In two previous articles (here and here) I suggested that banks use excess deposits and excess cash in general to make risky bets on a variety of securities including, but certainly not limited to, equities. I asserted that the primary mechanism by which they can do this is by running the cash through repo transactions. This process has the obvious advantage of allowing the banks to purchase high quality securities like Treasuries and agency debt (which U.S. financial institutions have a record amount of) and pledge it as collateral for cash loans while keeping the pledged securities on the books for regulatory oversight (and investor scrutinization) purposes.
These articles were generally well-received but encountered the usual charges of "conspiracy theory." Additionally, it was suggested that banks simply cannot buy stocks with excess deposits -- the implication is that if deposits are invested, they are only invested in "safe" assets. I attempted to show how institutions with market making lines hold inventories of securities as part and parcel of their market making businesses and that given the inherent difficulty in separating market making from proprietary trading and given the willingness of the Fed to accept billions in stocks as collateral in emergencies, it is a good bet that some large banks will trade stocks for their own accounts with excess cash.
The rest of the week proved rather fortuitous for proving my point. As I noted in an article published early Friday, Goldman Sachs' prop trading desk and market making business pulled in nearly $5 billion in revenue between them, helping the firm earn $5.60 per share while Wall Street's other banks struggled on a relative basis. However, the most clear-cut proof that banks' trade stocks (and virtually everything else) with excess deposits came in the form of a just-released report from JPMorgan's (JPM) CIO Task Force, the group charged with investigating the losses at the firm's London-based Chief Investment Office.
An Inconvenient Truth
The purpose of the report was to examine, at length, what exactly happened at CIO and how things could have gone so terribly wrong during the first quarter of 2012. For the purposes of this discussion however, the important takeaways can be found in the firm's own description of the CIO's mandate. Consider the following passage:
JPMorgan's businesses take in more in deposits that they make in loans and, as a result, the Firm has excess cash that must be invested to meet future liquidity needs and provide a reasonable return. The primary reposnsibility of CIO, working with JPMorgan's Treasury, is to manage this excess cash. CIO invests the bulk of JPMorgan's excess cash in high credit quality, fixed income securities, such as municipal bonds, whole loans, and asset-backed securities, mortgage backed securities, corporate securities, sovereign securities, and collateralized loan obligations.
That passage in unequivocal -- it is as unambiguous as it could possibly be. JPMorgan invests excess deposits in a variety of assets for its own account and as the above clearly indicates, there isn't much they won't invest those deposits in. Sure, the first things mentioned are "high quality fixed income securities," but by the end of the list, deposits are being invested in corporate securities and CLOs. This isn't necessarily bad, but it is important to note that the idea that desposits are invested only in Treasury bonds, agencies, or derivatives related to such "risk free" securities is patently false.
What the above chart shows is that nearly half of the $2 trillion deposit to loan gap in America comes from BAC, JPM, and WFC alone. Based on what we now know about JPMorgan's CIO unit, one can imagine where that $858 billion in excess deposits is being invested.
Disappearing Repo Risk
Another issue not mentioned in my previous two articles on banks and the repo market is the idea that in some cases, nearly half of what is going on in terms of the lending and borrowing of cash and securities is literally disappeared from the balance sheet. The following example owes much to Citi's Matt King.
Imagine you are a hedge fund and you want to buy $50 million in Stock X on margin. The bank which facilitates this transaction books the following entries: "Accounts Receivable +$50 million," "Cash -$50 million," and $50 million in Stock X goes in your account with the bank, as well as a concurrent obligation of course, to repay to $50 million loan. Now that you, the hedge fund, have your $50 million worth of Stock X, you decide you now want to short Stock Y. You borrow Stock Y from the bank, sell it, and repay the $50 million loan you took out for the purchase of Stock X with the proceeds. The journal entries for the bank offset and the accounts receivable and cash accounts are now unchanged (you, the hedge fund, have paid back your loan).
Consider though, how the bank obtained the $50 million worth of Stock Y to loan you for the short sale. It can, if it chooses, pledge the $50 million of Stock X that you bought initially as collateral for the $50 million of Stock Y that you shorted. Because securities-for-securities repos (or, "borrowed-versus-pledged" transactions) are not required to show on the balance sheet, this whole transaction simply zeros-out.
If this seems implausible, or too bad to be true, it can be proven quite easily. Take at look, for example, at page 42 of Morgan Stanley's third quarter 10Q. It shows around $273 billion in "Securities Purchased Under Agreements To Resell" and "Securities Borrowed" (the sum of those two items). Now have a look at the footnotes on page 47. The following passage appears:
At September 30, 2012...the fair value of financial instruments received as collateral where the company is permitted to sell or repledge the securities was $586 billion...and the fair value of the portion that had been sold or repledged was $438 billion.
Note the vast discrepancy here. The balance sheet shows only $273 billion of securities purchased under agreements to resell (reverse repos) and securities borrowed, yet somehow, the firm has actually repledged nearly twice that amount. This difference is likely attributable to transactions like the once described above and while, as Matt King notes, this is just client financing, it still equates to leverage and the pyramiding of repoed collateral:
The net effect, then, is for brokers to build up billion of dollars in reverse repo or stock borrowing transactions, on behalf of clients, of which only a fraction is recorded on the balance sheet... This looks like the explanation behind the footnotes...But why should investors care about all this? First, the pledging of collateral to brokers in such large sizes -- and the fingibility of pledged collateral with their own positions -- significantly improves their own ability to take short positions, make markets and provide liquidity in other markets generally. Second, these numbers imply a gross dependence on repo financing far larger than the on balance sheet numbers suggest.
This is entirely consistent (the fungibility argument) with the idea that the more excess cash available to firms with market making businesses, the more prone they will be to engage in these types of transactions, to pyramid collateral in repo, and to depend for repo financing for their very livelihood.
The point here (and it is a very important one) is that the JPMorgan CIO debacle exposed, for all to see, how big banks take risks with excess deposits. The discussion about the repo market shows how these transactions are facilitated, booked, and in some cases, disappeared altogether. This practice is very real and it goes on each and every day and indeed, it is partly responsible for the rally in equity and credit markets and it is facilitated by the Fed's injections of excess reserves via quantitative easing (remember the fungibility issue).
This market is inflated and it is held up by a flimsy scaffolding that requires Goldman-like expertise to manage and navigate effectively. It isn't to be trusted generally, and I reiterate my recommendation to focus on capital preservation going forward and to avoid artificially inflated U.S. equities (SPY) (QQQ). If this market were not sitting at 5-year highs, this piece and its predecessors would serve only as precautionary tales. As it stands currently however, this should serve as a reminder that what goes up, must come down, especially if the reason for the ascent is something other than actual investor interest in the market.