Why is the yield curve flattening?
Kevin Erdmann has an interesting post discussing the evolution of interest rate futures over time. In recent months the date of the first anticipated rise in rates has moved from late-2015 to mid-2015.
But, I expected this to correspond to forward rates for June 2016 Eurodollars of just over 2% and for June 2017 of just over 3%. Instead, rates have trended around a mean of about 1 5/8% and 2 5/8%. If that had been my target mean, it would have been perfect for this trade. (Basically, buying when the price declines and selling when it increases, profiting from random movements over time.) But, the farther the price moves from my target, the harder it is to profit from the position. Why haven’t the prices followed the market expectation?
The reason is that the slope of the yield curve has declined while the expected date of the rise has moved back. Late in 2013, the slope was up to around 33 bp. (Rates would be expected to rise 33 basis points per quarter after the initial rise.) For reference, in the past two cycles, during the rate recovery period, short term rates rose at a pace of 50 to 75 bp per quarter.
So, what’s going on? I think that the market has been surprised by how healthy the economy has remained in the face of the tapering of QE3. This diminishes inflation fears and also signals a hawkish intention from the Fed.
But, I think this reflects the Wizard of Oz view of the Fed’s interest rate policy. I think that there is a systematic underestimation of how much Fed policy chases the Wickesellian interest rate. I think the Wickesellian rate is probably already above zero. This is part of the reason that we have seen an acceleration in economic activity and employment this year. QE3 appears to have been only slightly accommodative, for reasons I don’t completely understand, and so its taper has probably not changed the objective stance of monetary policy that much. But, before QE3, a non-QE zero rate policy was probably disinflationary. The increase in the Wickesellian rate over the past 2 years means that at the end of QE3, a non-QE zero rate policy will probably be inflationary.
This might be right, but let me throw out another possible explanation. Perhaps the markets are reacting to the disconnect between the unemployment data and the GDP data. All throughout the recovery the unemployment rate has done better than the GDP data. Even back as far as 2011 I was doing posts entitled “A job-filled non-recovery.” But the disconnect has recently gotten worse, with NGDP growth coming in at under 4% over the last 6 quarters, and the unemployment rate falling by 1.7% over 18 months, roughly 0.1% per month. That’s a very fast decline in the unemployment rate, which suggests we’ll hit full employment fairly soon. But GDP growth has been really slow.
This pattern has gotten even more pronounced over the past 6 months. You can write off the first quarter due to bad weather, but weather certainly did not affect second quarter GDP. And yet NGDP in Q2 was up only 2.5% (annual rate) over 2013:4. That’s a shockingly low rate of NGDP growth, and yet the rapid decline in the unemployment rate continued. And we recently had the fastest 6 months of job creation since the late 1990s. One thing we know for sure is that job growth must slow at some point. Labor force growth will be exceedingly slow due to retiring boomers, and returning discouraged workers can only do so much. I’ve suggested 3% is the new normal for NGDP, but if you applied Okun’s Law to the recent data (and account for inflation) you’d get even a lower estimate.
So if you look at the unemployment data and the NGDP data together, rather than separately, you are forced to conclude that trend NGDP growth has slowed very sharply. If the markets are gradually figuring this out as more data comes in, it could explain the flattening of the yield curve. The rapid fall in unemployment explains why a rate increase is expected within less than a year, and the declining estimates of trend NGDP growth explain why rates are not expected to rise as far after that first rate increase.
PS. Kevin Erdmann also has a very good post linking school choice, banking regulation, and the right to exit. Highly recommended.
HT: TravisV
Tags:
9. August 2014 at 06:14
Most likely because of tapering http://static1.businessinsider.com/image/53d26d096bb3f7cf4817ada8-1200/hedy-mansour-langdon-p-cook-government-securities.jpg
9. August 2014 at 06:20
Ed, Wasn’t that priced in 8 months ago?
9. August 2014 at 06:29
Scott, there is huge resistance to the idea which only erodes slowly. Even today people explain QE as a way of lowering interest rates despite the overwhelming evidence of the past few years.
