Investigating captive mortgage reinsurance.

Lenders should do their homework before diving into reinsurance.

Banks and other mortgage lenders have recently begun participating more in the insurance of default risk on their originations. This interest can be attributed to several factors that relate to developments in both mortgage

Among the specific factors driving the trend are:

* Consolidation in banking and mortgage lending producing fewer and larger competitors that are more diverse and thus better suited to retain default risk and negotiate risk-sharing contracts with mortgage insurers;

* Mortgage insurance has recently been a profitable line of business; and

* Such arrangements move lenders further into the insurance industry in a coverage that is incidental to lending activities.

Captive reinsurance arrangements are becoming a popular vehicle for lenders to self-insure mortgage-insurance (MI) risk on mortgages they originate. In such an arrangement, the lender establishes a reinsurance company subsidiary (captive). The captive assumes MI risk written by a direct mortgage insurance company (direct writer) on loans originated by the lender. As consideration for the risk transfer, the direct writer cedes a portion of the MI premium to the captive.

As of late 1997, at least six national banks have received federal approval from the Office of the Comptroller of the Currency (OCC) to form a mortgage reinsurance subsidiary. Additional reinsurance subsidiaries have been established by mortgage lenders that are not subject to OCC oversight. As many as 50 or more of these companies may ultimately be formed by lenders and direct writers, according to Standard & Poor's (S&P), an agency responsible for rating the claims-paying ability of insurance companies.

If you are a sizable player in the mortgage lending market, the chances are good that these opportunities have attracted your attention. However, given the complexity of such arrangements and the variety of options available, the captive mortgage reinsurance arena should not be pursued without a carefully constructed strategy. Attention to the following eight considerations can help chart a course appropriate for a particular lender.

* Volatility of MI losses;

* Lender's appetite for risk;

* Performance of lender's loan portfolio;

* Risk profile of lender's loan portfolio;

* Reinsurance structures;

* Capital required;

* HUD compliance; and

* Reinsurance protection for the captive.

Each of these considerations is briefly discussed below.

Volatility of MI losses

During the 1990s, fueled by low losses and a strong economy, mortgage insurers' profits have soared. The five-year return on equity for the industry from 1992 to 1996 was 18.4 percent, according to S&P. During the same five-year period, the annual return on revenue for the industry peaked at 51 percent in 1996 and never fell below 25 percent.

However, the losses experienced by the MI industry during the 1980s are just as noteworthy as were its profits in the 1990s. Loss ratios represent a key measure of insurance underwriting results and are calculated by dividing incurred losses by earned premiums. Figure 1 displays a graph of the MI industry's calendar-year loss ratio for the 15-year period from 1980 to 1995. The industry saw losses rise sharply in the mid-1980s, peaking at a loss ratio of 192 percent in 1987. In other words, the industry incurred $1.92 of losses for every $1.00 of premium revenue in 1987. This period of heavy MI losses was largely a result of the boom and bust residential real estate market in the south central "oil patch" region of the United States.

Providing MI coverage is clearly a risky venture. Insurers set fixed premiums up front for coverage that frequently extends for seven to 10 years or more. Economic factors have a marked effect on mortgage default rates and therefore on MI losses. Lenders must be prepared for this risk if they intend to pursue a captive mortgage reinsurance arrangement.

Appetite for risk

Given the volatility associated with MI losses, it is critical that lenders assess their own appetite for risk before entering into a captive mortgage reinsurance arrangement. The large profits enjoyed by insurers in recent years will not continue indefinitely. MI margins compensate insurers for the risk associated with the coverage and allow for the accumulation of capital during the profitable cycles to establish a cushion for the high loss levels that can accompany adverse economic conditions. Lenders must be sure they are prepared to make a long-term commitment to the venture before spending the time it takes to establish the reinsurance subsidiary and negotiate the contract terms.

Many lenders have established subsidiaries to manage their expanding insurance services. Some already participate in the underwriting risk of other insurance coverages incidental to banking, such as credit life insurance and credit card unemployment coverage. MI reinsurance should be considered within the context of other insurance ventures being undertaken to determine the organization's appetite for the risk of reinsuring MI coverage. Lenders with a strong appetite for risk will welcome the opportunity to reinsure MI coverage and will prefer structures with larger reinsurance premium levels and correspondingly greater risk.

