After Thoughts & Questions on QE3

Great post from FT Alphaville on thoughts & questions of QE3.  (Reposted below)

1) We went into the presser wondering what “substantially” means; we left it wondering what “ongoing, sustained” means.

Bernanke gave a good answer for why he didn’t want to tie the end of the open-ended purchases to the unemployment rate: it can rise or fall for various reasons that don’t always reflect changes in employment (changes in the participation rate).

But his answer for what a labour market that is improving “substantially” means called to mind the old line about pornography; he’ll know it when he says it:

“We’re looking for ongoing, sustained improvement in the labor market… What we’ve seen in the last six months isn’t it.”

Bernanke added that although the FOMC had discussed “policy reaction functions”, the members hadn’t agreed on the right criteria.

The natural question is whether this lack of explicit targets undermines the message somewhat, as David Beckworth notes. The Fed is depending on the signaling strength of the other language in the statement (and reinforced in the presser) to be enough.

2) What else might be next for FOMC communications tools?

But we’re not ruling out the possibility of more explicit targets to be announced later. Bernanke brought up more changes in its communications after our colleague Robin Harding asked him what else the Fed could do as a follow-up if today’s announcements didn’t work. Here’s what he said:

We continue to work on how best to communicate with the public and how best to assure the public that the Fed will remain accomodative long enough to ensure recovery… Clarifying our response to economic conditions might be one way in which we could further provide accommodation.

Of course, Bernanke also might have been softening the ground for explicit targets: Evans Rule, NGDP targeting, etc…

Less importantly, we also wonder if he was referring to the ongoing project to arrive at a consensus forecast rather than continuing to produce frustratingly anonymous bubble charts and graphs. We’ll probably find out more about this in the minutes to this meeting.

3) Some have noted that the size of the new purchases, $40bn a month, is much smaller than both QE1 and QE2. So what?

For one thing, as bad as things are now, the circumstances leading up to those previous rounds were a lot worse (severe financial market strains before QE1 and a real threat of deflation before QE2).

But more to the point, we think the initial small size of the programme is feature rather than bug. Well, maybe.

We won’t repeat all of our thoughts from an earlier post. But the Fed’s previous easing programmes, though successful in their main objectives, were also flawed in that they both removed safe asset collateral from the financial system and failed to disabuse markets and economic agents of the belief that 2 per cent was an inflation ceiling. On Thursday, Bernanke tried the opposite script — smaller purchases to start and clear signs that 2 per cent is not a ceiling.

Will it work? We’ll see, but the hope is that the messaging itself is strong enough that the Fed will generate more economic activity per dollar of balance sheet expansion (precisely because the program is open-ended) than with lump sum purchases. In monetary policy parlance, the portfolio balance channel works better when accompanied by use of the expectations channel.

4) What will happen at the end of the year?

Operation Twist is scheduled to run only through the end of 2012, and the FOMC left itself the flexibility in the statement to make additional asset purchases later on as part of QE3. It’s quite possible that the Fed will continue to purchase longer-term Treasuries, but probably without the short-end Treasury sales that to this point have been neutralising the impact on the monetary base, as its supply of these Treasuries to sell will be close to exhausted.

The Fed has previously argued that the stock of holdings matters more than the flow of purchases, but…

… [the Fed] also recognizes that the stock of securities purchased, in 10-year note equivalent terms, will begin to decay on January 1 in the absence of additional Treasury purchases. …

We projected the Fed’s longer-term Treasury purchases through the end of Twist, then modeled the decay in duration that would occur on a monthly basis beginning in January if no additional Treasury purchases were to be announced. We find that the Fed would need to purchase approximately $12 billion per month in 10-year notes simply to prevent the stock effect of its previous purchases from leaking back into the market.

Comments and estimates from Credit Suisse strategists.

Anyways, the answer to the question of what happens at the end of this year also depends heavily on whether politicians have reached an agreement on the fiscal cliff during the lame duck session, more on which below.

5) There were no questions about market dysfunction in the presser, and Bernanke seems satisfied that this won’t be a problem. We’re still worried.

Our own questions — mainly about removing collateral from short-term lending markets and about liquidity in Treasury and MBS markets — remain. Here we’ll quote at some length from a couple of analyst notes in response to today’s meeting, each of which does a good job of explaining how this might get complicated later on.

RBC first:

While the $40bn of agency MBS purchases is smaller than the consensus expectations, it’s still large enough to create some unusual technicals in the MBS market.

Net MBS production this year has averaged $4bn per month. So the Fed purchases will take net supply deep into negative territory. The scarcity of new supply could create unusual delivery issues for the To-Be-Announced market, particularly if rates move abruptly.

At one point during QE1, the Fed had been buying the new current coupon (4s). Dealers assumed that they would be able to fill TBA sales with new production. However, rates rose abruptly and new production shifted to the 4.5s. At this point, the Fed owned more than 100 percent of the float of 4s. At the time, the price dislocations were modest because the cost of failing to deliver was small. But today there is a 200bp fails fee, meaning sellers will need to factor into the price their ability to source the cheapest to deliver pool and the price difference between the cheapest to deliver and the second cheapest to deliver pools. Again, these delivery issues will be most acute if surprise movements in mortgage rates shifts production to a different coupon.

