A swap as a loan
In his leaden exposition on the “bilaterality” of the ISDA Master Agreement, JC remarked that, despite looking like a bilateral, unfunded instrument, a swap is, in reality, an implied loan.
People come out from their jobs, most of which are meaningless to them, and they watch me jump 20 cars, and maybe get splattered. It means something to them. They jump right alongside of me. They take the bars in their hands and for one split second, they’re all daredevils. I am the last gladiator in the new Rome. I go into the arena and I compete against destruction and I win. And next week, I go out there and I do it again. And this time, civilisation being what it is, and all, we have very little choice about our life. The only thing really left us is a choice about our death. And mine will be glorious.
- —Evel Kneivel (1971), on jumping the Grand Canyon
This throwaway comment prompted explosions of indignance from friends, colleagues and people who JC greatly respects so he decided to double down on it. Look, if you are going to go down in an attempt, it might as well be at the Grand Canyon. So, here goes.
To recap the background to that post:
Whereas most finance contracts imply dominance of one party and subservience of the other — a loan, for example, has a lender who takes mortgages, sharpens knives and extracts excruciating covenants the way a dentist does teeth and a borrower whose mortal soul is traduced with the indignities of indebtedness but who must yet feign affection through gritted teeth and deep resentment — swaps are not like that.
Swaps, so conventional wisdom would have it, are exchanges among peers. “A swap is,” cognoscenti are given to say, “an equal-opportunity, biblically righteous compact between equals. Yes, the Transaction may go in and out of the money but it swings either way. As it does, each participant is an honest rival for Lady Fortune’s favour, however capricious may she be.”
That is the theory. But JC says, at least beyond the limited class of inter-dealer swap transactions, and even then, often within it, fiddlesticks. This conventional wisdom is not true. In the bigger picture, swaps are loans.
An “end user” swap is, in fact, a “synthetic” loan from a dealer to customer. [1]
JC is blessed with charitable friends who forgive intellectual softness.
“Oh, well,” they are prone to say when the old boy goes off on one, “I suppose you could analyse an interest rate swap as a pair of off-setting loans. Yes, that seems strictly true. But is it not rather to miss the point? Seeing each party lends to the other, and as notional principal flows in both directions at the same time, the loan, as you put it, cancels out. The parties to a swap are not really lending to each other, old thing.”
But this is not what JC means. When a dealer provides a swap to a customer, even after taking into account the “offsetting loans” of this traditional theory, economically the dealer is still lending outright to the customer. The dealer doesn’t actually disburse any money to the customer, but that doesn’t matter: the dealer applies the loan proceeds instead in financing its hedge on the customer’s behalf.
Customers and dealers
Now there is a boundless universe of “end user” swap Transactions. Here, one party is a “dealer” and the other — the “end user” — is a “customer”. Hence, the expressions “sell-side” — dealers — and “buy-side” — customers. These are the great majority of all swap arrangements in the known cosmos.
So what, in a bilateral arrangement, determines who is “dealer” and who is “customer”? It is not who is “long” and who “short” the swap exposure — one of the great swappist beauties of the ISDA framework is that customers can go long or short, as they please. Nor is it who pays “fixed” and who pays “floating”. Rather it is who is “on risk” and who is “flat”. Customers want to change their net economic exposure. Dealers do not.
The argument JC will mount for the remainder of this piece is this:
To change one’s economic exposure to any financial instrument involves capital investment. An investor must either make that capital investment itself from its own funds or it must borrow the money from someone else. In a swap, that “someone else” is the dealer. The way a dealer “makes that capital investment” is by hedging.
Therefore, the economic difference between the customer and dealer in a swap Transaction is that the dealer is lending and the customer is borrowing.
The capital cost of changing your position
For a customer, the object of any Transaction is to change its overall market exposure: to get into a position it did not have before, or get out of one it did. But dealers do not do this. Dealers stay flat.
“Hang on, though, JC: if a swap is bilateral, how can that be so? Does it not follow that if the customer changes its position one way, the dealer must change its position the other way?”
