On January 7, an article appeared on the Huffington Post entitled "Too-Big-To-Fail Banks Gamble With Bernanke Bucks." The piece was written by Bill Frezza, a venture capitalist whose columns also appear in Forbes. In the article, Frezza notes that since September 2008, commercial loans have declined by $120 billion while deposits have risen to a record $10.6 trillion as of the end of 2012. The result is a $2 trillion gap between deposits and loans:
Frezza is correct to point out the discrepancy. According to a recent piece in the Wall Street Journal, the loan-to-deposit ratio is now just 72% - conventional wisdom has it that banks would prefer that ratio to be as close to 100% as possible. Here is a chart which shows the declining loan-to-deposit ratio over the course of the post-crisis years:
Source: Wall Street Journal
After pointing out the fact that banks have quite a lot of excess cash, Frezza goes on to make a highly contentious claim. In short, he says that banks are using the cash to gamble with and that because quantitative easing contributes to the buildup of excess reserves, the Fed is providing banks with the dry powder they need to place speculative bets:
If you think [the excess cash cushion] makes the banks less vulnerable to shock, think again. Much of this balance sheet cash has been hypothecated in the repo market, laundered through the off-the-books shadow banking system. This allows the proprietary trading desks at these "banks" to use that cash as collateral to take out loans to gamble with. In a process called hyper-hypothecation this pledged collateral gets pyramided, creating a ticking time bomb ready to go kablooey when the next panic comes around.
The good news for Frezza is that he is entirely correct. The bad news (for him) is that he, like many others before him no doubt, has made a critical error: he has written an article based entirely on an earlier piece from ZeroHedge without fully understanding ZeroHedge's analysis. This became abundantly clear when, just a few hours after the piece was published, Frezza found himself on CNBC defending the article against the network's Pulitzer Prize-winning senior economics reporter Steve Liesman.
Frezza begins the exchange by saying that "according to ZeroHedge" the excess reserves are "being hypothecated and used to gamble in the derivatives market." Liesman, perhaps realizing that citing ZeroHedge does not necessarily mean one fully understands what ZeroHedge has said, quickly calls Frezza's bluff:
I just think it's factually inaccurate to say that the excess reserves are being hypothecated... and there is no evidence in the article that Bill wrote to substantiate that. The excess reserves are a byproduct of quantitative easing...they can't really do anything with it, and even if they went out and spent those excess reserves it would come back to them in another form. (emphasis mine)
Frezza seems to have only a vague notion of how to respond. He says that the banks "can't [do anything] with those dollars," implying that there are other dollars which the banks use to gamble. Frezza does not however, explain where the gambling dollars come from. Ultimately, Liesman corners him once again: "Where's the evidence of this?"
At this point, the conversation deteriorates quickly into an entirely asinine argument about loans and underwriting standards and ends abruptly with an unexplained reference to the now ubiquitous London Whale. The bottom line is that it certainly did not appear that anyone involved in the discussion fully understood the issue being debated.
In reality, there is evidence to support the notion that banks are using the excess cash on their books to gamble with. Consider the following chart which shows primary dealer holdings of U.S. Treasury bonds:
As you can see, as of late December, primary dealers' holdings of U.S. government debt were at an all-time high of around $140 billion. On a broader level, Bloomberg published an article in August which carried the title "Banks Use $1.77 Trillion To Double Down Treasury Purchases," in which it cited excess deposits as the main reason why banks' holdings of U.S. Treasury and government agency debt were hitting record highs.
There is no question then, that banks are buying U.S. government debt in record quantities. The question is what they do with it once they acquire it. The answer lies in the repo market. Recall that via repurchase agreements, high-quality, liquid collateral like Treasury bonds can be pledged for cash loans. This process can work as a kind of legalized money laundering service for banks that have excess reserve balances. That is, rather than simply take their excess reserves and use them to make big bets in the credit and equity markets (which would be too transparent even for U.S. regulators), they use the cash to purchase Treasury bonds. The Treasury bonds are then pledged as collateral for cash in repo transactions. The best part about the whole deal is that the Treasury bonds stay on the banks' books after they are pledged. From ZeroHedge:
The biggest benefit of Repo financing is that the bank can still hold the original pledged security on its books for Fed "supervision" purposes, even as it obtains fungible cash equivalents via repo.
This is critical: once the transaction is complete, the bank has Treasury bonds on its books and clean cash in hand to do with as it pleases even though the Treasury bonds themselves are pledged to another party. It is clear then that in theory, the more Treasury bonds one has, the more incentive there is for those assets to be pledged for cash given that the Treasury bonds stay on the books regardless.
Finding evidence of this is quite simple, especially if you've read the classic 2008 note from Citi's Matt King entitled "Are The Brokers Broken?" Here is an excerpt from that note:
Each broker's balance sheet has a line item on the asset side showing "Financial Instruments Owned". Either in brackets, or as a sub-item, they then report the proportion of this "pledged to counterparties," i.e. funded on repo... A further portion of the financial instruments owned - which is in many cases substantial - is reported in the 10-Q footnotes of "collateral pledged to counterparties which cannot be repledged."
In order then, for an investor to understand just how much of a firm's assets have been pledged as collateral for cash (i.e. funded in repo), that investor would need to know two things: 1) the meaning of the caveat to the "financial instruments owned" line item which shows how much of those instruments are pledged to counterparties, and 2) that buried in the Notes to Consolidated Financial Statements, there is another figure which shows how much of the firm's financial instruments are pledged to institutions which themselves cannot reuse the collateral.
