Collateral Shortage Update
After a small pause for April month-end, this year’s “collateral shortage” remains in full force. Each year there’s a season collateral shortage in April when the Treasury’s Cash Management Bills (CMBs) mature. This year, large CMBs matured on April 15, April 18 and April 25, but the collateral shortage did not begin until April 22, two days after the April 18 bills matured. The shortage began when General Collateral (GC) crossed below fed funds, trading at a negative spread. The “shortage” temporarily reversed itself on April 30, when the new 2 year, 5 year, and 7 year notes settled adding over $50 billion in net new securities back into the market.
It appears the same collateral shortage scenario is happening again in May. The Treasury is paying down about $80 billion in debt during the month, and again there’s a large net settlement at the end of month, which basically means the collateral shortage that began in April is still alive and well. Not that this is a major collateral shortage, but clearly GC rates are moving up and down based on Treasury issuance. Over the past week, each afternoon GC closed soft, with GC even trading slightly negative yesterday. Like in April, this trend with reverse itself at the end of the month. On May 30 there’s a net $49 billion and on May 31 there’s a net $35 billion settling in net new Treasury issuance. A total of $84 billion Treasurys hitting the market should bring the GC/FF spread higher with GC trading at least 5 basis points above fed funds again.
Details of the Treasury’s new Floating Rate Note (FRN) were announced with the quarterly refunding on May 1. The floating rate index choice had came down to either an overnight GC Repo rate or the auction rate of the weekly 13-week bill auction. The Treasury chose the weekly T-Bill auction rate. More details on the FRN: the final rules will be issued in coming months, the first issue will be a 2 year maturity with the size between $10 billion and $15 billion, and the first FRN auction is scheduled for the last quarter of 2013 or first quarter of 2014.
I’m still shocked that the Treasury chose 13 week bills rate and I believe it’s a mistake. A daily overnight Repo rate is a better floating rate which represents where U.S. Treasurys are funded in the market each day. Treasury Bills trade in their own world, based on the supply and demand for bills, which is not the same for other Treasurys. In addition, a floating rate based off a spread to a weekly auction rate just seems overcomplicated.
In the end, the spread where the FRN trades should be the spread between Treasury coupons and Treasury bills, adjusted for a couple factors. Suppose there was an FRN being auctioned with just 1 week left to maturity. The market would look at where 1 week GC traded and the 13 week bill auction rate, and that would be the spread. However, FRNs will trade a basis point cheap to regular Treasury collateral, just like TIPS. So now the price for the 1 week FRN rate is GC plus one basis point. Now, take that out 2 years, except there’s always a chance of a Flight-To-Quality (FTQ), supply distortions in the bill market, and quarter-end when bills trade at a premium. In that case, the FRN spread needs to price these risks and the market will want extra yield. I see the FRN based on Treasury bill auction rates trading relative cheap in the market, and certainly cheaper than an overnight Repo index.
At the end of this week on May 18 the U.S. government is theoretically supposed to run out of money and hit the debt ceiling when the “No Budget, No Pay Act of 2013″ expires. Technically, the Treasury is not authorized to borrow any additional funds after that date, but just like in the past, the Treasury is able to employ “extraordinary measures” to keep the government funded. Some might call them “accounting gimmicks,” but combined with strong tax revenues on June 15, the Fannie Mae payback of $59.4 billion at the end of June, and potentially strong September 15 tax revenue, the U.S. government is not expected to run out of funds until late October or early November.*
The current 10 year note is still going strong and trading at negative rates. 10 year notes are on a three month auction cycle with reopenings on the two intervening months. That means this current 10 year note originally settled on February 15. It was reopened in March, about doubling the size of the issue outstanding with the March 15 settlement. The final reopening settled on April 15, making the 10 year note a “triple issue,” after having been reopened twice. As a rule of thumb, 10 year notes trade the most special during the first month when it’s still a single issue. That makes sense, of course, based on pure supply and demand – there’s simply not much supply outstanding and if short-demand is significant, a 10 year note will trade very special, just like the current 10 year note did in March. After the first reopening settles, a typical 10 year might trade slightly special, and remember there are always exceptions under extreme circumstances. After the second reopening when it becomes a triple issue, a 10 year note rarely trades special.
Here’s the point: this 10 year note continues to trade special, even as a triple issue. That’s not typical and it means there’s a deep short-base in the 10 year note sector. But this market is going to get even more interesting after the WI 10 year note settles on Wednesday, May 15. Picture all the shorts required to make a triple issue trade special, then picture those shorts rolling into the new 10 year note, which is still a single issue. That’s a recipe for a pretty significant short-squeeze. Naturally not all of the shorts will roll forward, regardless, I believe there’s a high probability of the new 10 year note trading very special and fail during the first two weeks of June.
3.625% 5/15/2013 R.I.P.
The most special issue of all time matures tomorrow! And that’s even counting the Salomon 2 year note squeeze in 1991. Massive shorts and no supply in the market caused the 3.625% 5/13 to fail from the beginning of July 2003 through November 2003, it briefly cleaned up in December, and then failed again in January.
The events which led up to the massive squeeze are potentially relevant again today. The 3.625% 5/13 was issued in May 2003 and at the time the market was expecting a 50 basis point ease at the June FOMC meeting. The fed funds target rate was 1.00% and the market was expecting a new target rate of .50%, or at least .75%. The ease never occurred and the market took that as the turn in the easing cycle. It was. 10 year note yields which were trading at 3.15% in mid-June backed-up to 3.50% by the end of June and eventually hit 4.50% in mid-August.
As the bond market sold-off, traders quickly short-sold 10 year notes to hedge their long positions, especially mortgages. Many “retail” portfolios had bought the 3.625% 5/13 and now their positions were well underwater. With much of the issue in “hold to maturity” portfolios, those investors could not sell them or lend them back into the market. Short-demand was high and supply was low, the massive fail began.
Remember, back in 2003 it was before the Fail Charge and dealers still had the option of failing if they could not cover their shorts and they received no interest on their cash. After a month of fails, shockingly, the Fed chose not to reopening the issue at the August quarterly refunding.
As I discussed last week (http://scottskyrm.com/2013/05/negative-interest-rates-repo-rates/), the 3.625% 5/13 showed the market the true cost of fails. After it had failed for a period of time, dealers started to notice regulatory net capital charges, additional assets on their balance sheet, good clients were getting annoyed, and there was increasing credit exposure because you couldn’t call margin against fails. By September, the market needed a mechanism to price a delivery (clearing up a fail) and developed the Guaranteed Delivery (GD) Repo trade, allowing the 3.625% 5/13 to trade at negative rates. Unofficial negative rates in the Repo market were born from to the 3.625% 5/13.
As if the European Repo market wasn’t under enough pressure from the Financial Transaction Tax (FTT), now the EC (European Commission) is worried the Repo market is possibly a threat to the financial system. Currently, the ECB is setting up a Repo transactions database to collect data on European Repo market. The EC “will assess whether transparency at EU level has improved, while reserving the right to take legislative measures to remedy the situation”**
What exactly does that mean? My take is that EC is considering changing status of Repo from a sell/buy-back transaction to a secured loan. For more information on the difference between the two, (http://scottskyrm.com/2013/04/two-bankruptcies-that-created-the-modern-repo-market/). Of course, this would be the wrong solution at the wrong time and little thought is being paid to the unintended consequences of such an action.
* ICAP Wrightson; May 13, 2013
** “EU Weighs Repo Rules If ECB Plan Fails to Bolster Transparency”; 5/2/13; Jim Brunsden