Investment Committee: How Investment Decisions Are Governed

An investment committee is the group empowered to approve or reject an investment before money moves, deliberately separate from the people who found and proposed it. The committee works from a written memo stating the thesis, the base case, the downside, the sensitivities, the risks and the exact ask, and it decides by discussion in which dissent is recorded. The separation is the point: enthusiasm proposes, scrutiny approves.

The person who found a deal loves it. Loving the deal is not a flaw; nobody carries a project through six months of site visits, supplier quotes and spreadsheets without coming to believe in it. But belief is exactly the wrong state of mind in which to decide whether the money should move. So somebody who did not find it, and did not fall in love with it, has to decide, or every deal gets done. An investment committeeA standing group inside a company, fund or institution that must approve an investment before the money is committed. The group reads a written proposal and votes; it is not the team that wrote the proposal. is the structure that puts a second mind in the way of the money, working through a written memo, a meeting with a recorded vote, and a review afterwards of what the memo promised, all of it followed below on a Rs 90,00,00,000 capital project.

What is an investment committee, and how does it differ from the board?

The plainest version of the structure sits inside a household. A household where one person wants a new car and the same person signs the loan papers has no committee; it has a wish with a chequebook. A household where the person who wants the car has to explain the numbers to the other earner, who then says yes, no, or yes provided the top model is dropped, has one. Nothing formal has been created, but the decision now passes through someone whose enthusiasm is not already committed. The second earner is the committee. The grandparents who hear about it afterwards, and who would step in if the household kept buying cars it could not afford, are the board.

An investment committee approves individual investments; the board sets the mandate within which the committee may approve them and oversees the results. Inside a company such as Aravalli Agro Foods, an invented listed maker of packaged snacks and staples, the board decides that any capital project up to an approval thresholdThe largest amount a committee may commit on its own authority. Anything above the threshold goes up a level, usually to the board, with the committee's recommendation attached. of Rs 25,00,00,000 may be cleared by the investment committee alone, and that anything larger goes to the full board with the committee's recommendation attached. The committee meets monthly, works from written papers, and reports its decisions to the board each quarter. The board does not re-run the committee's work; it checks that the committee is doing it. Inside a fund the same shape holds: the fund's committee approves each investment, and the trustees or governing body oversee the committee. The mechanism is the same in both settings, and the company case carries numbers that are already familiar.

The membership matters as much as the mandate. Aravalli Agro Foods' committee has four members: Nikhil Sarin, the Chief Financial Officer, who chairs it; Suresh Menon and Kavita Rao, both independent directors; and Devika Rathore, the Managing Director. A quorumThe minimum number of members who must be present for a meeting's decisions to count. Below quorum, the meeting can talk but cannot decide. is three, and at least one of the three must be an independent director. Notice who is not on it: the head of operations who wants the new plant, the sales head who wants the new territory. Both present. Neither votes.

Where does the money threshold bite? Aravalli Agro Foods' committee can clear a Rs 12,00,00,000 packaging line on its own authority. The Rs 90,00,00,000 capital project for a third plant sits far above Rs 25,00,00,000, so the committee's job is to examine it and recommend, and the board's job is to decide. And notice something about the other item on the same agenda season, the warehouse Devika Rathore's cousin wants to sell the company for Rs 34,00,00,000: because a related party is involved, that one takes the audit committee route with the promoter recused, not the investment committee route. Different questions, different rooms.

Why separate the people who propose from the people who approve?

Think about a street vendor deciding whether to buy a second cart. The vendor has walked the new corner, counted the office workers, priced the cart, and can already see the queue. Ask that vendor whether the second cart is a good idea and the answer was settled weeks ago; the spreadsheet, if there is one, was built to confirm it. Now ask the vendor's sister, who lends the money and has never seen the corner. She asks what happens in the monsoon, whether the office building has a canteen, and what the vendor will do if the municipality moves the stalls. Her questions are the ones that decide whether the cart pays, and the vendor could not ask them, not from dishonesty but from love.

