Felix Salmon

Sovereign debt has long had a semi-fluid system of seniority, what Anna Gelpern calls a "seating chart". Back in the 1980s, countries were generally pretty happy to default on foreign bank loans. They occasionally defaulted to their fellow governments, too, but they generally never defaulted on foreign bonds: back then, bonds and loans were very different animals. At the top of the seniority spectrum were the "preferred creditors" – the IMF, World Bank, and other multilateral institutions who always got paid, no matter what.

Things changed over the years: when the Brady Plan turned loans into bonds, bonds lost a lot of their special status – although it’s worth noting that as recently as 1998 Russia never defaulted on its Eurobonds even as it was defaulting on everything else. Nowadays, no one considers bonds to be at all senior to loans, and there’s a fair amount of debate over whether foreign debt, generally speaking, is a safer proposition than domestic debt. Most interestingly, even the formerly-inviolate preferred creditor status is coming under attack.

The first chinks in the preferred-creditor armor started appearing in 2001, when the World Bank wrote a guarantee on public Colombian bonds. The structure was seemingly very clever: if Colombia ever defaulted on a bond repayment, the World Bank would pay up, and take on the debt. Since the World Bank was a preferred creditor, it would get paid back by Colombia, and the guarantee would then roll over to the next bond repayment. That way, a relatively small World Bank guarantee could underpin a large number of sequential Colombian bonds.

That structure was broken by Argentina in 2002, with the complicity of the World Bank. When Argentina defaulted on a bond guaranteed by the World Bank, the World Bank stepped in and paid up, as it had to – but then changed its repayment terms with Argentina so that the sovereign wouldn’t need to pay the World Bank back before the next bond was due. So much for the clever "rolling reinstatable guarantee", which hasn’t been used since.

Today, Alan Beattie has an excellent piece on another way in which people try to extend the World Bank’s preferred creditor status to private-sector obligations. It’s called ICSID, the International Center for the Settlement of Investment Disputes, and it’s an arm of the World Bank. If a company has a dispute with a country, it can take that dispute to ICSID for binding arbitration. (The country also needs something known as a Bilateral Investment Treaty, but those are extremely common and verge on the ubiquitous.)

Because ICSID is part of the World Bank, any ICSID award has the legal status of a treaty obligation: it’s considered a good notch or two senior to any other sovereign debt. Until now. Argentina has simply ignored that bit of international law, and hasn’t paid out on its ICSID obligations even though it’s remained current on most of its other debt. Dani Rodrik, for one, is sympathetic.

The lesson of this story is simple: if you think your sovereign debt is senior to some other sovereign debt, think again. There are norms and rules of thumb and generally-accepted principles, but ultimately sovereigns are sovereign, and they can and will act as they please. Even if that means violating international treaty obligations.

It’s like Telephone (Chinese Whispers) in the blogosphere today. First there’s a sober report on Venture Wire warning that "cracks of the economy" might affect the VC industry:

Perkins said VCs need to be aware of the "cracks of the economy" that might affect the industry, such as the failure of auction-rate securities. Several start-ups hold these obscure financial instruments, which many considered nearly as good as cash but are now illiquid after many bond auctions failed recently. Lawler said his firm surveyed his group of 60 portfolio companies and has so far identified 12 that have some amount of cash in auction-rate securities.

"Some amount of cash," of course, could be as little as $25,000, and by no means means "most" or "all" of their cash. But Penny Herscher immediately jumped to a worst-case scenario:

A very serious risk has emerged for some venture backed companies as a result of the credit crunch gripping the markets. Today some small VC backed companies have found that their cash is no longer liquid and this is the worst kind of crisis to a young company.

In turn, this was picked up by Michael Arrington, who posted a story under the ridiculously alarmist headline "20% Of Valley Startups Can’t Get To Their Cash".

Arrington’s 20% figure comes straight from the Venture Wire story: it’s 12 divided by 60. But of course Perkins never said those companies couldn’t get to their cash, he just said that they had "some amount of cash in auction-rate securities," which isn’t the same thing at all. For one thing, not all auction-rate securities are failing or illiquid.

So take a deep breath, everybody, you’re all going to be fine. Having money in auction-rate securities does involve taking a risk which most startups shouldn’t and needn’t take: this is a good warning. But things aren’t nearly as dire as the blogosphere would have you believe.

Vipal Monga takes a look at the CDO and CLO markets, and asks: are they really dead? Or just taking a breather? He draws the conclusion that they’re very different animals. CLOs, he says, are crucial to the loan market:

CLOs represented almost two-thirds of primary demand for loans in the syndication market over the past three years. The loan market likely cannot function without them unless lenders decide to hold loans on their books.

Mortgage-backed CDOs, on the other hand, were driven more by greed than necessity.

The mortgage-backed CDO simply takes already-securitized CMBS and RMBS and repackages them. Their popularity, however, soared after managers discovered they could profit from an arbitrage between interest paid to their lenders and the higher interest they harvested from securitized bonds in their portfolios.

According to CapitalSource’s Szwajkowski, about 75% of CDOs created over the past few years were such arbitrage vehicles. The rest were buyers of so-called whole loans, meaning they acquired the mortgages before they were securitized, playing a role similar to that of CLOs in the loan markets. That arbitrage play is now moot.

I think this is a very important distinction. Yes, the demand for CMBS and RMBS will fall if repackagers and arbitrageurs exit the market. But if that happens, there will be a healthy consequence: the investors who were blindly rushing in to the CDO market will now be forced to buy the underlying MBS instead, leaving less room for confusion and opacity – and also making the MBS market, if and when it recovers, that much more liquid and transparent.