The Concession: The Right to Operate and Collect Toll

A concession is the right, granted by a public authority, to build a stated asset, operate it, collect from the people who use it, and do all of that for a fixed period on stated conditions. Tapti Crossing Infrastructure Private Limited earns Rs 310 crore a year under such a right, so the project lenders are secured on something that expires rather than on concrete that lasts.

A picture that has nothing to do with roads makes the point. A woman sets up a tea stall on the pavement outside a large office building. She has a table, two flasks, a gas ring and a stack of glasses. The equipment cost her something real, and the honest answer to what her business is worth would have almost nothing to do with the table. Her real asset is permission to stand in that spot at that hour, in front of those particular two thousand people going in and out. With the permission taken away and the stall moved three streets back, the table is still a table, the flasks are still flasks, and the business has gone.

Everything here is that observation, taken seriously and given a balance sheet. A crossing built across a river is enormously expensive concrete. Concrete does not send anybody an invoice, so the crossing on its own is worth close to nothing. The invoice is sent under a piece of paper saying that this particular private company, and nobody else, may charge the people who drive over that crossing, and may do it until a stated date, after which it stops. The sheet of paper is the concession, and the concession is what the money was actually lent against.

Tapti Crossing Infrastructure Private Limited, an invented single-asset toll road company, carries the worked example throughout. The company was formed to build and operate one crossing, and it has no other business and no second source of cash. The record locks a small set of figures: project cost Rs 1,800 crore, funded Rs 1,260 crore of debt and Rs 540 crore of equity, and annual revenue of Rs 310 crore against operating cost of Rs 62 crore. Scheduled principal in the modelled year is Rs 63 crore. The record supplies nothing beyond those figures, and the working stops at the exact point where they run out.

The stopping point matters more than it sounds. The record for Tapti Crossing Infrastructure Private Limited contains no concession length. A great deal of what a reader instinctively wants to know about a toll road turns on that single missing number, and the last third of this guide is about what may and may not be done when it is absent. The short version is that the absence is written down, and the division that would fill it is resisted.

What is a concession, and who grants it?

A public authority grants a private party the right to build a specified asset, to run it, to collect from the people who use it, and to do all of that for a stated period on stated conditions. The grant does four separate jobs at once. A reader who holds all four in mind will never again be surprised by anything a project financing does.

The first is construction. The document names the asset. Not roads in general, not crossings in that district, but this crossing, to this specification. The private party takes on the job of getting it built.

The second is operation. Building something and then walking away is a construction contract, not a concession. The private party stays, runs the thing, keeps it working, and carries the cost of doing so. For Tapti Crossing Infrastructure Private Limited that cost is the Rs 62 crore of annual operating cost sitting under the Rs 310 crore of revenue.

The third is collection, and this is the one that turns a civil engineering project into a business. The private party may charge the people who use the asset, and may keep what it charges. Without this, there is no revenue line at all. The vehicle would instead be a contractor being paid by somebody, and a contract to be paid carries a different risk shape.

The fourth is the period. The right runs for a stated length of time and then stops. The grant is not a sale. The asset does not become the private party's property forever; it goes back. And because it goes back, everything financial about the arrangement has an end date built into it, whether or not anybody looking at the numbers has noticed.

Who grants it? A public body, called here the granting authorityThe public body that grants the right and takes the asset back when the period ends. On a project map it appears as a source of the revenue, not as a claim on it.. The record supplies no name for it. Project mapping draws that authority as a line pointing inward, into the vehicle. The authority is where the ability to earn comes from rather than a party queueing up to be paid. The inward-pointing line stands.

All four elements travel together, and removing any one of them breaks the financing rather than merely trimming it. Without the right to collect there is no revenue to service anything with. Without the stated period there is no horizon. No horizon sounds like a gift, except that the whole structure of a project financing is built in the space between the day the cash starts and the day the right ends. Without operation the party is a contractor. Without construction the party runs an asset it did not build. Running somebody else's asset is a real arrangement, but not this one.

Why is the right the thing that gets financed, and not the road?

Because a crossing nobody may charge for produces nothing, and a lender lends against production. Once that reframing lands, a great deal of otherwise puzzling behaviour by project lenders becomes obvious.

