Johnson EG2014 Securitizedcatriskandclimatechange
Johnson EG2014 Securitizedcatriskandclimatechange
Johnson EG2014 Securitizedcatriskandclimatechange
net/publication/260410900
CITATIONS READS
45 1,164
1 author:
Leigh Johnson
University of Oregon
23 PUBLICATIONS 1,098 CITATIONS
SEE PROFILE
All content following this page was uploaded by Leigh Johnson on 25 October 2018.
ECONOMIC GEOGRAPHY
Risk and the Implications of
Climate Change
abstract
Leigh Johnson This article analyzes the drivers and implications of
Department of Geography catastrophe bonds’ growing popularity as an alterna-
Winterthurerstrasse 190 tive asset class. As investor demand for bonds out-
University of Zürich paces their supply from reinsurers, the study asks
8057 Zürich how the place-based physical vulnerabilities of fixed
Switzerland capital have been rendered into assets deemed
[email protected]
increasingly desirable by growing blocks of financial
capital. Combining data from extended interviews
with industry datasets and market reports, the study
demonstrates how this securitization pathway allows
mobile capital on a search for yield to reframe spatial 155
Key words: liabilities as tradable assets, thus accessing new
risk transfer “returns on place.” By aggregating and analyzing
insurance data on approximately $37 billion in catastrophe
climate change
1
In its generic sense, securitization simply refers to the technique
of pooling the contractual liabilities held by a financial
Vol. 90 No. 2 2014
uneven results of the market’s capacity to fashion geographic liabilities into strategic
resources, analyzing the development of the nearly $40 billion in natural catastrophe
bonds issued from 1997 through 2011 and characterizing the highly particular geo-
graphic footprint of this securitization pathway. Finally it argues that given the competi-
tive search for yield in insurance-linked investing, securitizing the geophysical effects of
climate change could generate new unexamined risks and maladaptive economic
responses.
institution and selling them onward as a packaged financial product. Investors purchase a security based on
expectations of future income streams. In exchange, they contractually assume the liabilities associated
with that income stream, which are transferred off the original institution’s balance sheet.
2
Insurers purchase reinsurance to expand their own underwriting capacity and prevent massive payments of
claims from sending them into bankruptcy. Globally, roughly 10 percent of the non-life primary insurance
business (including property of all sorts, liability, and business interruption) is ceded to reinsurers (Group
of Thirty 2006, 18). In 2009, total global catastrophe reinsurance exposures were U.S. $205 billion (Swiss
Reinsurance 2009, 41; IAIS 2009).
3
Where applicable, this text uses the standard industry notation of “(re)insurance” to denote both the
reinsurance and insurance industries in cases in which their interests and structural positions in the
risk-transfer market are similar. The term also indicates that, in practice, it is not always possible to
distinguish clearly between the activities of reinsurers and insurers, since many large firms regularly
perform both roles (and may act as retrocessionaires selling reinsurance to other reinsurers). Examples
include Lloyd’s syndicates, Allianz, Berkshire Hathaway, and Swiss Reinsurance.
ECONOMIC GEOGRAPHY
part or all their principal. If not, they collect quarterly payments as well as a return of
their principal at the bond’s maturation date.
The most frequently securitized perils are the natural catastrophes with the greatest
potential to destroy insured property and generate associated business interruption and
workers’ compensation claims: hurricanes, earthquakes, and winter storms. Bond issu-
ances are always tied to specific perils in at least one geographic region (for example,
“California quake,” covering earthquakes in that state, or “European wind,” covering the
fierce extratropical cyclones that bear into Western Europe from the North Atlantic in the
winter months). Many are structured such that different tranches of the same bond cover
the sponsor for losses arising from separate perils in different geographic regions. A
growing number are explicitly multiperil across all tranches, meaning that the same
dollar of principal is exposed to two or more identified risks and locations. A bond’s
payout trigger may be based on the specific losses suffered by the sponsor, an estimate of
industry-wide losses, a modeled loss using event parameters fed into a catastrophe
model, or an indexed measure of event parameters such as average wind speed or ground
shaking (Hagedorn, Heigl, Mueller, and Seidler et al. 2009).
158 What is somewhat extraordinary in the risk-averse post-financial crisis climate is that
the cat bond market has experienced a growing and often unmet demand from the “buy
side.” A major portfolio manager complained in 2010 that overcapitalized reinsurers
were not passing enough risk onward to the capital markets: “Right now the market is
struggling with lack of issuance, not lack of investors” (Patel, quoted in Aon Benfield
2010, 33). In 2010 and again in 2011, Swiss private bank Clariden Leu temporarily
closed its popular ILS fund to new investors, blaming the insufficient issuance of new cat
bonds. The variety of investors accessing the market has also grown, with 2011 marking
the first year that institutional investors (including pension funds, sovereign wealth funds,
and foundations) outnumbered dedicated catastrophe funds (Aon Benfield 2011). This
overabundance of investors raises a question of broader significance for both the geog-
raphies of finance and the politics of environmental risk: how and why has investors’
interest in catastrophic risk remained so buoyant? This question is particularly puzzling
given that, first, the property insurance and reinsurance industries have typically been
bugbears for traditional equity investors, and, second, as reinsurers regularly warn,
climate change is generating greater uncertainty about the intensity, frequency, and
spatial distribution of extreme events.
The study presented in this article pursued three research questions to examine these
apparent contradictions: (1) As insurance exposures are rendered into “tradable poten-
tialities,” how and from where do these securities circulate, and what has been their
geographic development over time? (2) Why have these securities become an attractive
alternative asset class that is deemed increasingly investment-worthy by a growing
number of mainstream institutional investors, and what are the sociospatial ramifications
of this popularity? (3) Why do ILS market actors consistently frame potential changes in
extreme event patterns in terms of “opportunity,” and what are the implications of
securitization for risk management in areas that are increasingly vulnerable in a changing
climate?
