Here's a trade. I've got these bonds, see? I will sell them to you. You will pay me $100 and get $100 face amount of bonds (if you like, you can get $80 or $120 face value of bonds, depending on where the bonds are trading - but let's make them par bonds, to keep things simple).
But we're not done. I will also write a contract under which, if these bonds default prior to maturity, you can hand them back to me, and I will give you back your $100. My loss will be the $100, minus whatever I can get from the defaulted bonds. In exchange for this commitment from me, you will pay me a running payment. That payment will be equal to (1) the coupon payment on the bonds (remember, they're par bonds, for simplicity), minus (2) a risk-free rate of equal maturity, plus or minus (3) a basis driven by the cost of funding and differences in relative demand for different sorts of payments. Let's say the bonds pay 6%, the relevant risk-free rate is 2%, and our funding costs are 50bps. Then you pay me 350bps running. The payments, and the contract, expire at maturity of the underlying bond.
Ah, but you have an objection. You're paying me this running payment in exchange for my promise to cash you out if the bonds default - but how do you know I'm good for it? Fair. Why don't we do this. I will collateralize that promise. At first my collateral will be quite small, since default is unlikely so the expected value of my promise is small, but it will go up if the bonds decline in value and/or my credit deteriorates.
Okay, fine, now we've got a deal. So ... what is our deal? I've sold you bonds in a spot sale - that much seems clear - and we've got ... this other thing, this contract. What do we call that contract?
Well, we could certainly call it a "credit default swap." Right? You make quarterly payments to me, in exchange for which I pay off par (and get the defaulted bonds) if the bonds default. Your quarterly payments are in an amount more or less equal to the "credit spread" of the bonds. There are some differences from the normal ISDA-approved CDS - for one thing, our contract as I described it above is about one particular bond, not a cheapest-to-deliver bond of the issuer; for another, real ISDA contracts now have fixed spreads and up-front payments rather than running payments actually equal to a credit spread - but those are within the bounds of a stylized CDS contract. Certainly if we actually wrote each other that contract, we could call it "CDS" and not have anyone get too broken up about any misuse of terms.
But we can also, with very teeny adjustments that do not affect the economic substance, call it something else. Specifically a "repurchase agreement," or perhaps - since that term is more usually used for overnight trades - a "repo to maturity." The main differences, I think, are:
(1) Instead of you paying me a "credit spread" on the bonds, I keep the income on the bonds and pay you a "financing rate," which is kind of a risk-free rate plus some sort of funding cost. Like, 2.50%. So I keep 350bps of payments on the bond. So no economic difference from the CDS.*
(2) Instead of my agreeing to pay you par and take the bonds if the bonds default, I agree to pay you par and take the bonds at maturity (or prior default). If the bonds go to maturity, then I pay you par and get a thing that immediately cashes out at par - which is a transaction that has no substance, no different from the CDS just expiring worthless at maturity. (In fact, you'd probably net it so I don't even go through the motions of paying you par.) If the bonds default prior to maturity, our trade looks like CDS: I pay you money and get defaulted bonds. Again, no economic difference from the CDS.
(3) Various mechanical bits and bobs are different. Maybe our running payments are daily or monthly or one-time rather than quarterly. Probably our collateral mechanics are expressed differently, though they get at the same thing - my ability to pay you off if the bonds default, whatever my then-current state of credit is. These things might have substance but they're second-order; the basic gist of you get real(ish)-time mark-to-market collateral is mostly true in both circumstances.
Okay. Now. It is normally a good and noble endeavor to make fun of Gretchen Morgenson, because she really really doesn't care at all if what she writes about Big Evil Banks is true or not, so it's mostly not. Thus when she writes a column in the Times that says:
MF Global’s debacle was a result of complex swaps deals it had struck with trading partners. While those partners owned the underlying assets — in this case, government debt — MF Global held the risk relating to both market price and default.
... it is easy to interpret that as saying "MF Global was blown up by CDS," which is false: it was blown up by cash holdings of bonds, financed by fairly normal repo-to-maturity arrangements that got exploded by collateral calls. And getting that wrong would be sort of stupid, so lots of people - including, with majestic disdain, ISDA itself - pile on and call her a moron. Or at least:
Morgenson is one of those reporters who sees CDS beneath every rock, and even blamed CDS for Greece’s fiscal problems — twice. ... In the MF Global case, she’s seeing CDS when she was actually looking at bog-standard repos, which aren’t derivatives at all.
But she actually doesn't say that! She doesn't say anything about CDS blowing up MF Global. She just says that MF Global wrote trades where the counterparty owned the bonds and MF Global bore the market and default risk. And that is true!**
Now, no one can deny that using the term "swaps" for repo trades, while maybe layman-intuitive, is unfortunate - or that calling even the simplest things "complex" is a terrible affectation for someone pretending to explain complexity - or that she then goes off the rails and talks a lot about CDS and credit-linked notes and God-knows-what having nothing to do with MF Global or anything else good and true and sensible. I am not here to defend Gretchen Morgenson.
I'm just here to say that, if Gretchen sees derivatives under every rock - so do I. And so should you. If you get hung up on whether to worry about "derivatives" posing risks to the financial system, or just plain old "leverage," don't.
I grew up in a derivatives business, so when I see a thing that takes one asset and disaggregates some of the risks and costs of that asset and apportions them to different parties - I sometimes, just for fun, like to call that thing a "derivative." Our little trade above separates the funding of the bond from its credit risk - one of us has the ownership of the underlying asset, but we're both parties to a bilateral contract that has payoffs based on the state of that asset.
It seems to me that, if you see derivatives in all zero-sum bilateral contracts that disaggregate risks of underlying financial instruments - and if you think in cash flows and triggers rather than names - then you will have a clearer head in thinking about derivatives. On the one hand, you won't run around flapping your lips about how terrifying Evil Derivatives are and how they'll blow up the world. (And you won't worry about the difference between vanilla repos and CDS, because CDS is pretty pretty vanilla.) On the other hand, you will probably be a little less sanguine to learn that the thing that blew up MF Global was not a "derivative," but rather just "leverage," or "secured lending," or "bog-standard repos." Because deep down they're the same thing.
Sad Proof [ISDA blog]
CDS demonization watch, ISDA vs Morgenson edition [Reuters/Felix Salmon]
* These numbers - or rather, their relationship - are not arbitrary. Adjusting for technicals in different markets, the bond interest rate minus the CDS spread really should kind of equal the risk-free rate plus the cost of funding the bonds.[Update: a comment at Felix Salmon's blog points out that I'm being a bit cute in assuming that the risk-free rate on our trade is a fixed rate. That is, in effect, the way CDS works - you keep the bond and its rates risk; I get only the credit risk. Normally, however, in repo-to-maturity, I would pay you a floating financing rate, so I'd retain the interest rate risk. I was eliding this point, which does have economic substance, to focus on the splitting of financing costs from credit risk, so, mea culpa. This is not, I think, an essential point: you can certainly have a secured financing with a fixed interest rate, and you could even reasonably call it a "repo-to-maturity" though it would be a bit different from a regular-way repo trade (more so than our stylized CDS is a bit different from a regular-way ISDA CDS contract). In any case, you don't need a "derivative" to fix the rate, or to parcel out funding, rates and credit risk.]
** Or, true enough. You can debate the semantics of who "owns" the bonds in a repo trade. Legally Gretch is probably right; for accounting purposes it's murkier.