Markets are now expecting the Fed to end QE on schedule, in spite of recent market volatility and faltering foreign economies. I don’t see much reason to dispute that. The Fed wants to be rid of the QE straitjacket, even if it was a good fit last year. It’s better to be less constrained.
There have been some rough, scary days in the bond market recently, to be sure. But the Fed tends to have a “markets go up, markets go down” weary disdain for short-term volatility, at least so long as the mechanics of markets like clearing and settlement are working fine. Less liquidity in junk bonds is not enough to break the US economy, even if it might break some leveraged players with positions the wrong way round.
The central bank also tends to act with a lagged response compared to market players: officials tend to look at the broad average of data since the last meeting, rather than market mood-swings on the latest few days of data.
It would need much larger and more persistent market disruption to alter the Fed’s course. Indeed, you have to expect and even welcome some market upset as the Fed gradually returns to normality. A few market headaches now might reduce market complacency and tendency to bubble behavior. A little bit of two-way risk on Wall St is no bad thing.
That is not to deny the global real economy outlook is getting darker. Weakness in foreign economies is something which needs further monitoring, but the US has plenty of experience with being left as the demand locomotive which is supposed to pull the rest of the world along with it. Just about every US economic official of the postwar period has tried to persuade the Europeans to stoke demand a little more, for example. The Fed will probably want to see firm evidence of foreign weakness actually affecting the US, and is also conscious that falling oil prices will boost the US consumer. A rising US dollar will also boost foreign economies.
So the message will be a perennial Fed line: we will judge the pace of tightening in the light of incoming data, we have no precommitments to any course of action but will respond flexibly to new information.
The market can take the same message in different ways, though , and that’s where the main uncertainty lies. Investors may read that message as a willingness to delay rate rises next year, but there won’t be anything substantially new in it. But of the Fed chooses to reemphasize it has many tools to deal with contingent weakness despite the end of QE3, that might sound dovish to some ears. It’s all the same thing, however.
Could the Fed surprise with an extension of QE? The only reason would be to shock the markets with an additional injection of dovish credibility. But that runs into a “save the ammunition” argument: if there is any chance you may need to respond to a major negative shock later, it makes more sense to keep your powered dry until you can see such a target. So I think it is not the time to spring a major surprise.
Remember that Yellen is not a perpetual dove. It’s just that as long as domestic inflation seems muted or non-existent, she’ll err towards keeping rates low for longer. Look for a revival of the “first rate hike” versus “steepness of rate trajectory” debate before long, though. The longer you leave the first hike, the steeper subsequent rate hikes might have to be – and that could lead to bond market gyrations that make the last few weeks look like child’s play.
The collapse in European bond yields has been truly historic this year, with German 10-year bunds now hovering around 0.9%. Danger lights are flashing. There are obvious explanations: above all, growing deflation fears, as well as faltering economic data and Draghi’s comments last week about fiscal support and QE. Add to that some safe-haven related flows because of fighting in Ukraine. The ECB is now in full alarm mode because inflation expectations are dropping rapidly.
The terrifying thing about deflation is that expectations of falling prices can feed on themselves and become self-fulfilling, creating a chain reaction of deep problems in a modern economy.. The question is what can policymakers do about it.
The dirty secret about modern central banking is that monetary theorists understand very little about the process of expectation formation. That is why so much policy debate drifts into irrelevance.
Economic policymakers usually turn it into a debate about credibility, stemming from the Kydland-Prescott academic tradition which focuses on time consistency and credible commitments. It is often rational to break previous commitments, so why should anyone believe current promises? It also gets linked to another somewhat stale debate about “adaptive” versus “rational” expectations. For example, much of the difference in opinion on the FOMC come down in practice to disagreement about how forward-looking rather than backward-looking consumers are when it comes to inflation expectations. Do consumers and businesses just observe the recent trend, or anticipate problems before they arrive?
