Will Banks “Cough Up Executives” in the Treasury Bid-Rigging Scandal?

The Department of Justice may face an early test of its long-overdue policy change, that the government will seek to prosecute individuals, including executives, along with those of corporations. As Sally Yates, Deputy Attorney and author of the memo setting forth the new policy, put it, “We mean it when we say, ‘You have got to cough up the individuals.’”

As Bloomberg reported on Thursday, private plaintiffs have filed two suits alleging bid-rigging by the 22 primary dealers, adding pressure to an ongoing Department of Justice investigation. We’ve embedded the more recent filing, Cleveland Bakers and Teamsters Pension Fund v. Bank of Nova Scotia et al., at the end of this post. From the article:

The same analytical technique that uncovered cheating in currency markets and the Libor rates benchmark — resulting in about $20 billion of fines — suggests the dealers who control the U.S. Treasury market rigged bond auctions for years, according to a lawsuit….

The plaintiffs built their case against the 22 primary dealers who serve as the backbone of Treasury trading — including Goldman Sachs Group Inc., JPMorgan Chase & Co. and Morgan Stanley — using data from Rosa Abrantes-Metz, an adjunct associate professor at New York University who has provided expert testimony in rigging cases.

Bear in mind that investigations and litigation is underway, and no charges have yet been proven. However, in the last major Treasury bid-rigging scandal, in 1991, the Fed didn’t bother to wait for the Department of Justice to act. Our summary in ECONNED:

[Salomon Brothers trader Paul] Mozer and other government bond dealers routinely traded an arbitrage that involved Treasury bonds traded on a “when issued” basis. The Treasury would announce its funding calendar, and traders would start making a market prior to when the bonds would actually be auctioned. The arb play involved shorting the “when issued” bonds. A short seller eventually had to buy the instrument to close out his position, so he was committed to making a purchase at some point.

Mozer decided to squeeze the dealers doing this trade, but to do that, he needed to control the Treasury auction in the maturity he targeted, the two-year note. Primary Treasury dealers like Salomon submit bids at Treasury auctions, both for themselves and for customers.

Mozer first tried bidding for 100% of the offered amount for Salomon. A Treasury deputy secretary called and politely but firmly reminded him that the government had a gentlemen’s agreement that a buyer is restricted to a maximum bid of 35% of the total offered at any auction. The government, after all,
wants to raise money on good terms, not enrich dealers, particularly ones engaging in anticompetitive practices. Remarkably, Mozer not only got abusive with the official, but became openly defiant, next submitting a bid for 240% of the auction. That led to another unproductive call from the Treasury to Mozer, plus a request to the Fed, which runs the auctions on behalf of the Treasury, to lower the bid to 35%. Mozer tried yet again with a 300% bid, the Fed haircut it again, and the Treasury made the 35% informal understanding into an official limit. Mozer went on a rampage, submitting multiple 35% bids, making abusive calls to the Treasury, and trying to get media support.

The officialdom of the firm took notice, and the second most senior officer, Tom Strauss, ordered Mozer to apologize and take some time off.

But Mozer was not deterred. A few months later, he not only submitted 35% bids in Salomon’s name, but also submitted bids on behalf of unwitting customers and created phony customer trades after the fact to cover his tracks. He was clumsy about it. The Treasury got wind of what he was up to, conducted an investigation, and sent a letter to one of the clients that Mozer had falsely said was bidding, with a copy to Mozer.

When a firm crosses a regulator, the right response is to quickly roll over and show your belly: a massive display of contrition and swift punishment, at minimum a suspension, of the perp. That did not happen. Mozer showed the letter to [John] Meriwether, his boss, who showed it to Strauss. They agreed it was very serious, “career threatening.”

They were right, but the career it ended was [CEO John] Gutfreund’s. He was out of town; the matter was left until his return. Mozer, still in place, submitted yet another phony bid at the next auction. When Gutfreund got word, he agreed the matter was serious, and the top brass debated whom to notify (the Treasury or the Fed). There was no discussion of reining in Mozer, much less punishing him. The phony bids continued. Some hedge funds heard of Mozer’s ploy and started placing similar buy orders (except theirs were legitimate).

