Refinancing: Replacing Existing Debt and Why Timing Matters

Refinancing swaps one borrowing for another. Four things drive it: a cheaper rate, a later repayment date, easier terms, or a lender that wants its money back. Only the first is settled by arithmetic. The saving is the rate gap times the amount times the years left to run, and everything it costs to arrange falls due on day one.

Two facts underneath carry everything that follows, and both are covered separately. A borrowing carries a contracted rate and a date on which it has to come back. Where those two come from is set out under financial statements and lending institutions. And interest is a flow while an arranging cost is a single payment, so the two cannot be compared until the length of the flow is stated. Everything below is those two facts pushed hard.

The worked case is Harivansh Packaging Limited, whose published position gives borrowings of Rs 740 crore, a finance cost of Rs 60 crore, earnings before interest and tax (EBIT) of Rs 339 crore and cash of Rs 140 crore. The borrowings, the finance cost, the earnings and the cash carry the whole of what follows.

What actually changes when a borrowing is refinanced?

A household makes the shape plain. The arrangement is identical and the figures are small enough to hold in mind. A home loan has fourteen years left on it at one bank. Another bank offers a lower rate. The second bank does not hand the borrower a cheque; it pays the first bank directly, the first loan closes, and a new loan of the same size opens in its place. The household owes exactly what it owed the day before. Nothing was bought. Nobody got richer on the day of the switch. The rate changed, and so did the lender, the paperwork and possibly the number of years left.

A company does the same thing with a trancheA single slice of a larger borrowing, with its own amount, its own rate and its own repayment date. One company can run several at once and they need not resemble each other at all. of its borrowings. The new lender's money arrives and leaves on the same day, going straight to the lender being repaid. On the balance sheet the borrowings line does not twitch. On the profit statement, from the following period, the finance cost is different.

Refinancing is not raising money. A refinancing that brings in no new money at all can still be the most consequential thing a company does in a year. The claim runs against the instinct that a transaction has to move an amount before it can matter. A raise brings in cash the company did not have and gives away either shares or a claim in exchange. A refinancing brings in nothing and gives away nothing. A refinancing rewrites the terms on money the company already has, and terms are where a borrower lives.

Consider Harivansh Packaging Limited. Its Rs 740 crore of borrowings is one published number, and it is split here into two tranches so there is something to work on: a Rs 500 crore tranche and a Rs 240 crore tranche. Replacing the Rs 500 crore tranche with a Rs 500 crore facility from somebody else leaves the borrowings line still reading Rs 740 crore afterwards. The rate on two thirds of it moved, and with the rate went the finance cost, the interest cover, the covenants, the security and the date on which that two thirds has to be found again.

Why does a company refinance, and are the four reasons alike?

Four things send a treasurer looking for a replacement facility, and everything that follows turns on the fact that they are not variations of one another.

A lower rate. The replacement carries a smaller coupon than the borrowing it repays. A lower rate is the case everybody pictures, the one the arithmetic below computes, and the only one of the four with an arithmetic answer.

A later date. The replacement runs longer, so a repayment that was going to land next year lands in five years instead. The rate on the replacement might be higher, lower or the same. The thing being bought is time rather than money, so a borrower may well accept a higher rate.

Looser terms. The securityAssets a borrower puts behind a loan. A lender who is not repaid has something specific to look to. Which assets, and in what order, is a lending question rather than a refinancing one. given, the covenantsPromises written into a loan document, usually about keeping some ratio inside an agreed range. A breach of one is covered separately. accepted, or the restrictions on what the business may do with its own assets. A borrower that has grown into a facility written when it was smaller may want the restrictions rewritten more than it wants a cheaper coupon.

A lender that wants out. The existing lender declines to roll, or asks to be repaid, or sells the loan to somebody with a different appetite. Nothing about this reason is chosen by the borrower. The demand arrives, and the borrower has to find a replacement whether the replacement is cheaper or not.

Only the first of the four is judged by arithmetic, and treating all four as rate decisions is where the analysis goes wrong. A payback calculation run on a maturity refinancing produces a confident, precise, entirely irrelevant answer. A payback calculation run on a covenant refinancing cannot see the covenant. And a payback calculation run on a lender that has asked for its money back is answering a question that was never open. The payback arithmetic is a tool with one job, and the first skill is recognising which of the four situations applies before reaching for it.

How is the saving actually built?

Three things multiply together. A number produced with any of them missing is a number about something else.

The first is the rate gap. Not the old rate, not the new rate, but the difference between them. Harivansh Packaging Limited's Rs 500 crore tranche runs at a contracted 9.0 per cent and the replacement is offered at a contracted 7.5 per cent. The gap is 1.5 percentage points, and a lending desk would call that 150 basis pointsOne hundredth of a percentage point. Lending desks count in these because a hundredth of a point on a large amount is still real money..