The amount of actual empirical work on the matter in all this is negligible, however. It is mostly prescriptive theory rather than descriptive or experimental work. And thinking generally about credibility in policy debates tends to be sloppy, with dozens of traps. One major lesson is credibility is heavily dependent on the context, not something you can apply to any situation.
The importance of expectations has, however, led to much more emphasis on policy communication in the last few years, as a matter of practical necessity (and desperation). Monetary policy has become more like theater than math or engineering. How can you sound more credible? How can you make statements believable? How can you get people to understand your approach? Hence Yellen’s endless communication committee work on the FOMC before taking the top job.
But the deeper truth is academic economists just don’t have the skills or tools to understand much about communication, because of course it falls into psychology and organizational science and rhetoric and persuasion instead. Parsimonious mathematical models are not adequate guides in these realms. And people can reasonably doubt whether policymakers have the skill and capability to deliver, whatever their intentions may be.
Instead, it comes down to asking why people change their minds. That is my main focus in policy issues. Everyone knows from their own experience that attitude change is often a drawn-out, fraught, conflicted process. People often see only what they want to see for long periods of time. They can be influenced by networks and relationships and trust, by familiarity and the salience of issues within their larger sphere. They observe facts, but can explain them away or ignore them. (Watch any tv political debate.) There are long time lags and considerable inertia. And many people never change their beliefs at all, regardless of the evidence.
It is also a classic stock-and-flow systems problem. Inflation expectations in particular are usually very sticky, and take a long time to change. Think of a bathtub: it potentially takes a lot of drip-drip information (flow) to change the amount of water in the bathtub (stock), but the system can also suddenly change abruptly (the bathtub overflows.) People frequently forget that many policy issues have major stocks -i.e bathtubs, sinks, buffers – contained within them and so do not react in a linear way to marginal change. There are complex positive and negative feedback loops, and decisive events can change things rapidly. Expectations aren’t simply “adaptive” or “rational” but complex.
Policy tools like fiscal policy and QE most likely do not make much difference to consumer expectations, certainly in the short term. Just ask the Japanese how successful QE and massive fiscal spending has been in putting their economy back on a sustained growth path.
Because there is so much inertia in inflation expectations, it’s more likely that after a few months European expectations will drift back towards 2% again, and the ECB will claim the credit for something they had little to do with. But if inflation expectations really are becoming destabilized, it could take five or ten years and vast pain to fix the problem.
The Fed is all over the place in its discussion of policy strategy in the minutes yesterday, with a whole zoo of potential tweaks to policy operations under consideration.
But two clear themes stand out: 1) a desire to “alter the policy mix” or “rebalance” policy, as the FOMC worried about the market fixation on QE; and 2) a fervent desire to somehow explain to the market that altering the mix should not be seen as a signal that the committee is bringing forward the first hike in fed funds (I.e. The traditional short term interest rate.) They clearly want to find a way to present tapering of QE as largely a technical matter, an operational issue of finding the most effective way to deliver the chosen policy stance. The Fed is still traumatized by the market reaction to mere suggestions of tapering back in late spring.
They explicitly state that “trimming” QE could come in the “next few meetings” and may not necessarily be linked to labor market conditions.
So the possibility of change is live, although it's also partly just their usual desire to keep their options open. The minutes imply tapering Is now likely to be delivered as part of a package that will (intentionally) deliver several tweaks at once so a puzzled market is less likely to react in a binary and amplified way. The Fed will throw some sand in the market's eyes to make it less jittery.
I'm troubled by the communication discussion more broadly though, especially efforts by the Fed to claim it will keep fed funds lower for longer in a credible way. An example is Bernanke's emphasis that reaching 6.5% unemployment is not an automatic trigger for a first interest rate rise, and rates may be kept low for a long time after that.
This is in keeping with the “Woodfordian” idea that forward policy guidance can reduce real interest rates now by altering market expectations, even though nominal fed funds is stuck at the zero bound. By promising to keep policy loose for longer in the future, the Fed effectively loosens policy now.
