Back in September I noted that there were some rumblings at the Fed (notably from St. Louis Fed president James Bullard) about cutting the interest rate the central bank pays on excess reserves (IOER) to zero or perhaps even below zero in order to incentivize bank lending. Some, including San Francisco Fed chief John Williams, have argued that paying interest on excess reserves has discouraged banks from lending and in the process depressed the M2 multiplier.
While this has the desirable side effect of keeping inflation expectations somewhat anchored, it clearly has a deleterious effect on the availability of credit. In other words, banks simply aren't lending as the deposit to loan gap grew to an astounding $2 trillion in December. Apparently, the risk/reward calculus seems to favor collecting a riskless .25% from the Fed rather than chance a slightly greater return by lending to borrowers with less-than exemplary credit histories -- for the time being we will leave aside the issue of what banks actually do with deposits.
In the article cited above I argued that cutting the IOER rate to zero (or to negative 25 or negative 50 basis points) would spell the end for money market funds in terms of their ability to provide any kind of positive return whatsoever for their investors. Money market funds are of course part of the shadow banking system and one way they generate returns is by loaning money to banks in reverse repo transactions; that is, banks pledge collateral and money market funds loan money to banks against that collateral in exchange for a fee.
As long as the interest rate on the loan from the money market fund is less than the IOER rate, banks can fund assets in repo, park the funds at the Fed for .25%, and keep the difference between the IOER rate and the rate the money market funds charge for the loan. If the IOER rate falls to zero or below, that arbitrage opportunity disappears along with banks' incentive to fund assets in repo. This eliminates a critical source of yield for money market funds. Here's the Financial Times on the subject:
In a way, IOER represents a source of "supply" for money market funds because of the incentive for banks to play this arbitrage (between their funding rates they pay to MMFs and the IOER)...Lowering or eliminating IOER therefore creates a particular problem for money market funds, as many MMFs depend on lending to financial institutions in repo markets, through reverse repos, and in the market for commercial paper,
The IOER rate has of course remained at .25% and, ironically, the worry now is that as the Fed begins to hike rates, the IOER rate will rise and eventually the interest the Fed will owe banks on their excess reserves will exceed the fixed coupon payments the Fed earns on its SOMA portfolio thus technically bankrupting Bernanke and friends.
In any event, don't expect money market funds to provide any semblance of yield in the foreseeable future. In fact, the expiration of Operation Twist along with traders' anticipation of a bursting of the Treasury bubble represent a unique threat to money market funds and the explanation of that threat should serve as a nice exercise in dot-connecting for those interested in understanding how the system actually works.
Consider first the idea of "collateral scarcity." I have explored this concept in detail in previous articles and it has figured prominently in many a research note of late. There are several contributing factors to the scarcity of high quality collateral, the most obvious of which are central bank asset purchases, regulatory actions (Basel III, the centralized clearing of derivatives, etc.), and a reduction in the universe of AAA issuers. The following graph illustrates the latter point by showing the reduction in AAA rated government debt as a percentage of total sovereign debt and while foreign government bonds aren't necessarily germane to the current discussion, the idea of a shrinking universe of AAA collateral has clear implications for secured lending going forward:
Source: Gluskin Sheff
In the U.S., a distinction has been drawn between a generalized collateral "shortage" and the concept of collateral "scarcity." This distinction is deemed necessary because, as JPMorgan pointed out last fall, when the Fed buys Treasury bonds from primary dealers, it replaces one form of collateral (government debt) with another (CASH) so it might not be entirely appropriate to characterize the resulting situation as a "shortage." It would be proper however, to say that one form of collateral (Treasury bonds) is becoming more scarce compared to other forms of collateral. Here's JPMorgan:
QE is a good example of the distinction between collateral scarcity and collateral shortage. QE does not create a shortage, as the overall supply of collateral is unchanged, but it does create scarcity by making one form of collateral more expensive relative to another.
Now consider that in conjunction with the factors mentioned above (previous Fed purchases, a stricter regulatory climate, and a shrinking universe of quality collateral in general), the end of Operation Twist meant that the Fed once again became a net purchaser of U.S. government debt and the supply of short-dated Treasuries which were hitting the Street each month during Twist evaporated. This has effectively increased the severity of the collateral shortage (or "scarcity" or whatever you choose to call it) and, as the Financial Times noted last month, the scenario is now a classic case of "more money chasing less collateral, and that means lower collateralised rates."
