Claudia Sahm’s Recession Watch

The Days Ahead:

• Retail sales and housing starts

This Week:

• We look at the Sahm Rule that’s been right on every recession for 70 years
• It’s not showing any sign of stress
• The Fed can’t possibly be right in its estimates for unemployment
• Oatly, yes, that Oatly, claims some weird things
• Can we have a recession with this much cash around?
• There’s a lot of excess savings
• Banks calm down, but smaller banks will have earnings pressures
• Debt ceiling update. Don’t rely on the gimmicks being discussed. 

Claudia Sahm’s Recession Watch

There are many recession indicators, ranging from the popular version of two quarters of negative real GDP, to the official NBER’s gnomic definition of “significant declines” across the economy that show “depth, duration and dispersion.” You’ll note nothing is specifically spelled out and there are “no fixed rules.” So, it can be fiendishly difficult to define what a recession is and whether one’s coming.

There were two consecutive falls in GDP in 2022 but the NBER never came close to calling it a recession. The reason was that while personal consumption, at around 78% of GDP, rose by 1.5% in the first quarter, goods imports fell 2.3% and trade fell 3.1%, and in the second quarter, inventories fell by 1.9%. Imports are around 17% of the economy and inventory changes are around 0.5%. Put differently, both categories have outsized changes on GDP calculations. If we temporarily exclude both, then GDP in the first quarter of 2022 would have been 1%, not -1.6% and second quarter GDP would have been 1.3%, not -0.6%. The NBER was right not to warn of any recession as both quarters showed temporary corrections in small parts of the economy. Neither showed depth, duration or dispersion.

Similarly, there have been plenty of times when there weren’t two consecutive drops of GDP but a recession happened. In 1970, over four quarters, growth ranged from -4.2% to 3.7% but with no quarterly drops. The peak-to-trough drop was around 1%. In 1960 and 2001, there were similar patterns, no consecutive quarterly drops and shallow recessions.

There are as many recession indicators out there as there are economists, think tanks, analysts, banks and central bankers. But one recession measure that’s very reliable is unemployment. There’s never been a recession where unemployment didn’t fall (I know, no great insight) and there’s never been a recession where unemployment didn’t rise by at least 2%. The fastest unemployment rise, outside of Covid-19, was the doubling of unemployment to 10% in 18 months to September 2009. The slowest that it rose, was from 4.8% to 7.0% in the 14 months to April 1960.

This was observed by Claudia Sahm, a Fed economist working at the Brookings Institution, in 2019. She noted that when unemployment starts to climb, it climbs quickly and, specifically, if the 3-month average of the unemployment rate is 0.5% above its low from the prior 12-months, we’re in a recession. Her findings are rather conveniently called the Sahm Rule and it’s published by the St. Louis Federal Reserve Bank. We played around with the data and below we show the 3-month moving average of unemployment, currently at 3.5%, and the 12-month low point of unemployment, which was in April of this year at 3.4%.

Right now, the difference between the two is 0.1% but if the next three months show average unemployment at 3.9%, then, according to the rule, we’re in a recession. Not that a recession is coming. We’re in one. The model has been correct for the last 62 years and has thrown up no “false positives” unlike, say, the inverted yield curve.

Quick explanation of the yield curve. Normally, borrowing rates are lower for short term loans and higher for long term loans. If you lend money to Apple for 24 hours, you may only need a 3% annual rate. What could go wrong? They ship a few less iPhones? You don’t really care. You get your money back tomorrow. But if you lend money to my not-so-bright second cousin for 20 years for his “Dallas-the-revenge-of-Miss Ellie” theme park in Anchorage, you may need a little higher rate. After all, it’s for a long time and it’s a bit of a niche market. (I’m mixing credit quality with maturity here and the CFAs will be mad at me but oh well).

Anyway, short-term borrowing rates are less than long-term borrowing rates in a normal world and that is a “positive yield curve.” But when the world gets worried about the economy weakening, long-term rates fall faster than short-term rates. That’s mainly because the Fed gets to control short-term rates and if it want rates high, say to fight inflation, then that’s where they’ll go.

But the Fed does not control long-term rates. Investors do. And if they think that the economy is going to weaken and that the Fed will cut then they will accept lower rates for long-term bonds. So, now the yield curve is “inverted” because short-term rates are higher than long-term rates. It normally happens when people think a recession is coming.

So, if you hear on TV “Well what about the inverted yield curve?” you can say, “Well, yes, it is a bit weird. Still, it may just be a false positive.”

Back to the Sahm Rule. The 2-Year Treasury to 10-Year Treasury (2s/10s) curve is useful but it’s often early by up to three years so has only marginal predictive usefulness. One can play with other curves, such as the 5-Year and 30-Year Treasury, which has a longer history than the 2s/10s but that too can invert years before a recession.

The fun part of the story is that Claudia Sahm wasn’t looking for a recession indicator. She was writing about making direct payments to all individuals when unemployment rose faster than 0.5%. The argument was that it would provide a quick growth stimulus before the recession deepened and unemployment spread. What she saw was that there’s no such thing as a recession where unemployment creeps up gradually. What she saw was that it grows in 0.1% to 0.2% increments, and then, wham, it starts to move very quickly in following months. So, her answer was to get stimulus payments to people to head off higher unemployment, lower incomes and a deeper recession.

That… hasn’t happened. But the model has proved useful. It’s also very intuitive. A 0.5% increase in unemployment in May 2023 would mean 6.6 million people unemployed or 1 million more than the April rate. If we saw three job prints of negative growth of 330,000 for the next three months, we’d be rightly worried. As more people are let go, the fewer remaining workers spend less and save more. That reduces demand, and pretty soon employers aren’t seeing enough business and start to lay off more workers. After all, how many employers come out and say, “That’s our last round of layoffs”? They can’t because they can’t possibly know and they’re fearful that the unemployment rate will accelerate quickly once it gets going. It’s the proverbial stone rolling down a hill. Slow at first and then very fast.

So, what’s our takeaway? Job growth is almost bound to slow. You can see the real-time Sahm Rule here and it’s in safe territory right now. The Fed thinks unemployment will average 4.5% in 2023. They’re almost certainly wrong on this. To get to 4.5% by year end, the monthly job loss would be 159,000 (you can play with the numbers here)…and the Sahm Rule would be triggered sometime in early fall.

We don’t think the labor market is about to crack that badly. Yes, job growth may fall. The long effect of numerous rate increases, bank problems, the NFIB’s hiring decisions and the ongoing slowdown in manufacturing all point to less employment. But that’s not the same as triggering a big jump in unemployment. Jobs are still around. Labor participation has plenty of room to grow and some industries, especially Covid-19-effected ones like entertainment and food services, employ less people than they did in February 2020, while the rest of the working population is now 1.1% above its pre-Covid-19 level.

The Sahm-rule may become relevant in the fall. But for now, the signs are good that employment remains in good shape. We’ll keep a close eye on the changes.

Oatly. Not All That It Seems

There was no shortage of high-profile IPOs that went bad in 2020 and 2021. Rates were low, some businesses were booming and the talk was all about unicorns, which are private companies worth more than $1 billion.

Some came to market and disappeared. Casper Sleep went from IPO to delisted in less than two years. Lyft went from $90 to $8 in three years. McAfee was relisted and delisted in less than two years. GoHealth ran into regulatory problems and saw its IPO price fall from $380 to $10 in three years. They’re all part of the cut and thrust of innovation, spikey start-ups and growth. That’s hard enough at the best of times so there will be inevitable failures among the successes.

Then there’s Oatly. Oatly came to market in April of 2021 at $30 and is now around $2.

­It’s a Swedish company that sells non-dairy milk products. So, if you like your milk from oats, they’re the one for you. It’s made some pretty bold claims. It got into trouble in Sweden for saying it tasted just like milk but made for humans. The Swedish milk industry was not amused, sued and won. It then launched an “Are you stupid?” ad campaign that said the public can’t taste the difference between Oatley and animal-based products. But it turns out people could and EU advertising authorities frown on claims that are not, well, er, true.

It has trademarked brands it has no intention of launching. Why? Because the words sound like mjölk (milk in Swedish) in other languages.

