Fit chain ladder projection from a Triangle object

Fit a Mack (1993) chain ladder projection from an object of class "Triangle". The function works on long-form cumulative data and does not require a complete triangle. Age-to-age factors are estimated through as_link() and fit_ata(), then applied recursively. The point forecast follows the standard recursion, and prediction uncertainty is decomposed into process variance and parameter variance.

When weight is supplied (e.g. "premium"), age-to-age factors and their variance are estimated using the supplied WLS weights.

Usage

fit_cl(
  x,
  method = c("mack"),
  loss = "loss",
  weight = NULL,
  alpha = 1,
  sigma_method = c("locf", "min_last2", "loglinear", "mack", "none"),
  recent = NULL,
  regime = NULL,
  maturity = NULL,
  tail = FALSE,
  bootstrap = NULL,
  B = 999L,
  seed = NULL,
  conf_level = 0.95
)

Arguments

x

An object of class "Triangle".

method

One of "mack". Default is "mack". The argument is retained for future extensibility.

loss

A single cumulative loss variable (column to project). Typical choices are "loss", "premium", or "ratio".

weight

An optional column name passed to as_link() as the WLS weight variable. Typically "premium" when loss = "ratio". Default is NULL.

alpha

Numeric scalar controlling the variance structure in fit_ata(). Default is 1.

sigma_method

Method used to extrapolate sigma for links where it cannot be estimated. One of "locf" (default), "min_last2", "loglinear", "mack", or "none". "mack" applies the Mack (1993, Appendix B) tail estimator to the last unestimated link only, falling back to LOCF for any earlier ones with a warning. "none" performs no extrapolation; sigma stays NA and downstream variance terms drop those links via finite-value guards. Passed to .extrapolate_sigma_ata().

recent

Optional positive integer. When supplied, only the most recent recent periods are used for factor estimation. Default is NULL (use all periods).

regime

Optional regime specification for cohort cutoff. Accepts: NULL (default – no filter), a Regime object (from detect_regime() or regime_at()), the string "auto" (internal detect_regime(tri, loss = "ratio") call), or a function function(tri) -> Regime for deferred custom-config detection. When supplied, cohorts strictly before the resolved change date are excluded from factor estimation.

maturity

Maturity input forwarded to fit_ata(). Accepts four forms:

NULL (default): No maturity filtering.
Maturity object: Pre-built (e.g. from detect_maturity() or maturity_at()) – used as-is.
"auto": Internal detect_maturity() call with defaults (loss inferred from loss).
function function(tri) -> Maturity: Lazy spec, typically built with maturity_spec(), invoked on the triangle at fit time (leakage-safe for backtest()).

tail

Logical or numeric. If FALSE, no tail factor is applied. If TRUE, a log-linear tail factor is estimated from selected factors. If numeric, the supplied value is used as the tail factor.

bootstrap

Optional bootstrap specification. Accepts NULL (default, analytical Mack SE only), a BootstrapTriangle object produced by bootstrap() (replayed for SE / CI), or the string "auto" to run an internal nonparametric bootstrap at fit time.

B

Integer number of bootstrap replicates when bootstrap = "auto". Default 999L.

seed

Optional integer seed for reproducible bootstrap draws. Default NULL.

conf_level

Numeric in (0, 1). Confidence level used for bootstrap-derived CI columns. Default 0.95.

Value

An object of class "CLFit" containing:

call: The matched call.
data: The input "Triangle" object.
method: The method used ("mack").
groups: Character vector of grouping variable names.
cohort: Character scalar of period variable name.
dev: Character scalar of development variable name.
loss: Character scalar of loss column name.
full: data.table with observed and projected values, including process/parameter SE and CV columns.
proj: data.table identical to full with observed cells set to NA.
link: The "Link" object produced by as_link().
summary: Cohort-level summary with latest, ultimate, reserve, and Mack standard errors.
selected: data.table of selected factors used for projection.
factor: data.table of fitted factors from fit_ata().
maturity: Maturity diagnostics from detect_maturity(), or NULL when maturity filtering was not applied.
alpha: Value of alpha used.
sigma_method: Sigma extrapolation method.
weight: Weight variable name used, or NULL.
recent: Number of recent periods used, or NULL.
regime: Resolved Regime object, or NULL.
use_maturity: Logical; whether maturity filtering was applied.
tail: Tail factor argument supplied by the user.
tail_factor: Numeric tail factor applied.

Examples

if (FALSE) { # \dontrun{
data(experience)
tri <- as_triangle(
  experience[coverage == "surgery"],
  groups   = "coverage",
  cohort   = "uy_m",
  calendar = "cy_m",
  loss     = "incr_loss",
  premium  = "incr_premium"
)

# Mack chain ladder with process / parameter standard errors
cl_mack <- fit_cl(tri, loss = "loss", method = "mack")
summary(cl_mack)
plot(cl_mack)

# WLS factors for ratio (loss ratio) using premium as the weight
cl_ratio <- fit_cl(tri, loss = "ratio", weight = "premium")
} # }

Fit chain ladder projection from a `Triangle` object

Usage

Arguments

Value

See also

Examples