Greg Ip says the conditions for recession are not in place. I disagree. And I show where and how he went wrong. 
The chart itself explains where Ip went wrong. See if you can figure it out.
“This Doesn’t Look Like Recession” Says Greg Ip
Please consider This Doesn’t Look Like Recession. Here’s How One Could Happen.
Unemployment is rising, stocks are falling and bond yields are well below short-term interest rates. These are all telltale signs of recession.
But a closer look suggests that while recession risk has risen, the U.S. isn’t in one now. The distinction is crucial because it means it isn’t too late to head off a downturn. It all depends on the Fed, and on the unpredictable moods of investors, consumers and employers.
The increase in unemployment to date, according to a rule-of-thumb popularized by the economist Claudia Sahm, in the past has only occurred during recessions.
To decide if it’s raining, it’s better to stand outside than count umbrellas. Similarly, to determine whether recession has begun, better to look at the indicators the NBER uses than the Sahm rule. Three—payroll employment, industrial production and real (inflation-adjusted) incomes, minus government transfers—were all shrinking in the four months up to and including the month the Sahm rule was triggered, in 1990, 2001 and 2008. In all three, a recession had begun several months earlier.
In the four months through July, payrolls were growing, and in the three months through June, so were real incomes and industrial production. If a recession had already begun, it would be a very unusual one. (Sahm said last week she didn’t think a recession is imminent).
Background on the McKelvey (Sahm) Rule
Edward McKelvey, a senior economist at Goldman Sachs, created the indicator.
Take the current value of the 3-month unemployment rate average, subtract the 12-month low, and if the difference is 0.30 percentage point or more, then a recession has started.
Claudia Sahm, a former Federal Reserve and White House Economist, modified the indicator from 0.3 to 0.5.
Please consider The Sahm Rule: Step by Step written December 7, 2023 by Claudia Sahm.
I created the Sahm rule, and it’s on me to communicate it well. I try. If you have any questions, please add them to the comments.
Sahm claims to have invented the rule. However, credit should go to Edward McKelvey, at Goldman Sachs.
The Lag Effect
Sahm modified the McKelvey rule to eliminate false positive. But that was at the expense of being far less timely.
In the 2008 recession, the Sahm rule triggered three months late. In the 1973 recession, Sahm triggered 7 months late.
Ip’s Huge Mistake
Ip’s huge mistake is comparing conditions in place when the rule triggers instead of conditions when the recession began.
Ip shows 4-month sum pf payrolls as negative, when in fact, they were positive.
Change in Nonfarm Payrolls Oct 2007-Jan 2008
- Oct: +72,000
- Nov: +116,000
- Dec: +105,000
- Jan: +1,000 Recession Start
Every month was positive, including the start of the recession.
Change In Nonfarm Payrolls Using Sahm Trigger
Jan: +1,000
Feb: -71,000
Mar: -70,000
Apr: -219,000 Sahm Trigger (was this remotely useful?)
The Sahm rule did not trigger until three months into the recession when payrolls were -219,000.
And look at December 2007. Despite a report of +105,000. It was clear recession was unavoidable.
Bernanke did not see recession until April, when it obvious to the world.
Industrial Production Index Oct 2007-Jan 2008
- Oct: 101.6
- Nov: 102.2
- Dec: 102.3
- Jan: 102.1
Industrial production peaked one month before recession started.
Look at Ip’s chart for Industrial Production and note how negative (and wrong) that it is. September Industrial Production was 101.9.
The average of the four months at the start of recession is (101.6 + 102.2 + 102.3 + 102.1) = 102.1 That’s up from 101.9 not hugely down as Ip shows.
Now let’s take a look at things from the perspective of the Sahm indicator.
Industrial Production Using Sahm Trigger
Jan: 102.1
Feb: 101.8
Mar: 101.4
Apr: 100.7 Sahm Trigger (was this remotely useful?)
“Real-Time” Nonsense
Sahm labels “her” indicator as “real-time”.
With lags as long as 7 month and never leading in the history of the data to 1948, there is nothing “real-time” about it.
Ip enhances the mistakes by taking a lagging indicator then applying those conditions to other indicators with varying lags.
QCEW and Business Employment Dynamics (BED)
Greg Ip also ignores or is unaware of enormous discrepancies between the monthly nonfarm payroll reports, also called Current Employment Statistics (CES) and the QCEW reports.
BED is a large subset of the Quarterly Census of Employment and Wages QCEW which covers 11.6 million businesses
The nonfarm payroll reports are based off surveys of under 700,000 businesses. QCEW is based off 11.6 million businesses.
Expect the BLS to Revise Job Growth Down by 730,000 in 2023
On July 26, 2024, I commented Expect the BLS to Revise Job Growth Down by 730,000 in 2023, More This Year
Bloomberg’s chief economist, Anna Wong, arrived at a -730,000 overstatement in nonfarm payrolls. Using the same data, I calculated (in advance of her number), -779,000.
By her estimate, the BLS jobs overstated nonfarm payrolls by 81,111 jobs every month for 9 months. Click on link for more details.