The article proceeds stepwise to answer these questions. The remainder of the first
section reviews the literature on ILS within geography and cognate fields, suggests some
avenues of engagement with the geographies of finance, and introduces the study’s
methodological approach to researching capital markets. The second section reviews the
institutions, actors, and transactions that constitute the market, followed by an empirical
analysis of 15 years of cat bond issuance revealing patterns of geographic concentration
and recombination of so-called peak perils. The third section considers the specific
Vol. 90 No. 2 2014
4
A measure of underwriting profitability derived from the ratio of claims payments and expenses to total
premiums collected. Combined ratios for 2006 and 2007 were 92.4 and 95.6, respectively.
ECONOMIC GEOGRAPHY
they are by far the most visible products. Other securities, such as those linked to life or
automobile insurance policies, have been sold onward and a diversified ILS portfolio
often also includes non-catastrophic risks as well as catastrophe futures and options.5
Another popular subclass are privately traded swap contracts known as “industry loss
warranties,” whose payout depends on net industry losses exceeding a specified value.
Despite their rapid growth, relatively limited critical scholarly attention has been paid
to insurance-linked securities (Bougen 2003; Jagers, Paterson, and Stripple 2005;
Randalls 2009; Sturm and Oh 2010). Most of this work has focused on insurers’ and
reinsurers’ efforts to displace risk into capital markets, framing them within ongoing
discussions of Ulrich Beck’s (1992, 1999) famous postindustrial “risk society” in which
modernity’s technological developments become the source of catastrophic risks that
exceed the boundaries of calculability and control. As evidence of the “risk society,”
Beck has pointed repeatedly to the private insurance industry’s (alleged) refusal to write
coverage for risks from nuclear power plants, genetic engineering, and climate change,
among others. A significant debate has followed over the theoretical and empirical
validity of Beck’s uninsurability thesis, highlighting numerous techniques by which new
160 and seemingly “incalculable” risks are being rendered insurable and governable (Ericson
and Doyle 2004b; O’Malley 2003; Collier 2008). It is within this context that Bougen
(2003, 255) has asked whether catastrophe bonds may signal the emergence of a risk
network with “the necessary degrees of coherence between institutions and technologies
. . . for the practical functioning of catastrophe financing.” Sturm and Oh (2010) have
taken an explicitly geographic perspective, arguing that catastrophe bonds allow rein-
surers to solve the problem of ever-growing catastrophic losses through “scaled and
networked recovery schemes” that displace and spread risk. Similarly, Jagers et al.
(2005) have suggested that insurance-linked securities are strategic forms of privatized
environmental governance that are “becoming a dominant response” and “exit option”
for insurers who are seeking to manage the risks of climate change, allowing the industry
to displace the costs of climate impacts into the broader capital markets.6
These explanations are often accompanied by portrayals of capital markets as
bequeathing financial powers that free the insurance industry from its fatal ties to
absolute space. When the materialities of nature or space are considered, it is to highlight
them as problems, with all their messy specificity. The cat bond market is theorized as a
“way out” and “scaling up” for the insurance industry—a solution, albeit incomplete, to
the well-documented limitations of scale, capacity, and capital reserves within reinsur-
ance markets. Although these formulations usefully explain insurers’ and reinsurers’
motivations for seeking alternate sources of capital for underwriting, they do not offer a
convincing explanation of why investors would willingly buy into the market. This is
arguably because such accounts neglect the extent to which the (re)insurance industry’s
5
Compared to weather derivative markets, public trading in catastrophe derivative products has been staid.
Futures and options contracts that are indexed to measures of the intensity of individual storms are
currently traded on the Chicago Mercantile Exchange, but total volumes are quite small and parties
wishing to make large trades are often thwarted by the absence of market makers. Futures contracts based
on industry-wide losses to hurricanes (“IFEX” contracts that traded on the Chicago Climate Futures
Exchange from 2008 to 2012) are the newest addition to a long list of failed catastrophe derivatives and
options (Aon 2008; Csiszar 2007).
6
This conclusion is arguably overstated at present; at the market’s highest point in 2007, the $14 billion
of capital invested in natural catastrophe bonds represented less than 30 percent of global catastrophe
reinsurance limits; of this amount, only about 60 percent of bonds underwrote perils that were
weather-related.
Vol. 90 No. 2 2014
Wall Street, the City of London, and offshore financial centers; and the amplification of
mimetic forms of rationality by the financial media (French, Leyshon, and Thrift 2009).
Together, this body of work has begun to demonstrate how the practices of hedging,
arbitrage, and the maximization of yields produced new risks and volatility in the course
of everyday operations that went on to ricochet through the global economy as the credit
crisis broke (Engelen and Faulconbridge 2009). Given that the popularity of cat bonds
has only grown following the financial crisis, it is appropriate to ask whether this market
harbors its own unexamined sources of risk and volatility.
Finally, economic geographers have begun to reconceptualize the spatial characteris-
tics and footprint of finance, calling for a finer-grained geographic account of the
processes that constitute financialization as such (Christophers 2011, Engelen 2012).
Contra the analytical units of the regulation school (Boyer 2000), they have suggested the
inadequacy of conceptualizing “financialization through the geographical prism of
national economic space” (French, Leyshon, and Wainwright 2012, 9; see also
Christophers 2011). The growth of the catastrophe bond market provides an opportunity
to trace how geographically exceptional parts of the built environment are selectively and
162 opportunistically accessed by financial capital, such that risks and profits surpass the
territorial logic of nation states.
$350
Order in which tran-
ches are exhausted
Class A notes
$200 Class B notes
1,200
1,000
Securitized Through Cat Bond issuance
Losses in Millions (U.S. $)
retains
Insurerretains
800
Reinsurer C Reinsurer D
DirectInsurer
600
Reinsurer B
Direct
400
Reinsurer A
200
Direct Insurer
0
0 25 50 75 100
Percentage of Losses Held Within Each Layer
Figure 1. Hypothetical reinsurance program and catastrophe bond structure for a large insurer’s
natural catastrophe reinsurance coverage.
