
as service charges. 
 
Hurricane Katrina, with its catastrophic and unexpected consequences, evoked research interest 
in this field. Vigdor (2008) examined the economic aftermath of Hurricane Katrina in New 
Orleans. According to this paper, New Orleans lost more than 50% of its population and only 
recovered 20% after three years. The population became more aged and economically 
disadvantaged. The proportionate reduction in the housing stock exceeded the reduction in 
population, with signs indicating it would not return to its earlier pre-Katrina equilibrium. 
Hurricane Katrina also reduced both the number of workers and the number of firms operating in 
the city of New Orleans. Unlike Chicago and San Francisco, New Orleans had suffered decrease 
in population and housing prices decades before Hurricane Katrina, making a post-disaster 
trajectory difficult. In conclusion, Vigdor claims that New Orleans will converge to a new 
equilibrium which is below the pre-Katrina equilibrium. 
 
Apart from property damage to local communities, natural disasters could also threaten liquidity 
of municipal bonds. Property and casualty insurers have substantial holdings in municipal credits 
and are prone to liquidate those holdings to free up cash needed to pay claims in the wake of 
major natural disasters. Marlowe (2006) examines how Hurricanes Katrina, Rita, and Wilma 
(KRW) affected trading activity in the secondary market for municipal securities. Using 
Municipal Securities Rulemaking Board (MSRB) data from the second half of 2005, he finds 
little evidence of a widespread market response to these hazards. One important exception is that 
there was a sizable selloff of Louisiana credits and a subsequent increase in liquidity risk 
associated with those credits. These findings imply that a select group of investors responded to 
the threat KRW posed by selling out of their positions in municipal bonds. His findings show 
that the municipal market was largely unaffected by KRW, but the potential may exist for a 
large-scale sell-off if a similarly sized natural disaster were to affect a more robust geographic 
segment of the municipal market. 
 
Jacob Fowles, Gao Liu, And Cezar Brian Mamaril (2009) study whether underlying geologic 
earthquake risk affects interest costs for municipal bond issuers in California. Their results 
suggest that Hurricane Katrina seems to have changed how earthquake risk is perceived by 
investors in California municipal bonds. They find that before Katrina, underlying earthquake 
risk was not a significant predictor of borrower interest costs; while municipalities issuing debt 
after Katrina pay a premium to investors that is proportional to the municipality’s assessed 
underlying earthquake risk. 
 
 
 
 
  
DATA 
 
Data Source 
 
In this section, I describe the sources of my data and then review summary statistics. Data on the 
characteristics of municipal bonds comes from Bloomberg municipal bond database. In this 
paper, I only include bonds issued by a single county, with a maturity of greater or equal to 1 
year and have credit ratings offered by at least one of the three largest credit rating agencies, 
Standard & Poor’s, Moody’s and Fitch. For bonds with multiple ratings, I use the lowest rating 
for prudence.  
 
In this paper, I use the Bond Buyer Go 20-Bond Municipal Bond Index (BBI20) to capture the 
fluctuations of the municipal bond market on the whole. Starting from 1953, this index is 
published by Bond Buyer daily. BBI20 is a representation of municipal bond trends based on a 
portfolio of 20 general obligation bonds that mature in 20 years. The index is based on a survey 
of municipal bond traders rather than actual prices or yields. This paper uses data provided 
Federal Reserve Bank of St. Louis (FRED) weekly.  
 
I look at natural disasters from two dimensions, natural hazard risk and natural hazard damage. 
Hazard risk is the potential risk of occurrence of natural hazard, incorporating the magnitude of 
the hazard. Natural disaster damage measures the actual damages caused by natural hazards.  
 
Data on multiple hazard index comes from the National Center of Disaster Preparedness, 
Columbia University. The multiple hazard index for the United States counties was designed to 
map natural hazard relating to exposure to multiple natural disasters. The multiple hazard index 
was created by coding the individual hazard classifications and summing the coded values for 
each county (excluding counties in Alaska and Hawaii). Each individual hazard is weighted 
equally in the multiple hazard index. The index covers avalanche, earthquake, flood, heat wave, 
hurricane, landslide, long-term drought, snowfall, tornado, volcano and wildfire hazards. This 
index allows me to compare the hazard risks of counties across states and hazard types. 
 
