Negative Reserves and Minimum Floors

Why a Prospective Reserve Can Be Negative

A prospective reserve is the present value of future benefits and expenses less the present value of future premiums, valued at a specified time and on a stated basis. It can be negative when the value of expected future premiums exceeds the value of expected future outgo. This is possible, for example, when substantial premiums remain due on a policy whose expected future benefits are relatively small.

A negative value is not automatically a calculation error. It is the result of the selected cash flows, assumptions, and valuation time. A separate rule may require a reserve to be at least zero or at least some specified minimum. Applying that rule changes the reserve used in a profit test, and therefore changes the timing of reported profits.

Definition: Let $V_t$ be the raw prospective reserve at the start of year $t$, before that year's premium is received. A zero-floored reserve is $F_t=\max(V_t,0)$. A reserve with a specified minimum $M_t$ is $F_t=\max(V_t,M_t)$. Define the floor adjustment by $\Delta_t=F_t-V_t$, so $\Delta_t\geq 0$.

For the discrete annual calculations below, premiums are received at the start of a year, benefits are paid at the end, and the effective annual interest rate is $i$. The closing reserve $V_{t+1}$ is measured at the end of the year, immediately after that year's benefit payment and before the next year's premium. At a contract's terminal duration, the reserve is zero unless a different terminal obligation is explicitly specified.

Reserve recursion: If the start-of-year premium is $P_t$, the end-of-year benefit is $C_t$, and the closing reserve is $V_{t+1}$, then the compatible raw reserve is $V_t=(C_t+V_{t+1})/(1+i)-P_t$. This recursion assumes the stated timing, a single effective annual interest rate, and no additional cash flows. Include expenses or other benefits in the appropriate cash flows when they are part of the model.

The premium is subtracted after discounting because it is received at the start of the year. Discounting the premium as though it were paid at year-end would use inconsistent timing. The same timing convention must be used when calculating profit.

How a Floor Changes Annual Profit

Define the end-of-year profit for year $t$ as the accumulated opening reserve plus the start-of-year premium, less the end-of-year benefit and closing reserve. Under the stated timing,

\Pi_t=(V_t+P_t)(1+i)-C_t-V_{t+1}.

For an expected profit test with uncertain benefits, use the expected benefit in this expression, provided the reserve and profit projection use consistent assumptions. For a floored reserve, replace the raw reserves at the opening and closing durations by their floored values:

\Pi_t^F=(F_t+P_t)(1+i)-C_t-F_{t+1}.

The floor changes the reserve held, not the premium or benefit cash flows in this calculation. Subtracting the raw profit from the floored profit gives the annual effect directly.

\Pi_t^F-\Pi_t =(F_t-V_t)(1+i)-(F_{t+1}-V_{t+1}) =(1+i)\Delta_t-\Delta_{t+1}.

This result explains why an imposed floor does not simply add the floor adjustment to profit. A higher opening reserve contributes interest during the year, while a higher closing reserve must still be held at year-end. The profit effect is the accumulated opening adjustment less the closing adjustment.

Calculate the raw reserve.
Use future cash flows, valuation assumptions, and their actual timing to determine $V_t$.

Apply the floor at each duration.
Set $F_t=\max(V_t,M_t)$, where $M_t=0$ for a zero floor, and calculate $\Delta_t=F_t-V_t$.

Recalculate profit with floored reserves.
Use the floored opening and closing reserves in the annual profit formula. Do not alter the underlying cash flows unless the product terms themselves change.

Check the profit difference.
Confirm that $\Pi_t^F-\Pi_t=(1+i)\Delta_t-\Delta_{t+1}$.

Worked Examples

Worked Example: A Zero Floor at the Start of a One-Year Projection

A hypothetical one-year contract has a premium of $\$30$ at the start of the year and a guaranteed benefit of $\$10$ at year-end. The effective annual interest rate is $5\%$. There are no expenses, and the terminal reserve is zero. Calculate the raw opening reserve, the zero-floored reserve, and the year-end profit under each reserve basis.

Method. Use the prospective reserve recursion with the premium at the start of the year, then calculate profit from the same timing convention. This avoids discounting the start-of-year premium.