Yet, here we are, putting faith in nonfarm payroll numbers that are not only lagging, but outright nonsense.
My July 8 Recession Call
July 8: Weak Data Says a Recession Has Already Started, Let’s Now Discuss When
I’ve seen enough. A recession has started. Let’s discuss starting with a very good indicator that has few false positives and no false negatives.
Since then ….
ADP Change in Employment
Unemployment Rate
August 2: Unemployment Rate Jumps, Jobs Rise Only 114,000 with More Negative Revisions
August 2, 2024: The McKelvey (Sahm) Unemployment Rate Recession Rule Just Triggered
A recession indicator based off rising unemployment triggered in July. Claudia Sahm, a former Fed economist, takes credit for an indicator she did not invent. Let’s discuss.
Key Points
- Due to the lagging nature of the Sahm rule, recession is highly likely to already be underway when it triggers.
- Applying her rules to other indicators is a serious mistake.
- I like a trigger of 0.4, halfway between the original McKelvey idea and Sahm’s revised rule.
- Click on above link for details.
Also, please take a look at my July 31 post Small Business Employment Growth Is Now Negative (and What It Means)
ADP data shows year-over-year payroll growth is negative 88,000 for small corporations sized 20-49!
Payroll trends are negative in all but very large corporations (Given Intel’s 15,000 Mass Layoff and tech woes in general, how long will that last?)
Recession is underway.
Not clear to me why this post insists we are already in a recession, with big implications for the November elections. Obviously, as rightly noted in the post, the Sahm rule (or any rule based on unemployment) is a lagging indicator of when the business cycle turned (from peak to trough) as declared by the NBER (the lag for the Sahm rule in the post-war period relative to NBER cycle dating is on average a few months). Given the monthly data on income and production in different sectors over the last few months (which are the relevant concurrent business cycle indicators for the NBER), it’s hard to see how the NBER committee could decide in the future that the turning point has already happened. Moreover, we are talking about a business cycle turning point, not the trough or approaching the trough of the cycle, as refereed by the NBER business cycle committee. In practice, voters are likely to get only antsy about the cycle once the cyclical “downward journey” is well underway for a few months (in 2007-8 the NBER turning point was in December 2007, but most people didn’t pay attention until September 2008). Conclusion: this business cycle is pretty irrelevant for the November election.
Strip out the trillion in new debt every 100 days … and the US economy collapses.
Everything is fake. GDP… GDI… jobs reports… cuz it’s all based on something that must continue but cannot…
Even the election is fake https://fasteddynz.substack.com/p/what-if/
If IP still has the ear of FOMC members, “his opinion” reflects what he has heard.
Are we in one? I keep saying “soon” and I keep being wrong. My gut is still “soon” and “not quite yet”. I’d expect more recently updated data like WARN notices or Challenger layoff notices month over month to be increasing, but they’re moving the opposite way the past couple of months.
The stock market looked like it was giving a recessionary signal, but that quickly reversed. Bonds are going the opposite way this week.
GDP at 2.8% for Q2 likely has some momentum going into Q3 still. Maybe Q3? No question that consumers dialed it back a notch, but enough to tip into negative GDP? Maybe.
Did the birth/death model stop overestimating jobs? Maybe. UBS thinks so, but I’m doubtful. At some point it will & payroll data will be meaningful again. Jobs in July were in my mind likely flat accounting for that, much as in the HH survey. Bottom line – still not enough info to be all that definitive.
I’d still say Q3/Q4 and call bullshit on a soft landing. One day I’m going to be right. The stopped clock theory of forecasting works eventually.
Whatever it takes to keep the plane in the air until November, duct tape expenses be damned. If Trump gets in, point plane towards mountain, hands off the yoke.
Agree recession is underway.
Next question. How DEEP a recession is it likely to be?
Completely unknowable. Kamala with a Republican congress probably means deep recession as the stimulus checks won’t be too fast in coming. But we know that Trump has no aversion to his own idea of sugar high economics, and those will push the austerity measures into the future some more as well (and worsen the problem). I’m still voting for him though, hoping that Congress and/or his advisors keep his economy-goosing shenanigans to a minimum.
So you don’t think the supply side shock to the labor pool of about 10 million illegal immigrants could be distorting the Sahm rule? I believe that is what I am hearing you say, so why not?
Because Total Employment is declining. Adding more immigrants to a stagnant economy makes the recession worse (more mouths to be fed at the expense of those actually producing; larger deficits).
https://fred.stlouisfed.org/graph/?g=1rBd4
The question is, are illegal immigrants “real” as far as the economy is concerned? If they can’t find work, they can go home. In fact, that is what migrant labor has done every single year for decades.
Namesake Claudia Sahm recently said, “The Sahm rule is likely overstating the labor market’s weakening due to unusual shifts in labor supply caused by the pandemic and immigration,”
SPX will decide who wins the race.