The whole thing is just time inconsistent. It makes sense for the Fed to overemphasize its dovish inclinations at time A, in order to loosen monetary policy. But if it works and recovery is in full swing at time B, will the Fed keep its commitment to low rates when the market is screaming it is “behind the curve”?
Almost certainly not. You could almost guarantee an FOMC member would give a speech with Keynes' (alleged ) line: ” when the facts change, I change my mind. What do you do, sir?”
To put it in non-monetary terms, it is like governments promising to never negotiate with hostage takers – which is clearly the right policy at time A. But once you have an actual hostage situation at time B, with tearful relatives on tv, there is almost irresistible pressure to negotiate. The hostage takers know this at time A, too.
In practice, the more dovish in the short-term, the more chance of a steeper , more dramatic “change of mind” later. It is just hard to see any “commitment technology” that would realistically make the Fed stick to a past commitment rather than current best discretionary policy. Other central banks have tried things like long-term repo transactions that would entail a loss if they altered their commitment. But losses on such a transaction would mean nothing compared to choosing the best policy for a $15 trillion economy.
The Fed cannot credibly bind its future self. Past promises would be quickly forgotten.
The ECB surprised almost the whole market with its rate cut on Thursday. Just 3 out of 70 economists in the Bloomberg survey thought a move would come this week, although many thought December was a possibility.
I don't follow the ECB so closely in recent years, so I was detached and agnostic. But it's striking that all three of the majors -BoJ, Fed, ECB – have now massively surprised the market with rate moves this year.
It almost seems as if the more transparent central banks get, the more misunderstood they are. Or, alternatively, it simply shows that behind the scenes, making the sausage isn't a very clean or orderly affair. It's a transparent window into a very messy reality.
For all the talk of policy rules and effort devoted to forecasts, central bank decisions are still decisions, influenced by how issues get framed, shifting pros and cons and lack of full knowledge of the economy. Central banks muddle through rather than optimize.
It does show continuing jitters about deflation risk among the central banks, however. If you step back and think about it, we have now had three massive deflationary shocks in the last 5-10 years, which are still working their way through the system.
We have had 1) a historically large financial crisis, of course.
We also have had 2) a structural shock with the rise of a whole new general-purpose technology, the Internet. Of course, the dot.com bubble was the peak of excitement about new business models. But it usually takes much longer for technological change to affect old business models. Institutions, conventions, niches take time, a decade or more, to change. It's only the last five years that the Internet has fully hit journalism and legal services, for example. Barriers to entry get eroded, competition increases and whole sectors find themselves subject to downward pressure on pricing.
Technology also affects the meaning of slack or capacity. They become much more elastic. If you're a small business looking to expand, hit a switch to Amazon Web Services and you've got web capacity that the NSA wouldn't disdain, for example. You can rent what you need quickly rather than have to build or hire or learn whole new facilities from scratch.
This means that trying to calculate potential output – the core but unobservable concept in most central bank models – is getting substantially more difficult.
Finally, the third shock is 3) the lingering impact of doubling world labor supply with the effective addition of China, India and other EM countries to the world economy, another development that can take a decade or more to fully feed through.
So it's not surprising that signs of disinflation can make central bankers jumpy. The downside tail risks are still severe. It takes a long time for even flexible economies to adjust to such disruption.
But economies do adjust, and then extreme accommodation could also be an extreme mistake.
I've been thinking about yesterday's FOMC minutes, which try to explain why they shocked the markets by not tapering QE on September 18.
The discussion is unusually long and somewhat muddled, but I think this is the most important bit:
Moreover, the announcement of a reduction in asset purchases at this meeting might trigger an additional, unwarranted tight- ening of financial conditions, perhaps because markets would read such an announcement as signaling the Committee’s willingness, notwithstanding mixed recent data, to take an initial step toward exit from its highly accommodative policy. As a result of such concerns, a number of participants thought that risk-management considerations called for a cautious approach and that, in light of the ambiguous cast of recent readings on the economy, it would be prudent to await further evidence of progress before reducing the pace of asset purchases.
