How To Front-Run The Fed's Upcoming SOMA Limit Increase

With the bogey of a minimum QE announcement of $100 billion a month, leading to an in kind purchase of Treasurys, in addition to $30 billion a month from MBS Refis courtesy of QE Lite, a very likely announcement during next week's FOMC meeting, that nobody is talking about, is that the Fed may raise the existing 35% SOMA limit, or abolish it altogether, due to the imminent ceiling hit of purchasable CUSIPs. As a result, as Morgan Stanley suggests, possibly the most profitable Fed frontrunning trade if one wishes to bet on consensus QE, is to buy SOMA excluded CUSIPs as these will be telegraphed to be next in line to be monetized. Of course, in the apocryphal scenario that the Fed disappoints the market and decides to announce a less than $100 billion a month, or, gasp, nothing at all, MS' Igor Cashyn expects a complete bloodbath in rates (and most certainly in risk assets). Then again, the probability of the Fed doing the right and/or prudent thing ever is nil, so we would focus on buying out of favor SOMA issues, because as Morgan Stanley reports: "Net, we like buying what the Fed is buying."And how could one not: after all Morgan Stanley announces that in 2011 net Treasury issuance net of Fed Purchases will be zero!

In terms of looking at QE, Cashyn expects three scenarios: QE announcement of greater than $100 billion/month for 3-6 months; less than $100 billion, and no QE. Here are the three scenarios broken down.

Scenario 1: Fed Announces $100 Billion/Month for the Next 3-6 Months

Right on expectations: Our core view is that the Fed announces a specific figure at next week’s FOMC meeting of $100 billion / month in Treasury purchases for the next 3-6 months. The Fed chairman has had plenty of opportunities to back away from QE2, yet has taken none. In addition, a recent survey by the New York Fed polled bond dealers and other investors for their expectations of the size of QE2 along with its likely impact on yields (as reported by Bloomberg), which we view as a preparation for just such an announcement. Continued weakness on the inflation front, along with a persistently high unemployment rate, should also justify QE2 from the data front. We thus anticipate that the Fed will announce $250 billion over the next 3 months or $500 billion over the next 6 months (keep in mind that they are also already purchasing around $30 billion / month via their SOMA reinvestment program).

Such support would almost certainly be seen as bullish for the Treasury market, and we advise investors to position accordingly. That’s because the Fed’s purchasing pace will be on track to take down all of the $1.15 trillion in Treasury net issuance that our US economists expect for F2011, and should be reflected in yields accordingly (Exhibit 1):

Currently, we anticipate that the market is also expecting roughly $100 billion / month, but what’s contributed to the sell-off in Treasuries over the past couple of weeks is a subtle softening of the market’s call for a substantial program to a more data-dependent, fine-tuned approach. This has reduced the certainty of what will actually be announced, but if the Fed now delivers on the expectation of $100 billion / month, the slide in 10y Treasury yields should then reverse, in our view, and 10y notes should come right back down to the 2.35–2.50% range (a 15-30bp rally from here).

Further, while it can be argued that the implicit monetization of US government debt may ultimately prove inflationary (in fact, we like being positioned in 10s20s inflation breakeven steepeners to hedge this view), the initial impact on yields is  very clearly bullish, in our view. Investors who have reduced their longs in recent weeks will add back to those longs, and other investors that were previously on the sidelines will get back in. Net, we like buying what the Fed is buying and are bullish on Treasuries.

We also think that the belly will outperform and reverse its recent underperformance, and we continue to recommend staying in 2s5s flatteners, earning +8bp in rolldown + carry / 3-months, as we equate an expansion of the Fed’s balance to mean that the Fed will not be hiking anytime soon. Similarly, the 2s10s curve should also flatten.

We also think the Fed’s goal in this scenario is to drive inflation breakevens higher and real rates lower, and TIPS investors should also position accordingly. We specifically like buying breakevens in the front end of the curve in this scenario (i.e., <5y sector), as any announcement of QE2 should also be accompanied by renewed weakness in the dollar, resulting in a rise of the $-denominated prices of commodities, to which breakevens in the front end of the curve are most sensitive (see With QE2 All but Certain, a Look at the Treasury Market Implications, October 8, 2010).

