The summary appeared comfortable. A management liquidity exercise produced a weighted mean additional cash need of ₹12 crore, and the defined account contained ₹18 crore of usable capacity. The same exercise included a scenario requiring ₹36 crore. Its presence should have prevented the mean from being read as a statement that every scenario could be met. Instead, the comparison of two summary numbers encouraged precisely that impression. The question was not whether the expected-value calculation was wrong. It was whether the calculation had been asked to establish something it did not establish.
In the fictional 2024 exercise, the three additional-cash amounts were ₹0, ₹12 crore and ₹36 crore, with assumed weights of 40%, 40% and 20%. Multiplying and adding gives ₹12 crore: zero plus ₹4.8 crore plus ₹7.2 crore. The ₹18 crore capacity exceeded that mean by ₹6 crore, but it fell short of the ₹36 crore scenario by ₹18 crore. Both statements were arithmetically true. They described different relationships, and the decision needed to preserve the relationship relevant to the capacity management wanted to retain.
The weights were illustrative planning assumptions. They were not proven event probabilities, an actuarial claims estimate or a regulatory-capital calculation. The exercise did not identify the largest demand the insurer could ever face. Risk, actuarial, claims and investment owners retained their separate professional assessments. I wanted management to see the scenario account clearly enough to make a bounded liquidity decision, without borrowing authority from a weighted average or presenting the largest listed case as a complete description of uncertainty.
## Read the scenarios before the summary number
A mean can be useful because it compresses several conditional amounts into one number under stated weights. Compression necessarily removes detail. Two different collections of scenarios can produce the same mean while implying very different demands on resources. The average therefore needs to remain connected to the conditions from which it was calculated. In our paper, the ₹36 crore case had not disappeared mathematically, but it had become less visible in the recommendation. The board could easily focus on ₹12 crore versus ₹18 crore and miss the capacity question still open within the exercise.
I asked the team to show all three scenarios with the assumptions behind them before presenting the mean. That order was intended to make the conditional nature of the account easier to understand. It did not mean that the highest listed amount automatically determined the correct decision. Management still had to consider the purpose of the exercise, the wider obligations of the insurer, the usability of the resources and the cost of retaining more. A clearer presentation creates a better judgement; it should not disguise another automatic rule as the answer.
The meaning of usable capacity also required care. The ₹18 crore was the defined amount available within this management exercise, subject to the appropriate permissions and wider professional account. It was not a statement that every asset could be sold immediately or that funds held for another purpose could be redirected without consequence. A capacity measure can become misleading in the same way as an expected cash measure if the conditions behind it disappear. Both sides of the comparison need to describe resources and obligations on a basis appropriate to the decision being made.
## Capacity carries a cost that deserves to be seen
The response was to retain an additional ₹18 crore of liquidity through changes to future investment maturities within the approved process. On the stated exercise, that addressed the gap to the ₹36 crore scenario. The choice had a cost. Using a 4% annual planning yield difference, the comparison showed ₹72 lakh a year of forgone yield. This was an assumption for evaluating the decision, not an observed investment loss or a forecast that the difference would persist unchanged. I wanted that cost in the same discussion as the additional capacity.
It is tempting to describe a prudent-looking liquidity decision as though it has no tradeoff. That makes the recommendation morally comfortable but financially incomplete. Resources retained in one form may earn differently or be unavailable for another purpose. The board should be able to consider that consequence without being told that questioning the cost means disregarding resilience. Equally, an attractive yield should not make a capacity gap disappear. The finance role is to make the alternatives and their assumptions comparable enough for an informed decision, including when neither option provides a satisfying certainty.
Our action did not introduce an unapproved borrowing facility or assume that additional cash could be created at will after an event. It concerned future maturity choices and retained liquidity through the authorised process. Investment colleagues had to consider implementation alongside their mandates and the wider asset-and-obligation account. That professional work mattered as much as the amount in the management paper. A board approval can establish a direction and resource expectation, but it does not itself arrange every maturity or remove every constraint that affects practical availability.
## Do not replace one certainty with another
After exposing the limitation of the mean, a finance leader can be tempted to make the largest scenario the new reassuring answer. We now have ₹36 crore against a ₹36 crore case, so the problem appears solved. That conclusion would exceed the exercise. The scenarios were constructed for a bounded discussion, and the actual future can differ in amount, timing or combination. Addressing one identified gap is useful without becoming proof of universal resilience. The account should explain what the decision covers and what further monitoring or assessment remains necessary.
The same restraint applies to the assumed weights. A later event resembling one scenario would not prove that its assigned probability had been correct. An event that did not occur would not necessarily make the preparation wasteful. The decision should be judged against the evidence and purpose available when it was made, alongside subsequent learning. Otherwise, a useful planning tool can become a retrospective storytelling device in which every outcome is used to confirm the leader's preferred interpretation. I want the original assumptions retained so later review can examine them honestly.
For directors, the practical question is what a summary number is intended to support. Is management planning an average cost, deciding capacity for specified conditions, examining a stress or making a regulatory assessment through the appropriate process? Those questions may use related data but require different conclusions. Asking which question the paper answers can reveal a mismatch before the meeting becomes an argument about whether one number is conservative enough. It also helps the board respect the role of the professionals responsible for specialised estimates rather than treating a management exercise as their substitute.
I would use the fictional example to invite that discussion, not to prescribe ₹36 crore or any particular liquidity amount for another insurer. Its lesson is that a correct mean remains a conditional measure. The CFO has to preserve enough of the underlying account for decision-makers to understand what it can and cannot establish. In our case, the larger scenario and the cost of additional capacity both needed to return to view. The recommendation became stronger when it stopped asking one average to carry an assurance that belonged to a different question.