So far, so clever. However, things began to heat up recently. First up was the wide gap between Earnings Before Interest Tax and Depreciation and Amortization or EBITDA and adjusted EBITDA. Plenty of companies want you to look at adjusted EBITDAs because then they can exclude things like stock-based compensation, goodwill losses, acquisition expenses or impairments of asset losses. Uber famously excluded the cost of protective equipment for drivers. Oatly took a $397 million loss in 2022 and adjusted it to a loss of $267 million. In the first three months of 2023, it took a $76 million loss and adjusted it down to a $49 million loss. There were also some questions about whether its reported group sales match those recorded in individual markets. If true, that’s straight and simple revenue overstatement and, erm, is not good.

Second was a recent claim about “climate footprint” labels on its milk (sorry, not milk) cartons. It purported to show the entire carbon footprint of manufacturing and packaging over the whole lifecycle of “grower to grocer.” They called it CO2E, with the “E” meaning equivalent and shows how much greenhouse gases a carton of Oatly emitted over 100 years. It’s 0.41 for a regular carton of its Oat Drink.

There it is at the bottom right. The number comes from a company called Carbon Cloud which wants all companies to show their CO2E numbers. Lower numbers are good and higher numbers bad. Carbon Cloud wants to help companies get to lower numbers. The trouble is Carbon Cloud doesn’t reveal how they come up with the number, nor is there an agreed methodology on how to come up with the number. So how are we meant to know if it the lower number means anything?

Meanwhile, Oatly’s latest numbers showed they lose about $37 for every $100 in sales. In 2022 it was around $53.

This isn’t some big exposé of some shady stuff at Oatly. It’s more that a lot of hype came out in the markets in 2020 and 2021 and it’s only slowly unwinding. A company like Oatly may never be profitable. At its current rate of losses, it will blow through all of its equity buffer by the end of 2025. It’s a painful reminder, as we saw with Tupperware recently, that low rates and high expectations can keep a company going much longer than it deserves. Oatly will probably survive in some form. But for shareholders, the unicorn has been a nightmare.

Can You Have a Recession with This Much Cash Around?

Possibly, but it’s never happened. We’re familiar by now with the buildup in savings in the pandemic. The combination of government cash support and loans and the limited opportunities to buy stuff meant a rapid build up in the savings ratio, and cash holdings. From 2020 to mid-2021, personal savings went from $1.4 trillion to $6.0 trillion and the saving rate from 7% to 33%.

The savings rate is always a bit of a dodgy number as it’s the difference between two very large numbers. It takes all personal disposable income (currently $19.8 trillion), which is all income less taxes, and subtracts it from all personal outlays (currently $18.7 trillion), which is mostly spending and interest payment. The difference between the two is savings and is expressed as a percent of the first number. It’s subject to big revisions. In 2009, the Bureau of Economic Analysis (BEA) changed the methodology and found the savings rate was around 4% not 2%. They did it in 2017 and found it had been understated by around 3% almost every year from 1970. Today it’s around 5.1%, which is slightly below pre-Covid-19 levels.

But the savings rate doesn’t measure how much people have in savings. It just measures how much they’re not spending. If we look at bank deposits and money market funds (MMF), we see this:

The deposits and MMFs include both corporations and household accounts but the clear point is that they’re very high, both in absolute terms and as a percent of GDP. If we go further back, before MMFs grew to the $5.3 trillion they are today, we find that bank deposits were around 40% of GDP, grew to 60% pre-Covid-19 and are now around 65%. Savings tend to rise in a recession because workers worry about their jobs and cut back spending. But every way we look at it, current cash savings are at very high levels.

The next step is to figure out how much of the savings are excess savings. After all, we all keep some money at hand for bills, rainy days and spending. It’s a tricky number to come up with because you must adjust for drawdowns, moves to other savings vehicles (like stocks or bonds) and make some other calculations way over my head. Luckily the San Francisco Fed just did it and came up with $500 billion or 2% of GDP in excess savings.

Next, we need to find out who holds the excess savings and cash. After all, if it’s held by the top 1%, then it will probably stay in cash as that group tends not to spend much of every new dollar earned. But if it’s spread across other income groups, then there’s spendable cash out there that  would soften any downturn in the economy. It turns out they don’t exactly know but they do think it’s not all just with the top earners and frugal corporate treasurers.

For now, we think that there’s a large amount of cash, much higher than normal for this stage of the economy, and agree with the Fed that:

Despite recent rapid drawdowns of those savings, a large amount—around $500 billion—remains in the overall economy [and that households] have considerably more liquid funds at their disposal compared with the pre-pandemic period [and that] we expect the aggregate stock of excess savings will continue to support consumer spending at least into the fourth quarter of 2023.”

So, there’s a lot of money out there and historically, that’s a difficult backdrop in which to have a recession.

It’s a “glass half-full” version for sure. Others may say that cash hoarding is bad and that people won’t feel confident to go out and spend, especially if the debt ceiling turns into a bigger problem.  Fair enough, but we’d point to surveys like the NY Fed, and University of Michigan, which shows households seeing inflation coming down in the next year. That’s a good sign for spending.

We’ll know more in a month or two when the Flow of Funds (nnow known as Z.1 for some reason) come out in mid-June but for now, we’d take some comfort in the excess savings number. It may not be burning a hole in consumers’ pockets, but it’s a heck of a good buffer in the event of a slowdown.

Banks Calm Down

It’s been a rough couple of weeks with banks. We had solid and quick resolutions at the big two fails of Silicon Valley Bank (SVB) and First Republic (FRC), which had combined assets of over $429 billion, and of the smaller Signature Bank. Then, last week, this started to happen:

It shows the share price of some notable regional banks that range in assets from $33 billion to $90 billion.

As we’ve mentioned before, bank problems this time round are not bad loans (1989), bad securities (2007) or over leveraged balance sheets (1991). It’s a depositor problem. Basically, many banks pay around 0.4% to their depositors and make about 3.5% on their loans. That’s a fine business until the depositors leave to earn 4.5% on a government money market fund and the banks must fill the hole with expensive government borrowing at around 4.5% to 5.0%. At that point, banks are left with few options:

1. Sell the loans
2. Increase what they earn on the loans (but tough to do if they’re fixed-rate loans like mortgages)
3. Attract more low-cost deposits
4. Raise more capital to plug the gap
5. Throw in the towel and find a buyer

None of those is particularly attractive to a management that probably enjoys its independence.

We’d point out that none of the banks in the chart are under particular stress.

First, they each have a much broader deposit base. Some 94% of SVB’s deposits were uninsured and large. That always made them a risk and management were told so many times by the San Francisco Fed. They were ignored.

Second, they do not have the same reliance on government borrowings. At the end of March, 45% of FRC’s balance sheet was made up of lifelines from the Fed or the FHLB.

Third, they have a higher capital buffer when adjusted for book losses on loans and securities.

Finally, many smaller banks, like community banks with less than $50 billion in assets, have a stable source of funding. They have greater customer loyalty and deep knowledge of local borrowers. The New York Fed is a lot less worried about small banks than they are about mid-level banks with $50 to $250 billion in assets. It pays to be large and it pays to be small. The middle ground is a lot tougher.

Investors, however, know that the earnings power of many regional banks is low and not getting any better. Deposits are still leaving and few of the banks have a big enough book of variable-rate loans to offset higher funding costs. The banks need capital and they have only three sources.

1. Head over to the Fed (expensive).
2. Issue bonds (very expensive).
3. Raise equity (very, very expensive).

Or, of course, they can put themselves up for sale or just wait it out. The recent moves in their stock prices seem to show that if they are acquired, it will be at low prices, and if they do sit it out, it’s a long haul until they start earning decent returns again.

So, small bank stocks are probably not a great investment right now. They have to adjust to the new, higher rate environment and wait for loans made at low prices to run off. Shareholders should not expect much from them. We’d finish by pointing out that this is not remotely like 2007. The banks do not have large cross holdings of toxic assets. It is a national, not an international problem. The FDIC has moved fast and protected depositors. It’s all very dramatic but it’s not a crisis.

Debt Ceiling

The speaker of the House met with the President and other Congressional leaders. Not much came out of it but at least they’re talking. The quick version is that the House has a plan, which the White House doesn’t like and the Senate won’t pass. Everyone thinks default would be unthinkable and several economists have warned that the economy would drop by around 4% if some deal does not emerge.

In the last week, we saw the highest ever auction for a 1-month bill at 5.84%. The 1-month bill has only been around since 2001 but, still, it’s quite a price to pay. A bill maturing in early July yields 5.1% and one in October yields 5.0% so the price of the default in early June is around 0.7%.