Performance of lender's loan portfolio

A lender should examine the past performance of its portfolio of high loan-to-value ratio (LTV) loans when considering a reinsurance arrangement. Lenders whose mortgage underwriting quality has exceeded that of their peers will likely be more eager to participate in insuring their future loan performance. Furthermore, these lenders will find the MI companies enthusiastic about discussing reinsurance arrangements with their valued customers.

The lender's mortgages can be compared against benchmarks such as the performance of the average loans insured by the MI companies currently guaranteeing the lender's portfolio. Lenders can request reports from their current insurers that examine the lender's delinquency and claim rates by year of loan origination relative to the insurer's aggregate results.

Claim rates represent the percentage of loan originations for a book-year that have resulted in a claim as of a certain evaluation date. Likewise, delinquency rates represent the percentage of book-year loan originations that are currently delinquent. Both statistics are routinely monitored by insurers. Generally, claim rates are the more accurate measure of the actual performance of each book-year of insured loans. However, the claim rate for recent book-years will typically offer little value because the majority of MI claims usually occur from three to seven years after loan origination.

Therefore, delinquency rates are used as a barometer of future claim activity because some of the loans that are currently delinquent will eventually result in a claim. The relationship between origination-year age and the typical pattern of claim activity makes it critical that comparisons be made at comparable stages of maturity (i.e., on a book-year basis with a common evaluation date).

Generally, the lower a lender's delinquency and claim rates relative to the insurer's averages, the more profitable that business is to the insurer. Before drawing any definite conclusions about the quality of its insured loans based on a delinquency and claim rate comparison, a lender must also consider other characteristics of the loans in its portfolio of originations. For example, a lender insuring a disproportionate share of 85 percent LTV loans relative to the insurer's total book will likely have a lower claim rate since higher LTV loans are riskier. However, this lender will not necessarily be a more profitable customer to the insurer because the premium rates charged for 85 percent LTV loans are lower than for higher LTV loans. This highlights the need to examine the risk profile of a lender's loan portfolio.

Risk profile of lender's loan portfolio

As a lender's mortgage origination volume increases, the portfolio becomes more diverse and the risk of insuring (and reinsuring) the portfolio decreases. MI companies insure the loans of many lenders in order to reduce risk through diversification. However, a lender's captive is restricted to reinsuring only the lender's mortgages. Therefore, lenders with larger and more diverse origination volume are better suited to accept a larger piece of the risk pie.

Lenders should examine their loan distribution by LTV and loan type to assess the diversity of this risk. Lenders with higher concentrations than the industry of adjustable rate mortgages (ARMs) or loans with LTVs greater than 95 percent represent a greater risk than a more balanced portfolio.

However, there is probably no factor more important to the diversification of a lender's MI risk than the geographical distribution of the lender's originations. Geographical diversification is so critical that regulators have placed limitations on insurers' concentrations within a given Standard Metropolitan Statistical Area (SMSA). The National Association of Insurance Commissioners' (NAIC) Mortgage Guaranty Insurance Model Act limits an insurer's concentration to 20 percent of its insurance in force in any single SMSA.

The MI industry took a beating in the 1980s largely as a result of a regional economic event (the residential real estate depression of the oil patch states). Insurers' ability to withstand the losses of this period depended on their national diversification because the profitability of business in other regions partially diluted the catastrophic losses in the south central region. Likewise, larger and more geographically diverse lenders will be better suited to assume higher levels of risk.

Reinsurance structures

Myriad different reinsurance arrangements can be structured to meet a lender's particular appetite for risk. Contracts are typically structured to include mortgages originated by the lender during a three- to five-year origination period. The reinsurer receives premium revenue and is responsible for reinsured losses for a runoff period typically lasting 10 years for each origination year. The reinsurer may also be responsible for a ceding commission to the direct insurer. Ceding commissions are typically used in reinsurance contracts to compensate the direct insurer for its expenses associated with the underwriting and administration of coverage as well as claim settlement costs. While some captive mortgage reinsurance contract specify a separate ceding commission, others include a reinsurance premium quote on a net of ceding commission basis.

Generally, reinsurance structures can be broadly classified into the following two varieties: quota share and excess of loss.

