04
Oct
2014
Theory and Evidence That QE Pushes Down Long-Term Interest Rates
Thanks to Keshav for the Krugman link I used in the first half of this Mises CA post… Incidentally, if you thought, “Well duh, of course interest rates stayed low, because the Fed did Operation Twist and then QE3!” then you should really read this one. It’s not as open and shut as you’d like it to be. I can especially understand why Scott Sumner thinks the normal thing is for long-term Treasury yields to rise when the Fed buys bonds.
I think this is a really good summary.
“I can especially understand why Scott Sumner thinks the normal thing is for long-term Treasury yields to rise when the Fed buys bonds.”
Basically this just the reverse of what I said about Volcker and the 70ies. The question is just when which depends on when the market expects what to happen.
I mean look at the ECB. Draghi managed to push long term bond rates of many countries down by doing nothing, or technically by tightening. The only thing he said was “he will do whatever it takes” to stop sovereign defaults in the EU in 2012, then he did nothing and even let the ECB balance sheet shrink due to repaying of LTRO loans of banks. Yet long term rates across the board came down. Look at Italy and Spain for example.
Everything just based on a few magic words. Of course the market doesn’t forget what Draghi said, and at some point he will have to deliver.
That’s essentially a verbal guarantee of risk underwriting. Thing is, he doesn’t have either the authority or the money to provide that guarantee.
The English refer to this as a “cledge”.
Nations in the EU aren’t allowed to exceed 60% debt to GDP debt levels, and 3% deficits annually as well, so…
BTW the ECB already bought bonds in the primary market in 2011.. That was the reason Jürgen Stark resigned.
http://www.ecb.europa.eu/press/pr/date/2011/html/pr111103_1.en.html
And the market obviously believes Draghi. The tragic issue is that the market believing Draghi is what will make the situation at one point worse and the pressure on the ECB to actually do what Draghi promised even bigger.
It’s just amazing that the solution (Keynesian policy) to the problem that does not exist (market failure) introduces 700x more uncertainty into an already uncertain world which can then be blamed on unconstrained laissez faire capitalism.
It’s also amazing that all of this uncertainty regarding the impact of these policies (to cure the problem that does not exist) and the debates about this uncertainty do not cause the policies (which cause the uncertainty) to gain a bad reputation.
If we narrowly focus on the platinum age of Keynesianism, during lunchtime on March 6th, 1947, then it’s true that markets were stable. Compare that to the 1860s and 1870s, and the evidence is clear.
We cannot observe what otherwise would have happened to interest rates had government agents acted differently, or not at all in any capacity of government but a free market instead. No empirical facts can serve as evidence supporting or refuting one economic theory over another.
The supposed counter-arguments to this, which take the form of extreme events, and use phrases such as “strong likelihood”, as well as constrained to assumptions which make them a priori logical statements, are just outcomes that truth.
” And in any event, there’s no reason to expect a sudden swing in prices from a policy that was announced well in advance (such as halting bond purchases).”
Is not having any explanation for what else could have caused the change in prices the means by which we determine A correlated with B (with/without time lag) a revealing of A causing B!? I’m being serious. I don’t get how that works in economics specifically. Aren’t there always more factors than any one person can assimilate for why X is X and not, say, X+x?
I get the ceteris paribus arguments, they make sense to me, but if we look at X, in economics, how can anyone say X is X for “this” reason and no other?
We hope so! The whole reason for quantitative easing is that normal monetary expansion, printing money to buy short-term debt, has no traction thanks to near-zero rates. Gaining some traction — in effect, having some inflationary effect — is what the policy is all about.
Where is the underlying empirical evidence that economies have, need or lack “traction”? What we do know is the human beings cannot read each others’ minds and require the objective manifestations of subjective value judgments in the form of prices to guide them in a world of strangers and that these policies allegedly design to create “traction” (whatever that might be) impair that process.
The immediate effects of CB purchases of long term debt will be to increase their price and reduce their yield. But these purchases will have other effects such as increasing future inflation and growth expectations , and these effects may mean that CB purchases of long term debt may increase and not decrease interest rates.
I am pretty sure that Krugman and Gross would agree with this so thats not what the disagreement is about I think it more about the likely timing of economic recovery. Gross’s logic appears to have been good – long term government bond prices are high because interest rates are depressed – as the economy will eventually recover their prices will fall – they are therefore a bad investment. Obviously his timing was horrible wrong. DeLong gives a convincing explanation of why that might be.
Krugman called things differently to Gross and correctly – but as (in my opinion) the issue is more about the timing of the eventual recovery rather than any fundamentals in terms of economic theory I’m not sure why Krugman has chosen to spin this the way he has.