Data on natural disasters comes from Federal Emergency Management Agency (FEMA), 
National Oceanic and Atmospheric Administration (NOAA) and Spatial Hazard Events and 
Losses Database for the United States (SHELDUS). FEMA provides county-level data on 
disaster declarations. NOAA provides a list of billion-dollar natural disasters that occurred in US 
since 1980, the affected area and the CPI adjusted estimate cost of the disasters.  SHELDUS 
offers more detailed annually aggregated information on crop damages, property damages, 
injuries and fatalities caused by natural disasters on the county level.  
 
Summary Statistics 
 
[Table 1] 
 
[Graph 1] 
 
Table 1 presents summary statistics. The mean coupon rate is 4.71, with a maximum of 12 and a 
minimum of 0.5, and a standard deviation of 1.03. The mean BBI20 rate is 4.37, with a 
maximum of 5.85 and a minimum of 3.82. The mean BBI20 standard deviation is 0.07. Graph 1 
demonstrates the fluctuation of average monthly coupon rate and BBI20 from 2000 to 2010. The 
mean maturity is 16.57 years, with a maximum of 49 and a minimum of 7, and the standard 
deviation of maturity is 5.95. The mean market size is 6.50 million dollars; the maximum is 526 
million and the minimum is 3000 dollars. The mean real total GDP is 213.16 million dollars; the 
maximum is 1.3 billion and the minimum is 2 million dollars. 
 
[Table 2] 
 
[Table 3] 
 
About 48.8% of bonds in this sample are general obligation bonds; 9.3% are callable bonds and 
21.0% are insured. Investment level bonds account for 99.8% of all bonds in the sample, while 
speculation level bonds take up the remaining 0.2%. 8.3% of bonds are grade AAA; 0.4% of 
bonds are grade AA+; 5.33% of bonds are grade AA; 2.31% of bonds are grade AA-; 0.22% of 
bonds are grade A+, 78.48% of bonds are grade A; 2.42% of bonds are grade A-; 0.52% of bonds 
are grade BBB+; 1.38% of bonds are grade BBB; 0.53% of bonds are grade BBB-; 0.1% of 
bonds are grade BB; 0.02% of bonds are grade B+. 
 
[Graph 2] 
 
Graph 2 shows the multi-hazard indexes of US counties. Blue indicates low hazard index and red 
indicates high hazard indexes. The mean hazard index is 13.11; the maximum is 19; the 
minimum is 7; the standard deviation is 1.93. The counties with the highest hazard index are 
mainly located in California and Florida. California bears high risk of wildfire, heatwave, 
landslide, drought and earthquake, while Florida bears high risk of wildfire, heatwave, hurricane 
and flood. The mean property damage is 24.8 million dollars, with a minimum of 0 and 
maximum of 948 million dollars, and a standard deviation of 178 million. The mean crop 
damage is 589.6 thousand dollars, with a minimum of 0 and maximum of 234 million dollars, 
and a standard deviation of 5.6 million. The mean injury per capita is 1.4 in one million, with a 
minimum of 0 and maximum of 990 in one million. The mean fatal rate is 2.1 in ten million, with 
a minimum of 0 and maximum of 80 in one million. 
 
  
METHODOLOGY 
 
The interest rate of a tax-exempt bond can be viewed as the sum of after-tax risk-free 
interest rate and the risk premium: 
𝑅𝑖 = (1 − 𝜏)𝑟 + 𝜎𝑖 
 
where 𝑅𝑖 is the interest rate of a municipal bond; 𝜏 and 𝑟 are the marginal income tax rate 
and the market risk-free interest rate; and 𝜎𝑖 is the unsystematic risk premium of a particular 
bond. 
 
Building on this standard model, studies have explored factors that determine the risk premium 
𝜎𝑖 and expanded the original model to: 
 
𝑅𝑖 = 𝑓(𝑟𝑚) + 𝜎𝑖(𝑍𝑖(𝑋𝑖 , 𝑂𝑖), 𝑋𝑖 , 𝐵𝑖) 
 
where 𝑟𝑚 denotes the municipal bond market benchmark rate, and 𝜎𝑖 represents the municipal 
bond’s spread to the benchmark rate; 𝜎𝑖 depends on the issue’s credit rating 𝑍𝑖, the focal factor 
𝑋𝑖 and other control variables. The issue’s credit rating 𝑍𝑖 is also a function of 𝑋𝑖 and other 
determinants (𝑂𝑖). 
 
