Alternative risk transfer (ART) includes those contracts, structures, and solutions that enable firms either to finance or transfer some of the risks to which they are exposed in a non-traditional way.2 ART is all about “convergence” – the convergence of capital markets and insurance, the convergence of corporate finance and risk management, the convergence of swap dealers with (re-)insurance companies, and so on.3
The increasingly diverse set of offerings in the ART world has broadened the range of solutions available to corporate risk managers for controlling undesired risks, increased competition amongst providers of risk transfer products and services, and heightened awareness by corporate treasurers about the fundamental relations between corporation finance and risk management.
This chapter presents a descriptive overview of the major products and solutions in the ART market today. After reviewing in Section 2 the origins of the term “ART” in the form of “captives,” Section 3 then describes “finite risk” programs (including some recent controversy that has surrounded these innovative structures). The products discussed in Sections 2 and 3 tend to be used primarily for risk finance (i.e., raising funds at a fixed price to smooth the cash impact of a loss) rather than risk transfer (i.e., actually shifting the adverse impact of a risk to another firm). Sections 4 and 5 review two ART forms used for true risk transfer – multi-line, and multi-trigger programs. Section 6 then explains how the structured finance world is increasingly becoming a part of the ART universe, both for risk finance and risk transfer applications. Section 7 summarizes a relatively new ART form known as contingent capital, and Section 8 concludes.4
2. Self-Insurance, Captives, and the Emergence of ART
ART first gained widespread acceptance as an industry term in the 1970s to describe organized self-insurance programs. As insurance markets hardened and produced rising premiums and declining capacity, corporations wanted to emphasize to insurers that they could often seek the protection they needed through al
2 Not all risk, of course, is “undesired.” Some risk is necessary for the profitable operation of a business. How to distinguish between core risks that a company is in business to bear and other risks is the subject of lengthy discussions in Culp (2001, 2002a, 2004).
3 The convergence of corporate finance and risk management that the proliferation of ART products has encouraged is not emphasized in this article, purely for reasons of length. For that analysis, interested readers are requested to consult Culp (2002a, 2002b, 2002c, 2002d).
4 Although the material here is original, portions of this chapter draw heavily from Culp (2002a).
5 Captive-friendly domiciles exist both on-shore (e.g., the State of Vermont) and off-shore (e.g., Bermuda, Singapore, and the Channel Islands).
Alternative Risk Transfer
ternative means, the most obvious of which was self-insurance. Even today, more and more corporate treasurers are reminded that their weighted-average cost of capital – i.e., the cost of raising capital to self-insure – should be approximately the limit they are willing to pay for external capital provided through an insurance program.
The decision to self-insure is called a retention decision. A planned retention is a risk to which a firm is naturally subject that the firm’s security holders are prepared to bear on an ongoing basis. A planned retention may occur either because the alternative – risk transfer – is too expensive, or because the risk is considered integral to the firm’s core business activities and operating profits. (Culp 2001, 2004)
A funded planned retention is a retained risk for which a firm sets aside funds to smooth the cash flow impact of future losses. A firm may wish to obtain funds now to cover a subsequent loss, for example, if post-loss funding costs are expectedly to rise dramatically in response to the announcement of the loss. The practice of pre-funding retained risks is known broadly as risk finance.
Self-insurance is a form of risk finance. Not all self-insurance, however, is ART. Many self-insurance schemes are either indirect or are informal, such as earmarked reserves. Funds allocated to such loss reserves neither get the favorable tax and accounting treatment afforded true insurance or qualified self-insurance, nor are such funds immune from the temptation managers may have to use the funds in some other way.
The major ART forms designed to facilitate the pre-loss funding of planned retentions are significantly more “formalized” than simply earmarking balance sheet reserves. These ART forms are discussed in the sections below. Although many were originally and still are primarily intended only as a source of risk finance, some of the more recent structures discussed below also include some degree of risk transfer, as well.
2.1 Single-Parent Captives
To set aside funds as self-insurance against future losses in a way that does not arise investor suspicions that the money could be spent in some other way, the alternative risk financing structure called a captive emerged in the 1970s. In its most basic form – called a single-parent captive – the sponsoring firm sets up a wholly owned subsidiary that is also a licensed (re-)insurance company and then purchases insurance from itself by way of the new captive subsidiary. The equity capital of the captive is usually minimal – just enough for the captive to obtain an insurance or reinsurance license in a captive-friendly domicile.5 Having obtained this license, the captive then writes explicit insurance contracts to the ceding spon
sor to cover the risks the sponsor wishes to pre-fund in exchange for explicit premium payments. In certain circumstances, the premium paid is tax deductible.
Captives are commonly used by firms to insure high-frequency, low-severity loss events for which the ceding sponsor has a relative stable historical loss experience. The expected losses in such cases are usually relatively easy to estimate, and the premium collected by the captive for providing insurance is set equal to those expected losses in present value terms over some risk horizon– usually a year. If the present value of actual losses exceeds the present value of expected losses charged by the captive to the sponsor as a premium, the risk borne by the captive – and ultimately the ceding sponsor – is called underwriting risk.
Because loss claims do not necessarily arrive in the same time period (e.g., year) that premium is collected, however, the captive also faces both timing and investment risk. Timing risk is the risk that the assets acquired by the captive using premium income to fund future claims grow at a rate that may be perfectly correct in a present value sense – i.e., after a year the assets may be worth exactly the total claims paid – but that may be too low to finance the unexpected arrival of a lot of large claims early in the insurance cycle. In other words, timing risk is the risk that the captive’s assets are inadequate at any discrete point in time to fund its current liabilities. Investment risk is the related risk that market risk on the assets acquired by the captive to fund its claims results in an unexpected shortfall of assets below insurance liabilities.
Like any other insurance company, a captive manages its underwriting risk by attempting to price its insurance to cover expected losses but manages its timing and investment risk through its technical reserves. Technical reserves represent the future claims expected on the insurance contracts the captive has written to its ceding sponsor corporation and come in two types – unearned premium reserves, and loss reserves.
In most insurance lines (e.g., liability and property), policy coverage lasts one year and premium is payable at the beginning of the policy year. Although premium is collected in advance, it is earned only as time passes if a claim has not occurred. Unearned premium is premium that has been collected that may still need to be used to cover an as-yet-unsubmitted claim. The unearned premium reserve is thus the proportion of premium that must be set aside to honor future expected claims.
The technical reserves an insurance company maintains to honor any future claims – known or unknown – above the unearned premium is called the loss reserve. Loss reserves may be set aside for losses that have been reported and adjusted, reported but not adjusted, incurred but not reported (IBNR), or for loss adjustment expenses.
Like more traditional insurance companies, captives engage in one of two types of reserve management methods for financing the claims arising from their liabilities. (Outreville 1998) Under the first method – the capitalization method – the captive invests the premium collected from the ceding sponsor in assets and then uses those assets plus the return on those assets to finance subsequent insurance claims. Captives using the capitalization method usually attempt to keep assets funded by premium collections linked to the technical reserves of the liabilities for
Alternative Risk Transfer 5
which premium was collected. Technical reserves at captives using the capitalization method tend to be medium- or long-term, as are the assets invested to back the corresponding liabilities.
The compensation method, by contrast, is a “pay-as-you-go” system in which all premiums collected over the course of a year are used to pay any claims that year arising from any insurance coverage the captive has provided to its sponsor. Under this method, no real attempt is made to connect assets with technical reserves. All premium collected is used to fund mainly short-term assets, and those assets collectively back all technical reserves for all insurance lines.
One important implication of the differences in reserve management styles is the captive’s potential demand for reinsurance. If the captive becomes concerned that the funded retention should actually have been transferred rather than retained, the captive structure makes it easy for the firm to acquire selective reinsurance for risks about which the ceding sponsor and the captive may have been especially worried.
A major attraction of the captive structure – like pure self-insurance – is the retention of underwriting profits and investment income on assets held to back unearned premium and loss reserves. If the actual losses underwritten by the captive are lower than expected, the sponsor can repatriate those underwriting profits – plus any investment income – in the form of dividends paid by the captive to its sole equity holder, the sponsor.
Local laws, regulations, or tax requirements often require firms to obtain local insurance coverage. In this case, firms may opt for a captive structure in which the captive is incorporated and chartered as a reinsurance company rather than an insurance company. In this case, the ceding sponsor then buys its coverage from a locally recognized insurer (called a fronting insurer), which then reinsures 100% of the exposure with the captive. In some cases, multiple fronting insurers are required to provide recognized cover to different operating subsidiaries of multinationals, as illustrated in Figure 1.
Figure 1. Single-Parent Captive with Fronting Insurer 2.2 Other Captive-Like Structures
The costs of setting up a captive are often surprisingly low. Most of the costs go to the fronting insurers (if required) and to the captive manager (i.e., the firm
retained to run the captive, process claims, and the like). Despite the relatively affordable nature of a captive, not all firms wishing to self-insure find single-parent captives to be the ideal solution. Some firms, for example, self-insure cyclically based on how high external premiums are. For such firms, the costs of constantly setting up and dismantling captives (or allowing one to remain open but idle) can get prohibitive quickly.
Firms that opt not to establish single-parent captives have several alternatives available. The difference between most of these alternatives is the degree to which the structures facilitate pure risk financing versus actual risk transfer. This distinction will become clearer as the alternatives are discussed below.
2.2.1 Mutualized Structures
One way to enjoy the benefits of a captive without setting up a single-parent captive is to share captive ownership through a multi-parent or group captive – i.e., a captive insurer whose equity ownership is held by several firms rather than just one. Such structures are also similar in design and operation to mutual insurance companies, risk retention groups, and other cooperative-style insurance companies.
A group captive, for example, is a captive reinsurance company that collects premium from multiple sponsors and in turn agrees to underwrite certain risks of those sponsors. The premium, investment income, expenses, and underwriting risks are all pooled. The loss sharing regime, in turn, may be proportional to the premiums paid into the captive or fully mutualized. In either case, group and multi-parent captives involve some degree of risk transfer through the pooling of claims by the multiple sponsors.
Group captives are often set up by industry trade associations on behalf of their members. Energy Insurance and Mutual Limited, for example, is the group captive representing numerous U.S. electricity and gas utilities. The benefits of pooling premiums and risks allow the group captive to achieve a smoother time profile of loss payouts than would be possible in any individual participant’s situation.
2.2.2 Rent-A-Captives and Protected Cell Companies
If a firm does not wish to establish a single-parent captive but also does not wish to engage in risk transfer through the pooling arrangements typically associated with mutuals, two alternative structures are still available in the self-insurance realm. The first is called a rent-a-captive. A rent-a-captive is essentially similar to a multi-parent group captive except that the participating corporations relying on the captive for insurance do not actually own any part of the rent-a-captive and do not pool their risks with one another.