Transformer wrote:
I am pretty sure that Krugman and Gross would agree with this so thats not what the disagreement is about I think it more about the likely timing of economic recovery.
Transformer I’m not trying to be a jerk, but it sounds like you haven’t actually been reading what Krugman has been writing about Gross. In the post I linked to, he specifically says something like, “Brad DeLong tries to rationalize Gross’ mistake by saying he had a story about recovery that turned out to be wrong, but no, that’s not what his mistake was. Gross’ error was far more crude than a mistake about the timing of recovery.”
Well, Krugman said at the time (2011) “And so I don’t expect rates to spike when QE2 ends unless there’s good economic news that gives us a reason to believe that the zero-rate policy on short-term rates will end sooner than expected.” So, I do think this is partially about the timing of the recovery.
But I agree (having now looked at the posts more carefully) it was mostly that Krugman thinks that Gross was an idiot for believing that the QE purchases themselves were what caused rates to be low.
An on this point I think Krugman is right and Gross wrong.
Krugman changes his mind so often it’s impossible to figure out his position. Sometimes he says that low interest rates are the cause and “liquidity trap” is the effect. Sometime he says “liquidity trap” is the cause and low interest rates are the effect.
This presumes there are households all over the USA just hankering to buy some government debt at near zero interest. Seems unlikely to me. If such household “savers” exist, why is the lion’s share of new government debt purchased by the Fed?
I’ve highlighted the key point… there hasn’t been much economic recovery in the USA. The employment participation rate continues to fall.
Essentially though, Gross was playing double-guess the Fed. Let’s suppose the Fed just made a rule from tomorrow that it would stop buying any government debt at all, end of story, and it would absolutely not expand its balance sheet, not one iota more. Does anyone (even Krugman) dispute that the primary effect of this would be higher rates on government debt? There may be other secondary effects down the track, but at very least we would see a higher bond rate.
When QEII was announced, the Fed put a specific number and, by implication, a specific end date to the operation. The time to short bonds was right after that announcement in Jackson Hole which the market had already priced in during the Summer of 2010 following the end of QE I. If Gross waited until after the program’s first purchase in November 2010, then he waited too late. The market had already priced in the end of the program and perhaps had even anticipated Twist or a QE III.
When QEII ended, the market was already pricing in Twist, which, if memory serves, I first read about from Gross himself. The reaction in the 10Y yield, however, was pretty muted in the Summer of 2011 with a small spike down when the program was officially announced
I think the bond market’s reaction to QEIII is instructive. It wasn’t given a specific purchase amount, nor end date when announced, and the 10Y treasury traded pretty damned flat from the end of QEII until QE III was announced, and continued to trade pretty flat until Bernanke gave it an end date in May 2013. Now, the question is- what has the market priced in since it knows QEIII end later this month? Long rates have trended down all year to date, clearly fooling a lot of people who had forecast higher rates at the year’s beginning. Gross certainly has good company in being wrong.
Yancey I strongly agree with most of the things you bring up here. The issue of “the end of QEII” would be more about the announcement, not the actual day they stop buying. And since QE3 was open-ended, that’s why it was a much better test case.
However, even here I think there are some nuances. When the program officially ends, you can see what the Fed is saying about ratcheting up a new program etc. So that’s the sense in which it wouldn’t shock me if there’s something that happens when the flow trickles to a halt, because people would have been guessing beforehand about possible announcements around that time, and then they’d observe what announcements actually occur.
Bob, why isn’t the high inverse correlation between the 10-year yield and M2 part of the empirical evidence in our favor?
Somebody, I don’t get into that because M2 is partly endogenous. The Fed can’t directly control M2, but it has been able to control its bond purchases.
Somebody can you point me to something on this? I don’t mean to say if you’ve got smoking gun correlations that they’re irrelevant, I’m just saying it would be hard to directly map that onto the Krugman/Gross debate. Krugman is trying to use this as a “teaching moment” to show that interest rates are low because of depression, not because of QE, and so I wanted to hit him directly on that point.
Bob Murphy is wrong.
Be more specific TravisV I’m wrong about a lot of stuff.
No, no. he has a point. He just forgot to share it.
It seems pretty clear to me that long term rates went up with the QEs. One response is that there was an anticipatory liquidity effect, so that the decline in rates happened before the announcement, and the rise in rates after the announcement was anticipating an end to QE.
How there could there be an anticipatory liquidity effect? The effect itself is based on the fact that the buying from the Fed overwhelms the anticipation of higher inflation and economic growth. To get an anticipatory liquidity effect, the market would have to anticipate its own lack of ability to anticipate.