9. August 2014 at 06:57
Prof. Sumner, can I ask you an unrelated question? If, instead of targeting NGDP growth, the Fed started targeting S&P500 growth, would this be equivalent in your view? Thanks
9. August 2014 at 07:15
Prof. Sumner,
Thanks! I think you might also like this recent post by Erdmann:
“What a disappearing middle class doesn’t look like”
http://idiosyncraticwhisk.blogspot.com/2014/08/what-disappearing-middle-class-doesnt.html
9. August 2014 at 07:29
Any opinion on whether declining marginal productivity may be the culprit for suppressed NGDP (and RGDP BTW) growth?
9. August 2014 at 08:39
ssumner wrote:
Bingo! The latest counterfactual we can offer for what NGDP would look like in the absence of QE suggests that it is still what is making the difference between expansion and contraction in the first half of 2014. (That wasn’t the case in the second half of 2013, but there appears to have been unique circumstances that contributed to stronger organic economic performance at the time.)
Since the end of 2013, we’ve also seen something of a diverging trend for total employment and establishment (nonfarm payroll) employment (here’s the data nationally, and specifically for Illinois, which we’ve been tracking following the expiration of federal extended unemployment insurance benefits since the state collects a unique dataset for those affected by the program’s end). The divergence shows increasing total employment as there has been a dramatic increase in the number of marginal jobs being filled but falling establishment employment, which is consistent with a weak economic performance (at least according to this very recent research).
Finally, if short term rates rise as long term rates fall, that also suggests an increased likelihood of weaker economic performance in the future.
9. August 2014 at 12:06
Good points, Scott.
I have updated my post with a few graphs comparing employment growth and real GDP growth.
9. August 2014 at 12:13
And still the theory that real, healthy growth is impaired because of inflation, and not because it wasn’t large enough, continues to be avoided/ignored.
9. August 2014 at 12:18
Because that theory is insane
9. August 2014 at 13:06
Edward:
No it isn’t. It is very much sane.
You’re avoiding/ignoring it as well.
9. August 2014 at 15:43
Nope. I’ve examined it and I think it’s crazy.
9. August 2014 at 16:35
Scott – look at the hours worked full and part-time workers Fred St Louis series, now at about 230 thousand million hours per year—the same level of hours worked as in the year 2000!
Sure, some demographics going on, but mostly weak aggregate demand. 12 million Americans collecting “disability” payments from SSDI or VA.
Crank down on “disability” and crank up the presses!
Re Erdman: Remember supply and demand in capital markets. The supply today exceeds demand.
9. August 2014 at 17:15
Ed, Maybe.
Maurizio, No, bad idea.
TravisA, I’ll use that one later.
Dustin, It’s partly productivity, partly slow labor force growth.
Ironman, Interesting graphs.
Kevin, I focus more on the unemployment rate than employment growth, when considering what sort of NGDP growth is sustainable.
Ben, Yes, that’s a puzzle, but how much can be solved with more NGDP? Part of it, yes, but how much?
9. August 2014 at 17:35
Prof. Sumner,
I’m TravisV. 🙂
9. August 2014 at 17:42
Jeff Gundlach Warned Us Rates Could Fall, And They Did “” Here’s What He’s Saying Now
“”It’s really hard for me to identify why rates should go higher,” said Jeffrey Gundlach of DoubleLine Funds.
In a phone call with Business Insider, Gundlach reiterated his expectation for the 10-year yield to trade between 2.2% and 2.8%, with the risk that it goes below 2.2%.
Most of Wall Street’s interest rate gurus came into 2014 confident that rates would actually rise as the U.S. economy was expected to reach escape velocity. In January, the consensus year-end forecast for the 10-year yield was 3.4%.”
http://www.businessinsider.com/jeffrey-gundlach-on-rates-2014-8
9. August 2014 at 17:51
TravisV, Have you considered changing your last name to make it easier for me to remember? 🙂
I fixed the HT
9. August 2014 at 19:51
– The yield curve has been flattening since early 2013 and NOT only since the beginning of 2014.
– There’s no relationship between a weakening NGDP and a flattening yield curve.
– I still expect long term rates to go higher. But NOT as a result of “the economy reaching escape velocity”. But short term rates will\could go lower for a bit. In other words, the yield curve will steepen again. Actually when I look at the yield curve right now it seems it’s in the process of turning to steepening.
