Why Test Forecast Assumptions?
A forecast-based internal capital assessment depends on assumptions about the business the insurer expects to write and how that business will perform. Growth may increase earnings, but it can also increase the capital needed to support the business. Higher expenses can reduce resources available to absorb losses. Investment returns and dividends can also change forecast capital resources. A conclusion that appears adequate under one set of assumptions may therefore be sensitive to modest changes in the forecast.
The purpose of an assumption test is not to assert that a particular alternative forecast will occur. It is to find out whether the internal capital conclusion is robust to specified, plausible changes and to identify which assumptions matter most. The test must use a consistent projection basis: a changed assumption should flow through every affected part of the calculation, rather than being applied only where it makes the result look worse or better.
A positive margin says that resources exceed the selected target at the assessment date; it does not establish that every risk is adequately managed or that the assumptions are likely. The interpretation also depends on how \(R\) and \(K\) are defined, whether capital is fungible where needed, and whether the projection captures relevant constraints. Those definitions and constraints must be held consistent across the base forecast and each sensitivity.
A disciplined assumption test
Begin with the approved forecast basis and record the assumptions that feed the resource and capital-need projections. Change a selected assumption or a clearly defined set of assumptions, recompute both \(R\) and \(K\), and compare the resulting margin with the base margin. An assumption that affects only resources can still change the adequacy result; an assumption that affects both resources and capital need must be reflected in both.
State the assessment date, resource and capital-need definitions, target multiple, and forecast assumptions. Calculate the base margin.
Change an assumption by a specified amount or to a specified adverse level. Keep unrelated inputs fixed for a single-assumption sensitivity.
For example, a growth change may affect both forecast earnings and the capital need associated with business volume.
Check whether the sign of the margin changes, how much headroom remains, and which assumption caused the change. State the limits of the test.
For a fixed target multiple \(q\), the change in margin from a base forecast to a revised forecast is
This follows directly by subtracting the base margin \(R-qK\) from the revised margin \((R+\Delta R)-q(K+\Delta K)\). The revised expression expands to \(R-qK+\Delta R-q\Delta K\); the common base terms cancel. Thus, an increase in resources improves the margin, while an increase in capital need reduces it by \(q\) times that increase. If the target multiple itself changes, this fixed-\(q\) formula is insufficient: the margin must instead be recalculated using the revised \(q\).
Worked Example: Growth Raises Both Earnings and Capital Need
A hypothetical insurer has a one-year forecast with all amounts in millions of dollars. Its internal target is \(R\geq 1.20K\). Forecast premiums are \(P_1=100(1+g)\), where \(g\) is the assumed growth rate. The forecast equations are \(R_1=150+0.08P_1-3\) and \(K_1=80+0.30P_1\). The 8% contribution to resources is the assumed operating contribution per premium dollar, and the 3 is a fixed dividend. Compare growth of 10% with growth of 20%, holding all other assumptions fixed.
Method. Recalculate forecast premiums, resources, capital need, and the margin under each growth assumption. This avoids the mistake of treating higher growth as an increase in resources without also allowing for the associated capital need.
At 10% growth: \(P_1=100(1+0.10)=110\). Resources are \(R_1=150+0.08(110)-3=150+8.8-3=155.8\). Capital need is \(K_1=80+0.30(110)=80+33=113\). The target resources are \(1.20(113)=135.6\), so the margin is \(M=155.8-135.6=20.2\). The target is met.
At 20% growth: \(P_1=100(1+0.20)=120\). Resources are \(R_1=150+0.08(120)-3=150+9.6-3=156.6\). Capital need is \(K_1=80+0.30(120)=80+36=116\). Target resources are \(1.20(116)=139.2\), so \(M=156.6-139.2=17.4\). The target is still met, but the margin is lower.
Check. The growth increase raises resources by \(156.6-155.8=0.8\) and capital need by \(116-113=3\). The margin change is \(0.8-1.20(3)=0.8-3.6=-2.8\), matching \(17.4-20.2=-2.8\). The conclusion remains adequate in this test, but the higher growth assumption does not strengthen the margin.
Expense Assumptions and Thin Margins
An expense sensitivity often changes the projected resources directly through lower earnings. Whether it also changes capital need depends on the chosen capital model and the expense measure being varied. The analyst should not automatically hold \(K\) fixed if the model links expenses to the amount or risk profile of the business. Conversely, if the stated projection defines \(K\) independently of expenses, changing \(K\) without a modeled reason would be inconsistent.
Worked Example: A Small Expense Change Reverses the Result
A hypothetical insurer projects one-year resources in millions of dollars as \(R_1=190+15-E-4\), where \(E\) is forecast operating expense and 4 is a planned dividend. Its internal capital need is \(K_1=155\), and its target multiple is \(q=1.25\). Compare the base expense assumption of 7 with an expense assumption 20% higher. Assume, for this example, that the capital model does not change \(K_1\) when expenses change.
Method. Recalculate resources using the revised expense, then compare them with the unchanged target requirement. The common mistake is to apply a percentage increase to total resources rather than to the expense assumption itself.
Base forecast: Target resources are \(1.25(155)=193.75\). Resources are \(R_1=190+15-7-4=194\). Therefore \(M=194-193.75=0.25\), just above the target.
Higher-expense forecast: The revised expense is \(7(1.20)=8.4\), an increase of \(8.4-7=1.4\). Resources are \(R_1=190+15-8.4-4=192.6\). Capital need remains 155, so \(M=192.6-193.75=-1.15\). The target is missed.
