Bailout

The Greek derivatives aren?t Goldman?s fault

By Felix Salmon

The first thing you have to know about the Greece-Goldman story is that the definitive account was published by Nick Dunbar of Risk magazine in July 2003. He kicked off by saying that the deals he was writing about “are likely to prove controversial” — he probably never guessed just how controversial they would end up being, nor how long it would take them to achieve that status.

The details of the deal are more or less what I suspected, and indeed the Spiegel story hews so closely to Dunbar’s account that the Risk article was clearly a primary — if not the primary — source. Here’s Dunbar, explaining what went on:

The cross-currency swaps transacted by Goldman for Greece?s public debt division were ?off-market? ? the spot exchange rate was not used for re-denominating the notional of the foreign currency debt. Instead, a weaker level of euro versus dollar or yen was used in the contracts, resulting in a mismatch between the domestic and foreign currency swap notionals. The effect of this was to create an upfront payment by Goldman to Greece at inception, and an increased stream of interest payments to Greece during the lifetime of the swap. Goldman would recoup these non-standard cashflows at maturity, receiving a large ?balloon? cash payment from Greece…

It seems the total credit risk incurred by Goldman Sachs was roughly $1 billion. Effectively, Goldman Sachs was extending a long-dated illiquid loan to its client.

Dunbar also goes into a lot more detail about the fees on the transaction than the NYT does. I think it’s reasonable to accuse the NYT of being unfair to Goldman when it writes that Greece “paid the bank about $300 million in fees for arranging the 2001 transaction”. The point here is that because this is a cross-currency swap rather than a bond deal, the interest payments are going the wrong way: Goldman is sending Greece a steady stream of payments over the course of the deal, and then being repaid with a big balloon payment at the end. Essentially, Goldman is continually lending Greece money, and getting no interest payments in return, until maturity a long way out.

So the fee associated with the deal isn’t a fee for arranging the transaction, as the NYT would have it: instead, it has to cover all of the credit and market risk that Goldman Sachs is taking on in lending the money to Greece. What’s more, a very large chunk of the fee was immediately paid to Depfa, which sold Goldman a $1 billion 20-year credit default swap on Greece to hedge its credit risk. And for what it’s worth, Dunbar’s article puts Goldman’s total charge for the transaction at $200 million, not $300 million: I have no idea which figure is more reliable.

It’s also a bit disingenuous of the EU to start saying now that Eurostat was not aware of the transaction. Put aside for one minute the fact that Eurocrats have been known to read Risk; Dunbar’s article actually goes into some detail about how the Eurocrats knew exactly what was going on:

The drafting of ESA95?s section on derivatives was the subject of fierce arguments between the government statisticians and debt managers of certain eurozone countries.

The statisticians wanted derivatives-related cashflows to be treated as financial transactions, with no effect on deficit or interest costs, and with the derivatives? current market value stated as an asset or liability. The debt managers opposed this, insisting on having the freedom to use derivatives to adjust deficit ratios. The published version of ESA95 reflects the victory of the debt managers in this argument with a series of last-minute amendments.

In particular, ESA95 states in a page-long ?clarification? that ?streams of interest payments under swaps agreements will continue? having an impact on general government net borrowing/net lending?. In other words, upfront swap payments ? which Eurostat classifies as interest ? can reduce debt, without the corresponding negative market value of the swap increasing it.

In other words, Eurostat knew that Greece, Italy, and others were planning this kind of deal even before they happened, thanks to their successful lobbying efforts with respect to ESA95, and it was inevitable that they would structure deals with investment banks doing exactly what they did.

So while it’s entirely fair to blame Greece for trying to hide its debt, and to blame Eurostat for letting it do so, I think that blaming Goldman is harder. It was surely not the only bank involved in these transactions, and the swaps were simple enough to be shopped around a few different banks to see which one could provide the best deal. Structuring swaps transactions is one of those things which investment banks do. If countries like Greece buy swaps in order to hide their true fiscal status, then that’s the country’s fault, not the banks’. No self-respecting bank would decline such a transaction because they felt it was unfair to Eurostat.

Yes, I’m sure that Goldman put a team of people onto the Eurostat rules and made that team available to the Greeks. But let’s not blame the advisers here, for structuring something entirely legal and which the Greeks and Italians clearly wanted to be able to do all along. This is a failure of European transparency and coordination; Goldman is a scapegoat.

Full post as published by Macroprudential on February 16, 2010 (boomark / email).