9. August 2014 at 20:10
Short term rates WILL be raised in an attempt to stem the outflow of money from the US. Not because the economy is recovering.
9. August 2014 at 20:13
Benjamin, I thought you made a good point about hours worked, so I added a graph at my original post. I was surprised to find that it didn’t strengthen the correlation between employment and real GDP growth, and current GDP is still lower than the linear correlation would predict. Although, in either case it’s not outside the typical range.
Scott, I don’t understand what you mean by this:
” I focus more on the unemployment rate than employment growth, when considering what sort of NGDP growth is sustainable.”
9. August 2014 at 20:23
TravisV, I replied at my blog. Short version: The yield curve is unusual now, so I prefer looking at forward rate contracts. The 10 year rate could remain low even if short term rates increase faster than currently expected. 5 year rates have been much firmer than 10 year rates, for instance.
9. August 2014 at 23:13
O/T: Why epidemics will happen less and less often. http://reason.com/archives/2014/08/08/ebola-dress-rehearsal-or-show-closer
9. August 2014 at 23:41
Noah Smith has an interesting new post on Wallace Neutrality: http://noahpinionblog.blogspot.jp/2014/08/can-fed-set-interest-rates.html
(The guy on the right of the header image always wins in the end, so he’s the central bank I guess. The guy on the left is my favorite though – is this going to make me start sympathizing with the Treasury???)
10. August 2014 at 00:45
For anybody who wonders what I am talking about, it is this FRED chart:
http://research.stlouisfed.org/fred2/series/B4701C0A222NBEA
It shows workers in the U.S. are working less hours today (cumulative and total) than in 2007, and about the same as in 2000.
Americans work now about 230,229 million hours a year, down from 227,100 or so in 2007.
So, in the United States, in terms of total work hours per year, we are at a lower level than in 2007—seven years ago.
Scott Sumner asked me why this is, and mostly I do not know.
Of course, productivity is part of it, and worker disincentives, such as food stamps, unemployment insurance and the 12 million people collecting “disability” from SSDI and the VA.
My gut tells me all of the disincentives existed before, although maybe people learned how to exploit them, not having jobs.
I come back to it: The Fed is monetarily asphyxiating the economy, reducing hours demanded.
Speaking anecdotally, I have worked in many small businesses, and ran two. When you get orders, then you do what you can to fill them. Usually, a small business never turns away business. I think it is a fiction that businesses are sitting on their hands, due to certain disincentives (I dislike Obamacare, but it doesn’t even apply to businesses with less than 50 employees). Businesses are improving productivity as they always have, and meeting demand.
It is demand that is weak.
No doubt some people are retiring. But if demand were rising, we would see remaining workers put in more hours, or some workers drawn back into the labor pool.
10. August 2014 at 04:14
I don’t think this explains much of the move in long term rates, but it’s worth pointing out that six months, or a year ago, many people seemed to think it was likely that we would get immigration reform this summer. In fact, six months ago, one might have looked at the increasing numbers of child migrants to the US and concluded that immigration to the US was going to increase over the next decade. Now it’s looking like more people coming here in the short term might have the opposite effect over the medium term.
Personally, I always thought the hopes of immigration reform this year were pie in the sky, but many seemed to disagree … And what has happened certainly made me more pessimistic for our chances to do it any time in immediate future.
10. August 2014 at 06:45
Willy, You said;
“There’s no relationship between a weakening NGDP and a flattening yield curve.”
That wasn’t my claim, I said lower NGDP growth expectations might be reducing long term rates.
Kevin, The relationship between employment and GDP tells us something about productivity. But productivity growth doesn’t determine sustainable growth in GDP, it’s productivity growth plus labor force growth. Comparing the unemployment rate and NGDP growth is a better way of figuring out what sort of growth is sustainable.
Nick, Yes, immigration could have a impact on NGDP.
10. August 2014 at 06:51
Saturos, Yes, that’s a good post.
10. August 2014 at 07:41
“That wasn’t my claim, I said lower NGDP growth expectations might be reducing long term rates.”
