Working Capital Financing: The Cost of Stretching Payables

Working capital financing asks who funds the gap between paying suppliers and being paid. Sankalp Industrial Systems Limited, invented, waits 65.70 days to collect, holds inventory 73.00 days and pays at 91.25 days, a cash conversion cycle of 47.45 days. One day of receivables is Rs 3,28,76,712. Refusing a supplier discount of 2 per cent for payment within 10 days costs 21.28 per cent a year.

The shape of this subject is easier to feel at a street corner than in a treasury note. Begin there. A vegetable seller buys a sack of onions on Monday morning for Rs 1,200/-. She sells the last of them on Thursday. Two of her customers are the small canteens on the same road, and they settle at the end of the week. So her money goes out on Monday and comes back on Saturday, and for five days there is a hole where the money used to be. Nobody sends her an interest bill for those five days. The hole is real all the same. On Tuesday she cannot buy the second sack until somebody hands her cash.

Now change one detail. The onion trader she buys from lets her take the sack and pay on Friday. The hole shrinks from five days to one. Nothing about her selling has changed, nothing about her customers has changed, and yet the amount of her own money tied up in the business has fallen by four days of purchases. Somebody has to fund the stretch between paying for goods and being paid for them, and the only question a business ever really answers is which of three people it is going to be: the supplier, the lender, or itself.

Funding that stretch is the whole subject, and it scales without changing shape. A manufacturer with a Rs 12,00,00,00,000 revenue line is running the same sack of onions through a bigger yard. One idea holds up everything that follows: trade creditBuying now and paying later, which is borrowing from the supplier. is borrowing, and the fact that no interest line ever appears for it does not make it free. A supplier who offers something off the bill for early payment has already priced that credit. A customer who takes 65.70 days to pay is being funded by the seller at the seller's own cost of money. Once every line of the cycle is read as a loan with a rate attached to it, the rest of this guide is arithmetic.

Who is actually funding the gap between paying and being paid?

Sankalp Industrial Systems Limited, invented, makes industrial valves, precision castings and the aftermarket parts and service that go with them. In Year 0 it sold Rs 12,00,00,00,000 of goods and the direct cost of goods soldThe direct cost of what was sold in the period. against that was Rs 7,20,00,00,000. Cost is 60.0 per cent of revenue. Revenue and cost are the flows. Three balances sit against them. ReceivablesMoney customers owe the business for goods already delivered. of Rs 2,16,00,00,000, being money already earned and not yet in the bank. Inventory of Rs 1,44,00,00,000, being castings and valve bodies sitting in a yard. PayablesMoney the business owes suppliers for goods already received. of Rs 1,80,00,00,000, being bills received and not yet settled.

Adding the first two and taking away the third gives net working capital of Rs 1,80,00,00,000, exactly 15.0 per cent of revenue. The 15.0 per cent is the ratio the whole five year forecast holds, and it is the reason this matters to anybody building one: every extra rupee of revenue drags a fixed slice of capital behind it. But the total on its own hides the split that decides everything. The Rs 1,80,00,00,000 is not one number, it is three numbers pulling in two directions, and only the third of them is somebody else's money.

Following one purchase all the way through shows who is holding the bag at each moment. A casting arrives on day 0. The casting sits in the yard, gets machined, gets assembled and finally leaves as a finished valve on day 73.00. The customer takes it, and pays for it on day 138.70. So between the moment the company committed to that casting and the moment cash came back for it, 138.70 days went by. Waiting 138.70 days for the money back is a long time to be out of pocket. Except the company did not actually pay for the casting on day 0. Its suppliers wait 91.25 days on average, so payment went out on day 91.25. So the true hole runs from day 91.25 to day 138.70, and it is 47.45 days wide.

Look at what the picture says about ownership of the problem. Of the 138.70 days that money is committed, the supplier carries 91.25 of them without charging a printed rupee for it. Ninety-one days of unbilled funding is not a small favour. Supplier credit is the single largest source of short term funding this company has, larger than the Rs 1,00,00,00,000 drawn on its working capital line at Year 0. The business itself funds only the last 47.45 days. Trade credit is doing two thirds of the work of financing this operating cycle, and it appears nowhere a reader would look for financing.

How long is the gap, and what is each piece measured against?

Three balances, three measures, and one detail that decides whether any of them means anything. Each measure asks the same question in the same shape: how many days of flow is this balance worth? Take the balance, divide it by one day of the relevant flow, and the answer comes back in days. The trap is in the word relevant.

Days sales outstandingReceivables divided by daily revenue: how long collection takes. runs on revenue. One day of revenue is Rs 12,00,00,00,000 over 365, or Rs 3,28,76,712. Receivables of Rs 2,16,00,00,000 divided by that daily figure gives 65.70 days. Revenue is the right denominator for a simple reason: a receivable is what a customer was billed, and a customer is billed at selling prices. The balance and the flow are on the same basis.

