Businesses entering the extended warranty field tend to think of extended warranties as large moneymakers because the repair costs appear small compared to the actual sales or premium—the amount paid by the consumer for the warranty. However, extended warranties can be a money-losing proposition if management does not maximize the use of all possible resources, including actuaries. An actuarial perspective has been used successfully in certain lines of business with extensive warranty exposure, such as the auto industry, and is available to any company that is either considering a new warranty business or looking to rectify an unprofitable warranty line.

Before getting into the core components of the actuarial approach, it may be important to define some terms. There are two main types of contracts often referred to as "warranties": a manufacturer’s expressed warranty from an original equipment manufacturer (OEM), sometimes referred to as an "OEM warranty," and an extended service contract (ESC), often referred to as an "extended warranty." We will use the term "warranty: throughout this article to encompass both types unless otherwise specified.

Both types of warranty should have the same broad underlying goal: to maximize profit. In general, companies selling extended-service contracts tend to focus on top-line revenue with little regard for the bottom-line results. Extended service contracts can be and are an important profit source for many retailers. OEM warranties, on the other hand, can be a significant cost to a manufacturer, hindering profitability. The actuarial point of view aids in program design by providing management with unique insight into the company's warranty costs.

The actuarial approach

How can actuaries help? At the outset, an actuary can advise on how best to manage a particular risk. For example, should the company pursue typical insurance or self-insurance? The situation for each company is unique. In general, typical insurance is less complex for the insured. By reducing complexity, the insured forfeits some profit, tax, and cash-flow advantages that it could receive from a self-insurance arrangement, such as a captive.

When the company has determined what vehicle to use to manage its warranty risk, it can turn to actuaries for specific insight into the financial mechanisms of that risk, including:

• estimating the expected costs associated with a new warranty
• projecting expected payments by year
• estimating the necessary reserves for outstanding claims in order to determine booked liability or buyout cost
• estimating the cost of goodwill payments
• calculating earnings patterns to properly align revenue and expected losses
• pinpointing poorly performing products earlier
• helping in the design of warranty coverage and length
• helping management in risk-financing decisions

Actuaries also bring a valuable insurance perspective to a warranty. Commercial codes allow manufacturers and retailers to issue warranties or service contracts without regulation. In this manner, the warranty business is largely unregulated, and in many states warranties are not regarded as insurance; thus, there is no requirement that an actuary review warranty exposures. That is not to say that companies should not strengthen their warranty business with insurance principles. An actuary can infuse a warranty with sounder risk-management principles, providing the company with a competitive advantage.

How does the actuarial approach differ from a mechanical statistical approach?

Statisticians typically use a Weibull distribution to predict warranty costs. The Weibull distribution models failure rates and is particularly good at modeling breakdowns over time, and thus is a go-to method for many industries. But there is a problem with using a Weibull distribution to model warranty behavior, because such an analysis assumes that warranty claims correlate only to product failures. This approach doesn't take into account consumer behavior and other external factors.

By contrast, an actuary will use a loss triangle, which uses historical data to track development over time of a group of policies (or warranties). This allows a better consideration of behavioral and other external factors. Thus the actuarial approach allows the use of both quantitative and qualitative information. We have identified some of the qualitative considerations below.

Adverse selection - Adverse selection (also called "antiselection" or "negative selection") is the tendency of people with a significant loss potential to buy insurance (or warranties). Information asymmetry—where one party, the consumer in this case, has much more information about possible loss potential than the other party—factors in to adverse selection.

For example, assume there are two types of consumers, high and low risks. If a service agreement is priced to cover the average cost of repairs for both high and low risks, more high risks than low risks will buy the service agreement, resulting in inadequate premiums. The price will then go up to reflect greater-than-average costs, compounding the problem by causing the low risks to drop out of the market, leaving only the high risks to buy these service agreements and resulting in a vicious cycle.

Providers must take care when designing their program(s) to reduce the chances of adverse selection. Adverse selection can be avoided by:

• Incorporating deductibles, so that consumers share in the loss. For example, a $50 deductible on a cell phone will motivate consumers to be more careful with their equipment.
• Implementing a time limit before coverage begins. If the consumer must wait for, say, 60 days until coverage begins, he will be less likely to purchase the extended service contract, knowing he will immediately abuse the product.
• Selling the extended service contract at the point of sale. This approach ensures that the extended service contract is purchased at the time when the equipment is whole and functioning. Such a provision is not unlike the preexisting condition provision often attached to individual health insurance policies.
• Increasing knowledge of both product and consumer. Adverse selection emerges from information asymmetry. Does the consumer have better information than the seller? For example, a snowboard company with an extended warranty that didn’t know its product was popular in terrain parks would face some serious warranty implications as snowboards came back damaged. (It would have to be said that a snowboard company that offered an extended warranty without very clear damage provisions would be a textbook example of knowing neither its product nor its consumers!)

