Risk Retention And Funded Losses

Introduction

Risk cannot always be eliminated, and retaining some exposure can be a deliberate business choice. The important distinction is between accepting the exposure with a considered plan for paying losses and allowing losses to fall unpredictably on cash flow. A retention decision therefore has two parts: deciding how much risk the organization will bear and deciding how it will finance the resulting losses.

Funding does not make a loss less likely or reduce its severity. It can, however, make payment more manageable and reduce the chance that a loss creates a sudden liquidity problem. This tutorial examines unfunded retention, internal funding approaches, and how loss predictability affects the amount of support that may be needed. The discussion is aligned with the CAS Exam 9 (Risk Management for Actuaries) syllabus.

Definition: Risk retention is the deliberate acceptance by an organization of some or all of an exposure to loss. Under unfunded retention, losses are paid from cash or operating resources available when they occur, without a specific fund set aside in advance. Under funded retention, the organization plans in advance to provide resources for retained losses, for example through a dedicated pool of liquid assets, budget contributions, or an internal capital allocation.

Retention Is a Financing Choice as Well as a Risk Choice

Retaining a risk does not mean ignoring it. The organization should understand the losses that remain after its controls, decide whether it can bear them, and consider when funds might be needed. The earlier material on risk avoidance and reduction decisions distinguishes removing an exposure from reducing its likelihood or impact. After those decisions, some residual exposure may remain; retention is one way to address that residual risk.

The expected loss is useful for planning, but it does not describe the full range of possible outcomes. A predictable series of modest losses may be budgeted as operating expense. A volatile exposure can produce losses far above its average, so a fund set equal to the expected loss may be inadequate in an adverse year. Funding decisions should therefore consider loss frequency, severity, variability, liquidity, and the organization's ability to replenish resources.

Approach	How losses are met	Key consideration
Unfunded retention	Cash or operating resources available when the loss occurs	Payment may compete with other cash needs in a bad period
Budgeted retention	Planned operating contributions, often based on expected losses	A budget based only on an average may not cover unusually high losses
Dedicated internal fund	Assets are set aside or earmarked for retained losses	The fund's amount, liquidity, permitted use, and replenishment plan matter
Internal capital allocation	Resources are identified as available to support the exposure, without necessarily being physically segregated	An accounting allocation alone does not ensure cash will be available when needed

An accounting reserve or provision should not be confused with a pool of cash. A booked amount can recognize an expected obligation, but it does not by itself create liquid assets. Conversely, a dedicated fund can support payment even though the underlying loss still affects the organization's economic results. A sound funding plan states what resources are available, who can access them, and how they will be replenished after use.

Expected Loss and the Risk of Reserve Shortfall

Suppose aggregate retained losses over a defined period are represented by $L$, and the organization has $R$ dollars available for those losses. A shortfall occurs when $L>R$. The expected loss $E[L]$ summarizes the average over repeated periods under the assumed model; it does not guarantee that any particular period's losses will be near that average.

For independent loss amounts $X_1,\ldots,X_n$, aggregate loss is $L=\sum_{j=1}^{n}X_j$. Linearity of expectation gives $E[L]=\sum_{j=1}^{n}E[X_j]$, whether or not the losses are independent. If they are independent and have finite variances, then $\operatorname{Var}(L)=\sum_{j=1}^{n}\operatorname{Var}(X_j)$. Independence matters for the variance result: losses that tend to occur together can create a more volatile aggregate than the independent model suggests.

Worked Example: A Reserve for Repeated, Modest Losses

A hypothetical service operation expects five independent opportunities for a small retained loss during the coming year. Each opportunity has probability $0.20$ of producing a loss of $10,000 and probability $0.80$ of producing no loss. The organization has a $20,000 fund. Find the expected aggregate loss, its standard deviation, and the probability that the fund is insufficient.

Model and assumptions. Let $N$ be the number of losses. Then $N\sim\operatorname{Binomial}(5,0.20)$, and aggregate loss is $L=$10{,}000N$. The opportunities are independent, the one-year loss amounts are fixed as stated, and there is no investment return during the period ($i=0\%$). No mortality assumption is relevant to this scenario.

Expected loss. The expected number of losses is $E[N]=5(0.20)=1$. Therefore,

$$ E[L]=$10{,}000E[N]=$10{,}000. $$

Variability. The binomial variance is $\operatorname{Var}(N)=np(1-p)=5(0.20)(0.80)=0.80$. Thus $\operatorname{Var}(L)=($10{,}000)^2(0.80)=$80{,}000{,}000$, and the standard deviation is $\sqrt{$80{,}000{,}000}=$8{,}944.27$.

Probability of a shortfall. The fund is insufficient when $L>$20{,}000$, which means $N\geq 3$. The exact binomial probabilities are

\begin{aligned} P(N\geq 3) &=P(N=3)+P(N=4)+P(N=5)\\ &=\binom{5}{3}(0.20)^3(0.80)^2 +\binom{5}{4}(0.20)^4(0.80) +(0.20)^5\\ &=0.0512+0.0064+0.00032\\ &=0.05792. \end{aligned}

Rounded to four decimal places, the shortfall probability is 0.0579. The fund covers outcomes with zero, one, or two losses, but not three or more. This calculation shows why an expected loss of $10,000 does not by itself determine a sufficient fund: the chosen $20,000 fund still has a positive probability of being exceeded.

Method and mistake avoided. The method is to model aggregate loss through the binomial claim count, then compare the fund with each possible aggregate outcome. A common mistake is to call the fund adequate merely because it is twice the expected loss. Adequacy depends on the organization's tolerance for the chance and size of a shortfall, not on that ratio alone.

Funding Volatile Losses

For a low-frequency, high-severity exposure, a reserve near the expected loss can look small relative to the amount that might be needed in an adverse outcome. A useful analysis distinguishes at least three quantities: the average loss, the probability that available funds are exceeded, and the size of the additional funding need if they are exceeded. The right funding choice also depends on whether the organization can obtain cash promptly without disrupting other obligations.

Worked Example: A Fund Facing a Rare Severe Loss

A hypothetical retained exposure has the following one-year aggregate loss distribution: $0 with probability 0.95, $50,000 with probability 0.04, and $300,000 with probability 0.01. The organization has $50,000 earmarked for these losses. Calculate the expected loss, the probability of a shortfall, and the expected amount by which loss exceeds the fund.

Model and assumptions. The aggregate loss $L$ has the discrete distribution stated above, with no other possible outcomes. The period is one year, $i=0\%$, and there is no mortality component. The fund is available in cash at the time of loss.

Expected loss. Weight each possible loss by its probability:

$$ E[L]=(0)(0.95)+($50{,}000)(0.04)+($300{,}000)(0.01) =$2{,}000+$3{,}000=$5{,}000. $$

Shortfall probability. The fund covers the $0 and $50,000 outcomes exactly or in full. Only the $300,000 outcome exceeds it, so $P(L>$50{,}000)=0.0100$.

Expected excess over the fund. Define the excess as $\max(L-$50{,}000,0)$. It is $250,000 in the severe outcome and zero otherwise. Its expectation is

E[\max(L-$50{,}000,0)] =($250{,}000)(0.01)+($0)(0.99) =$2{,}500.

The $5,000 expected loss is one tenth of the available fund, but that average does not prevent a $250,000 funding gap in the 1% severe outcome. A plan could identify additional liquid resources or a contingency source for such a gap; it should not imply that the $50,000 fund itself covers the severe loss.

Method and mistake avoided. The method is to evaluate the shortfall outcome by outcome, rather than infer protection from the ratio of fund to expected loss. A common error is to say that the fund is sufficient because it equals the $50,000 moderate loss. It is sufficient for that outcome, but not for every outcome in the stated distribution.

Projecting a Fund Over Time

A funded-retention plan needs a time convention. If $B_{t-1}$ is the fund at the start of year $t$, $i$ is the effective annual rate earned during that year, $C_t$ is a contribution made at year-end, and $L_t$ is the loss paid at year-end, a simple balance projection is

B_t=B_{t-1}(1+i)+C_t-L_t.

This formula assumes the opening balance earns interest for the full year and that the contribution and loss occur at year-end, with the contribution available for the payment. If contributions or payments occur at different times, the timing and interest accumulation must be changed. The formula is a cash-balance projection, not a model of how likely the losses are.

Start with available resources.
Identify the opening fund and whether it is actually liquid and available for the retained exposure.

Apply the time convention.
Accumulate the opening fund for the period at the stated effective rate; place contributions and payments at their specified dates.

Compare resources with losses.
A negative projected balance signals that the plan needs additional cash, a larger opening fund, higher contributions, or a different risk decision.

Test adverse outcomes and replenishment.
Repeat the projection for plausible loss scenarios and specify how the fund will be restored after it is used.

