QVMS - Weekly Indicators: The Price Of Gas Continues To Be Decisive

Summary

• High frequency indicators can give us a nearly up-to-the-moment view of the economy.
• The metrics are divided into long leading, short leading, and coincident indicators.
• The long and short leading indicators are decisively negative, while consumer spending and employment coincident indicators are holding up for now.
• Gas prices have likely been, and will continue to be, decisive in 2023 as they were in 2022.

Purpose

I look at the high frequency weekly indicators because while they can be very noisy, they provide a good nowcast 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.

A Note on Methodology

Data is presented in a "just the facts, ma'am" format with a minimum of commentary so that bias is minimized.

Where relevant, I include 12-month highs and lows in the data in parentheses to the right. All data taken from St. Louis FRED unless otherwise linked.

A few items (e.g., Financial Conditions indexes, regional Fed indexes, stock prices, the yield curve) have their own metrics based on long-term studies of their behavior.

Where data is seasonally adjusted, generally it is scored positively if it is within the top 1/3 of that range, negative in the bottom 1/3, and neutral in between. Where it is not seasonally adjusted, and there are seasonal issues, waiting for the YoY change to change sign will lag the turning point. Thus I make use of a convention: data is scored neutral if it is less than 1/2 as positive/negative as at its 12-month extreme.

With long leading indicators, which by definition turn at least 12 months before a turning point in the economy as a whole, there is an additional rule: data is automatically negative if, during an expansion, it has not made a new peak in the past year, with the sole exception that it is scored neutral if it is moving in the right direction and is close to making a new high.

For all series where a graph is available, I have provided a link to where the relevant graph can be found.

Recap of monthly reports

December data started out with declines in both the ISM manufacturing and non-manufacturing index, the latter a sharp decline. The leading new orders subindex in manufacturing declined further into contraction. The household segment of the jobs report was the best in over half a year, as labor force participation increased, jobs increased sharply, and unemployment decreased. But the establishment portion was more mixed, as jobs growth decelerated further, the manufacturing workweek declined sharply in recessionary territory, temporary jobs declined, and aggregate hours worked declined. Still, both the leading manufacturing and construction employment sectors showed continued growth.

November data consisted of construction spending, which increased, although residential construction declined and factory orders declined.

Long leading indicators

Interest rates and credit spreads

Rates

• BAA corporate bond index 5.74%, down -0.10 w/w (1-yr range: 3.51-6.59)
• 10-year Treasury bonds 3.56%, down -0.32 w/w (1.66-4.25)
• Credit spread 2.18%, up +0.22 w/w (1.76-2.42)

(Graph at Moody's Seasoned Baa Corporate Bond Yield | FRED | St. Louis Fed )

Yield curve

• 10 year minus 2 year: -0.70%, down -0.15% w/w (-0.85 - 1.59)
• 10 year minus 3 month: -1.06%, down -0.56% w/w (-1.06 - 2.04) (new low)
• 2 year minus Fed funds: -0.07%, down -0.17% w/w

(Graph at 10-Year Treasury Constant Maturity Minus 2-Year Treasury Constant Maturity | FRED | St. Louis Fed )

30-Year conventional mortgage rate (from Mortgage News Daily ) (graph at link)

• 6.20%, down -0.34% w/w (3.52-7.38)

Although they have declined significantly since October, corporate bonds remains near the top of their 10 year range, so are negative. The same applies to long term treasury bonds and mortgage rates. Corporate bond yields, which often peak even before the Fed finishes raising rates, may have already made their high.

While the spread between corporate bonds and Treasuries remains positive, all three of my yield curve indicators have now turned negative. Only the Fed funds and 3 month Treasury are not fully inverted vs. the rest of the curve, where the 6 month Treasury is the highest yielding maturity.