In a quota-share arrangement, the primary insurer and reinsurer share all losses and premium on a pro-rata basis according to the specified quota-share percentage. In an excess-of-loss arrangement, the reinsurer is responsible for all losses once the primary insurer's losses reach a specified level referred to as the attachment point. The reinsurer pays the primary insurer for all losses in excess of the attachment point up to the reinsurer's overall policy limit. No losses are reimbursed by the reinsurer if losses do not exceed the attachment point. As of late 97, most captive mortgage reinsurance arrangements have been on an excess-of-loss basis.

The corridor of losses reinsured by a lender can be defined in several ways. The primary insurer's direct loss ratio for loans subject to the contract can provide the basis for the reinsurer's layer. For instance, the reinsurer might cover losses exceeding 75 percent of the direct insurer's premium up to 110 percent of direct premium; (i.e., between direct loss ratios of 75 percent and 110 percent). Alternatively, the reinsured layer can be specified based on the direct risk insured by the primary insurer.

Regardless of how the reinsurer's layer of risk is specified, it is typically set at a level sufficiently higher than expected losses so that the reinsurer is expected to incur no losses in the majority of years. For example, the reinsurer may be expected to be loss-free for three out of four years of mortgage originations. However, the reinsurer's losses may be expected to consume the entire reinsured layer roughly 1 out of every four years. The one adverse origination year may produce losses up to four or five times as large as the reinsurance premium. In other words, the reinsurer is typically participating in a loss layer penetrated only in adverse loss cycles.

By contrast, a quota-share arrangement provides a reinsurer with a pro-rata share of risk that basically behaves identically to the direct risk insured by the mortgage insurer. The exposure covered by the direct insurer and reinsurer have the same risk profile just in different sizes reflecting the quota-share percentage. Unlike excess-of-loss participation, the reinsurer participates in all insured layers, including those associated with adverse underwriting cycles and layers of expected loss levels.

This feature may be particularly appealing if the lender believes average loss levels can be managed through mortgage lending underwriting standards but that catastrophic loss levels are virtually uncontrollable due to economic forces outside the lender's control. Such a lender would likely want to participate in the more manageable layers of loss included in a quota-share agreement, and possibly purchase aggregate excess insurance for the captive's exposure in the catastrophic claim layers. Quota-share arrangements are relatively new and less common than excess-of-loss arrangements. The appropriate maximum allowable quota-share level reinsured by lenders is a hot topic of debate by regulators. Vermont, which regulates many captives including several mortgage reinsurers domiciled in the state, has recently indicated that it may permit arrangements where the quota share is 25 percent or lower. The insurance commissioners for the states of North Carolina and Wisconsin have recently taken a similar view. However, the OCC has given banks approval to reinsure up to quota share levels of 50 percent. The OCC has indicated it would separately consider any banks seeking quota-share arrangements of more than that percent.

Capital required

Lenders must be prepared to contribute capital to the captive to support the risk of reinsuring a coverage as volatile as mortgage insurance. The capital must be committed to the reinsurer on a long-term basis due to the lengthy runoff period associated with the exposure. While minimum capital levels vary by state of domicile, statutory minimum capitalization for a Vermont captive is $250,000. However, lenders must be willing to contribute additional capital to provide a cushion for adverse years when losses exceed premiums.

At a minimum, the NAIC model act specifies that mortgage insurers are required to maintain capital so that aggregate insured liability (i.e., risk) does not exceed a factor of 25 times the insurer's capital. Risk is defined as coverage on all insured mortgages currently in force. For example, required capital associated with $1 billion of insured loans in force would be approximately $10 million. A lender with a 25 percent quota-share reinsurance contract on these loans would need at least $2.5 million to support this risk (i.e., 25 percent of $10 million).

Capital requirements for captive mortgage reinsurers tend to be more strict than the 25-to-1 standard for the following reasons:

* Lender captives are reinsuring a less geographically diverse portfolio than the aggregate insurance written by primary insurers. The additional risk associated with reinsuring this portfolio requires additional capital;

* Mortgage insurers are typically capitalized above the minimum level (i.e., at a ratio at or below 20 to 1). Additional capital is required to maintain a sufficient financial strength rating to be acceptable primary insurance providers on mortgages pooled by Fannie Mae and Freddie Mac. Primary insurers may similarly require their reinsurers to be sufficiently capitalized so that their rating is not jeopardized by potential insecurity of reinsurance collectibility;

* Lender captives typically reinsure on an excess-of-loss basis. As mentioned earlier, the reinsured layer tends to be above expected losses in the more volatile excess layers. The additional risk associated with such layers of coverage may require additional capital.

Mortgage insurance is a capital-intensive business. However, a portion of the capital required of the reinsurer may be met through sources other than cash, such as a letter of credit. Furthermore, during profitable years, capital will be generated from the reinsurance operations through the accumulation of retained earnings and a contingency reserve.

Mortgage insurers are required to establish a contingency reserve to cover potential loss. This reserve is also required of captive reinsurers. When computing an insurer's capital for purposes of required risk-to-capital thresholds, both the insurer's statutory surplus and its contingency reserve are included.

Under statutory insurance accounting, 50 cents of every mortgage insurance premium dollar must be set aside for 10 years in a contingency reserve. Reserve contributions cannot be released before the 11th year unless the insurer's losses exceed a threshold loss ratio of 35 percent (with state insurance commissioner approval). Net annual contributions to the contingency reserve are tax deductible as long as the deferred tax (which will be earned as revenue upon release in year eleven) is funded with noninterest-bearing tax and loss bonds.

The contingency reserve and capital requirements emphasize the long-term commitment required to reinsure mortgage insurance risk.

HUD compliance

A lender will want to be comfortable that its reinsurance arrangement does not violate section 8 of the Real Estate Settlement Procedures Act (RESPA). On August 6, 1997, the U.S. Department of Housing and Urban Development (HUD) issued a letter clarifying the applicability of RESPA to captive reinsurance arrangements. HUD concluded that these arrangements are permissible "so long as payments for reinsurance... are solely payment for goods or facilities actually furnished or for services actually performed."

HUD outlines several factors which will cause additional scrutiny to be given to a captive-reinsurance arrangement and HUD presents the following two-part test to determine if a violation exists.

Test 1. The arrangement meets three requirements that establish that reinsurance is actually being provided; and

Test 2. The compensation paid for the reinsurance shall not exceed the value of the reinsurance.

The factors leading to additional scrutiny and HUD's two-part test are both discussed in detail in the December 1997 Mortgage Banking article "Being Held Captive by HUD." As noted in that article, the most difficult criteria that must be satisfied to establish that reinsurance is being provided (Test I) is that there must be real transfer of risk.

HUD acknowledges that the transfer of risk requirement is clearly satisfied by a quota-share arrangement but states that the transfer of risk requirement can be met in the case of an excess-loss arrangement "if the band of the reinsurer's potential exposure is such that a reasonable business justification would motivate a decision to reinsure that band." Therefore, excess arrangements must be scrutinized more closely to ensure that no RESPA violation exists.

Based on the guidelines outlined by HUD, lenders must be comfortable that their captive reinsurance arrangements do not violate RESPA.

Reinsurance protection for the lender captive

While the notion of a reinsurer purchasing reinsurance of its own may initially seem strange, it is a common practice in other lines of insurance. Known as retrocessions, such coverage allows the reinsurer to assume more risk for a given level of capital.

Lenders may want to consider purchasing reinsurance protection to limit the risk reinsured by its captive, particularly if the lender is pursuing a quota-share arrangement. For example, a lender may favor a quota-share arrangement due to:

* Its definite transfer of risk and the correspondingly stronger case against a RESPA violation; and

* The inclusion of the more predictable and manageable loss layer in the risk reinsured by the lender.

However, the reinsurer may desire reinsurance protection in order to:

* Protect the lender against catastrophic loss scenarios that present a greater risk to lenders with less geographic diversification;

* Reduce the volatility of the financial performance of the captive; and

* Reduce the amount of capital required to support the risk reinsured by the captive.

There are several reinsurers based in the United States and elsewhere (some of whom have served primary mortgage insurers in the past), that represent a third-party option for retrocessional protection to a lender captive. As an unrelated third party to the transactions between the lender and the primary insurer, such a reinsurer could provide protection to the captive while preserving the clean RESPA status afforded by HUD to quota-share arrangements.

As mergers and acquisitions in banking and mortgage lending create larger and more diverse lenders, and as banks continue to increase their insurance operations, captive mortgage reinsurance is an idea whose time has come. However, given the nature of the risk, the complexity of the arrangements and the options available, lenders will want to do their homework before they plunge into the captive mortgage reinsurance waters.

Michael C. Schmitz is an associate actuary in the Milwaukee office of Milliman & Robertson. His areas of expertise include consulting to mortgage insurance companies and to lenders exploring captive mortgage reinsurance.