Guest Post: Monetary Miscalculations From A World Captive To Models
Submitted by Atlantic Capital Management
Monetary Miscalculations From A World Captive To Models
Looking through the Federal Reserve’s newly released Discount Window data fills in some missing pieces surrounding the credit crunch in 2008. We now know why Senator Chuck Schumer was so concerned about IndyMac. In the three business days after June 19, 2008, IndyMac had to double its discount window borrowings from $200 million to $400 million. Four short days later, Schumer’s leaked letter forced IndyMac to ask the Fed for $1 billion.
Beyond some of these little details that end up providing granularity to the whole picture, there is still one piece of data that stands out as a singularity. Although it had become public knowledge over a year ago, the Lehman Brothers activity on September 15, 2008, still flashes a deepening warning as our economy and markets depend more and more on central banks.
On the surface, Lehman’s use of the Primary Dealer Credit Facility (PDCF, the investment banker’s discount window) seems to be insignificant. It was a momentous day, after all, with turmoil in every corner of the global marketplace. Why shouldn’t Lehman borrow $28 billion from the Fed on that Monday? It had filed for bankruptcy at about 1:30am that morning, so clearly it was in need of financing.
A lot has been published already about that volatile week. However, I still believe there is a hole in the “official” story as it relates to overall monetary policy. What is truly striking is not that Lehman used the Fed that Monday; rather the significance was that it was Lehman’s first use of the PDCF since April 16, 2008. Lehman Brothers did not use the Fed’s liquidity until after it had declared bankruptcy.
Conventional wisdom has seemed to settle on some version of the government allowing Lehman to fail as an example against moral hazard. But access to the PDCF would not have been viewed as a “bailout” any more than IndyMac’s $1 billion discount window activity. It is a fact, shown by internal JP Morgan documents, that $24.6 billion in fresh financing for Lehman would have been enough to avoid bankruptcy had it been done on Friday, September 12, 2008.
As it turns out, Lehman Brothers is not an exclusive term. The Lehman Brothers that went bankrupt was the holding company for the bank, LBHI. The rules of the PDCF state that holding companies are not eligible for “primary dealer” status. So LBHI’s US investment bank subsidiary, LBI, was the legal entity designated as a primary dealer.
But this should not have hindered access to the PDCF either. As I mentioned above, LBI did access the Fed’s program on five occasions in early 2008: $1.6 billion on 3/18, $2.73 billion on 3/24, $2.13 billion on 3/25, $2.13 on 3/26, and $2 billion on 4/16. All five were overnight loans.
Lehman even went so far as to make a dramatic spectacle of its first transaction. It put its CFO, Erin Callan, on CNBC on March 18 to declare to the world that it was going to use the PDCF in an effort to take some of the stigma out of it. It had clearly used the borrowing program and did it in plain sight.
Even the venerable Brian Sack (then the senior economist at Macroeconomic Advisers LLC, now the New York Fed’s head of open market operations) speculated to Bloomberg News on September 10, 2008 that:
“The PDCF could be used to keep Lehman operating until a broader solution was found. The challenge is figuring out what that broader solution is.”
So why did they wait until after the bankruptcy to finally borrow?
In short, it seems that the New York Federal Reserve got impatient.
In the weeks before Lehman’s end, the bank had pinned its survival hopes on the Korea Development Bank (KDB). As KDB ran through its due diligence process, some of Lehman’s counterparties began to get nervous, especially JP Morgan.
JP Morgan acted as Lehman’s clearing bank, meaning JP Morgan had some minor risk associated with that relationship, as well as extensive inside knowledge. But because the clearing capabilities were absolutely essential to LBI, Lehman kept JP Morgan up to date with all of its survival plans, including KDB.
On September 5, KDB emailed several executives at JP Morgan, including Jamie Dimon, to notify them that no deal would be done by Lehman’s September 10 deadline. That same day JP Morgan forced Lehman to post additional collateral for additional perceived clearing risk.
On September 9, KDB officially and publicly ended its interest. The stock tanked and Lehman’s management decided to pre-announce the third quarter’s results, a hefty loss, on September 10. Lehman’s credit default swaps spiked 200 basis points, and JP Morgan again got more collateral from Lehman (as did Citicorp).
