by John Lounsbury
A major scandal that affected many major broker-dealers in recent years has gotten less publicity than it should have. I refer to the use of a “safe” investment vehicle known as ARS (Auction Rate Securities). These were peddled by registered representatives of brokerage houses to trusting retail investors as alternatives to money market funds with similar low risks but higher yields. These representations turned out to be far from correct; court ordered restitution by brokers to investors so far have exceeded $67 billion. This came from a market that peaked at a value of $330 billion in early 2008.
To put this in perspective, the giant Madoff scheme was smaller, with a peak value variously estimated from around $50 billion to as much as $65 billion and recoverable losses estimated to be between $18 billion up to over $40 million.
To be sure the Auction Rate Securities situation is not a Ponzi scheme, as the Madoff case was, but when you lose either $18 billion or $67 billion, the question of whether it was in a Ponzi scheme or some other form of misrepresentation becomes an academic question to the investor.
What is an Auction Rate Security?
Here are definitions from Investopedia:
An Auction Rate Security (ARS) is a debt security that is sold through a dutch auction. The auction rate security (ARS) is sold at an interest rate that will clear the market at the lowest yield possible. This ensures that all bidders on an ARS receive the same yield on the debt issue.
In essence, an auction rate security is a long-term debt issue, but it acts as if it were a shorter term issue. This is because interest rates are reset approximately every month, depending on whether the issue is tax exempt, and the interest is paid either shortly after the auction yield is settled or every quarter or half year.
The idea is that the interest rate reset allows the higher long-term yield, or at least most of it, to be paid on securities that are highly liquid because they are sold at auction every month, or at other intervals that ranged from 7 to 35 days. The auctions served the dual purpose of maintaining liquidity and keeping interest rates current to market conditions. The principal is stable, supported by the fact that all the securities eventually mature a face value. Or at least that is the design concept.
The Flaws of the ARS Design
The flaws involve both liquidity and stability of principal. A review paper summary from Securities Litigation and Consulting Group (by Joe Prendergast, Craig McCann and Eddie O’Neal) gives a complete accounting of the history, operation and shortcomings of Auction Rate Securities.
The assumption is that the auction will be essentially fully supported and the entire demand for redemption can be met by auction sales each month. In 2007 and earlier, auctions auction failures were hidden by the brokers themselves buying the unbid bonds. This hid from investors the weakening market demand and declining liquidity that was developing. In early 2008 the auction failures were accelerating and the brokers were unable (or unwilling) to continue absorbing the unwanted securities.
If there are securities that are not bid in auction then that portion of the ARS is illiquid until such time as the necessary numbers of securities are bid to satisfy all investor demands for payment.
When the securities not bid at market (or contract allowed rates) then they may be sold at less than par value. In that case the short-term nature of the ARS is lost and it reverts to behaving like the underlying long term bonds. Liquidity can be achieved but only with loss of principal in such a situation. Liquidity with stability of principal is not maintained. The “essentially a money market or cash equivalent” description, as promoted by the brokers, simply did not exist.
Stability of Principal – There are two flaws here.
The first flaw has already been described. The short-term stability of principal is lost when the auctions fail.
The second flaw derives from the assumption that all the securities will mature – there will be no defaults. It is possible to design an ARS which reserves from coupon payment flows for a low amount of default. But when defaults rise above that reserved amount, principal is not preserved. That is what happened to many ARS because some contained elements of the toxicity that infected credit assets in the financial crisis:
Munis and other bonds with guarantees from insurers who were suffering from downgraded creditworthiness due to sub-prime mortgage and other poorly performing debt securities;
Inclusion of CDO securities (collateralized debt obligations) with obscure default risks; and
Inclusion of investment company and corporate preferred stock which have no maturity dates like bonds do.
The SLCG review paper identifies two built-in bad incentives.
The first of these was the implication to investors that the broker-dealers would support the auctions. This was not an explicit commitment.
Secondly, the ARS prospectuses specified a maximum rate for each ARS. This put a floor under the auction price that could not be exceeded (to the downside).
These perverse incentives, or inducements, were the fatal handcuffs that destroyed the entire investment concept. Investors were sucked in by expressed and implied guarantees that made ARS investment seem better than alternatives. In any case when something seems better than everything else, it is wise to ask if it is too good to be true.
The SEC actions for recovery have been made on behalf of retail investors but not corporate buyers, presumably on the theory that large corporate investors should have known what they were getting into. The corporate buyers who have been caught holding the bag have three options:
- Negotiate a settlement with the ARS underwriter;
- Sell the securities in a private secondary market, often at a huge discount; or
- Change the classification of the security from short-term to long-term and wait for maturity to recover principal.
The video below from Bloomberg nearly two years ago discusses how the loss of liquidity at the beginning of 2008 was still a problem more than 18 months later.