I expected improved margins, but was uncertain about volumes. It seems my expectations and uncertainties were exactly the same as the market’s. While some carriers reported near-record operating ratios, volume was generally flat. Consequently, everyone sold off but XPO (which was my identified short, fml).

I also overestimated the market’s response to high diesel prices. Counterintuitive to the uninitiated, LTL revenue and margins improve with high diesel through fuel surchages. This was not the secret I thought it would be, and no carrier was rewarded for margin improvements on the “junky” revenue of fuel surcharges. This is not that surprising, because valuations were broadly high (and still are, despite the subsequent sell-off) and required a sign of increased demand, not just the sustained operational discipline that most public carriers already have.

Can I say that my identified KPIs were the correct triggers? I don’t know: I can only say they were not falsified.

I’m working on a much deeper article on the LTL industry that has charts of capital structure, operational performance and leverage, network density, and formulas for the relationship between KPIs. In the meantime, here’s my report card comparing my prediction against the actual results and price movement.

ARCB.

• Trigger: Improved weight (reduced linehaul miles is good, too)

• Quote: “The big question mark for this quarter is that ARCB is also going to have a 70 million dollar impairment and 5 million dollars of added costs for downsizing. For these reasons, I do not think the set up is that favorable for today.”

• KPI performance:

  • Weight/shipment: +8%

  • Tonnage/day +4.9%.

  • Linehaul: Slight increase

• Earnings day price action: -5.4%

• Postmortem: Because of impairments, I did not think the setup was good, and it wasn’t. But the LTL KPIs actually look pretty good; at last ArcB is getting heavier freight, though still hauling it for way too long (diluting density). As the runt of the litter, it has the best way to come back up. One to watch.
  
  Also, ArcB has a brokerage business, and truck brokerages are in a weird time. I didn’t account for how this would affect consolidated performance.

XPO

• Trigger: Margin expansion

• Quote: “XPO is also likely the most sensitive to fuel surcharges, so it may mechanically have the best beat.”

• KPI performance:

  • OR: -300bps

  • Yield ex-fuel: +4.4%

• Earnings price action: -0.1%

• Postmortem: XPO outperformed its peers here, but its valuation is exceedingly high. I have no strong opinion on XPO at the time of writing.

ODFL

• Trigger: Improved volume

• Quote: “However, investors are already paying richly for ODFL’s quality and I see few scenarios with much upside in terms of share price. Really, though, ODFL is one for the compounder bros; if that’s you, keep it on a watchlist and feast whenever the industry finally bottoms out.”

• KPI performance: Tonnage improved very modestly, with fuel surcharges driving near historic ORs.

• Earnings price action: -1.5%

• Postmortem: This is inline with expectations. Without improved volume, ODFL’s valuation doesn’t have anywhere to go.

Saia

• Trigger: realized operating leverage ex-fuel.

• Quote: “The key untested constraint is how the network density will perform with a number of new nodes. There are a lot of new operational kinks to work out with such added volume in the northeast. …Given fuel surcharges, revenue should grow from 2025 Q2 by 8% at the quarter’s OR is likely below 86%. What would really be bullish for SAIA is improved yield ex-fuel with stable or improving volumes.”

• KPI performance: modest shipment improvement with no ex-fuel yield improvement

• Earnings price action: -11.1%

• Postmortem: Saia did not meet volume improvements or exfuel OR gains. Further comments in the earnings call indicate downward adjustments for the year. These were my triggers and they weren’t met. Saia was my pick of the week, so I cringe.

Bonus tables

Here are some tables that are from my research in progress. These are more expanded views of the data in snippets above. Pay $8 or GTFO.

The next article will visualize this data.

Summing up

As I reviewed this, I realize my prediction weren’t as bad as they first seemed. But what was very stupid were my long/short picks: long ArcB and SAIA, short ODFL and XPO. As it turned out, my longs did the worst on price changes and my shorts did the least bad. Fortunately the shorts didn’t actually go long or I’d have entered a pure long/short liquidity ouroborous:

Saia is excusable, I suppose. I identified a trigger that wasn’t met and it fell. ODFL call was correct based on trigger not meeting. What’s strange is that I felt the need to add ArcB to the long and XPO to the short for that day: by my own comments, I should have flipped this! XPO had the best margins to gain in the shorterm and ARCB had the most to lose from EPS. To be fair, I added the L/S section as hasty tack-on, but it seems I picked ArcB and XPO purely based on their multiples, against my own prediction.

To be transparent, I did take a real, though miniscule, position: 1.5x short ODFL and 1x long Saia. This didn’t work out and I closed with a tiny, but permanent loss of capital.

Since the prints, the whole sector has sold off a bit more, and PMI index continues to climb.

Why play earnings?

I’m of the opinion that retail investors should not play earnings. If a stock has good liquidity, institutions will pay a lot to beat other traders on earnings day. They know more about demand conditions because they subscribe to expensive alt-data services. In the month before the print, they’ll spend $100k on expert networks and deploy a team of analysts to scrutinize the technical and fundamental factors. They have better trading infrastructure, so they’ll beat you off the starting line. And all these people are competing with not just you but with each other. It’s like a race between Usain Bolt, Justin Gatlin, and the winner of a Sunday fun run.

Earnings are attractive trading events, obviously, because they have a high degree of volatility. But volatility means high statistical variance. So even if you win, are you sure you have an edge? You can win at the craps table, but you can’t beat craps.

That said, forecasting is an important skill to develop. Playing earnings has three elements:

1. Identify the essential KPIs or second derivative.

2. Predict how the market price responds when these KPIs come in, at, or above expectations

3. Forecast the actual KPIs that will be reported

All three are difficult, but valuable exercises. So I will continue to make predictions as long as I cover LTL and other stocks that interest me. Incrementally I hope to better understand what makes the price move up and down.

Today, one share of sells for $19.73. Per its announced liquidation plan, this price puts it at a healthy discount to the low end of the expected distribution range of $24.73 and $32.69 (price adjusted for payout of first $5 special dividend).

This range presents an upside between 25.5% and 65.9%. However, there is also a timing risk, and of course, a downside risk. What I like about events though, is they can be modeled in a much simpler set of probability trees than the life of long-lived businesses. That’s what I do here.

Background

The liquidation range is based on an external valuation by Eastdil Secured. An investor presentation breaks out Eastdil’s work by property class and development status.

Since announcing the liquidation, the company has sold two properties and paid a special dividend of $5 per share. That leaves the remaining distribution of between $24.73 and $32.69.

The liquidation plan in the 2026 proxy states that Stratus aims to sell the assets in 2 years. If it can’t after 3 years, it will turn what remains over to a liquidation trust. For US investors, distributions reduce the cost basis, rather than count as income.

On August 10th, the stock will delist from the NASDAQ. Henceforth, shareholders should expect low liquidity and minimal communication. Since the ex-dividend day and de-listing announcement, the stock price has fallen a good bit and the spread between it and the new range has substantially widened.

