Skip to main content

methods.bornhuetter_ferguson()

Bornhuetter-Ferguson: the unreported share of an a priori ultimate.

Usage

Source

methods.bornhuetter_ferguson(
    cells,
    *,
    premium,
    expected_loss_ratio,
    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
)

Each origin’s ultimate is its latest cumulative loss plus premium * expected_loss_ratio * (1 - pct_reported), where pct_reported comes from the chain-ladder development pattern.

cells, the development options and the tail options are as in chain_ladder(); with a tail, pct_reported includes it.

premium is keyed by origin period, never by position: either a table with columns origin_period and premium (a polars DataFrame or a pyarrow Table, for example), or a dict from origin period to amount. The origin periods may be written in any of the ways the cells’ may, and need not be written the same way as the cells (integer years in the cells and year-end dates in premium, for example). It needs exactly one positive amount for each origin of the triangle and no others.

expected_loss_ratio is the a priori loss ratio, one number for every origin, applied to premium. benktander() iterates this method.