In a blog post last Thursday, former Fed Chairman Ben Bernanke made the case for delaying run-off of the Fed’s balance sheet. Along the way, he offered some support to (what had previously been) my heretical take on the taper tantrum. According to Bernanke, bond yields went up during that period largely because of the signaling effect. It had little to do with the portfolio balance channel, Bernanke implied.
As I recall from the early taper tantrum period, Fed officials were then warning that the bond market was overheating and that they meant to do something about it. That was quite the signal. The market reaction, then, had little to do with investors fretting they would have to take down more default-free rates duration at some point. I would have liked to rely on Bernanke’s authority on this issue a bit earlier. But better late than never — to be so clear.
A fun fact here is that there is actually no inconsistency between Bernanke’s description of the taper tantrum as having “much” to do with signaling and his longstanding faith in the importance of QE’s portfolio balance channel. During the taper tantrum, the expected stock of assets on the Fed’s balance sheet – say five years forward – was arguably increasing, because the projected date of the initial taper was not being draw dramatically forward, and yet the Fed was being increasingly explicit about the idea that the taper would not be followed quickly by asset sales or perhaps even passive run-off.
I recall that an academic had attempted to quantify the two influences and found that the expected stock was indeed rising slightly over the tantrum period. But, I’ll be damned if I can find (again) his report. For me, this is like the mythical Yeti. All I have is the grainy image from memory, which obviously undermines my credibility here.
Leaving aside the importance of signaling in the tantrum, I see a separate set of issues in Bernanke’s case for delaying run-off. Briefly, he overstates the importance of the portfolio balance channel, now as in the past; he accordingly worries too much about balance sheet contraction and unintended consequences; and he falls into an inconsistency that is separate from the issue of signaling discussed above.
Let’s start with the inconsistency, just because many people find evidence of inconsistency more compelling than arguments that go more directly to the question at hand. I don’t think this is a particularly admirable aspect of human nature, but I mean to exploit it to get your attention.
Bernanke has argued that the expected stock of assets on the Fed balance sheet is the business end of the portfolio balance channel. And he worries now that it is hard to estimate how the market might react when the Fed decides to allow passive run-off. From this, he concludes two things. First, the Fed should delay run-off until after the fed funds rate has risen significantly further. This will give the Fed room to ease conventionally – by cutting rates – if there is an untoward reaction to balance sheet contraction.
Second, the Fed should wait until it is in the position to commit to a sustained run-off (until the equilibrium balance sheet is achieved) before beginning the process, because the odds of market disruption are reduced if the Fed’s actions are “predictable.”
I have no problem with the first point, because it is fully consistent with my view that QE works mostly through signaling and behavioral channels and that the knee-jerk reaction to the first sale might be disproportionate to the official story of how QE works. It is probably a minor point, but the cost of taking it on board seems low to me.
More importantly, Bernanke’s second point is directly at odds with his conception of how portfolio balance works. If it were the expected stock of assets on the balance sheet that delivered the portfolio balance effect, and if that effect were important, then the Fed would not want the initial move to be viewed as a watershed.
Rather, the Fed would want the expected stock to move only incrementally, which would actually be best served by ambiguity about the importance of the first move. I am not making the case for ambiguity here. I am making the case for logical consistency, and drawing an inference from its absence in this case.
I have seen it argued that the Fed does not want to use QE as a “marginal” tool of monetary policy. I think I saw that point raised in comments after my guest post to the Noahpinion blog. (I thought it was a good one – even though offered against my main argument there.) On this view, QE should remain in place until the Fed judges it safe to be removed and then it should be removed, much like a band-aid being ripped off. Never to be touched again.
That view rhymes with Bernanke’s, obviously. But so far as I can tell, the motivation of it is politics, not economics. QE is a controversial tool and the less attention brought to it the better. The Fed would not want to draw attention to QE by allowing run-off and then later having to stop run-off or, worse, resume net purchases.
I have a lot of sympathy for this take, as mentioned. I don’t think the worry is actually valid, but I accept that it heavily influences policy . More importantly, though, it has nothing to do with what Bernanke is arguing in his recent post — beyond also concluding that the Fed should delay asset sales.
The bigger issue here is that QE is seemingly not an important tool of monetary policy, beyond its signaling and behavioral effects. As Bernanke himself emphasizes in his recent post, the signaling effects can easily be separated from the portfolio balance effects, just as the Fed successfully did in the fall of 2013. The behavioral effects are tougher, which I think is actually an argument against QE. Monetary policy should not work mostly by fooling people, IMV. On the other hand, those behavioral effects seem to be ephemeral, so I don’t think they are a big deal beyond short horizons. By the time they might affect the economy, as opposed to a trader’s bonus, they are probably gone.
FWIW, my view on QE remains that it is largely harmless, * rather than dangerous or a cure all — or indeed cure anything. Mechanically, it is nothing more than shortening the maturity of the federal government debt, with Fed rather than Treasury as agent. And when the Fed allows run-down, the rising trend in the average effective maturity of the federal debt will steepen slightly / resume. ** There will probably be some market reaction around that, but it should be of more interest to short-term investors than policy makers, I would think.
The purpose of this post was to torture Bernanke until he seems almost to say as much. I didn’t quite get all the way, but made some progress, I think.
* I see that Nick Rowe of the Worthwhile Canadian Initiative believes that QE prevented deflation by raising the money stock. I am not in that camp, although I sympathize (purely on the basis of intuition) with Nick’s desire to use monetarist concepts to lean into Neo-Fisherianism. I will leave that debate to my betters! The only point I would assert here is that Nick’s mechanism seems to have nothing to do with portfolio balance, at least a la Bernanke. So I feel within my rights bracketing that discussion. Maybe one day I will be up to speed on it.
** The average maturity of the federal debt has nominally been rising since late 2008, roughly when QE began. While QE was ongoing, the Fed was implicitly swapping bonds for overnight debt acting as agent for Treasury. Accordingly, once we pierce the Treasury/Fed veil, we see that the effective maturity went down slightly during the QE period. Summers et al have documented this, although indirectly, looking at the various sources of net duration supply. Recently, the nominal maturity has stopped rising as Fed actions have gone to neutral. (I am never quite sure if that is fully coincidence.) Assuming the Treasury maintains the current mix and the Fed allows run-off, then the effective maturity should start rising again. And of course Mnuchin is talking about radically lengthening maturity, which is an entirely separate discussion.