Case 081Structured finance and securitisationHard

An expressway operator wants to borrow against ten years of future toll collections. Size the issue at 1.5x cover, then show what a 20% traffic shortfall does to that cover.

The situation

Varadha Expressway collects about Rs 150 crore of tolls a year and its traffic study projects 5% growth a year. It wants to raise money now by selling bonds backed by the next ten years of toll collections, which will be paid into an escrow account controlled by a trustee before Varadha sees them.

Investors require debt service cover of 1.5x on the base projection. The bonds pay 8.5% and amortise over ten years, with each year's debt service set at the projected toll divided by 1.5, so debt service grows with the projection. The rating agency will test a stress case in which traffic starts 20% below the base and grows only 2% a year.

Your task

How large an issue do the base projections support, what does the stress do to cover, and how much would you actually raise?

Quick check

Debt is sized at 1.5x cover on the base tolls. Traffic comes in 20% lower than projected. What is cover in year one?

Worked solution

Try it on paper, then open one step at a time.

30-second answerThe answer to give first

The base case supports about Rs 799 crore, but I would raise about Rs 738 crore, sized on the stress case. A 20% shortfall alone takes cover to 1.2x. Add growth of 2% instead of 5% and, because debt service was sculpted to rise 5% a year, cover falls below 1.0x from year 8 and to 0.92x by year ten. Sizing so the stress never breaks 1.0x cuts the issue by Rs 60 crore.

Step 1How do you size a deal on cash that has not been collected yet?

Set each year's debt service at what the tolls can carry, then ask what that stream of payments is worth today. With cover of 1.5x, debt service is Rs 100 crore in year one, growing 5% a year with the projection to Rs 155 crore in year ten. Discounted at 8.5%, that sculptedA repayment schedule where each year's debt service is set to a fixed fraction of that year's projected cash flow, rather than being equal every year. schedule supports an issue of about Rs 799 crore. It is like a shop owner borrowing against next year's rent from a tenant: the lender cares about the rent actually arriving, not the building's paper value.

The relationship

\text{Issue} = \sum_{t=1}^{10} \frac{150 \times 1.05^{t-1} / 1.5}{1.085^t} \approx 799

150	year one toll collections, Rs crore
1.05	projected traffic growth, 5% a year
1.5	required debt service cover
1.085	the bond rate, 8.5%

What it says in wordsEach year's debt service is two thirds of that year's projected tolls, and the bonds are worth the present value of those payments, about Rs 799 crore.

Step 2What does a 20% traffic shortfall do to cover?

On its own, not much. Tolls fall 20% but debt service does not, so cover drops from 1.5x to 1.2x every year. Cover of 1.5x means tolls can fall a third, 33.3%, before they fail to meet debt service. The more dangerous stress is on growth, because this schedule assumed it. If traffic starts 20% low and grows 2% instead of 5%, tolls rise from 120 to only 143 while debt service climbs to 155. Cover slides every year and falls below 1.0x in year 8.

Base tolls cover sculpted debt service 1.5 times every year, but in the stress case tolls start at Rs 120 crore and grow only 2% a year while debt service keeps rising 5% a year, so tolls fall short of debt service from year 8 and cover ends at 0.92x.

Year	Base tolls	Debt service	Stress tolls	Stress cover
1	150.0	100.0	120.0	1.20x
3	165.4	110.2	124.8	1.13x
5	182.3	121.6	129.9	1.07x
7	201.0	134.0	135.1	1.01x
8	211.1	140.7	137.8	0.98x
10	232.7	155.1	143.4	0.92x

Rs crore. Debt service is two thirds of base tolls every year. Under the stress, cover falls from 1.20x in year one to 1.01x in year seven and 0.92x in year ten, so the escrow runs short in the last three years.

Step 3How much would you actually raise?

Size on the stress, not the base. The weakest year in the stress is year ten, so solve for the year-one debt service that keeps year-ten cover at 1.0x: Rs 92.4 crore instead of Rs 100 crore. That supports an issue of about Rs 738 crore, Rs 60 crore less, and base cover rises to 1.62x. A level schedule, the same Rs 100 crore every year, supports only Rs 656 crore but never assumes growth at all. The choice between the two is a view on how reliable the traffic study is.

Sizing on the base case gives Rs 799 crore but leaves stress cover as low as 0.92x, sizing so the stress never drops below 1.0x gives Rs 738 crore, and a level schedule that assumes no growth gives Rs 656 crore.

Say the limitation. Future-flow deals are sized on collections that depend on traffic, on toll rates set under the concession, and on the concession itself surviving ten years. A reserve account holding six months of debt service, and a trap that keeps surplus cash in escrow when cover drops below an agreed level, are the usual defences, and the rating leans on them as much as on the cover ratio.

Where candidates lose it

The common error is stressing only the level of traffic and stopping at 1.2x, which looks comfortable. The sculpted schedule quietly assumed 5% growth for ten years, and it is the growth assumption that breaks cover in the later years.

The second is treating cover as falling point for point with traffic. Cover scales: a 20% shortfall takes 1.5x to 1.2x, and it takes a one-third shortfall to reach 1.0x.

What the interviewer asks next

How big a debt service reserve would carry the stress case through years eight to ten?
Why might a toll rate linked to inflation make this deal safer than traffic growth alone?
What happens to the bondholders if the concession is terminated early?
Would you rather hold the senior bonds of this deal or a plain bond of Varadha itself?

Company names and figures are illustrative.