Section 5 - Losses Flashcards
What is the Overlap Fallacy in loss development and loss trend analysis?
The Overlap Fallacy is the false conclusion that loss development and loss trend are overlapping adjustments. Loss development brings immature losses to their expected ultimate level, while loss trend restates historical losses to the level expected for similar losses that will occur in the forecast period. Even in the absence of one, the other would still occur.
How do catastrophes and large losses differ in ratemaking?
- Catastrophes: Affect multiple policies, occur infrequently, and are usually defined based on a threshold for aggregate losses in an event (e.g., hurricanes, earthquakes).
- Large Losses: Impact a single policy and exceed an insurer-defined threshold (e.g., high-severity liability claims).
Adjustment Method:
- Catastrophes are replaced with a long-term catastrophe load.
- Large losses are capped at a threshold, and an excess factor is applied.
4 critical adjustments that must be made to losses during the ratemaking process
1) Remove extraordinary events (Large losses and CATs)
2) Restate losses to the coverage or benefit level expected during the future policy period
3) Develop losses to ultimate levels
4) Restate losses to the cost level expected during the future policy period with Loss trends
Explain why an indicated rate increase of 5% is not necessarily indicative of deteriorating experience
We are told that rates were adequate at the time of the rate change. Therefore, if experience does not get better or worse after the change, then experience should change with expected net trend.
example : Net trend = (claim trend)/(premium trend) – 1 = (1 + 0.04) / (1 + 0.01) – 1 = 2.97%
Time from the change to the effective date of the new rates = 1.5 years
Therefore, experience should change with respect to net trend = (1 + 0.0297)^1.5 – 1 = 4.5%
Since this is close to the rate change implemented at that time, this is as expected and does not indicate deteriorating experience.
What is the difference between loss development and loss trend?
Loss Development adjusts immature claims to their ultimate value.
Loss Trend adjusts all claims to reflect future economic and frequency/severity changes.
What are examples of factors that drive loss trend?
a. Inflation
b. Legal environment changes
c. Advancements in safety technology
d. Medical cost increases
e. Societal changes such as changes in court practices and legal precedents
What are the steps to determine large losses in ratemaking and what is their proper order?
- Cap historical losses at a selected threshold.
- Apply development and trend to capped losses.
- Add back trended, developed excess loss costs from external sources (e.g., ILFs or reinsurance).
This avoids skewing historical averages due to volatility of large losses.
When should you choose an exponential trend vs. a linear trend in ratemaking?
Use exponential when costs change at a percentage rate (e.g., inflation, medical costs), or when modeling claim frequency, which often trends downward — exponential form preserves the proportional decrease.
Use linear when changes occur at a constant dollar amount per period.
Why should severity and frequency be analyzed separately when trending?
Separating them provides clearer diagnostics. Total loss cost may rise due to either:
* An increase in average claim size (severity)
* An increase in number of claims (frequency)
Suppose overall losses increase by 10%. If you split it:
Severity rose by 15% (e.g., higher repair costs), Frequency dropped by 5% (e.g., safer cars). Looking only at combined losses would hide this insight, leading to misleading assumptions.
What are key considerations when selecting loss trend factors?
Stability of the data
Large losses removed
Length of time used in fitting the trend
Recent shifts in trend patterns
Credibility of the data
Type of trend (severity, frequency, or pure premium)
To select your loss trend, how many points should select for a short tail LOB vs a long tail LOB
For rapidly changing trend environments (ex. private passenger auto),
more recent data is often relied upon (such as an 8 point). For lines highly impacted by weather (ex. homeowners), long-term trends are often relied upon (such as a 20 point fit).
What is the issue and the solution of Using CY Data to Measure Loss Trend
CY data does not match claims with exposures, in your CY you can have a very old claim paid.
One solution for this issue is to use generalized linear models to measure trend. GLMs can account for changes in the size of the
portfolio as well as changes in the mix of business, leaving the true trend intact. (Often used in Commercial Lines )
How does applying a trend affect large losses differently from large losses capped? Include an example
Applying a trend increases all losses proportionally, but capped losses are only trended up to the cap. This means the impact of trend is smaller on capped losses.
