Skip to main content

methods.benktander()

Benktander: Bornhuetter-Ferguson repeated, each time from the last ultimate.

Usage

Source

methods.benktander(
    cells,
    *,
    premium,
    expected_loss_ratio,
    n_iters=1,
    dev_grain_months=12,
    average="volume",
    history_periods=None,
    drop_high=False,
    drop_low=False,
    preserve=1,
    drop_above=None,
    drop_below=None,
    exclude=(),
    exclude_valuations=(),
    trim_ties="volume",
    unsupported_factor="raise",
    exhausted_exclusions="keep",
    zero_cells="missing",
    tail=None,
    tail_factor=None,
    tail_decay=None,
    tail_attach_lag=None,
    tail_fit_lags=None,
    tail_steps=None,
    tail_rows=None
)

Mack (2000), “Credible claims reserves: the Benktander method”. With q = 1 - pct_reported at each origin’s latest age, L its latest cumulative loss and E = premium * expected_loss_ratio, the ultimate after n iterations is U_n = L + q * U_(n-1), starting from U_0 = E. n_iters=1 is bornhuetter_ferguson() exactly, and as n_iters grows the ultimate moves to the chain ladder’s. n_iters must be a whole number from 1 to 10,000; 0 would be the expected loss method, which ignores the reported losses, and each iteration is one more pass of a loop, so the upper limit bounds the time one call can take.

cells, the development options and the tail options are as in chain_ladder(), and premium and expected_loss_ratio as in bornhuetter_ferguson(). The result has the same tables and columns as Bornhuetter-Ferguson’s.