How to Build an Issuer Credit Snapshot, Step by Step

An issuer credit snapshot is a one sheet record of what a borrower's price implies, built in a fixed order. The instrument and its terms are named, then the reference rate at the matching maturity; the two are subtracted and both units recorded; the spread measure is named; the recovery assumption is written down; the division is run and the arithmetic checked backwards; and a reason is written beside every row that could not be filled. The empty rows are part of the output.

A procedure earns its place when two people doing the same work produce results that can be laid side by side, and when a reader six months later can still tell which figures were observed and which were supplied by whoever held the pen. Every step below is a subtraction, a division, a check or a label, and nothing else. The steps that look like housekeeping are the ones that make the arithmetic auditable, and they are the steps most people drop first.

The move a snapshotA fixed-order record of what a borrower's price implies, built so that two people working separately produce results that can be compared. makes is the move credit measurement makes everywhere. Take the price apart. A borrower pays a rate for a period; a reference borrower pays a lower rate for the same period; the difference is a quantity that somebody is being paid to bear something. Put arithmetic on that difference. Then ask the harder question. What else could be sitting inside the same difference wearing no label at all?

Most households already run a rough version of this. A cooperative bank down the road offers more on a five year deposit than the post office does for the same five years, and nobody has to point out that the extra is not a gift. Almost nobody writes down, in the same place every time, what was assumed when the extra was judged worth taking. The snapshot is that habit turned into a form with numbered rows.

What goes into an issuer credit snapshot, and in what order?

Eight steps, run in order, with no step allowed to start before the one above it has produced its label. Every step consumes something the step before it wrote down, so the order is the teaching. Step three cannot subtract until step two has named what is being subtracted. Step six cannot divide until step five has written the divisor down and stamped it as supplied rather than observed. Step seven cannot list what is missing until steps one to six have shown what was reachable.

1. Name the instrument and its termsIssuer, maturity, coupon rate, face amount, payment frequency and compounding convention, all written out before a single figure is computed. Palash Cements Limited, invented, five years, 9.10 per cent annual coupon on Rs 1,000.00/- of face, one payment a year, annual compounding.
   Checking: could a stranger reprice this instrument from this row alone?
2. Name the reference rate and its maturityTake it from the government SPOT curve at the same maturity as the instrument. Five years here, so the five year government SPOT rate of 6.90 per cent a year, written with the word SPOT in it and with the maturity beside it.
   Checking: does the curve actually carry a point at this maturity, or is the match being stretched?
3. Subtract, and record both units9.10 less 6.90 is 2.20 percentage points, which is 220 basis points. Write both. A snapshot carrying one unit forces the next reader to convert, and conversion is where this material loses people.
   Checking: are both units on the row, and does the row name the rate the spread is measured over?
4. Record which spread measure this isA G-spread here, being one yield less one government SPOT rate at the matching maturity. The row exists so a snapshot is never compared against a spread built a different way.
   Checking: is the measure named, and would the next reader know how it was constructed?
5. Write the recovery assumption downIts own row, with the word ASSUMED in it and with its base named. Assumed recovery 40 per cent of the amount owed, so the loss given default is 60 per cent of the same amount owed, written 0.60.
   Checking: does the row say ASSUMED, and does it say a percentage of what?
6. Divide, then multiply back before writing anything down2.20 divided by 0.60 is 3.6667 per cent a year. Multiplied back by 0.60 it returns 2.2000 percentage points. Only once the loop closes does the figure enter the document, and it enters as an implied annual default rate.
   Checking: did the multiplication come back to the spread it started from, to four decimals?
7. Fill the empty rows with the reason they are emptyRating, measured default frequency, measured recovery, the liquidity component of the spread, a second issuer for comparison, and spread history. Six rows, six reasons, written in.
   Checking: is every blank carrying a reason rather than a space?
8. State what the finished snapshot supportsOne line naming what this document can be used for and what it cannot. It travels with the rows, because the rows travel further than the person who built them.
   Checking: does the line refuse the things this document cannot carry?

Which reference rate does the snapshot start from, and how is it chosen?