In this paper, I use OLS estimation to estimate the overall correlation between natural hazards 
and municipal bond coupon rate. The first two models are basic OLS model.  
 
Model 1: 
𝐶𝑜𝑢𝑝𝑜𝑛 = 𝛽0 + 𝛽1ℎ𝑎𝑧𝑎𝑟𝑑𝑠 + 𝛽2𝐺𝑂 + 𝛽3𝑐𝑎𝑙𝑙𝑎𝑏𝑙𝑒 + 𝛽4𝑐𝑜𝑛𝑡𝑟𝑜𝑙 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒𝑠 + 𝜀 
 
Model 2: 
 
𝐶𝑜𝑢𝑝𝑜𝑛 = 𝛽0 + 𝛽1𝑐𝑟𝑜𝑝 + 𝛽2𝑝𝑟𝑜𝑝𝑒𝑟𝑡𝑦 + 𝛽3𝑖𝑛𝑗𝑢𝑟𝑖𝑒𝑠 + 𝛽4𝑓𝑎𝑡𝑎𝑙𝑖𝑡𝑖𝑒𝑠 + 𝛽5𝐺𝑂 + 𝛽6𝑐𝑎𝑙𝑙𝑎𝑏𝑙𝑒
+ 𝛽7𝑐𝑜𝑛𝑡𝑟𝑜𝑙 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒𝑠 + 𝜀 
 
 
These models include three types of control variable: market conditions, issuer’s characteristics, 
and bond attributes. I use BBI20 to control for systematic change in the municipal bond market. I 
also use the standard variation of BBI20 in the previous 8 weeks to control for market volatility. 
Other control variables include bond’s issue date; maturity; log of market size; log of total real 
GDP; credit rating and state of issuance, and state and year dummies. Prior to 2014, US census 
bureau only reported GDP on the metropolitan level, therefore, I use the metropolitan to 
approximate county GDP. Crop and property are log of crop damage and property damage 
adjusted to 2018 dollars; injuries and fatalities are per capita data.  
 
Model 3 and 4 include dummy variables that disaggregate the bonds in the sample into two 
groups: those issued before Hurricane Katrina and those issued after. These two models intend to 
test whether Hurricane Katrina changed the market’s view on climate risk. If the coefficients of 
dummy variables are significantly positive, then Hurricane Katrina might have caused a shift on 
investors’ mindsets. 
 
Model 3: 
𝐶𝑜𝑢𝑝𝑜𝑛 = 𝛽0 + 𝛽1ℎ𝑎𝑧𝑎𝑟𝑑𝑠 + 𝛽2𝐺𝑂 + 𝛽3𝑐𝑎𝑙𝑙𝑎𝑏𝑙𝑒 + 𝛽4𝑎𝑓𝑡𝑒𝑟 𝐾𝑎𝑡𝑟𝑖𝑛𝑎
+ 𝛽5ℎ𝑎𝑧𝑎𝑟𝑑𝑠 × 𝑎𝑓𝑡𝑒𝑟 𝐾𝑎𝑡𝑟𝑖𝑛𝑎 + 𝛽6𝑐𝑜𝑛𝑡𝑟𝑜𝑙 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒𝑠 + 𝜀 
Model 4: 
 
