Skip to contents

Build a total development summary from experience data

Source: R/triangle.R
build_total.Rd

Aggregate loss and rp by group and compute the corresponding total loss ratio over a selected period window.

This function is intended for high-level portfolio comparison across groups such as products, coverages, or channels. It summarises:

  • the number of observed periods (n_obs)

  • the first and last observed periods (sales_start, sales_end)

  • total loss and total rp

  • total loss ratio (lr = loss / rp)

  • each group's share of total loss and risk premium

If period_from and/or period_to are supplied, the input data are first restricted to that period window before aggregation. This is useful when comparing groups on a common period basis.

Usage

build_total(
  df,
  group_var,
  cohort_var = "uym",
  dev_var = "elap_m",
  value_var = c("loss", "rp"),
  period_from = NULL,
  period_to = NULL,
  fill_gaps = FALSE
)

Arguments

df

A data.frame containing experience data.

group_var

Grouping variable(s).

cohort_var

A single period variable. This may be an underwriting period (uym, uyq, uyh, uy) or a calendar period (cym, cyq, cyh, cy). Default "uym".

dev_var

A single development variable used to count observed periods. Default "elap_m".

value_var

Value variables to aggregate. Must include both "loss" and "rp". Default c("loss", "rp").

period_from

Optional lower bound for cohort_var. Only rows with cohort_var >= period_from are kept. May be supplied as Date, character, or any value coercible to Date. Default NULL.

period_to

Optional upper bound for cohort_var. Only rows with cohort_var <= period_to are kept. May be supplied as Date, character, or any value coercible to Date. Default NULL.

fill_gaps

Logical; if TRUE, zero-fill missing (group_var, cohort_var, dev_var) cells before aggregation so that every cohort has a consecutive dev_var sequence. Default FALSE. Note that filling inflates n_obs (counts filled rows as observed periods); use validate_triangle() to inspect first.

Value

A data.frame with class "Total" containing:

n_obs

Number of observed development periods

sales_start

First observed period

sales_end

Last observed period

loss

Total loss

rp

Total risk premium

lr

Total loss ratio (loss / rp)

loss_prop

Share of total loss

rp_prop

Share of total risk premium

Examples

if (FALSE) { # \dontrun{
build_total(df, cv_nm)

build_total(
  df,
  cv_nm,
  period_from = "2023-01-01",
  period_to   = "2023-12-01"
)
} # }