This weekend in the US we celebrate Labor Day. It is the unofficial end of summer, as the students are all back in school and school sports, from the youngest ages to the now-paid college athletes, are in full swing.
It is also a time after which Wall Street comes back from family holiday trips and focuses on maximizing their bonuses. There will be industry conferences in September, more earnings in October, an election in November, and a beta run into the end of the year to try to make more money. Investment bankers will bring every deal they can in order to generate more fees.
However, September is also seasonally the worst month of the year. Whether I look the past 10, 20 or 30 years, September is always the most negative, one of the few negative, months for the SPX even with Q4 being solidly positive:
Suffice to say, a lot could happen. What would surprise the market? It won’t be rate cuts. There are 9 of those now priced in thru the end of 2025. Jay Powell is incredibly unlikely to NOT deliver now in September, though I think 25 bps is the most likely (still believe we don’t need it but the Fed won’t disappoint markets).
What could surprise the market, which is now focused on growth and not inflation, is that the landing we are having is hard and not soft. A vast majority of the market sees us as having a soft landing. What is the difference between them? Jobs. Labor.
The post last week was a video interview with
, of It was very thought-provoking on the future of AI and jobs and if you haven’t watched it, I would. Amrita is also a good follow. Another reader, who writes the (another very good follow) commented on the post and steered us to a link “layoffs.fyi”. It shows us that in 2023-34, the tech industry has laid off about 400k workers. These are very talented and very high paid workers. No wonder Wal-Mart has seen higher income people trading down to Wal-Mart. If people with technology skills are being laid off, it will be tough for everyone. I can tell you on campus, I constantly hear from students that the investment banks and consultants are not hiring like they used to. Whether firms are seeing AI productivity or not, they are certainly preparing for it. But that can have some negative repercussions.Speaking of AI, this week was all about NVDA. I wrote about it before the earnings:
Chart of the Day - NVDA earnings
It is all about NVDA earnings today. There is no way to avoid that. The biggest stock in the biggest theme rippling thru global mkts. Yesterday was the lowest volume day of the year by a long shot. Clearly the only thing on people's minds at this point is NVDA
With a holiday coming up in the US, one might expect a day of flurry tomorrow & then everyone heads out of town too
There is almost no doubt NVDA will put up good numbers on an absolute basis. The trends are pretty clear in AI & NVDA has a moat around its business for now
However, markets move at the margin. It is how these numbers come out relative to expectations that really matters. A high growth rate that falls short of that expected would see downside in the stock
The charts today show you what is expected today on a number of levels. First, the options mkt has an implied move of 9.81%. That is, if you trade options for today's expiration, the stock needs to move 10% just to breakeven. That is a big ask
However, we can see the average move in NVDA around earnings is close to 16%. Certainly over the last 5 quarters, a 9.81% move would be on the low side of reality
We can also see the mkt is expect 65 cents this qtr & $2.70 for the full fiscal 2025. Just 3 years ago NVDA only made 23 cents for the entire fiscal year. This explosive growth in earnings has driven the stock up. Will the expectations be for continued growth at that level?
The bottom chart shows how the stock, will initially leading earnings expectations, has now followed it higher. As earnings stair step higher each quarter, the stock follows suit. Every time earnings forecasts go higher, the stock looks cheaper
But notice that those stair steps are getting smaller & smaller. Still positive, yes, but smaller. This qtr expectation is only 3 cents higher than last qtr actual. The expectation right now is earnings will continue to be 6-7 cents per qtr higher. What if that comes down
This year's actual P/E is 75x however the P/E on 1 yr fwd is only 47x. Investors are relying on the continued beat and raise of earnings to make the case of staying in the stock. If it does, the trend continues. If it doesn't?
NVDA is just one stock, the biggest of course, in a powerful theme. As we see each qtr, the results & narrative from NVDA will drive every other name even tangentially related. It will drive every etf with NVDA in it, which impacts those other stocks
The story today after the close will be NVDA earnings. This will be bigger than the Fed days or the non-farm payroll days or any economic news days. This will be one to watch, but remember to watch the moves at the margin
NVDA’s negative reaction to earnings, as well as other news like SMCI delaying its financial reports after a short report comes out, had a negative feel to all of tech but certainly high-fliers. This has put the Nasdaq in a precarious position, sitting on the long-term trendline as well as at the lower end of the ichimoku cloud. The MACD hasn’t made up its mind yet so this week will be critical to see if the level holds and we get an all clear, or if the bears re-assert control.
However, the NVDA post also got me thinking about productivity. This is meant to be the holy grail of AI after all. So I took a look at that too:
Chart of the Day - AI
With the NVDA earnings, headlines & reactions, the focus today will be all about AI. What inning are we in for this cycle? How much growth is left? Who will dominate it?
