Expected Ultimate Losses For Partially Mature Periods

Why Estimate Ultimate Loss Before Claims Mature?

A recent accident period may have little reported or paid claim experience even though its policies have already earned premium or accumulated exposure. For ratemaking and reserve analysis, an actuary may still need an estimate of the period’s ultimate losses: the losses expected to be attributable to the period once its claims are fully known and valued. An a priori estimate uses the period’s premium or exposure together with an expected loss assumption, rather than relying on claim experience that has not yet emerged.

The central distinction is between how much loss is ultimately expected and how much of that loss has appeared in the data so far. A period can have low reported losses because it is immature, not because its expected ultimate losses are low. Claim maturity therefore does not, by itself, change the a priori estimate of ultimate losses.

Two Bases for an A Priori Estimate

The premium-based method begins with earned premium on the same basis as the selected expected loss ratio. The expected loss ratio (ELR) is the selected expected ultimate loss amount divided by the corresponding premium amount. If \(EP\) is the period’s earned premium and the selected ELR is expressed as a decimal, the estimate is:

Here, \(\widehat{L}_{U}^{\,\mathrm{exp}}\) denotes estimated expected ultimate losses. The formula is appropriate when the premium and ELR are aligned in coverage, loss and LAE scope, time period, and rate level. For example, an ELR selected for on-level premium should be applied to premium on that same on-level basis. If those bases differ, first make the relevant adjustment or select a compatible assumption; multiplying mismatched quantities does not produce a sound estimate.

The exposure-based method uses the period’s exposure units and an expected loss cost per unit. An exposure unit might be a vehicle-year, building-year, or another measure appropriate to the coverage. If \(E\) is the number of exposure units and \(ELC\) is the selected expected loss cost per unit, then:

The loss cost and exposure definition must match. A loss cost per vehicle-year cannot be multiplied by building-years, for instance. The loss cost must also reflect the intended loss and LAE scope and be appropriate for the period’s coverage and risk characteristics.

Why the Formulas Work

The premium formula follows directly from the definition of an expected loss ratio. Suppose the selected ELR represents expected ultimate loss per dollar of compatible premium. For one dollar of premium, the expected ultimate loss is \(ELR\) dollars. For \(EP\) dollars of premium, linear scaling gives the expected amount \(EP \times ELR\). This scaling assumes the selected ratio is suitable for the premium being analyzed; it does not require each policy to have identical losses.

The exposure formula has the same logic. If each appropriate exposure unit carries an expected ultimate loss cost of \(ELC\), then \(E\) units have an aggregate expected loss of \(E \times ELC\). More formally, if \(L_j\) is ultimate loss attributable to exposure unit \(j\), and each unit has the same conditional expected loss \(E[L_j]=ELC\), linearity of expectation gives:

This derivation does not require losses from different exposure units to be independent. It does require that the assumed expected loss cost reasonably represents the units being counted. If expected costs differ across meaningful groups, calculate the expectation for each group and add the results rather than forcing a single inappropriate cost onto all exposures.

Neither formula says that actual ultimate losses will equal the estimate. Random claim occurrence, claim severity, and changes in the composition of risks can cause actual results to differ. The formulas state how the selected expectation is applied to the period’s premium or exposure.

Worked Applications

Claim Maturity Is Not an Ultimate-Loss Multiplier

A maturity measure describes the amount of claim experience that has emerged under a defined measure and at a defined valuation date. It is not automatically a percentage by which an a priori ultimate estimate should be reduced. For example, if a selected ELR implies $1,548,000 of expected ultimate loss, multiplying that amount by 40% merely because the period is described as “40% mature” gives $619,200. That calculation estimates neither the stated expected ultimate nor, without an appropriate emergence assumption and definition, a valid observed amount.

The following comparison separates the quantities:

A proper emerging-loss or reserving analysis may use claim data and development assumptions to assess ultimate losses. That is a different source of information from the a priori estimate developed here. The two types of information should not be confused: the a priori estimate starts from selected expected cost, while an experience-based estimate depends on the observed claims and the method used to interpret them.

Common Basis Checks

Before applying either formula, verify the inputs rather than treating multiplication as the whole task. Ask whether the premium is earned for the period being estimated, whether its rate level matches the selected ELR, and whether the ELR includes the intended loss and LAE components. For an exposure-based calculation, confirm that the exposure count uses the unit underlying the loss cost and that the cost reflects the same coverage and loss scope.

A common exam trap is to use a rate or cost that is numerically plausible but belongs to a different basis. For example, a loss-only ELR multiplied into a premium basis intended to estimate loss and LAE omits LAE. Likewise, a per-policy loss cost cannot be applied to a count of insured locations unless the selected cost is explicitly defined on that exposure basis. Resolve such mismatches before calculating.