Rent-a-captives are set up, maintained/managed, and owned by market participants like (re-)insurance companies or insurance brokers for the benefit and use of corporate customers. The customers in turn remit premium payments to a fronting
6 The rent-a-captive may also require collateral from participants in excess of premium paid to pre-fund later losses.
Alternative Risk Transfer
insurer that then cedes the premium to the rent-a-captive through facultative reinsurance to give the customer coverage for losses on the risks they wish to retain. The rent-a-captive itself typically maintains “customer accounts” for participants in which premiums are credited and claims booked. In addition, investment and underwriting income are tracked and may be returned to the participants, usually when the rent-a-captive contract is terminated. Unlike a multi-parent captive, the individual customer accounts in a rent-a-captive are segregated.
From the perspective of a self-insuring participant, the rent-a-captive works much like a single-parent captive except that ownership rights and dividends now accrue to a third party.6 Figure 2 illustrates.
Figure 2. Rent-a-Captive
Some have expressed concerns, however, that rent-a-captives do not achieve true customer account segregation – specifically, that the customer accounts are on paper only, but that the actual funds are commingled. Participants then worry that the commingled assets of the captive may be mis-invested, yielding reserve losses, or that the loss of one firm could be ex post mutualized in the event of the captive’s insolvency.
As a result of some of these concerns about rent-a-captives, captive management organizations have been offering a second alternative for corporations wishing to self-insure in a non-mutualized captive structure without setting up a singleparent captive. Called protected cell companies (PCCs), these entities are set up essentially like a rent-a-captive except that customers have ring-fenced and genuinely segregated, bankruptcy remote accounts.
3. Finite Risk
Once known as financial reinsurance and unique to the insurance industry, finite reinsurance or finite risk products now provide insurance companies and nonfinancial corporates alike with an important source of risk finance. Like captives, finite risk products are primarily intended to help firms pre-fund a retained risk that the firm wishes to self-insure. Increasingly, however, finite risk programs also
contain some degree of risk transfer, as well, thus offering corporate risk managers a way to pre-fund certain losses and transfer others.
A major distinction between the types of finite risk ART forms available in the market today is whether or not the liability whose timing risk is being managed with a finite risk product has or has not already been incurred. Retrospective finite risk products are intended to help firms manage the timing risks of existing liabilities of the firm, whereas prospective finite risk solutions cover contingent liabilities that have not yet been formally assumed by the firm. In the case of an insurance company seeking financial reinsurance through finite risk products, retrospective finite risk products cover past underwriting years and prospective products cover current or future underwriting years. For a corporation, the distinction is essentially the same except that the liabilities being managed are not acquired through an underwriting process but are instead the result of some business decision(s) made by the firm that alter its natural risk profile.
3.1 Typical Finite Risk Structures
A Loss Portfolio Transfer (LPT) is the cession by a firm of all remaining unclaimed losses associated with a previously incurred liability to a (re-)insurer. In addition to paying an arrangement fee, the cedant also typically pays a premium equal to the net present value of reserves it has set aside for the transferred liability plus a risk premium to compensate the (re-)insurer for the timing risks of the assumption. A LPT thus enables a firm to exchange an uncertain liability in the form of a stream of unrealized losses over time for a certain liability whose present value is equal to the expected NPV of the unrealized losses plus a risk premium and a fee.
The principal risk that the cedant transfers to the (re-)insurer through a LPT is the timing risk that losses or claims arrive at a much faster rate than expected. In that case, the investment income on the reserves – and perhaps the reserves themselves – may be inadequate to fund the losses. A time series of losses that occurs more slowly than expected, by contrast, will represent an opportunity for a net gain that the (re-)insurer would typically share with the cedant. LPTs thus are risk financing mechanisms through which firms can address the timing risk of a liability.
LPTs usually include aggregate loss limits, as well as exclusions for certain types of risks not arising directly from the ceded liabilities. Per loss deductibles are sometimes also included in LPTs by (re-)insurers. Because the timing of losses ceded in an LPT can sometimes be extremely long-term, the cedant may also demand some kind of surety from the assuming (re-)insurer. Letters of credit, collateral, or bank guarantees may be requested by a cedant to prove the financial integrity if the (re-)insurer has questionable credit quality.
LPTs can be attractive sources of risk finance for various reasons. LPTs can also benefit non-insurance, corporate customers seeking to swap an uncertain liability stream for a fixed payment today. LPTs can help corporations with cap-
Alternative Risk Transfer
tives, for example, wind up certain self-insurance lines if the firm alters its retention decision for certain risks. LPTs are also useful to non-financial corporations in securing risk financing for run-off solutions, especially in the area of environmental claims and clean-up cost allocation.
In order to qualify as legitimate insurance transactions for tax and accounting purposes, however, finite risk structures must involve some degree of underwriting risk for the (re-)insurer. Accordingly, LPTs are often coupled with finite reinsurance contracts known as adverse development covers (ADCs) – essentially just excess-of-loss coverage above a certain minimum attachment point and up to a limit. The lower attachment point of an ADC is usually close to the cedant’s current reserves.
Consider, for example, a firm with a workers comp exposure to asbestos claims from its personnel over the next five years. The firm estimates the terminal value of its five-year liability at €10 million five years hence and has set aside the present value of that amount in reserves – say, €9 million. The firm remains concerned, however, about two risks: that a large claim will occur earlier than planned, and that total losses will exceed estimated losses. A typical finite transaction would involve a LPT and ADC in which the firm cedes the €9 million to a reinsurance company. In return, the reinsurer agrees to cover €12.5 million in losses. The insurer is exposed to timing risk on the first €10 million in claims and underwriting risk on the remaining €2.5 million.