Goodwill - In order to maintain a good relationship with its customers (goodwill), a company may continue coverage after the warranty has expired or cover causes not indemnified by the warranty. Goodwill can include any payments outside of the contract term or coverage. This is a common practice and can be an expensive one, accounting for between 5% and 20% of total costs, based on our historical analysis.

Goodwill costs can be contained by keeping comprehensive data to track warranty-policy inception, including information about the date of purchase. Many companies provide warranties before determining how to administer the returns. When a customer submits a claim without date of purchase (having thrown away the sales receipt), insisting that the product was bought within the warranty period, the selling company may feel obliged to honor the return in order to maintain goodwill with the consumer.

Seasonality - When estimating warranty costs, which are usually based on monthly or quarterly data, one must take into account periodic fluctuations due to seasonal patterns — lawnmowers, for instance, are more likely to have a claim in spring or summer than in fall or winter.

The geography of exposure may increase complexity. Summer temperatures last longer in Texas than in Illinois, so lawnmowers are likely to be used more often and over a longer period of time in Texas. We would expect a higher frequency of claims in Texas, with those claims occurring more evenly throughout a greater portion of the year.

Product mix and underlying warranty -  Triangular methods require consistency. If a product mix changes over time, development will also change. Development of the current product mix going forward must be adjusted from the historical data.

Similarly, if the underlying manufacturer's warranty changes, the actuary will have to make an adjustment to the analysis. An extended service contract may overlap the manufacturer's warranty. For this reason, the extended service contract may experience no losses for the initial period, but if the underlying warranty changes, the extended service contract will gain or lose additional exposure to loss.

Pipeline claims - Warranty exposures experience a lag between when a claim is incurred (or occurs), reported, recorded in the system, and finally paid. The loss triangles must be adjusted for these lags, so that they show "apples to apples" developments for each quarter. Sometimes, a claim may be reported in the warranty period, but not paid until after the warranty policy has expired.

Earning pattern - The earnings pattern for warranty exposures is typically based on an incurred (as opposed to paid or reported) basis. Once the triangles are adjusted for lag (see pipeline claims, above), the earnings pattern can be calculated from the development pattern. Earnings patterns can be based on actual payments, depending on the accounting used.

Premiums or fees must be earned to reflect special accounting rules for warranties. Traditional lines of insurance, such as workers' compensation, assume that losses occur evenly throughout the policy year (usually one year). But warranty losses do not occur evenly; they can occur for many years into the future, which means the premium is not earned evenly over the life of the policy. It is crucial to reflect the appropriate liability for warranty losses; a mismatch can show unprofitable warranties as profitable.

Unearned premium reserve versus loss reserve - There is a distinction when referring to the components of the total warranty liability. A loss reserve is for a claim that has already occurred (is incurred), but has not yet been paid. Traditional insurance products have a long lag between when a claim occurs and when it is paid. Thus most of the liability is a loss reserve.

A warranty policy, though, has a long lag between when the policy is sold and the claim occurs, but often has a short lag between occurrence and payment. A warranty claim is paid quickly after the loss occurs, usually within weeks. That means warranty liability is almost entirely an unearned premium reserve, with very little actually being a loss reserve.

Length of warranty - When a product breaks, consumers may wait until the warranty coverage is about to expire to file a claim. This can cause a slight increase in claims, or "bump" in development, in the loss triangles. The effect can vary, depending on length of warranty. A longer period has less effect; a shorter period has more.

click to enlarge

Changes in consumer behavior  - Claim development can be affected by consumer awareness, particularly when consumers suddenly become aware of their product's warranty (if the press has been reporting the product is defective, for example) or become aware of problems with the business holding the warranty (such as a potential bankruptcy).

The value of the warranty can affect behavior, as well. Consumers are more likely to pay attention to a warranty for an expensive item like a car than a warranty for a household gadget with a small repair (or replacement) cost.

Obsolescence - Some products, such as computers, become obsolete before the warranty coverage expires, so that even if the product breaks, filing a claim is not worthwhile for most consumers. Again, this affects the tail development or the emergence of claims for older products. With some products, the "bump" is offset by obsolescence.

Other common pitfalls

There are many other factors that can affect the success of a warranty program, such as:

• repeat breakdowns (which may need to be analyzed separately because of a leveraging effect on first returns)
• nondefective returns (which should be removed from the exposure base—e.g., sales units)
• trends (such as changes in frequency and severity of breakdowns, and inflation)

An actuarial approach to warranty management can help companies handle these risks. The ultimate goal of any warranty program, after all, is to maximize profit. The actuary can be the key to unlocking the profit potential.

Michael Paczolt is an actuary with the Chicago office of Milliman. He has expertise in property and casualty insurance, including loss reserving and ratemaking. He also has experience in commercial lines, including workers' compensation, professional liability, auto liability, general liability, and both manufacturer and extended warranty exposures.

Douglas K. Nishimura is a consultant with the Chicago office of Milliman. He has extensive experience in warranty contracts, workers' compensation, general liability, product liability, auto liability, directors and officers liability, and other commercial lines. His clients include commercial insurers, large corporations, healthcare institutions, and risk-retention groups.