Go back to the tea stall. If a bank were somehow to lend against that business, what would it be worried about? Not the table. The table can be replaced for a small sum and is not what generates the cash. The bank would be worried about whether she keeps her place on that pavement, whether somebody can move her, on what terms and how quickly, and what happens to the office building she is standing outside. Every real question is a question about the permission. The equipment is a rounding error in the analysis, even though it is the only physical thing there.

Now scale that up by a factor of several thousand and put it in a project financing. Tapti Crossing Infrastructure Private Limited earns Rs 310 crore a year. The company earns that because a public authority has said it may charge the people crossing the river. The concrete is identical whether that permission exists or not. Take the permission away, and the concrete stands there earning Rs 0. Not less. Nothing.

The project lenders are secured on a right rather than on concrete, so a project lender spends its time on the terms of that right and comparatively little on the engineering. None of that is a slight on the engineers. The claim is only a statement about where the cash comes from. A road that is beautifully built and cannot be charged for services no debt. A road that is adequately built and can be charged for services all of it.

A second-order consequence follows, and it explains why enforcement in a project financing looks so different from enforcement against an ordinary company. If a packaging manufacturer stops paying, a lender can in principle reach machines, land and inventory, and those things have buyers. If a single-asset road company stops paying, reaching the concrete is not much of a remedy. Concrete over a river has almost no buyers and cannot be moved. The lenders need the ability to step into the right, or to put somebody else into it, and the collecting then continues. Which assets a lender can actually reach is covered separately. Only the reason belongs here: the value is in the right, so any remedy that does not reach the right is not much of a remedy.

The same reasoning explains why the document, rather than the asset, is what a project team reads first. An outside reader looking at a road financing usually starts with traffic and construction. A project lender starts with what the right actually permits, for how long, and on what conditions. Those three settle whether there is a business at all, before anybody argues about how big it is.

What conditions ride on the right, and which of the project's own lines do they touch?

A concession is not a bare permission. A concession is a permission with obligations wrapped around it. The obligations are the mechanism by which somebody outside the project can move numbers inside it.

Three kinds of condition turn up in general terms. The three are these.

First, a tariff mechanismThe rule inside the document that sets, or puts limits around, what the operator may charge the people using the asset.. The document typically says something about what may be charged and how that may move over time. The document might set the charge, it might cap it, it might tie it to something, it might require an approval before it moves. Whatever form it takes, the effect is the same: the amount the project may collect is not purely the operator's decision.

Second, a service standardThe condition the asset must be kept in while it is operated, written into the document as an obligation rather than left to the operator's judgement.. The asset has to be maintained to a stated condition while it is being run. Potholes, lighting, signage, safety, availability: whatever the document specifies, it becomes an obligation rather than a management choice about how much to spend.

Third, a condition at the end. The asset has to be in a defined state when it goes back. The handback condition is the one readers most often skip, and it quietly creates a cost with no revenue behind it. The work has to be done at the point when the right is about to stop.

Now the part that matters for the numbers rather than for the lawyers. Look at what each of those three touches on the project's own statements. The tariff mechanism sits directly on top of the Rs 310 crore of revenue. The service standard sits directly on top of the Rs 62 crore of operating cost. Two of the project's own lines can therefore be moved by somebody who is not the operator, and there is no third source of cash to absorb the movement. The exposure is a structural feature of a single-asset vehicle rather than a comment on any particular authority: a company with five businesses has four other places to look when one line moves, and Tapti Crossing Infrastructure Private Limited has none.

The third condition, the state at handback, does not sit on either of those lines in the modelled year. The handback cost sits somewhere later, and this record does not say where or how much. The obligation is therefore real and carries no figure, and both halves of that have to be said together.

How far a granting authority's own actions can go, and what happens to a project when they do, is covered under political risk. The mechanism is what matters: the conditions exist, and two of them are the channels through which the project's own revenue and cost can be moved from outside.

Why must the debt be repaid inside the period rather than at the end of it?