I judge things by what is happening around me. The Home Depots and Lowes are not as busy as a few years ago. I often see folks have less things in their grocery carts. Many restaurants are not as full as a few years back. But maybe things are booming in other areas? So how slow does it have to get before it’s a recession?
Canada employment
full employment 61.6K
Part time employment -64.4K
net change -2.8K
Employment availability is tightening as part time work is lessening. Those working two jobs full time and part time are running out of wiggle room to make the Bills..
Hasn’t wall street and the government have outlawed them indefinitely?
Just keep buying “cheep” stocs.
Yet, I keep hearing that the numbers used above for unemployment rising has a lot to do with the supply of workers outpaced hiring, both due to a higher participation rate and immigration. If this is true, perhaps, this time is different?
No, because the separate measurement of total Employment Level shows the usual stagnation-and-decline pattern typical of the early stage of recessions (and not other times in the cycle).
https://fred.stlouisfed.org/graph/?g=1rBd4
Paramount to lay off 15% in the coming weeks. Wells Fargo announced it was closing 30+ branches (not sure of the headcount). Last week it was Intel announcing 9-10% layoffs. The mass layoffs are here because the economy has slowed too much to support the current employment levels. Who is next?
R.I.P. Doug Sahm. The Texas Tornadoes actualy performed at Bill Clinton’s inauguration in 1993.
McDonald’s revenue is down – and that’s in inflated dollars.
You know things are bad when people can no longer afford to eat at McDonald’s.
Businesses know when recession is on the way even if government is full of BS on that front. The $5/meal fast food blitz began last fall during the NFL regular season. My own perception was things began slowing almost immediately around Labor Day 2023.
It all depends what you deign to measure. Government will likely never admit a recession again. Many market commentators will still be saying the economy is good even when it’s a deep recession/depression.
Atlanta’s gdpnow is currently at 2.9%. There are short-term money flows that are a proxy for R-gDp. There are long-term money flows that are a proxy for inflation.
Short-term flows have been up since 2023. Long-term flows have been down since 2023. N-gDp will fall in August.
How many actual recessions have your “proxies” been backtested against?
the soft science of economics is sort of silly. a recession is when my neighbor or family member loses their job. a depression is when i lose my job. i lived through a depression in a hood in a city that prices of homes crashed from top to bottom by 70%. one in 3 neighbors lost their homes due to short sales………….the austrian economics get these things usually correct. the boom of currency and bust is gonna be epic
This week I read NY Gov Hochul suspended the congestion pricing program that would charge drivers entering NYC $15 during business hours (about $3500 annually). There was enough pushback that threatened the NYC economy. The money was supposed to go toward the subway system. The subway budget has such a deficit that at least one official lamented Climate Change initiatives would not be achieved, and possibly would be cut from the budget. That’s one more indicator of a recession.
The NatSec State never goes into recession. That’s a problem for the peasants.
the department of war, is the term i enjoy best. the old and more honest label
When Greg Ip used to be the Fed Whisperer … good times.
Mish rips Ip’s tip, Ip’s blip not hip, means zip. Ip’s verbal slip-and-trip gives Ip fat lip.
Ip quips for market rip… Ip loses grip in next dip?
The best forecasting tool for recession over recent history is the consumer credit card and loan delinquency rates. With ~70% of economy dependent on consumer spending/banks continuing to lend, past recessions arrive when over-levered consumer loan delinquencies rise to levels indicating a fragile financial system has developed. Every recession has begun once this condition had been reached and then an unexpected event triggered banks to withdraw further credit expansion. Having over-spent without a cash reserve, consumers are forced to retrench which creates a spiral of retrenchment i.e., recession. The level of credit card delinquency indicating credit fragility is 4.7%. The data is published every 3mo by the St Louis Fed on its FRED site. We have some distance to travel yet before we get to this level.
So good info, but what’s more important, the most recent number, or the trend? July #is 4.3%. What are odds we get to 4.6 when Aug is released?
This data only goes back to 1992 so “every recession” is a sample size of 2-3 (and missed the 2020 event, so not “every”).
This data lags badly – it is now August and the most recent quarterly data point available is only for Q1 (Jan-March) which itself wasn’t released until May 21. In a couple of weeks we’ll hear about April-June just before it’s September.
For the absolute level to matter, one has to make hairsplitting threshold arguments, such as that a 1.5% increase to the 4.8% delinquency rate in 1996-1998 wasn’t a recession signal, but somehow a 0.5% increase to a 4.9% rate in 2001 was a recession. I don’t think things work that way.
The “level which indicates or triggers fragility” will change over time depending on the state of the rest of the system. If banks are under-reserved against credit losses, the absolute level of delinquency could be lower and still trigger a credit pullback.
Finally, the demographic which is in financial trouble now may be under-represented in the credit card data. The fraction of delinquent cards (or individuals) is more important to estimating consumer stress, but that is what gets reported. What is reported is only the fraction of delinquent balances (which matters more for the banks).
My three most reliable recession indicators: 1. More than one person driving around at night with a busted headlight. 2. Only thin/attractive women present at the airport. 3. Older, attractive, blonde women present at the dollar tree.