Adding an ADC to a LPT in a finite structure is not merely a question of tax and accounting. The ADC can also play an important role for the cedant by protecting the firm against the risk that realized losses on an existing liability are higher reported and forecast. ADCs are commonly used, for example, to cap old liabilities that are of concern in a merger or acquisition. When the acquiring firm or merger partner is concerned that a liability could be much greater than the target firm has planned for in its reserve holdings, the cession of risk through an ADC can provide the target firm with a good remedy to such concerns on the part of its suitor.
3.2 Potential Benefits to Corporates
The potential benefits to corporates of finite risk programs are significant. As this range of benefits has become better understood, interest in these products has significantly increased in the past several years. In particular, finite gives corporates an intermediate solution for situations in which the retention level is uncomfortably high but a pure risk transfer solution is unavailable or too expensive. Finite products also provide an alternative to risk transfer solutions that directly impact a firm’s working capital layer. By partially funding retentions outside the working capital layer, firms can increase their debt capacity.
Finite risk programs can also have other potentially significant benefits for corporates. Finite can help firms stabilize an insurance budget and, when properly constructed and accounted for, reduce earnings and/or cash flow volatility. Finite
also allows firms to create off-balance-sheet provisions for unusual risks (e.g., extreme “tail” risk events, exotic risks, operational risks, etc.).
In addition, finite risk programs are widely regarded as important devices for combating adverse selection problems through positive signaling. A firm that enters a charge-off against its earnings for a liability that has not been fully realized, for example, may be suspected of possessing superior information about the liability that leads to under-reporting. A firm wishing to counter such fears by investors can take out an ADC to lock in its liability at the charge-off amount and thus signal its confidence that the charge-off was indeed correct. (Shimpi 2001)
Turner & Newall, a United Kingdom motor components manufacturer, utilized an ADC for signaling purposes – i.e., to combat a concern amongst investors and analysts that it had inadequately reserved against a major liability.7 The liability for Turner & Newall was a series of asbestos claims associated with some of its discontinued operations.
Turner & Newall self-insured its asbestos claims by establishing a captive and then reinsured some of that underwriting risk with an ADC for $815m XS $1,125mn. The ADC had a 15-year tenor and, like other finite risk products, contained an agreement for a partial premium rebate if actual loss developments were favorable relative to its reserve holdings after the 15 years.
In a more general case, the multinational firm Hanson PLC was concerned when it acquired building materials company Beazer PLC that Beazer’s discontinued U.S. operations would create an impediment to growth for the new conglomerate. Hanson self-insured the liabilities of Beazer’s U.S. operations through a captive, and the captive, in turn, acquired $80mn. XS $100mn.8 in an ADC in perpetuity. In this manner, Hanson ringfenced the liabilities of Beazer’s discontinued U.S. operations using an ADC.
Although all of the above examples involve retrospective coverage for liabilities already incurred, finite reinsurance can also be applied prospectively to liabilities that have not yet been incurred – e.g., a policy line that an insurance company intends to offer but has not yet, or an environmental liability that a firm will incur when a factory goes into future production.
3.3 The AIG/Brightpoint SEC Settlement
The origin of finite risk traces back to the “time and distance” policies once commonly used in Lloyd’s by insurance syndicates to smooth the volatility of their earnings and premium income. In a time and distance policy, one insurer makes a premium payment to another insurer in exchange for insurance coverage that exactly equals the terminal value of the premium income stream. This allows
7 The details of this example are discussed in GGFP (2000).
8 The notation $A XS $B refers to an excess-of-loss reinsurance treaty with a lower attach
ment point of $B and a coverage level of $A. In the Hanson/Beazer example, the ADC
thus reimbursed the firm for any losses above $100 million up to $180 million.
9 Actually, AIG went out of its way never to call the product a finite product. And in fact, it was not a finite product in the legitimate sense of the term.
Alternative Risk Transfer
the cedant to stabilize the volatility associated with any claims by swapping a cash flow stream with uncertain timing for a certain cash flow stream.
Time and distance policies are no longer considered legitimate insurance outside of Lloyd’s. In the United States, for example, an insurance contract must include some element of underwriting risk and true risk transfer to be distinguished from a financing or depository arrangement. Prior to the failure of Enron, most considered a “10:10” rule reasonable – viz., if there is at least a 10% chance that at least 10% of the policy risk is borne by the insurer, the contract is “insurance.” Today, a “20:20” rule is generally applied. On a €10 million finite policy, for example, the insurance provider must essentially face at least a 20% chance of incurring a €2 million underwriting loss. Otherwise, the deal is considered a financing arrangement.
A recent enforcement action by the U.S. Securities and Exchange Commission (SEC) reminds us of the importance of these requirements. The SEC announced in September 2003 that it had reached a settlement with American International Group (AIG) for payment by AIG of a civil penalty in the amount of $10 million. The penalty was based on AIG’s role in helping a firm called Brightpoint perpetrate an accounting fraud using a finite product.9
In October 1998, Brightpoint – a distribution and outsourcing firm – announced that it expected to take a one-time charge-off associated with the closure of its U.K. division in the range of $13 to $18 million. By December of 1998, the estimate of the loss had grown to $29 million. But rather than restate, Brightpoint and AIG entered into a transaction that ostensibly allowed Brightpoint to report actual losses in the estimated range.
