Being in certain rooms at certain times seems to be a good predictor of selling a book. Bin Laden's bedroom on the night of his death is an obvious one, and various days in the Oval Office have or may soon have their chroniclers, though the world still awaits the unabridged memoirs of the guy who cleaned out Jeff Gundlach's office at TCW. But the Treasury Department conference room where regulators imposed TARP on eight big banks seems to have been especially fecund; by my count Hank Paulson has already published his account, somebody in the room seems to have contributed to this account, and now Sheila Bair has written a book that includes hers, which was excerpted in Fortune today.
This is weird because - well, one, because a bunch of guys (and Sheila Bair) in suits discussing the terms of a preferred stock purchase in a conference room is not necessarily the first place you'd look for thrilling literature, but also, two, because the accounts are all pretty similar. Here's Bair's take on the bankers' reaction to the TARP terms:
I watched Vikram Pandit scribbling numbers on the back of an envelope. "This is cheap capital," he announced. I wondered what kind of calculations he needed to make to figure that out. Treasury was asking for only a 5% dividend. For Citi, of course, that was cheap; no private investor was likely to invest in Pandit's bank. Kovacevich complained, rightfully, that his bank didn't need $25 billion in capital. I was astonished when Hank shot back that his regulator might have something to say about whether Wells' capital was adequate if he didn't take the money. Dimon, always the grownup in the room, said that he didn't need the money but understood it was important for system stability. Blankfein and Mack echoed his sentiments.
Coincidentally, Dimon is always the grownup in these accounts, too; what is new here is really Bair's take as the head of the FDIC:1
The fact remained that with the exception of Citi, the commercial banks' capital levels seemed to be adequate. The investment banks were in trouble, but Merrill had arranged to sell itself to BofA, and Goldman and Morgan had been able to raise new capital from private sources, with the capacity, I believed, to raise more if necessary. Without government aid, some of them might have had to forgo bonuses and take losses for several quarters, but still, it seemed to me that they were strong enough to bumble through. Citi probably did need that kind of massive government assistance (indeed, it would need two more bailouts later on), but there was the rub. How much of the decision-making was being driven through the prism of the special needs of that one, politically connected institution? Were we throwing trillions of dollars at all the banks to camouflage its problems? Were the others really in danger of failing? Or were we just softening the damage to their bottom lines through cheap capital and debt guarantees?
So what can one make of this? One place to focus is the phrase "cheap capital," which Vikula uttered and Bair endorsed.2 One rule of thumb in banking is that you should always take cheap capital - you want to avoid capital because it's usually expensive. So why did Wells Fargo's Richard Kovacevich bitch about it? And why, other than tactical protesting-too-much, were Jamie & friends ambivalent?
One thing I've pondered recently is the cost or profitability of various bailouts, and I suppose we can apply that knowledge to the question of: how cheap was TARP?3 We have after all discussed the fact that the TARP capital purchases signed at this meeting were all paid back at a profit, with double-digit (and, in the case of GS and MS, over 20%) IRRs for the government. That's great, or something, but in hindsight it's not exactly cheap. Goldman wouldn't have been thrilled to issue debt at 20%, even in October 2008.
But hindsight is always, um, 20% IRR. Was it cheap at the time? Obviously October 2008 was not a market particularly susceptible to answering that - as Bair says, Citi might not have been able to raise $25 billion at any price. On the other hand markets were at least in theory open to most of the banks, and from the comfort of 2012 why not just blithely apply that theory? As long as you let your assumptions do most of the work, you can stupidly rough out a valuation for the TARP capital purchases; here's a simpleminded approach4 and here is the chart it generates:
What this is meant to show is:
- On October 13, 2008, when Treasury gave these banks $120 billion, how much was it subsidizing them? How much was it overpaying compared to where the market would have valued these securities? Imagine this as Vikram Pandit's napkin-scribbling, if his napkin had a Black-Scholes calculator.5 So this chart suggests that Treasury gave Citi $25 billion in exchange for preferred stock and warrants worth a little under $21 billion, or about a 20% premium to actual value. So a $4 billion day-one subsidy: cheap capital! JPMorgan, on the other hand, gave Treasury back stuff worth about 94 cents on the dollar, making it just a 6% subsidy.