Step two takes the reference rateThe government SPOT rate at the same maturity as the instrument being measured, used as the base the borrower's yield is subtracted from. from the government SPOT curve at the same maturity as the instrument. Palash Cements Limited has issued at five years, so the reference is the five year government SPOT rate of 6.90 per cent a year, on the invented curve this material works from. A mismatched reference sweeps part of the government curve's own slope into what is then called the borrower's spread, so the maturity is matched, or the failure to match is recorded.

The curve carries a five year point and the bond runs five years, so the pick is easy. The pick is not always easy. If the curve has nothing at the instrument's maturity, the honest response is to write that absence into the reference row rather than to reach for the nearest available rate and say nothing about it. A three year government SPOT rate of 6.55 per cent a year set against a five year bond does not produce a smaller spread; it produces a spread with an unlabelled piece of curve slope inside it.

The labelling rule this material enforces hardest

Every rate written anywhere on a snapshot carries the word SPOT or the word FORWARD. A SPOT rate is the rate for money placed today and returned on one stated future date. The five year government SPOT rate of 6.90 per cent a year is one, and a SPOT rate is the only thing step three subtracts from. A FORWARD rate is the rate for money placed on one future date and returned on a later one. A FORWARD rate is not a separate opinion about the future; it is already sitting inside the SPOT curve and can be pulled out of it with arithmetic.

The reason the labels are compulsory rather than tidy is that the two objects read almost identically. The invented curve behind this material carries a one year SPOT rate of 5.90 per cent a year and a two year SPOT rate of 6.25 per cent a year. The one year FORWARD rate for the year beginning one year from now, pulled out of those two, is 6.6012 per cent a year. The three year SPOT rate on the same curve is 6.55 per cent a year. The FORWARD rate and the three year SPOT rate sit 5.1157 basis points apart and are completely different objects.

So the snapshot never carries a bare number for a rate. The snapshot carries a rate with its label and its maturity, every time. A reader who meets 6.6012 and 6.55 in the same document with no labels will merge them into one idea within about a week.

How is the spread recorded so the next reader can check it?

Step three subtracts and writes both units. 9.10 per cent a year on the Palash Cements Limited bond, less the five year government SPOT rate of 6.90 per cent a year, is 2.20 percentage points, or 220 basis points. Both go on the row. One basis point is one hundredth of a percentage point, so the two figures are the same quantity written two ways, and a snapshot that carries only one of them makes every later reader do the conversion in their head at the exact moment they are trying to think about something else.

A spread is always over something and always for a stated period, so no spread is ever written as a bare number. The row reads 220 basis points over the five year government SPOT rate, for five years. Strip either half of that away and the figure becomes uncheckable. Nothing is left to subtract it back from.

The reason that sounds pedantic and is not: a spread with no maturity beside it cannot be undone. Somebody who reads 220 basis points on its own has three unknowns and one figure. That reader does not know which government SPOT rate it was measured against, does not know at which maturity, and does not know whether the borrower's yield behind it was 9.10 per cent a year or something else entirely against a different reference. Every one of those is recoverable from a row that says 220 basis points over the five year government SPOT rate, for five years, and none of them is recoverable from a row that says 220. The extra eleven words are the difference between a figure and a measurement.

Step four, the row that stops a comparison going wrong

Step four names the measure. Step three built a G-spreadA spread measured as one yield less one government SPOT rate at the matching maturity, which is the simplest of the three common constructions., one yield less one government SPOT rate at the matching maturity. The measure row exists for one reason: so this snapshot is never laid beside a spread that was built a different way. A Z-spread is a different construction rather than a different opinion, and on this same Palash Cements Limited bond, on the same day, at the same price, it reads 227.1722 basis points instead of 220.00, a difference of 7.1722 basis points. The 227.1722 figure is settled under the spread measures and appears here to show why step four is a row rather than a footnote.

Two people who compare a G-spread against a Z-spread and call the gap a finding about the borrower have read a measurement convention as though it were news. The measure row makes that mistake take effort.

Where does the recovery assumption go, and how is it labelled?

Step five gets its own row and the row carries the word assumedSupplied by whoever is doing the arithmetic rather than observed in a market or measured in a study. inside it, together with the base the percentage is a percentage of. The row reads: assumed recovery 40 per cent of the amount owed. Then the loss given defaultOne hundred per cent less the recovery rate, measured on the same base, which here is the amount owed. falls straight out of it. Subtracting 40 from 100 leaves 60 per cent of that same amount owed, written 0.60.

An assumption written in a fixed place gets questioned and an assumption mentioned in passing does not, so the step that produces nothing new is the most important step in the procedure. Every misreading of an implied default rate anywhere on this platform starts the same way, with a recovery figure that was perfectly true when somebody said it out loud and carried no label at all by the time it reached the third reader.

A recovery figure that was not assumed would have to come from somewhere, and it is worth being blunt about where. Such a figure comes from a study of what holders of similar claims actually received after similar borrowers stopped paying, over a period long enough to include a bad stretch, with the claims sorted by where they sat in the queue. A recovery study of that kind is a real body of work, and it is done. The 40 per cent here is not a weak version of that study. The figure is a placeholder occupying the position where such a study would sit, and the row says so in the word ASSUMED rather than leaving the reader to infer it from silence.

Four quantities, four different bases and periods

The row is fussy about its base because the quantities around it all look like percentages and none of them share a base. A recovery rate is a percentage of the amount owed, not of the price paid and not of the coupon. A loss given default is a percentage of that same amount owed. A default rate is a rate per year, not a chance over the whole life of the instrument and not a count of anything. An exposure is a rupee amount at a stated moment. An expected credit loss is a rupee amount per year on a stated exposure.

Five quantities, and the multiplication that joins them only means something once each one has been labelled first. Those unshared bases are why every row on the snapshot names its base or its period inside the row rather than in a note underneath, and why the procedure spends a whole numbered step on a subtraction a schoolchild could do.

How is a default rate implied, and how is the arithmetic proved to close?

Step six divides, then multiplies back before anything is written into the document. 2.20 percentage points divided by a loss given default of 0.60 gives 3.6667 per cent a year. The 3.6667 per cent is the implied annual default rateThe annual rate of default the price would need for the spread to be exactly compensating, given the assumed recovery. The rate is solved backwards, not counted., and it goes into the row under that name and never as a default probability or a default risk.

Now the back-checkRunning the arithmetic in reverse to confirm it closes on the same figure it started from, before the result is written down.. 3.6667 multiplied by 0.60 is 2.2000 percentage points to four decimals, the spread from step three. The loop has closed, so the figure may be written down. The rounded 3.67 multiplied by 0.60 gives 2.2020 percentage points and does not come back to 2.20, so carry four decimals inside the multiplication.

What goes in the rows that cannot be filled?

Step seven is the one that separates a snapshot from a summary. Each empty rowA field that cannot be filled from what is available, recorded with the reason rather than left blank or deleted. gets the reason it is empty written into it, and it keeps its number like every other row. Six of them come out of this particular build.

The rating row is empty because Palash Cements Limited has no rating anywhere on this platform, and because a scale and the meaning of each of its steps are set by the rating agencies and by the Securities and Exchange Board of India (SEBI) at sebi.gov.in rather than by whoever is holding the form. The measured default frequency row is empty because no default study was read for this work. The measured recovery row is empty, and that emptiness is exactly why step five says assumed. The liquidity component row cannot be separated out at all, so the whole 220 basis points has been treated as payment for credit. The second issuer row is empty because no second issuer exists here to compare against. The spread history row is empty because no spread series exists here either.

The reason step seven earns a number of its own is what happens to the document afterwards. A snapshot gets forwarded. A snapshot gets pasted into a note, quoted in a meeting, and read a year later by somebody who was not in the room. At every one of those handovers, a blank row is silently reinterpreted by whoever is reading it: as an oversight, as a figure somebody forgot to type, or worst of all as a zero. Nothing is left to interpret in a row that reads "none, no default study was read for this work", so it survives all three handovers unchanged. The blank cannot travel. The reason can, and travelling is the work it does.

What does the finished snapshot look like written out?

Here is the whole thing for Palash Cements Limited, invented, run through all eight steps. Every figure in it comes out of the two rates named at the top, and the six rows at the bottom are as much a part of the output as the six at the top.

The price row, which makes the size of a spread real

A rate is easy to nod at and hard to feel. So the snapshot carries one more row that turns it into money. Palash Cements Limited's five payments are Rs 91.00/- at the end of each of the first four years and Rs 1,091.00/- at the end of the fifth. Discount them at 9.10 per cent a year and they come to Rs 1,000.000000/-, the price that makes this an issue at par. Discount the identical payments at the five year government SPOT rate of 6.90 per cent a year and they come to Rs 1,090.446383/-.