At the outset of September, Lehman’s management declared a $40 billion liquidity pool for the company. Because of these collateral calls and the ongoing loss of short-term financing, Lehman’s actual, available liquidity pool was only $2 billion by September 10. Informally and quietly, the company began investigating bankruptcy options.
By the weekend of September 13 & 14, the Federal Reserve and the US Treasury were already actively seeking a solution. Treasury Secretary Hank Paulson was working with the Federal Reserve Bank of New York (FRBNY) President Tim Geithner to gather Wall Street CEO’s and get them to find a private solution. Paulson was adamant that no Federal guarantees were going to be used this time in a clear departure from the Bear Stearns episode (though Paulson later admitted that the Treasury would have financed some of the deal, he only wanted the CEO’s to come up with a solution ex-Treasury first).
Lehman itself was fully investigating a sale to Barclays. That sale came very close to fruition on September 14, except that the FRBNY wanted Barclays to guarantee Lehman’s obligations until the sale was finalized. Barclays agreed in principal but was rebuffed by the FSA (the UK’s equivalent of the SEC). The FSA stated that the guarantee would require shareholder approval first, and since there was no precedent it would not waive that requirement.
Dick Fuld, Lehman CEO, wanted Paulson to call British Prime Minister Gordon Brown to press the FSA, but Paulson denied the request. Fuld even tried to contact Jeb Bush, then a consultant for Lehman, to get his brother President Bush to pressure Brown.
It was clear that a deal was very much reachable at that moment, the “broader solution” that Brian Sack alluded to. All that was needed was some more time – time that could have been provided by the PDCF.
Apparently the FRBNY was unwilling to give any. It was decided by Sunday, September 14 that Lehman Brothers was to either have a concrete deal in hand or declare bankruptcy. The FRBNY’s directive was that Lehman would not open for business on Monday.
From the Valukas Report, we know that Lehman’s outside bankruptcy council, Harvey Miller, received a phone call on the way to meet with FRBNY that Sunday and was told by another attorney at his firm that “Citibank had been told that Lehman was being liquidated”.
Shortly after that, FRBNY issued a press release changing the terms of the PDCF to accept a much broader range of collateral. Again from the Valukas Report, “Upon learning of the expansion of the PDCF window, Lowitt [Lehmans’ CFO] and Fuld [Lehman’s CEO] initially believed that Lehman’s problem was solved and that Lehman would be able to open in Europe by borrowing from the PDCF.”
Lehman’s team responded to the change in PDCF as if it was a lifeline, but FRBNY only changed the terms for the other 16 primary dealers. For Lehman, it placed special restrictions on its ability to access the window – now known as the “Friday criterion”.
Not long after the press release Herbert McDade, Lehman’s President, with Miller at the FRBNY meeting, called Fuld to tell him that “the Fed has just mandated that we file for bankruptcy.” FRBNY’s general council, Thomas Baxter said that the filing needed to happen before midnight. McDade and Miller tried to protest but were told simply, “that it was decided and there were cars available to return them to the Lehman building.”
During its investigation, the Financial Crisis Inquiry Commission (FCIC) asked Baxter about the “Friday criterion”. He responded that the Lehman board minutes from that fateful Sunday essentially established LBHI was insolvent and its own bankruptcy council had advised, “it was likely the corporation ultimately have to file for protection under Chapter 11.”
Since LBI was the only entity allowed to access the PDCF, Baxter relayed that FRBNY believed the “Friday criterion” was needed. First, LBI’s collateral would be discounted by a steeper haircut regime than the other primary dealers due to LBHI being in bankruptcy – FRBNY believed that it had to discount the value of the parent’s support. Secondly, FRBNY required that LBI certify that any securities pledged were actually owned by LBI as of September 12 and that they were not subsequently transferred from LBHI. Again Baxter clarified that this was a necessary legal protection for taxpayers, ensuring that there would be no legal recourse to the collateral by the parent in bankruptcy.
These two restrictions effectively closed the PDCF to LBI without a bankruptcy filing for LBHI. FRBNY knew that Lehman was somewhat uniquely structured so that any securities or collateral were, in fact, “owned” by LBHI.