In the narrow confines of a specific ISDA Transaction — ignoring any hedging arrangements — this is so. But the narrow Transaction is not the whole picture. There are hedging arrangements. In the wider context of the parties’ overall net risk positions, customers invest: they take on risk. They change their positions. Dealers don’t. Their interest is only in commission. They do not have a market position.[2]
The dealer “provides” exposure by sourcing it in the market, delta-hedging it, and charging its customer a commission. There are all kinds of enterprising and funding-efficient ways it can do this, but fundamentally, the dealer stays market-neutral. The customer’s credit risk for the life of the trade is all the excitement the dealer wants. More, actually. As long as the dealer’s market-side hedges work, the only market risk the dealer takes arises if the customer fails. That is to say, the dealer has customer credit exposure for as long as the customer stays in its risk position. As long as the customer remains solvent, it chooses when to exit. The dealer is (morally) committed to staying in.[3] If the customer wants to exit before a stated term, even where it has no legal right to do so, it is an odd swap dealer indeed who will make a price.[4]
Now: the thing about being net long, or net short, a financial exposure is that someone needs to acquire that exposure. Even if it is an “unfunded” derivative like a rate or index, in the real world one can only achieve that rate or index return by making an actual capital investment in the underlying products that generate that rate or index. Even if it is not the investor, someone has to commit capital to generate that return. That someone is the dealer.
This is the same capital expenditure that a bank must make when extending a loan. The difference is only that a bank commits that capital directly to its customer, wheras a swap dealer commits it to its own hedging programme and then charges the financing cost of that capital to the customer.
This is the fundamental difference, for example, between an delta-one equity swap and a margin loan. This is why equity swaps are also called “synthetic prime brokerage”.
Worked example
Imagine Hackthorn Capital Partners holds USD10m shares of Lexrifyly, Inc. and wants to gain exposure to Cryptöagle, GmbH. Hackthorn is fully invested and has no spare cash.
Hackthorn can do one of three things:
(i) sell Lexrifyly and buy Cryptöagle — that is, make an outright long investment out of the proceeds of sale;
(ii) hold Lexrifyly and borrow to buy Cryptöagle — that is, take a margin loan;
(iii) hold Lexrifyly and get synthetic exposure to Cryptöagle via an equity swap from its dealer, without, apparently borrowing any money.
To make it easy, let’s say the quantity and strike price for each is 10m shares at USD1, and the market value (MV) for Lexrifyly is USD1.2 and for Cryptöagle is USD1.4. Let’s say the investment period was 1 year and interest accrued at 1% per annum.
Here are the positions:
Sale
If it sells Lexrifyly outright and buys Cryptöagle outright:
- Sold: USD10m Lexrifyly @ USD1.
- Borrowed: 0.
- Bought: 10m Cryptöagle @ USD1.
- Current MV of Cryptöagle: USD1.4.
- MV(10m Cryptöagle) equals USD14m
Loan
If it keeps Lexrifyly and borrows to buy Cryptöagle:
- Held: USD10m Lexrifyly.
- Borrowed: USD10m.
- Bought: USD10m Cryptöagle @ USD1.
- MV(10m Lexrifyly) + MV(10m Cryptöagle) - USD10m and accrued interest equals:
- USD12m + USD14m - USD10.1m equals USD15.9m
Swap
If it keeps Lexrifyly and buys a swap on Cryptöagle struck at USD1:
- Held: USD10m Lexrifyly.
- Borrowed: Zero.
- Swap outgoings: Interest on USD10m
- Swap incomings: (MV(Cryptöagle - Strike) * 10m)
- MV(10m Lexrifyly) + ((MV(Cryptöagle - Strike) * 10m) - accrued interest on USD10m
- USD12m + USD4m - USD100,000 equals USD15.9m
No surprise that switching out of one stock and into another yields a different result. But the economics of the “loan” and the “swap” are identical. The investor pays a floating rate and has a USD10m notional loan value deducted from its pay-out. Like a loan, the equity swap gives the investor exposure to Cryptöagle whilst keeping its exposure to Lexrifyly, which it uses to fund cashflows on its new capital asset.
This is a form of leverage. As it would have in a loan.
The floating rate Hackthorn pays is implied funding. The dealer will only accept this if it is satisfied Hackthorn has enough capital to finance its swap payments and settle any differences at termination. This is the same risk calculation a bank would make on a loan. (To keep it simple, I have ignored the scope for synthetic margin loan and rehypothecation.)
But, hang on: this is a bilateral arrangement, right, so isn’t the converse also true for the dealer? Isn’t the dealer, in a sense, “borrowing” by paying the total return of the asset to get “exposure” to the floating rate in the same way? Is not a “short” swap position, for a dealer, exactly the same as a “long” swap position for a customer?
