The Journal had an article this morning about how cash equities traders are getting used to having computers as coworkers but I say unto you: can a computer do this?1

52. On March 31, 2010, Customer A, an investment adviser to a private fund, asked Jefferies to find buyers for several MBS, including Lehman XS Trust Series 2007-15N 2A1 (LXS 2007-15N 2A1) and Harborview Mortgage Loan Trust Mortgage Loan Pass-Through Certificates, Series 2006-10 2A1A (HVMLT 2006-10 2A1A). [Jefferies trader Jesse] Litvak approached a representative at AllianceBernstein about buying the MBS.

53. Litvak told the AllianceBernstein representative that the seller had offered to sell the HVMLT MBS at 58-00 and the LXS MBS at 58-8:


he will sell to me 20mm orig of hvmlt 0610 @ 58-00 but he is being harder to knock back on the lxs bonds … said that he thinks that one is much cheaper yada yada yada … he told me he would sell them to me at 58-8 (30mm orig) … I would be fine working skinnier on these 2 … but think you are getting good levels on these …


is he paying u or am I?


all the levels I put in this room are levels he wants to sell me … I will work for whatever you want on these …. so to recap levels he is offering to me:
hvmlt 06-10 2a1a (20mm orig) @ 58-00
lxs 40mm orig at 58-8…

Bot em


Can u wash the hvmlt and [add] 5 ticks to lxs?…


thats fine.

54. Litvak misrepresented to AllianceBernstein the prices at which Jefferies had acquired the MBS for re-sale. Litvak bought the HVMLT MBS at 57-16 (not the “58-00” he told Alliance Bernstein) and he acquired the LXS MBS at 56-16 (not “58-8” he represented).

55. Litvak also misrepresented the compensation that Jefferies would receive for these trades. AllianceBernstein purchased the $20 million HVMLT MBS at 58 and $40 million of the LXS MBS at 58-13. As a result, on the HVMLT trade, Litvak made 16 ticks for Jefferies; he did not work for free (or “wash” the trade) as he had agreed. And, on the LXS MBS, Litvak made 61 ticks for Jefferies; he did not work for “5 ticks” as agreed.

56. As a result of his misconduct, Litvak made over $600,000 more for Jefferies on the LXS trade and over $50,000 more on the HVMLT trade.

That’s from the SEC’s complaint against former Jefferies trader Jesse Litvak, who apparently made a habit of this sort of thing. He would (allegedly!) tell a potential buyer (seller) of RMBS bonds that he had a seller (buyer), but he would inflate (deflate) the price that he was supposedly getting from the other side in order to inflate his spread. This worked 25 times – that the Feds caught – and allegedly made Jefferies $2.7 million in deceptive profits. This is particularly lovable: Read more »

  • 11 Dec 2012 at 5:47 PM

AIG’s Remaining Bailout Reduced To Rounding Error

A while back I built a spreadsheet to do math about AIG, and it took me a long time and led to basically one short post with what I still think was a rather lovely blobby picture, so I’m just going to shamelessly reuse that spreadsheet with slight updates and be all OOH LOOK AN IRR:

So yeah: as the AIG bailout saga comes to its sort-of conclusion, we can sort of conclude that the government made a 5.6% return on its money. Assumptions etc. in the original post; the accounting profit ties out reasonably well, if you squint, with the Treasury’s official math.

Herewith some random observations and questions on AIG:1 Read more »

  • 26 Mar 2009 at 5:07 PM

What If No One Came?

It strikes us that the PPIP plan requires a certain faith by the administration. Specifically, that balance sheets are not actually so underwater that even a 30% subsidy is a hollow gesture. What’s more, how sure is the administration that actual price discovery is something that any of these institutions actually want? Clearly, given the seller-financing leverage shell-game baked into the plan, the hope is that bids will buoy up. The problem, however, was perfectly highlighted on today’s FDIC call.
What, a banker effectively asked, if his participation were to “blow a hole in the capital?” Would capital requirements be waived or adjusted to keep the institution from running afoul? (Probably not). The meaning was somewhat veiled, but the broader implication was that actual price discovery would so impact the balance sheet and impact equity capital so negatively as to reveal this particular institution to be liver sausage.
What about bids or asks that resulted in no actual transaction? Would they, one voice trembled, constitute… (gulp, deep breath)… pricing data sufficient to trigger mark-to-market treatment? (Could be!)
As if on cue, another questioner wondered if the FDIC could force participation. (Probably not). You could almost feel the exhale of held breath.
Would participation exempt an institution from special examination? (Laughter). No exhale on this one.
Could it be that the biggest problem confronting the nation isn’t that Goldman Sachs might make money buying assets in the PPIP because Tim “The Safecracker” Geithner is in league with the devil? What if almost no one participated at all? One side of us thinks that we are reading too much into all this. Another thinks that if you can read between the lines you can almost hear the cracks widening.