The second is the amount actually being refinanced. The Rs 240 crore tranche is untouched and keeps paying what it was paying, so the amount here is the Rs 500 crore tranche and not the Rs 740 crore of total borrowings. The first two multiplied together give a rate of saving: 1.5 per cent of Rs 500 crore is Rs 7.5 crore a year, before tax.

The third is how long the new terms will actually run. The time term is the one that gets dropped. Beside two numbers sitting in the term sheet in bold, the years left feel like an afterthought. A rate of saving with no duration attached cannot be compared with anything. Leaving out the time term does not make the answer approximate, it makes the answer meaningless. Rs 7.5 crore a year is a true statement that answers a different question from the one being asked.

Two footnotes on that table, both of which matter later. The Rs 7.5 crore a year is stated before tax. After tax at the 25.0 per cent effective tax rateWhat the reported tax charge works out to as a proportion of profit before tax, rather than any rate written in a statute. On this company the effective rate lands on 25.0 per cent. the company keeps Rs 5.625 crore of it, which across 18.00 crore shares lands on Rs 0.31/- a share, or 2.5 per cent of the Rs 12.50/- the company earns. The after tax number is worth knowing, and it is not the figure the payback is built on. How the costs of the swap are treated in the same computation is settled by the tax rules. And the four years is an assumption stated on its face, not a fact from the record. The record carries no maturity date for either tranche.

What does the swap cost, and when does that bill fall due?

Two kinds of cost, and both of them land at the beginning.

The first sits on the borrowing being repaid. A lender that has priced a loan expecting to hold it for years does not usually agree to be repaid early for nothing, so the document carries a prepayment charge. On Harivansh Packaging Limited's Rs 500 crore tranche the charge is 1.0 per cent of the amount repaid, or Rs 5 crore. The percentage is a term of this facility and nothing else, and another document carries whatever that document says.

The second sits on the borrowing being taken. Arrangement, legal and documentation costs on the replacement come to Rs 2.5 crore here. Together the swap costs Rs 7.5 crore to execute.

Both costs are paid once and at the start. The saving arrives in slices over years, and that mismatch in timing is the entire shape of the decision. If the cost arrived in the same slices as the saving there would be nothing to think about: Rs 7.5 crore a year of saving against some annual charge, and the answer visible in one line. The cost does not behave like that. The cost is a step. The step lands in full before a single rupee of saving has been collected.

The household version is exact. Switching a home loan carries a processing fee at the new bank and possibly a foreclosure charge at the old one, both payable at the switch, against a lower instalment that shows up month by month afterwards. A household with two years left on the loan and a household with eighteen years left face the same two charges and are in completely different positions. Nobody would tell them the same thing.

How long before the saving overtakes the cost?

Divide. The cost is Rs 7.5 crore and the saving is Rs 7.5 crore a year, so the payback is 1.0 year. From the day the swap is done, Harivansh Packaging Limited spends twelve months getting back what it spent, and only after that does the refinancing start putting money in.

Now the central test. A payback longer than the time left on the borrowing being replaced means the refinancing loses money with certainty, and that is knowable before anything is signed. Not likely to lose. Not risky. Certain, in the same way that spending Rs 100/- to collect Rs 80/- is certain, and knowable on the day the term sheet arrives rather than discovered two years later.

Say it as a comparison of two durations and it becomes hard to get wrong. On one side, how long the saving takes to repay the cost. On the other, how long the saving will actually run. If the first is shorter, the swap is worth doing on rate grounds. If the first is longer, it is not, whatever the headline says about basis points.

What does all of this look like on Harivansh Packaging's own numbers?

The record publishes borrowings of Rs 740 crore and a finance cost of Rs 60 crore. Divide the second by the first and the blended cost of that borrowing recomputes to 8.11 per cent. The blended cost is the only rate the record actually yields, and it is an average rather than a rate anybody agreed to.

To have something to refinance, the Rs 740 crore is split into two tranches. Both amounts and both rates are constructed, and the constraint they satisfy is that they reconcile exactly to the published finance cost.

Now the swap. The Rs 500 crore tranche is replaced by a facility at a contracted 7.5 per cent, so its interest falls from Rs 45 crore to Rs 37.5 crore and the company saves Rs 7.5 crore a year before tax. The cost of getting there is a prepayment charge of 1.0 per cent on the amount repaid, Rs 5 crore, plus Rs 2.5 crore of arrangement, legal and documentation costs, so Rs 7.5 crore in all, paid at the start. Rs 7.5 crore of cost over Rs 7.5 crore a year of saving is a payback of 1.0 year.

Now the finding that everything above has been building towards. The identical swap is run at three different tenors, changing nothing except how long the tranche had left.

The rate cut is the same 150 basis points in all three rows, the amount is the same Rs 500 crore, the cost is the same Rs 7.5 crore, and the answer changes sign purely on the time left. A treasury team reporting the first row and a treasury team reporting the third row would write the same sentence about the rate. One of them made the company Rs 22.5 crore and the other cost it Rs 3.75 crore.

