The Investment Decision: Where Capital Should Go

The investment decision is where a company puts capital and on what test. Sankalp Industrial Systems Limited, invented, commits Rs 1,00,00,00,000 of net new capital in each of five forecast years and expects 18.00 per cent on it against capital costing 12.00 per cent. Every project decision has the same four parts: an outlay now, cash after tax later, a rate to compare them at, and a rule.

A photocopy shop next to a college gate is where this decision is least disguised, so start there. The owner has one machine and a queue outside it during admission season. A second machine costs Rs 3,40,000. He will need another Rs 40,000 of paper and toner sitting on the shelf so the second machine never idles, and that money is gone from his pocket the moment he buys it even though nobody has consumed a sheet yet. Against that, he thinks the second machine brings in about Rs 90,000 a year of extra takings after the extra electricity, the extra rent for the corner it stands in, and the tax he pays on what is left.

He has just built the whole of the investment decision without using a single technical term. Rs 3,80,000 goes out first. Roughly Rs 90,000 comes back each year afterwards, and it is money he can actually take home rather than a figure on a paper. Somewhere in his head is a comparison: his cousin offered him a share of a transport business at a return he can name, and the money for the machine can only be in one place at a time. The four parts of a project decision are an outlay, a stream of cash after tax, a rate to compare them at, and a rule, and the photocopy shop has all four whether or not anyone writes them down.

Named properly, those four parts turn a proposal into something that can be checked against them, and they hold just as well at the size of a listed manufacturer, where the outlay is Rs 1,00,00,00,000 a year and the sums do not fit in anybody's head. The company throughout is Sankalp Industrial Systems Limited, a listed maker of industrial valves, precision castings and the aftermarket parts and service that go with them. Its last completed year is called Year 0.

What does the investment decision decide, and where does it stop?

The investment decision decides which uses of money get money, and it stops the moment somebody asks where that money is going to come from. Where the money comes from is a decision of its own, with a different set of people asking it, a different set of instruments, and a different failure mode. Keeping the two apart is not pedantry. The separation stops a bad project being approved because the bank happened to offer cheap credit that quarter, and it stops a good project being refused because the treasury side is having a difficult year.

The separation has a practical shape. A project is assessed on the assumption that it is funded at whatever the company as a whole is funded at, and the funding is made invisible in the cash counted. The interest does not appear in the stream. If it did, it would be counted twice, once as a cash outflow and once inside the rate the stream is compared against, and the project would be punished for something that has nothing to do with it.

The investment decision asks whether a use of money is worth more than the money, and it asks that question in a form that does not change if the funding mix changes tomorrow. That is the whole of the design. A project that is worth doing on borrowed money is worth doing on money from shareholders, and if the two answers differ, something has been counted in the wrong place.

What are the four parts every project decision has?

Whatever a proposal is called, whatever it is for, and whoever wrote it, it has four moving parts and nothing else. A new production line, a warehouse, a software rollout and an effluent plant all reduce to the same four. The four parts are more useful than they sound. The fastest way to find what is wrong with a proposal is to look for the part that is missing, and something is almost always missing.

The four split in a particular way. The first two belong to the project and nobody else can supply them: what it costs and what it brings in. The third belongs to the company and arrives from outside the project entirely. The fourth is no kind of fact at all. The rule is a choice of method, and different methods can rank the same two projects in different orders. Because of that split, an argument about a project is usually an argument about part one or part two, and an argument about part three is an argument about the company rather than about the proposal itself.

Part one, the outlay: what has to go out before anything comes back?

The outlay is everything that leaves before the returns begin, and the reason it deserves a section of its own is that most people write down only the obvious half of it. The obvious half is capital expenditureBuying or building something that will keep working for years: a shed, a press, a fitting bay. The books record an asset acquired rather than a bill for the year.: the machine, the building, the installation, the commissioning. Capital expenditure appears on a quotation and somebody signs it.

The half people miss is working capitalCash that daily trading swallows and holds onto: unpaid customer bills and goods on the shelf, netted against how long suppliers are content to wait.. A new line needs raw material sitting on the floor before it can run, and it needs finished stock waiting to be shipped, and once it is shipped the customer takes sixty days to pay. All of that is cash the company has and cannot touch. The working capital comes back at the end of the project, when the stock is sold and the last invoice is collected, but for the whole life of the project it is out. Go back to the photocopy shop. The paper is still there, so the Rs 40,000 of paper and toner is not a cost. The paper is capital, and it is as much out of the pocket of the owner as the machine is.

