Government Securities: The Market and Who Takes Part

A government security is the state's own borrowing, written down so that the right to be paid can pass from one holder to another. One borrower stands behind it, the amounts and their dates are settled when it is created, and the market that follows has two sides. On one side the security is made and the borrower is present. On the other it only changes hands.

What is a government security, and what makes it one instrument rather than a category?

The plainest version of it is the one that survives every later complication, so start there. A government securityThe state's own borrowing written as a transferable instrument: stated amounts, stated dates, and a right to receive them that can be passed to somebody else. is a promise by the state to pay stated amounts on stated dates. The promise is the whole of it. There is no cleverness hidden underneath, and a reader who has already met a bond as an object has met most of this instrument.

Two features carry every argument in this guide, so it is worth separating them before either does any work. The first is that the borrower is the same at every length. The state promising to repay one lender next year is the same state promising to repay another in three decades. Nothing about who is making the promise differs between the shortest instrument on the record and the longest one. One borrower at every length sounds obvious written down, and yet almost every mistake described below comes from forgetting it.

The second feature is the one that does the heavy lifting. The promise is transferableAble to be passed to another holder without the borrower being asked, told or involved. The new holder simply steps into the old holder's place.. A holder who wants their money before the final date does not go back to the state and ask for it early. The holder finds somebody else who is content to wait, and hands the promise over. Everything that happens to a government security after it is created happens between holders, and none of it reaches the borrower at all.

The mechanism is familiar on a smaller scale. Suppose a lender advances a neighbour Rs 20,000/- and writes the terms on a slip of paper. If the slip says pay me, then only that lender can be paid, and if the money is needed sooner the only route is to go back and ask. But if the slip says pay whoever is holding this slip, something quite different has been built. A cousin can take it off the lender for a price the two of them agree, and the neighbour still pays the same Rs 20,000/- on the same day to whoever turns up with it. The neighbour does not need to be told, does not need to consent, and owes exactly what was written down at the start. The second slip is a government security in miniature.

One thing has not been said. Why the state borrows, how much it borrows in a year, and what that borrowing means for a national account belong to the borrowing programme, covered separately. The instrument is a smaller and much more tractable thing than the policy behind it, and treating the two as one subject is how readers end up unable to explain either.

Where does a government security come from, and who is standing there?

The market has two sides and they are so different that reading them as one thing is the commonest error a newcomer makes. Take the first side on its own terms. The creating side is the primary sideThe side of a market on which a security is created. The borrower is present, the instrument did not exist before the transaction, and money moves to the borrower.: the side on which a government security is created.

Three things are true here and nowhere else. The borrower is present in the room. The security does not yet exist before the transaction and does exist after it. And money moves towards the borrower, the entire point of the exercise from the borrower's end. Every other transaction in a government security for the rest of its life fails at least two of those three tests.

The creating side produces a set of terms that are then fixed for the life of the instrument. The amount promised. The dates on which amounts fall due. The coupon, if the instrument carries one. And the horizonThe length of time between today and the date an amount falls due. A three year horizon means the money comes back three years from now., meaning how long the wait runs. Once those terms are fixed they never move again, and every price struck for the rest of the instrument's life is a price for that unchanged set of terms.

The route by which a government security is offered, the mechanic by which it first reaches a holder, and the schedule on which government borrowing is offered at all are set by the Reserve Bank of India at rbi.org.in. Each of those rules is real, precise, and liable to change. A reader would carry a stale mechanic into a real situation and be confidently wrong, so a mechanic written out from memory would be worse than useless. The block near the foot gathers every one of them as a labelled row routed to that source.

There are two kinds of statement in a subject like this one. There are statements that can be checked by rerunning the arithmetic, and there are statements that are true only because a specific authority currently says so. The first kind belongs in the text. The second kind belongs at its source, with a pointer to it. Mixing them teaches a reader to trust both equally. Trusting both equally is exactly the habit that gets people into trouble.

What happens on the side where a government security changes hands?

Now the second side, and the empty chair in it. The changing-hands side is the secondary sideThe side of a market on which an existing security changes hands between holders. The borrower is not a party to the transaction and is not asked about it.: the side on which a government security that already exists moves from one holder to another.

Reverse each of the three tests. The borrower is absent. The security existed before the transaction and exists in exactly the same form afterwards. And the money moves between two holders rather than towards the state. A holder who wants money now sells the promise to somebody content to wait, and the price the two of them agree is whatever discounts the remaining payments at the rate being asked for that length of wait on that day.

Nothing in that transaction reaches the borrower. The transfer alters no date. No amount changes. Nothing is added to or taken from what the state has undertaken to pay. The creating side settles the promise once; the changing-hands side re-prices that same settled promise every single day anybody looks at it. The two sides are not two versions of one event. They are events of two different kinds, and merging them is what produces the failure examined later.

The household version is a wedding hall. A household books a hall for a date next winter and pays for it. Six months later their plans change and another household takes the booking off them, at whatever the two of them settle on: possibly more than was originally paid, possibly less. The hall does not care. The hall owes one hall, on one date, to whoever holds the booking. The price at which the booking changed hands reports something about what a winter date is worth in the middle of the year. The price reports nothing whatever about the hall.