Here is where a reader who has only ever looked at ordinary corporate borrowing has to put one habit down. In the corporate world, a maturity is rarely a cliff. A maturity is usually a conversation instead.

Take Harivansh Packaging Limited, the invented packaging manufacturer this subject area uses elsewhere, with borrowings of Rs 740 crore against earnings before interest, tax, depreciation and amortisation (EBITDA) of Rs 477 crore. When one of those facilities reaches its maturity, the company does not generally have to find Rs 740 crore in cash on the day. The company goes and borrows again, from the same lender or another one. The business it is borrowing against is still there: still making packaging, still selling to food and personal care customers, still generating cash next year and the year after. RefinancingReplacing a borrowing that has come due with a new one, usually from the same or another lender, so the cash does not have to be found on the maturity date. works because there is something continuing to lend against.

Now ask the same question of Tapti Crossing Infrastructure Private Limited, positioned one day after its right has ended. A new lender would need something to lend against. Not the crossing. The crossing has gone back to the granting authority. Not the revenue. No right to collect it survives. Not another business. A single-asset vehicle has no other business by definition. There is a company with no asset, no revenue and no prospect of either.

An ordinary company reaching a maturity can borrow again because the business continues. The concession that produced a project's cash has gone back at the end of the right, so a project reaching that point cannot borrow again. The contrast is the largest single structural difference between financing a project and financing a business, and everything else about project debt follows from it.

So the tenorThe length of time a borrowing runs, measured from when it is drawn to when the last rupee of principal is due. of the debt has to sit inside the period. Not touching the end of it, and in practice comfortably short of it. Nobody wants the last instalment falling due in the same week the revenue stops. And because the debt has to be extinguished rather than rolled, project debt is normally repaid by amortisationRepaying principal in instalments across the life of a borrowing, so the balance falls step by step rather than sitting there until one large payment at the end. rather than sitting at its full balance waiting for a refinancing. The Rs 63 crore of scheduled principal in the modelled year for Tapti Crossing Infrastructure Private Limited is one such instalment. The balance is meant to walk down to zero while the right is still running.

The constraint is a severe one. In an ordinary corporate structure, the amount that can be borrowed is roughly a question about how much cash the business throws off and how comfortable everybody is. In a project, there is a second constraint stacked on top: whatever the cash can service, the schedule also has to finish before a date, and the date is set by a document rather than by the market. Two constraints bind instead of one, and the tighter of the two wins.

What is handback, and what do the sponsors hold afterwards?

At the end of the period the asset returns to the granting authority in whatever condition the document requires. The return is called handbackThe return of the asset to the granting authority when the period ends, in whatever condition the document requires it to be in., and it is the last thing that happens in the sequence.

An ending that leaves nothing behind is unusual enough that ordinary corporate intuition gets it wrong every time. At the end of a normal company's life there is a company left. The company has factories, customers, a brand and a balance sheet, and whatever all of that is worth, it is worth something, and it belongs to the shareholders. At the end of Tapti Crossing Infrastructure Private Limited there is a company left with nothing in it. The crossing has gone back. There is no residual asset for the sponsors to sell, wind up or carry forward.

So what do the sponsors hold at the end? Only the cash they took out along the way. Nothing else. No other asset exists to hold. In the modelled year, Tapti Crossing Infrastructure Private Limited had Rs 65.3 crore available to its sponsors, being EBITDA of Rs 248 crore less debt service of Rs 182.7 crore. The build-up of that debt service figure is set out under debt service. Against equity of Rs 540 crore, Rs 65.3 crore is a cash return of 12.09 per cent in that year.

One qualification has to travel with that 12.09 per cent wherever it appears. The figure is one year's cash on one year's equity, and it is not a return over the life of the right. The record carries no period over which a life return could be struck. A reader who mentally attaches the word "annual" to it and pictures it repeating has made exactly the error that distinction guards against.

The consequence for how the equity has to be read is sharp. In an ordinary company, an equity holder's return has two parts: the cash received along the way, and the value of the holding at the end. Here the second part is zero by construction. The residual is the whole of the return, so every rupee the sponsors are ever going to see has to come out during the period and not after it. That is why sponsors in this structure care so intensely about the years in the middle, and why anything that delays a distribution costs them more than the same delay would cost a shareholder in a continuing business.