Days inventory outstandingInventory divided by daily cost of goods sold. runs on cost of goods sold. One day of cost is Rs 7,20,00,00,000 over 365, or Rs 1,97,26,027. Inventory of Rs 1,44,00,00,000 divided by that daily figure gives 73.00 days. Inventory is not carried at what the company hopes to sell it for. Inventory is carried at what it cost to buy and make. Dividing a cost balance by a revenue flow compares two things measured on different rulers.

Days payable outstandingPayables divided by daily cost of goods sold. runs on cost of goods sold too, and for the same reason. A payable is what a supplier invoiced, and a supplier invoices the company for inputs at cost, not at the price the finished valve eventually fetches. Rs 1,80,00,00,000 over Rs 1,97,26,027 a day gives 91.25 days. The denominator is not a house convention that a firm may pick; it is fixed by the requirement that the balance and the flow be measured on the same basis.

What happens if all three days measures are run on revenue?

The result is a set of numbers that adds up, sits neatly in a table, and describes a business other than the one in the accounts. Revenue is the figure everybody has to hand, and dividing everything by it feels tidy. The wrong denominator is the quieter of the two failures here, and the more common one.

Run it and see. Inventory of Rs 1,44,00,00,000 over Rs 3,28,76,712 of revenue a day gives 43.80 days rather than 73.00. Payables of Rs 1,80,00,00,000 over the same denominator gives 54.75 days rather than 91.25. Collection days were already on revenue, so they do not move. Put the three together and the cycle reads 65.70 plus 43.80 less 54.75, and that comes to 54.75 days.

Nothing about that output looks broken. The output is internally consistent, it foots, the three inputs are all real balances, and 54.75 days is a perfectly plausible number for a manufacturer. A reviewer scanning the table has nothing to catch. A wrong denominator applied consistently produces a wrong answer that is internally consistent. Consistency is the only thing most reviews test for, and the error never announces itself.

The cost of the wrong denominator is every conclusion built on top. On the wrong basis this company appears to hold six weeks less stock than it does, and appears to pay suppliers in under two months when it actually takes three. Someone benchmarking it against a peer set computed correctly would conclude the company holds stock tightly and pays promptly, and both readings are backwards.

Is the cycle 47.45 days or 47.4 days?

Both, and the honest thing is to say which one was used. The cash conversion cycleCollection days plus inventory days less payable days. is collection days plus inventory days less payable days. On the unrounded figures that is 65.70 plus 73.00 less 91.25, giving 47.45 days. But payable days are very often printed to one decimal place, and 91.25 printed to one decimal is 91.3. Adding the three printed figures gives 65.70 plus 73.00 less 91.3, and that comes to 47.4.

Five hundredths of a day is not a material amount of money. On this company one day of cost is Rs 1,97,26,027, so the difference between the two figures is under Rs 10,00,000 on a balance sheet carrying Rs 1,80,00,00,000 of net working capital. The reason it matters has nothing to do with size. Exactly one wrong move is available here, and it is not choosing either figure. The wrong move is to shave the payable days to 91.20 so the printed column adds to the printed total. A reader can check either basis, but nobody can check a number quietly adjusted to agree with the line above it.

The cycle here is computed on the unrounded days and reported as 47.45. Anybody adding the three printed figures will land on 47.4 and now knows why. The discipline is simple: a rounded column presented as parts of a whole either states its basis or misleads, and the same discipline holds wherever a set of rounded parts is drawn as though it summed exactly to a rounded total.

What is one day actually worth in rupees?

Here is where most conversations about working capital go wrong, and it is a mistake of language rather than of arithmetic. Somebody says the business needs to take ten days out of the cycle. Everybody nods. Nobody asks ten days out of which line, and the answer differs by two thirds.

Receivables are measured against revenue, so one day of receivables is one day of revenue. Rs 12,00,00,00,000 over 365 is Rs 3,28,76,712. Inventory and payables are both measured against cost of goods sold, so one day of either is one day of cost of goods sold. Rs 7,20,00,00,000 over 365 is Rs 1,97,26,027. The two figures differ by Rs 1,31,50,685 a day.

The ratio between them is 1.667, and it is not a coincidence or a quirk of this particular company. The ratio is one divided by 0.60, and 0.60 is the cost of goods sold ratio. Any business where cost is 60.0 per cent of revenue will find that a day of receivables is worth exactly one over 0.60 times a day of stock or of supplier credit. A business at 80.0 per cent cost would find the ratio is 1.25. The gross margin sets the exchange rate between a day on one line and a day on another. The same ten day improvement is worth Rs 32,87,67,123 or Rs 19,72,60,274, depending entirely on where it came from.