Vol. 90 No. 2 2014
Insured Policy
Holders
Insurance Company
Catastrophe
“Cedant”
Modeling Firm
(fee
s) Rating Agency
Collateral Special
often deposited in Treasury Purpose Deal Manager
money market funds,
or with Total Return Swap
Vehicle
counterparty (as
is n sumin
ot t g b
rig o
ger nd
ed )
-Hedge funds Cat Bond
Initial funds exchanged
Regular payments
Contingent payments
-Dedicated cat funds
-Reinsurers
-Institutional investors
(pension funds, foundations,
} Investors
165
handling practices, reinsurers typically mandate that the primary insurer remain respon-
sible for some significant proportion of losses above the reinsurance attachment point.
Here the primary insurer remains liable for 10 percent of any losses between $200
million and $1.15 billion and all losses above that amount.
In this hypothetical case, the primary insurer has sponsored a $350 million cat bond in
three tranches that substitutes for a top layer of reinsurance coverage and is triggered only
after losses exceed $800 million. In the existing market, roughly half of all bond
issuances come from primaries. State Farm and U.S. military insurer USAA are currently
the most frequent sponsors, with Zurich Financial, Allianz, Chubb and AXA also acting
as repeat players. In the case of bond issuance by reinsurers, the tranches and trigger
structure remain similar. Global reinsurance giants Swiss Reinsurance and Munich
Reinsurance account for the largest share of these issues, with French firm SCOR also
sponsoring consistently.
If the bond is triggered, the capital to compensate the sponsor for losses comes first
from the most “junior” tranche, in this case the Class C notes. As losses mount, the more
senior notes are tapped. Some senior tranches guarantee to return part or all of investors’
principal; such “protected” tranches may be rated as investment grade by ratings agen-
cies. The unprotected tranches offer higher rates of return, but feature no such guarantees
on the principal.
A relatively standard institutional and legal structure has been developed to issue cat
bonds, schematized in Figure 2. The sponsoring (re)insurer—the “cedant” of the risk—
creates an independent and legally distinct special purpose vehicle (SPV) that is domi-
ciled offshore, most often in the Cayman Islands.7 The SPV acts as the functional
equivalent of a reinsurer to the sponsor, who pays it an initial premium that is deposited
7
The Caymans gained popularity as a domicile for the SPVs of collateralized debt obligations because there
is no withholding tax on coupon payments to investors from the bond’s collateral account.
ECONOMIC GEOGRAPHY
in a trust. In exchange, the SPV issues notes to investors, whose principal is also
deposited in the same trust. The trust safeguards the bond’s collateral, often investing the
collateral in a money market fund of U.S. Treasury bonds.
The cedant also pays significant fees to third parties, including (1) a deal manager to
oversee the marketing of the bond, (2) a catastrophe modeling firm to model expected
losses on the portfolio, and (3) a rating agency to produce an external evaluation of the
bond’s credit-worthiness. Analysis by a catastrophe model, which joins thousands of
stochastic event sets with engineering damage modules and highly granular data on the
underlying geographies of exposure in the built environment, is an absolute prerequisite
for the development of pricing guidelines based on expected losses (Johnson 2013). Cat
model output is also a major factor considered by bond rating agencies—which do not
themselves perform any catastrophe modeling—when assigning ratings to catastrophe
bonds (cf. Standard & Poor’s 2009).8
Given the scale of risks, the technicality of the offerings, and the expertise required,
most investors in ILS are quite specialized. From 2008 to 2011, 35 percent to 45 percent
of the total investments in cat bonds were made by roughly 25 dedicated “cat funds”
166 —privately managed funds that were launched in the years following September 11 and
Hurricane Katrina when there were large returns on capital. These are essentially single-
strategy hedge funds that invest only in insurance-linked securities. Quite a few of these
dedicated funds are either direct spin-offs from reinsurance or insurance companies or
were launched with seed money from insurers who retain some financial interest. The
segmentation of the remaining investor base varies dramatically depending on the con-
ditions within the larger capital market and the point in the reinsurance pricing cycle.
Aon Benfield (2011) reported the following breakdown for deals it managed in 2011:
institutional investors, 44 percent; dedicated catastrophe funds, 34 percent; mutual funds,
10 percent; reinsurers, 7 percent; and multistrategy hedge funds, 5 percent. Notably, 2011
marked the first year that the institutional investors, including public-sector pension
funds, sovereign wealth funds, and foundations, surpassed dedicated catastrophe funds as
the largest block of investors, up from 31 percent the previous year. There is also a
distinct geographic signature of the demographics of cat bond investors; again for 2011
Aon Benfield deals, 47 percent of investors came from the United States, 33 percent from
Switzerland, 7 percent from Bermuda, 5 percent from the United Kingdom, and 8 percent
from elsewhere.
In the 15-year history of the catastrophe bond market, the direct connection between the
financial capital invested in cat bonds and the at-risk landscapes they underwrite has rarely
been evidenced through bond defaults. Although (re)insurers handle millions of claims
annually, by the end of 2011 only a handful of cat bonds out of nearly 200 issued had been
triggered by insurance losses. One of the most notable defaults to date was Kamp Re,
sponsored by the primary insurer Zurich Financial Services to cover U.S. hurricane and
earthquake claims if the firm’s total losses exceeded $1 billion in a three-year period. The
deal closed on July 28, 2005, just one month before Hurricane Katrina began to wreak
havoc on the Gulf Coast. By October, Kamp Re notes were trading at pennies on the dollar
in the expectation of a total loss (Lane and Beckwith 2010). The administration of
outstanding claims finally concluded at the end of 2010, and Kamp Re returned roughly
8
Other components of this research explored the tremendous influence of catastrophe modeling and its
associated epistemic practices within the (re)insurance industry as a whole (Johnson 2011). I do not detail
these arguments here except to note the importance of models as calculative devices (Lepinay 2007;
Muniesa, Millo, and Callon 2007) that create an operable equivalence between perils of different onto-
logical orders and allow for the commodification of insurers’ contingent exposures.
