Friday, June 29, 2018

May personal spending returns to typical late cycle pattern

 - by New Deal democrat

The consumer continues to do okay. That is the message from personal income and spending as reported for May this morning.

First of all, let's compare the YoY% growth in real personal spending (blue) with real retail sales (red):

For the last 50 years, during all business cycles but one, retail sales rose faster than, and ran ahead of, spending during the first part of the cycle, and declined below spending in the latter part of the cycle, so much so that the comparison is a good mid-cycle indicator. The crossover point occurred in 2014 during this cycle, so it is an excellent bet that we are in late cycle.

But wait! Real retail sales have recently been rising again YoY, and last month surpassed spending.  Have we reset? I very much doubt it.

First of all, let's compare the two metrics m/m since the beginning of 2017:

In general, the growth occurred late last year after the hurricanes and California wildfires, and was likely in response to those events. This year so far, real spending has averaged only +0.1% per month, and most of that in March -- which may be related to the recent withholding tax cuts. Retail sales for May look like something of an outlier. So the likelihood is the recent YoY uptick is transitory.

Also, let's again compare spending (blue) with the personal saving rate (red):

A steep decline in the savings rate also tends to be something that happens in mid- to later-cycle. Note the big decline in 2004-05, and again in 2016.

When we zoom in on the last 18 months, we see that the bump higher in spending last autumn was coincident with a dip in the savings rate:

Now both have returned to their previous rates. My suspicion is that this will continue until the last 6 months or so before the next recession, when YoY spending will decelerate, and the savings rate may begin to increase, as has been the case historically.

Thursday, June 28, 2018

The yield curve is already signaling a slowdown

 - by New Deal democrat

Throughout this expansion, I have had a sneaking suspicion that the yield curve (the difference in interest rates between short and long term bonds) would be the indicator most likely to fail.

Originally that was because we are in a very non-inflationary period similar to that which prevailed between the 1920s and 1950s. After 1929, at no time during the 1930s, 1940s, or early 1950s did the yield curve invert, even though there were five recessions during that time, including the very deep 1938 recession.

Now I have a second concern: too many people are paying too much attention to it. As a result, they are "anticipating" an inversion and may be altering their behavior in response, in a way they have not done in the past. Humans are very sneaky primates, and when you observe their behavior, they always observe back.

That being said, it would also be a mistake to ignore such a trustworthy signal in the past. So, as usual, I K.I.S.S. and consider it among the array.

And in that vein, as I write this the spread between 2 and 10 year bonds is down to +.32%. That is narrow enough that, if past is prologue, a signal is being sent.

Here is the graph showing this spread for the last 40 years (blue), wherein I have subtracted -.35% so that the current spread is at the "zero" line, compared with real YoY GDP growth (red):

Needless to say, the spread crossed this line every time on the way to an inversion, which signaled a recession to come.

But on three other occasions -- 1984, 1986, and 1994, the spread got to +.35% or lower without actually inverting.  On those occasions there was no recession.

But there was a slowdown about 1 year later in each instance. YoY GDP growth *decelerated* by 5% from 1983 to 1984; by 1.5% from 1986 to 1987; and by 2% from 1994 to 1995.

So it's a fair bet that there will be a similar deceleration of GDP growth from this quarter at some point in the next 12 months or so, even if there is no yield curve inversion.

Wednesday, June 27, 2018

House prices continue to outpace buyers' incomes

 - by New Deal democrat

Now that the Case Shiller indexes have been reported for May, let's take an updated look at house prices.

The below graph compares average hourly earnings for all workers (including managers)(red) with the Case Shiller 20 city index, and the FHFA's purchase only nationwide index:

As you can see, both house price measures have been rising at roughly 6% YoY for the last two years, while wages have been rising at less than 3% YoY. House prices have been rising faster than wages all the way back in 2012, when prices hit bottom. 

Because the series is so volatile, I am showing the median price for new homes separately compared with wages, below:

The pattern here is the same: prices have continuously increased faster than wages.

The only significant house price measure I am not able to include in graphic form is the N.A.R.'s median sales price for existing homes, which increased 4.9% from $252,500 to $264,800.

