Business case · scenario
What a behavioural rating factor could do to the kasko loss ratio
A transparent scenario on public TSB data. Start from the sector or a company, then replace any input with your own. Every formula is on this page.
Same period a year earlier (2024): 67.8% → 69.2%, +1.4 pts.
Loss ratio: all kasko, gross, excluding the earthquake pool (DERK), 2025 full year (TSB publishes no loss ratio by vehicle type). Premium: 2025 car kasko only, commercial vehicles and trucks excluded. Source: TSB.
Incurred claims / earned premium, gross
Denominator: car kasko policies (7.51 M in 2025), not all vehicles
Loss cost per exposure. Field experiment, CMT estimate from measured driving: −7.4% (frequency −4.4%)
Your pricing decision. Turkish precedent: 8.5–10% of the policy at renewal (Aksigorta, Fiat connected cars)
Cars only. Leave empty for percentages only
Programme segment loss ratio
64.1%
5.1 pts better
Portfolio loss ratio
68.7%
0.5 pts better
Behaviour effect at half / estimate / 1.5×
Half
68.9%
0.3 pts better
Estimate
68.7%
0.5 pts better
1.5×
68.4%
0.8 pts better
Segment LR = base × (1 − pure-premium drop) ÷ (1 − discount). Portfolio LR weights participants and the rest by exposure.
Why read this by segment
2.6%
of the 18.0 M cars on the road are electric (stock, end of Aug 2026, TÜİK)
18.1%
of the 601 k cars registered in Jan–Aug 2026 are electric (new cars only, TÜİK)
Two different measures, not a trend: 2.6% is the whole car park, 18.1% is this year's new cars. Even if every EV had kasko, EVs would be at most 6.2% of car kasko policies. So the portfolio moves by tenths of a point; the EV segment by about five, and it grows with every new-car month.
Evidence on loss ratio, and what each one measures
−7.4%
Loss cost, estimateUS randomised field experiment, 1,449 drivers with feedback against a control group. Measured: safer driving (hard braking −16–21%). CMT, the trial's telematics partner, estimates −7.4% loss cost and −4.4% claim frequency: a model estimate, not observed claims.
Accident Analysis & Prevention 225 (2026)≈−15 pts
Loss ratio, company levelItalian insurers with ≥50% black-box policies vs those without, 2019 (−10 pts in 2018). Correlation, not causation: selection is part of it.
IVASS, Bollettino IPER 2020 Q48.5–10%
Tariff discountAksigorta's Fiat Yol Arkadaşım Connect Kasko (since Oct 2020): a connected-car device sold with a subscription gives a driving score and days of use. At renewal, a score above 75 earns 8.5–10% off the policy, plus up to 250 TL by days of use. Stated reason: good drivers should not carry the average risk cost.
Aksigorta · AA, 2 Oct 2020↓
Loss ratio, early adoptersUBI lowers loss ratios, mainly among early adopters (selection plus behaviour).
North American Actuarial Journal, 2022
Where the starting data comes from
- Kasko gross loss ratio by company, 2025 and Jan–Jun 2026 (TSB 'Hasar Prim Oranları', all vehicle types)
- Car kasko premium and policies by company, 2025 (TSB 'Kasko Prim Adet', vehicle type 1-Otomobil)
- Registered cars 18.0 M, EV share 2.6%; Jan–Aug 2026 registrations 601 k, EV share 18.1% (TÜİK, 16 Sep 2026)
- Sector loss ratio 69.2% (2025) and the call for risk-based driver pricing (TSB 2025 sector report, p. 27 and p. 62)
Read before quoting
- A scenario, not a forecast. The behaviour effect is measured in the US.
- EV kasko is severity-heavy (parts, batteries, labour). Driving behaviour mostly moves frequency.
- Selection effect: participants may already be better risks. The pilot measures this with a control group.
- Seven cars are not credible. A rating factor needs enough vehicle-years of exposure.