Quarterly rate & reserve review
Close the books on a real insurer's triangle: roll the reserves forward from the prior evaluation, test actual against expected emergence, re-estimate ultimate loss ratios and produce a credibility-weighted rate indication — then hand the decision to the Decision Studio.
Quarterly rate & reserve review
A period-end reserve roll-forward and rate indication on a real CAS Loss Reserve Database triangle. The database is annual (year-end evaluations); a live quarterly cycle applies the same roll-forward logic at each quarter-end instead.
Prior evaluation compared below: 1996.
Reserves — as of 1997 vs. prior (1996)
| AY | Age | Reported | CL ultimate | BF ultimate | Selected | IBNR | Δ ultimate vs prior | Hindsight ultimate |
|---|---|---|---|---|---|---|---|---|
| 1988 (CL) | 10 | 6,815,646 | 6,815,646 | 6,815,646 | 6,815,646 | 0 | 6,837 | 6,815,646 |
| 1989 (CL) | 9 | 7,712,077 | 7,719,821 | 7,718,254 | 7,719,821 | 7,744 | 2,365 | 7,721,911 |
| 1990 (CL) | 8 | 8,364,955 | 8,396,601 | 8,390,893 | 8,396,601 | 31,646 | 1,711 | 8,394,117 |
| 1991 (CL) | 7 | 8,215,810 | 8,288,545 | 8,281,125 | 8,288,545 | 72,735 | -2,292 | 8,288,143 |
| 1992 (CL) | 6 | 8,876,813 | 9,043,728 | 9,025,374 | 9,043,728 | 166,915 | -16,258 | 9,026,329 |
| 1993 (CL) | 5 | 9,337,099 | 9,702,726 | 9,661,140 | 9,702,726 | 365,627 | -40,611 | 9,673,610 |
| 1994 (CL) | 4 | 9,640,098 | 10,422,621 | 10,340,414 | 10,422,621 | 782,523 | -73,688 | 10,375,605 |
| 1995 (CL) | 3 | 9,006,113 | 10,571,471 | 10,475,533 | 10,571,471 | 1,565,358 | 30,967 | 10,512,108 |
| 1996 (BF) | 2 | 7,486,113 | 10,490,872 | 10,436,791 | 10,436,791 | 2,950,678 | -275,688 | 10,387,245 |
| 1997 (BF) | 1 | 4,344,144 | 10,933,658 | 10,662,535 | 10,662,535 | 6,318,391 | 10,662,535 | 10,165,481 |
| Total | — | 79,798,868 | — | — | 92,060,485 | 12,261,617 | — | — |
Total unpaid / IBNR
$12,261,617k
Mack std. error (total)
$550,736k
Mack CV
4.4%
Rate indication
| AY | Ultimate LR | Trend years | Trended LR |
|---|---|---|---|
| 1988 | 87.3% | 10.0 | 117.3% |
| 1989 | 88.1% | 9.0 | 114.9% |
| 1990 | 85.7% | 8.0 | 108.6% |
| 1991 | 78.2% | 7.0 | 96.2% |
| 1992 | 78.9% | 6.0 | 94.2% |
| 1993 | 79.3% | 5.0 | 91.9% |
| 1994 | 78.5% | 4.0 | 88.3% |
| 1995 | 74.8% | 3.0 | 81.8% |
| 1996 | 71.2% | 2.0 | 75.5% |
| 1997 | 71.4% | 1.0 | 73.6% |
3-yr weighted trended LR
87.1%
Indicated: 24.4%
5-yr weighted trended LR
90.0%
Indicated: 28.6%
Credibility Z (premium rule)
1.00
Complement: 28.1%
Credibility-weighted indication (3yr)
24.4%
Assumptions & limitations
- The CAS Loss Reserve Database records annual (year-end) diagonals only; each 'quarterly' evaluation here re-runs the same year-end roll-forward logic, standing in for what a quarterly cycle would repeat four times a year.
- Selected ultimate uses chain-ladder for accident years at least 3 years mature and Bornhuetter–Ferguson (a-priori = prior-evaluation weighted ultimate LR) for younger years.
- Full credibility is assumed at $250,000,000 of premium (a stated proxy for the classical 1,082-claim standard, since Schedule P carries no claim counts).
- The complement of credibility is the prior evaluation's own 3-year indicated change, trended forward one year; it is 0 when no prior evaluation exists.
- Tail development beyond age 120 months is not observed in Schedule P and is assumed to be 1.0.