Vol. 90 No. 2 2014
$0.25 on every dollar of principal to investors (Trading Risk 2011a). Hurricane Ike’s
colossal damage to offshore energy assets in 2008 also looked likely to trigger a loss on
junior notes covering Glacier Reinsurance for U.S. wind and quake, but investors disputed
the legitimacy of Glacier’s proof of loss in arbitration and eventually recouped the entirety
of the principal (Trading Risk 2013a). Of the six bonds with any exposure to the Japanese
megaquake and tsunami of 2011, only two triggered. The largest, Munich Re’s $300
million Muteki cat bond structured for Japanese insurer Zenkyoren, was a complete loss
for investors (AIR Worldwide 2011). The unprecedented losses from U.S. thunderstorms
and tornados in the summer of 2011 also triggered a total loss to American Family Mutual
Insurance Company’s $100 million Mariah Re cat bond, but investors have filed suit to
reclaim the capital, claiming that the company falsified its loss report (Trading Risk
2013b). Hurricane Sandy also precipitated smaller losses on several Swiss Re bonds that
have yet to mature.
These events have not discouraged investors’ interest in the high-paying risks,
however, and in industry circles many consider it fortunate that cat bonds were finally
hit with some high-profile losses. They suggest that the orderly default of bonds and
fulfillment of obligations to the original cedants would demonstrate the reliability of 167
cat bond protection to both risk cedants and insurance regulators, some of whom are
as yet unwilling to treat cat bond facilities as the equivalent of reinsurance when
6
Billions U.S. $
0
1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011
Figure 3. Data provided by Guy Carpenter Securities. Following standard industry reporting
168 practices, monetary values have not been adjusted for inflation.
show later, this dynamic is a crucial reason why ILS market boosters frame climate
change as a major opportunity for the future expansion of the market.
9
Note that this article includes only data on natural catastrophe securitizations linked to property insurance;
thus, the numbers presented here may be lower than accounts that include extreme mortality securitizations
of life insurance portfolios. For an analysis of how life and non-life securitizations are made fungible, see
Johnson (2013); for an exhaustive account of the vast array of insurance-linked securitizations, see Barrieu
and Albertini (2009).
Vol. 90 No. 2 2014
climate change and the season’s “never-ending loss records” (Munich Reinsurance 2006,
17). It was the most economically destructive season in history, yielding roughly $170
billion in total damages in the U.S. and neighboring countries, $69 billion of which were
insured (Swiss Reinsurance 2006). The result was a dramatic draw down of reinsurance
capital reserves and 76 percent higher average rates for U.S. property catastrophe rein-
surance during annual renewals in 2006 (Guy Carpenter 2006, 16).
The effects on the cat bond market were immediate and dramatic. In 2006, (re)insurers
sponsored a total of 28 bonds in an effort to access capital more cheaply than could be
had in traditional reinsurance and retrocession markets. The $4.7 billion of notes issued
in 2006 found an eager audience of investors hoping to take advantage of the high
post-disaster returns. These investors included multistrategy hedge funds, dedicated
catastrophe hedge funds, and even some institutional investors. The dramatic pile-on
continued in the following year with 27 bonds and the largest volume of issuance ($7
billion) in history. “Taking the development from 2004 to 2007 and using this as a basis
for further projections, some people had already forecast that capital markets would take
over traditional reinsurance” (Wallin 2010 n.p.).
The unfolding credit crisis in 2007 purportedly made catastrophe bonds appear more 169
attractive as an uncorrelated alternative asset class. In the spring of 2008, market boosters
were still giddily optimistic about the growth potential of ILS. Despite—and in fact
10
Doing so jeopardized about $335 million of reinsurance coverage for U.S. hurricanes purchased by Allstate
and Munich Reinsurance, $100 million for California earthquake coverage purchased by Aspen Reinsur-
ance, and $225 million for U.S. earthquake and hurricanes purchased by Catlin Reinsurance (Trading Risk
2010b). All four bonds partially defaulted; losses totalled $116 million. (Trading Risk 2011b).
ECONOMIC GEOGRAPHY
80%
Other
60%
Japan typhoon
Japan quake
Europe wind
U.S. hurricane
40%
U.S. quake
20%
0%
1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011
Figure 4. Data provided by Guy Carpenter Securities. In Figure 4a, “bonds on risk” refers to all
bonds susceptible to losses in a given year; that is to say, they are between issuance and
maturation. Note that it is industry practice to tally the total principal of bonds covering two or
more perils (in which the entire principal is exposed to both perils) in all of the peril categories
they cover; therefore, the sum for each year exceeds the notional value of bonds on risk. Refer
to Table 1 for exact breakdowns of such recombinant issuances.
ance and reinsurance coverage. Because of this place-specific premium, cat bonds
covering peak perils generally offer investors the largest relative return per unit of risk.
Large institutional investors and multistrategy hedge funds are especially apt to focus on
the peak perils with the highest returns, since their assets are typically already widely
diversified. Contrary to some early suggestions that cat bonds could facilitate the provi-
sion of affordable coverage in regions outside reinsurers’ traditional markets, the years
following the 2008 crisis have seen a reconcentration of capital into peak peril regions.
Table 1 documents the sequence in which various regions and perils have been secu-
ritized through cat bond structures. Based on a data set of 197 publicly announced natural
catastrophe bond issuances from 1997 through 2011, the table provides an annual
breakdown of the region and monetary values of securitized perils. The “peril combina-
tion” categories listed in the table were developed by coding each tranche of every
issuance according to the perils and geographic locations that were securitized.11 The
progression of issuance detailed in the table make several dynamics apparent. The first,
which was noted earlier, is the continuing concentration of issuance in U.S. peak perils.
