Loss development factors: a primer for captives

There is a widespread and growing interest in the use of captive insurance as a tool for risk management, transfer, and financing. Naturally, this rising interest is accompanied by an influx of corporate decision-makers who may have little or no experience in the realm of insurance, especially with regard to risk retention issues.

While the SIGMA team has written in the past about the ways in which one might track loss development for captives, we believe an opportunity has presented itself to acquaint new entrants to the captive insurance landscape with one of the core concepts that must be understood when retaining risk at any level: loss development factors.

Loss development is a term often used by those in the insurance industry to describe the high-level activity of claim experience that occurs over time. The majority of claims in property casualty insurance tend to have a period of time in which their associated costs “develop” until they reach their ultimate value at closure. This timeframe can vary significantly depending on the risk being observed—property claims are typically reserved and paid over a relatively short amount of time, whereas workers’ compensation claims might take months or years to reach their ultimate value.

This high-level change is quantified through loss development factors, which measure the aggregate improvement or deterioration in groups of claims. Usually, the claims being measured are grouped based on the specific risk being analysed and the policy period in which they occurred or were reported. These factors drive a large portion of actuarial analytics, and the amounts used in a specific report might be based on industry benchmark data or claims experience unique to the entity being analysed. The ultimate values (or ultimate loss estimates) they produce usually drive the results of loss projections and estimates of required reserves, two of the most frequently used components of an actuarial report.

It is difficult to overstate the importance of loss development factors and their impact on numerous aspects of retained risk, whether that is through more traditional insurance policies or captive structures. As such, the remainder of this article will hopefully serve as a primer for those unfamiliar with their use by answering several common questions about loss development factors.

Question 1: How are loss development factors calculated and used?

As mentioned above, loss development factors are used to project the additional cost expected on claims associated with occurrence or reporting periods. While these factors quantify the late developing aspects of certain losses, they also account for losses that occurred during the period but are not reported until a later date. Both possibilities are often combined into a singular concept known as incurred but not reported (IBNR).

The first step of calculating a loss development factor is the construction of a loss development triangle, which organises loss data based on an associated policy period and a specific point in time, often referred to as an “evaluation date”. The next steps involve several mathematical calculations that measure the change in total claim costs for each policy period from one evaluation date to the next. While a detailed description of these calculations is outside the scope of this high-level article, numerous educational resources can be used to review this process.

Once calculated, loss development factors are applied elsewhere in an actuarial report to estimate the ultimate loss value of claims in specific policy periods. In recognition of the uncertainty inherent in such calculations, multiple “development methods” (and in many situations, other actuarial methods) are used to produce a range of estimates of ultimate loss values based on the various aspects of insurance claims (payments, reserves, etc). Consider below an example of a loss development factor being used to estimate ultimate losses in a period using incurred losses.

01/01/17-18 Incurred losses =	$1,000,000
Incurred loss development factor =	x 1.100
Estimated ultimate losses =	$1,100,000

As a final note, consider that the process of calculating development factors isn’t limited exclusively to loss values, as they are often used to analyse claim counts and other quantifiable amounts as well.

Question 2: How can a captive develop unique loss development factors or find industry benchmark factors?

If a captive has sufficient historical data, the loss development triangles referenced above can be constructed allowing the calculation of loss development factors which are unique to the insured entity (or entities). Theoretically, the use of unique factors as opposed to industry averages produces a more accurate calculation of estimated ultimate losses.

Doing so requires multiple loss runs at evaluation dates in equally cadenced intervals, such as 12/31/17, 12/31/18, 12/31/19, etc.

Industry benchmark factors can be determined based on published loss development triangles from various organisations or rating bureaus. Often, this information is available only through a fee or subscription basis. Actuarial and other types of firms which have access to this information might use it as part of their own analytics, and this access (or lack thereof) could be an important piece in the decision of which third-party firms a captive partners with.

Question 3: What special considerations do captives have regarding loss development factors?