There is a household version of this. Buying a flat differs from taking a long lease on one. In the first case, whatever the buyer paid, at the end of the time in it there is still a flat, it belongs to the buyer, and it has a value. In the second, at the end of the lease the keys go back, and everything the tenant got out of the arrangement was the living done in it. Nobody confuses those two when it is a flat. People confuse them constantly when it is a road.

What does this record actually say about how long the right runs?

Nothing. The silence is the whole answer, and it is deliberate.

The record for Tapti Crossing Infrastructure Private Limited carries a project cost, a funding split, a revenue line, an operating cost line, a debt balance and one year's scheduled principal. The record carries no concession periodThe stated length of time for which the right runs. When it ends, the right to operate and to collect ends with it., no debt tenor, no traffic forecast and no year by year schedule.

A reader's first instinct is that this is a gap in a teaching example, the sort of thing that would obviously be filled in on a real transaction. A concession length is frequently one of the most commercially sensitive figures in a project, and often the last thing an outside reader ever gets, so the absence is not a defect in the example. Reading published material about infrastructure and finding that the period is described loosely, or in a way that does not tie to the numbers presented, or not at all, is an ordinary experience rather than an unlucky one.

Which is why the discipline being taught here is a working skill rather than a classroom rule. Anybody who reads project material for a living spends a considerable part of their time holding an incomplete record and deciding what may still be said. The wrong response is to stop. Plenty can still be said. The other wrong response, and far the more expensive one, is to quietly manufacture the missing piece so the analysis can proceed as though it were complete.

A missing concession length has a useful name. The absence is a known unknownA figure that can be named as missing, as against one whose absence has gone unnoticed. The first can be asked for; the second silently corrupts whatever is built on top of it.: a figure identified as absent, and an identified absence can be pointed to and asked for. A named absence is a workable position. The dangerous position is the one where the figure has been silently replaced by an estimate. Then nobody knows to ask.

The temptation, when the period is missing, is to divide one number by another.

What does the missing period forbid the analyst from computing?

Four things go at once. A list of them can be handed to somebody, and a gesture cannot.

First, a return over the life of the right. Any measure that spreads the sponsors' cash across the whole period needs the whole period. Rs 65.3 crore in the modelled year is a fact about that year. Turning it into a return over the life needs to know how many years there are and what happens in each of them, and the record contains neither.

Second, a payback on the Rs 540 crore of equity. Payback asks how long it takes for the cash out to add up to the cash in. Payback has an answer only if the cash in each year is known, and knowing the cash in each year requires the years.

Third, the total collected before the crossing goes back. Rs 310 crore of annual revenue multiplied by an unknown number of years is an unknown number, and it stays unknown no matter how confidently it is written.

Fourth, and this is the one that actually matters commercially, whether the debt retires before the right ends. Whether the debt retires in time is the single question the entire structure turns on, as the comparison with refinancing showed. Checking it needs two dates: when the last instalment falls due, and when the right stops. The record supplies neither.

Every one of those four is a question a reader will genuinely want answered, and the correct output is the question written down as unanswerable rather than an estimate wearing a confident tone. That distinction is the professional habit. An estimate wearing a confident tone reads like work. A written absence reads like a gap. And yet the written absence is the more valuable object. The person who receives it knows exactly what to go and ask for. The person who receives the estimate does not know there is anything to ask.

The four absences are not a reason to say nothing. The figures the record does supply carry a real computation through to a finding. Refusing to invent and refusing to work are different things. A reader who confuses them ends up producing nothing at all.

What can still be computed, and what does it actually establish?

Now do the work the record does support, and stop visibly where it stops.

Tapti Crossing Infrastructure Private Limited borrowed Rs 1,260 crore. Its scheduled principal in the modelled year is Rs 63 crore. Take the first thing that is simply true: Rs 63 crore against Rs 1,260 crore is exactly 5.0 per cent of the amount borrowed. And take the second: the balance in the modelled year falls from Rs 1,260 crore to Rs 1,197 crore. Both of those are arithmetic and neither needs anything the record does not contain.