𝐶𝑜𝑢𝑝𝑜𝑛 = 𝛽0 + 𝛽1𝑐𝑟𝑜𝑝 + 𝛽2𝑝𝑟𝑜𝑝𝑒𝑟𝑡𝑦 + 𝛽3𝑖𝑛𝑗𝑢𝑟𝑖𝑒𝑠 + 𝛽4𝑓𝑎𝑡𝑎𝑙𝑖𝑡𝑖𝑒𝑠 + 𝛽5𝐺𝑂 + 𝛽6𝑐𝑎𝑙𝑙𝑎𝑏𝑙𝑒
+ 𝛽7𝑎𝑓𝑡𝑒𝑟 𝐾𝑎𝑡𝑟𝑖𝑛𝑎 + 𝛽8𝑐𝑟𝑜𝑝 × 𝑎𝑓𝑡𝑒𝑟 𝐾𝑎𝑡𝑟𝑖𝑛𝑎
+ 𝛽9𝑝𝑟𝑜𝑝𝑒𝑟𝑡𝑦 × 𝑎𝑓𝑡𝑒𝑟 𝐾𝑎𝑡𝑟𝑖𝑛𝑎 + 𝛽10𝑖𝑛𝑗𝑢𝑟𝑖𝑒𝑠 × 𝑎𝑓𝑡𝑒𝑟 𝐾𝑎𝑡𝑟𝑖𝑛𝑎
+ 𝛽11𝑓𝑎𝑡𝑎𝑙𝑖𝑡𝑖𝑒𝑠 × 𝑎𝑓𝑡𝑒𝑟 𝐾𝑎𝑡𝑟𝑖𝑛𝑎 + 𝛽12 × 𝑎𝑓𝑡𝑒𝑟 𝐾𝑎𝑡𝑟𝑖𝑛𝑎
+ 𝛽13𝑐𝑜𝑛𝑡𝑟𝑜𝑙 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒𝑠 + 𝜀 
 
 
Model 5 and 6 estimate the correlation between natural hazards and municipal bond coupon rate 
by year.  
 
Model 5: 
𝐶𝑜𝑢𝑝𝑜𝑛𝑡 = 𝛽0𝑡 + 𝛽1𝑡ℎ𝑎𝑧𝑎𝑟𝑑𝑠𝑡 + 𝛽2𝑡𝐺𝑂𝑡 + 𝛽3𝑡𝑐𝑎𝑙𝑙𝑎𝑏𝑙𝑒𝑡 + 𝛽4𝑡𝑐𝑜𝑛𝑡𝑟𝑜𝑙 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒𝑠𝑡 + 𝜀𝑡 
 
Model 6: 
𝐶𝑜𝑢𝑝𝑜𝑛𝑡 =   𝛽0𝑡 + 𝛽1𝑐𝑟𝑜𝑝𝑡−1 + 𝛽2𝑝𝑟𝑜𝑝𝑒𝑟𝑡𝑦𝑡−1 + 𝛽3𝑖𝑛𝑗𝑢𝑟𝑖𝑒𝑠𝑡−1 + 𝛽4𝑓𝑎𝑡𝑎𝑙𝑖𝑡𝑖𝑒𝑠𝑡−1 + 𝛽5𝐺𝑂𝑡
+ 𝛽6𝑐𝑎𝑙𝑙𝑎𝑏𝑙𝑒𝑡 + 𝛽7𝑐𝑜𝑛𝑡𝑟𝑜𝑙 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒𝑠𝑡 + 𝜀𝑡 
 
The yearly estimation provides an opportunity to examine if major natural disasters would 
impact municipal bond coupon rates in the following years, and whether this influence last. 
  
EMPIRICAL ESTIMATION 
 
In this section, I present my empirical results for the six models. 
 
[Table 4] 
 
Table 4 reports the OLS result of hazard index on municipal bond coupon rates. Column 1-4 
report the results of model 1-4. The coefficient of hazard index is 0.003 but not significant. The 
coefficient of property damage is negative but not significant. The coefficient of injuries is 
positive but not significant. However, the coefficients of crop damage and fatalities are 
significantly negative.  
 
As mentioned earlier, model 3 and 4 are specified similarly to model 1 and 2 but also include 
dummy variables that disaggregate the data set into bonds issued before and after Hurricane 
Katrina, and a variable interacting climate risk with the post-Katrina dummy variable in order to 
allow climate risk to impact bonds issued before and after Hurricane Katrina differentially. In 
column 3, the coefficient of post-Katrina dummy is insignificant, and the coefficient of post-
Katrina×hazard risk is slightly negative and insignificant. In column 4, the coefficient of post-
Katrina×crop damage is significantly negative; the coefficients of post-Katrina×property 
damage, post-Katrina×injuries, post-Katrina×fatalities and post-Katrina dummies are negative 
and insignificant. These results indicate that natural hazard risk did not affect municipal bond 
interest rate differently after Hurricane Katrina. Moreover, Hurricane Katrina seems to increase 
the negative effects of crop damage on municipal bond interest rate. One possible explanation is 
that after Hurricane Katrina, crop insurance coverage increased, and therefore diluting the impact 
of crop damage. Overall, despite the catastrophic effects of Hurricane Katrina, the market did not 
seem to change its perception of natural hazards. 
 