- Ultimately, Treasury got paid back. What was the value of what it got paid back, as a percentage of what it invested? So, for instance, Treasury gave Goldman $10 billion and got back $11.4 billion ($11.1 billion accounting for time value), making an 11% profit.6
So, I dunno. This certainly shows that the government was pretty undiscerning in its investment terms. But that's of course the point. A peril of this sort of approach is that it assumes you're thinking like a market-oriented investor, and Treasury wasn't. Part of the purpose of TARP was to paper over the difference between the banks, so the fact that Treasury vastly overpaid for its Morgan Stanley investment while getting roughly fair value for its JPMorgan investment was unavoidable.
And the shape of the chart suggests, a little bit, that it worked, no? Treasury invested in big banks with very different prospects at very different premiums to market value, with Morgan Stanley's cash coming very cheap and Wells Fargo's and JPMorgan's coming relatively meh. And yet it got similar, or similar-ish, rates of return on all those investments: TARP had not only the theory but also the effect of leveling the banks. In fact, the banks that were more heavily subsidized ex ante - MS, GS, and Citi - paid back the most in the end, exactly because the government's stabilization had a bigger effect on their stock prices and so made the government's equity more valuable.
That doesn't really answer Bair's question, though. I particularly like her theory that all the other banks got cheap or cheap-ish capital just because Citi needed a bailout. Given that the big winners of TARP, in terms of cheap fundraising, were Goldman and Morgan Stanley, this reminds me a little of the theory that AIG got its quasi-bailout as a way to funnel money to Goldman Sachs as its derivative counterparty. Maybe the best business model is to just make sure you're in the room with someone with a terrible business model.
1.Though, also, some pretty good catty characterizations of bank CEOs! Ken Lewis "was viewed somewhat as a country bumpkin by the CEOs of the big New York banks, and not completely without justification." After wrapping up Lewis, Bair goes on "Other CEOs were smarter. The smartest was Jamie Dimon, the CEO of J.P. Morgan Chase, who stood at the center of the table, talking with Lloyd Blankfein, the head of Goldman Sachs, and John Mack, the CEO of Morgan Stanley. Dimon was a towering figure in height as well as leadership ability. ... Blankfein and Mack listened attentively to whatever it was Dimon was saying." There's sort of a Jamie and the two dwarfs vibe here, isn't there?
2.You could focus elsewhere. There are a lot of depictions of Sheila Bair being very FDIC-focused and protective of her deposit-insurance funds; if that's the case, why exactly is she objecting to more robustly capitalizing the banks out of Treasury's pocket? I feel like "capital levels seemed to be adequate ... to bumble through ..." is kind of weak praise from the FDIC?
3.Obviously things like cheap FDIC guarantees for all bank debt are cheap. Just talking the capital purchase program here; presumably so was Pandit when he said "hey this is cheap."
4.Like, quite simple-minded. For the "estimated value at TARP time," this model (1) values the preferred as a 5% stream of payments for five years, followed by a 9% perpetuity, all discounted at a random roughing-out of the bank's preferred discount rate based on senior CDS on 10/13/08 and the rough spread between a JPMorgan pref and JPMorgan CDS, and (2) values the warrants as a Black-Scholes option with inputs including 3-month at-the-money option-implied vol haircut by 60% for no reason other than that it looked plausible. Those right there are bad assumptions, but it gets worse; I've done things like ignore the fact that half the warrants could be cancelled on a qualifying equity offering, or that the pref was callable. All of this tends I think to overstate the initial value - i.e. if my math shows that Treasury paid $1 for something worth 90 cents, it was really probably worth something more like 80 or 85 cents - but in, like, a plausibly consistent way. For the "final value," I've just taken the actual cash flows (again from ProPublica's Bailout Tracker) and discounted them at the 10-year treasury rate, just because. I've combined Bank of America and Merrill Lynch because, y'know, reality did.
5.As Lloyd's does.
6.Here is another way of putting it, showing the value (in cents on the dollar) of the stuff that Treasury paid a dollar for: