For anyone still confused what Operation Twist is (covered here first about 4 months ago), here is SocGen's Aneta Markovska, charting just what the flawed duration extension will look like (as a reminder, unless the 2s10s is steepened, and at that substantially, we may as well bury the banks: nobody is taking on new mortgages now regardless of where the 10 year is, just look at weekly MBA numbers. However, to make sure the US banking system expires, just flatten the curve completely, and it is game over for NIM). In a nutshell, SocGen believes that the Fed will dump $420 billion worth of 1.5-4 year USTs and use them to purchase bonds with a maturity longer than 4 years - ideally, this would be 20 Years (yes, they would need to be reinstated, and this is our view, not SocGen's) and 30 Years, and sell the 10 years. But since the Fed has zero practical world experience, one can only hope, knowing full well the end result will be yet another TARP to bail out the banks. From SocGen: "The next step from the Fed will almost certainly be for more easing and it will almost certainly be duration extension. The only question is September or November? Prior to the August employment report, the market was split 50/50 on the timing of the announcement. The report pushed the odds in favour of September which is our central scenario.We estimate that at the upper limit, the extension could amount to as much as $420bn in duration purchases, which would make it comparable in size to QE2. However, the Fed may not announce the full amount up front but instead give a monthly run rate and reevaluate at each meeting. Matching the previous run rate, we would expect the Fed to do roughly $55bn per month. This could take the Fed as far as April 2012, at which point inflation should have receded enough to put QE3 back on the table." We are not too sure just who will buy the 1.5-4 year bonds at current yields, but certainly some greater fool than us does and always will exist. What is important, is that dry powder for about $55 billion in POMO recycling will suddenly allow the banks to flip assets to greater fools yet again. As to whether this will work, like last year, we very much doubt it, especially since everyone will be buying gold and crude.
More from the very troubled French bank:
The Fed is likely to ease again. The most likely next step will be duration extension which was officially put on the table last week with a mention in the FOMC minutes. Importantly, Fed officials considered an active duration extension involving outright sales of short-term paper; this is in contrast with a passive extension where they Fed would reinvest MBS proceeds into the longer end of the Treasury curve. The only question that remains is when.
September or November?
Prior to last Friday’s employment report, the market was split 50/50 on September vs. November. Activity data, from retail sales to production and durable orders, have generally held up well through July, suggesting that the Fed may want to wait until November. However, the August employment report tilted the odds significantly in favour of a September announcement which is our central scenario.
There are many potential scenarios but for illustrative purposes we show an extreme scenario where the Fed liquidates all of its holdings with maturities between 1.5 and 4 years and uses the funds to purchase further out the curve. Currently, this amounts to $420bn which would be roughly in line with the amount of Treasuries that the Fed allocated to the 4y-30y segment of the curve under QE2. Under the Fed’s own estimates, QE2 reduced the 10yr Treasury yield by about 20-30 basis points. Is equivalent to 100 bps of rate cuts. Operation twist, if taken to an extreme, could have a similar impact.
The ‘twist’ will likely be spread out over the course of several months. Rather than announcing the total size upfront, we believe that the Fed is more likely to give a monthly run rate and simply say that they will re-evaluate it at the next meeting. Under QE2, the Fed increased its holding by $75bn per month of which $55bn was allocated to the long end of the curve (>4 years). Assuming the same monthly run rate and the earlier assumption that the Fed would be willing to liquidate everything in the 1.5y-4y bucket, QE 2.5 could stretch over 7-8 months. If announced in September, it would take us through the end of April.
Will MBSs be included?
There is also the possibility of reinvesting back into the MBS market. By doing so, the Fed may actually get more ‘bang for the buck’. MBS spreads have widened out by about 30 bps this year, particularly following the US sovereign downgrade. By investing back into this sector, the Fed could make a more direct impact on mortgage rates. For now, however, this is not seen as the most likely scenario.