Seriously, who in the bond markets integrates NGDP in their pricing models? The dominant “nominal” factor is consumer price changes, not total spending.
Is it possible that maybe what you think is reducing long term rates is reallu just what your chosen model states, and not what bond market participants actually think?
10. August 2014 at 08:11
Benjamin Cole:
Demand is not weak. The desire to have more wealth and consume more and higher quality goods and services is practically infinite.
The Fed is not “asphyxiating” the economy. There is more than enough money and spending to buy every single good produced and every single labor hour offered – goods and labor types just have to change, and their prices just have to fall without limit. Once this occurs, or as we get closer to it, every laborer and good will tend to find able and willing buyers.
More inflation cannot solve this problem, because more inflation will prevent prices from falling, and more importantly it will hamper the market’s ability to change goods and labor types, to the extent that the reason for the need for changes is due to previous inflation.
Your theory blinds you to the counter-productive consequences of the very activity you believe is the cure. The market does not want more toilet paper as a store of value and means of exchange. If it did, then it would have created it itself. The market is desperately trying to completely eradicate the government’s fiat currency, and the symptoms of this white blood cell eradication is what you percieve as a lack of fiat currency.
The central bank is not a free market institution. Free markets and central banks are fundamentally antagonistic and incompatible. Ignoring the market’s weakening caused by inflation, and pretending that the incompatibility can be solved by just changing one non-market absolutist rule for another, is without a doubt a quest of the absurd. It is a desire for destruction and blaming the victim when it doesn’t react the way you want it to react.
The Soviet Union was an 80 year long economic depression. Market forces, i.e. individual desires constrained to homesteading and trade, were for 80 years trying to eradicate the government’s food, clothing, textiles, stores, farms, and mines. Just like you believe right the hell now with government toilet paper money, the communists perceived the widespread economic problems as a result of insufficient government activity, and not because of their activity itself.
You are a faith based zealot whose religion is monetary socialism. You choose to not comprehend the very destructive consequences of what you incorrectly perceive to be insufficient.
The market wants and needs a market based money. A market based money will solve so many of the problems you have not the foggiest idea is the cause because to understand them is to admit you were wrong the whole time. Since you can’t be wrong, the problem cannot be the legalized counterfeiting operation itself. It has to be the wrong number of times CTRL-P is pressed. More is double plus good.
Excellent blogging(tm)
10. August 2014 at 08:13
Benjamin,
Thanks for the Fed chart. Jobs or no, people are still expected to be responsible for the fact they are alive. Some responsibility “nudges” are more coercive than others.
10. August 2014 at 08:27
Seriously, who in the bond markets integrates NGDP in their pricing models?
Everyone. That’s like asking “how many consumers integrate inflation expectations into their spending decisions?” They all do, even if they don’t think of it in precisely those terms.
10. August 2014 at 09:01
Talldave:
Consumer price inflation expectations are not NGDP expectations. They can move differently, and even in opposite directions, depending on productivity.
Saying everyone integrates NGDP into their bond pricing models “even if they don’t think in precisely that term” is really just refusing to admit that bond traders don’t actually integrate NGDP into their pricing models but pretending they do by claiming they are thinking of something they are not.
I integrate consumer price inflation into my spending decisions, and I do think in that term. Inflation of the money supply does not affect my income and the prices of goods I buy equally in terms of rate of change, or percent change. I am not better off if goods prices rise faster than my income. I would rather you be unemployed because your asking wage is far too high, than my purchasing power being reduced by the inflation which might enable your asking wage to find a willing buyer because their income was raised before mine due to inflation.
I have never seen any bond pricing model that contains NGDP as a factor. If you could show me any university/college textbook, or course material, or even a white paper from a finance or economics PhD that hasn’t even been accepted in any journal, that has NGDP as a factor in a bond pricing model, then I’ll take your claim to be something other than wishful thinking.
10. August 2014 at 15:19
Scott, do you think the yield curve will invert before the next recession as it has ahead of every downturn since the mid 1950s?
10. August 2014 at 15:51
Major Freedom:
Egads, how do explain the tremendous American prosperity of the 1960s, or of the 1982 to 2007 period? Good enough for me—print more toilet paper!