[Table 5] 
 
Till now we observe not significant impact of natural hazards and municipal bond interest rate. 
But if we take a closer look at yearly regressions of model 5 and 6, we could see that the 
significance and magnitude of coefficients of hazard risks vary from year to year. Table 5 reports 
results from estimating yearly regression of model 5. In year 2002, 2006 and 2007, the 
coefficients are significant. In year 2002, the coefficient is 0.101; from year 2006 to 2007, the 
coefficient decreased from 0.062 to 0.058, and the p-value increased from 0.017 to 0.082. This is 
in line with severe natural hazards’ occurrences in the US. 2001 witnessed Tropical Storm 
Allison whose persistent remnants caused severe flooding and significant damage in Texas and 
Louisiana. In 2002, large portions of 30 states, including the western states, the Great Plains, and 
much of the eastern U.S underwent moderated to extreme drought. In 2005, Hurricane Katrina 
caused 170 billion dollar damage in the east coast, and was followed by Hurricane Rita and 
Wilma, each causing 25 and 26 billion dollar damage.  
 
[Table 6] 
 
Table 6 reports the results of model 6. In 2006, the coefficient of property damage is positive and 
statistically significant. In 2003, the coefficient of crop damage is positive and significant. 
 
The results indicate that natural hazards have significant but short impacts on municipal bond 
coupon rates. Even with disasters such as Hurricane Katrina, the impact only lasted for one to 
two years. 
  
CONCLUSION 
 
Hurricane Katrina has brought to the forefront the importance of research in the relationship 
between the municipal bond market and natural disasters. Using county level municipal bond and 
natural hazard data, this paper analyzes the impact of natural disaster risk and disaster damage on 
municipal bond coupon rate. Results show that during 2000-2010, there are no significant 
impacts of natural hazard risk and disaster damage on municipal bond interest rate. Yearly 
regression shows that municipal bond interest rate is only affected by natural hazard risk and 
damage shortly after occurrence of natural disasters. My estimates demonstrate that the market 
seems to have a short memory of natural disasters. 
 
Further study is warranted in order to evaluate the validity of these findings. Additionally, there 
are other possible extensions to this study that would provide additional insight into the 
relationship between natural disaster risk and the municipal bond market through studying the 
effects of natural disasters on municipal bond prices and trading. Finally, limitations of my data 
set prevent me from estimating the indirect impact of natural disaster risk on interest costs 
through insurance and bond ratings. 
 
  
REFERENCE 
 
COLE, C.S., LIU, P. & SMITH, S.D. (1994), The issuer effect on default risk insured municipal 
bond yields. Journal of Economics and Finance, 18: 331. 
 
DENISON, D. (2006), Bond Market Reactions to Hurricane Katrina: An Investigation of Prices 
and Trading Activity of New Orleans Bonds, 27: 39-52.  
 
FOWLES, J., LIU, G. and MAMARIL, C.B. (2009), Accounting for Natural Disasters: The 
Impact of Earthquake Risk on California Municipal Bond Pricing. Public Budgeting & Finance, 
29: 68-83. 
 
HANDLEY, D.M. (2006), Hurricanes on the Alabama Gulf Coast: The Manageable Impacts of 
Ivan and Katrina. Municipal finance Journal, 27: 95-111. 
 
KLOMP, J. (2014), Financial fragility and natural disasters: An empirical analysis. Journal of 
Financial Stability, 13: 180-192. 
 
MARLOWE, J. (2006), Volume, Liquidity, and Investor Risk Perceptions in the Secondary 
Market: Lessons from Katrina, Rita, and Wilma, 27:1-37. 
 
VIGDOR, J. (2008), The Economic Aftermath of Hurricane Katrina. Journal of Economic 
Perspectives 22: 135–154. 
 
YAWITZ, J.B. (1978), Risk Premia on Municipal Bonds. The Journal of Financial and 
Quantitative Analysis, 13: 475-485. 
 