One further caution before the control below. A release of this kind happens once. Cutting collection days from 65.70 to 55.70 hands the business Rs 32,87,67,123 in the year it happens and then nothing more. The shorter cycle is the new normal, and next year starts from there. Interest saved on money no longer borrowed is what continues afterwards. At the 7.60 per cent this company pays on its short term line, Rs 32,87,67,123 not borrowed is Rs 2,49,86,301 a year of interest not paid. The release is a one-off; the interest is the annuity.

What does the money the company is owed cost it?

Receivables: the cost of being owed

Start with the household version again. Being owed money is easy to feel and hard to price. A tailor stitches forty school uniforms for a shop and hands them over in June. The shop settles in August. For those two months the tailor has already bought the cloth, already paid the two people who helped, and has nothing back. If she borrowed to buy the cloth, she is paying interest on it while the uniforms sit on somebody else's shelves. If she did not borrow, she is short of the money she would otherwise have used for the next order. Either way the credit she extended has a price, and she is paying it.

Sankalp Industrial Systems Limited, invented, is doing the same thing at scale. Rs 2,16,00,00,000 is out with customers at any moment. The Rs 2,16,00,00,000 has to come from somewhere, and the company runs a working capital facilityA revolving loan secured against receivables and inventory. secured against exactly these receivables and its inventory, on which it pays 7.60 per cent a year. Funded there, being owed Rs 2,16,00,00,000 costs Rs 16,41,60,000 a year.

Break it down per day and the number stops being abstract. One day of the 65.70 costs Rs 3,28,76,712 of balance, and carrying that at 7.60 per cent for a year is Rs 24,98,630. So every single day of average collection time this company grants its customers costs it roughly Rs 25,00,000 a year in interest. Ten days of it is Rs 2,49,86,301. Credit terms are almost always decided by a sales conversation and almost never priced by a finance one. A company ends up giving away a discount it never wrote down.

Notice where that cost surfaces in the accounts. The cost does not appear as a reduction in revenue, or as a line called cost of customer credit. The cost appears in the interest expense, mixed in with the borrowing that funded a factory, and nothing in the presentation separates the two. Real costs arriving under the label of something else is the general shape of this whole subject.

What is a company buying when it stretches a supplier?

Time, and it is paying for it. Almost nobody believes that on first reading. The payment is invisible.

The term worked here belongs to one supplier of this invented company: 2 per cent off the bill for payment within 10 days, otherwise the whole amount at 45 days. Written on an invoice it looks like an administrative note. Read as finance it is a loan offer with everything in it except the word interest.

Consider Rs 100.00 of invoice face value. Payment on day 10 hands over Rs 98.00. Payment on day 45 hands over Rs 100.00. So the choice is between Rs 98.00 now and Rs 100.00 in 35 days. The buyer is being offered 35 days of credit on Rs 98.00, and the price of it is Rs 2.00. Stretching a payablePaying later than the term allows or than the discount invites. is exactly this transaction, repeated across every invoice the practice touches.

One honesty note before the arithmetic. A careful reader will already have spotted a tension. Sankalp pays at 91.25 days on average, twice the 45 day net term on this one supplier. The average and the term are not in conflict, and they are not the same measurement. The 91.25 days is a whole-company average across every supplier, on every term, including whatever arrangements sit behind the longest of them. The 2 per cent within 10 days, whole amount at 45 days is one supplier arrangement, and the arithmetic below prices that arrangement rather than claiming the entire Rs 1,80,00,00,000 balance sits on it. A days measure is an average of many different arrangements and can never be read back as the term on any particular invoice.

How does a discount term turn into a rate?

In two steps, and the second one is where the surprise lives. A discount termAn offer of a percentage off for paying early instead of at full term. gives a charge and a length of time, and a rate is a charge divided by an amount, scaled to a year.

Step one, the charge as a percentage of what is actually kept. The buyer gives up Rs 2.00 in order to hold on to Rs 98.00, so the charge is 2 over 98, or 2.0408 per cent. The denominator matters. The charge is not 2 over 100. Payment on day 10 would have parted with Rs 98.00 and no more, so the Rs 100.00 is not what is being borrowed. The Rs 98.00 is the loan.

Step two, scale it to a year. The charge bought 35 extra days, from day 10 to day 45. There are 365 over 35 such stretches in a year, being 10.4286 of them. On the simple annualisationScaling a period cost to a year by multiplying, without compounding. that is 2.0408 per cent times 10.4286, giving 21.28 per cent a year. The 21.28 per cent is the figure used throughout. Compounding the 2.0408 per cent over the 10.4286 periods instead gives 23.45 per cent. Both are locked and both are standard; the simple version is the more common convention and the compounded one is the more complete description of what happens if the practice repeats all year. The two figures differ by 2.17 percentage points and neither changes the conclusion. The honest thing is to name the one used rather than to argue about the choice.