10. August 2014 at 19:55
Is China’s monetary policy easing? See new story below:
China Loosens Monetary Conditions in Test of Credit Power
“China loosened monetary conditions last quarter at the fastest pace in almost two years, a Bloomberg LP gauge showed, testing the waning effectiveness of credit in supporting economic growth.”
http://www.bloomberg.com/news/2014-08-10/china-loosens-monetary-conditions-in-test-of-credit-power.html
10. August 2014 at 22:26
“”There’s no relationship between a weakening NGDP and a flattening yield curve.””
“That wasn’t my claim, I said lower NGDP growth expectations might be reducing long term rates.”
OK. But “Growth Expectations” don’t drive long term interest rates. It’s merely a matter of demand & supply that determines rates, NOT expectations.
It’s the amount of risk people want to take on. That determines how steep the yield curve is.
11. August 2014 at 02:40
Ironman: It’s sticky wages, of course. The Non-Farm Payroll has more sticky wages than Total Employment, so you would expect a bigger drop in payrolls than total employment as Aggregate Demand falls. A great way to show it.
11. August 2014 at 07:30
Tommy, I have no idea, but I will say the yield curve is not as good at forecasting recessions as people assume. It gives some false positives, and misses some recessions.
Willy, You said:
“OK. But “Growth Expectations” don’t drive long term interest rates. It’s merely a matter of demand & supply that determines rates, NOT expectations.”
Is this supposed to be a joke? Of course it’s supply and demand, which are determined by expectations.
11. August 2014 at 08:34
I don’t think the unemployment rate is as meaningful as it once was – along the lines of capacity utilization rates. If you look at total hours worked (full and part time) and divide into population, you find that we are working now at barely above the rate back to 1980 when the % of women in the workforce was substantially lower than it is today
11. August 2014 at 13:28
Why is the Treasury yield curve flattening?
How about Ukraine, Russia, Syria, Iraq, ISIS, renewed US bombing, Gaza, Israel, Egypt, Libya, with some South China Sea just for flavoring.
US Treasurys and German Bunds are the flight to safety sovereign bonds, and both curves have flattened as risks have increased.
An important lesson is interest rates are not driven strictly by domestic considerations, including NGDP.
Also, bouts of global weakness. Check GDP growth in Italy, Ukraine, Russia, and Japan. Disinflation in many countries. Bias toward easy policy in many central banks.
Plenty of reasons for low safe bond yields, even if NGDP growth is rising in the USA.
11. August 2014 at 14:10
US Yield curve has gone from expecting a “V” shaped recovery to expecting an “L shaped” recovery.
If the yield curve is arbitrage neutral, then the forward curve represents a probability weighted average of all possible paths that fed funds may take. Currently it is pricing in the most likely date for the first hike around June of 2015. But it is not pricing the second hike until December of 2015.
This is suggesting only weak confidence that the fed actually starts tightening by June.
But, if you look at Real GDP growth, NGDP growth, Capacity utilization, and Labor force participation, all seem to be agreeing that we are in this “muddling through” rate of economic expansion. Yes, there is growth, but it is pretty anemic. The yield curve seem to be predicting growth a 2% RGDP pace for the foreseeable future.
In 2009 and 2010 there was an expectation that the economy would “snap back” and we would get several quarters above 4% before settling into a longer term growth trend in the 2 1/2 to 3 range… and then that didn’t come to be and the yeild curve flattened. It steepened last year, as markets seemed hopeful that the recovery would sustain something a little faster, and that hope is fading.
12. August 2014 at 05:25
BB. That doesn’t quite work. While yields spreads have fallen, stocks have risen sharply over the past 12 months. In any case, what makes you think those factors would not reduce expected NGDP growth?
Doug, I mostly agree. I would add that markets expect 2% RGDP growth as long as the unemployment rate is falling. When it hits bottom (around 5%) growth will slow to 1.2%.
14. August 2014 at 04:44
Prof. Sumner, you suggest that growth will slow once the employment rate hits 5%. Wouldn’t improved employment numbers and higher growth go in tandem?
14. August 2014 at 07:27
No, IMPROVING numbers, not improved numbers. Growth is fast when unemployment is falling. When it stops falling (at 5%) growth slows.