Housing

Mortgage applications (from the Mortgage Bankers Association )

• Purchase apps down -12.2% w/w to 155 (154-349) ((SA))
• Purchase apps 4 wk avg. down -3 to 166 ((SA)) (341 high Jan 29, low 158 Nov 25)
• Purchase apps YoY -42% ((NSA))
• Purchase apps YoY 4 wk avg. -39% ((NSA))
• Refi apps up-16.3% w/w ((SA))
• Refi apps YoY down -87% ((SA))

*((SA)) = seasonally adjusted, ((NSA)) = not seasonally adjusted

(Graph at here )

Real Estate Loans (from the FRB )

• Up +0.1% w/w
• Up +11.3% YoY (-0.9 - 11.3)(new high)

(Graph at Real Estate Loans, All Commercial Banks | FRED | St. Louis Fed )

The highest mortgage rates in 12 years killed both purchase and refinance mortgage applications, the four week averages of which continue at or close to 8 and 20 year lows, respectively. On the other hand, mortgage rates, like bond yields, appear to have made their peak for this cycle, and are getting “less negative.”

From 2018 until late in 2020 real estate loans with few brief exceptions stayed positive. Earlier last year they varied between neutral and negative, but for the past several months have been very positive. This is being helped by inflation in house prices; thus the turn in the indicator will be when that cools.

Money supply

The Federal Reserve has discontinued this weekly series. Data is now only released monthly. November data was released one week ago:

• M1 m/m down -0.8%, YoY Real M1 down -8.8% (20 year low)
• M2 m/m down -0.3%, YoY Real M2 down -7.1% (40 year low)

No recession has happened without a YoY real M1 negative, or YoY real M2 below +2.5%. Real M2 fell below that threshold in March. Real M1 also turned negative as of May.

Corporate profits (Q3 actual S&P 500 earnings from I/B/E/S via FactSet at p. 30 )

• Q3 actual unchanged at 55.65, down -1.8% q/q
• Q4 estimated at 53.90, down -3.1% q/q

FactSet estimates earnings, which are replaced by actual earnings as they are reported, and are updated weekly. The "neutral" band is +/-3%. I also average the previous two quarters together, until at least 100 companies have actually reported. While the cumulative decline since the recent Q2 peak would be -4.9%, the average is -3.3% - which is still negative.

Credit conditions (from the Chicago Fed ) (graph at link)

• Financial Conditions Index unchanged (loose) at -0.21 (-0.03 - -0.62)
• Adjusted Index (removing background economic conditions) unchanged (loose) at -0.21 (+0.16 - -0.59)
• Leverage subindex up (tighter) to +0.28 (+0.51 - -0.35)

In these indexes, lower = better for the economy. 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. Leverage is near its highest since the Great Recession, obviously very negative.

Short leading indicators

Economic Indicators from the late Jeff Miller’s “Weighing the Week Ahead”

• Miller Score (formerly "C-Score"): up +3 w/w at 271, +9 m/m (125 6/24/22 - 374 on 2/25/22)
• St. Louis Fed Financial Stress Index : down -0.0873 to -0.0134 (0.4997 5/27/22 - -.8325 9/16/22)
• BCIp from Georg Vrba: up +0.8 to 18.4 iM's Business Cycle Index (100 is max value, below 25 is recession signal averaging 20 weeks ahead)

The Miller Score is designed to look 52 weeks ahead for whether or not a recession is possible. Any score over 500 means no recession. This number fell below that threshold at the beginning of August 2021, so not only is it negative, but we are now well into the “recession eligible” time period.

The St. Louis Financial Stress index is one where a negative score is a positive for the economy, and during its limited existence, has risen above zero before a recession by less than one year. It has just done so, warranting a change of rating to “neutral.” It will turn negative if it stays above +0.2 for at least 4 weeks.

The BCIp, which remained very positive until very recently, deteriorated sharply in the past month. It is now below its recession-signaling threshold for the second week.