Now look at the shape rather than the single number. The shape is the part that generalises. The charge is fixed at 2 per cent whatever the terms say, but the time it buys is not. Buy 35 days with it and it annualises to 21.28 per cent. A term of 2 per cent within 10 days against the whole amount at 30 days buys only 20 days, and the same 2 per cent annualises to 37.24 per cent. Buy 60 days with it and it falls to 12.41 per cent. The rate is the charge spread over the time bought, so the shorter the stretch the term grants, the more expensive the same discount becomes.

Where does that rate sit against everything else the company borrows?

At the top, and it is not close. Sankalp has three numbered borrowings at Year 0, and every rate is its own contracted rate. A secured rupee term loan of Rs 3,00,00,00,000 at 7.80 per cent. Listed unsecured debentures of Rs 2,00,00,00,000 at 8.50 per cent. A working capital facility of Rs 1,00,00,00,000 drawn at 7.60 per cent, secured against the very receivables and inventory measured above. Blended across the three, the pre-tax cost of debt is 8.00 per cent.

Against that ladder, 21.28 per cent a year is 2.80 times the cheapest money the company has and 2.66 times its blended cost of debt. On the compounded basis, 23.45 per cent is 3.09 times the facility rate. Two notes on that last figure, both of them the same discipline as the 47.45 against 47.4 above. The multiple is computed once from the full values, 23.4524 over 7.60. Dividing the printed 23.45 by 7.60 gives 3.0855, and 3.0855 rounds to 3.09 as well, but only by luck. Doing arithmetic on printed figures is the habit that produces a wrong last digit somewhere else. The rounding happens once at the end, so the multiple is 3.09 and not 3.10.

The comparison in rupees is more persuasive than the comparison in rates, so it is worth working on Rs 100.00 of invoice face value. Taking the discount means paying Rs 98.00 on day 10. If that Rs 98.00 has to be borrowed on the facility, carrying it from day 10 to day 45 costs Rs 98.00 at 7.60 per cent for 35 days, or Rs 0.71. The total outlay is Rs 98.71. Refusing the discount means paying Rs 100.00 on day 45. The gap is Rs 1.29 on every Rs 100.00 of invoice, and it goes in the same direction on every invoice, in every month, for as long as the practice continues.

Scale it, with the assumption named. If every rupee of the Rs 7,20,00,00,000 of annual purchases sat on this one term and the discount were taken on all of it, the 2 per cent would be worth Rs 14,40,00,000 a year and the facility interest needed to fund the early payment would be Rs 5,14,21,808, leaving Rs 9,25,78,192. Those figures show the size of the prize rather than what this company actually earns. Nothing says the whole payables balance sits on this one term, and an average of 91.25 days is evidence that it does not.

What does the same term look like from the other side of the invoice?

Identical. The symmetry is the tidiest fact in the whole subject and the one most worth carrying away. There is one term, one charge and one length of time, so there can only be one rate. The buyer gives up Rs 2.00 to keep Rs 98.00 for 35 days and pays 21.28 per cent a year. The seller gives up nothing except 35 days of waiting, keeps the Rs 2.00 it would otherwise have discounted, and earns 21.28 per cent a year on the money it was owed. Same term, same rate, opposite signs.

The symmetry is the check on any discount policy, and it cuts both ways in a way that catches people out. A company may refuse discounts as a buyer, on the grounds that holding cash is prudent, and offer discounts as a seller, on the grounds that collecting early is prudent. Such a company is paying 21.28 per cent in one direction and earning it in the other, on entirely different balances. The same arithmetic that makes one side attractive makes the other side expensive, so a discount policy has to be priced from both ends at once.

Which line should a company attack first?

The arithmetic and the answer point in different directions, and being honest about that is more useful than a clean rule. Per day, receivables release Rs 3,28,76,712 and the other two release Rs 1,97,26,027, so on pure arithmetic collection days rank first by a factor of 1.667. The simulation above shows exactly that, and the arithmetic is not in dispute.

But the arithmetic is not the constraint. Collection days are set by customers, and a manufacturer whose customers are larger than it is does not adjust them by deciding to. Inventory days are set by how long a casting takes to become a valve and by how much finished stock the aftermarket business has to hold to answer a service call, both of which are operating facts rather than treasury choices. Payable days can be moved unilaterally, and being movable is exactly why they are the line that gets moved. The arithmetic above shows what moving them costs.

So the ranking that survives contact with a business is: the line that can actually be moved, priced. The constraint is never the arithmetic and always the counterparty, so a ranking of the three lines by rupees per day alone answers a question nobody inside the company is facing. Everything here prices the move rather than recommending it.

Who reads a cash conversion cycle for a living?