Home > XE Currency Blog > Updating the mid-cycle indicators


XE Currency Blog

Topics7692 Posts7737
By New_Deal_democrat September 14, 2015 1:10 pm
  • XE Contributor
New_Deal_democrat's picture
New_Deal_democrat Posts: 547
Updating the mid-cycle indicators

While we have a host of economic indicators telling us if a recession is near, is upon us, or has recently ended (i.e., leading, coincident, or lagging indicators), the fact is that expansions tend to be much longer than recessions.  Thus I have tried to develop "mid-cycle indicators," which are data series that can help tell us if we are in the earlier or later part of an economic expansion.  

Because the economy is growing throughout an expansion, mid-cycle indicators are usually about the second deriviative, that is, whether growth in a given metric is accelerating or decelerating. Most are somewhat noisy, most obvious in retrospect. What they have in common is that the peak, or the change in trend, tends to happen somewhere between 1/3 and 2/3 of the way through an expansion.

So far I have identified 7 data series with promise.  They are:

1. Real retail sales YoY vs. Real consumption expenditures YoY
2. Real durable goods consumption expenditures YoY vs. real nondurable consumption expenditures YoY.
3. YoY% change in real personal consumption expenditures.
4. YoY% gain in employment
5. U. Michigan consumer sentiment
6. Real personal savings rate
7. YoY% change in gross fixed capital formation

Let's look at each of them in order

1. Real retail sales YoY vs. Real consumption expenditures YoY

I have been tracking this comparison for about 4 years.  It is a very reliable and consistent midcycle indicator, having been triggered during 11 of the last 12 expansions. In the earlier part of an expansion real retail sales tend to grow at a fester rate than real personal consumption expenditures.  In the later part of an expansion, real personal consumption expenditures grow faster. Ultimately the growth rate for both declines, and real retail sales YoY turn negative before the onset of a recession.  Here's what they look like now:

For the last 9 months, real personal consumption expenditures have decisively grown more quickly than real retail sales.

2. Real durable goods consumption expenditures YoY vs. real nondurable consumption expenditures YoY.

Real consumption expenditures for durable goods tend to turn south before expenditures for nondurable goods, which may never turn negative even during a recession.

Expenditures for durable goods are holding up well.  Just like in 1986, when oil prices also collapsed, since late last year expenditures for nondurable goods have gone to zero YoY and even turned negative. Arguably, this mid-cycle indicator has not been triggered, although there were several months last year where there was a pattern change.

3. YoY% change in real personal consumption expenditures

Almost always this peaks in the earlier part to midpoint of an expansion, and if the growth rate declinea by half, that is confirmation of the prior peak. 

This cycle gave a false signal after an initial peak in 2010.  The series made another peak in the first quarter of this year, with only a slight decline in the second quarter.

4. YoY% gain in employment

This is a very valuable indicator because it has such little noise. Usually there is one well-defined peak during an expansion, although sometimes such as in the 1990s, there are two peaks, on early and a second later in an expansion, usually reflecting a mid-cycle raising and the lowering of interest rates by the Fed.

This indicator appears fo have peaked last November

5. U. Michigan consumer sentiment

Officially this is one of the 10 items in the index of leading indicators.  Most often, however, consumer confidence has typically peaked somewhere near the middle of expansions.

The FRED data is always 6 months behind, per their agreement with U. Michigan.  Here's the more recent trend from Briefing.com:

Confidence peaked last January coincident with the low in gas prices. 

6. Real personal savings rate

This is a noisy series best used in confirmation of others. Typically in the middle to later part of an expansion, consumers dig deeper into their wallets, and their YoY savings rate minus the rate of consumer inflation decreased by about 5% or more.

In this expansion that event appears to have been triggered by the end of the 2% payroll tax holiday that ended in December 2012. Even though 2% more of their paychecks were withheld starting in January 2013, consumers simply dug deeper into their savings and spent more.

7. YoY% change in gross fixed capital formation

this is another noisy series, best used as confirmation of other data.Sometimes there are multiple peaks during an expansion, but most often the peak is somewhere near the midpoint.

This appears to have made its peak for this expansion early in 2011.

To summarize, of our 7 mid cycle indicators, 3 (#s 2, 3, and 6) give equivocal readings.  The remaining 4 (#a 1, 4, 5, and 7) appear to have peaked.

Could the mid-cycle readings be false? Yes, of course, but for that to be so, I would expect to see a fresh low in interest rates, especially for mortgages and/or a surge in real wage growth. The first looks unlikely, and there is no sign that employers are willing to open their wallets sufficiently for the second to happen.  The most likely conclusion, therefore, is that while the next recession is not likely at any point in the next year (per the long leading indicators), most likely we are closer in time to the next recession than the last one.

Aside from just listing and discussing mid-cycle indicators, can we mathematically estimate how reliable they are as a group?  I'll take a look in my next post.

Paste link in email or IM