No, because in providing these swap exposures whether long or short, the dealer simultaneously delta-hedges. It does not change its market position. The customer buys an exposure: that is, starts without and ends up with a “position”; the dealer manufactures and then sells an exposure: it starts without a position, takes an order, buys a hedge, and then sells an offsetting position to the customer, ending up where it started, flat, and without a position.
Provided the dealer knows what it is about, its main risk in running a swap portfolio is not market risk — it should not have any — but customer credit risk. The moment a customer fails, the dealer’s book is no longer hedged: what was a delta-hedge is now an outright long or short position. when the reason your customer has blown up is that its highly levered investment in the same segment of thinly traded illiquid stocks has just gapped downwards, this is enough of a risk to be a real bummer, as those involved in the Archegos omnishambles will tell you.
Fixed/floating swaps
Nuncle: ’Tis none so mincey as a Farrington chop
And nowt so loansome as a fixed rate swap.
Ok; that’s a delta-one equity swap. But synthetic prime brokerage is, surely, an unusual case?
Aren’t “normal” swaps truly bilateral? How about good old-fashioned interest rate swaps? Surely simply paying a fixed rate and receiving a floating rate has none of these same loan-like characteristics?
The first thing to say here is that in the real universe of actual, non-derivative instruments, interest rate cashflows do not exist independently of an investment in principal.[5] This is because an interest rate is, by definition, the income on a capital investment.
Oh, sure, you can detach and sell a strip of coupons off a bond: okay. But to do that, there must first be a bond, and you must buy it, cut it up and sell the stripped bond principal back into the market. Once you’ve done that, you have your disembodied interest cashflow, all right — but someone else has its dark inversion: this weird, mutilated, principal-only, zero-coupon instrument that trades at a heavy discount to its fully-limbed equivalent. It will exist, but unhappily: like Weird Barbie or one of those intercised children with no daemon in His Dark Materials. Once you have sold the principal you might not be able to see it any more, but it is still there.
“In the real world interest rates do not exist independently of principal investments. This is because an interest rate is, by definition, the income on a capital investment.”
Repeat: in the real world, interest rate cashflows depend on income-generating assets. It stands to reason. A rate without principal is like a shadow without a boy.
Do swaps change all that?
No: because at some point, swaps must be based in the reality from which they are derived. This is not bitcoin, folks.
Derivatives as “engines of hypothesis”
Derivative
/dɪˈrɪvətɪv/ (n.)
FINANCE: (of a product) having a value deriving from an underlying variable asset. (emphasis added)
When the Children of the Forest wrought their wristy magic on the First Men and the Way of the One Agreement passed into common understanding our leaden, earth-bound notions of “necessary principal” were swept away. Only then did the swap market take wing, upon the nuclear power of leverage. Income could flow, at last, unshackled of its leaden principal host, and was free to nudely frolic in ISDA’s glittering starlight.
The “synthetic” world is an alternative, magical realm. Normal rules of space-time do not apply. There are amulets, magic instruments and imaginary tools with which even ordinary mortals can do impossible things. As we have seen, we can isolate income from principal and trade them hypothetically, as discrete instruments.
But gravity is not banished; only postponed. At some point, this fantasia must alight on planet Earth and engage with real-world instruments, because that is what it is all derived from. Ultimately, somewhere, someone needs to construct each enchanting payoff from grubby, weighty, principal-laden corporate rights and obligations. Those rights and obligations are — on our mortal coil, must be — embedded in a scaly crust of principal.
And that principal must be financed.
So if you want to earn floating rate on a notional of a hundred bucks in the real world, you pony up a hundred bucks. That means selling an investment you already own: going off some other risk. Even free cash deposited with the bank is an investment: it is a loan to the bank. If you don’t want to sell down another investment, you must instead borrow from someone. We are somewhere near the foundational mythologies of modern capitalism, here, by the way. this sort of talk gets bitcoiners quite agitated.
If that someone is the dealer from whom you bought the floating rate note, consider the final cashflows: you pay a fixed rate on your loan; you finance that from the income generated by your asset portfolio, the principal on the note you’ve bought cancels out against the principal of your loan and bingo: you have an interest rate swap.