The raw opening reserve is

V_1=\frac{10+0}{1.05}-30 =\frac{10}{1.05}-30 =9.5238095-30 =-20.4761905.

The zero-floored opening reserve is $F_1=\max(-20.4761905,0)=0$, so $\Delta_1=20.4761905$. The terminal reserve is zero under both bases.

Using the raw reserve, the year-end profit is

\Pi_1=(V_1+30)(1.05)-10 =(9.5238095)(1.05)-10 =10-10=0.

Using the zero-floored reserve, the year-end profit is

\Pi_1^F=(0+30)(1.05)-10 =31.50-10 =\$21.50.

The increase is $\$21.50$, which checks against $(1.05)\Delta_1=1.05(20.4761905)=21.50$. The reserve floor shifts profit into this projection year; it does not change the premium or benefit.

Worked Example: A Minimum Floor Across Two Years

A hypothetical two-year contract has a $10\%$ effective annual interest rate. Premiums of $\$8$ and $\$12$ are received at the starts of years 1 and 2. Benefits of $\$4$ and $\$5$ are paid at the ends of those years. There are no expenses, and the terminal reserve is zero. At the starts of years 1 and 2, apply a minimum reserve of $\$2$. Calculate the raw reserves, floored reserves, and the profit effects in each year.

Method. Work backward from the zero terminal reserve using the prospective recursion, then apply the minimum and calculate year-by-year profit. The minimum applies at the starts of the two in-force years, not after the contract has terminated.

At the start of year 2,

V_2=\frac{5+0}{1.10}-12 =4.5454545-12 =-7.4545455.

At the start of year 1,

V_1=\frac{4+V_2}{1.10}-8 =\frac{4-7.4545455}{1.10}-8 =\frac{-3.4545455}{1.10}-8 =-3.1404959-8 =-11.1404959.

Both raw reserves are below the $\$2$ minimum, so $F_1=F_2=2$. The floor adjustments are $\Delta_1=2-(-11.1404959)=13.1404959$ and $\Delta_2=2-(-7.4545455)=9.4545455$. The terminal reserve remains zero.

Raw year-2 profit is $(V_2+12)(1.10)-5=(-7.4545455+12)(1.10)-5=5-5=0$. Floored year-2 profit is $(2+12)(1.10)-5-0=15.40-5=\$10.40$. The difference formula gives $1.10(9.4545455)-0=\$10.40$.

Raw year-1 profit is $(V_1+8)(1.10)-4-V_2=(-11.1404959+8)(1.10)-4-(-7.4545455)=-3.4545455-4+7.4545455=0$. Floored year-1 profit is $(2+8)(1.10)-4-2=11-6=\$5.00$. Its difference from raw profit checks as $1.10(13.1404959)-9.4545455=14.4545455-9.4545455=\$5.00$.

The floor produces $\$5.00$ of additional year-1 profit and $\$10.40$ of additional year-2 profit in this projection. Neither amount should be interpreted as a change to the contract's premiums or benefits.

Worked Example: A Negative Expected Reserve With Uncertain Benefits

A hypothetical one-year policy has a premium of $\$10$ at the start of the year. At year-end it pays $\$20$ with probability $0.25$, and $\$4$ with probability $0.75$. The effective annual interest rate is $5\%$, there are no expenses, and the terminal reserve is zero. Find the expected benefit, raw reserve, and expected profit with a zero floor.

Method. First calculate the expected year-end benefit. Use that expected amount in the one-year prospective reserve and profit calculations, which must use the same probabilities and interest assumption.

The expected benefit is

E[C]=0.25(20)+0.75(4)=5+3=\$8.

The raw reserve is $V_1=8/1.05-10=7.6190476-10=-\$2.3809524$, so the zero-floored reserve is $F_1=0$. Under the raw reserve basis, expected profit is $(V_1+10)(1.05)-8=(7.6190476)(1.05)-8=8-8=0$. Under the zero floor, expected profit is $(0+10)(1.05)-8=10.50-8=\$2.50$. The check is $1.05(2.3809524)=\$2.50$.