They are still a little traumatized by the sharp market reaction in May and June, which they still struggle to understand. Even if the markets were expecting a taper, the committee didn't want to risk a big reaction last month, at least without further efforts to decouple the idea of tapering QE from raising rates.
I had thought the FOMC would take a small first step to minimize the risk of such a major disruption, and if they did not take a cautious step, then the disruption would be much larger when they did eventually move. The first move is inherently dangerous, but you were never going to get better circumstances for it than having the market calmly and confidently expecting a small taper.
But they felt that they were risking too much of a major disruption if they began tapering, given the events of the spring, and were too nervous to risk it when other data was still mixed. They worried given the mixed data they could be misread as more aggressive. It's a judgment about market behavior.
What the committee can't understand, or are unwilling to admit here, is that there are limits to the effectiveness of their forward guidance, and they have undermined much of the effectiveness it did have. Markets do not adjust smoothly. There is a risk of a “door-shut panic” when the direction of policy turns, regardless of what beliefs are held about the Fed's timing for raising Fed funds. The best thing the Fed can do is handle the market in a way which doesn't make a dramatic rise in long-term yields (and hence mortgage rates) more likely.
Clearly the Fed decision not to taper came as a surprise to almost everybody in the market, including me. I had thought the data was too mixed to justify moving in itself. But if the Fed had the market calmly expecting as taper, the advantages of avoiding market disruption with a tiny move were worth it.
Instead, the Fed focused on the mixed data, and ignored the market expectation. Why did they make that judgment?
In normal circumstances, the Fed pays attention to market expectations and tries to shape them, but sees it as at best one factor. “Markets go up, markets go down”, they typically say with a shrug. It's their job to focus on the real economy instead of worrying about Wall St. Other central banks are even less inhibited about surprising the markets.
So was this a case of the blind spot of the market thinking it was more important that it in fact was, or not enough “outside view” of how typically the Fed pays attention to markets? Maybe.
Instead, I think it's evidently at root a different judgment of the risks of market disruption, compared to the lingering risks of economic weakness. Why should market expectations be more important than mixed data in this case?
Tiger or kitten
It is easy to dismiss claims of hypothetical negative market reactions, or point out “bond market vigilantes” have not reacted to massive government fiscal deficits and historically loose monetary policy since 2008. This is the Paul Krugman view, for example. The bond market is a kitten, that must defer to the views of experts such as himself.
But we are at a historically dangerous point with bond markets. We are at the end of thirty years of generally falling interest rates, which of course raises the price of bonds. Rising yields means losses on bond market portfolios. That means there is more risk that people will dump them , driving yields up even further.
And this comes at a time when there is chronic oversupply of government debt, because of the stimulus programs and deficits of recent years. That hasn't led to a steep rise in yields in large part because QE means the Fed has been buying up a large proportion of the total US bond market.
Banks, especially in Europe, have also been stuffed to obesity levels with their government debt, so much so that a steep rise in yields will cause problems in the banking system.
Fortunately, inflation is not a worry for now (inflation is disastrous for fixed -rate bonds). But the Fed is so far away from normal interest rates that any sign of strong growth or inflation could make people dump bonds.
In other words, the bond market is more stretched than it has been in decades, perhaps ever. A historically vast bubble could pop at some point.
The bond markets are more a sleeping tiger than a kitten, and it could suddenly wake up with a very cranky sore head. The Fed just opened the cage door, and made it a very confused and angry beast.
The US is not Greece, or Italy. Solvency is not an issue. But markets often gap , and behave in very binary risk-on, risk-off ways, as I argued here. A snowball can turn into an avalanche. Greenspan liked to point out how markets adjusted 400bp over a few months in 1979.
The Fed actually can't do much about unemployment directly. The chain of the transmission mechanism is very long and tenuous, especially when credit is gummed up. But it affects bond markets instantaneously. The 10-year plunged 12 bp at 2pm yesterday. Equities and emerging markets soared, on signs of looser policy.
The chance of a major financial disaster also rose significantly yesterday. I had thought the Fed would be more sensitive to that.
