Workers' compensation, retrospective and experience rating
hardhat aims to bring workers' compensation calculations into a single Python package by converting the algorithms from the (historical) CAS Exam 8 material into code. We will start with Table M construction (since it's one of the easier papers).
from hardhat import TableM
from hardhat.utils.utility_functions import load_sample
df = load_sample(key='brosius')
table_m = TableM(
data=df,
experience='Actual Loss',
index="Risk"
)
print(table_m)
Actual Loss Entry Ratio
Risk
1 20000 0.2
2 50000 0.5
3 60000 0.6
4 70000 0.7
5 80000 0.8
6 80000 0.8
7 90000 0.9
8 100000 1.0
9 150000 1.5
10 300000 3.0
print(table_m.phi(r=1.2))
0.21000000000000002table_m.lee_diagram().show()Brosius, J. Eric, "Table M Construction," CAS Exam Study Note, 2002, pp. 1-14.
Teng, Michael T.S., "Pricing Workers' Compensation Large Deductible and Excess Insurance," Casualty Actuarial Society Forum, Winter 1994, pp. 413-437.
Feldblum, Sholom, "Workers' Compensation Ratemaking," CAS Exam Study Note, 1993.
Fisher, Ginda Kaplan; McTaggart, Lawrence; Petker, Jill; Pettingell, Rebecca, "Individual Risk Rating," CAS Study Note, 2019.
Gillam, William R., "Workers Compensation Experience Rating: What Every Actuary Should Know," PCAS LXXIX, 1992, pp. 215-254.
Meyers, Glenn G., "An Analysis of Retrospective Rating," PCAS LXVII, 1980, pp. 110-143.
Lee, Yoong-Sin, "The Mathematics of Excess of Loss Coverages and Retrospective Rating—A Graphical Approach," PCAS LXXV, 1988, pp. 49-77.
Gillam, William R., "Retrospective Rating: Excess Loss Factors," PCAS LXXVIII, 1991, pp. 1-40.
Cook, Charles F.; Lommele, Jan A.; Retterath, Ronald C., "Reserving Issues for Workers' Compensation," CLRS Transcripts, 1994, pp. 83-130.
Skurnick, David, "The California Table L," PCAS LXI, 1974, pp. 117-140.
Robertson, John P., "NCCI's 2007 Hazard Group Mapping," Variance 3(2), 2009, pp. 194-213.
Venter, Gary G., "Transformed Beta and Gamma Distributions and Aggregate Losses," PCAS LXX, 1983, pp. 156-193.
If you happen to have experience in workers' compensation (*cough*, ahem), your participation will be greatly appreciated. This is intended to be a collaborative effort open to the general public.