Sankalp Industrial Systems Limited puts Rs 18,00,00,000 into working capital in Year 1, and Rs 1,20,00,00,000 of extra revenue is expected that year, so the tie-up runs at 15.0 per cent of it. The 15.0 per cent is held right through the forecast, so every rupee of new revenue drags fifteen paise of capital along behind it. An outlay that counts only the equipment understates what the project actually takes, and the size of the understatement grows with how fast the project grows.

Part two, the stream: cash after tax, in the year it arrives

The second part is the flow of money that comes back, and three words in that sentence are doing all the work: cash, after tax, and arrives.

Cash rather than profit, for reasons the next section takes apart properly. After tax next. The company does not get to keep the part the tax authority takes, so a stream measured before tax is measuring money that never belongs to anybody here. And in the year it arrives. A rupee that shows up in Year 5 is not the same thing as a rupee that shows up in Year 1, and the machinery for handling that difference is a subject of its own. Each amount belongs in its correct year and stays there.

One thing does not go into the stream, and it is the thing people put in most often. Interest is not a project cash flow. The cost of money is already inside part three, the rate. Put it in the stream as well and the project pays for its funding twice. Who gets paid out of the stream is a different decision entirely, so the stream is what the project produces before anybody is paid for lending or investing.

Part three, the rate: where does the number being compared against come from?

The rate is the return the project has to clear, and the single most important thing about it is where it comes from. The rate does not come from the project but from what the capital costs, a fact about the company and about what people supplying it could get elsewhere at the same risk. At Sankalp Industrial Systems Limited it stands at 12.00 per cent. How it is built, from a cost of equity, a cost of debt and the weights between them, is a subject with its own long treatment, covered separately. Here the number is simply used.

Why does this matter enough to keep saying? Because the rate is the one part of the four that a project sponsor would most like to influence and has the least right to. If a proposal is short of the rate, there are exactly two honest responses: change the project so the cash improves, or accept that the money is better used elsewhere. Lowering the rate because the project is strategic is a third response and it is not honest. Lowering the rate converts a rejection into an approval without anything about the world changing.

There is a real complication here. A single company rate treats every project as though it carried the risk of the company average, and a project riskier than the average is then flattered by it. In principle the rate should reflect the risk of the cash being valued, not the risk of the entity doing the valuing. In practice a great many companies run one rate for everything. When that is defensible, when it is not, and how a project rate would be built instead are questions carried in full by a separate subject.

Part four, the rule: what does the comparison have to produce before the answer is yes?

Three parts give an outlay, a stream and a rate. An answer needs a rule that says what the comparison has to produce, and three parts do not supply one. Several such rules exist. Some ask whether the stream is worth more today than the outlay. Some ask what rate the project itself is earning and compare that with the rate the company demands. Some count the years until the money has all come back. Each has a use, each has a blind spot, and they do not always agree with one another about which of two projects to prefer.

Every one of those rules, and what to do when two of them disagree, is carried by a separate subject. Fluency in the four parts comes first. Somebody fluent in them can pick up any rule quickly, and somebody who has learned one rule and not the parts underneath it will apply that rule to a badly built stream and get a confident wrong answer.

Three of the four parts are facts to be gathered and the fourth is a method to be chosen. Two competent people can therefore agree on every number in a proposal and still disagree about it.

Why is the test built on cash and not on profit?

Because a company cannot spend profit. Cash is what it spends. The distinction sounds like a slogan until the two numbers move in opposite directions in the same twelve months, and in the year a company invests they do exactly that.

Here is the arithmetic on Sankalp Industrial Systems Limited, and it is worth going slowly. Net operating profit after taxTrading profit once tax has been deducted but before a single lender is paid, so the figure describes the business itself and not the way it was funded. at Year 0 is Rs 1,80,00,00,000. In Year 1 it is Rs 1,98,00,00,000, so it improves by Rs 18,00,00,000. Every number in that sentence is true and a proposal built on it would look very good.

