Given the recent improving economic data in the US, the consensus likelihood of getting QE3 later this year seems to have gone from probable to uncertain.
Obviously it all depends on what happens in the next few months, with particular emphasis on employment trends and inflation expectations. For his part, Bernanke certainly hasn’t done anything to dismiss the notion that he’d like to keep going, emphasising the influence of the depressed housing market and arguing that the unemployment rate understates the true severity of problems in the labour market.
As for the rest of the FOMC, the language in the latest minutes was vague.
But something we didn’t see in the minutes and haven’t heard discussed much — strangely, now that we think about it — is the possibility of extending Operation Twist beyond its expiration on June 30.
Let’s say, hypothetically, that we end up with a situation where job gains slow from their January pace and headline inflation rises unexpectedly faster than core because of oil prices — in other words, a situation where Bernanke would want to further ease but would face pressure not to. Then we think the idea of extending Twist might come up: it would be less politically controversial than QE3 because it sounds like a harmless parlour dance it doesn’t involve an expansion of the Fed’s balance sheet, just another change in the composition of its Treasury holdings.
Extending Twist would also make sense as a QE3-alternative if the Fed wants to maintain its emphasis on housing. But rather than the MBS purchases you would get with QE3, instead you keep long-term Treasury yields low.
As for how to go about it, Credit Suisse offers some helpful detail in a recent note, including some possible complications. The assumption is that the Fed would extend Twist through the end of the year. Short version: to get a similar impact as the first version of Twist, in the extension the Fed would have to consider selling Treasuries with a later maturity date than in Twist 1.0 …
Twist 1.5: Options for the Fed
We investigate the potential for the Fed to extend ”Operation Twist” through the end of thecalendar year and find that it could continue to extract the same net duration from the market as it is currently, provided the Fed is willing to expand its population of sellable securities out to the 4-year maturity point.
When the Fed first announced its maturity extension program (Operation Twist), it stated that it would liquidate approximately 75% of its Treasury coupon holdings with between 3 months and 3 years to maturity to meet its $400 billion target. Those figures were easily reconciled when considering a snapshot of the Fed’s holdings at the time. However, when one considers that the Fed’s 3-month to 3-year maturity range has been a rolling bucket through the program’s execution, the full range of maturities ultimately exposed to sales will be closer to 3.5 years than the rolling 2.75-year-wide maturity range.
There is a catch to this: some issues at the very front end of the maturity range tend to slipout of the sellable range before they can be sold. Nevertheless, with only about $200billion in Treasury sales remaining for the Fed to sell in ”Twist”, the Fed still holds close to $376 billion in Treasuries that are currently within the sellable maturity range or will be within the maturity range by the end of June.
Even if we allow for a modest quantity of Treasuries to slip through the bottom of the maturity range without being sold, the Fed should still have about $150 billion in Treasuries within its sellable range by the end of the program on June 30.
If the Fed decided to extend the program another six months to span the election cycle,the rolling nature of the sellable maturity range would introduce about $70 billion in additional sellable supply from issues rolling down in maturity. Combined with the $150 billion remaining at the end of the program, the Fed would have approximately $220 billion available for sale, which would provide for another $165 billion in additional ”twist”, assuming the Fed continued to limit sales to 75% of its available supply.
Unfortunately, this would result in a much less effective program in terms of the net duration that could be removed from the market each month, provided purchases were allocated across the curve similarly to Twist 1.0. With only $27.5 billion in purchases eachmonth (assuming sale of 75% of available holdings), the program would remove just $32.5 billion per month in 10-year equivalents compared to $51.8 billion per month in Twist 1.0.
However, if the Fed were willing to extend its sellable range out to the 4-year maturity point, it could introduce another $220 billion in sellable issues that mature between December 2015 and December 2016. All together, the Fed would have approximately $440 billion in sellable supply. Assuming a hold-back of 25%, that would leave over $300 billion the Fed could sell in an extended “Twist,” which would allow $50 billion per month in sales that could be offset with purchases of longer-maturity issues.
Despite the fact that the duration of the issues in the 3- to 4-year maturity range would be higher, and essentially double the DV01 of each month’s sales, the slightly higher par value of purchases each month ($50 billion vs. $44 billion) would offset the impact. The result would be a program that on a net basis would remove the same duration from the market each month as the current incarnation of ”Twist”.
All very speculative and nothing to worry about for a while, just something to keep in mind if the debate heats up again.
Related links:
Was Twist an “anti-stimulus”? – FT Alphaville
How to serve front-end yields up with a Twist – FT Alphaville