This calculation is an expected-profit calculation, not a claim that the year-end benefit is always $\$8$. The actual benefit has two possible amounts; $\$8$ is its expectation under the specified probabilities.

What the Floor Does—and Does Not Do

A floor affects profit through reserve movements. If the floor adjustment is larger at the start of a year than at the end, the accumulated opening adjustment can exceed the closing adjustment, increasing that year's profit. If the closing adjustment is large, some profit is deferred because more reserve remains held at year-end. The sign and size of the annual effect therefore depend on both durations, not just on whether the opening raw reserve is negative.

For a zero floor, a reserve that is already positive is unchanged at that duration. A minimum floor can also affect a positive raw reserve when that reserve is below the required minimum. In either case, calculate the reserve adjustment from the actual floor applied; do not assume it equals the absolute value of a negative reserve unless the floor is zero.

A floor does not, by itself, alter the contractual cash flows or the assumptions used to project them. It changes the reserve basis used in the profit calculation. If an initial cost or funding requirement is part of the model, it must be included consistently when comparing total economics; otherwise, a higher projected annual profit may be mistaken for newly created value.

Key result: With the same premiums, benefits, and interest rate under both bases, the annual profit change is $\Pi_t^F-\Pi_t=(1+i)\Delta_t-\Delta_{t+1}$. A floor can redistribute profit across projection years. Always check both the opening and closing floor adjustments.

Common Timing and Interpretation Traps

Discounting a start-of-year premium. With premium $P_t$ at the start and benefit $C_t$ at year-end, use $V_t=(C_t+V_{t+1})/(1+i)-P_t$, not $(C_t+V_{t+1}-P_t)/(1+i)$.
Applying the floor only to the opening reserve. The profit calculation also includes the closing reserve. A floor adjustment at $t+1$ must be included when calculating year $t$ profit.
Forgetting the interest earned on the opening adjustment. The opening floor adjustment is accumulated by $1+i$ in the end-of-year profit formula.
Imposing a floor after termination without a stated obligation. At expiry, use the terminal reserve specified by the contract or valuation model; a minimum for in-force policies does not automatically create a post-termination reserve.
Treating higher annual profit as changed cash-flow economics. A reserve floor changes the reserve basis and profit emergence. It does not change premiums or benefits unless the terms of the hypothetical product also change.

Check Your Understanding

For each problem, use premiums at the start of the year, benefits at year-end, and the stated effective annual interest rate. Round money to the nearest cent.

A one-year contract has a start-of-year premium of $\$20$, an end-of-year benefit of $\$18$, and an interest rate of $4\%$. Find the raw reserve and its zero-floored value. (Answer: raw reserve $-\$2.69$; zero-floored reserve $\$0.00$.)
A one-year contract has a premium of $\$15$, a benefit of $\$12$, an interest rate of $5\%$, and a minimum opening reserve of $\$2$. Find the raw reserve and floored year-end profit. (Answer: raw reserve $-\$3.57$; floored profit $\$5.85$.)
A two-year contract has an interest rate of $10\%$, premiums $P_1=\$5$ and $P_2=\$8$, and benefits $C_1=\$3$ and $C_2=\$9$, with terminal reserve zero. A zero floor applies at the starts of years 1 and 2. Find the raw opening reserves for years 1 and 2, and the floored profit in year 1. (Answer: $V_2=\$0.18$, $V_1=-\$2.11$, and year-1 floored profit $\$2.32$.)
A one-year policy pays $\$25$ at year-end with probability $0.40$, or $\$5$ with probability $0.60$. The start-of-year premium is $\$14$, the interest rate is $4\%$, and the terminal reserve is zero. Find the raw opening reserve and expected profit with a zero floor. (Answer: expected benefit $\$13.00$; raw reserve $-\$1.50$; floored expected profit $\$1.56$.)
A one-year contract has interest rate $6\%$, a start-of-year premium of $\$15$, and an end-of-year benefit of $\$5.30$, with terminal reserve zero. The raw opening reserve is $-\$10$; apply a zero floor. Find the change in year-end profit. (Answer: $\$10.60$, equal to $1.06(10)$.)

What You'll Learn