That's why I thought even if the data was mixed, a reduction of $10 bn or so in tapering would take much of the risk of bond market disruption off the table for now. A difference of $10bn is far too small to make any perceptible difference to unemployment. But it would have made an enormous perceptual difference in bond markets, reducing the chance of sharp rises in yields which would do disproportionate damage to the real economy. And the Fed could then have waited several months before moving again, if it wished to respond to labor market worries.
Remember, markets gap. They ignore risks and then go slightly nuts later.
The Fed has already succeeded today. The market is centering around expectations of a $10bn taper, i.e a cautious first step in reducing the amount of additional monetary stimulus. And despite that expectation, the S&P is sitting at 1704. The fears in late spring that equities were on a pure sugar rush high because of the Fed – and so would plunge when tapering came – seem to have washed out of the market. The economic data are still mixed, but more people are prepared to believe gradual recovery is still on track.
That means the vaccination strategy for financial markets earlier this year seems to have worked. The patient tossed and turned a bit n May and June, causing some alarm, but now appears to be more resistant to full scale cardiac arrest. Tapering is much less likely to cause violent market seizures.
I’d agree with the consensus of $10-15bn of tapering, concentrated in Treasuries. The Fed would be taking much more of a risk by surprising the market, and so will be inclined to validate the consensus. One little model I’ve been building suggests there’s more chance of an upside surprise, perhaps on another factor like misunderstanding Fed intentions about the first actual rate rise down the road, but it’s still being calibrated and I think caution will prevail.
Bernanke will want to reinforce the message that tapering does not entail an automatic desire to raise Fed funds any time soon. One option that has been discussed is to lower the unemployment threshold for tightening, although that seems very clunky and restrictive. If markets seem to be currently taking things calmly, they may defer it until there is more evidence such a crude indicator is needed.
Despite the appearance that tapering is priced in and firmly expected, it pays to be cautious and alert for financial dislocation or instability in the next few days and weeks. Things seem calm., but this is still a major turning point for the bond markets, a dangerous moment, and any technical problems could ramify into something more serious. The biggest risk could be overseas. The German election will be over soon, which could focus market attention on Europe again at time when an upward drift in US bond yields could complicate issues.
So the payrolls number came in fractionally below expectations (169k compared to 180k) and the revisions were not good. However, the markets always get excited about things which are essentially within the margin of error for this number. For the establishment survey, the number is plus or minus 90,000 with a 90% confidence interval. The total civilian labor force (household survey) is 155,486,000 this month, so a miss of circa 11,000 is a sliver of a sliver of a sliver of a very large number.
The Fed looks at it and takes it seriously, sure. It’s the most important number they have. But they are not quite as binary about it as the markets. They look at a range of data, and how it changes over a period of months, to avoid getting blindsided by any one number.
So what does it mean for tapering? I still think the economic data is a little too weak to justify tapering this month, if that was the only consideration. But it isn’t. The Fed also wants to minimize potential instability and volatility when it tapers. $10bn off bond purchases each month will make an infinitesimal difference to monetary accommodation for the real economy. But timing the first move correctly could stop financial markets overreacting, driving up long rates disproportionately, and damaging the recovery.
Tapering isn’t guaranteed on September 18. But the burden of proof will be in favor of getting the first step out of the way, to reduce the chance of major volatility later. As of today, I’d still incline to think they’ll taper, so long as the markets do not swing so far as to make it a shock which causes instability.
Bernanke is doing his best to assure the markets that discussion of adjustments to quantitative easing (QE) does not mean the Fed is hawkish, or inclined to raise fed funds in the next two years. There are two main points from his NBER Q&A and the minutes yesterday.
First, making short-term decisions about QE is a separate matter from eventual tightening. Scaling back bond purchases – reducing the degree of additional accommodation – should not be read as a sign the Fed will actually affirmatively tighten monetary policy any time soon. It is a matter of having less of a lead foot on the accelerator, not actually hitting the brakes.