Meanwhile, median household income as measured by Sentier Research YoY from April 2017-18 rose 4.6% from $58,771 to $61,483:

This is almost certainly a better metric than wages, because after all, it is a household that purchases a house! Even so, household income is still lagging house price increases.

The bottom line is, even taking out the affect of higher mortgage rates, the price of housing continues to rise faster than the incomes of potential buyers.  There is some evidence, from the decline in the last three months of permit issuance for single family homes, and the deceleration this year in the growth of purchase mortgage applications, that lack of affordability is making a dent in home sales,  but as I pointed out yesterday, that hasn't shown up in the (very volatile, and heavily revised) report of new home sales.

Tuesday, June 26, 2018

New home sales continue very positive

 - by New Deal democrat

New home sales, along with purchase mortgage applications, are the most consistently positive of all of the measures of the housing market. In fact, as I have written many times before, usually new home sales are the very first measure of housing to peak, leading even permits.

The problem with new home sales is that they are very volatile and are among the most violently revised of all data releases.

But, that caveat in mind, once again new home sales in May rose to the highest level during the entire expansion save for one month:

Averaging over three months cuts down on the volatility. Measured that way, they did make a new expansion high, as shown in the below for the last year:

May 607
Jun 604
Jul 592
Aug 577
Sep 584
Oct 603
Nov 656
Dec 655
Jan 661
Feb 644
Mar 656
Apr 660
May 669

Something new I thought I'd look at this month is the comparison between new home sales and mortgage rates. Interest rates have had a very regular relationship with permits, leading them by between 3 to 9 months.

When we do the same comparison with new home sales, measured quarterly to cut down on the volatility, we see that new home sales (which are recorded as of the date the contract is signed) react almost instantly to mortgage rates (inverted in the graph below):

Zooming in on the last 7 years, and measuring monthly, we can continue to make out the two moving more or less simultaneously, but once again the demographic tailwind has been very strong:

Even though mortgage rates have been higher YoY almost consistently for the last 18 months, they haven't had much of an effect on new home sales.

Let's also compare new homes sold with new homes for sale, measured by the same survey. We can see that the former has consistently led the latter, as the number of new houses are offered for sale continues to increase even after sales have peaked:

The amount by which houses sold exceeds houses for sale, measured quarterly to cut down on volatility, if anything has tended to peak near mid-cycle:

Here is the same metric measured monthly over the last 7 years:

The peak may have been in the 4th quarter of last year, since Q1 of this year was below that, and through the first two months, Q2 is even lower.

But the bottom line is, for now new home sales remain very positive, even while permits, as I mentioned last week, may have taken a hit from higher interest rates. I continue to suspect that it will take rates on the order of 5.25% for a sustained period of time like half a year for housing to truly roll over.

Monday, June 25, 2018

Weekly Indicators for June 18 - 22

May data was focused on housing. Housing starts rose to a new expansion high, but the less volatile and more leading permits made an 8 month low, as did single family permits. As a result of the decline in permits, the Index of Leading Indicators rose less than forecast, at +0.2.

My usual note: I look at the high frequency weekly indicators because while they can be very noisy, they provide a good Now-cast of the economy, and will telegraph the maintenance or change in the economy well before monthly or quarterly data is available.  They are also an excellent way to "mark your beliefs to market."
In general I go in order of long leading indicators, then short leading indicators, then coincident indicators.
NOTE that I include 12 month highs and lows in the data in parentheses to the right.

Interest rates and credit spreads
  • BAA corporate bond index 4.85% up +.03% w/w (1 yr range: 4.15 - 4.94)  
  • 10 year treasury bonds 2.90% down -.02% w/w  (2.05 - 3.11) 
  • Credit spread 1.95% up +.03% w/w (1.56 - 2.30)
Yield curve, 10 year minus 2 year:
  • 0.34%, down -0.02% w/w (0.34 - 1.30) (new expansion low)
30 year conventional mortgage rate
  • 4.69%, up +0.05% w/w (3.84 -  4.79)
BAA Corporate bonds remain neutral. If these go above 5%, they will become a negative. Mortgage rates and treasury bonds are still both negatives. The spread between corporate bonds and treasuries has now gone above 1.85%, and so I have switched it from positive to neutral. The only remaining positive is the yield curve, and if that declines to +0.25%, that too will become a neutral.


Mortgage applications  
  • Purchase apps up +4% to 257 w/w 
  • Purchase apps 4 week avg. +3 to 250
  • Purchase apps YoY +3%, 4 week YoY avg. +3%
  • Refi app up +6% w/w
Real Estate loans
  • Unchanged w/w  
  • Up +3.8% YoY ( 3.3 - 6.5) (re-benchmarked, adding roughly +0.5% to prior comparisons) 
Refi has been dead for some time. Purchase applications were strong almost all last year, began to falter YoY in late December, but rebounded during spring, ultimately making new expansion highs. The 4 week average declined from the peak enough to qualify as neutral for the last several weeks, but this week rebounded to positive.

With the re-benchmarking of the last year, the growth rate of real estate loans changed from neutral to positive. If the YoY rate falls below +3.25%, I will downgrade back to neutral.

Money supply
  • -1.4% w/w 
  • +1.5% m/m 
  • -0.4% Real M1 last 6 months
  • +3.4% YoY Real M1 (1.6 - 6.9) 
  • +0.1% w/w  
  • +0.7% m/m 
  • +1.4% YoY Real M2 (0.9 - 4.1)
Since 2010, both real M1 and real M2 were resolutely positive.  Both decelerated substantially in 2017. Real M2 growth has fallen below 2.5% and is thus a negative.  Real M1 growth this week was again below 3.5% YoY, and on a 6 month basis it has returned to a negative, so real M1 overall is scored as neutral.

Credit conditions (from the Chicago Fed) 

  • Financial Conditions Index unchanged at -0.81
  • Adjusted Index (removing background economic conditions) unchanged at -0.52
  • Leverage subindex up +.03 to -0.34
The Chicago Fed's Adjusted Index's real break-even point is roughly -0.25.  In the leverage index, a negative number is good, a positive poor. The historical breakeven point has been -0.5 for the unadjusted Index. All three metrics presently show looseness and so are positives for the economy.
Trade weighted US$
  • Up +1.31 to 123.92 w/w +1.7% YoY (last week) (broad) (116.42 -128.62) 
  • Down -0.25 to 94.54 w/w, -2.84% YoY (yesterday) (major currencies)  
The US$ appreciated about 20% between mid-2014 and mid-2015.  It went mainly sideways afterward until briefly spiking higher after the US presidential election. Both measures had been positives since last summer, but in the last two weeks the broad measure turned neutral.

Commodity prices
  • Down -3.01 to 109.57 w/w
  • Up +8.23 YoY 
BBG Industrial metals ETF 
  • 133.22 down -3.46 w/w, up +20.14% YoY (108.00 - 149.10)
Commodity prices bottomed near the end of 2015. After briefly turning negative, metals also surged higher after the 2016 presidential election.  With the exception of one week ago, ECRI has been neutral for many months.  On the other hand, industrial metals have been strongly positive and recently made a new high.

Stock prices S&P 500
  • Down -0.9% w/w at 2754.88
Stock prices are a neutral, having not made either a new 3 month high or low in the last three months.  They made a string of new all-time highs beginning in summer 2016.
Regional Fed New Orders Indexes
(*indicates report this week) 
  • Empire State up +5.3 to +21.3 
  • *Philly down -22.7 to +17.9 
  • Richmond up +25 to +16
  • Kansas City up +1 to +38
  • Dallas down -0.2 to +27.7
  • Month over month rolling average: down -5 to +24
The regional average has been more volatile than the ISM manufacturing index, but has accurately forecast its month over month direction. After being very positive for most of this year, it has moderated slightly in the last few weeks.

Employment metrics
 Initial jobless claims
  • 218,000 unchanged
  • 4 week average 221,000 down -3,250 
Initial claims have recently made several 40+ year lows and so are very positive. The YoY% change in these metrics had been decelerating but is now back on its multi-year pace. 

  • Unchanged at 97 w/w
  • Up +1.5% YoY
This index was generally neutral from May through December 2016, and then positive with a few exceptions all during 2017. It was negative for over a month at the beginning of this year, but returned to a positive since then.

