It is the first Sunday of April and the annual review is spread across the dining table: statements, a spreadsheet, last year's notes. The conversation covers the fund that was switched, the stock that was sold in November, the SIP that was increased, the holding that has doubled. Four decisions, four numbers, two hours. What does not come up at all is the six decisions that produced nothing to look at — the cover renewed, the reserve left untouched, the leveraged idea a colleague described in July and did not take up, the tip from the cousin that was ignored, the second property not bought, the account limits tightened in March. Some of those were good and some were probably not, and the review has no way of telling them apart, so it silently drops all six. This is not carelessness. A review is built to read results, and these decisions did not produce any — which means the half of the year with no numbers goes ungraded every single year, and it contains most of the things that determine whether the household is still solvent in a bad decade.
A careful doctor writes the working diagnosis in the file before the scan comes back — what she thinks it is, what would change her mind, and what she would do either way. It takes a minute and it seems redundant, because the scan will settle it. Its value shows up only afterwards: whatever the scan says, she can now tell whether her reasoning was sound, instead of discovering that she had "always suspected" whatever turned out to be true.
A decision that produces no result cannot be graded afterwards, because there is nothing to grade and the mind will supply a story with whatever ending is convenient. It can only be graded against something recorded before the period began. Written first, the reason survives the year. Written afterwards, it is a reconstruction with the answer already in it.
Two sentences that sound identical and are not
Almost the whole difficulty is contained in a pair of sentences that any household will use interchangeably at a review, and which make entirely different claims.
- "I did not need it." An observation about the year that occurred. It is very likely true and it is nearly always available, because most years are quiet ones.
- "I would not have needed it." A claim about a year that did not occur. Nothing in your experience supports it, and nothing ever will, because the world in which the event happened is not one you can visit.
The first slides into the second without anybody noticing, and once it has, the decision looks retrospectively unnecessary — which is the mechanism by which protection gets dropped, reserves get deployed and constraints get relaxed, always with a feeling of clear-sightedness rather than of risk-taking. It is the same slide that makes every past crisis look predictable: what actually happened acquires an air of inevitability, and everything that did not happen becomes something that was never going to.
What can be graded, when the result cannot be
An earlier module in this track separated process failure from outcome failure — the audit for a decision whose result you already know. This is the case that audit cannot reach, because there is no result at all. Three things remain gradable, and none of them requires knowing what would have happened.
| Gradable | The question | What a bad answer looks like |
|---|---|---|
| The exposure | Does the thing this decision addressed still exist, and is it still the size it was? | "The loan is repaid and both of us are earning now, and the cover has not been revisited since it was taken." The exposure has changed and the arrangement has not — the honest finding may be that it should be reduced |
| The price | What is this costing, in money and in what else the money could have done? | "I have never worked it out." A protection decision has a price and it is allowed to be too high — a review that can only ever conclude "keep it" is not a review |
| The reachability | On the day this is needed, can it actually be got at, by whoever will need to? | "It is in a scheme with a lock-in, and the login is on my phone." A reserve that cannot be reached in the week it is wanted has the cost of a reserve and not the function |
The register, and what goes in it
- 1Write the line when the decision is taken, not at the review
One line: what was decided, what it is protecting against, what it costs a year, and what would make you change it. Thirty seconds. The whole value is in the date on it, because a reason recorded before the outcome is the only kind that cannot be edited by the outcome.
- 2Include the decisions not to act
The leveraged idea declined, the tip ignored, the property not bought. These are decisions with consequences and they leave no trace anywhere else — no transaction, no entry on a statement, nothing for the review to find. They are also where a household's discipline actually lives.
- 3Write down what would count as it working, in advance
Not "returns", because there will not be any. Something like: "if either of us stops earning, the household runs for eight months without selling equity." That sentence can be checked in April against the actual balance, and it can be checked without a crisis ever having occurred.
- 4Review it on the same day as the numbers, in the same sitting
Separating the two guarantees the second never happens. Read the register first, before the statements — once the year's returns are in the room, everything that produced no return is being compared against them, which is the substitution the whole exercise exists to prevent.
- 5Let it conclude "reduce" or "stop" sometimes
A register that only ever ratifies is a comfort object. The exposure genuinely does shrink — children start earning, a loan is cleared, a second income arrives — and when it does, the honest entry is that less of this is now needed. That is what makes the entries that say "keep" mean anything.