Now the honest stop, and it needs stating in full because the temptation to keep going is strong. The record does not say whether Rs 63 crore is the instalment in every year. Whether it steps up is not stated. Whether there is a larger final payment is not stated. Neither the date the debt is due to be fully repaid nor the length of the right appears anywhere in it. So the tenor cannot be derived, the period cannot be derived, and the relationship between the two cannot be checked at all. The relationship between them is the question the whole structure turns on.

And yet something real survives, and that makes the result a finding rather than a shrug. The balance is coming down at 5.0 per cent of the original amount in the modelled year, a slow rate, so the schedule in the years after this one has to do a great deal of work. Rs 1,197 crore is still outstanding after a full year of scheduled repayment. Whatever shape the rest of that schedule takes, it has to remove all of it, and it has to do so before a date that this record does not disclose.

The finding is worth handing to somebody, and notice what it does. The finding does not answer the original question. The finding tells the reader precisely which document to go and ask for, and what to expect inside it: a schedule that gets steeper. If it turns out to be flat all the way through, that is now a question rather than an assumption. The finding has converted an absence into a specific request, and specific requests get answered.

Note also what has not been done. No year has been counted. No date has been named. No shape has been assumed. Every figure used here appears in the record, and every conclusion is a restatement of those figures.

Why no control can be built on an absent period

An interactive control needs something a reader can move: a slider, a set of inputs, a picture that redraws as a number is pushed around. None of that fits a record that carries no period.

Such a control would have to move the length of the right, the shape of the repayment schedule, or the year the debt retires. Every genuinely interesting control on this subject runs along time, and this record carries no period and no schedule. Any such control would therefore have to be built on figures that were invented for the purpose of making the control work.

A slider built on an invented period would teach a reader to picture a timeline the record never supplied, and that picture is exactly the habit the missing period forbids. Moving it and watching a bar redraw leaves behind a mental image of this project's life. The image would be a product of the slider rather than of the record, and it would be far more persuasive than a sentence saying the period is unknown. Pictures always are.

So the refusal is the lesson. The correct move is sometimes to build nothing at all. No timeline can be drawn here, and the reason is the same reason the four measures in the earlier list cannot be computed.

How do a lender, a sponsor and an analyst actually use this?

Three readers pick up the same concession and go straight to different clauses of it, and which clause each one turns to shows what each of them is afraid of.

A project lender goes to the period first and to the conditions second, and it does both before it looks seriously at the traffic. The period comes first because the entire repayment schedule has to be constructed inside it, and no version of the analysis starts anywhere else. The conditions come second because the tariff mechanism and the service standard are the two channels through which the cash it is relying on can be moved by somebody who is not its borrower. Only once those are settled does the size of the cash become the interesting question. A reader who assumes lenders start with traffic forecasts has the order backwards.

A sponsor reads the same document as a distribution question. Since the residual at the end is zero, everything the sponsor is ever going to receive has to come out during the period. So a sponsor is unusually sensitive to anything that delays or blocks a distribution: a reserve that has to be topped up first, a condition that has to be cured before cash may leave, a maintenance obligation that lands in a heavy year. In a continuing business, a delayed distribution is a timing annoyance. Under a finite right, a delayed distribution can be a permanent loss. There may not be enough years left for it to come back.

An analyst outside the transaction reads it as a set of questions to ask, and the useful skill is knowing which absence is expensive. If the traffic forecast is missing, a range can still be reasoned about. If the operating cost is missing, the margin gives a handle on it. If the period is missing, four separate measures go at once, as the earlier list showed, and no amount of cleverness elsewhere puts them back. So the period is what an analyst chases first, and the honest interim output is the finding that the balance fell 5.0 per cent, real enough to hand over while the chasing continues.

The household version pulls all three together. If a relative had taken a shop on a lease and were borrowing to fit it out, two questions follow almost without thinking. How long is the lease? And does the borrowing finish before it does? The lease question and the finish-first question, asked in that order, are the whole of a project lender's opening position, and everybody already knows to ask them about a shop. The only thing that changes at Rs 1,260 crore is the number of zeros.