At quarter end, the demand for the Federal Reserve's Overnight Reverse Repo Program exceeded the $300 billion cap that was established by the Fed and went into effect September 22. The RRP hit record demand of $407 billion at the end of the third quarter, surpassing the previous record of $339 billion at the end of the second quarter (prior to the cap). As a result, the repo rate fell to zero. "A Federal Reserve tool designed to set a lower boundary on short-term rates didn't work that way Tuesday, reinforcing concerns that new limits could thwart its effectiveness. The Fed plans to use the tool, known as overnight reverse repurchase agreements, when it starts raising rates, likely sometime next year," wrote The Wall Street Journal Wednesday in "In Significant Test, Fed Facility Fails to Defend Short-Term Rate Floor.

The piece explains, "Through these trades, the central bank takes in cash from money-market mutual funds and other nonbank financial institutions in exchange for one-day loans of Treasury securities and pays them interest in return. Before Tuesday, the Fed paid a rate of 0.05%. Fed officials say they expect the so-called reverse repo rate will serve as a floor under short-term interest rates when they start lifting borrowing costs. But on Tuesday, when heavy demand for the facility exceeded new caps on its use, the repo rate fell to zero." (Note: Crane Data will publish its Sept. 30 Money Fund Portfolio Holdings next week, which will show how much of the $300 billion was purchased by money market funds.)

According to a statement from Fitch Ratings, "[S]ome participants in the Fed's auction on 9/30/14 were pressed to invest their cash, and bid as low as (-0.20%) to ensure they receive a sufficient allocation at the auction. Regular participants in the Fed's RRP, including money market funds, were caught somewhat off-guard by the imposition of the cap, as the program served as a reliable last resort source of supply for the short-term markets. Since the Fed's announcement of the cap on 9/17/14, money funds have been looking for alternative sources of supply, like Treasuries, to stay invested at quarter-end."

Wells Fargo Strategist Garret Sloan offered a timeline of events on Wednesday. "Minutes after the 8:30am RRP close, the Fed published its results, showing that just over $407 billion in bids had been submitted from 102 bidders. Because total bids exceeded the $300 billion cap, bid prices came into play and some investors did not receive allocations. The stop-out rate, or the rate at which all investors were either completely or partially filled, was 0 bps. This means that any investor bidding below 0 bps was fully allocated at auction, and those investors that bid 0 bps were allocated on a pro-rata basis. Of the 102 counterparties that submitted bids, 81 investors received allocations," he writes.

Sloan explains, "Looking at the composition of the 140 Fed counterparties, it should be expected that the six GSEs and the 18 banks would not have chosen to bid below 0 basis points given the alternatives that they have at their disposal (i.e. Federal Reserve deposits and IOER). The 22 primary dealers may have had an interest in bidding if they thought they could re-hypothecate collateral out to investors at more attractive rates later in the day. However, the idea that the primary dealers would want to grow their balance sheets at quarter-end and leverage their positions when it is precisely the opposite behavior that is precipitating the quarter-end squeeze makes that idea unlikely. No, it is the money market funds that have driven the RRP thus far, and money market funds that drove the market yesterday."

He adds, "The average bid per counterparty equated to $3.99 billion, well below the $30 billion maximum bid per counterparty. The range of bid yields started at the high Fed offered rate of 5 basis points and got as expensive as -20 basis points. By 10am, the repo market had backed up from those rich levels, and by the end of the day the average Treasury repo rate was a miniscule 1.8 basis points. However, the fact that the average rate remained positive might be considered a small victory. Moreover, the fact that the market average remained positive for the day could be ammunition for the Fed to only make small tweaks to the RRP, rather than increase its size dramatically."

Finally, Sloan says, "The flip side to this argument is that from the time that investors first heard about the cap they were making plans to avoid quarter-end pressure, buying T-bills and other short-term products at extremely expensive levels. So, what are the leakage points in the Fed's monetary policy program? In other words, what would cause money market rates to remain sticky once the Fed increases IOER to 50 bps and the RRP to 25 bps? The top-end of the range does not seem to be an issue. Demand from banks at the higher IOER would likely remain stable or grow, keeping that cash stockpile from imposing its will on lower-yielding products. The question is what happens at the lower-end of the range now that the RRP is no longer designed as a fixed-rate full-allotment facility? [T]here is simply no question that the Fed will adjust the size of the RRP to address this situation, not because they wish to support money market funds per se, but because the FOMC will find it necessary to do so from an efficient monetary policy perspective."

Finally, in his weekly "Short-Duration Markets" report, JP Morgan money market strategist Alex Roever predicted the outcome. "Still, given the sharp supply/demand imbalance that usually takes place during these highly technical times, we suspect the demand for ON RRP could still exceed $300bn, initiating a Dutch auction where the stop-out rate could clear below 5bp. The good news is that post quarter-end some of that pressure to find liquidity will quickly pass. The bad news is that these changes could remain in place. At the heart of the issue is how effective these changes to the ON RRP facility will be in lifting rates when it comes time to normalize monetary policy."

Email This Article




Use a comma or a semicolon to separate

captcha image