The AI discussion has permeated the economic data as well. When one presents a set of bad economic leading indicators - yield curve, Sahm Rule, continuing claims, ISM new orders etc - the argument is that these traditional indicators don't matter because of the AI boom
The presumption is that the AI boom is a productivity miracle that will not only save the day, but take the day to heights never seen before in the history of the world
The question I would have is, are we seeing this productivity miracle that everyone is talking about? When I posted my interview about whether genAI is a bubble on my Substack, an astute reader sent me a link to layoffs.fyi
This shows that in 2023, 264k techies were laid off and a further 134k so far in 2024. I can see from the investment banks & consultants on campus, hiring is not the same as it was a couple years ago
Is company behavior impacted by this expectation of productivity gains? Are we seeing any productivity gains? The chart today shows non-farm business productivity over the last 35 years
As you can see, we came well off the lows that we saw in 2022, which was really an unwind of some seismic gains seen in the year prior. As we are normalizing here, the level of productivity in the US right now looks ... average
The horizontal line is the average productivity of the last 30 years. It comes in at 2%. Right now, in the midst of an AI productivity boom that will save the world, we are at 2.7%. Sure, it could get better, maybe much better. However, in the last reading, it is heading lower & not higher
Thus, while I acknowledge the power of the AI rally that has made millionaires of a lot of people, I am a little more hesitant to to say that AI is going to change the near-term direction of the economy
If anything, in anticipation of a boom (that may not come), this actually may be a negative for labor market dynamics
I think I will stick with those old school indicators that work because they indicate how and where money flows throughout the economy
I decided the next day to take it a step further, and see what this means for GDP, after all the market cares about growth now:
Chart of the Day - GDP
Yesterday we got a GDP number. Ignore it please. I like to rant against it but let's face it, this was the second revision of a 2nd quarter number. We will get another revision. This is all backward-looking. Markets look forward
Yesterday, I also posted about labor productivity. This type of data is more useful. You see, we can figure out what GDP is going to be by looking at the formula: GDP = total hours worked * labor productivity
In this formula, total hours worked = total number of workers * avg hours worked. Thus but looking at a few components, and importantly the direction of those components, we can determine the future direction of growth, not the past direction of growth that GDP itself tells us
As I discussed yesterday, productivity had a nice bounce last year but is rolling back over. In spite of the presumed AI productivity boom, this number has not exploded higher by any stretch. That's the orange number
Now let's look at the other parts. The white line is the total labor force size. This number has been stalling out. Perhaps this is because of the 10k Boomers that retire each day
Some suggest this number is flawed because it doesn't capture the number of immigrants that have come into the labor mkt. Given these are undocumented workers, I am not sure how many jobs they actually are taking right now but this is possible
More importantly to me is looking at the average hours worked. This number is in blue. This number has the most near-term flexibility. When companies see more orders coming thru, they ask their workers to put in more hours before they look to hire anyone
Conversely, when times are getting tough, they cut the number of hours before they start to fire. In fact, a good recession indicator is the ratio of full time to part time workers. This ratio is collapsing
Putting it together, the total number of workers is stagnating most likely. The hours worked has fallen quite a bit and the productivity is rolling over. This all suggests that growth, which the mkt cares about now, is heading lower and not higher
That doesn't necessarily mean a recession or a hard landing. We could get a soft landing as 76% of investors think. The avg hours worked, the continuing jobless claims, the 818k jobs the BLS took away, & the stalling productivity all suggest there could be more downside than we think tho
Look at how the average hours worked leads the unemployment rate. This is what you would expect, companies cut hours before firing people but once they go down that path, it is hard to stop. We also see hours recover before the jobs do. Thus, the trend in hours is NOT a good trend.
Along the same lines, I saw this chart from Francois Trahan at The Macro Institute. When the ratio of full-time to part-time jobs plummets, we tend to have a recession. I look at all of these data points, and while Claudia Sahm herself tries to downplay the Sahm Rule, there are plenty of other indicators that suggest we are seeing meaningful deterioration in labor markets that is NOT being offset by productivity gains from AI. That has me worried.
It has me worried because many investors are focused solely on the rate cuts & think these are some sort of panacea. Sure, this will lower the cost of capital. It will lower mortgage rates. However, as I have shown before, it isn’t mortgage rates alone, but whether or not people have jobs. My simple index of mortgage rates and unemployment rate (inverted) has a better fit vs. the leading housing data than mortgage alone. When I compare to the NAHB Real Estate Index or to the Conference Board Leading Indicator of Building Permits, I understand why the housing market is so weak right now. This matters. This is the start of H.O.P.E. This has me worried.
So I will keep my eye on the 20-day moving average of the put-call ratios. This is the best short-term market indicator. Right now, the demand for puts is falling, giving a good sign for the market. That could indicate that futures will hold the important levels and move higher. I am not naive, however. I know we are at a precarious level in a seasonally poor month with deteriorating data. This could be unwound in a short period of time. My sense is the first part of the month may resonate with investors coming back to work and putting risk on, bankers doing deals, and a market focused on the benefits of rate cuts.
This first week of September will be chock full of labor market data. Productivity, JOLTS, ISM Employment, Challenger Job Cuts, Jobless Claims and finally Non-Farm Payrolls. There will be many looks. Maybe the market will focus on these others since the NFP number has been so massively revised. Even with their trouble assessing labor, the BLS keeps trying. In the chart above which I got from colleague Taek Pae, the BLS shows the very high unemployment rates in the major cities in the US. This definitely sets a negative tone heading into this holiday-shortened week.
From Labor Day to Labor Week to maybe Labor Concerns. This week will be critical.
Stay Vigilant
Thanks for the shout out, Richard. Good luck for your first week back at school (teaching).
Regarding AI: In the coming days, I will be publishing a series of articles on AI from an investor point of view...what is the current state of AI given what we have learned from recent Q2 earnings reports (not just NVDA's report LOL), where are the best places to invest now and leading into 2025 and which are the so-called AI sectors to avoid. Cheers!
wondering about this “I think I will stick with those old school indicators that work because they indicate how and where money flows throughout the economy” i was wondering if you could share these indicators?