Second, Bernanke and the FOMC are insistent that policy is still conditional on the state of the economy. Unemployment rate thresholds are just indicative of FOMC thinking, not strict commitments or requirements for Fed action. So small adjustments in current policy should not be taken to entail a massive shift in the planned path of policy path.
The committee clearly believes markets have overreacted at the long end to their earlier talk of adjusting QE before the end of the year, and are a little anxious about it. So the focus is separation anxiety, to emphasize the Fed is not going to tighten prematurely as a separate matter from QE.
However, it's the thread of discussion about deflation which is most interesting.
I am particularly sensitive to talk about deflation in recent speeches and minutes. It is more important than even labor market data because it represents such an important tail risk. And there is a tick-up in deflation worry in the minutes yesterday. Inflation expectations have been stable, as the minutes note, but any sign of drift there will get a near-panic dovish reaction from the FOMC.
That increased tone of concern about falling inflation, together with an increased risk of turbulence in Europe and China, makes me a little less certain that September will see the first adjustment of the pace of QE purchases. I'd still say it is most likely in terms of timing if payrolls continue their moderately good recent pace. The Fed still needs to defuse the potential for a bond market crash, and the longer it delays the trickier it could be. That still argues for a small token change in policy. But for the same reason it would take very little for the FOMC to defer token action for one or two meetings this fall.
So markets have stabilized for now after a rough ten days, ever since Bernanke suggested in the press conference that Quantitative Easing (the Fed’s continuing purchases of $85bn of bonds each month) would not continue indefinitely.
There have been some casualties. Emerging markets have been hit hard. Bond funds are seeing record outflows.
Nonetheless, it has all gone as well as could be expected so far. Think of it as the Fed letting some air out of the balloon of market obsession over QE. It is gingerly doing its best to make sure the balloon does not get pricked as if by a blunt needle and burst in one spectacular meltdown.
A series of rough and turbulent weeks to let the air out of the balloon is far better than seeing yields rise 200bp in a matter of a week or two in an abrupt way that could kill the housing market and the economy again.
What we have been seeing so far is just the downside of the upside. The broad outlook is good. The US economy is recovering, even if slowly and unevenly.
The natural consequence of recovery is long-term yields will rise from their historically low recent levels. At some point in 2015 or later, the Fed will raise short rates.
Those are all desirable things. But to get back to a normal, flourishing economy we have to deal with overextended asset markets and falling bond prices along the way. To throw in another metaphor, to get healthy you might have to endure some nasty days at the gym along the way. Adjustment in markets is not going to be smooth, especially as a thirty-year bull market in bonds reverses.
The door shrinks as well
Things could still go wrong. The biggest danger will come when the Fed actually makes its first tangible move, locking in timing and forcing market adjustment. That is the point of maximum risk of a “door-shut panic” , as Charles Kindleberger called it. A huge crows of people tries to rush for the small exit door at once. Some of them get crushed. Markets freeze. Liquidity disappears.
And someone reminded me this morning of a wise saying by my former boss, Richard Medley. “As people go through the door, the door shrinks, as well.”
I still think September is the most likely point for a first move by the Fed. Better in their view to go small and early rather than wait and have to go big and late, especially if you stress that the move is tentative and reversible. And the more you can talk about gradually reducing QE purchases, the more air you can hopefully let out of the balloon.
However, it will cost the Fed very little to wait a few meetings if other negative shocks hit. China looks fragile. Egypt and Syria raise the risk of serious turbulence in the Middle East. The US municipal market is vulnerable to event risk. There is still a lingering if small risk of deflation.
International shocks are more likely to derail timing than big surprises in the domestic economy. I don’t put a great deal of faith in automatic triggers for Fed actions at particular levels of US unemployment. The Fed will still make decisions on the basis of all the information available, and those thresholds should be seen as rough calibrations of when the Fed would consider conditions right, rather than a specific commitment or promise.
For now, though, the outlook is tilted towards the upside. We just have to cope with the downside of the upside.