Leverage is a state of mind (or balance sheet)
One last way to look at this: an interest rate swap is a levered investment in a debt instrument. Interest rate swaps are, in this sense, “synthetic fixed income prime brokerage”: a margin loan to buy a fixed income asset.
We can see this by considering the parties’ respective economic positions before and after trading. The customer changes its net position; the dealer does not. Swapping a fixed cashflow for a floating one is to keep the “asset” funding that fixed cashflow, and to borrow the funds required to buy the new floating-rate asset. Because that borrowing has the same principal amount as the purchased floating-rate asset, the principal amounts cancel out, and the customer is left with just the floating rate cashflow, for which it must pay the fixed rate cashflow it has agreed.
Without that implied loan, the customer would have to sell an asset to raise the proceeds to buy the floating-rate bond outright from the dealer. That is, pay the principal amount to the dealer, and acquire the interest and principal cashflows of a floating rate asset. Here the customer is certainly not borrowing anything. It is making a fully-funded long investment.
“But, but, but JC: can’t you see? If you pay someone 100 and they pay you the return of an instrument worth 100 and interest, you have loaned them the money?”
Quite so: but that is the nature of a floating-rate bond. It is a loan. But it is not a loan to the dealer. It is a loan to the issuer of the floating-rate bond. If the dealer is paying you the return of a floating-rate bond you may be assured it has used your money to buy the floating-rate bond, to hedge itself. It is flat. You have not, net, lent the dealer anything.
The bigger picture
This is all part of a bigger picture, more sort of meta-observation about the world, that all business can be boiled down to financing assets: in a sense, every business is the simple quest to earn a better rate than you pay. This is just what borrowers do (if you can’t earn a better rate than you pay, you don’t borrow).
What a business does to achieve its “return on equity” — whether it is to make widgets, sell them, provide professional advice, or clip a ticket in the military-industrial agency problem complex that is the financial services industry — all of these activities boil down to ensure income reliably outperforms outgoings. The same is true for every recipient of those outgoings, and so on: the market is an overlapping complex of “borrowings” — whether it is labour, expertise, services, land, or plant — where, curiously, no-one wants to be the ultimate lender, because to be an ultimate lender is to hold physical cash, and to hold physical cash is to earn nothing, and since acquiring cash through lawful means generally involves incurring cost, holding physical cash is the quintessential dead-loss investment.
(Contra the bitcoiners, this, rather than “degenerate fiat capitalism”, is the inherent reason “trad-fi money” loses value. Fractional reserve banking may also cause losses in value, but they are incidental. Money, in itself, is uninvested capital. It is meant to lose value. That is your incentive to buy some productive capital with it. Then its losing value becomes someone else’s problem.
Thus, any number of distinct parts of the financial system when you look at them up close, begin to resemble each other, and a large part of the value proposition of any financial services firm is optimising its financing position:
- Swap dealers buy physical hedges but then lend them in the market to raise funds against them that they can return to their treasury for use as operational capital.
- Hedge funds borrow on margin from prime brokers, who hold their assets as collateral and, to optimise their funding, rehypothecate the assets and lend them into the market against cash or high-quality assets they can return to their treasury department.
- Private Wealth investors and asset managers put their assets into agent lending programmes to earn a return raise funds against them that they
|
Premium content
Here the free bit runs out. Subscribers click 👉 here. New readers sign up 👉 here and, for ½ a weekly 🍺 go full ninja about all these juicy topics👇
|
See also
References
- ↑ Where regulations require dealers to post variation margin outright against their own out-of-the-money swap exposures (rather than simply calling for it when their customers are out-of-the-money), the regulations make the financial system less stable, more risky, more leveraged, and more prone to the market calamities that fueled the global financial crisis. Bilateral variation margin is a category error. Swap dealers should not collateralise their customers. There. I said it.
- ↑ To be sure, dealers may have a “trading axe”, as they delta-hedge across their whole derivatives trading book. But — especially since the global financial crisis, and in large part before it, this axe is largely aimed at flattening the risk in their trading book and optimising their financing costs of all that hedging.
- ↑ Not legally, necessarily: there may well be regulatory capital reasons that a dealer must have a termination right on say 30 or 60 days’ notice — these give negotiation teams the thrill of arguing about the gilding of a lilly — but do not expect a dealer ever to exercise those rights, short of a market meltdown that threatens its own existence.
- ↑ Swap break costs are easy enough to calculate.
- ↑ This is just as true of dividends on equities, of course.