The "Optimal QE Exit" Sequence Proposed At Jackson Hole

In the absence of Ben Bernanke or any other notable keynote speaker, the broader attention at this year's Jackson Hole symposium has fallen almost exclusively on one researcher, Arvind Krishnamurthy, professor from Northwestern University, and his paper on "The Transmission of Unconventional Monetary Policy." The reason is that in the paper, as can be seen from the attached handout to it titled "The Ins and Outs of Large Scale Asset Purchases" he attempts to answer the $64,000,000,000,000 question of "How to optimally manage an exit", and specifically what securities would be and should not be bought.

Here is the "optimal" exit sequence advocated by Arvind:

  1. Cease Treasury purchases;
  2. Sell Treasury portfolio;
  3. Sell older MBS;
  4. Cease new MBS purchases

With the following two sub-questions:

  1. Mechanics: Which asset prices will be most affected? Are salesand cessation of purchases conceptually different?
  2. Expectations: How do exit announcements affect dynamics ofasset prices?

As further recapped by Fed watcher Annalyn Kurtz, in the paper, the authors call for these three "exiting" steps:

  • Step #1: Stop buying Treasuries, and even start selling some of the government debt accumulated over the last four years. The Fed currently buys $45 billion in Treasuries each month, but that's not having much of an impact on the real economy, Krishnamurthy and Vissing-Jorgensen argue.
  • Step #2: Start selling the oldest mortgage-backed securities. The Fed currently buys $40 billion in mortgage-backed securities each month in an effort to boost the housing market, but selling its older-dated bonds would have very little impact on current mortgage rates, they say.
  • Step #3: Continue buying newer mortgage-backed securities. This is the Fed's most powerful tool and "most beneficial source of economic stimulus," the economist argue, and should be the last part of the program to be brought to an end.

Hardly anything groundbreaking, and it is well-known by now that the Fed has long accepted the TSY->MBS sequences as the generic way out as the impact of MBS monetization is substantially higher on financial conditions than merely TSys.

As we reported yesterday when it comes to Primary Dealer expectations, the way QE is expected to taper into nothingness (at least until it is forced to untaper), is starting in September with a $15 billion reduction in monthly flow ($10bn TSYs, $5bn MBS so in line with Arvind's views), until we hit $0 in June.

One thing that is notable in today's paper is that Arvind does discuss the "Scarcity Channel" - further confirmation that the Fed is recognizing the crowding out effect. As to whether or not the Fed perceives such scarcity as no longer benign and impairing endogenous private sector liquidity, we will find out next month when Bernanke will or won't announce the Taper.


* * *

As to "optimally exiting" QE in practical terms, without causing an epic market dislocation, in a world in which every asset class, and portfolio manager has been habituated to falling back on Chairman Ben if and when things turn bad, and where "bad" is only "good" thanks to the fed, well... good luck.

The complete handout from the Arvind presentation: