Ponamus

Software for P&C insurers. Reserving now; pricing and capital to follow.

Suppose

The IBNR would be 62,791 against 52,135 as published, +20.4%

A single judgement like this moves the reserve. Reservoir shows what it moved, figure by figure, and keeps a signed record of who made it and why.

Request trial access Reservoir, for loss reserving

Link ratios, RAA data accident years 1981–1990

A tail factor of 1.050 on every year.

Chain-ladder IBNR, as published52,135 With the supposition62,791
General liability, cumulative incurred, from the Reinsurance Association of America, as published by Mack (1994)1. The published figure is reproduced to the unit; the others are recalculated in your browser.

What we make

  • Reservoir

    Loss reserving

    Property and casualty reserving from link ratios through selections to a signed report, with IFRS 17 and Solvency II outputs.

    Available nowWindows and browser

  • Tributary

    Pricing

    Rate-level and rating-factor work: what the book needs if trend or mix changes, and what that does to the indicated rate.

    In development

  • Highwater

    Capital modelling

    The 1-in-200 year: how much capital it takes, and how that moves with reinsurance or a change in the reserve.

    In development

Checking the work

  • Go to source

    Select a figure and press Ctrl+[, as in Excel. Reservoir walks it back to the method, the factors and the cells it came from. An example

  • Audit trail and sign-off

    Each change, down to a single excluded ratio, is a signed entry in a log that shows any tampering, with its author, time and reason. Preparer, reviewer and approver are kept apart and sign with multi-factor authentication.

  • Deterministic results

    The same data and choices give the same figures to the last digit on any machine, desktop or browser, and each result carries a fingerprint you can compare.

  • Your data

    The desktop app works on your own files. In the browser trial the calculations run on your computer and uploaded data never leaves it.

  • The assistant

    It answers questions about reserving and about your project, and declines the rest. When it suggests a change you see the effect first; nothing is applied until you accept, and the change goes on the record under your name.

The Ratios view

Reservoir's Ratios view for Workers Comp, paid: link ratios coloured against the selected factors, eight averages and a fitted exponential curve beneath, and the selected factors with a 1.0184 tail in a highlighted row.
Workers compensation, paid. Ratios coloured against the selection, eight averages and a fitted curve to choose from, an exponential tail of 1.0184, and the cumulative factors that follow. Public CAS Schedule P data; company names are from the dataset, not clients.

Checked against the papers

QuantityPublishedReservoir
Chain-ladder IBNR, RAA data152,13552,135
Mack standard error, RAA126,90926,909
Chain-ladder reserve, Taylor–Ashe data218,680,85618,680,856
Mack standard error, Taylor–Ashe22,447,0952,447,095
ODP prediction error, RAA318,19318,195

Rounded as the papers print them. The ODP difference is England and Verrall’s iterative fit against an exact one. All reconciliations

Contact

Send a triangle and we will take you through it in Reservoir, from the ratios to the signed report. Ponamus is new; a person, not a queue, answers this address.

info@ponamus.com

  1. Mack, T. (1994). Measuring the variability of chain ladder reserve estimates. CAS Forum, Spring 1994.
  2. Mack, T. (1993). Distribution-free calculation of the standard error of chain ladder reserve estimates. ASTIN Bulletin 23(2). Data from Taylor and Ashe (1983).
  3. England, P. D. and Verrall, R. J. (2002). Stochastic claims reserving in general insurance. British Actuarial Journal 8(3), table 3.