We also saw that tax receipts continue to come in low:

We can understand why the non-withheld section is down. It’s mainly because capital gains taxes and stocks were a lot lower in 2022. The lower income tax is a puzzle but may be because of more IRS employees and so quick response times and fewer over payments.

Last week we put these as the likely outcomes:

1. Grand Bargain: A last minute deal with the White House, the Senate and the House. Until then everyone waits around to see who blinks. Markets won’t like it and we’ll see more volatility. Likelihood: 75%
2. Discharge Petition: House Democrats try to end run the process by moving a debt suspension bill out of committee and to the floor. Likelihood: 1%
3. Stop Gap: All agree to suspend the debt limit for a few months and link it to the September budget process. Then everyone plays their favorite game of closing national parks, partial government shut downs and a deal is done. Likelihood: 19%
4. The White House: invokes the 14^th Amendment about paying public debt and tells the Treasury business as normal. Likelihood: 1%
5. The Fed: steps in and dusts off the 2013 very cunning plan where they buy all bills at par or swap defaulted bills for a new ones or issue a “make-up” payment, the terms of which they get to choose. That’s part of their “promote orderly market functioning” role and don’t be surprised if they use it. Likelihood: 1%
6. Default: or more like a technical default with late bill maturity payments. Likelihood: 1%
7. Something Else: we haven’t thought of. Likelihood: 2%

The only change we’d make is that Option 5 is even less likely. The White House does not want to do any of the five quick fixes. These are:

1. Invoking the 14^th Amendment.
2. Minting the trillion dollar platinum coin.
3. Instructing the U.S. Treasury to pay debt interest over paying anyone else.
4. Issuing premium bonds, where face amounts would remain at $100 but coupons would be very high. For example, a $100 10-year bond with a 10% coupon would raise well over $200 for the Treasury but only the $100 counts to the debt limit.
5. Issuing Consols, which are interest-only perpetual bonds. They’ve been around for hundreds of years and they pay an income stream only with no principal repayment. As in #4, because there is no principal there’s no hit to the debt ceiling.

They’re all possible but the ensuing political and legal battles could get ugly. The Fed doesn’t really want to get involved either. Chair Powell was emphatic in last week’s press conference when asked what the Fed would do if the debt defaulted (our emphasis):

And, again, I would just say we don’t — no one should assume that the Fed can do — can really protect the economy and the financial system and our reputation globally from the damage that such — such an event might inflict.”

So, there’s no immediate respite. Expect volatility and “who blinks first” to continue.

The Bottom Line

Inflation came in low. Headline inflation has now declined every month since June 2022. The numbers would have been better if not for some strange goings on in the used car market. It’s down to fleet buyers restocking ahead of the summer rental season and a shortage of cars. Used car prices were up 4.4% on the month. The all-important shelter index, which is 43% of the core-inflation index, grew 0.5%, way down from its December peak of 0.8%. As we mentioned in earlier notes, it’s tracking down with the Zillow rent index.

Other signs of cooling include unemployment claims, which came in at 264,000 and the highest since October 2021. There’s probably more to come as the Challenger job cuts last week showed 337,000 layoffs in the first four months of 2023 compared to just 79,000 in 2022. Producer Prices were also less than expected with 2.3% on the headline and 3.2% on the core. Again, that last number peaked at 9.7% in early 2022. So, the Fed can pat itself on the back. Inflation is coming down.

The S&P 500 was up for the week as we write on Thursday morning, some hours before the close. We’re now 18% above the lows of last October and 14% below the record high of early January 2022.

As we mentioned last week we like what we’ve seen coming from the earnings season. Around 75% of companies have reported an upside earnings or sales surprise and overall earnings are down 2.2% compared to estimates of a 7% decline just a few months ago.

The 10-Year Treasury remains in a narrow trading range of 3.4% to 3.6%.  Few people expect the Fed to raise rates next month, especially with this week’s inflation numbers. The debt ceiling looms of course. If the problem escalates, we’d expect longer Treasuries to rally. After all, it remains the ultimate haven asset class.

Subscribe below for our investment updates.

---

Art: Stephanie Holman (b. 1967)