Trade weighted US$

• Down -0.58 to 121.51 w/w, +5.6% YoY (last week) (broad) (114.23 - 128.58) (Graph at Nominal Broad U.S. Dollar Index )
• UP +0.41 to 103.91 w/w, up +8.5% YoY ( major currencies ) (graph at link) (94.63-114.78)

In early 2021, both the broad rating and the USD against major currencies turned higher YoY, and so changed to neutral. With both measures now well above +5% YoY, these ratings are negative. Note that both have improved considerably - i.e., “less bad” - in the past few weeks, and the broad index in particular is on the verge of turning neutral.

Commodity prices

Bloomberg Commodity Index

• Down -4.70 to 108.11 (100.86-136.61)
• Up +6.7% YoY (Best: +52.3% June 4)

(Graph at BCOM | Bloomberg Commodity Index Overview | MarketWatch )

Bloomberg Industrial Metals ETF (from Bloomberg ) (graph at link)

• 162.89, down -2.45 w/w (135.97-327.84)
• Down -8.2% YoY (Best +69.0% May 7)

During the Boom of 2021, commodity prices soared, and total commodities were very positive. Total commodities (which include oil) remain in the middle 1/3rd of their range, so are neutral. Industrial metals have declined into the bottom 1/3rd of their 52 week range, so have also turned negative.

Stock prices S&P 500 (from CNBC )(graph at link)

• Up +1.4% to 3895.08

Since January 3 of last year, there have been ongoing new 3 month and even 1 year lows. Since in the past 3 months we have had no new high, this indicator remains negative.

Regional Fed New Orders Indexes

(*indicates report this week) (no reports this week)

• Empire State down -0.3 to -3.6
• Philly down -9.6 to -25.8
• Richmond up +10 to -4
• Kansas City down -5 to -17
• Dallas up +11.7 to -9.2
• Month-over-month rolling average: up +4 to -12

The regional average is more volatile than the ISM manufacturing index, but usually correctly forecasts its month-over-month direction. These had usually been extremely positive ever since June 2020, but since last spring, gradually declined to neutral and then negative.

Employment metrics

Initial jobless claims

• 204,000, down -19,000 w/w
• 4-week average 213,750, down -6,750 w/w

(Graph at St. Louis FRED )

New claims made new all-time lows on a 4 week average in April. Once this metric failed to make a new 3 month low, its rating changed to neutral. It will not turn negative unless and until the 4 week average is higher YoY. In the last several months, it has trended sideways to moderately higher.

Temporary staffing index (from the American Staffing Association ) (graph at link)

• Down -1 to 105 w/w
• Up +0.4% YoY

This gradually improved to neutral at the beginning of 2021, and has been positive since then. There is a great deal of seasonality in the numbers, which typically rise slowly throughout the year except for certain holiday periods. The comparisons in the past several months have deteriorated sharply, and having fallen below +3.0% YoY, turned neutral.

Tax Withholding (from the Dept. of the Treasury )

• $339.9 B for the month of December vs. $351.1 B one year ago, -$11.2B or -3.2%
• $370.3 B for the last 20 reporting days this year vs. $343.0 B one year ago, +$27.3B or +8.0%

YoY comparisons turned positive in the beginning of 2021, remained that way until a short time ago. The YoY% change fell below 5% several times since summer, then rebounded, and oscillated between neutral and positive for several months. This week they are neutral again, as they were down YoY for the second month in a row in December, but the 20 day total is positive.

Oil prices and usage (from the E.I.A. )

• Oil down -$7.64 to $73.67 w/w, up 27% YoY ($71.46 - $123.70)
• Gas prices up +$.13 to $3.22 w/w, down -$0.06 YoY
• Usage 4-week average down -7.0% YoY

(Graphs at This Week In Petroleum Gasoline Section - U.S. Energy Information Administration [EIA] )

Gas prices are in the middle 1/3rd of their 3 year range, and so have returned to neutral. Oil is also in the middle of its 3 year range, and so it remains neutral.

Mileage driven remains negative.

Note: With gas and oil prices so volatile in the past 12 months, I believe the best measure is against their 3 year average. Measuring by 1 year, both have turned positive.