Now put the capital back in. In the same Year 1 the company spends Rs 1,34,80,00,000 on capital expenditure, gets Rs 52,80,00,000 of that back in the sense that depreciationA yearly slice charged against profit to recognise that machinery wears out. Nothing moves when the slice is charged; the money moved when the machine was bought. is a charge with no money attached to it, and ties up a further Rs 18,00,00,000 in working capital. Net of all that, the Year 1 free cash flow to the firmWhat remains in cash for every provider of money once tax is paid and the year investment is made, measured before a rupee of it reaches a lender or a shareholder. comes to Rs 98,00,00,000. Against Year 1 profit of Rs 1,98,00,00,000, cash is Rs 1,00,00,00,000 lower.

Look at what the picture says. The company reinvests the same Rs 1,00,00,00,000 every year, so the gap between the two lines is Rs 1,00,00,00,000 in every single year. Both lines climb at the same Rs 18,00,00,000 a year, so by Year 5 profit after tax has reached Rs 2,70,00,00,000 while the cash the business throws off has reached Rs 1,70,00,00,000, and the distance between them has not moved a rupee. So this is not a company where something is going wrong. The company is investing, and investing always looks like this from the outside: profit up, cash down, at the same time, for the same reason.

Profit reports what a period earned. Cash reports that same figure once the money the period had to sink back in has been taken out of it, and only the second of the two is money anybody can do anything with. A decision made on the first number alone is a decision made on half the facts, and the missing half is precisely the price.

What does incremental mean, and why does it decide what goes into the stream?

Incremental means caused by the decision. A cash flow is incremental if it happens because the answer was yes and would not have happened had the answer been no. Cause decides what belongs in the outlay and what belongs in the stream, and the test is stricter than most people expect, disqualifying costs that are entirely real.

Sankalp Industrial Systems Limited faces exactly such a case. A proposed line will run in a regional warehouse the company keeps open in any case, for other reasons, and would go on keeping open if the line were never built. The warehouse costs real money. Rent is paid, staff are paid, the electricity bill arrives. None of it is caused by the project, so none of it belongs in the project outlay. Charging the project with it makes the project look worse than it is, and the company may then refuse something that would have made it better off.

The test cuts the other way just as hard. Money already spent, before the decision is taken, is not incremental either. Saying no does not bring it back. A feasibility study that cost Rs 40,00,000 last year is gone whichever way the vote goes, and putting it into the outlay to justify continuing is the oldest bad argument in the room. And a cost that appears nowhere on any invoice can still be incremental: if the new line takes over a shed the company was about to rent out, the rent it will now never collect is caused by the project as surely as the machine is.

The incremental test is a subtraction between two futures, the one where the project happens and the one where it does not, and anything identical in both futures drops out of the arithmetic entirely. Both directions of the error cost the same. Loading a project with costs it did not cause makes good projects look bad exactly as reliably as leaving out real costs makes bad ones look good.

After tax, and at whose rate?

Everything in the stream is measured after tax, and the reason is the one that governs the whole subject: the company can only use what it keeps. A share of operating profit goes in tax before anybody sees it, so a project that adds Rs 100 of operating profit does not add Rs 100 of usable cash.

The rate that matters applies to the extra profit the project creates, and it need not be the company average or the rate printed anywhere. Every figure above is built on a 25.0 per cent effective tax rateHow much of pre-tax profit a company genuinely parts with once allowances and timing differences have done their work, and it frequently differs from the rate in the headlines.. The 25.0 per cent is the assumed rate of Sankalp Industrial Systems Limited and describes no tax law anywhere. Real charges are governed by the authorities listed at the foot, whose conditions move without warning.

The rate works on Sankalp Industrial Systems Limited in a single line. Year 1 earnings before interest, tax, depreciation and amortisationTrading profit measured before funding costs, before tax, and before the charges that recognise assets wearing out, so several items are stripped from it on purpose. (EBITDA) is Rs 3,16,80,00,000 and depreciation is Rs 52,80,00,000, so operating profit is Rs 2,64,00,00,000. A quarter of that goes away in tax, leaving Rs 1,98,00,00,000, the profit number that has been running through the worked example from the start. The tax is not a detail applied at the end; it is the difference between a stream the company can spend and a stream that merely passes through it.

What does the investment decision look like for a whole company?

Everything so far has been about one project. Step back. A company reveals what it is doing with capital in three lines of its own accounts, so the same decision is easier to see at the level of the whole business.