And Credit Suisse, which notes how Treasury market liquidity might also be impaired if additional purchases are announced in December:

Another aspect of the announcement was the Fed’s decision to let its extended maturity extension program (“Twist”) run separately from, and in parallel to, its new unsterilized MBS purchase program for the next few months. There are a number of reasons we believed the Fed would choose this path.

First, the current pace of purchases in the Treasury market is already substantial in the “Twist” program. Monthly purchases consume more than 48% of new Treasury issuance in the 7-year sector, 64% of issuance in the 10-year sector, and 93% of monthly bond issuance. If one considers the Fed’s self-imposed per-issue holdings limit, the purchasable supply being introduced to the market each month is only 70% of the actual auction sizes. As a result, monthly purchase volume comprises an even more imposing proportion of new purchasable supply, with Fed purchases consuming 70% of purchasable issuance in the 7-year sector, 91% of 10-year supply, and 148% of bond supply.

Of course, the Fed is not limited to purchasing new issuance, with the majority of purchase volume consisting of far-off-the-run issues. Indeed, on its face, remaining purchase capacity for the Fed still appears relatively robust. However, much of this seasoned supply is stashed away by investors that aren’t necessarily interested in selling to the Fed, at least not at prices close to the market.

As a result, the brisk pace of purchases relative to new supply has the potential to impair liquidity over time. The steadily declining participation in Fed purchase operations, in our view, is evidence that the Fed is chewing through the inventory of bonds investors are willing to sell. If purchases were to be increased in order to facilitate QE3, the potential for purchases to impair liquidity would only grow.

These issues matter because they go right to the idea of these purchases being open-ended. If difficult market dynamics assert themselves, then the Fed will have to reconsider what else it can do; and it would also risk some damage to its credibility.

In the most recent minutes, and then during Jackson Hole, Bernanke suggested that further asset purchases were unlikely to lead to market disruption problems. The NY Fed also tries to head off some possible concerns in this Q&A. Today’s announcement essentially confirms that Bernanke isn’t yet worried about this.

But that could change. Consider a scenario where a fiscal cliff deal isn’t reached by the start of the year, and to credibly signal that the Fed will do all it can to make up for the aggregate demand shortfall, the FOMC is compelled to announce more asset purchases than it had expected. This remains, of course, a low-probability scenario for now.

Oh, and do check out this stat from SoberLook.

6) Lowering or eliminating IOER won’t happen anytime soon.

This was one idea that Bernanke didn’t mention in response to Robin’s question (and others). We might also find out more about these discussions in the minutes, but at this point we’d be willing to bet that IOER won’t be lowered until NGDP has begun to accelerate meaningfully, when the Fed can safely do this without disrupting money markets.

7) On credibility and NGDP targeting, just what did Bernanke mean?

He made some fascinating remarks in the presser when asked about Michael Woodford and NGDP targeting:

The thrust of [Woodford's] research is that forward guidance is in fact… the most powerful tool that central banks have when interest rates are close to zero. He advocates policies like nominal GDP targeting that would essentially require credibility lasting many years, the implication being that the Fed would target the level of nominal GDP and promise to do that years into the future even inflation rose as a result of that policy.

His own perspective is that credibility is the key tool that central banks have to get traction at the zero lower bound.

Whether we have the credibility to persuade markets that we’ll follow through is an empirical question. The evidence… is that when we’ve announced extended guidance, financial markets have responded to that, private sector forecasters have changed their estimates of what unemployment and inflation will be when the Fed begins do accomodation. So the empirical evidence is that our announcements do have credibility.

There’s a good reason for that. We’ve talked a lot, both publicly and privately, about the rationale for keeping rates low even as the economy strengthens, and the basic ideas are broadly espoused within the committee. So there is a consensus that even as personnel changes and so on going forward that this the appropriate approach and that by following through we will have created a reserve of credibility that we can use in any subsequent episodes that occur.

We’re not exactly sure what Bernanke was getting at here, but we can think of a few possibilities:

– He is saying that the Fed has the credibility to do what it is doing now, but not yet adopt explicit targets.

– Relatedly, he is saying that a buildup of credibility that will result from following through on today’s announcement could allow the Fed to make the bigger leap to NGDP targeting later.

– He is suggesting that what he is doing now is pretty close to NGDP level targeting anyways.

– None of the above.

Also notable were his thoughts on personnel changes. He could be talking about worries that a less-doveish FOMC voting membership (and certainly this year’s is very dove-ish) would be more likely to tighten, which might cast doubt on today’s commitments to keep rates low for years into the future.

But he also might talking about the fact that his own chairmanship ends in January 2014, and he is unlikely to be reappointed. This might lead some to question the Fed’s commitment given that the language guidance now goes to mid-2015. By saying that his views are widely shared by the committee, he minimises the impact of such doubts.

 It’s impressive how much Bernanke was able to sway the committee from previous meetings.

Related to the point above, this too could mitigate somewhat the concern that he’ll probably be gone come January 2014.

It’s amusing to see those two single members on the extreme flanks. Best guess: Lacker is one, Evans the other.

9) Continue keeping an eye on how mortgage rates track MBS yields.

That this spread has widened because of impairment in mortgage markets is commonly known. Don’t forget about it. But as we noted here, one reason banks might have been hesitant to ramp up origination capacity was that they didn’t believe the low-rate environment would last. The commitment today to lower rates for longer could change some of their minds.