The separation of proposal and approvalThe rule that the person who brings an investment forward is not the person, or the group, that decides whether to do it. The proposer argues; a different group judges. exists because the person who found an investment has already decided, and a decision made once cannot be made again by the same mind. A sponsorThe person or team who brings an investment proposal to the committee, argues for it and answers the committee's questions. The sponsor writes the memo but does not vote on it. is not asked to be neutral; that would be asking for the impossible and would lose the energy that found the project in the first place. The sponsor is asked to be complete and honest, and to accept that a different set of eyes decides. The committee is not asked to be brilliant; it is asked to be unpersuaded until persuaded, and to ask the questions the sponsor's belief has quietly stopped asking. Three things follow. The sponsor writes; the committee reads before the meeting, not during it. The sponsor presents and answers; the sponsor does not vote. And where a committee member has a stake in the proposal, that member steps out for the vote, on the same logic as recusal anywhere else.

At Aravalli Agro Foods the sponsor of the capital project is Vikram Sethi, the Chief Operating Officer, an invented executive who has spent a year on the third-plant plan. He is right to want it. He is also the wrong person to approve it, and the company's structure says so by keeping him off the committee that hears him.

What must a committee memo contain?

A committee cannot ask good questions of a presentation; it can only ask them of a document. The memoThe written proposal a sponsor sends to the committee before the meeting: what the investment is, what it should earn, what could go wrong, and exactly what approval is being asked for. turns the sponsor’s belief into claims that can be tested one at a time. Its structure is a checklist, and each section is a question the committee would otherwise have to ask out loud. A memo carries seven things: the thesis, the base case, the downside, the sensitivities, the stress case, the risks, and the exact ask, and a memo missing any one of them is a pitch wearing a cover sheet. Four of those sections are the analysis tools set out under risk and return: the base case is the expected path; the scenario work gives the downside; the sensitivity work shows which single assumption moves the answer most; the stress case pushes the bad things together and asks what breaks first. The memo does not reteach them. The memo uses them, in that order, on one project.

The thesis is one paragraph: why this, why now, why this firm. The base case for Aravalli Agro Foods' third plant is a new extruded-snacks line costing Rs 90,00,00,000, commissioning twelve months after approval, adding Rs 24,00,00,000 of earnings before interest, tax, depreciation and amortisation (EBITDA) a year at full run from the second year, and, after Rs 6,00,00,000 of depreciation and Rs 8,10,00,000 of interest on the new borrowing at an illustrative 9 per cent, lifting profit before tax by Rs 9,90,00,000 against today's Rs 72,00,00,000. The downside scenario is the one everyone in the company can name: the single national retail chain that takes 18 per cent of sales cuts volume by a fifth, and the uplift falls from Rs 24,00,00,000 to Rs 16,00,00,000. The sensitivities take four assumptions one at a time. The stress case lands the bad things together and asks about the covenant. The risks section lists what is not in the numbers: a competitor's plant opening across the road, a change in the edible-oil duty, the new line's operators taking a year longer to reach yield. The ask is exact: approve Rs 90,00,00,000 of capital expenditure, funded by a Rs 90,00,00,000 term loan, with authority to sign the equipment contracts within ninety days.

The sensitivity section does a job the base case cannot. Look at it on its own for a moment. The base case is one story. The sensitivities tell the committee which part of the story to doubt first. For the third plant, moving each assumption on its own gives four very different hits to the Rs 24,00,00,000 uplift: raw material and packaging cost up 10 per cent takes Rs 12,00,00,000 off it, the retail chain cutting a fifth of its volume takes Rs 8,00,00,000, commissioning slipping by a quarter takes Rs 6,00,00,000, and power and freight up 10 per cent takes Rs 3,00,00,000. Input cost dominates, so input cost is the assumption the committee should spend its questions on, and the memo says so before anyone asks.

How does the committee decide, and what makes a good discussion?

Picture a wedding budget meeting where the person who wants the bigger venue talks for the whole hour and everyone else nods; then picture one where an aunt who is paying half the bill has read the quotes the night before and asks why the caterer's number assumes 400 guests when the list has 310. The second meeting is uncomfortable and it is the one that protects the household. A committee meeting is designed to be the second kind. The quality of a committee decision is set by whether the questions the sponsor did not want to hear get asked, answered and written down, and by whether disagreement survives the vote on paper.

The sequence runs the same way every month. First, the memo is circulated at least a week ahead and members read it; a member who has not read it does not vote. Second, the sponsor presents briefly and takes questions, and the chair's job is to make sure the questions reach the weak sections rather than the strong ones. Third, a pre-mortemA discussion held before a decision in which the group assumes the investment has already failed and works backwards to the most likely reasons. A pre-mortem gets doubts said aloud before the vote rather than after.: the chair says, it is two years from now and this project has failed; tell me why. The answers that come out are the risks nobody wanted to write in section six. Fourth, the sponsor leaves the room and the committee votes; a simple majority carries, but the chair may set conditions on the approval, and conditions are how a committee says yes and no at the same time. Fifth, dissentA member's disagreement with the decision the group reached, stated and written into the minutes so that it can be checked against what actually happens later. is recorded by name and reason. Sixth, the decision, the conditions and the dissent go into the minutes and, above the threshold, into the recommendation to the board.

Why insist on the pre-mortem when there is already a risks section? Because a risks section is written by the sponsor, and the sponsor's list is the list of risks the sponsor has already answered. The pre-mortem asks the people who did not write the memo to imagine the failure, and their list is different. At Aravalli Agro Foods it was Suresh Menon, in the pre-mortem, who said the failure he could see was not the plant but the bank: the money is drawn on day one and the earnings arrive in year two, so what does the covenant certificate say in the quarter after drawdown? Vikram Sethi's memo had a stress case; it stressed volumes and costs and had never once mentioned the loan agreement. The loan agreement is the gap the simulation below is built around.

What happens after: tracking and post-mortems?

A household that borrows for a shop fit-out and never looks at the shop's takings again has not made a decision; it has made a wish and walked away. The vote was made on a memo full of predictions, and predictions can be checked. So the committee’s work does not end at the vote. Every approved memo becomes a promise the company tracks against, quarter by quarter, and every finished project gets a post-mortemA review held after an investment has run for long enough to judge, comparing what the memo predicted with what actually happened, and asking what the committee should learn. Named for the medical examination held after a death; here nothing has necessarily died. in which the memo is reread against what happened. The tracking is mechanical: the base case said Rs 24,00,00,000 of EBITDA a year from the second year, so each quarter's report to the committee shows the actual against Rs 6,00,00,000 a quarter, and the variance is explained in a paragraph. The post-mortem, held about two years in, is the interesting part. The post-mortem asks three questions. Was the thesis right? Which assumption was wrong, and was it the one the sensitivities flagged? And what did the dissent say?

The third question is why dissent is recorded by name and reason. Kavita Rao's dissent on the demand assumptions was not a protest; it was a prediction, and a recorded prediction can be tested. If, two years on, the line is running at 70 per cent of the assumed volume, the post-mortem does not say the project was unlucky. The post-mortem says the demand assumption was the weak one, a member said so at the time, and the committee's next demand assumption from the same sponsor gets a harder look. A committee that records dissent and rereads it learns; a committee that only records approvals cannot tell luck from judgement.

How does the capital project go through Aravalli Agro Foods' committee?

All of it comes together on the one project. Vikram Sethi's memo reached the committee eight days before the meeting. Thesis: both existing plants are running near capacity and the distributor network of 2,200 can place the volume, so a third plant for extruded snacks. Base case: Rs 90,00,00,000 of capital expenditure, twelve months to commission, Rs 24,00,00,000 of EBITDA a year at full run, Rs 9,90,00,000 of extra profit before tax. Downside: the retail chain at 18 per cent of sales cuts a fifth of its volume, and the uplift falls to Rs 16,00,00,000. Sensitivities: input cost dominates. Stress: input cost and the chain cut together, and the uplift is Rs 4,00,00,000 in a bad year. Risks: a rival plant, a duty change, a slow ramp. Ask: approve Rs 90,00,00,000, all debt, sign within ninety days.