Gutfreund waited more than a month to leash and collar Mozer. His profits during this period were substantial, and Gutfreund thought he could hang on to Mozer’s ill-gotten gains. When the Salomon chief finally sat down with the Treasury, he argued that the firm’s conduct had been proper, even though the
squeeze had become visible and costly to competitors. The Treasury was not satisfied. An investigation ensued and the results did not support Gutfreund’s claims. When the Federal Reserve, the regulator in this matter, found out via reading the story in the press, the end came quickly. Gutfreund, Strauss, Meriwether, and the firm’s general counsel resigned in a matter of days.

Similarly, when Barclays tried shifting blame for its misconduct in the Libor scandal †o the Bank of England, the Bank forced the resignation of the chairman, CEO, and president. But similar shows of spine from American regulators are almost antique. That is in part due to the fact that the New York Fed disbanded its primary dealer surveillance unit in 1992.

Despite the fact that the efforts to look into the Treasury market are gaining momentum, I have serious reservations about the statistical analysis in the suit embedded below. It does not provide direct evidence of collusion, but tries to use statistical approaches to infer that the price anomalies around the time of auctions can only be explained by collusion. By contrast, the Department of Justice appears to be first looking for direct evidence of collusion, as in dealers sharing information in chat rooms and coordinating bidding strategies. One would assume that if they find this behavior (and press leaks suggest they have), statistical evidence would then serve to demonstrate impact, i.e., harm.

I admit I’ve only skimmed the case, and it does not include exhibits. However, I didn’t see any sign in the narrative that the analysis allowed for the fact the trade sizes in Treasury auctions are vastly larger than in secondary market trades. You would thus expect them to be at lower prices, meaning higher yields. By contrast, the sole expert who did the analysis in this filing, as far as I can tell from my cursory look, appear not to have considered this idea at all.

Similarly, we see unduly confident statements like this:

This pattern of laying off risk that yields were going to move against the primary dealers suspiciously ceased when there was an auction. Rather than hedging by buying in one and selling in the other, or vice-versa, Defendants instead were, to a statistically significant degree, seen only around auctions going in the same direction in both the spot and futures markets. Defendants’ volume of activities also increased markedly in these periods. The only plausible explanation for such “doubling down” behavior is that Defendants were confident that they knew how the markets were going to move at the time of auctions, because they were manipulating the auction prices.

In corporate underwriting (stocks and bonds), it is well understood by issuers that the dealer will price the underwriting (the “price” being the price at which he buys the stock or bonds from the company) so as to assure a profit to investors who buy the day the security is sold. In the stone ages of my youth, the Goldman syndicate department viewed a 15% first day appreciation as a good level to hit (as in not so high as to annoy the company selling the stock but high enough to keep investors signing up for Goldman deals). Later in the 1980s, when I was at McKinsey, one team studied the pricing pattern of various firms on stock offerings, and some were more company-friendly (as in less first-day price appreciation) than others. Put it another way, it has long been seen as normal and necessary to price new issues at below the market price to get underwriter to underwrite them and investors to buy them.

The larger point is that to get these auctions off, the Treasury and Fed expect the dealers and the investors to make a certain amount of profit to reflect the risk of taking down large orders time and time again when all their competitors are doing just the same thing. So you’d expect the pricing to be a bit lower and the yields to be higher. The question is whether the “juice” is in line with what is needed. The plaintiff’s expert, Abrantes-Metz, contends that 69% of the auctions appear to have been rigged. The average yield premium she cites, 91 basis points (close to 1%), does seems awfully rich. But by taking the implicit position that any discount in price, meaning higher interest rate, is suspicious, Abrantes-Metz looks to have overstated her case and made it easy for banks to assail her analysis.

Bear in mind, however, that this is just the initial filing. Even with my reservations, it looks certain to pass summary judgment. That means the plaintiffs get to do discovery and dig into e-mails and phone logs and depose witnesses. As a result, the Department of Justice and the private litigant will be on parallel paths, and thus will make it harder for the DoJ to go easy on the banks if there are any smoking guns. Stay tuned.

gov.uscourts.nysd_.446741.1.0
gov.uscourts.nysd.446741.1.0