Who may deal on this side, and under what conditions, belongs once again to the Reserve Bank of India at rbi.org.in. Named here, not written here.

A second asymmetry is worth pulling out, and it is the one that surprises people who have only ever met the first side. The creating side is an event. The creating event happens once for each security, at a moment, and then it is over. The changing-hands side is not an event at all: it is a condition that persists for the whole life of the instrument. Every day the instrument exists is a day on which somebody could look at it and put a number on it, and mostly somebody does.

Why does one borrower face six different rates on the same day?

Here is where participant categories come from, and the answer is more interesting than a list of names would be. The same borrower is offering money back at one year and at thirty years, at the same moment, on the same terms of promise. A holder whose own obligations fall due next year and a holder whose obligations fall due in decades are not looking at the same instrument, even though they are looking at the same borrower.

A range of tolerable waits is the whole reason a market like this has categories rather than one undifferentiated crowd. Different lenders have different waits they can tolerate, and a wait is the one thing that genuinely differs across a government's borrowing. The market shows not six borrowers and not six qualities of promise, but one borrower met at six different lengths of waiting.

The invented SPOT curve used throughout this sequence carries six recorded points, all on annual compounding: the one year SPOT rateThe rate for money placed today and returned at one stated future date. One rate for each date, with no re-investment step in between. of 5.90 per cent a year, the two year SPOT rate of 6.25 per cent a year, the three year SPOT rate of 6.55 per cent a year, the five year SPOT rate of 6.90 per cent a year, the ten year SPOT rate of 7.35 per cent a year, and the thirty year SPOT rate of 7.60 per cent a year. Every one of those six rates is assumed rather than observed. A real curve would sit at other levels and would move every day, and the arithmetic below would run in exactly the same way on either set.

Subtract the ends and see how much room the shape covers. One basis point is one hundredth of a percentage point. The thirty year SPOT rate less the one year SPOT rate is 7.60 less 5.90: 1.70 percentage points, or 170 basis points. A gap of 170 basis points is a wide spread of pricing for a borrower whose creditworthiness is by construction identical at both ends. Nothing about the promiser accounts for a single one of those 170 basis points. The waiting accounts for all of them.

The everyday version of this sits in any market street. Ten shops in one mall all borrow from the same lender. One wants the money back off its books within a year, another can leave it for a decade. The lender quotes ten different rates for the ten different lengths without forming a single new view about any of the ten shops. The rate varies because the wait varies, not because the lender has changed its mind about anybody.

Categories of holder are defined by an authority rather than derived from arithmetic. Which categories exist at all, and what conditions apply to each of them, is set by the Reserve Bank of India at rbi.org.in, and it moves. So is any measure of what each category holds. A figure invented to fill that gap would look exactly like a fact and would be worth less than nothing.

Why does every rate here carry the word SPOT?

Because a rate without that word is ambiguous in a way that costs money, and this particular set of recorded numbers has a trap built into it on purpose. Take the one year SPOT rate of 5.90 per cent a year and the two year SPOT rate of 6.25 per cent a year, both on annual compounding, and pull out of them the rate that applies to the second year alone.

The working runs as follows. One point zero six two five squared is 1.12890625. Divided by 1.0590 that comes to 1.06601157. Taking away one, the one year one year FORWARD rate is 6.6012 per cent a year. The record rounds that to 6.60 per cent. Nothing was forecast, guessed or supplied. The figure was already sitting inside two numbers present from the start.

Now put that beside the three year SPOT rate of 6.55 per cent a year and look at how close they are. The difference is 6.6012 less 6.55: 0.0512 percentage points, or 5.12 basis points. Two figures five basis points apart, and they are completely different objects: one is the price of waiting three years, and the other is the price of waiting through the second year alone.

The closeness is not an accident in the record and it is not something to engineer away. On any smooth curve, FORWARD rates land near SPOT rates. A reader who meets both without labels will merge them, and once merged they cannot be separated again by care alone. So the discipline is mechanical and it applies throughout this sequence: every rate is written as a SPOT rate or as a FORWARD rate, spelled out, every time, with no exceptions for brevity.

One more thing about that 6.6012 per cent a year, the figure most likely to be misread. The FORWARD rate is arithmetic pulled out of today's recorded SPOT curve. It is not anybody's expectation of what the one year SPOT rate will be a year from now.

Where does a holding actually live?

Here is the field readers are most often wrong about, and the wrongness is charming rather than careless: people picture a certificate. Something printed, something with a border, something that could be put in a cupboard and handed over across a table.

A holder of a government security ordinarily holds no such object. The holder has instead a holding entryA line in a record that constitutes the ownership rather than evidencing it. There is no separate object it is a receipt for.: a line in a record. And the crucial part, the part that a reader has to sit with for a moment, is that the entry is not evidence of an ownership stored somewhere else. The entry is the ownership. There is no more real version of it kept in a safe.

The same arrangement is accepted everywhere else without anybody noticing. A bank balance is not a stack of notes with a name on it in a vault somewhere. A balance is a line in a record. When one person pays another, no notes move; a record is edited in two places. Nobody finds this troubling, and nobody should find it troubling here either.

The consequence is the one worth carrying. If a holding is an entry, then transferring it is a change to a record rather than a handing over of anything. And that is precisely how the promise can move between holders without the borrower being told, asked or involved. There is no object that has to be located, produced, endorsed and delivered. There is a line that reads one thing and then reads another.

The form a government security is held in, and where the holding is recorded, are set by the Reserve Bank of India at rbi.org.in. So is the convention that decides when a purchase is paid for and delivered. Both are confirmed at that source, and no settlement cycle appears anywhere above in any form.

What does a price change alter, and what does it leave exactly where it was?

Where a price comes from, and how a yield is solved out of one, are covered separately. The separation between the two sides is worth working through, carefully and once. Take the simplest claim on the record: Rs 1,000.00/- promised by the government in three years, with nothing paid in between.

A sum cannot be reproduced without the compounding convention, so the convention sits inside the arithmetic rather than in a note beneath it. Every rate and every price here is struck on annual compoundingOne discounting period a year. An amount due in three years is divided by one plus the rate three times, rather than six times at half the rate.: one discounting period a year. So Rs 1,000.00/- due in three years is divided by 1.0655 three times, and it comes to Rs 826.684201/-. The same six recorded numbers on a semi-annual convention would give different prices and a different set of FORWARD rates, and a reader who is not told which convention is running cannot check a single figure here.

Now walk that one claim through both sides of the market. On the creating side, the Rs 1,000.00/-, the date it falls due and the three year horizon were settled when the security came into existence, and none of the three can be reopened by anybody. On the changing-hands side, the holding can pass to somebody else at whatever price the rate being asked for that horizon implies on the day, and it can pass again the day after, and the borrower is not a party to any of it.

Suppose the rate being asked for the three year horizon is different tomorrow and the same claim changes hands at a different figure. Read the two columns. The price at which the promise transfers between two holders has changed. The Rs 1,000.00/- has not, and neither has the date it falls due. The borrower's position is word for word identical either way. Merging those two columns is the single error at issue, and it is the error the failure block below is built around.

The right-hand scale in that drawing repays a moment's attention. Only one level on it is marked, and that is not a design shortcut. The record carries one rate for the three year horizon and therefore one price for that claim. The record does not carry a second day, a second rate or a second price, so no second mark can honestly be drawn. A scale left mostly blank is better than one filled with figures that came from nowhere.

What cannot be established about this market?

More than can be established, and that is the honest shape of a record like this one. The absences are worth setting out plainly rather than working around.

The record holds no count of participants, no share of holdings by any category, no traded volume, no issuance schedule and no result of any offering. Every one of those is a real, measurable thing that exists somewhere, and not one of them exists here. Supplying any of them from memory would be producing figures from nowhere, and a figure produced from nowhere is the most damaging thing a reference can carry. In print it is indistinguishable from a figure that was checked.

The second absence is arithmetic rather than data, and it is the more important of the two. The invented SPOT curve carries six recorded points and nothing between them: one, two, three, five, ten and thirty years. There is no four year SPOT rate here. There is no nine year SPOT rate. There is no twenty nine year SPOT rate. There is nothing shorter than one year at all.

So no line is drawn between the points and no value is read off one. Interpolating looks harmless, so the reason is worth stating carefully. Reading a four year rate off the gap between the three year SPOT rate of 6.55 per cent a year and the five year SPOT rate of 6.90 per cent a year requires a method to be chosen, and different methods disagree. A straight-line reading and a curved reading give different figures. Two treatments working from the identical record would then print two different numbers for the same object, and a reader who noticed would be right to distrust both.

Where a reader expects a rate in between, the answer is that the rate is not in the record. Drawing the gaps as gaps is more honest than filling them, and it happens to be the thing most curve pictures get wrong.

A nodeOne of the horizons at which a rate is actually recorded, as against a value read off a line drawn between two recorded horizons. and a point on a drawn line are not the same kind of thing, however similar they look once the line is inked. The six nodes here were recorded. Everything else on that stretch of road would be a construction, and constructions belong to whoever is willing to name their method. The construction and publication of a benchmark government curve is routed to the Clearing Corporation of India Limited at ccilindia.com.

The last point is a writing rule rather than a data one, and every sentence above obeys it. No rate goes up or down anywhere above. Up and down mean the price in one sentence and the yield in the next, and mixing the two says the opposite of the arithmetic without anybody noticing. So a move is written as a rise in the yield or a fall in the yield, every time.

So what is a government security good for, once the routing is stripped out?

Quite a lot, and less than a newcomer hopes. A government security is a promise from one borrower, settled once and never reopened, that can be moved between holders without the borrower being disturbed. The combination is unusual, and it is why this instrument became the reference every other bond in this subject area is measured against. The promise gives it one identity across every length of wait; the transferability gives it a price that anybody can observe and argue with.

The one thing it cannot do is stand still. A promise that can be transferred has a price, and a price that is struck fresh between holders will move whether or not anything about the borrower has changed. A reader who takes only one thing away should take that one. A fixed promise and a moving price, held apart, keep the rest in order.