 
 
 
Mean Std Min Max
Coupon 4.710073 1.029575 .5 12
WSLB20 4.379571 .3808107 3.82 5.85
Maturity 16.57375 5.945183 7.33744 48.92813
Market Size 6504580 2.06e+07 3000 5.26e+08
GDP 213.1649 301.6554 1.9890 1303.393
Hazard Index 13.11346 1.929653 8 19
Crop Damage (Million) 24.8 178 0 182
Property Damage (Million) 589.6272 5645.861 0 3210
Injuries (Per million) 1.41 14.3 0 990
Fatalities (Per million) 0.213 1.93 0 0.8
N 8276
Table 1: Summary Statistics
Rating Percentage
AAA 8.3
AA+ 0.4
AA 5.33
AA- 2.31
A+ 0.22
A 78.48
A- 2.42
BBB+ 0.52
BBB 1.38
BBB- 0.53
BB 0.1
B+ 0.02
Table 2: Rating
1
Figure 1: Average coupon rate and BBI20
0 1
Insured 79.0 21.0
GO 48.8 51.2
callable 90.7 9.3
Table 3: Dummy Variables
2
3
Figure 2: US Hazard Index
Model 1 Model 2 Model 3 Model 4
Hazard index 0.003 0.013
(0.01) (0.02)
Crop damage -0.008* -0.004
(0.00) (0.01)
Property damage -0.005 -0.007
(0.00) (0.01)
Injuries 158.114 1788.185*
(244.44) (721.24)
Fatalities -6145.163* -4163.509
(2815.01) (15825.47)
After Katrina -0.125 -0.280*
(0.22) (0.12)
Hazard×After Katrina -0.011
(0.02)
Crop damage×After Katrina -0.007
(0.01)
Property damage×After Katrina 0.003
(0.01)
Injuries×After Katrina -1874.968*
(766.55)
Fatalities×After Katrina 241.973
(16095.46)
BBI20 0.764*** 0.773*** 0.768*** 0.778***
(0.04) (0.04) (0.04) (0.04)
BBI20 standard deviation -0.413* -0.492** -0.430* -0.510**
(0.18) (0.18) (0.18) (0.18)
Original Maturity (Years) 0.050*** 0.051*** 0.050*** 0.051***
(0.00) (0.00) (0.00) (0.00)
Market size 0.089*** 0.090*** 0.088*** 0.089***
(0.01) (0.01) (0.01) (0.01)
Insured -0.074** -0.067* -0.078** -0.070*
(0.03) (0.03) (0.03) (0.03)
General Obligation -0.085*** -0.083*** -0.086*** -0.083***
(0.02) (0.02) (0.02) (0.02)
Callable -0.101** -0.103** -0.100** -0.102**
(0.04) (0.04) (0.04) (0.04)
GDP -0.017* -0.027** -0.016* -0.028**
(0.01) (0.01) (0.01) (0.01)
Table 4: Model 1-4
4
5
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
Hazard index 0.035 0.101 -0.057 -0.004 0.016 0.062* 0.058 -0.065 -0.033 -0.007
(0.07) (0.05) (0.04) (0.05) (0.03) (0.03) (0.03) (0.04) (0.02) (0.01)
BBI20 -1.008 1.205** -0.312 -0.545 2.137*** 0.059 1.377** -0.713*** 0.256 0.828***
(1.17) (0.40) (0.28) (0.34) (0.41) (0.20) (0.42) (0.21) (0.13) (0.04)
BBI20 standard deviation 7.472 -0.620 0.749 0.742 4.433 -1.561 1.066 3.033*** -2.424** -0.226
(4.55) (2.05) (1.78) (2.13) (2.50) (2.11) (1.97) (0.83) (0.78) (0.17)
Original Maturity (Years) 0.053** 0.006 0.090*** 0.010 0.016 0.020** 0.019* 0.039*** 0.073*** 0.062***
(0.02) (0.02) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.00)
Market size -0.074 0.047 -0.017 0.052 0.126*** 0.100*** 0.181*** -0.261*** -0.043 0.152***