The committee's questions went first to input cost, as the sensitivities said they should, and Vikram Sethi had good answers on hedging film prices and pass-through to shelf price. Then Suresh Menon asked the covenant question. The memo did not contain an answer. Here is the arithmetic the memo had left out. Aravalli Agro Foods' net debt is Rs 2,40,00,00,000 against EBITDA of Rs 1,05,00,00,000, leverage 2.3 times against a covenant ceiling of 3.0 times. Draw the full Rs 90,00,00,000 on day one and net debt is Rs 3,30,00,00,000. The plant earns nothing for twelve months. So at the first covenant test after drawdown, trailing EBITDA is still about Rs 1,05,00,00,000, and leverage of Rs 3,30,00,00,000 over Rs 1,05,00,00,000 comes to 3.1 times. The memo's stress case had stressed the plant and never stressed the balance sheet, so the one covenant the project breaks was the one thing the memo did not mention. Interest cover holds: with Rs 8,10,00,000 of new interest, total interest is Rs 25,10,00,000 and cover is 4.2 times against a 3.0 times floor. Leverage is the covenant that binds.

The decision was neither yes nor no. The committee recommended approval subject to a condition: Rs 36,00,00,000 of the Rs 90,00,00,000 to come from equity, through a rights issue in which Devika Rathore takes up her 52 per cent share. New borrowing is then Rs 54,00,00,000, net debt at drawdown is Rs 2,94,00,00,000, and leverage on today’s EBITDA is Rs 2,94,00,00,000 over Rs 1,05,00,00,000, or 2.8 times, inside the ceiling with 0.2 times of room. The 0.2 times of room means trailing EBITDA can fall to about Rs 98,00,00,000, roughly 7 per cent, before the ceiling is touched during the building year. Once the plant is at full run and EBITDA is Rs 1,29,00,00,000, leverage falls back to about 2.3 times, where it started. Kavita Rao voted against, on the ground that the volume assumption for the new line rested on the same retail chain that already takes 18 per cent of sales; her dissent was minuted by name and reason. The vote was three to one. The recommendation, the condition and the dissent went to the board, and the board approved on those terms.

How do lenders, analysts, investors and households actually use the idea?

A lender reads a borrower's investment committee as part of the credit file. When Aravalli Agro Foods' bank asks how capital projects are approved, it is asking whether somebody other than the enthusiast decides, whether the memo tests the covenants, and whether the minutes exist. A borrower whose committee caught the 3.1 times problem before drawdown, and came to the bank with an equity plan, is a borrower whose next request gets read faster.

An analyst reads the committee through outcomes. The annual report will not print the memos, but it will print the capital expenditure and, two years later, the segment earnings, and an analyst can ask on the call whether the third plant is running at the volume the company guided to. A company that can answer with the memo's number and the actual next to it has a committee that tracks; a company that changes the subject may not. Practitioners rarely ask whether a company has an investment committee; they ask whether the people who approve are different from the people who propose, and whether anyone rereads the memo afterwards.

An investor, including the retail investor Farida Shaikh who holds a few hundred Aravalli Agro Foods shares, uses the idea in a smaller way: a board that requires written memos, records dissent and holds post-mortems is a board that is harder to walk a bad project past, and that is worth more than any single quarter's number. And a household can run the same discipline at kitchen-table scale. The person who wants the thing writes down what it costs, what it earns or saves, what happens if the income that pays for it stops, and exactly what is being asked; the other earner reads it before the conversation, not during it, and says yes, no, or yes if. A household that runs it that way has an investment committee of two.

Aravalli Agro Foods Limited, Devika Rathore, Suresh Menon, Kavita Rao, Nikhil Sarin, Vikram Sethi, Ritu Iyer, Sameer Kulkarni, Manav Das, Farida Shaikh, the retail chain, the lender and the capital project are invented.
Educational material. Not advice on any investment, tax, budget or market position.