How Securitization Arbitrages Bond-Market Inefficiencies

How can you dislike any column which includes the word "importunate"? John Kay's latest column is yet another salvo in the war against securitization, however, and despite his eloquence I'm afraid he's misguided. Here's the nut of his argument:

The risk characteristics of a bet can never be eliminated, but can be changed by division and aggregation. In every B-rated bond, there is A- and C-rated paper waiting to get out. Discard the C and your B has become an A.
This scheme works, but at a cost: in an efficient market, the value of the risk reduction will be precisely offset by a reduction in return.

Kay compares the investment banks behind securitizations to alchemists, who "created" more gold in the world than had ever been mined. "And so it came to pass," he says, "that the volume of investment grade securities far exceeded the value of investment grade credits."

And there he leaves it. But in fact, his conclusion is the very reason why securitization does work.

Think, for a minute, about your average bond investor. There are many fewer bond investors than there are stock investors, so you might not know very many of them. But they're not typical investors, by any means. They're extremely risk averse: if they buy a bond at 99 and it falls to 97, that constitutes a catastrophe in their world, while it would be completely normal volatility for a stock investor. What they are looking for is safety, and guaranteed returns.

Now the main skill of bond investors is to manage interest-rate risk. When rates rise, the value of bonds falls – so a bond investor's biggest fear is always that yields will go up. And bond investors are typically so concentrated on managing interest-rate risk that they often don't want to worry at all about other risks, like credit risk – the risk of default. As a result, there's enormous demand for AAA-rated securities – bonds which have a negligible probability of defaulting. Indeed, that demand vastly oustrips the natural supply of such securities – bonds issued by a handful of large industrialized nations, and a few other entities such as the World Bank or UPS.

When you have a systemic excess of demand over supply, that's precisely the point at which markets are no longer efficient. And indeed that's what's happened in reality. AAA-rated bonds, it turns out, yield much less than the mere reduction in credit risk can explain on its own. If you take a small increase in credit risk, and buy AA-rated bonds instead, then you have historically been able to get a much higher return – even after accounting for defaults.

Enter the alchemists. By splitting a B into an A and a C – or, more to the point, into a AAA and a C – you get to tap the demand for AAA-rated securities, which is unquenchable by natural supply. Because your synthetic AAA is hard to understand and illiquid, it will yield more than a natural AAA like a Treasury bond. But even so, you're taking advantage of a natural market inefficiency, and everybody wins. The investors get the AAA debt they crave, and you get more for your AAA and C combined than you would for just the B on its own. Sometimes the parts are worth more than the whole.

Now the investors are taking on risks, buying synthetic AAA bonds, which are not present in natural AAAs. The main one is market risk: the risk that the value of those bonds will fall dramatically. And there's also the risk of human error: that the people who carefully structured the AAA bonds to be bankruptcy-remote got confused, and made mistakes, and ended up creating a security with a higher risk of default than they originally intended.

But the underlying arbitrage – the reason why it makes sense to create AAA securities in the first place – remains. The bond market, dominated as it is by risk-averse bond investors, is simply not a perfectly efficient market. And securitization can help to arbitrage some of those inefficiencies.