Fundamentally, we cannot lose sight of the fact that QE2 is intended to inflate asset prices, and to that end, equities and bonds should both rally. But the rally in equities will only have a secondary effect for bond yields, which will still be pushed lower over the near term. This is in fact what the Fed wants to accomplish, driving Treasury yields low enough to promote investors to get out the credit spectrum and increase the valuations of riskier assets – but this can only be accomplished if yields stay low.

And here is the key trade that is most profitable in case of scenario 1: buy "SOMA-excluded" Cusips:

SOMA Ceiling Rise Possible but Unnecessary (Yet): Apart from the size of the purchases, next week’s announcement could also be accompanied by an increase in the Fed’s SOMA limit from 35% currently to, say, 50%. The implication of this is that Treasury notes that are currently ineligible for purchase by the Fed (e.g., mostly high coupon bonds) reverse some of their recent cheapening on the curve versus the low coupon bonds. Exhibit 2 shows where such bonds are concentrated on the UST curve:

An area of the curve where high coupon bonds are likely to outperform in the Exhibit above include rolled-down 30y bonds in the 2018-22 year sector (although the 2026-27 year bonds already seem to be a bit rich on the asset swap curve).

We will compile a list of the most convex 2016-2020 SOMA excluded CUSIPs soon and present it to readers to determine which are the Treasuries most likely to benefit from Bernanke's insanity.

Continuing on, here is Scenario 2, one which will see a major move down in assets from bonds to stocks, and everything inbetween. In a nutshell: expect the 10 Year to sell off to 2.75% if the Fed does not do monetize at a $1.2+ trillion a year runrate.

Scenario 2: Fed Announces <$100 Billion/Month for the Next 3-6 Months

Below expectations: A risk to our view is if the Fed tries to be too flexible in its approach to QE2, driven by the uncertainty on the fiscal front. Specifically, whether the Bush tax cuts get extended, as well as in what form, may lead the Fed to hold back for now. Further, a recent article by Jon Hilsenrath in the WSJ highlighted that three regional Fed bank presidents – Narayana Kocherlakota of Minneapolis, Richard Fisher of Dallas, and Charles Plosser of Philadelphia – have expressed skepticism about QE2, and a smaller program may be needed to pacify some of this dissent (although truth be told, they will not be taking voting positions at the FOMC next year).

In any case, we think the Treasury market would be disappointed, leading 10y yields to rise back toward 2.75% and the belly of the curve to underperform (as it is directional with yields). CFTC positioning data of speculative investors (i.e., non-hedgers) reveals that market participants are currently long in both the front end and the back end of the curve (Exhibit 3):

With both the front-end and back-end longs at their 2-year highs, any disappointment from the Fed is likely to drive yields higher from here.

And now, for the last scenario, one that will cause untold destruction in stocks, and is thus impossible. But here it is anyway:

Scenario 3: Fed Does Not Announce Treasury Purchase at This Time

Complete disappointment: Fed language promises support to the economy if conditions continue to worsen but backs away from providing any new stimulus at this meeting. We personally view this as a <5% probability event, a tail risk to our view if you will.

In such an event, we are likely to see a major backup in Treasury yields, with 10y notes going back to the 2.75-3.00% range. Exhibit 3 above already demonstrated the fact that investors are long the market, and any disappointment is likely to be met with swift selling of Treasuries, led by the back end.


However, the one part of the curve that we do not expect to be materially affected is the front end of the Treasury curve, as the Fed will remain on hold for the foreseeable future (read: disappointing inflation trends / high unemployment). Thus, the  2s10s curve will remain directional with movements in the 10y note, and will flatten / steepen to the above moves accordingly.

Investors should think in terms of asset inflation when evaluating the effects of QE2 – if no purchases are announced, both equities and bonds are set to underperform. Our certainty on this is quite high, as equities have rallied 12% since Bernanke’s Jackson Hole speech (near their 1y highs) and 10y yields are still near their 1y lows (Exhibit 4):

And while some might think that an underperformance of risk assets leads to an automatic outperformance of Treasuries, which may eventually keep Treasury yields from rising much higher over the long-term, Treasury yields will still sell off over the near term.

Since this is the Fed we are talking about, whose only mandate is to keep artificial stock price levels as high as possible, you can forget about Scenarios 2 and 3. Which is why the SOMA trade appears most attractive. And don't forget to fund it by shorting the carry currency of choice these days... the dollar of course.