Baxter also states later in his response to the FCIC that, “after LBHI filed for bankruptcy, the FRBNY did extend on September 15 an aggregate amount of credit of approximately $60 billion to LBI, which enabled LBI to continue in business…To keep all of Lehman operating as a going concern, rather than only the U.S. broker dealer, would have required a vastly greater amount of credit and collateral.” [emphasis added]
That is quite a statement to make knowing full well that FRBNY had just expanded the range of collateral accepted. And even bolder considering the worthless junk that we now know was actually accepted from the other primary dealers as they got caught up in the panic.
We know that FRBNY had a radically different philosophy when it came to Morgan Stanley (MS). By September 29 Morgan Stanley was drawing $61.2 billion from the PDCF alone, and $100 billion from the Fed in total. That accounted for 10% of MS’s total funding needs and 25% of its short-term fundings.
We also know that as early as September 17 Morgan Stanley was already in trouble and drew a massive $27 billion in financing from the PDCF. There were unsubstantiated rumors that MS’s CEO John Mack told Citigroup’s CEO that same day “we need a merger partner”, and then he actually circulated a memo to employees blaming short sellers for the company’s problems.
From FRBNY’s perspective, it seems clear that the difference in treatment was the perceived insolvency of Lehman. But was that FRBNY’s call to make?
The timing of that call was also suspect, particularly in light of Barclays’ interest in merging. If Lehman and Barclays just needed more time to work out a deal and get the right approvals, why was FRBNY so against giving it? It had already established the means.
I think the answer lies in the exact timing of the bankruptcy filing. FRBNY wanted Lehman in bankruptcy that Sunday night – before the Asian markets opened for business. Remember that in the week of trading leading up to this point, the markets (particularly short-term credit) were pounded by ceaseless rumors and news about Lehman. Investor fear was rising due to the unending uncertainty.
So it seems the FRBNY went into that weekend with the idea that there would be a resolution by Sunday night, no matter which way it went. They believed, in my opinion, that getting Lehman off the front pages was the answer to the uncertainty problem, and by extension the short-term fear. If Lehman could be unwound in an orderly fashion, it would surely calm the markets.
It was a colossal miscalculation.
The fact that PDCF collateral terms were loosened for the 16 non-Lehman primary dealers shows that FRBNY expected some kind of funding tightness in the wake of Lehman’s bankruptcy. But despite that apparent loosening of monetary policy, FRBNY was still sterilizing its liquidity efforts. We know that the Fed shut its repo desk completely in mid-September just when liquidity levels were getting serious.
I believe these miscalculations are the obvious result of too much adherence to monetary models. I think that FRBNY ran many simulations that weekend and determined that Lehman’s bankruptcy would cause a quantifiable amount of disruption, and that the Fed would easily absorb the difficulties. It seems that the Fed believed that any short-term trouble would be more than balanced by the longer-term benefits of Lehman’s orderly demise (which was, in fact, orderly). Because of these mathematical, random walk constructs, a 1% probability is no concern at all.
We know that the Fed models were so far off in 2008 that the estimated probability of the Fed funds rate being near zero was at least a six standard deviation event – an impossibility in their minds. Even if they had updated their models for the building crisis (which I am fairly certain they did not) how much more accurate would they have been? A five standard deviation event?
The Fed is a creature of ineffective models. Every single monetary decision is based on them. But we have yet to see its full explanation for the actions taken on September 14, 2008. Counterfactuals are mostly unhelpful, but in this case could it at least admit that it was unequivocally wrong, and that its modeling assumptions fatally flawed? Any benefit from taking Lehman out of the news was absolutely overwhelmed by the abrupt, outright panic the bankruptcy filing created. It is not coincidence that Morgan Stanley needed emergency funding on September 16, nor that several more institutions failed in short order thereafter.
I am certainly not suggesting that Lehman bore no responsibility for any of this, or that FRBNY was the only problem here. Certainly JP Morgan and Citigroup should accept some blame for their blatant, extortive collateral demands. And the entire concept of credit creation through investment banks and securitized means were major contributors. Rather, I am simply trying to isolate this overall adherence to the mathematical modeling of financial behavior at the center of repeated monetary miscalculations.
In the current context of quantitative easing (QE), I have no doubt that the Fed has run the numerous random walk simulations to quantify any effects of ending or extending QE. I also have no doubt that those calculations will be as “effective” as they were in March 2010 and, unfortunately, September 2008.