solarseven/iStock via Getty Images

solarseven/iStock via Getty Images

I have several systems for forecasting the economy. One is the high-frequency "Weekly Indicators," which, as the name implies, is updated weekly and is thus very timely. For the short term, up to about 6 months out, a second, K.I.S.S. method, relies on the Index of Leading Indicators, is a perfectly adequate tool with the inconvenient habit of being right more often than most highly paid Wall Street forecasters.

To forecast the period over 6 months out, I turn to "long leading" indicators, relying for that on monthly and quarterly data which has been extensively vetted in for decades already, and as during that time going back 50 years has a sustained record of turning one year or more before the onset of a recession.

Most of what you read which seeks to forecast the economy falls into two categories of leading you astray:

Most observers simply project existing trends into the future, overlooking that there are some data-points that reliably turn before others, and thus completely miss turning points until they are already in the rear view mirror.

A great deal - perhaps most - of what you read about the economy starts off with an ideology and a conclusion based on that ideology already baked into the cake. The writer then marshals a series of data-points that confirm that already-determined conclusion. The very same utterly decisive economic metric from last year that was moving in one direction is completely ignored this year when it is moving in the opposite direction, and vice versa.

By contrast, my method is to look repetitively at a bank of indicators that have decades - in a few cases, over a century - of history, and a demonstrable record of being reliable harbingers of the continuation or reversal of an existing trend.

Now let's return to my main topic at hand. Geoffrey Moore, who for decades published the Index of Leading Indicators, and in 1993 wrote "Leading Economic Indicators: New Approaches and Forecasting Records," identified 4:

A variation of the above is Paul Kasriel's "foolproof recession indicator," which combines real money supply with the yield curve, i.e., the difference in the interest rate between short- and long-term treasury bonds. This turns negative a year or more before the next recession about half of the time.

Another long leading indicator has been described by UCLA Prof. Edward E. Leamer, who wrote "Housing IS the Business Cycle." In that article, he identified real residential investments as a share of GDP as an indicator that typically turns at least 5 quarters before the onset of a recession.

Several other series appear to have merit as long leading indicators as well. Real retail sales in several forms also have value as a long leading indicator and, in particular, real retail sales per capita. Additionally, the tightening of credit conditions also appears to have merit as a long leading indicator.

That gives us a total of 7 categories of long leading indicators. All of these economic series have a long-term history of turning a year or more before a recession.

When I last looked at these indicators in January of this year, I concluded that:

"the long leading index . . . forecast[s] an outright stall by year end 2022. It would not take much - an inverted yield curve, or a further softening of business profits - to tilt the forecast negative (for 2023). But we're not there yet.”

Because there have been significant developments since then, an update looking ahead into Q2 2023 is important.

On a monthly basis, corporate bond yields data goes back over 100 years to 1919. With the exception of the 1981 "double-dip," the fiscal contractions underlying the 1938 and 1945 recessions, 1926, and - excusably I think! - 2020's pandemic - corporate bond yields have always made their most recent low over 1 year before the onset of the next recession, as shown in the graph below of the last 40 years (although the correlation goes back much further):

Now here is a close-up of the last 10 years:

AAA-rated bonds last made a new low in July 2020; BAA-rated bonds in December 2020. Mortgages reached all-time lows in early January 2021. In the past few months, corporate bonds have approached 10 year high yields, and mortgage rates are at 10 year highs. In the past 40 years, only in 1994 did a conquerable move not result in a recession. Thus corporate bonds are very negative.

Housing, which typically follows interest rates with a few months' lag, has typically declined by over one year, and by 20% or more before the onset of a recession (the mild recession of 2001 being the exception, with only a 10% decline), as shown in the below graph of housing permits, both total (not shown) and single family (red in the graph below) as well as new home sales (gold):

Housing followed the pattern of interest rates in 2021, peaking early before declining nearly 20% by early autumn and then recovering somewhat. This year so far, sales (the most leading indicator but very noisy) have declined by over 20%, while single family permits (the least noisy) have declined, but only 9%. By contrast, the three month moving average of housing starts (blue) as of April is at a 10-year high:

This is due to the 50 year record of houses authorized but not yet started (not shown). Since it is starts that constitute real economic activity, I am according them extra weight in this forecast.

Additionally, housing as a share of GDP, after declining for three quarters in a row after making a 3-year high in Q1 of 2021, rebounded in Q1, but was still below its cycle peak:

This is still a negative. Taking the caveat about housing starts into account, overall housing is a weak negative.