What is a refinancing done for maturity actually asking?

Everything above assumed the borrower had a choice. The tranche was going to sit there at 9.0 per cent, somebody offered 7.5 per cent, and the question was whether taking it beat leaving it alone. Leaving it alone was a real option.

A maturity refinancing removes that option. The borrowing falls due. The amount has to be repaid, from cash the company has, from cash it can generate, or from money somebody else lends it. There is no version of the situation where the tranche simply stays where it is, and once the alternative to refinancing disappears, so does the arithmetic that compared the two.

Harivansh Packaging Limited's own figures set against the Rs 500 crore tranche show why. Cash on the balance sheet is Rs 140 crore, or 28.0 per cent of the tranche, and the remaining Rs 360 crore has to come from somewhere else. The company's EBITDAEarnings measured before interest, tax, depreciation and amortisation are taken off. The figure appears here only as a size marker for what a year of trading throws up., earnings before interest, tax, depreciation and amortisation, is Rs 477 crore. The figure sounds like enough until what that Rs 477 crore is already committed to is set against it: Rs 60 crore of interest, Rs 75 crore of tax, and whatever capital expenditureMoney spent on the plant, machinery and buildings a business will use for years, as opposed to money spent running this year's operations. the business needs to keep its plants running. The record publishes no cash flow statement and no capital expenditure figure for this company, so the exact residue cannot be computed. One year of a Rs 477 crore EBITDA plainly does not free up Rs 360 crore after all of that.

The question on the table is whether the company continues, not whether it saves. Judging a maturity refinancing by the rate saved is judging the wrong thing. A treasurer who replaces a maturing Rs 500 crore tranche at a rate 200 basis points higher than the old one has not done bad work. Somebody found Rs 500 crore. Finding it was the job.

And a company refinancing under pressure usually does pay more. Paying more is a fact about the situation rather than a failing of the people in it. A borrower with a date closing in and no alternative source has less to negotiate with than a borrower with three years of runway, and lenders price what they see. Reading a higher replacement rate as evidence of a weak treasury team gets the causation backwards; the rate is a reading of the position, and the position is what the refinancing exists to fix.

Which is why the question is opened years before the date. Not because anybody enjoys early paperwork, but because the only lever the borrower has over the price is time, and time is the one thing that cannot be added later.

Where does the saving show up in the reported numbers?

Take the rate refinancing again and follow it into the statements. The Rs 500 crore tranche now costs Rs 37.5 crore instead of Rs 45 crore, the Rs 240 crore tranche still costs Rs 15 crore, so the finance cost line drops from Rs 60 crore to Rs 52.5 crore.

The one movement shows up in three reported places at once. The finance cost itself is down Rs 7.5 crore. The blended cost of borrowing, being the finance cost over the Rs 740 crore borrowed, falls from 8.11 per cent to 7.09 per cent, a move of 1.02 points. And interest cover, being EBIT of Rs 339 crore over the finance cost, rises from 5.65 times to 6.46 times, a move of 0.81 times.

None of those three is a saving in its own right: they are the same Rs 7.5 crore appearing in three places, and counting them separately triples a number that only happened once. The reason is mechanical rather than subtle. All three readings are built on the finance cost line, so anything that moves that line moves all three, and the three movements are not independent facts about the company. The three readings are one fact, reported through three different denominators.

A sheet of before and after ratios is exactly how a refinancing gets written up. The triple count matters more than it sounds. Three improved lines look like three achievements. A reader who adds them, or who treats the cover improvement as something extra sitting on top of the finance cost saving, ends up describing a transaction three times better than the one that happened.

What about the terms that are not the rate?

A facility agreement is a long document and the coupon occupies one line of it. Everything else in there is also a term, also negotiated, and also capable of moving in either direction when the paper is rewritten.

The security given can widen, so assets that were free of any charge now sit behind the loan. The covenants can be set with less headroom, so a ratio that had room to move before is now close to a level the borrower must not cross. The amortisation scheduleThe timetable saying how much of the principal comes back to the lender in each period, rather than the whole amount arriving on the last day. can front load, so principal starts coming back sooner and the borrower has less to work with each year. The currency can change, and a change of currency introduces an exposure that had nothing to do with the rate. And the right to prepay again can be locked for a period, so the flexibility the company just used is the flexibility it has given up.

A refinancing can cut the rate and tighten the covenants in the same document, and where it does, the payback arithmetic above has answered only part of the question. The calculation compares rupees against rupees. A covenant is not a rupee amount. A covenant never enters the sum and never shows up as a cost. Two years later, when the business needs room the document no longer allows, a covenant can be the term that matters most.

None of that means the arithmetic is useless. The arithmetic settles the part of the question that has a number attached, and reading the document settles the rest. A borrower who does the first and skips the second has a precise answer to a subset.