Tax Withholding 
  • $177.3 B for the last 20 reporting days vs. $181.1 B one year ago, down -$3.8 B or -2.1%
  • 20 day rolling average adjusted for tax cut [+$4 B]: up +$0.2 B or +0.1%
With the exception of the month of August and late November, this was positive for almost all of 2017. It has generally been negative since the effects of the recent tax cuts started in February.

I have discontinued the intramonth metric for the remainder of this year, since the kludge to guesstimate the impact of the recent tax cuts makes it too noisy to be of real use.
I have been adjusting based on Treasury Dept. estimates of a decline of roughly $4 Billion over a 20 day period. Until we have YoY comparisons, we have to take this measure with a big grain of salt.

  • Oil up +$4.85 to $69.18 w/w,  up +61% YoY 
  • Gas prices down -$.03 to $2.88 w/w, up $0.55 YoY 
  • Usage 4 week average up +0.9% YoY 
 The price of gas bottomed over 2 years ago at $1.69.  With the exception of last July, prices generally went sideways with a slight increasing trend in 2017.  Usage turned negative in the first half of 2017, but has almost always been positive since then. Because the YoY change has gone back above 40%, the rating is now negative.

 Bank lending rates
  • 0.452 TED spread down -0.006 w/w 
  • 2.090 LIBOR unchanged w/w (tied for new expansion high)
 Both TED and LIBOR rose in 2016 to the point where both were usually negatives, with lots of fluctuation.  Of importance is that TED was above 0.50 before both the 2001 and 2008 recessions.  The TED spread was generally increasingly positive in 2017, while LIBOR was increasingly negative. This year the TED spread has also turned negative, then turned positive in the past month before returning to negative in the last two weeks.
Consumer spending 
  • Johnson Redbook up +4.7% YoY
 Both the Goldman Sachs and Johnson Redbook Indexes generally improved from weak to moderate or strong positives during 2017 and have remained positive this year.

Railroad transport
  • Carloads up +1.9 YoY
  • Intermodal units up +6.3% YoY
  • Total loads up +4.1% YoY
Shipping transport
  • Harpex unchanged at 678 (440 - 678) (tied for 6 year high)
  • Baltic Dry Index down -27 to 1377 (~700 - 1700)
Rail was generally positive since November 2016 and remained so during all of 2017 with the exception of a period during autumn when it was mixed. After some weakness in January and February this year, rail has returned to positive.
Harpex made multi-year lows in early 2017, then improved, declined again, and then improved  yet again to recent highs. BDI traced a similar trajectory, and made 3 year highs near the end of 2017, declined early this year, but recently has hit multiyear highs.
I am wary of reading too much into price indexes like this, since they are heavily influenced by supply (as in, a huge overbuilding of ships in the last decade) as well as demand.
Steel production 
  • Down -0.7% w/w
  • Down -0.4% YoY
Steel production improved from negative to "less bad" to positive in 2016 and with the exception of early summer, remained generally positive in 2017. It turned negative in January and early February, but with the exception of three weeks (including the last two weeks) has been positive since then. 

Among the long leading indicators, the more leading Chicago Fed Financial Conditions Indexes and real estate loans remained positive, rejoined this week by purchase mortgage applications. Corporate bonds and mortgage rates are neutral, rejoined this week by real M1. Treasuries, refinance applications, and real M2 are all negative. That the most leading monthly housing data for May declined to seven month lows is also noteworthy.

Among the short leading indicators, industrial metals, the regional Fed new orders indexes, financial leverage, jobless claims, gas usage, and staffing are all positive. Stock prices, gas prices, and the spread between corporate and Treasury bonds are all neutral, rejoined this week by the ECRI commodities index. The US$ is mixed. Oil prices turned negative.

Among the coincident indicators, positives include consumer spending, Harpex, and rail, joined this week by tax withholding. The Baltic Dry Index is neutral. LIBOR and TED remain negative, joined this week by steel.

There were a number of small moves this week across the spectrum, but (except for the unreliable tax withholding) all but one were to the downside.  As a result, while both the nowcast and the short term forecast remain positive, they are a little less so than in the last few months. The longer term forecast crossed the line ever so slightly from positive to neutral several weeks ago, and remains neutral.