Bank lending rates

• 0.310 TED spread down -0.159 w/w (0.02 -.685)
• 4.400 LIBOR up +0.0300 w/w (0.10130- 4.400) (graph at link) (new 12 month high)

TED was above 0.50 before both the 2001 and 2008 recessions. Since early 2019 the TED spread had remained positive, except the worst of the coronavirus downturn, until this spring. It has been very choppy recently, varying between neutral and negative. This week it was neutral again.

LIBOR has been increasing consistently well into its negative range.

Coincident indicators

St. Louis FRED Weekly Economic Index

• Up +0.40 to +1.82 w/w (+0.73 12/10/22 - +6.16 2/19/22)

After a very positive 2021, this measure declined to less than half its best YoY level, thus changing to neutral. Last week was the lowest level since February 2021. The only time in its 15 year history it has been this low not in or coming out of a recession was in January 2016. I will continue to treat it as neutral unless the number turns negative.

Restaurant reservations YoY (from Open Table)

• December 29 seven day average -9% YoY (Best +% Jan. 5, 2023)
• January 5 seven day average +15% YoY (Worst -29% Jan 13, 2022)

The comparison year for this metric is 2019 and not 2021. This year the metric gradually improved to neutral, and for one week, positive. But it has since changed back mainly to neutral, and then negative. But this week was the most positive YoY since the pandemic began.

Note I am now measuring its 7 day average to avoid daily whipsaws.

Consumer spending

• Johnson Redbook up +10.2% YoY (high 15.8% in Jan. 2022; low 5.7% Dec. 9, 2022) United States Redbook Index - 2023 Data - 2005-2022 Historical - 2024 Forecast

The Redbook index remained positive almost without exception since the beginning of 2021 until October. With two exceptions the past 7 weeks have been the lowest YoY comparisons in many months. The new link I have added above goes to a 5 year graph to best show the comparison.

I recently downgraded this metric to neutral, but it has rebounded sharply in the past several weeks.

Transport

Railroads (from the AAR )

• Carloads down -12.5% YoY
• Intermodal units down -0.5% YoY
• Total loads down -6.8% YoY

(Graph at Railfax Report - North American Rail Freight Traffic Carloading Report )

Shipping transport

• Harpex down -21 to 1208 (1208-4586) ( new 18-month low )
• Baltic Dry Index down -369 to 1146 (965-3369) (graph at link)

Rail carloads turned positive early in 2021, before gradually fading to negative from August through the end of the year and the beginning of this year. The total loads index has been consistently negative for the past four months. In the past several months, comparisons have hovered near the zero line, varying between neutral and negative. This week they were negative again.

Harpex increased to near record highs again earlier this year, but has since backed off all the way to new lows. BDI has traced a similar trajectory, warranting a change to negative.

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 ( American Iron and Steel Institute )

• Down -0.7% w/w
• Down -9.4% YoY (worst -10.0% Dec 2, 2022)

Since the end of March 2021, against terrible comparisons, this metric had been positive, typically running at a double digits higher YoY percentage growth. This past spring, after almost continuous deterioration, it turned negative, and has remained so.

Summary And Conclusion

Below are this week’s spreadsheets of the long leading, short leading, and coincident readings. Check marks indicate the present reading. If there has been a change this week, the prior reading is marked with an X:

Seven weeks ago I went from “Recession Watch” to “Recession Warning,” as all three of my primary systems are consistent with a near-term or imminent recession.

Consumer spending and employment are the two big coincident indicators I am watching for confirmation that a recession has likely begun. Spending has had a big rebound in the last few weeks. Temporary employment is on the cusp of turning down - but hasn’t yet. Tax withholding has been choppy, but since the beginning of the fiscal year on October 1 is up only 1.8% compared with the same period one year ago (this while average wages are up 5% YoY).

I continue suspect the decisive factor in the direction of the economy since June has been the decline in gas prices from $5 to $3 in December. In the next few months it will likely continue to be decisive.

For further details see:

Weekly Indicators: The Price Of Gas Continues To Be Decisive