Week 2 of Loss Ratio | Looking Beyond the Default
- Rika Taute
- Mar 26
- 3 min read
The Decision Question
Week 1 answered:
Are we performing above or below expectation?
The next natural question is:
What is driving the change in loss ratios?
Loss ratios can move because of:
claim frequency
claim severity
large losses
Without separating these drivers, pricing decisions risk targeting the wrong lever.
A loss ratio of 75% can mean very different things:
Stable, credible portfolio
Large exposure base, predictable development, stable frequency
→ 75% is information
Thin, volatile book
Low exposure, driven by a small number of large losses
→ 75% is noise
Post rate increase
Premium uplift with severity trending upward
→ 75% may be temporarily flattering
Offsetting trends
Deteriorating frequency, improving severity
→ Same loss ratio, but a different underlying risk signal
Suggested Chart - Frequency vs Severity View
A practical approach is to show frequency and severity trends alongside the loss ratio.
The very real complication in this approach is that we are dealing undeveloped claims - especially in the more recent periods. Showing frequency and severity (paid or incurred) over time might therefore not give the full picture. This is solved by showing these values per development period.
The first set of graphs shows a GI book that typically has lower severity claims and more stable loss ratios.
The loss ratio shows improvement from 2017 to the more recent 2025. With year 1-3 development, the loss ratio is expected to land around 45%.
This is driven by the decrease in claim frequency in the most recent years.
Severity does not show observable trends, but the most recent year is in the higher cluster.
This suggests underwriting or exposure improvement, not pricing change.
The second set of graphs shows a volatile GI book with very high severities.
The loss ratio again shows improvement from 2017 to the more recent 2025.
This is driven by the decrease in claim frequency in the most recent years.
Severity is highly volatile with significant development in year 2.
The loss ratio alone is not decision-ready here. Volatility in severity obscures the signal.
What this makes visible
This view separates the mechanics behind the loss ratio.
Instead of reacting to a single number, it decomposes a single outcome into its underlying drivers.
More importantly, it distinguishes between different types of deterioration:
Systematic (frequency trends)
Structural (severity / inflation)
Incidental (large losses)
Each driver points to a different action:
Frequency-driven deterioration
→ Review underwriting standards, exposure growth, or risk selection
Severity-driven deterioration
→ Review pricing adequacy, limits, deductibles, or inflation assumptions
Large loss-driven change
→ Separate signal from noise before acting
Offsetting trends
→ Investigate which driver is dominating and whether trends are temporary or structural
The loss ratio alone does not tell you which of these is happening.
A loss ratio is a result.
Different drivers can produce the same result - but require completely different responses.
Without separating them, decisions risk targeting the wrong lever.
Trade-offs & Risks
The main limitation is complexity.
Compared with the standard loss ratio chart:
it requires more explanation
it introduces additional metrics
it may overwhelm non-technical audiences
It also depends on stable exposure measurement.
If exposure changes materially, frequency comparisons can mislead.
Decision Boundary
A chart becomes decision-grade when it helps identify the driver - not just the outcome.
Use this when…. | Don’t use this when… |
|
|
Comments