Most often, the following considerations should be reviewed for any entity seeking to enter the captive insurance space or better understand their existing captive:

New captives should consider building loss development triangles in the time period shortly following captive inception. This year-to-year process can be much more efficient than trying to construct unique loss development triangles at a later date.
Accounting Standards Update disclosures require some triangles and payment development patterns to be accumulated each year. The paid loss development disclosures, for example, should produce paid loss development information which could then be reused in other captive analytics.
If a captive owner decides to add an insured entity, change the retention level for specific risks, or make other changes impacting the captive’s risk portfolio, the effects on loss development should be reviewed closely. This process can be done as part of or separately from standard actuarial reporting to ensure the actuarial team is involved in the early stages of understanding the potential impact.

Question 4: What are some common mistakes made when using loss development factors?

In our history as an actuarial consulting firm, we have seen several mistakes made by those reviewing or using loss development factors. The most common of these are:

Using a factor for the wrong coverage. Loss development factors for workers’ compensation, for example, are very different from those used to analyse automobile liability or products liability, and using these factors interchangeably produces significant risk of over or under-estimating ultimate losses. For workers’ compensation, factors may even differ significantly by state. If unique, credible data is available, loss development factors by coverage should be developed, but in lieu of such data, an appropriate industry factor for that coverage may be used.
Using a factor on the wrong type of loss. Paid loss development factors are often very different from incurred factors. More specifically, paid factors tend to be much larger, as payments almost always lag the reserving changes impacting incurred losses. Industry loss development factors based on paid losses and incurred losses are available and should be applied only to the appropriate type of loss.
Selecting a loss development factor that does not properly reflect the coverage trigger. This type of mistake might involve the use of a claims-made loss development factor when the coverage being analysed is occurrence-based.
Applying loss development factors to individual claims. The construction of loss development factors relies on the law of large numbers to produce reasonable estimates. That being the case, applying such factors to a single claim is not in line with the intended scope of their use.

Question 5: How does my captive’s retention change affect my historical loss development patterns?

We generally recommend that, as a starting point, loss development triangles should be constructed on an unlimited (or ground-up) basis. While there are certainly situations where triangles at specific retention levels should be considered and used, having unlimited triangles allows for a base line of unadjusted loss summaries through time. In general, losses at lower retention levels tend to have more predictable loss development patterns, relatively lower development at each evaluation, and a smaller tail.

Question 6: How do loss development factors impact the analysis of emerging risks?

Many emerging risks are of a low-frequency, high-severity nature. In these situations, loss development factors and the associated loss development methods may not be a reasonable approach for estimating ultimate losses. This is largely because of the difficulty in constructing credible loss development triangles, as losses may be highly volatile and claim history can be limited.

Another issue is that benchmark factors may not yet be available for emerging risks, as the types of firms which gather and analyse this data on a widespread scale don’t have enough information to produce reasonable results. As a result, other actuarial methods which do not rely on development patterns are normally relied upon.

Current examples of emerging risks matching this profile include cyber risk or breach liability. A company facing such risks may expect to experience a claim once every five to 10 years, and the associated severity of an event could easily be in the millions of dollars. In addition, there may be a significant lag in loss-producing events between their occurrence date, the date at which they become known, and the date at which they are eventually reported. Regardless of the impact on loss development, it is always worth noting that coverage triggers for emerging risk policies should be carefully reviewed prior to captive placement.

Conclusion

The concept of loss development is certainly more nuanced than can be covered in a single article, but a high-level comprehension should always be the objective of those involved in captive insurance. Its reach and impact are simply too great to be ignored or misunderstood, especially by those tasked with making large-scale decisions impacting their organisation.

That said, it is important to keep in mind that loss development is only one of a number of concepts surrounding actuarial analytics, and a focus should be maintained on the numerous ways in which actuaries and their work can help captive owners better understand their captive and accomplish their long-term goals relating to risk and finance.

Actuaries are always happy to help those using our reports get the most value possible from their contents, so potential captive owners new to this exciting industry should never hesitate to reach out for assistance.

Enoch Starnes is an actuarial consultant at SIGMA Actuarial Consulting Group.