The AIG/Brightpoint agreement was a finite program with both retrospective and prospective components. The policy specified coverage limits of $15 million on each component and called for a total premium payment of around $15.3 million. About $15 million of the premium was allocated to the retrospective component, and the policy included language that AIG’s retrospective liability could never exceed the premium collected.
Under U.S. GAAP, a firm is allowed to net the benefit of an insurance program against the associated loss as long as the recovery is considered “probable.” At the same time, the premium paid can be expensed over the life of the policy. But if it is “possible” but not “probable,” the insurance cannot be used to reduce the size of the loss. And if the recovery is “known with certainty,” then the recovery can be netted against the loss, but in that case the entire premium must be expensed in the same quarter that the loss and recovery are recognized and netted.
Brightpoint expensed the premium monthly over the life of the policy and was thus able to net a “probable recovery” of $11.9 million against the $29 million loss to keep the total reported loss in the estimated loss range. Because the contract exposed AIG to absolutely no underwriting risk, however, the SEC concluded that the contract was not “insurance” and thus should have been accounted for as a pure deposit of cash. Even if the contract had been true insurance, the retrospec
tive loss appears to have been known by the time the policy was put in place, which should have required Brightpoint to expense all of its premium in 1998Q4. Instead, Brightpoint tried to expense it monthly over the next several years, as if the whole policy were prospective.
4. Multi-Line Programs and Risk Bundling
Risk transformation products are often distinguished from one another along three dimensions: length of coverage; sequencing of losses borne by risk transfer counter parties; and types of risk resulting in potential losses. The relation between the sequencing of losses borne by risk transfer counter parties and the types of risk giving rise to those losses is often different for traditional and alternative risk transfer products. The former are usually characterized by a “layered” approach in which individual risks are placed into silos and the risks are then transferred in layers – either “horizontal” or “vertical” layers or both.
Consider a hypothetical non-financial U.S. corporation ABC that buys coffee beans from Brazil and then sells ground coffee and coffee products in Europe and the United States. ABC thus is exposed to exchange rate risk, coffee price risk, and credit risk. Now suppose that ABC also faces property and liability risks.
Under a traditional insurance program, ABC would manage its risk in “silos” – i.e., one risk at a time. Classical excess-of-loss reinsurance is typical of a vertically layered risk transfer solution, where vertical refers to the sequence in which the risk transfer counter parties absorb losses the firm incurs. The company might, for example, vertically layer its property loss coverage. Up to $100 million is retained, and then two different insurers provide excess-of-loss coverage in two different vertical loss layers. One insurer provides protection for every property loss above $100 million up to $600 million (i.e., the $500mn XS $100mn layer), and another provides protection for property damage in the $600 million to $800 million vertical loss layer.
ABC’s exchange rate risk, by contrast, might be horizontally layered, or shared with two firms that provide shared protection for every dollar loss up to $500 million regardless of the order in which the loss occurs. If ABC loses $1 from exchange rate movements, each hedge counter party owes $0.50. Similarly, a total annual loss of $500 million results in a $250 million payment by each counter party.
ABC’s liability risk, in turn, might be subject to a blended cover. The firm could buy two insurance policies, each with a $50 million deductible and $350 million policy limit, for the first $300 million of its non-retained liability risk. A third policy might then compensate ABC for $200mn. XS $400mn. with a $50 mn. deductible. If ABC loses $375 million in one year on property damage, the third policy is “out-of-the-money” and the first two are completely exhausted. The whole program is shown in Figure 3.
Alternative Risk Transfer
4.1 Overcoming Silo-by-Silo Inefficiency
One major inefficiency often associated with vertically and/or horizontally layered silo-by-silo risk transfer programs is the over-commitment of capital such programs can engender to each risk silo. In the previous example, firm ABC has secured too much risk transfer capacity unless the occurrence of large losses across risks are perfectly correlated over time. It has paid for $500 million in exchange rate protection and $700 million in property protection. But in a multi-year context, ABC will not possibly need all $1.2 billion unless the maximum covered losses occur in both risk silos at the same time. On a correlation-adjusted portfolio basis, the total loss exposure of ABC is lower than the sum of the two individual risk silos.
Multi-line programs can, in principle, help firms address this problem in two ways. First, firms can allocate less capital to their risks at a lower total cost when correlations across both time and risk types are factored into the premium charged for the policy. Second, firms can achieve a more customized, tailor-made blanket of coverage that includes only those risks with which the firm is truly concerned
about transferring to another party. At the same time, a multi-line program can also be an effective way to catch “gaps” between silos arising from unplanned retentions and failures in the risk identification process.
The supposed benefits of a multi-line program do not just emphasize enhanced efficiency in capital utilization. In addition, most multi-line policies are provided by a single carrier, thus reducing transaction costs and total arrangement fees. That multi-line products are also usually multi-year policies also simplifies renewals and/or reinstatements.
The basic structure of a multi-line program can, of course, be modified in a number of ways to suit the needs of corporate risk transfer customers. A common modification, for example, is to allow firms to optionally attach a risk silo-specific catastrophic excess-of-loss layer to the program for any risk with which the firm is inordinately concerned. Figure 4 shows a basic multi-line program for ABC with a single $100mn. deductible and an aggregate policy limit of $500mn. In addition, the program includes a $300mn. XS $500mn. layer of catastrophic excess-of-loss property coverage.
Figure 4. Multi-line program with optional excess-of-loss property cover.
Comparing Figures 3 and 4, the differences in the programs are clear. In the multi-line program, all losses count toward the same deductible and the same limit, thus preventing the inefficient allocation of capital to one silo when it is badly needed in another.
Some multi-line products also combine risk- or occurrence-specific limits with aggregate limits to help firms further customize their exposure. One such program is Swiss Re’s Multi-line Aggregated and Combined Risk Optimization (MACRO), a product aimed at non-financial corporations to help them bundle and tailor their
Alternative Risk Transfer
exposure profiles and retention decisions. The MACRO product is a multi-line, multi-year structure that has a single annual deductible and a single per-year exposure limit, both defined across all risks covered in the program. In addition, the program has an aggregate exposure limit across all years. Users may also opt for risk silo-specific catastrophic excess-of-loss supplements, as well as requesting automatic or optional reinstatement.
At the more specialized end of the multi-line product offering spectrum are coverage programs sometimes known as “twinpacks” that only bundle two related risks. A popular such product was the joint offering by Cigna and XL Capital. The Cigna/XL twinpack covers high-layer property and casualty losses. At the other extreme are “integrated risk management” (IRM) programs that have literally claimed to provide earnings per share protection to their buyers. In the case of AIG’s Commodity-Embedded Insurance (“COINsm”), for example, the insurer’s objective is to provide a product that delivers total EPS protection by including essentially all the major risk exposures that a firm might face. AIG’s STORMsm program is a similar EPS insurance structure with a bias toward helping firms manage adverse weather-related events.
4.2 A Mixed Record
Some multi-line programs have been very successful, whereas others have been dismal failures. In some cases, multi-line products marketed by large and reputable reinsurance firms were never bought and subsequently taken off the market entirely, whereas in other cases the failures involved the actual dismantling of multiline programs by their buyers. These failures have led many to question the viability of multi-line policies.
Practitioners, commentators, and even providers of multi-line products have given several reasons for the failures and successes of multi-line products to date. Perhaps the most often cited reason for failure is that many extremely comprehensive multi-line programs did not achieve the cost savings they promised. Specifically, (re-)insurers rarely retain 100% of the risk exposures they provide, especially when financial risk is included in the picture. Consequently, (re-)insurers offering IRM protection are still faced with hedging, reinsuring, or retroceding the risks they are not prepared to retain. Although integrating imperfectly correlated risks in the same policy allows the firm to charge a lower premium in principle, that premium will result in a loss for the (re-)insurer when the costs of hedging and reinsurance or retrocession are taken into account. And if the (re-)insurer passes on those costs, so much for cost savings.
In other words, many multi-line products allow a (re-)insurer to offer an integrated solution but in turn merely push the unbundling problem back one level. A highly publicized illustration of this problem was the placement of a multi-line solution with Honeywell that covered traditional insurance risks plus the foreign exchange risk facing the company. When Honeywell merged with Allied Signal, an assessment of the IRM program revealed that had Honeywell purchased separate
insurance policies and engaged in classical hedging solutions to address its foreign exchange risk, it would have ended up with a cheaper risk transfer solution. Accordingly, the program was terminated and dismantled. Mobil Oil also dismantled a multi-line program in 1999 for the same reasons. And Utah-based petrochemical company Huntsman claims this was the reason it opted not to buy the “Risk Solutions” product offered by XL and Cigna in the first place, claiming that its coverage with 30 different insurers was simply cheaper than the proposed combined policy. (Banham 2000)
Nevertheless, some multi-line policies do appear to have succeeded. Union Carbide recently renewed a major multi-line product, and both Mead Corp. and Sun Microsystems claim to have saved over 20% by consolidating their numerous risk transfer policies into a single structure. (GGFP 2000)
Apart from these examples of firms that claim to have achieved major cost reductions through multi-line programs, many successful multi-line structures to date have appealed to customers not because of the premium reduction they facilitated, but rather because they allow their customers to achieve a more “optimal risk coverage” through a much more customized enterprise-wide risk management solution. Twinpack programs that couple related risks like property and casualty are an example, as is the celebrated United Grain Growers multi-line, multi-trigger
5. Multi-Trigger Programs
All risk transfer products contain a first or single trigger – that is, whether or not the transaction is in-the-money. Returning to Figure 4, annual losses of less than $100 million across all covered risk types means that the first trigger of the program – losses in excess of the deductible – is inactive. The program does not pay out. But especially when the payout is of the kind that exposes the (re-)insurer to significant moral hazard problems, the use of a second trigger that is not based on the actual economic damage sustained by the customer can make significant sense.
By making a multi-line policy conditional on a trigger or event whose outcome the risk transfer purchaser cannot influence, the (re-)insurer can be comfortable that losses have not been deliberately caused or loss mitigation mechanisms underutilized. Importantly, second triggers of this sort affect the ability of the insured party to make a claim but do not generally affect the amount of the claim itself. Multi-trigger products thus manage to control moral hazard without exposing their users to significant basis risks.
Multi-trigger structures also tend to be cheaper – often significantly – simply because more conditions must be met in order for the policy to be drawn on. In
Alternative Risk Transfer
this sense, multi-trigger structures are very much like knock-in barrier options and have essentially the same cost-saving appeal to users.