More simply, if you are a dealer and you have what everyone wants (in this case Treasury bonds), you're not going to be willing to pay a high interest rate to pledge it as collateral for cash. The more in demand your collateral becomes, the less willing you are going to be to pay any amount of interest to loan it out, even if you get cash in return. In fact, your collateral may become so scarce that you would only exchange it for cash at par (that is, with no interest paid).
In light of these considerations, it makes sense that since the Fed announced in December that it would conduct unlimited, unsterilized purchases of Treasuries (thus sucking collateral out of the system), the average rate for borrowing and lending Treasury bonds has collapsed as the following chart from the DTCC shows:
(click to enlarge)Source: DTCC
This leads us to the idea of negative repo rates. Put simply, there may indeed come a special situation wherein the collateral you have is so scarce that not only will you not pay an interest rate to pledge it for cash, you will actually require the counterparty with the cash to pay you an interest rate for the privilege of borrowing your collateral. Last month for instance, the overnight repo rate for the on-the-run 5, 7, and 10-year Treasury bonds turned negative.
First, consider what this means. In a repurchase agreement, the general assumption is that the counterparty lending the cash will receive interest when the borrowing counterparty repurchases the pledged collateral. That is, the transaction is really just a collateralized loan: one party pledges securities in exchange for a cash loan and upon the repurchase of those securities, repays the lender the original cash loan plus an interest rate. When the lender of the cash receives an interest rate for his or her trouble, the repo rate is positive.
Now, consider what a trader may want to do as the market begins to chatter about a bond bubble and as yields on Treasury bonds begin to creep up: the temptation might well be to short Treasury bonds. Dealers short Treasuries the same way one might short a stock; that is, the dealer sells the bond, borrows that bond to make delivery to the buyer, then hopes to buy it back later at a lower price and pocket the spread. As Bloomberg noted, this type of trade is often executed before Treasury auctions:
Many times traders short, or sell securities they've borrowed in the repo market, before a Treasury sale to profit if prices of the securities fall after the auction.
Clearly, if one has sold a specific bond that one doesn't own, just any old bond won't do for delivery to the buyer. The New York Fed explains:
..."special collateral" [repo] is a repurchase agreement in which the lender of funds designates a particular security as the only acceptable collateral. Dealers and others lend money on special collateral repos in order to borrow specific securities needed to deliver against short sales. A short sale is a sale of securities that the seller does not own and that it has to borrow to make delivery. Dealers sell Treasury securities short in the expectation that prices will be lower in the future. (emphasis mine)
Knowing this, it is easy to imagine the type of desperate situation which might quickly materialize in the event some entity with unlimited purchasing power starts buying all the specific notes you need to deliver on your shorts. This is precisely what began taking place last month, especially in the case of 7-year Treasury bonds. Unable to locate the notes to cover short positions, traders were forced to borrow the paper from the only player left in town willing and able to lend it -- the Fed.
With a 7-year auction looming on January 30, traders borrowed $15.75 billion in 7-year bonds from the Fed on January 28 presumably to cover shorts ahead of the auction. In a situation such as that, one might expect that the party who needs the notes to cover his shorts would be willing to pay up to get the specific notes needed. Indeed, the repo rate on the 7-year was negative .1%.
Note that now, the roles are effectively reversed in the repo transaction. Instead of the counterparty with the cash receiving a positive interest rate from the counterparty looking to pledge securities for a loan, it is now the counterparty with the securities that is receiving the interest rate from the counterparty with the cash. So the traders are paying a fee to borrow the Treasury bonds as opposed to the counterparty with the Treasury bonds paying a fee to borrow cash. This is the definition of a negative repo rate and when it occurs on a specific note (the 7-year Treasury bond in our example) it is called the "special rate"
Quite a bit more can be said about this dynamic, but for now, consider what this means for yield-strapped money market fund managers. With repo rates depressed (the average Treasury repo rate hit a low of just 8 basis points on January 22) the business model is broken. Recall that money market funds lend cash to dealers against collateral like Treasury bonds. On a good day, those loans are now yielding half of what they were last September.
In addition to paltry yields for investors, this could lead to an increase in systemic risk as money market funds reach for yield in places they shouldn't (especially considering their perceived safety). Here's Barclays' Joseph Abate speaking about the issue two years ago:
The longer repo rates stay [low], the more likely money funds are to reallocate toward higher-yielding assets.
Expect repo rates to remain pressured as the Fed continues to pull $45 billion in Treasuries and $40 billion in MBS out of the market each month. This will put further pressure on money market funds and may meaningfully increase the level of systemic risk residing in the shadow banking sector by forcing money market fund managers to take on riskier assets which may not be easily liquidated in the event of a run.