You Misplace 5 Or 6 Billion Dollars And All Of A Sudden People Stop Trusting You To Keep Track Of Your MoneyBy Matt Levine
When JPMorgan’s whale drowned a lot of people asked “where were the regulators?” and that was a silly question, because the people with the most incentive and ability to keep the whale afloat were, in descending order, (1) the whale, (2) the whale’s bosses, (3) the whale’s bosses bosses, (4) the regulators, and (5) the people asking “where were the regulators?,” so if categories 1-3 missed the problem then there’s no reason to get all mad at category 4. “If X’s could do Y they wouldn’t be X’s” is an important tool to keep in your mental toolkit, and if regulators could distinguish good from bad trades they’d be at least risk managers and probably, like, Warren Buffett.
What regulators are supposed to do, ideally, is not pick trades but rather set up systems to prevent bad trades from having ruinous systemic effects, and a major method of doing so is capital regulation. JPMorgan lost $5.8 billion on whale-failing, and if you or I lost $5.8 billion we would probably be scaling back our vacation plans, but Jamie Dimon isn’t because JPMorgan had lots and lots more money where that came from. Capital!, in both senses of that exclamation.
This is a pleasing use of regulatory intelligence:
Two regulators—the Office of the Comptroller of the Currency and the Federal Reserve Bank of New York—told J.P. Morgan Chase Wednesday that the firm’s capital cushions should be amended for the first half of the year, the bank said in a filing with the U.S. Securities and Exchange Commission. The changes were requested because of concerns about models used to measure risk taken by the Chief Investment Office and other areas of the bank. The firm’s so-called Tier 1 common capital fell to 9.9% of risk-weighted assets as of June 30, from 10.3% previously reported.
Now capital is, for our purposes, the amount of stuff you have minus the amount of money you owe. (Our purposes are limited.) And in fact JPMorgan’s capital has not changed since its July 27 earnings announcement, when it got its house in order total-whale-loss-wise and nailed down exactly how much stuff it had.* Rather, its capital ratios have changed. Capital ratios are capital divided by your risk-weighted assets, and JPMorgan’s risk-weighted assets swelled gloriously in the last two weeks. Back on July 27, JPMorgan told us that RWAs were $1.269 trillion as of June 30, 2012, and $1.235 trillion as of March 31, 2012. Now they’re telling us that RWAs are $1.319 trillion as of June 30 (a $50 billion increase) and $1.3 trillion as of March 31 (a $65 billion increase).** The same capital divided by higher RWAs = a lower capital ratio, down from 10.3% to 9.9% as of the end of Q2.
While having more assets is good, having more risk-weighted assets is bad, because it requires you to “hold more capital,” a horrible phrase that translates loosely into “have a lower return on equity and give less cash back to shareholders.” Which is what is happening here: due to its 40bps of vanished capital, JPMorgan will delay its long-dreamt-of share buyback from 4Q to early 2013 and be generally sad and remorseful.
But the OCC’s and Fed’s move is not mainly about punishing JPMorgan, or about the vague sense of “if capital is good more capital is better.” It’s about hedging JPMorgan’s now somewhat tarnished ability to keep track of its own risks. From JPMorgan’s 10-Q:
On August 8, 2012, the Firm was informed by the Office of the Comptroller of the Currency (“OCC”) and the Federal Reserve Bank of New York that they had determined that the Firm and JPMorgan Chase Bank, N.A. should amend their respective Basel I risk weighted assets (“RWA”) at both March 31, 2012 and June 30, 2012. The determination relates to an adjustment to the Firm’s regulatory capital ratios to reflect regulatory guidance regarding a limited number of market risk models used for certain positions held by the Firm during the first half of the year, including the CIO synthetic credit portfolio. The Firm believes that, as a result of portfolio management actions and enhancements it will be making to certain of its market risk models, these adjustments will be significantly reduced by the end of 2012.
If you are a little bank, your risk-weighted assets are basically computed by taking each asset’s value, multiplying it by some regulatory risk-weighting based on a rough perception of its risk (0% for Treasuries, 20% for munis and GSE bonds, 50% for mortgage loans, 100% for commercial loans, etc.), and adding up the result. But if you are a great big bank, your risk-weighted assets are calculated in large part based on models that you develop yourself. This actually makes a lot of sense: regulator-written rules are a blunt instrument and unlikely to give you appropriate credit for hedging and diversifying, whereas your internal models are built to do exactly that. If you buy $100 of X and $100 of Y and those two are perfectly inversely correlated, you don’t really have $200 of risk, you have $0 of risk, so why treat it like $200 for capital purposes?
This works great until your model tells you that you hedge a long credit exposure by being massively long … umm … basically the exact same credit exposure. Then you lose money,*** yes, but you also should lose some confidence in what your models are up to, shouldn’t you?
The Fed and the OCC lost confidence, anyway. It’s not clear exactly what changes they demanded from JPMorgan – that excerpt above is all we’ve got – though to the extent that the changes are concentrated in a “limited number of market risk models” they look rather drastic.**** Perhaps the regulators gave JPMorgan tips on how to make the models more conservative; conceivably they even told it to abandon some of its internal models and use the standardized, little-bank approach until the internal models redeemed themselves. (This is kind of what “as a result of portfolio management actions and enhancements it will be making to certain of its market risk models, these adjustments will be significantly reduced by the end of 2012″ sounds like.) As a result, RWAs went from 55.4% of total assets to 57.6% – in line, incidentally, with the average for North American banks, but well above the worldwide average for systemically important banks.
This strikes me as a notable regulatory success: the capital rules allow a lot of judgment and flexibility, and the regulators are now exercising their judgment to limit JPMorgan’s. When JPMorgan was the bank that could do no wrong, it was allowed to do its own math to decide how much capital to maintain; now that it’s done wrong, the regulators will be checking its work for a while – and erring on the side of more capital to cushion it from a resurfacing of the whale. The only black eye is that it took this long: though the Fed and OCC have been thoroughly informed about JPMorgan’s model failures and whale misery for months, this change in capital modelling was lobbed in hastily on August 8, two weeks after the earnings release and just a day before the 10-Q was going to be filed. But then, if the regulators could keep up with events in real-time, they wouldn’t be regulators.
* Actually it went up a teeny bit. As of the earnings release, total Tier 1 capital at June 30, 2012 was $148,399mm (and $155,352mm at March 31) and Tier 1 common was $130,095mm ($127,642mm at March 31). In the second quarter 10-Q and first quarter 10-Q/A filed yesterday, those numbers stayed the same except that total Tier 1 capital increased by $26mm to $148,425mm for reasons that are a not particularly interesting mystery to me.
*** Or make it. Or, as with the whale, make and then lose it. You don’t stay particularly “hedged” though.
**** For instance – JPMorgan has about $179bn of credit derivative notional outstanding, net of purchased/sold protection on identical underlyings (page 143 of the Q). The increase in RWAs, if concentrated only in synthetic credit, would be almost a third of the total possible exposure.