From 1997 through 2011, the total volume of bonds issued to cover stand-alone U.S.
hurricane risk ($7.9 billion) was by far the largest of any peril; it was followed by U.S. 171
hurricane plus quake ($4.7 billion); U.S. multiperil ($4.5 billion); California quake
($3.75 billion); and European windstorm ($3.75 billion).
11
This coding generated 33 initial peril combinations, which were reduced to 23 using the following
conditions: peril combinations’ cumulative issuance had to be greater than $350 million, and each peril
combination had to be sold in at least three different years. Combinations that did not meet these
conditions were recategorized as “other multiperil,” “other quake,” or “other hurricane.” To portray recent
trends in recombination, new peril combinations developed in 2008 or later were included, regardless of
their cumulative volumes, provided that they had been issued at least twice.
172
Table 1
Annual Volume of Catastrophe Bond Issuance by Location and Peril Covered
Total for Peril Combination
Peril Combination 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 (Millions U.S. $)
US Ha 395 450 280 417 150 219 58 180 0 1322 1828 370 950 881 417 7914
Cal Qa 112 0 0 150 265 474 386 323 0 698 619 0 375 0 350 3751
Jpn Qa 90 0 200 0 0 126 215 125 0 203 260 300 0 0 0 1519
US multia 30 90 46 0 0 0 0 0 176 228 1759 550 0 705 875 4458
Euro W 6 0 160 0 0 76 276 35 217 515 579 378 212 618 677 3748
Fla Ha 126 0 90 0 0 0 0 0 0 130 210 150 0 0 706
US H + Q 100 182 0 120 0 160 295 205 68 0 318 1016 1280 985 4728
Jpn Ha 80 0 0 0 0 0 0 0 200 120 0 0 0 0 400
US Q 100 0 0 67 59 0 300 490 138 0 14 350 0 1518
Euro W + Jpn Q 17 0 0 0 0 0 0 157 235 0 108 99 0 616
Other multi 483 150 0 278 0 405 100 465 0 68 170 0 2118
US H + Q + Euro W 162 0 0 0 200 30 75 0 0 0 150 617
US H + Euro Wa 120 0 0 0 100 0 0 0 0 150 250 620
US H + Q + Euro W + Jpn Q 258 8 100 63 0 200 200 0 0 100 929
US H + Cal Q + Euro W + Jpn Q 190 86 0 263 0 0 0 0 0 539
Other Q 100 0 75 0 125 0 0 0 0 300
US multi + Euro multi 250 0 125 0 0 0 0 375
ECONOMIC GEOGRAPHY
fragmentation and recombination of risks from peak peril regions in a ceaseless search
for new risk-return profiles. The fragmentation and recombination of peak perils to create
hybrid multiperil, multiregion securities underscores that financialization is a selective
process that ties specific, often far-flung, places and geophysical events into circuits
through which finance capital can move, rather than “rolling out” evenly across econo-
mies and physical landscapes.
Media reports often attribute the majority of investors’ interest in cat bonds to the
purposive search for an uncorrelated asset class, particularly following the credit crunch
and financial crisis. From the simple perspective of returns, it was more or less the case
that cat bonds passed what one modeler called “the mother of all correlation tests,” the
financial meltdown of 2008. This is not to say that cat bond returns did not suffer, but they
weathered the financial storm in the fall of 2008 with higher values and less volatility
than either the S&P 500 or index portfolios of high-yield bonds. The relative success of
the catastrophe bond market in comparison to other equities is one of the primary reasons
why the ILS market continues to attract interest from “real money” (i.e., unleveraged)
funds, such as pension and sovereign wealth funds.
Nevertheless, the ILS market also demonstrates the impossibility of an entirely zero
beta financial product. Some correlation with the wider market is unavoidable due both to
ILS’s position as one of many alternative asset classes that share a common infrastructure,
and to the market’s shifting returns based on interest rates and (re)insurers’ cost of capital.
Prior to the collapse of Lehman Brothers, the ILS sector showed a 4 percent to 8 percent
correlation with the S&P 500; this correlation rose to 30 percent in the months following
174 September 2008 (Insurance Insider 2009). Even if one considers only a bond’s loss
exposure independent from its financial structuring, wider economic conditions may
influence human behaviors that drive total losses. These behaviors include the propensity
to file claims, exaggerate claims, litigate, or commit insurance fraud. An ILS consulting
firm thus stated in a retrospective analysis of the correlation around the financial crisis:
“going forward the correlation is not going to be as low as the 4 percent prior to Lehman,
ever. We anticipate [it] will settle . . . in the mid-teens” (Lane 2009, 16).
Despite these inevitable correlations, the ILS market’s credentials as a relative diver-
sifier multiply due to the second advantage of scalability. Among one of the common
drawbacks of many alternative asset classes are the relatively small size of individual
offerings and the limited total capacity for investment. This drawback makes it difficult
for institutional investors who are trying to invest large volumes of funds; high transac-
tion fees and the piecemeal nature of deals can scuttle otherwise attractive alternative
investments. “The beauty of cat bonds,” the manager of a large fund explained, “is that
you can put $500 million in!” In 2010, for example, State Farm issued a $250 million
U.S. multiperil cat bond, the entirety of which was purchased by a single multiasset
manager, thought to be the Ontario Teachers Pension Plan (Trading Risk 2010a).
Finally, insurance-linked securities allow investors to engage with place-based risks
selectively and opportunistically, rather than invest in either the insurance or reinsurance
industries as a whole. This approach avoids the recalcitrant problems of (re)insurers’
capital structure and the material encumbrances of their sales networks. In the traditional
market for catastrophe coverage, reinsurers typically expend resources to maintain offices,
sales personnel, and long-term underwriting relationships with insurers. Such ties also
prevent reinsurers from abandoning relatively unrewarding risks. One multiasset invest-
ment manager bluntly summarized: “I’ll probably always have a heavy tilt towards Florida
wind because that pays the most. Reinsurers, on the other hand, continue writing business
like Australia wind that pays them 1 percent or something like that. Those are stupid risks
at shitty returns. Why make fixed capital investments when I don’t have to? Reinsurers in
some sense do have to in order to maintain offices, clients, networks. But we don’t have to.”
By comparison, because cat bond contracts are written for particular regions and
perils, they allow investors to select risks that are based on geography and relative
return—a much more enticing prospect than investing the same capital in a reinsurer that
is bound to underwrite less profitable risks. The same manager articulated this rationale:
“Where the reinsurance market is broken is in its capital deployment abilities. It’s crazy
Vol. 90 No. 2 2014
to keep betting the same odds if the price you’re getting paid for it is going up and down.
How much you bet on heads should change based on how much you’re paid for it in a
cyclical market.”
Commentators and fund managers often use the metaphor of “accordion capital” to
describe the injection of funds from the capital markets into ILS when reinsurance capital
is scarce and rates are high and the subsequent removal of these funds when returns fall
below those available in other sectors. The influx of this capital is sometimes credited for
the dampening of reinsurance price cycle amplitudes beginning in the early 2000s,
visible in Figure 5 in the reduced spread between the peaks and troughs of pricing cycles.
The opportunistic dynamics of such accordion capital have profound implications for
how and with what effects the capital markets “access” climate change risks. This is the
focus of the next section.
Some preliminary conclusions about the dynamics of catastrophe risk securitization
and its position within the political economy of the (re)insurance industry can be made.
Despite the rhetorical formulation of the “convergence” of insurance and capital markets,
transferring risks from one market to the other is useful only insofar as the two have not
entirely converged. Rather than convergence, this relationship may be more precisely 175
called “interdigitation.” Such interdigitation allows strategic access to risks and their
recombinations, rather than undifferentiated mixing. Currently securitization is used
Figure 5. The “rate on line” refers to the relative price the buyer pays for a given amount of
coverage. Because the expected losses–and thus risk prices–for different perils and regions can
be vastly divergent, industry reports typically use an index to chart relative price movements
instead of calculating a single market-wide average rate on line. Source: Guy Carpenter &
Company (2011).
ECONOMIC GEOGRAPHY
risk by tying it to absolute space and geophysical nature, while on the other hand, they
facilitate a strategic avoidance of the material encumbrances and associated costs of less
profitable places and business relationships maintained by (re)insurers. This cleaving
allows investors to access place-based premiums and risks—returns on place—without
requiring them to take on the fixed costs required to originate those premiums. The
remainder of the article follows the principle of maximizing “returns on place” to its
logical conclusion and suggests that efforts to write catastrophe bonds to capitalize on
growing climate change risks are a cause for concern.
13
An exhaustive review of current studies on tropical cyclones and climate change written by 10 leading
scholars in the field concluded that “it remains uncertain whether past changes in tropical cyclone activity
Vol. 90 No. 2 2014
increasing climate-related losses in the future will eventually result in higher rates (cf.
ABI 2004; Munich Reinsurance 2006; Geneva Association 2009; Schwierz et al. 2010,
among many others; see Johnson 2010).
The potential for higher premiums and the promise of a greater demand for insurance
in a changing climate keep ILS boosters optimistic about the prospects for market
growth. These possibilities constitute ILS capital’s “ecological fix” (Bakker 2009),
which was reflected in the epigraph with which this article began, a comment made by
the managing principal of the prominent cat fund Nephila Capital: “climate change could
be a major driver for growth in the insurance and reinsurance industry, but only if prices
change to reflect the changing risk. . . . Over the long term, it presents more of an
opportunity than a threat” (Schauble, quoted in Kent 2010, n.p.). Higher (re)insurance
rates on line would drive even greater competition from alternative sources of risk
financing that are able to provide equivalent coverage for lower—yet still very
remunerative—rates. As was demonstrated in the previous section, price undercutting
from the capital markets is possible because of the high fixed costs built into the
(re)insurance business.
Thus the cat bond market’s exposure to changes in extreme events has perhaps 177
ironically become another item in a growing list of financial qualities that recommend it.
Hedge fund manager John Seo enumerated the strategic attraction of cat bonds to the
have exceeded the variability expected from natural causes” and that “detection of an intensity change of
a magnitude consistent with model[ed climate] projections should be very unlikely at this time” (Knutson
et al. 2010, 157, 160; see also Emanuel, Sundararajan, and Williams 2008).
ECONOMIC GEOGRAPHY
bailouts for banks that were deemed too big to fail amounted to a massive transfer of
wealth from households and governments to financial capital.
Although retrospective analyses of the financial crisis have the benefit of hindsight, the
present research was conducted midstream as discussions of “writing to securitize” were
unfolding. Thus, it is unclear how insurance-linked securitization of climate risks will
proceed. This should become a priority for future research. At this point, I can only make
inferences on the basis of the market dynamics identified here.
If cat bonds and ILS products in general do reach a “tipping point” of the kind some
seek, in which underwriting to securitize climate risks becomes a principal strategy for
insuring vulnerable landscapes, the prospects for climate-appropriate development and
adaptation would seem dim. If high yield is a function of high risk, then it is credulous
to imagine that ILS instruments will ever systematically encourage less risk taking. As I
demonstrated in the second section, the cat bond market is characterized by the highly
segmented and stratified provision of coverage in which “peak perils” attract an inordi-
nate amount of the total capital that is invested. There is no reason to expect that this
dynamic will change fundamentally.
Hence, investors’ ongoing interest in high-paying peak perils could ironically make 179
more capital available for paying claims, rebuilding, and new underwriting in places like
Miami, Florida, where the highest concentrations of value are being made even more
14
Thanks to an anonymous reviewer for framing the issue in these terms.
ECONOMIC GEOGRAPHY
is a more than passing resemblance to the panic witnessed in the credit crisis of 2007–08.)
Primary insurers who suddenly found themselves without investors willing to buy their
bonds might then become dependent on the state as an insurer of last resort. As Ericson and
Doyle (2004a, 2004b) have pointed out, the insurance industry typically reveals its uneasy
dependence on the financial apparatus of the state in times of uncertainty and crisis:
Insurers are eager to gamble on uncertainties as long as they keep reaping profitable returns.
However, an extraordinary catastrophic loss leads them to hop on the pass-the-exposure
express and seek refuge until they can restructure the market and gain confidence that they will
again be on the winning side. The refuge is provided by governments, who help underwrite
faltering insurance markets. This socialism for business enterprise can supersede any social-
izing effects on behalf of the insured (Ericson and Doyle 2004b, 169).
Although the authors based this description on the behavior of (re)insurers following
the September 11 terrorist attacks, their conclusion has broader applications, given how
well it describes the behavior of financial capital following the subprime crisis. In cases
of extreme crisis, the entire financial services industry depends on state intervention to
180 provide an economic backstop, either through direct infusions of capital—the bailout
model—or by acting as an insurer of last resort—the subsidy model. The ultimate
economic burden is devolved to taxpayers on the one hand, and sovereign debt markets
on the other. The market in securitized geophysical risks is unlikely to prove an exception
to this rule.
Conclusion
This analysis of the catastrophe bond market has enumerated the drivers and impli-
cations of the securitization of fixed capital’s place-based vulnerabilities. I have dem-
onstrated how parts of (re)insurance markets have been rendered investment-worthy for
financial capital through enabling purposive, scalable, and selective engagements with
insurance risks. This securitization pathway allows mobile capital on a search for yield
to reframe spatial liabilities as tradable assets, thus accessing new “returns on place.”
Although the search for new risk-return profiles has fueled the simultaneous parsing and
recombination of risks from peak peril regions and the extension of securitization to new
locales and geophysical events, the geographic footprint of cat bond investments remains
extremely uneven and concentrated on peak perils in the United States and, to a far lesser
degree, in Europe and Japan.
Investors’ continued interest in the ILS market hinges on the articulation of a funda-
mentally ambivalent relationship with geographic space. Although the commodification
of catastrophe risks requires tying financial capital to absolute spaces, the high rates of
return on catastrophe bonds are possible only because these instruments enable financial
capital to avoid the fixed costs and relational entanglements borne by (re)insurers. This
selective engagement with catastrophe risks is exemplified in investors’ reformulation of
the impacts of climate change on extreme events as new growth opportunities.
Recent proposals to underwrite new insurance policies with the express intention of
securitizing them have hypothesized that the cat bond market could parallel the expan-
sion of asset-backed securities and credit derivatives in the 2000s. Documenting the
effects and outcomes of these proposals is a task for future research. But if catastrophe
bonds share much in common with other alternative assets like residential mortgage-
backed securities, then the similarity of the risks posed by these instruments to particu-
larly vulnerable places and populations should give us pause. As Christophers (2009,
821) has pointed out with respect to asset-backed securitization and credit derivatives,
“the more institutions there are passing finite amounts of money back and forth, with
Vol. 90 No. 2 2014
each such institution determined to extract its own return on capital, . . . the greater the
cumulative demand that is obviously made on that finite money—requiring greater risks
to be taken to sustain returns while underlying markets are rising, and leading to
significantly more (and more widespread) damage when those markets collapse.” In the
case of cat bonds, the same search for yield that leads investment to concentrate in
high-paying peak peril regions could make more capital available for rebuilding and new
underwriting in areas where high concentrations of value are becoming even more
vulnerable to extreme events. Incentivizing post-catastrophe rebuilding in peak peril
areas would magnify the peaks and troughs of preexisting uneven development rather
than dampen them. In this context, if a series of major catastrophes then prompted
investors to retreat, the state would find itself an insurer of last resort for a built
environment that had ironically become less adapted to climate extremes. If securitiza-
tion discourages climate-appropriate adaptation and development, the end result may be
landscapes that are even more vulnerable than those of today.
Aalbers, M. 2008. The financialization of home and the mortgage market crisis. 181
Competition and Change 12:148–66.
References
——. 2009. The sociology and geography of mortgage markets: Reflections on the
Hall, S. 2013. Geographies of money and finance III: Financial circuits and the “real economy.”
Progress in Human Geography 37:285–92.
Hartwig, R. 2008. 2007 year end results, April 9. Insurance Information Institute. Available
online: www.iii.org/articles/2007_year_end_results.html
Harvey, D. 1982. The limits to capital. Oxford, U.K.: Basil-Blackwell.
Herweijer, C., Ranger, N., and Ward, R. 2009. Adaptation to climate change: Threats and
opportunities for the insurance industry. Geneva Papers on Risk and Insurance—Issues and
Practice 34:360–80.
Ho, K. 2009. Liquidated: An ethnography of Wall Street. Durham, N.C.: Duke University Press.
IAIS. 2009. Global reinsurance market report 2009. Basel: International Association of Insurance
Supervisors.
Insurance Insider. 2009. ILS shows 60% correlation to wider markets. Insurance Insider May 11:17.
Jaffee, D. M., and Russell, T. 1997. Catastrophe insurance, capital markets, and uninsurable risks.
Journal of Risk and Insurance 64:205–30.
Jagers, S., Paterson, M., and Stripple, J. 2005. Privatizing governance, practicing triage: Securiti-
zation of insurance risks and the politics of global warming. In The business of global environ-
mental governance, ed. D. Levy and P. Newell, 249–74. Cambridge, Mass.: MIT Press.
Johnson, L. 2010. Climate change and the risk industry: The multiplication of fear and value. In
Global political ecology, ed. R. Peet, P. Robbins, and M Watts, 185–202. London: Routledge.
183
——. 2011. Insuring climate change? Science, fear and value in reinsurance markets. PhD dissertation,
Martin, Randy. 2006. American war and the financial logic of risk management. Durham, N.C.: Duke
University Press.
Martin, Ron. 2011. The local geographies of the financial crisis: From the housing bubble to
economic recession and beyond. Journal of Economic Geography 11:587–618.