(0.07) (0.05) (0.05) (0.04) (0.03) (0.03) (0.03) (0.04) (0.03) (0.01)
Insured -0.030 0.358* -0.184 -0.081 -0.234 -0.541*** -0.018 -0.133 0.166 -0.165***
(0.29) (0.17) (0.15) (0.20) (0.12) (0.11) (0.15) (0.17) (0.10) (0.03)
General Obligation -0.067 -0.087 -0.532** 0.006 -0.117 -0.658*** -0.213 0.311 -0.110 -0.070**
(0.40) (0.23) (0.20) (0.24) (0.14) (0.11) (0.14) (0.16) (0.10) (0.03)
Callable 0.108 0.463* 0.229 0.560* -0.093 -0.207 0.141 -0.475* 0.354 -0.063
(0.30) (0.21) (0.18) (0.22) (0.15) (0.11) (0.15) (0.19) (0.21) (0.05)
GDP -0.265* 0.214** 0.091 0.007 0.001 0.056 0.069 0.115* -0.090** -0.039***
(0.10) (0.07) (0.08) (0.06) (0.04) (0.03) (0.05) (0.06) (0.03) (0.01)
Table 5: Model 5
6
2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
Crop damage -0.020 -0.029 0.039 0.008 -0.027 0.012 -0.078*** 0.005 -0.010 0.000
(0.02) (0.02) (0.02) (0.03) (0.02) (0.01) (0.02) (0.02) (0.01) (.)
Property damage 0.004 -0.007 -0.003 -0.028 0.005 0.027** 0.003 0.010 0.006 0.000
(0.03) (0.03) (0.02) (0.01) (0.01) (0.01) (0.01) (0.02) (0.01) (.)
Injuries -7254.291 7285.842 798.291 13348.021 3156.391 356.154 -2839.489 1486.581 1517.835 0.000
(10874.71) (5530.16) (632.51) (12588.94) (2657.15) (14907.83) (2273.96) (3445.23) (795.05) (.)
Fatalities 28311.568 126386.153 - 51020.104 -55620.830 10526.192 -21574.592 4937.507 - 0.000
47220.481* 24019.738*
(55860.43) (108933.06) (18893.66) (55120.21) (32047.19) (13851.85) (20040.14) (42203.59) (11191.08) (.)
BBI20 -1.314 1.591*** -0.370 -0.454 1.858*** -0.162 1.178** -0.721** 0.290* 0.828***
(1.14) (0.44) (0.26) (0.34) (0.44) (0.22) (0.44) (0.22) (0.13) (0.04)
BBI20 standard deviation 8.742 -1.258 0.515 0.756 6.486* -3.047 1.488 2.584** -2.457** -0.221
(4.62) (2.20) (1.68) (2.05) (2.83) (2.23) (2.15) (0.88) (0.78) (0.17)
Original Maturity (Years) 0.046* 0.008 0.097*** 0.022 0.011 0.021** 0.012 0.037** 0.071*** 0.061***
(0.02) (0.02) (0.02) (0.01) (0.01) (0.01) (0.01) (0.01) (0.01) (0.00)
Market size -0.094 0.059 -0.047 0.038 0.103** 0.092** 0.216*** -0.266*** -0.047 0.152***
(0.07) (0.05) (0.05) (0.03) (0.03) (0.03) (0.04) (0.05) (0.03) (0.01)
Insured 0.190 0.370 -0.302 -0.097 -0.026 -0.592*** -0.026 -0.199 0.160 -0.170***
(0.31) (0.19) (0.15) (0.19) (0.16) (0.11) (0.15) (0.18) (0.10) (0.03)
General Obligation -0.031 -0.086 -0.963*** 0.043 -0.440* -0.818*** -0.235 0.314 -0.067 -0.072**
(0.37) (0.26) (0.21) (0.24) (0.17) (0.12) (0.15) (0.17) (0.10) (0.03)
Callable 0.221 0.612* -0.043 0.420 -0.110 -0.206 -0.078 -0.394 0.447* -0.062
(0.30) (0.23) (0.17) (0.22) (0.16) (0.12) (0.16) (0.21) (0.21) (0.04)
GDP -0.309** 0.216* 0.079 -0.062 -0.040 0.097* 0.039 0.130* -0.095** -0.038***
(0.11) (0.08) (0.09) (0.07) (0.05) (0.04) (0.06) (0.07) (0.04) (0.01)
Table 6: Model 6