Here are the three lines for Sankalp Industrial Systems Limited in Year 1. Capital expenditure of Rs 1,34,80,00,000. Depreciation of Rs 52,80,00,000, the charge for wearing out what it already had. And Rs 18,00,00,000 more tied up in working capital. Subtracting the second from the first and adding the third gives the capital that is genuinely new.

Take Rs 52,80,00,000 off Rs 1,34,80,00,000, add Rs 18,00,00,000 back on, and the answer lands on Rs 1,00,00,00,000 with nothing left over. The exact landing is not a coincidence of one year. Run the same three lines through the whole forecast and the answer is Rs 1,00,00,00,000 every time. Capital spending and depreciation both rise, and the difference between them stays put.

Stated plainly, the decision this company has taken is to put Rs 1,00,00,00,000 of new capital to work in each of five successive years, Rs 5,00,00,00,000 altogether, on the expectation of Rs 18,00,00,000 more profit after tax each year. By Year 5 the improvement is Rs 90,00,00,000 a year above Year 0. A company that spends heavily on capital and depreciates almost as heavily is not investing much, and the only way to see that is to run the subtraction rather than to read the spending figure.

Why is 18.00 per cent an assumption rather than a result?

Divide the reward by the price. Rs 18,00,00,000 of extra profit after tax for Rs 1,00,00,00,000 of new capital is 18.00 per cent, and the very same 18.00 per cent turns up in all five forecast years. Both halves of the fraction repeat. Against capital costing 12.00 per cent that is six points of spread, and six points sounds like a company doing well.

Stopping there means being fooled by one's own arithmetic. Nothing was measured. Somebody chose the profit line and somebody chose the capital line, and 18.00 per cent is what those two choices imply. The 18.00 per cent is an output of the forecast in the same way that the answer to a sum is an output of the numbers put into it, and calling it a result makes it sound like evidence.

The comparison sitting next to it is what makes the assumption live rather than harmless. The capital already in the ground at this company earns 15.00 per cent: Rs 1,80,00,00,000 of profit after tax on invested capitalEverything working inside the business: the amount tied up in trading plus the amount sunk into premises and equipment, whoever happened to put it there. of Rs 12,00,00,00,000. So the forecast is quietly claiming that the next Rs 1,00,00,00,000 will do better than everything before it, by three points, in every year, and the record holds nothing at all that would support the claim.

Could it be true? Certainly. New equipment often is more productive than the equipment it stands beside, and a company adding capacity into demand it already has may well earn more on the increment than on the average. But that is an argument somebody has to make, with something behind it. A document that prints 18.00 per cent without saying it is an assumption has presented a choice as a fact, and every figure downstream inherits the disguise.

What does a set of projects look like when there is more than one?

Real companies do not decide one project at a time in a quiet room. Companies decide a list, and the items on it are related to one another in ways that change what the decision even is. Sankalp Industrial Systems Limited has five under consideration, with outlays running from Rs 30,00,00,000 to Rs 2,00,00,00,000 and Rs 4,15,00,00,000 in total if every one of them went ahead. Three different relationships are sitting in that list.

The middle group is where the decision changes shape. When two projects need the same floor, clearing the rate stops being sufficient for either of them. Both may clear it comfortably and only one can be built, so they have to be set against each other rather than each against the rate. How that head-to-head comparison is resolved, and what happens when different rules resolve it differently, is carried by a separate subject.

Two separate pools of capital are in play here, and they do not add up to each other.

Both pools face the same 12.00 per cent, and nothing in the numbers settles whether approving an item would enlarge the run rate or come out of it.

Is every project chosen on a value test?

No, and pretending otherwise wastes time and damages credibility. The effluent treatment plant is required by the conditions attached to the consent under which the plant operates. The company cannot run without it. Put it through a value test and out comes a number that decides nothing. No future exists in which the company refuses and keeps its gates open.

The right question about that project is a different one, and it is a real question with real money in it: which of the available ways of complying costs least, over the whole life of the obligation rather than at the moment of purchase. Complying cheaply is a comparison between compliant options, not a comparison against the rate. The comparison has the same four parts as any other decision, but the rule at part four is minimise the cost of complying rather than clear a return.

Two other kinds of item sit in the same category. Replacement of something that has failed, where the alternative is stopping production, is not really a choice about whether. Safety work is not a choice about whether. In each case the decision is genuine but the question is which and how, never whether. Mistaking a mandatory item for a candidate is how a compliance project ends up being argued about for two quarters instead of being designed properly in one.