McDowell, L. 1992. Valid games? A response to Erica Schoenberger. Professional Geographer
44:212–15.
——. 1998. Elites in the City of London: Some methodological considerations. Environment and
Planning A 30:2133–46.
Michel-Kerjan, E., and Morlaye, F. 2008. Extreme events, global warming, and insurance-linked
securities: How to trigger the tipping point. Geneva Papers on Risk and Insurance–Issues and
Practice 33:153–76.
Moore, J. 2000. Environmental crises and the metabolic rift in world-historical perspective.
Organization and Environment 13:123–57.
Munich Reinsurance. 2006. Hurricanes—More intense, more frequent, more expensive: Insurance in
a time of changing risks. Munich: Munich Reinsurance.
Muniesa, F., Millo, Y., and Callon, M. 2007. An introduction to market devices. In Market devices,
ed. M. Callon, Y. Millo, and F. Muniesa, 1–12. Oxford, U.K.: Blackwell.
Newman, K. 2009. Post-industrial widgets: Capital flows and the production of the urban.
184 International Journal of Urban and Regional Research 33:314–31.
O’Malley, P. 2003. Governable catastrophes: A comment on Bougen. Economy and Society
32:275–79.
O’Neill, P. 2009. Infrastructure investment and the management of risk. In Managing financial
risks, ed. G. Clark, A. Dixon, and A. Monk, 163–89. Oxford, U.K.: Oxford University
Press.
Parry, B. 1998. Hunting the gene-hunters: The role of hybrid networks, status, and chance in
conceptualising and accessing “corporate elites.” Environment and Planning A 30:2147–62.
Pielke, R., Gratz, J., Landsea, C., Collins, D., Saunders, M., and Musulin, R. 2008. Normalized
hurricane damage in the United States 1900–2005. Natural Hazards Review 9:29–42.
Pike, A., and Pollard, J. 2010. Economic geographies of financialization. Economic Geography
86:29–51.
Pollard, J. S., Oldfield, J., Randalls, S., and Thornes, J. E. 2008. Firm finances, weather derivatives
and geography. Geoforum 39:616–24.
Pryke, M. 2007. Geomoney: An option on frost, going long on clouds. Geoforum 38:576–88.
Rajan, R. 2010. Fault lines: How hidden fractures still threaten the world economy. Princeton, N.J.:
Princeton University Press.
Randalls, S. 2006. Firms, finance, and the weather: The UK weather derivatives market. PhD
thesis, School of Geography, Earth and Environmental Sciences. Birmingham, U.K.: University
of Birmingham.
——. 2009. Managing financial risks in urban environments. In Managing financial risks: From global
to local, ed. G. Clark, A. Dixon, and A. Monk, 209–29. Oxford: Oxford University Press.
Schoenberger, E. 1991. The corporate interview as a research method in economic geography.
The Professional Geographer 43:180–89.
Schwierz, C., Köllner-Heck, P., Zenklusen Mutter, E., Bresch, D., Vidale, P.-L., Wild, M., and
Schär, C. 2010. Modelling European winter wind storm losses in current and future climate.
Climatic Change 101:485–514.
Sinclair, T. 2005. New masters of capital: American bond rating agencies and the politics of credit-
worthiness. Ithaca, N.Y.: Cornell University Press.
Smith, N. 2008 [1984]. Uneven development. Athens: University of Georgia Press.
Solomon, S., Qin, D., Manning, M., Alley, R. B., Berntsen, T., Bindoff, N. L., Chen, Z.,
Chidthaisong, A., Gregory, J. M., Hegerl, G. C., Heimann, M., Hewitson, B., Hoskins, B. J., Joos,
F., Jouzel, J., Kattsov, V., Lohmann, U., Matsuno, T., Molina, M., Nicholls, N., Overpeck, J.,
Raga, G., Ramaswamy, V., Ren, J., Rusticucci, M., Somerville, R., Stocker, T. F., Whetton, P.,
Wood, R. A., and Wratt, D.. 2007. Technical Summary. In Climate Change 2007: The Physical
Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovern-
Vol. 90 No. 2 2014
mental Panel on Climate Change. ed. S. Solomon, D. Qin, M. Manning, Z. Chen, M. Marquis, K.
B. Averyt, M. Tignor, and H. L. Miller, 19–91. Cambridge, U.K.: Cambridge University Press.
Standard & Poor’s. 2009. Methodology and assumptions for rating natural catastrophe bonds. 12
May. New York: Standard & Poor’s.
——. 2011. Mariah Re Ltd. Series 2010-1 Notes Downgraded to “CC(sf)” From “CCC(sf).” 13
October. New York: Standard & Poor’s.
Sturm, T., and Oh, E. 2010. Natural disasters as the end of the insurance industry? Scalar
competitive strategies, alternative risk transfers, and the economic crisis. Geoforum 41:154–63.
Swiss Reinsurance. 2004. Focus report: Hurricane season 2004—Unusual, but not unexpected.
Zurich: Swiss Reinsurance Company.
——. 2006. Natural catastrophes and man-made disasters 2005: High earthquake casualties, new
dimensions in windstorm losses. Sigma 2/2006. Zurich: Swiss Reinsurance.
——. 2009. The role of indices in transferring insurance risks to the capital markets. Sigma
4/2009. Zurich: Swiss Reinsurance.
——. 2010. World insurance in 2009. Sigma 2/2010. Zurich: Swiss Reinsurance.
Thornes, J., and Randalls, S. 2007. Commodifying the atmosphere: Pennies from heaven?
Geografiska Annaler: Series A, Physical Geography 89:273–85.
Tickell, A. 1996. Making a melodrama out of a crisis: Reinterpreting the collapse of Barings Bank.
Environment and Planning D: Society and Space 14:5–33.
185
Trading Risk. 2010a. Secretive $250mn Merna Re III sold to single fund. In Trading Risk, 4.