The line item nobody can defend
Four years ago you set aside a reserve equal to eight months of household expenses, in a liquid form earning modestly. Equities have run hard since. Your brother-in-law points out, correctly, that the reserve has cost you a substantial amount in foregone returns and that nothing has gone wrong in four years. Both incomes in the house are secure.
The five shapes in this module
Each of these is a decision with no result to point at. What is actually being asked?
Module checkpoint: when nothing happens
5 questions. Answers are revealed once you submit all of them.
1.A household has paid ₹14,000 a year for eleven years for pure protection against ₹1 crore, with no claim. Which reading of that record is sound?
2.A family lets a six-year-old health policy lapse during an income gap, intending to buy fresh cover once work resumes. Beyond the months without cover, what has the lapse cost?
3.An approach carries roughly a 5% chance each year of an outcome the household could not recover from. It has been run for four years without incident. What follows?
4.After three strong equity years the defensive portion of a portfolio has badly lagged. Which of these is a sound reason to reduce it?
5.At an annual review, which of these can be graded for a decision that produced no result at all?
April ki review mein chaar faislon ke number hain, aur chhah faislon ka koi number nahi — cover renew kiya, reserve ko haath nahi lagaya, udhaar wala idea chhod diya, cousin ki tip ignore kar di. Review chhah ke chhah chup-chaap chhod deti hai, kyunki unka koi result hi nahi bana. "Mujhe zaroorat nahi padi" ek beete saal ka observation hai; "mujhe zaroorat padti hi nahi" us duniya ka daawa hai jahaan aap kabhi gaye hi nahi. Teen cheezein phir bhi jaanchi ja sakti hain, bina yeh jaane ki kya ho sakta tha: khatra abhi bhi utna hi bada hai ya nahi, saal ka kharch kitna hai, aur zaroorat wale hafte woh paisa haath mein aa bhi paayega ya nahi. Aur wajah faisle ke din likho, review ke din nahi — baad mein likhi hui wajah mein jawab pehle se ghusa hota hai.
- "I did not need it" is an observation; "I would not have needed it" is a claim about a world you never visited.
- Silent evidence: outcomes that did not occur leave no record, so any tally built from what you can see is skewed.
- A prevention decision cannot be graded on results. Exposure, price and reachability can all be graded without them.
- Write the reason and the test when the decision is taken — afterwards, the reason is reconstructed with the answer in it.
- Read the register before the statements, and let it sometimes conclude that less of this is now needed.
Mark it done to track your progress through the curriculum.
Common questions
Short, direct answers to what people ask about this topic.
- silent evidence meaning
- Silent evidence is the set of outcomes that did not occur and therefore left no record, which makes any tally built from what you can see systematically skewed towards the things that happened. It is survivorship bias arriving from the other direction: in a list of successful investors the failures are missing from the sample, and in your own financial history the disasters you were protected against are missing from yours. Either way the visible record is not a fair sample of what you are trying to judge.
- a claim about what would have happened in a year that never occurred is called a
- A counterfactual — a statement about a world you did not live in, which is why no amount of experience can confirm or refute it. The everyday version is the slide from “I did not need it” to “I would not have needed it”: the first is an observation about the year that happened and is usually true, the second is a claim about a year that did not. Once the slide has happened the decision looks retrospectively unnecessary, which is how protection gets dropped and constraints get relaxed.
- how do I review a decision that produced no result
- Grade the three things that do not require knowing what would have happened: whether the exposure it addressed still exists at the size it did, what the arrangement costs in money and in what that money could otherwise have done, and whether it can actually be reached on the day it is needed by whoever will need it. All three are answerable at a review from documents alone. What can never be graded is whether the event would have occurred, and no method available to anybody reaches that.
- what should a financial decision register contain
- One line written on the day each decision is taken: what was decided, what it is protecting against, what it costs a year, and what would make you change it. Include the decisions not to act — the leveraged idea declined, the tip ignored, the property not bought — because those leave no trace on any statement and nothing else in a review will find them. Read the register before the statements at the annual sit-down, so that the things which produced no return are not being read next to the year’s winners.
- is an emergency fund still worth keeping if nothing has gone wrong in four years
- Four quiet years are the expected output of a reserve rather than evidence about its size, because a reserve exists for events that are rare by definition — that record is equally consistent with it being essential and with it being far too large. What can be checked is whether the exposure has moved since it was set: notice periods, the number of earners, fixed obligations, whether a parent has become dependent, whether both incomes now come from the same sector. Resizing from those numbers is a different decision from abandoning it because nothing happened.