Check. Resources fell by exactly 1.4, so the margin also fell from \(0.25\) to \(0.25-1.4=-1.15\). The calculation shows that the base result is sensitive because its headroom is smaller than the tested expense increase. It does not show that expenses will in fact rise by 20%.
Investment Returns, Dividends, and Break-Even Values
Investment-return and dividend assumptions affect capital resources through different mechanisms. Returns contribute to investment income or asset values under the projection’s specified basis; dividends are distributions that reduce resources available to support the business. The forecast must be explicit about the period, the asset base to which a return is applied, and whether the dividend is paid during that period. A rate applied to the wrong asset base or period can materially distort the sensitivity.
A useful way to summarize sensitivity is to solve for the assumption value at which the margin reaches zero. This break-even value shows how much deterioration the forecast can absorb before it crosses the chosen target. It is conditional on the other assumptions and the specified projection equations; it is not a probability forecast.
Worked Example: Finding the Break-Even Investment Return
A hypothetical insurer has \(90\) million dollars of starting capital resources, expected operating earnings of \(8\) million dollars, and \(250\) million dollars of invested assets. It pays a dividend of \(8\) million dollars at the end of the year. Assume the one-year investment return \(r\) is applied to the full 250 million dollar asset base, with no other investment effects. Projected capital need is \(K_1=80\) million dollars and the target multiple is \(1.20\). Find the minimum return that meets the target, then test a 2.0% return.
Method. Write resources as a function of the return, set the margin to zero, and solve for \(r\). This avoids confusing the return on invested assets with the return on starting capital resources.
The target resources are \(1.20(80)=96\) million dollars. Forecast resources are \(R_1=90+8+250r-8=90+250r\). The target is met when \(90+250r\geq96\). Thus \(250r\geq6\), and \(r\geq6/250=0.024\), or 2.4%.
At a 2.0% return, investment income is \(250(0.020)=5\) million dollars. Resources are \(90+8+5-8=95\) million dollars. The margin is \(95-96=-1\) million dollars, so the target is missed. At the break-even return of 2.4%, investment income is \(250(0.024)=6\), resources are \(90+8+6-8=96\), and the margin is zero.
Check. The 2.0% return is 0.4 percentage points below the break-even value. On 250 million dollars of assets, that difference is \(250(0.004)=1\) million dollars, exactly the shortfall. The calculation is consistent with the direct margin result.
Reading Sensitivities Without Overstating Them
A single-assumption test helps isolate a driver, but it does not necessarily describe a plausible business outcome. Growth, expenses, investment returns, and dividends may be related. For instance, a growth plan might require additional expenses, while a lower-return environment could also affect management’s dividend decision. Test these as a combined set only when the assumptions form a coherent hypothetical scenario, and state the relationship between them. Do not add unrelated adverse changes merely to produce a more severe result.
A conclusion is more robust to the tested assumptions when reasonable variations leave the result on the same side of the stated target with meaningful headroom. A result that changes sign under a small variation is sensitive, even if the base forecast narrowly meets the target. Neither observation alone determines the final assessment: the relevance of the tested range, the quality of the model, and the organization’s stated decision criteria also matter.
| Test feature | Good practice | Common weakness |
|---|---|---|
| Assumption definition | Specify the changed input and its amount or level | Use a vague label such as “adverse” without stating what changed |
| Projection mechanics | Recalculate every affected resource and need | Change earnings but ignore the associated change in capital need |
| Interpretation | Report the margin and whether the target is met | Describe a positive base margin as proof of resilience |
| Combined changes | Explain why the assumptions can occur together | Combine inputs without a coherent business rationale |
The key exam trap is confusing a sensitivity with a forecast of what will happen. A test answers a conditional question: “What is the capital result if these specified assumptions apply, holding the other stated conditions as defined?” It does not by itself assign likelihood to the changed assumptions. State the conditions, calculate the revised margin, and limit the conclusion to what the test demonstrates.
Check Your Understanding
For each hypothetical case, calculate the requested margin or threshold using the stated assumptions.
- A forecast uses \(P_1=100(1+g)\), \(R_1=120+0.06P_1-2\), \(K_1=50+0.20P_1\), and \(q=1.10\). Calculate the margin at \(g=10\%\) and at \(g=15\%\). (Answer: \(M=45.4\) million at 10% growth; \(M=44.6\) million at 15% growth.)
- Resources are \(R_1=100+12-E-3\), capital need is \(80\), and \(q=1.20\). Calculate the margin for expenses \(E=10\) and \(E=14\), and state whether the target is met in each case. (Answer: \(M=3\) at \(E=10\), target met; \(M=-1\) at \(E=14\), target missed; amounts are in millions.)
- Starting resources are 70 million dollars, operating earnings are 5 million dollars, invested assets are 100 million dollars, and the dividend is 6 million dollars. Capital need is 60 million dollars and \(q=1.25\). Under \(R_1=70+5+100r-6\), find the break-even return. (Answer: \(r=6.0\%\).)
- A hypothetical insurer has starting resources of 140 million dollars, operating earnings of 9 million dollars, invested assets of 200 million dollars, and a dividend of 11 million dollars. Capital need is 110 million dollars and \(q=1.20\). If the return is 1%, calculate resources and the capital margin. (Answer: \(R_1=140\) million; \(M=8\) million; target met.)