I once came across a definitional difference between financial speculation and gambling. This might seem rather trivial, or only of interest to some moral philosophers. But it was of great significance when the Chicago Board of Trade was trying to get permission to trade financial futures that were cash settled, rather than settled by delivery, alongside the long established commodity futures markets. The latter were seen as of obvious business value because they allowed farmers and manufacturers to reduce their exposure to price uncertainty. That they were settled by actual physical delivery seemed evidence in their favour. The fact that the markets could also be used for purely speculative trades too was just an inevitable by-product.
But financial futures that were settled by cash payment appeared to many people as though they simply allowed traders to take bets on more things than before. If the new products were merely opportunities for gambling then they would fall foul of the state of Illinois’s anti-gambling laws.
The famous Chicago economist Milton Friedman wrote a note advising that financial futures and other financial derivatives were of real economic value, and would allow CBOT to offer risk management services to various businesses that otherwise had to face interest rate and currency volatility. The State of Illinois accepted this argument and fully legalised financial futures trading, to the great benefit of the CBOT and investment banks.
The definition that I came across was that gambling involved the deliberate creation of risks purely for the purpose of allowing people to bet on them. But financial speculation meant traders assuming already existing risks. So although a speculator might be “taking a punt” on say the value of Greek bonds in the same way that a gambler was risking money on a horse at Newmarket or the fall of the dice in a casino, he or she was also taking that risk away from somebody else by buying the bond.
If speculators are more informed or rational or better capitalised than the existing holder of the risk then we could say that the overall balance of risk is improved and people who wish to lay off some of that risk can, in a liquid market, find a willing counterparty to trade with at a reasonable fair value.
Whether this is a sufficient moral justification, I leave to others to consider, but it seems to have some substance. Or at least it did till we started hearing about synthetic CDOs and CDSs. A conventional Collateralised Debt Obligation (CDO) is a company set up to buy mortgage backed securities, financed by the sale of new securities to investors. It therefore transfers risk (voluntarily) from the originators of the mortgage, such as banks or savings and loans, to new investors. As we know, this ended in disaster for many investors but the problem was misunderstanding of the risk, not the CDO as such.
A synthetic CDO depends on the existence of CDSs – credit default swaps. A CDS is a bilateral contract like any other swap, involving an exchange of cashflows between two parties, one of which is typically a market maker, usually a bank. With a CDS one cash flow is a fixed set of payments but the other is contingent – if the underlying credit is declared in default, a payment by the protection seller is triggered to the protection buyer.
A CDS allows insurance to be purchased against default, which sounds reasonable. But it also allows a protection buyer to become “short” (profit from the potential default of) the underlying credit. Unlike going short an actual bond or stock, where you must borrow the security and then sell it, buying a CDS doesn’t require any ownership of the underlying credit. The origins of credit derivatives were in banks trying to reduce their exposure to particular borrowers they had lent to. But the tradeable CDS breaks the link between owning the credit and insuring against its default.
As a result, there is no limit to the amount of CDSs that can be issued, if there are willing writers and purchasers. And there is no limit in the number of synthetic CDOs that can be created. For a synthetic CDO is designed to mimic the payoffs of some reference portfolio of credits but it multiplies the risks rather than redistributing them. The pot of sub-prime mortgages is and was fixed but it was and is possible for an unlimited number of trades to be created that are linked to the value of those mortgages.
This is one reason why there are calls for banning “naked” CDSs. The analogy is that naked shorting (selling a security you don’t actually own) is illegal on most stock exchanges. Since you must borrow a share to short it, there is a limit to how much shorting can go on and it is all based on redistributing risk. Short selling is a legitimate and potentially socially valuable aspect of securities markets, since better informed or smarter traders can use shorting to make the market more efficient as well as part of risk-management in pairs trades (where you go long an undervalued security and short the one you believe is overvalued).
The legality of the sort of transactions highlighted by the SEC civil suit against Goldman Sachs is hard to determine. Goldman denies any wrongdoing. But it is quite easy to see how the general public and a lot of sophisticated financial commentators see synthetic CDOs as a bad idea which involve no social value and appear only to exist to allow German banks to speculate with other people’s money, for the benefit of investment banks.