Corporate profits deflated by unit labor costs have an excellent track record going back over half a century as well. Below I show them both in log scale with and without inventory adjustments:

These were reported for Q1 as part of the revised GDP report yesterday. In nominal terms without inventory adjustment they rose; after inventory adjustment they declined. But after deflating by unit labor costs, which rose by 2.8% in Q1 compared with Q4 in 2021, they declined by -1.3% and -6.9% for the quarter, respectively. Further, they are down from their respective peaks in Q2 and Q3 of last year, respectively, by -3.6% and -7.0%:

As a result, this indicator, which was neutral in January, is now a negative.

No recession has ever started without at least real M1 turning negative or real M2 declining to under +2.5%. Here is this relationship going back over 60 years on a YoY% basis:

As of March, Real M2 turned negative YoY. As of April, real M1 has also declined for 2 months in a row, is at a 9 month low, and is only up 0.7%:

This is sufficient for real money supply to be a negative.

This has been an excellent long-range forecasting tool in times of inflation (although I do not think a positive yield curve is definitive in low interest rate deflationary environments). In the last 60 years, typically, a recession has begun after the Fed raises rates to combat inflation, and sufficiently so that the yield curve inverts.

Here is what the yield curve across the spectrum of Treasury maturities looked like at the beginning of January (dark red) compared with earlier this week (light red):

The current curve is typical of the last few months, with a positive slope out through 7 years, while there has been a 7 vs. 10 year inversion, and another 20 vs. 30 year inversion.

Here is what 3 popular measures of the yield curve, the 10 year minus 2 year Treasury, 10 year minus 3 month Treasury, and 2 year Treasury minus Fed funds rate, look like for the past 5 years:

Although the 10 year minus 2 year spread did very briefly invert 2 months ago, and remains tight, the others have not, and indeed have very positive spreads. Despite what is happening at the very long end of maturities, this indicator remains a positive.

The loosening or tightening of credit also appears to be an important component of changes in the economy over one year out. Although it only has a 30-year track record, two components of the Senior Loan Officer Survey have had a good track record. The report for Q1 was released earlier this month.

The first component, the percentage of banks tightening vs. loosening standards for commercial loans to large (blue) and small (red) firms, was almost exactly neutral. (Note that a "positive" number actually means tightening. Don't blame me!):

The second component, demand for loans by large (blue) and small (red) firms, remained positive:

It is worth noting that the Chicago Fed leverage (gold) and Adjusted Financial Conditions (red) subindexes, which have a good record anticipating the Senior Loan Officer Survey’s result re tightening credit, both turned above zero in the past month (again, a positive number is a negative for the economy):

This indicator remains weakly positive on net for the next 12 months.

These peaked more than a year before the onset of the last two recessions, but made troughs within six months of each recession bottom.

These made record highs last March and April, during the stimulus spending spree, then declined into autumn before rising again since December:

These are up 1.4% since last May, and the highest of all time except for March and April 2021. While for that reason I can’t call these a positive, the recent rebound in no way resembles their declining pattern before the last two recessions, and switches the long leading signal from negative to neutral.

Let me point out, first of all, that this is a more comprehensive set of indicators, and in some cases, indicators with longer and better track records than are found in my more timely "Weekly Indicators" columns.

The indicators have generally slowly deteriorated in the past 12 months.

As of now, two are still positive: the yield curve and (weakly) credit conditions.

One is neutral: real retail sales per capita.

Four (all of the indicators in Prof. Moore’s model) are now negative: corporate bonds, housing, corporate profits, and real money supply. The first two had their best readings in January 2021; corporate profits in Q3 of 2021; and real money supply in January 2022. Twelve months later is January 2023, and that according to past history is the beginning of the window within which a recession may occur.

As a result, for the first time since the pandemic began the long leading forecast in my premiere system is now negative, warranting a “Recession Watch” for the period 12 months past the last date, i.e., beginning with Q1 of 2023.

This should be taken in similar fashion to a “hurricane watch” in weather forecasting. A hurricane is not certain, but a storm is brewing in the Atlantic, conditions are favorable for further development, with a significant chance of landfall in the forecast cone. So as of now we watch for (1) the long leading indicators to be sustained over a longer period, and (2) for the short leading indicators to begin to turn, which may not happen for a number of months yet. The more that happens, the more the “watch” turns into a “warning.”

This article was written by

Disclosure: I/we have no stock, option or similar derivative position in any of the companies mentioned, and no plans to initiate any such positions within the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.