The difference is Rs 90.4464/- on Rs 1,000.00/- of face, or 9.0446 per cent of the face amount, and that is what 220 basis points for five years is worth in rupees. The convention sits inside the arithmetic rather than in a note beneath it: annual compounding, one discounting period a year, so an amount is divided by 1.0910 once for every year. The same coupon, the same maturity and the same yield on a semi-annual convention give a different price from figures that look identical in print. With the convention written beside the price, a reader can reproduce the sum. Without it, no amount of other detail will serve.

What does a finished snapshot let somebody do, and what does it not?

The rows will outlive the conversation that produced them, so step eight writes the scope into the document in one line. A finished snapshot supports a comparison between two readings of the same price under two different stated assumptions. A finished snapshot also supports a conversation about which of the six empty rows would change the picture most if somebody could fill it. A finished snapshot does not support a statement about how likely this borrower is to fail, it does not support a ranking against a borrower whose snapshot was built any other way, and it does not support a view about whether the spread is generous.

Three limits sit under that line, and none of them is optional.

First, the 40 per cent recovery is an assumption

No recovery study was read for this work, and no market quotes a recovery rate the way it quotes a yield. Move the assumption and the answer moves with it, and it moves a long way. Hold the 220 basis point spread completely still and only the assumption changes:

Same price, four answers, and the range runs from 3.1429 to 7.3333 per cent a year. The assumption is doing more than half the work, and step five gives it a row of its own instead of a mention for precisely that reason.

Second, the whole spread is being treated as payment for credit

In a real market some part of any spread pays for not being able to sell the instrument easily on demand. Every basis point of that read as credit makes the implied annual default rate too high. Split 0.40 percentage points off the 2.20 as payment for something other than default, and the remaining 1.80 percentage points divided by the same 0.60 gives 3.0000 per cent a year instead of 3.6667. One price cannot be split between credit and liquidity, so the snapshot says so in the liquidity row, and the error therefore runs in one direction only.

Third, an implied annual default rate is what the price says

An implied annual default rate is not a forecast and not a measured frequency of anything. Nobody counted defaults to produce 3.6667 per cent a year. The figure was solved backwards out of one spread and one supplied assumption, and reading it as the probability that Palash Cements Limited fails is a misreading of the arithmetic that produced it. The word implied is attached to the figure in every row and every sentence for exactly this reason, and a snapshot that drops the word has handed the next reader something it does not contain.

Who actually builds one of these, and when?

Somebody builds a snapshot at the moment a price arrives and a decision has not yet been made. A lending desk offered a private placement runs the eight steps before the meeting. The argument in the room is going to be about the spread, and the room needs to agree on the size of the spread before it can argue about whether it is enough. A treasury team recording a holding runs it so that the file explains itself to whoever opens it after the person who bought it has moved on. An analyst covering an issuer runs it every time the price moves enough to matter, and the value of the routine is that the earlier snapshots are still comparable.

The practical payoff is not the implied annual default rate at all but the six empty rows, and those rows are the agenda for whatever gets done next. A team that can see at a glance that it has no measured recovery, no spread history and no second issuer knows exactly which three phone calls would improve its position, and knows that no amount of further arithmetic on the numbers it already has will.

There is a second thing the routine buys, and it only shows up on the second or third build. Because the row order is fixed and every assumption sits in the same place, two snapshots on the same issuer taken three months apart can be laid on top of each other and the difference read off in one pass. If the implied annual default rate moved, exactly one of three things did it: the borrower's yield moved, the government SPOT rate moved, or somebody changed the assumed recovery. A snapshot written as free prose hides which. A snapshot written to this order cannot, and that is most of the reason it is worth the extra ten minutes.

The household version is the same shape. The deposit rate the cooperative bank offers is compared with what the same money would earn for the same five years elsewhere, the difference is written down, and then what is being assumed about getting the money back is written down too. The written assumption is the whole discipline. Most people do the subtraction and skip the line, and the line is the only part of it that stops the extra half a percentage point from looking free.

Palash Cements Limited, its five year bond and the government SPOT curve used here are invented.
Educational material. Not advice on any investment, tax, budget or market position.