One example of a highly successful multi-trigger IRM program was the United Grain Growers EPS insurance program mentioned earlier. In that program, the actual payment on the policy was based on losses actually sustained by the firm, but payment was triggered only after a decline in an industry index of grain prices that were correlated with but not unduly influenced by the customer’s own actions. (Harrington, Niehaus, and Risko 2002)
6. Structured Finance Solutions
For many years, structured finance was considered the domain of investment and commercial banks. Increasingly, structured finance is part of the ART world for several reasons. The extensive use of insurance products in the structured financing process, the increasing participation in structured finance by major (re)insurance participants, and the actual securitization of insurance risks all have contributed to the integration of classical structured finance into the ART community.
In addition, much of the structured finance realm also shares the important commonality with many ART forms of integrating risk management and financing decisions. Whereas an ART form may be considered non-traditional risk transfer with capital structure implications, structured finance is the other side of the same coin and includes non-traditional securities issuance and design with risk transfer implications.
Structured financing arrangements typically come in one of four forms, each of which is briefly discussed below.
6.1 Asset Securitization
In a traditional asset securitization, a firm sells or conveys assets to an independent special purpose entity (SPE), and those assets are then used by the SPE as collateral to back the issuance of new securities. In some cases, the cash flows on the original assets are simply passed through to the owners of the new securities, whereas in other cases the SPE repackages the cash flows into new risk/return bundles.
A typical asset securitization is depicted in Figure 5 for a loan portfolio. The loans are sold or conveyed to an independently owned and controlled SPE, where they are deposited with a trustee on behalf of investors in the securities issued by the SPE. Those securities pay interest and principal based on the interest and principal received on the loan portfolio.11
11 A typical securitization may also include a swap designed to exchange a stream of irregular interest and cash flows for a more stable stream of interest to service the new securi-
Figure 5. A typical asset securitization involving a loan portfolio.
Traditional asset securitization has led to the development of several major markets, including the asset-backed securities market (including securities backed by assets like credit card receivables, capital leases, auto loans, and the like), the mortgage-backed securities market (both residential and commercial), and other specialized markets like asset-backed commercial paper conduits. The principal objective of traditional asset securitization is generally to facilitate an asset divestiture.
Using securitization to sell off assets is often driven by a more fundamental corporate finance and risk management objective, such as increasing debt capacity. A typical mortgage banker, for example, originates four to five times the amount of loans that are actually on its balance sheet. By selling loans that are then turned into mortgage-backed securities, the mortgage banker can focus on the fees from origination and servicing without having an overblown balance sheet.
6.2 Risk Securitization
In a risk securitization, a firm does not engage in the sale of assets, but instead creates a financing structure designed to promote risk transfer directly to participants in the capital market. Unlike asset securitizations, risk securitizations generally are not a source of fund-raising for the original asset owner.
A typical risk securitization is illustrated in Figure 6, where a firm enters into a risk transfer agreement with an independent SPE to cover some specific risk event. The SPE issues securities that pay an unusually high interest rate if the specified risk event does not occur. If the risk event does occur, the SPE uses the proceeds from the sale of securities, the investment income on those securities, and the fee paid by the original firm to compensate the original firm for its loss. In that case, security holders may receive little or no interest and/or principal.
ties. Such swaps are purely for liquidity enhancement purposes and do not absorb credit risk. In addition, an external guarantor may be engaged to provide credit enhancement.
12 A retrocession is simply the cession of risk by a reinsurer to another reinsurer – i.e., insurance purchased by a reinsurer.
Alternative Risk Transfer
If a bank wishes to manage the default risk of a loan portfolio but does not necessarily wish to sell or securitize the loans, it can engage in a risk securitization. Figure 6 illustrates. Instead of selling the assets in exchange for cash, the bank retains the assets and enters into a credit default swap (CDS) with an independent SPE. Proceeds from a security issue plus fees on the CDS are deposited in trust and invested in low-risk securities. As long as no loans default, the investment income on the securities plus the CDS fees are paid to the security holders in the form of a high coupon. In the event of default, interest and/or principal on the securities can be withheld to reimburse the bank. The bank has thus achieved the same risk management result as in Figure 5 but has neither sold its assets nor raised any funds.
Figure 6. A typical risk securitization involving a loan portfolio.
Risk securitizations are frequently undertaken by insurance companies and reinsurance companies when reinsurance and retrocession 12 coverage is either not available or perceived to be too expensive. In these cases, the SPE is a licensed (re-)insurance company that collects premium from the (re-)insurance firm in return for providing reinsurance or retrocession cover. Claims are paid first out of the investment income of the SPE – based as before on premium collected and the proceeds of insurance-linked securities sold to investors – and then by liquidating the portfolio of high-quality assets purchased using the proceeds from the insurance-linked note issue. Purchasers of insurance-linked notes receive a very high coupon in exchange for bearing the risk that the SPE must withhold interest and principal to finance unusually large (re-)insurance claims. Risk securitizations of this kind have involved a wide range of insurance – catastrophic property, life, trade credit, mortgage default, and the like.
6.3 Future Flow Securitization
Both asset securitization and risk securitization is often undertaken with the principal objective of risk transfer. Yet a third type of structured finance transaction – a future flows securitization – is intended instead primarily as a mechanism
13 Culp (2002a) discusses one of the more controversial examples of this, known as the “Hollywood Funding” debacle.
for raising funds. The structure is essentially the same as an asset securitization, except that the “asset” is still relatively intangible at the time it is sold for cash. The cash raised from the sale is used by the original owner to turn the asset from intangible into tangible.