The recurrence follows directly from the cash flows. The opening amount becomes $B_{t-1}(1+i)$ after earning interest. Adding the year-end contribution gives $B_{t-1}(1+i)+C_t$; paying the loss then leaves $B_t=B_{t-1}(1+i)+C_t-L_t$. If this result is negative, it represents the extra cash required at that date to meet the modeled payment, rather than a physically negative cash balance.

Worked Example: Testing a Multi-Year Internal Fund

A hypothetical department begins with $20,000 in a dedicated fund. At each year-end it contributes $10,000 and pays the year's retained losses. The opening balance earns an effective annual rate of $i=3\%$. Project the fund through year 1, when losses are $8,000, and year 2, when losses are $35,000. Find the year 2 shortfall, if any.

Model and assumptions. The stated losses are deterministic projections for this two-year cash-flow test; no probability distribution is assumed. Interest is earned on each year's opening balance for the full year. Contributions and losses occur at year-end, with the contribution available before the loss is paid. No mortality assumption is relevant.

Year 1. Interest on the opening fund is $$20{,}000(0.03)=$600$. The balance before the contribution and loss is $$20{,}000+$600=$20{,}600$. After the $10,000 contribution and $8,000 payment,

B_1=$20{,}600+$10{,}000-$8{,}000=$22{,}600.

Year 2. Interest on the year 2 opening balance is $$22{,}600(0.03)=$678$, so the balance before the year-end contribution and loss is $$22{,}600+$678=$23{,}278$. Adding the contribution gives $$23{,}278+$10{,}000=$33{,}278$. The projected loss is $35,000, so the shortfall is $$35{,}000-$33{,}278=$1{,}722$.

The fund is not sufficient for the specified year 2 cash flows without an additional $1,722 at the payment date. If that amount is contributed to meet the loss, the fund ends at zero. The calculation does not imply that $1,722 is adequate for other possible loss scenarios.

Method and mistake avoided. The method is to roll the fund forward year by year using the stated timing, then compare the resources available immediately before payment with the loss. A common error is to apply interest to the year-end contribution as though it had been invested for the full year; under the stated timing, it earns no year 2 interest.

Common Mistakes and Practical Checks

Treating expected loss as a guaranteed annual amount. An average is not a cap. A reserve equal to expected loss can be exceeded, sometimes substantially.
Calling an accounting allocation cash. Confirm that the resources are liquid, accessible, and not already committed elsewhere.
Ignoring timing. A contribution made after a loss cannot finance that payment. State when interest, contributions, and losses occur before projecting a balance.
Assuming a reserve eliminates risk. The loss still occurs; funding primarily supports the ability to pay it. The exposure's severity and uncertainty remain.
Failing to plan replenishment. A fund that can cover one event may be depleted afterward. Decide how and when resources will be restored, especially when losses may recur.

Key takeaway: Retention is deliberate only when the organization understands the residual exposure and can explain how losses will be financed. Expected losses help set a budget, while volatility, liquidity needs, adverse scenarios, and replenishment capacity help assess whether that budget or fund is enough.

Check Your Understanding

For each problem, state the key funding implication as well as the requested calculation.

A hypothetical unit faces four independent opportunities for a $8,000 loss, each with probability 0.25; otherwise the loss is zero. With $i=0\%$, find the expected annual loss and the probability that a $16,000 fund is insufficient. (Answer: Expected loss $8,000; shortfall probability 0.0508.)
A one-year retained loss is $0 with probability 0.80, $20,000 with probability 0.15, and $100,000 with probability 0.05. Find the expected loss and the probability that a $25,000 fund is insufficient. (Answer: Expected loss $8,000; shortfall probability 0.0500.)
A dedicated fund starts at $5,000, earns an effective annual rate of 5%, receives a $4,000 contribution at year-end, and pays a $2,000 loss at that same year-end. Assume the contribution is available before payment. Find the ending balance. (Answer: $7,250.00.)
Ten independent exposures each produce a $1,000 loss with probability 0.10 and zero otherwise. Find the aggregate expected loss and the probability that a $2,000 fund is insufficient. (Answer: Expected loss $1,000; shortfall probability 0.0702.)
A fund begins with $12,000, earns 4% on its opening balance for the year, receives a $7,000 year-end contribution, and must pay a $20,000 year-end loss. The contribution is available before payment. Find the additional cash required to avoid a negative balance. (Answer: $520.00.)

What You'll Learn