To clarify, the disclosure on the balance sheet which reveals what portion of the firm's assets are pledged as collateral only accounts for collateral that has been pledged to parties which can themselves repledge the collateral. The portion of the firm's financial instruments that have been pledged to institutions which cannot, for whatever reason, reuse the collateral is only disclosed in the Notes section of firms' 10-Qs.
When taken together, these figures can be divided by the total amount of the firm's financial instruments to reveal what percentage of those instruments are funded in repo. Bringing it all together, the excess cash pumped into the system via quantitative easing combined with customer deposits is clearly leading to record purchases of Treasury bonds and government sponsored debt by banks. Theoretically, these securities could be pledged as collateral for cash in repo transactions, providing the banks with clean cash to use for bets in equity and credit markets. Meanwhile, the pledged assets remain on the balance sheet while investors not willing to dig into the 10-Q notes remain uninformed about exactly what percentage of those assets are funded in repo.
If this is taking place, one would expect to find that the percentage of banks' financial instruments funded in repo would be greater now than before deposits began to outpace loans in 2008. This would be the 'evidence' Liesman was looking for.
Let's take Morgan Stanley as a test case. As of May 31, 2008, 50% of Morgan Stanley's financial instruments were pledged as collateral in repo transactions:
Source: MS 10Q, Matt King
A look at Morgan Stanley's latest 10-Q shows that as of September 30, 2012, the firm had around $133.8 billion in financial instruments pledged to counterparties who could repledge that collateral and around $251.3 billion in total financial instruments:
Source: Morgan Stanley 10-Q
On page 47 of the 10-Q, there is a table which shows the amount of the firm's financial instruments pledged to institutions which cannot repledge the collateral. The figure is $51.7 billion:
Source: Morgan Stanley 10-Q
Taken together, this means that $185.5 billion of the firm's $251.31 billion of financial instruments are funded in repo. That's 74% compared to only 50% in 2008.
Next, consider that as of May 31, 2008, 39% of Goldman Sachs' financial instruments were pledged (second column in the following table):
Source: GS 10-Q, Matt King
Goldman's latest 10-Q shows that, at least as far as the balance sheet is concerned, only $66.75 billion of the firm's total $415.3 billion in financial instruments are pledged as collateral. However, on page 54 of the 10-Q, one finds the following table which, in addition to showing the aforementioned $66.75 in collateral pledged to counterparties who can repledge the assets, shows the amount of the firm's financial instruments which are pledged to counterparties who cannot repledge the collateral. The figure is $118.63 billion with an additional $2.6 billion pledged in other economically similar transactions:
Source: Goldman Sachs 10-Q
Taken together, this means that around $188 billion of the firm's total $415.3 billion in financial instruments are pledged as collateral in repo transactions. That is 45.3% compared to 39% in 2008.
These are but two examples. In order for this analysis to be complete, a study would need to be conducted across the entire industry to determine if this trend holds. However, what this does show is that there is indeed evidence that since 2008, at least two systemically important financial institutions have meaningfully increased the percentage of their financial instruments funded in repo.
Why It Matters
Understand that the above suggests that financial institutions may be actively seeking to get cash in exchange for a greater portion of their assets than they were before the Fed began pumping money into the system. This makes sense: you want to do something with your excess cash, not just sit on it, so you buy liquid assets such as Treasury bonds and agency debt (which is more desirable in the repo market), pledge it as collateral in repo transactions, then come away with the assets still on the books and a pocket full of clean cash that you can legitimately say didn't come from the Fed or from depositors. To top it all off, it is difficult even for those with half a clue to figure this out because thanks to FAS 140, "collateral pledged to counterparties which can be repledged to other counterparties" is the only type of collateral that need be mentioned as "pledged" on the balance sheet, meaning that investors who do not read the Notes will not understand the true extent of the firm's repo funding.
This money from repos is the money firms can use to gamble with. What Frezza and others are trying to say (even if they don't know it) is that it is no coincidence that stocks have rallied as the Fed has pumped money into the coffers of the primary dealers while ICI data shows retail investors have pulled nearly a half trillion from U.S. equity funds over the same period. It is the banks that are propping stocks.
The above suggests two simple arguments for why investors should steer clear of U.S. equities in 2013. First, as Matt King pointed out four-and-a-half years ago, secured repos are hazardous for the borrower:
...given potential exposure to some counterparties running into the billions of dollars, and given the psychological fear of censure by senior management and shareholders, the temptation for lenders is often to pull back whenever a borrower begins to look as though they are in trouble. The withdrawal might consist of an increase in haircuts, a refusal to repo more illiquid types of collateral, a reduction in maturity, or a cessation of trading with a counterparty altogether.
Investors should consider that with the S&P 500 at a five-year high, any small hiccup in the repo market in terms of its collective willingness to continue to fund a large firm's equity bets via reverse repos could lead to forced selling or, at the very least, a cessation of ravenous buying. That's the nuanced argument.
The simple argument goes something like this: if the Fed and corporations' deposits are funding the large scale purchase of government debt by banks who are in turn pledging that debt for cash, they then use to boost risk assets like equities (which the data presented above suggests), then even the mere suggestion that quantitative easing may be scaled-back or that corporations may stop depositing money and start making capital expenditures or raising wages, will leave banks without the dry powder they need to prop-up equities. In short, it is never wise to buy at the top of an artificially inflated market and that is precisely what this market has become. I believe the above has shed some light on an important debate and has given investors reason to beware of U.S. stocks (SPY) (QQQ) in 2013.