A few years ago, for example, Hollywood film studios running over-budget on movie production used future flows securitizations to help fund the completion of their films. They essentially sold the rights to the revenues that the films were expected to generate once they were completed, and used the proceeds from the sale of those rights to investors in order to raise the money required complete the films.13 Future flows securitization have also been used extensively in project finance – e.g., securitizing the revenues from uncompleted oil or gas fields to raise funds that are then used to bring the fields into production. (Culp and Kavanagh 2003; Culp 2004)
6.4 Structured Liabilities
A final category of structured finance that can help corporations manage their risks is structured liabilities. Structured liabilities are any corporate securities whose features are deliberately re-engineered to facilitate some form of risk transfer either between the issuer and investors or across different types of investors.
Most hybrid securities, for example, are intended to help investors manage their risks from investing in the firm. Convertible bonds protect debt holders against the risk that the firm pursues actions designed to increase the value of its shares by reducing the value of its outstanding debt. Callable and puttable bonds are intended to accomplish a similar result. Numerous such hybrid and convertible structures are available today – see Coxe (2000) and Kat (2002).
Structured notes are also examples of how firms can embed risk transfer solutions into corporate securities. Essentially the combination of straight debt and derivatives, structured notes have long been used by firms to help manage their interest rate, equity price, foreign exchange, and commodity price risks. Consider, for example, the 10-year notes issued by Magma Copper Company in 1988. Instead of paying fixed coupons to investors, the “interest” on the Magma bonds was based on the price of copper. Each quarter investors received a payment based on the average price of copper over the previous quarter less a pre-defined fixed strike price. The Magma bonds thus were equivalent to traditional 10-year bonds plus a series of call options on copper. (Culp and Mackay 1997) As a copper mining firm, including call options on copper allowed Magma to manage its funding risk – its “interest service burden” would rise only when its revenues were also rising.
Alternative Risk Transfer
7. Contingent Capital
Although all ART products have some capital structure implications for their purchasers, perhaps the most obvious example of the integration between financing and risk management solutions is the ART form known as contingent capital. Whereas “paid-in capital” is a claim in which investors pay cash now in exchange for a proportional claim on the net cash flows of the firm beginning on the issue date of the claim, contingent capital, by contrast, is a right (but not obligation) to issue paid-in capital later. In other words, contingent capital is essentially a type of put option on paid-in capital or traditional securities.
Contingent capital products can be distinguished from outright options on corporate securities primarily through their reliance on a second trigger. As in the second triggers typical of the multi-trigger programs discussed in Section 5, the second trigger of a contingent capital facility is usually defined in terms of some clearly specified risk or event that is beyond the control of the purchaser of the facility.
Most insurance structures involve a reimbursement to purchasers for actual economic damages sustained. But because a direct reimbursement causes moral hazard, insurance also includes various additional contractual terms like deductibles, coinsurance, limits, and policy exclusions. Derivatives contracts, by contrast, typically include no such additional features because their payoffs are tied to a market price beyond the control of the derivatives user. For the same reason, however, derivatives users are subject to the “basis risk” that payoffs on derivatives hedges are not perfectly correlated with the exposure being hedged. The tradeoff between derivatives and insurance thus is between moral hazard and basis risk, and contingent capital is often deliberately designed to minimize both.
Contingent capital can involve a wide spectrum of structures, many of which are discussed in more detail in Culp (2002b, 2002c). Perhaps the most basic such structure resembles a put option that gives the purchasing corporation the right to issue new securities (usually equity or highly subordinated debt) at a pre-specified price in the event that a triggering event occurs. Such structures have been issued with a wide range of second triggering events, some based on variables under the capital purchaser’s control and others on proxies.
Swiss Re, for example, offers a contingent capital facility called Contingent Long-Term Capital Solutions (CLOCSTM ). Swiss Re concluded in December 2001, for example, a CLOCS transaction with MBIA Insurance Corporation (MBIA), a monoline insurance company that specializes in providing credit insurance or “wraps” for bonds that guarantee their timely payment of interest and principal. The Swiss Re CLOCS provide MBIA with US$150 million in subordinated debt that converts to perpetual preferred stock over time in the event that MBIA sustains significant losses on its existing guarantees. MBIA thus has access to additional capital on pre-loss terms after taking a major hit on its guarantee business.
Similarly, a “CatEPut” is a contingent capital facility (exercisable into preferred stock) where the second trigger is a natural disaster or a catastrophe-related prop-
erty insurance loss. Designed by the Chicago-based insurance broker Aon and offered by Centre Re, CatEPuts have been bought mainly by reinsurance companies with catastrophic property exposures either seeking excess reinsurance capacity or an alternative to “hardening” primary reinsurance and retrocession markets.
Perhaps the most obvious benefit from the growth in the volume and diversity of alternative risk transfer products is the much wider range of solutions now available to corporate risk managers. The expansion of the ART universe has also increased competition amongst providers of risk finance and risk transfer products, all the while still fostering greater cooperation amongst derivatives and insurance participants seeking to develop products that represent the best of both worlds.
Perhaps the most exciting impact of ART has been the steadily increasing number of corporate treasurers who now profess to be “thinking about” corporate finance and risk management as two sides of essentially the same coin. Capital structure and capital efficiency are indeed at the root of both financing and risk management decisions, and as more and more treasurers and CFOs come to realize this and to see how ART products can be used for capital management, the appeal of ART products likely will continue to grow.