These are quality properties: high-end shopping centers and residential housing and lots. Management has a good track record over thirty years. Since the Austin real estate hangover, they have been buying up stock and rejecting low ball offers, and the CEO, William H. Armstrong, owns 8.1% of standing shares. A good payday for him is a good payday for all shareholders. An institutional investor, Oasis Management, also has an active stake in about 10% of the company, though this has been trimmed by ~400 basis points since 2024 (when a failed tender that Oasis supported didn’t work out).

Estimating the expected value

My model simplifies the range of outcomes into a binary scenario of “good” and “bad.” I am less worried about quantifying the entire range of possible distributions than I am about understanding the assumptions behind the p of success.

In this simplification, I also bound outcomes to a fairly pessimistic set of parameters:

• The bottom end of the Eastdil valuation correctly assesses today’s fair market value today.

• In the success scenario, the liquidation yields this exact fair market value (i.e. bottom of range).

• In the failure scenario, the gross assets decline by 30%. Because of leverage, this decreases net assets by 40% (this extreme decline is a brute approximation, ignorant of tax, debt, and JV considerations)1.

• In either outcome, the liquidation distributes in one lump sum at the end of 3 years.

• The outcome is binary: either success scenario or failure scenario, one lump payment.

With these scenarios, if the success scenario is realized, the 3-year CAGR is 7.82%

If the failure scenario is realized, it is -9.38%.

An expected value model sums the probability-adjusted CAGRs of the two outcomes. If p denotes the probability of success:

Using a probability of success of 15% and our computed CAGRs, this formula gives an expected CAGR of 5.24%.

With treasuries approaching the 5% mark, this expected value is terrible for an illiquid security. However, it is entirely possible that the actual scenario could work out much more favorably for investors. This calculation compresses many conditions into two outcomes, both of which may err on the side of caution.

At the end of this article, I’ll paste my Python code so you can play with some scenarios. But be careful with greedy predictions! While I may have understated the upside, a number of tail risks could widen the downside too.

So I’ll outline the conditions that tilt the expected value to one side or the other.

Conditions that would improve expected return

My model intentionally takes a pessimistic view. Some conditions that would raise the expected value:

• The success outcome could, by Stratus’ own evaluation, be much better.

• The liquidation plan guides for 2 years of windup, not three.

• A failed outcome may return more than the floor I set, even if it returns less than the purchase price.

• The distribution will surely not happen in a lump sum. Getting 99% of the money in the first year changes the IRR significantly.

• Probably most significantly, the probability of this failure case is quite extreme―assigning a 15% probability of a 30% decline in gross asset value.

  
  Really, 30% is extreme. Looking at data from FRED, this is the greatest amount by which commercial real-estate fell year over year in the GFC, though median house values fell by “only” 19.1% from its peak in 2007 to its trough in 2009. I do not consider a crisis of that magnitude so likely in the next three years! However, location and type of property mixing means this is not an apples-to-oranges comparison. Austin is a high-beta real-estate market, and deep crashes are more frequent on a regional level than they are on a national level2. On the other hand, Austin real estate already had a boom in 2022, from which prices and occupancy have already fallen significantly. While the trend is obviously bad for Stratus from the last few years, supply and demand may already have already started clearing.

Still, in Q1 of 2026, multifamily vacancies are still at 13%. Demand side is not great, with hugely slowed net migration. This leads me to the unaccounted for range of downsides.

Conditions that would depress expected value.

• The failure outcome could be worse than my expected range. While my failure outcome is severe, worse property declines have happened. I also may underestimate severe tail events—flooding, expensive litigation, liquidation incompetence, political revolutions, and twin tornados. While unlikely, none of these are unprecedented, and extreme cases will wipe out common equity.

• The success outcome could be worse than the expected range. Austin real estate has had a prolonged supply glut. While indications seem to be that the market is turning, indications are not certainties.
  
  Rising interest rates could drag even a recovering Austin market, which would also increase the opportunity cost of foregoing liquid short term cash equivalents.. A bit more vagueposting by Warsh may be enough to do the trick.

• Perhaps most significantly, the time to full liquidation could drag on longer than three years. These early asset sales were successful, but already planned. As we move farther down the asset list, some are still in construction or “planned.” I assume the planned lots will be sold before construction, but that’s an assumption.

Unfortunately, reassessed outcomes seem fairly common for real-estate windups. The Special Situations Investors site has many pitches on liquidations, which seem to end with disappointment more often than not. Scanning the comments over different articles, I see a lot of updates of reduced ranges and extended timelines.

The New York REIT pitch is particularly harrowing: it started looking like a slam dunk with a generous range and gorgeous real estate, yet the prices whittled down and down as the process dragged on and on with re-financing and a host of other problems. That post is from 2016: it seems the liquidation is still ongoing ten years later!

In uncertainty lies the spread

If information were perfect and future was known, there would be no spread. Whether you like a bet here comes down to your tolerance to hold illiquid assets and your measure of faith: faith in management’s ability to flip properties quickly, faith in the quality of Stratus properties, faith in the Austin real estate market. Liquidity considerations come down to style and portfolio constraints, but the real estate considerations require expert judgement.

How appealing this is today is up to you; I have tried to present a simple framework that doesn’t obscure decision-making in a procrustean bed of parameters. However, as the stock delists and time drags on with no news, there is a good chance the the price slides further. At some discount to asset value, things don’t have to go terribly right to go well.

Appendix and code

As always, I appreciate all feedback. I will be glad to hear any critique of the model or assumptions.

Code

Here’s the code I did for calculations, you can play with the parameters yourself.

price = 19.73
min_range = 24.73  
good_value = min_range
years = 3

debt = 176.8
gross_value = 675.8
shock = 0.30  # assumed decline in underlying property values in the bad case
nav_pretax = gross_value - debt
nav_shocked = gross_value * (1 - shock) - debt
pct_nav_decline = (nav_shocked - nav_pretax) / nav_pretax

bad_value = min_range * (1 + pct_nav_decline)
roi_good = good_value / price
roi_bad = bad_value / price
cagr_good = roi_good ** (1 / years) - 1
cagr_bad = roi_bad ** (1 / years) - 1

p_bad = 0.15
p_good = 1 - p_bad
expected_cagr = p_good * cagr_good + p_bad * cagr_bad
# p /g

rows = [
    ("good-case value", f"${good_value:.2f}"),
    ("good-case cagr", f"{cagr_good*100:.2f}%"),
    ("bad-case value", f"${bad_value:.2f}"),
    ("bad-case cagr", f"{cagr_bad*100:.2f}%"),
    ("bad-case value", f"${bad_value:.2f}"),
    (f"CAGR if good ({p_good:.0%})", f"{cagr_good*100:.2f}%"),
    (f"CAGR if bad ({p_bad:.0%})", f"{cagr_bad*100:.2f}%"),
    ("expected CAGR", f"{expected_cagr*100:.2f}%"),
]
label_width = max(len(label) for label, _ in rows)

print()
for label, value in rows:
    print(f"  {label:<{label_width}} = {value}")
print()

Events to watch

(in proxy)

It’s an interesting time for LTL, because demand indicators are bullish―PMI is consecutively high and truckload capacity is tight―and the high diesel prices should mechanically increase revenue and operating income. However valuations are already high across the industry, and the volatility and unevenness across sectors in the US may net out to a relatively flat demand for freight transport.

In this post, I go down the list of the companies that report this week and try to identify the triggers that could result in improved short-term valuations. This is short-term forecasting based on multiples and margin ranges―not for value investors!

To emphasize, these triggers are brief and informal. I am working on a more rigorous financial analysis using a variety of data. When I’ve finished with that, I’ll be able to make an updated list with an actual range of numbers and some sense of catalyst probabilities. But this earnings week provides an early place for me to test my loosely held assumptions. Besides, it’s fun to predict.

Since this information has almost no value to people who are not specifically looking at the valuation dynamics of this space in this exact moment in time, the analysis presently is behind a paywall. For people who are interested, the cost of a monthly subscription is $8. Pay or don’t, I don’t really care―you won’t be missing the opportunity of a life time. More and better analysis is coming.

You can read my other posts on LTL to determine my grasp of the landscape:

• 1. Density is the name of the Game. An overview of the industry.

• 2. What drives LTL top and bottom lines? A map of factors that contribute to revenue and operating expenses.

• Appendix: The Anti-Guide to LTL, by Alain Bedard. A company-specific look at dysfunction arising from an LTL acquisition.

Comp tables

July 27 comps.

A chart of ORs from my research.

TFI

The week kicked off with TFI International reporting its earnings for Q2 2026 on Monday. It was a good quarter for TFI: every division substantially grew operating income over H1 2025:

• LTL operating income increased 17%

• Truckload increased 50%

• The asset light division increased by 32%.

Of course, this report is most concerned with the long-beleaguered LTL division. However, the truckload gains are worth highlighting, since truckload tightness causes freight to flow to LTL.

TFI’s CEO Bedard reports that truckload market today is something that he hasn’t seen in 30 years. While very tight, the tightness is a matter of supply constraint, not improved demand. That’s half good for LTL (tight TL), half bad (no improved demand).

Bedard has not seen truckload capacity overflow into LTL, but he mentioned speaking to a CEO of another company that had seen this.

The LTL business had much improved volume, but TFI admittedly underpriced itself in the market. They are raising prices, which may depress volume. Improving service levels is the key improvement for TFI.

Trigger: Sustained volume with better pricing

Now for the carriers who haven’t reported…

ArcBest

Arc Best is consistently the laggard of the group, but it also has the most to gain. The charts I made of KPIs across the industry shows an interesting story of ArcB’s weakness.

Read more

“We say that the tiger is always the last one to survive in the jungle. And at TForce Freight today, I mean, we’re a bunch of fat cows, okay? We are way too fat.”

I have never seen earnings calls like the ones given by Alain Bedard, the founder and CEO of Canadian trucking firm TFI International. The gruff French-Canadian curses, self-deprecates, and diagnoses issues with a blunt and folksy candor.

Sometimes, the best way to learn is by anti-example. Bedard and TFI have a long history of successful operations and acquisitions in Canada across a variety of transport and industrial segments. However, TFI’s US LTL sector has been fraught with problems since it acquired the division from UPS in 2021. Because Bedárd is actually an experienced and accomplished operator in transport and LTL, and because he is so candid, his comments provide a sharp picture of how things can go wrong.

In fact, for almost every point that I make in my article about LTL revenue and costs, What Drives the LTL Top and Bottom Lines, Bedard has a statement about how things can go wrong on that exact point. If my article was a guide to LTL operations, let’s call this Alain Bedárd’s anti-guide.

Acquiring disfunction

TFI has a long history of M&A. Its main driver of growth is the rollup: it looks for targets that trade at distressed multiples and tries to improve their performance by improving the operation and reducing overhead. A special target of interest is what Bedárd calls “the orphan,” neglected divisions that exist in larger organizations.

First, our focus has always been, can we find an orphan somewhere? Okay, because I like orphan. Because nobody care about an orphan, right? If you think about UPS Freight, it was an orphan of UPS, because UPS is a fantastic company, but for them, LTL was, "We don't really care about that." … It's difficult to do, so we don't have much competition on an orphan deal. Nobody's stupid enough to try to buy an orphan. That's our focus. That's number one.

— TFI Investor Day 2022 (Nov. 10, 2022)

However, things often aren’t neglected for their superior performance. As it turned out, UPS Freight was an organization that is full of issues. Here’s Bedard speaking in the early days following the acquisition:

The UPS truckload division that we got with the acquisition of UPS Freight was a disaster of a division because it’s a dedicated division, and the pricing with customer didn’t make any sense. The commitment didn’t make any sense. For instance, they committed to all freight, but they didn’t have the power or they didn’t have the driver... We will lose money because of the poor management that was there previously. — TFI Q3 2021

As you’ll see from the following quotes, this early warning understated the many issues to come.

Service quality

In What Drives the LTL Top and Bottom Lines, the primary way that LTL carriers can differentiate themselves is through superior service. Fortunately for carriers, shippers of freight aren’t as finicky about service levels as are the customers of high-class hotels.

Good service in LTL boils down to a few measurable criteria:

1. Pick up freight on time.

2. Don’t damage it.

3. Deliver it on time.

4. Bill correctly.

On all four of these points, TForce has suffered.

Pickups

UPS Freight was not doing well on pickups.

…like I said many, many times, missed pickup is a disaster in the sense that the customer is waiting for you to pick up the freight and you don’t show up. I mean, that doesn’t help your reputation. That doesn’t help the churn into your business.”

— TFI Q1 2025

The same point reiterated on the next call:

Our missed pickup, which was a cancer for us, a cancer because like nobody cared, we were all the way up to 4% three years ago.

… You got to be stupid to miss 3% of your shipment…because that's 3% of shipping that you're not going to have because you just miss it because you're stupid.

— TFI Q2 2025

A few years prior to these comments, Bedárd explained how the culture of missed pickups happened. In LTL, drivers deliver freight in the morning and pick it up in the afternoon. To get home on time, pickup was an afterthought for drivers and dispatchers.

you think about an LTL company, the focus is on delivery, right? Because that's the first thing you do in the morning. The guy leaves and does his delivery. And then if he has time left, then he does the pickup.

…yes, we have to deliver the freight in the morning, but your role number one is to pick up freight. Why? Because if you don't pick up freight, then you don't have the revenue.

LTL carriers recognize revenue when they deliver freight. The obvious corollary to this is that, to deliver freight, it must first be picked up. In busy markets like Los Angeles, missing pickups not only eroded customer trust but it caused lost revenue, since these shippers rebook with another carrier instead of waiting another day.

Damage

LTL freight touches many hands, so the chance of damage is much higher than it is in simpler modalities.

When freight is damaged, consignees need to report, shippers need to submit claims, and carriers ultimately reimburse. No one wants this lose-lose-lose situation, so LTL carriers aim to keep the claims ratio as low as possible.

When you have claims, customers are not happy because you have claims because either you break the guy’s stuff or you lose it, right?

—TFI Q2 2022

But what “low” means is relative. For UPS freight, the bar was pretty high. At one point, the claims ratio was around 4%, and TForce was paying more than 1% of its revenue out to claims!

“Our claim ratio per dollar revenue, okay, it’s a major improvement. You know, we used to be at 1% of revenue, which is a disaster, okay? You know, the guys told me,‘Well, we used to be at 2%, now we’re at 1%.’ I said, ‘One is terrible,’ okay? You gotta be at 0.2, 0.25.”

— Q3, 2022

There was a particular issue with dock workers packing spillable commodities:

As an example, you know, we spend about $350,000 a month right now for spills. I mean, I fell off my chair when I learned that. I said, “What?” She said, “Yeah, that’s what we spend.” “Impossible. No, there’s a mistake.” “No, it’s not a mistake. It’s reality.” Now we are, you know, working with the team to address. This is not normal, guys.

I mean, what are we doing, right? It’s the way you load the trailers, and there’s many things like that.

— TFI Q2, 2021

Eventually the ratio did get under 0.8%, and then reversed. The reason for the reversal is surprising:

“So the reason we’re at 0.9% is that we used to break so much stuff that we have a store in Richmond that was selling the stuff that we were breaking. But now we break less stuff, so we don’t have this revenue that goes against the expense... And now we’re just expensing, okay, the stuff that we’re breaking now, okay, or losing now.”

— TFI Q4 2024

Deliveries

On the other side of pick up is delivery. TForce did not do well there, either:

I would say we were a little late to the party on on-time service... the market softened a little bit. Our competitors picked up their on-time service, and we were few months behind, I think, on that.

— TFI Investor Day 2022]()

The situation has improved, though Bedard stays realistic:

Q: Just any other update on the on-time deliveries or other service metrics that have been kind of elusive?

A: Oh, yeah. Yes. Yes. Yes. Yes. Yes. Yes. Yes. Yes. Well, the on-time delivery, we are improving. Absolutely. But we’ll never be a 99.9 guy. So the goal is to be something around a 98.

— TFI Q4 2024

Billing

Transparency in pricing and billing is a part of customer service. Billing dysfunction is another example of a lose-lose situation, where carrier and customer both waste time on administration and value is destroyed for all involved. As you have probably guessed, UPS Freight had some problems in its billing.

Credit is the lubricant of the economic machine. Through accounts receivables and payable, buyers and suppliers can abstract away payment from delivery, saving time through specialization and reduced administration. UPS Freight, unfortunately, hadn’t quite modernized. While it seems hard to imagine, truckers were responsible for collecting cash from each delivery:

I’ll give you another stupid example, COD. We’re not in 1965, we’re in 2021. COD, we’re not gonna do that because it’s a mess, right?

—Q2 2021

The disorganization caused by messy billing sprawled across the organization:

“I mean, this is something I’ve never seen in my life, how bad of a system that we have, how many mistakes we make. I mean, I think that for, let’s say, 22,000 shipments, we will probably issue like 35,000 invoices because we bill, we credit, we rebill. I mean, it’s just a nightmare. This is not something new. This has been going on for years and years and years, but it’s never been addressed.”

— TFI Q4 2023

To drive home the point of this section, billing irregularities causes customer dissatisfaction.

“You know, if you don’t bill a customer properly because you’re so disorganized that your master file has been kept up by a bunch of guys that don’t know what they’re doing, okay, in the Philippines, as an example, well, for sure the customer is not gonna be happy because you can’t bill the guy properly. Then he’s got another issue with you because then you’re trying to collect something that don’t make sense.”

— TFI Q3 2022

Underinvestment in tangible and operational drag

It’s interesting to note that all the service dysfunctions in the past section relate directly to under-investment in capital expenditures and the use of assets that are likely well beyond their depreciation schedule.

“The P&D software that we’re using dates back about 20 years, right? So it was good 20 years ago, okay?” — TFI Q4 2024

Having assets that are mismatched to the present moment drags an operation, bloating expenses while reducing service quality. Underinvestment in trucking fleet and IT contributed substantially to dysfunctions of the previous section. On the other side, too much investment in terminal capacity diluted the density of the network. But before I get to that all-important topic of density, let’s look at how the legacy fleet and IT systems bloated operating expenses.

Aging Fleet

Inspiring words at the 2022 investor day:

“We just weren’t good at managing equipment at UPS. I mean, I would say we were borderline hoarders, you know, possibly need some therapy… Our equipment is very, very old… Our maintenance costs are high. Our fuel economy, worst in the industry, I guarantee.”

— TFI Investor Day 2022

Trucks are generally maintenance capex for LTL carriers; their depreciation schedule is fast and every new generation improves fuel efficiency. UPS Freight’s fleet had an average of 8 years. Compared to ODFL or Saia’s average age of 3.7, this is ancient. The aging fleet caused a number of issues for T-Force:

“The other issues we have is UPS truckload fleet was really, really old. A piece of shit, you know, running 2014, 2015. No safety in the truck, no cameras on the truck. I mean, nothing. It was like a ragtag fleet.” — TFI Q3 2021

A year later, the trucks were still “pieces of shit.” In the context of worker pride:

….everything is falling apart, and the guy’s got a 2008 truck, and it’s like a piece of shit, excuse the expression. I mean, that doesn’t help, you know, the pride of your employee, and that reflects also to the customer to a certain degree

— TFI Q3, 2022

Besides the pride factor, this ragtag fleet also directly affected service levels through breakdowns:

“The fact that we run an old fleet is not helping us. It’s not helping us with our drivers. It’s not helping us with our customers… You got an old 2008 trucks and it breaks down, well, the service is gonna suffer, right?… Today we drive 2006, 2007. Think about that. I mean, we’re in 2022.” — TFI Q4 2021

Service quality affects the revenue side, but the old trucks also increase operational expense through poor fuel economy and increased maintenance costs.

“Fuel economy is the same. It’s the shit. It’s bad. Why? We have the same trucks.” — TFI Q2 2021

Bedard gives some numbers on cost per mile:

“Equipment has been delayed. That’s a little bit of an issue because our MPG on old equipment is just the shit. Our maintenance cost on an old truck is $0.45 a mile, which is about $0.40 more than the normal.” — TFI Q4 2021

Old equipment and purchased transport

An interesting side effect of the “ragtag fleet” is how it forced TFI to rely on purchased transport. In the LTL industry, some linehaul miles are outsourced to rail and full truckload companies. The poor fuel efficiency and reliability of the legacy fleet forced TFI’s hand:

“From day one, we were not able to move away from rail because our fleet, our trucks were so bad that it was just a problem.” — TFI Q2 2025

Using rail for purchased transport can decrease linehaul costs, but leaves a carrier control vulnerable to service issues that are out of their control:

“If you’re late, and the reason you’re late is because your linehaul is late. And the reason your linehaul is late is because you use the rail. Well, you got a big problem…

…you can’t [move to road] when your fleet average age is eight years, right, because you got old trucks. And old trucks, they break down all the time. They’re always in the shop.” — TFI Q1 2024

For comparison, the best-in-class carrier, ODFL, spends less than 2% of its on purchased transport.

Legacy IT and Operational drag

I mentioned how legacy IT systems were causing billing issues, which cause customer aggravation and drive revenue away. However, the legacy systems also created maintenance cost and administrative bloat.

At TForce Freight today, I mean, we’re a bunch of fat cows, okay? We are way too fat. We have too many clerks. We have too many because our technology is so old, okay, that it takes an army of people to service our customer.” — TFI Q1 2024

Along with the P&D software that was good twenty years ago, the linehaul software also caused unnecessary administrative work:

We have a very poor system to manage our line haul network, right?… It’s very manual. We have five different systems that we have to use to go and optimize our network. It takes about a week to do it. — TFI Investor Day 2022

Poor data and visibility compounds inefficiencies. An LTL operation transports freight across a network of nodes, where each node is represented by a terminal. For the network to function optimally, each node needs to have ownership of its individual performance. This responsibility falls to the terminal manager. At TForce, however, the poor data integration and lack of visibility impeded the means for their managers to have agency over their operations:

“We know what’s going on. We don’t have that TForce Freight today, right? The manager is not responsible for all the costs. He doesn’t know. He’s got no financial information. Now that we are running on TFI’s financial, now we are in a position to slowly implement that at the terminal level so that our guys could start managing the business the way it should be managed: at the terminal.” — TFI Q4 2022

Density, or lack thereof

“In the U.S., I mean, the density of our LTL operation is the shit, okay? We’re working on it.” — TFI Q1 2023

Density is the most important feature of an LTL network. The more operational time spent picking up and delivering freight, the more revenue a carrier earns per run. Tforce has been plagued by density issues since it was acquired into existence.

Take terminal capacity. Generally, an LTL carrier measures capacity by its number of terminals, and, on a more granular level, the number of doors. Unlike fleet size, superior firms invest in terminals as growth assets, with the intention to build up capacity to absorb demand when the market is tight.

However, if a firm doesn’t have enough trucks or demand to fill the network, too much terminal real-estate dilutes density. Here Bedard compares the TForce US density against the superior operations of TFI’s Canadian LTL operation.

Right now we are operating a terminal with 40 bills a day, okay? You know, guys, does it make sense if you’re a 30,000 bills a day company to be running a terminal in a 60,000 people city? Does that make sense?… If we have 20,000 trucks in the U.S., it makes sense. If you have only 3,000, you need more density, right?… We need to operate terminals where we have density of population and density of customers… If you look at what we do in Canada, I mean, our costs are way better than our costs of what we’re running in the U.S. Why is that? Because we’re so dense.” — TFI Q3 2021

Density leaves a network through multiple forms. First, if too many terminals are clustered together and there isn’t demand to fill capacity, the linehaul trucks that haul consolidated freight across the country are required to either make multiple stops in one city, or run with half-empty trailers.

“If I look at a city, I’ll take the example of Chicago. I mean, to service Chicago, do you need four or three or four or five terminals? You don’t need that. I mean, it’s a nightmare for the line haul.” — TFI Q2 2021

On the other hand, having a giant terminal with insufficient demand results in massive under-utilization:

We have a terminal in Rialto in California, a fantastic terminal, a diamond terminal… I’ve got 60 shipments a day. Think about that, 60 shipments a day in a huge terminal”

However, real density issues are revealed in the pick-up-and-delivery runs, which traverse the last and first miles of every LTL haul. For a carrier, every mile that is not spent picking up or dropping off freight is just fuel and labor cost.

A sparse P&D route results in empty miles.

“We have guys driving 150 miles, okay? That doesn’t make any sense. You know? Because if this guy drives at 40 miles an hour, that means that he’s driving about 3-4 hours a day. When he’s driving, he’s not picking up freight, and when he’s driving, he’s spending money. He’s burning fuel. He’s burning tires.” — TFI Q4 2021

Bedard provides some numbers on the miles driven.

“We drive more than 6 million miles a month to deliver freight, us, excluding the line haul. To me, we should do no more than 3, maybe 3.5… Why would you go 75 miles away [from a hub]?” — TFI Q1 2022

Here, the freight deliveries are described as 300-mile “milk runs:”

Another leg is to, you know, pick up more freight around your terminal so you’re not stupid and having milk run of 300 miles or 200 miles that the guy keeps driving for about 65% or 70% of his time spending money and not picking up any freight. Also it’s the weight per shipment, so you got to be focused on the freight that fits.

You don’t want freight that is so light that it weighs 200 pounds because you’re paid by the pound, by the hundred pound, right? — TFI Q2 2022

Density as expressed of stops per mile:

They have to drive an average of about 10 miles between each and every stop, which is nonsense. We drive less than five miles in Canada.— TFI Q4 2023

The problem of density is not only a route optimization problem. It is also an issue of controlling the quality of freight. The culture of UPS Freight / TForce has been to accept all freight, no questions asked. This results in uneconomically sparse routes. In fact, an early task in the acquisition was to slash the amount of freight being delivered:

Let's say today we're doing about 23,000-24,000 shipments a day in, you know, in 1st month of the year, January. The plan right now sits at about 25 by the end of the year, okay? I mean, it's not a very tall order for our sales department because don't forget, when we bought the company, we were at 32,000 shipments a day, right? Now we're down to 23, 22, 23, 24. We took a hell of a beating, but we had to. We had to get rid of that freight that didn't fit. —TFI Q4 2022

So, questions remains: if TFI knows what the problem is, why haven’t things gotten better? Can they?

Turning around and around

Consider the interconnections between the topics mentioned here. The old trucks caused issues in fuel economy and maintenance, to the point TFI had to overspend on purchased transport. Breakdown in old fleet and delays from purchased transport delay delivery of every commodity in the trailer. Delayed deliveries and a casual culture lead to insufficient time for pick-ups. Poor visibility of the system made these oversights hard to control. This visibility caused more costly administrative work and reduced manager agency across the terminals, contributing to poorly run docks and damaged freight. And all these dysfunctions scaled with volume, and since the sales team accepted too much low-quality freight, volume had the perverse function of scaling service failures while diluting density. And diluted density leaves less cash funded by operations to invest in improvements.

Today, in 2026, TForce represents a sizable part of TFI’s total operations, and it remains a turnaround play. Five years post merger, the task of simultaneously rationalizing terminals, upgrading fleet, reducing volumes, improving freight quality, and changing the culture has made TFI’s acquisition of UPS freight rather painful.

Perhaps part of the reason for the slow turnaround is that culture is slow to change. Perhaps TFI, which dominates Canadian LTL, was unprepared for the competitive nature of the US LTL market:

The culture is the culture of, you know, the old days of laissez-faire and, I don’t really care. I mean, we’re changing that culture. You’re gonna say, “Alain, you bought the company five years ago.” I mean, five years ago, think about that. We’re still, okay, working on changing this culture of we’re not a monopoly anymore, okay? We are a LTL company in North America, and we compete with good peers. I mean, we’re competing with good companies in North America, so we have to be good

To end on a positive note, at the time of writing (July 2026), conditions across the US LTL industry are bullish: the indices suggest consecutive months of heightened industrial activity, tightened truckload capacity, and high diesel prices. The market knows this: LTL carriers, including TFI, are all carrying historically high valuations. Yet, numerically at least, improvements are much easier when starting from low baselines. And, as all these quotes show, the management of TFI is obviously aware of its issues: it has spent five years trying to fix them. With a newer fleet and most of the old real estate, will 2026 be the year TForce turns around?

Appendix

• Some TForce freight managers apparently dislike Bedard.

• The SEC fined UPS Freight $45 million dollars for overvaluing its UPS Freight division. I don’t recall seeing TFI mention this.

• My other LTL articles:

This is Part 2 of my ongoing series on LTL freight. If you have no idea what LTL freight is, I recommend reading Part 1 first.

LTL Freight: Density is the Name of the Game

This article examines the main drivers of revenues and margins in the less-than-truckload (LTL) business. To frame the discussion, I’ll start by mentioning the most important metric in all the transport industry: the operating ratio. The operating ratio, or OR, the percentage of revenue consumed by operating expenses. In other words, it’s the complement of the operating margin:

The lower a carrier’s OR is, the more efficiently revenue flows to the bottom line. The concept is basic to all businesses―superior firms spend a little money to make a lot of money―but wide disparities in ORs exist across LTL carriers. This article examines the broad factors that go into the numerator and denominator of this all-important ratio.

• From the revenue side, I’ll look at factors that contribute to demand, pricing, and carrier differentiation.

• From the cost side, I’ll look at the major classes of capital and variable expenses.

In the next article, I’m going to directly compare firms and try to discover why the best carrier has an OR range in the 70s, the good ones have ORs in the 80s, and the laggards never break the mid-90s.

1. The LTL market landscape

The good times in the LTL industry happen when industrial activity is high and truckload capacity is tight. LTL is a consolidated industry and has only gotten more so since a major carrier, Yellow, went bankrupt in 2023.

LTL Customer base

LTL is almost a purely B2B industry. As the literal mover of things, LTL’s reach extends to categories of industrial activity, commerce, and manufacturing.

At the broadest perimeters of its TAM, the LTL customer base includes anyone that needs to transport a shipment that is large and likely palletized, but not so large that it fills a truck, nor so fragile or time-sensitive that paying for a dedicated service is preferable.

Within this set of requirements are a broad range of industrial activities. Getting more specific, some big categories of LTL customers include:

• Manufacturers and OEMs. Inventory turns over and equipment gets moved and replaced. These routine industrial actions require transport. LTL services are especially harmonious with “lean” style manufacturing, which favors constant movement of inventories over inventory buffers. The fabled American manufacturing renaissance, if it ever arrives, would be a bonanza for LTL.

• Construction. Commercial and residential building operations require a lot of material, equipment, and equipment parts, which are transported in and out of the site over the project lifecycle.

• Medical supplies. Hospitals and pharmaceuticals often ship goods with special handling and time-critical instructions. These incur assessorial charges, another important source of revenue (more on that in the pricing section).

• Grocery. Grocery shipments are high volume and require special handling and temperature controls, so capturing a large contract to ship regularly to a grocery distribution center or store can provide a nice stream of revenue for an LTL carrier.

• Retailers. Big-box stores need to move heavy inventory and equipment in and out of warehouses.

• E-commerce. Historically not an LTL market, e-commerce has become an important revenue stream since 2020, the time when people became more comfortable buying large, heavy goods online.

For some specific customers and segments, I can provide examples from investor decks:

• XPO’s 2026 deck has logos from giants like CAT, 3M, Ford, and Deere.

• Saia’s deck crosses industrials, technology, oil services, and big-box stores, with logos that include Dell, Halliburton, Home Depot, Starbucks, and Honeywell.

• FDXF’s investor day does not mention specific customers, but they have a slide discussing the markets that they want to capture: small-medium businesses, healthcare, grocery, and data centers.

  
  It’s interesting to note that FedEx Freight’s January 1, 2026 prospectus did not mention data centers. Whether FedEx Freight identified a new opportunity in 5 months, forgot to mention it in the original prospectus, or added it to sprinkle their investor day with a bit of the magic of the moment is an open question. But it’s not hard to imagine that LTL carriers will capture some revenue from a capex boom that requires filling massive buildings with heavy, shock-sensitive, and extremely expensive freight.

For some concrete examples of what freight moves through LTL, YouTube is a mine of primary sources. Search for something like “Day in the Life LTL” and you’ll find countless vlogs from P&D and linehaul drivers, and videos of dockworkers playing Tetris with pallets and packaging peanuts. For examples of shipments, this ODFL P&D driver’s deliveries included a mattress, cabinets, generators, and TV stands. I watched another video of a P&D driver who dropped chemicals for a lumber yard, heavy tires for a construction site, and a boom hoist.

Mutable market demand

We believe that 2023 was a terrible, very tough year for us, okay, for the industry in general. 2022 was the party, 2023 was the afterparty, okay?

— Alain Bedárd, TFI CEO | Q2, 2023

With its fortunes so embedded in the fortunes of many other industries, the conditions that create LTL demand are volatile. Indeed, in looking over earnings calls and histories, I’m surprised how sentiment whips up and down. It seems every year involves at least one manager claiming things are either “the worst we’ve seen” or “the best we’ve seen”.

To gauge this demand volatility, I did a crude sentiment synthesis and analysis over ten years of calls. This visualization shows the results; it’s not to be taken too seriously, but I think it captures some broad subjective feeling. Some context on the conditions is provided for each year.

While the shifts aren’t quite as sudden I first perceived, there is substantial oscillation. Of course, this timeline just summarizes and ranks subjective statements. For quantitative data on the macro drivers LTL fortunes, the end of this section includes links to important indices.

Despite the volatility of demand, however, the industry has broadly trended towards better operational efficiency and sustained operating profits. The bad years still provide cash flow from operations, and the good years provide a chance to realize great operating leverage. That leads me to the competitive landscape.

The competitive field

The LTL industry is broadly consolidated. For comparison, in the fragmented truckload industry, the top five carriers account for 8% of revenue; in the consolidated LTL space, the top five carriers capture 51% of revenue.

This consolidation has multiple causes. The most significant and natural is that LTL is far more complex and capital intensive than truckload: it’s hard to enter and even harder to do well. Secondly, the top firms are all reasonably well-run (or perhaps, compared to management of all businesses, excellently run).

Because the barriers to entry are so high, the carriers also enjoy a quasi-oligopoly. Through the last decade, all major public carriers have maintained price discipline on their base rates. After looking at ten years of history, I haven’t seen any significant entrants or “crazy” actors looking to undercut prices. 2009 seems to mark the last major period of price competition, which occurred when FedEx Freight tried to run Yellow out of business and caused rates to spiral across the industry.

With this price discipline comes operational robustness. Despite a whippy and unpredictable freight market, all major carriers have maintained annual ORs below 100% for at least the last ten years. In other words, no carrier―even the defunct Yellow―had a year of negative operating income.

Looking at cash provided by operations tells a similar story. While the laggards, ArcB and Yellow, had negative operating cash flows in some first quarters, even their annual OCF was always positive.

Finally, while the years between 2023 and 2025 were markedly soft in the post-COVID freight crash, all carriers have been buoyed by the Yellow bankruptcy (whose collapse, as the preceding charts show, is more about the importance of the I in EBIT than it is about dismal industry economics). The timing of Yellow’s exit was exceptionally fortunate for its competitors, occurring right when LTL demand turned from scorching to freezing. So FedEx Freight never needed to undercut prices in 2009―it could have just waited 14 years and let Yellow bankrupt itself.

A note on the Amazon in the room

After all my words about price discipline and insinuations about oligopolies, I must mention the Amazon overhang. For years, Amazon has dipped its toes near the LTL space, and analysts have long feared when it would jump in. In June of 2026, that day arrived.

When I started looking into LTL, I thought this moment would be cataclysmic for the existing firms; now, having researched how these firms operate, I’m cautiously skeptical of Amazon’s threat. At the moment, it’s unclear how far Amazon LTL can reach, and it lacks the terminal infrastructure of the established players. Nevertheless, Amazon is obviously never a player to underestimate, and it surely has the intellectual capital to drive new levels of data-driven optimization. Yet, LTL is a specialist’s game.

This story is still in its first chapter…

Some takers and givers of revenue

LTL exists in the broader transport industry, and so its revenue passes to and from players in the broader field. Two important places where revenue overlap are truckload and third-party logistics (3PL).

Truckload

LTL transports only 10-15% of total road freight. When truckload capacity is tight, shippers become more likely to use LTL. This effect is discussed constantly across 15 years of earnings calls for each carrier. Some samples across the years:

• 2014.

  • Saia. “Tightest environment they’ve seen in careers; customers shifting heavier shipments to SAIA due to TL scarcity” (Q2 2014)

  • ARCB “Greatest sequential growth in 15 years” driven by “service disruptions in intermodal and driver shortages limiting TL capacity” (Q3, 2014)

• 2016.

  • FDX/FDXF. “Some heavier weight shipments shifted back to truckload as TL capacity eased” (Q2, 2016)

• 2017

  • XPO. “Customers moving from 3-4 pallets to 4-5 pallets per shipment” as TL tightened (Q3 2017)

• 2024

  • TFII. “Q: I wanted to also key in on, you called out, truckload continuing to steal some LTL freight…when would you expect that to improve? A: The minute the truckload guys get busier” (Q3, 2024)

• 2026

  • ODFL “Truckload tightness driving LTL volume recovery; retail leading, industrial to follow” (Q1 2026)

3PLs

Third-party logistics (3PLs) brokers handle LTL shipping on behalf of clients. Since they have access to rates from multiple carriers and understand the industry well, 3PL can help infrequent shippers manage the complexity of LTL pricing at presumably better costs. On a per-customer basis, shippers who use 3PLs are unlikely to be large sources of revenue, since larger customers are more likely to negotiate long-term rates directly. However, since 3PLs capture volume from many customers, they can negotiate rates on volume and capture a bit of the difference.

For an example of how much 3PL contributes to overall revenue volume, ODFL reported in a call in 2026 that about a third of its overall revenue comes from 3PLs. This seems consistent across time and carriers: Saia reported a similar range back in 2013.

For LTL carriers, the volume from 3PLs can contribute to ROIC or dilute it. If carriers and brokers have a constructive relationship, both can benefit from the service that a broker provides. 3PLs can pre-consolidate volume and educate less sophisticated shippers on the details of how to pack and document their commodities (large sources of pain across all points of the industry). As long as carriers can maintain price discipline and achieve leverage through density, the extra volume from 3PLs provides marginal contribution dollars. ODFL, as usual, provides the exemplar.

Yeah. About a third of our business overall is related to 3PLs. You know, as mentioned earlier, you know, we’re pretty consistent with what we target for increases every year, you know, be it with our general rate increase that applies to our tariff-based business, and that’s about 25% of our revenue overall, as well as, you know, what we try to achieve as we go through contract renewals. Obviously every account is different and we look at each account on its own merits and what their profitability measurements are. We’ve been pretty consistent with getting increases and, you know, it’s a different approach that I think that we take versus some of our competitors.

—Adam Satterfield (EVP and CFO, Old Dominion Freight Line), Q1, 2026

On the other hand, when carriers accept aggressive pricing and low-quality freight, volume from 3PLs dilutes margins. TFI’s Bédard shows typical candor in how providing a “blanket rate” resulted in 3PLs using TFI’s LTL service only when it was the absolute cheapest option.

Yeah. You know what the commercial team has done, as an example, you know, with our 3PL, you know, what we gave those guys, let’s say a year ago, was mostly blanket rates, which is the worst that you could do, right? Because then you give the guy a blanket rate that mean they will use you when you’re the cheapest and lowest guy in the world, right? We said, “This doesn’t make any sense.”

— Alain Bédard (Chairman, President, and CEO, TFI International), (Q1, 2026)

Finally, like truckload, 3PL pricing and activity are industry barometers. Since 3PLs deal with many customers and carriers, they occupy a position of information advantage. Carriers report that an early sign of softening demand occurs when 3PLs start asking for lower prices.

Inorganic growth and M&A

Carriers can expand their reach through acquisitions. FedEx Freight was formed entirely of acquisitions of carriers with networks in strategically dense areas. XPO entered the LTL market when it acquired Interride in 2013, and became a major player when it acquired Conway in 2015 (and has since spun off its other businesses to become a pure play LTL firm). Saia recently has broadly expanded its network not through mergers per se but through buying up old Yellow terminals (as have others to lesser extents). ODFL remains aloof from the behavior of its peers and has done little M&A.

When done well, acquisitions can be accretive and strategic. XPO provides the best example among its peers (no surprise, since it was founded by Brad Jacobs, who made a legendary career from industrial rollups and roll outs).

When done poorly, however, M&A dilutes network density. A current example is TFI’s acquisition of UPS’s LTL division in 2021. While the acquired assets have some quality service centers, the sparse network and undisciplined sales have chewed away at the operating ratio.

“I said it many times, we got 12,000 doors over there. We don’t need 12. We need maybe, I don’t know, 6,000, 7,000 or 8,000. Okay, we have way too many doors.”

— Alain Bédard, CEO, TFII | Q2, 2021

Macro indicators

Finally, I’m going to list a few key indexes that inform the LTL industry.

For measures of overall industrial activity:

• The ISM PSI reports survey data on industrial activity. A number above 50 indicates an expanding manufacturing economy. All carriers monitor this index.

• Inventory to sales ratio. When inventories to sales are low, companies likely need to restock, which requires material movement.

• Durable goods. High orders of durable goods indicates businesses are moving heavy things from point A to point B.

Trucking indicators:

• Cass truckload linehaul rates. Rising rates for TL can indicate demand flowing soon to LTL.

• Tender rejection. High rejections of truckload tenders also indicates tight TL supply.

2. How LTL carriers charge

LTL pricing has many variables. Determining a shipment price can be complicated. But before I get into the main determinants of price, I think it’s helpful to look at the metrics that the carriers themselves consider most important.

Most public LTL firms file a helpful monthly 8-K reporting some key performance measures. Generally, these are some variation of the following:

• Revenue per hundredweight. Also known as yield, this metric is a key indicator of the LTL industry. However, it doesn’t tell the whole story. Because light, bulky freight is typically more expensive to ship, when average weight per shipment increases, revenue per hundredweight likely decreases.

• Average weight per shipment. Heavier average weight indicates increased industrial demand.

• Average length of haul. Longer hauls cost more money.

• Revenue per shipment. How much the carrier makes on average per shipment. This metric combines the prior three.

In the next post, I’ll compare these metrics across carriers. For this post, the key element to consider is that high revenues comes from some mix of high volume and high prices, with shipments that are picked up and delivered with density. The previous section was about broad drivers of volume. So what goes into the price?

General base rate

The general base rate is what every carrier charges as a “sticker price” depending on a composite of pricing factors. Negotiated freight usually ships with a lower price than this base (besides the added assessorial fees).

A general rate increase (GRI) happens when carriers increase their base rate. Dreaded by shippers and celebrated by carriers, the GRI is typically an annual event, announced by all carriers in October. ArcBest, however, bucked the trend in 2026, announcing a GRI in May (I presume the first of two for the year).

Freight class

LTL freight is classified on a scale of 50 to 500. Higher-classed freight incurs higher charges. These classifications are determined by the LTL trade association, NMFTA. The variables are density, handling, stowability, and liability.

As with routes for carriers, for shippers, density is also the name of the game. While weight is a price determinant, 2,000 pounds of bricks takes up less space than 2,000 pounds of ping-pong balls. Irregular shapes are harder to handle and pack. Fragile freight requires more service.

For carriers, shipping mixed freight classes, like so many other things in LTL, is an optimization problem. The firms that do it best are the ones that can pick up and deliver the most profitable mix in the most predictable revenue streams along the densest routes.

Origin/Destination

Longer hauls and routes to and from out-of-the-way zip codes cost more.
For carriers, route density, penetration, and customer selection make the difference here. Firms that can ship the most stuff to the most places at the fastest pace capture the most revenue.

Assessorials

Special-handling requirements increase the delivery variability. To make up for this, carriers charge assessorials. Assessorials include things like delivery to places that don’t have a loading dock, white-glove service, hazmat, and so on.

Assessorials can make a surprisingly large contribution to the top line. In the 2024 Q2 earnings call, XPO’s CEO mentioned that assessorials made up about 10% of revenue, which already sounds like a lot, but it’s 500 basis points worse than the best-in-class peer (almost surely ODFL):

Mario Harik (CEO, XPO) So when you look at it, the difference on assessorial revenue as a percent of total from where we are today and where we want to be, where best-in-class is, there’s roughly a 5-point difference. So we have grown this last quarter assessorial revenue double-digit on a year-on-year basis, and we are making inroads towards getting to that 15%.

2024 Q2 earnings

Fuel surcharges

LTL prices have a surcharge based on the DOE’s weekly index of diesel prices.

Unlike truckload carriers, fuel surcharges are not pure pass-throughs of cost. In other words, revenue from high surcharges contributes to the bottom line. Saia’s CFO mentions this explicitly in the 2024 Q3 call.

Matthew Batteh (EVP and CFO, Saia) Fuel is not a pass-through. I mean, we invest. Every part of this business is inflationary, and we expect to make a return on every part of it. The fuel surcharge mechanism is similar to what’s out there for other LTLs.

—2024 Q3

However, while fuel surcharges provide a way for carriers to “effortlessly” capture more revenue, I don’t consider it a quality signal like the other factors in this list. When reporting metrics, most carriers strip out the surcharge costs to provide a more direct view of operational performance. Furthermore, sustained high diesel prices likely indicates demand-destroying pressure everywhere in the industrial economy.

John Smith, CEO FedEx Freight. Our fuel surcharge is never truly a pure pass-through, but it’s typically a net positive for FedEx Freight above our cost of fuel. The fuel surcharge can be a year-over-year headwind or a tailwind in any given year…

— 2026 FDXF investor day

3. Differentiation among carriers

For all its complexity, the LTL business is fundamentally a transport business. While moving palletized objects from point A to point B seems like a business destined for commoditization―indeed, the industry jargon for shipments is “commodities”―the firms in the sector can nonetheless differentiate through service quality and specialization.

Service differentiation

Superior service is the best way that a carrier can achieve pricing power. Service quality ultimately is what customers pay a premium for, and it is the key to ODFL’s dominant margins.

Service quality requires intention across the carrier’s network. Freight is repeatedly broken down and re-consolidated in the course of an operation, creating deep inter-dependencies along the points of the network. Freight is transported from node to node, and disruptions ripple across the system. Brad Jacobs, of XPO, colorfully describes these butterfly effects:

[When capacity is tight, service] suffers a little bit more because it’s a network, and you tickle it here and it laughs there, you pinch it here and it cries there.

Freight can only move if drivers, dock workers, and dispatchers coordinate constructively to load and unload freight without damage and without dead time. Dysfunctional coordination causes late pickups, late deliveries, and linehauls that leave either late or under capacity.

Quantitative service measures

In the world of logistics, where failures can cascade, service quality is mainly a function of moving commodities on time and undamaged. These aspects of service quality are measurable. Metrics include:

• On-time performance. This measure can be broken into sub-measures for on-time pick up and delivery.

• Claims ratio. When freight arrives damaged, shippers and consignees not only have to replace inventory but also submit claims in a triangular communication between shipper, carrier, and consignee. Carriers with the lowest claims ratios minimize these frustrating and costly incidents.

• Re-rates. Since LTL freight pricing can be complex, it’s not uncommon for shippers to be surprised by added final charges. For obvious reasons, frequent rerates are a source of customer aggravation and cost uncertainty.

Customer relationships

LTL is a business where relationships matter. High-quality customers and 3PLs appreciate long-term sales relationships and a carrier’s understanding of industry particularities. Good relationships are also how carriers and shippers can lock in a level of price stability through long-term deals.

Customer relationships are also built with every pick up and delivery. P&D drivers typically have a set “run,” so they’ll interact with frequent customers over years. Shippers and brokers who interact with terminal workers pick up on the general vibe of professionalism and efficiency. Compounded across thousands of drivers and hundreds of terminals, these micro-interactions can make a big difference on customer retention.

Differences in customer quality are also significant. A virtuous cycle exists where well-educated customers receive superior service. When shippers understand freight pricing, package freight properly, and fill out bills of lading correctly, they experience fewer re-rates, damages, and miscommunications about P&D. For this reason, carriers can also achieve superior service through customer education and pricing transparency (this “education” effect also can come from 3PLs who work with smaller companies that don’t have sophisticated logistics departments).

Price mixing

Carriers that consolidate freight with different service and handling requirements can theoretically improve density through throughput. Regional transport from small-to-medium businesses often has the best margins but insufficient volume. National accounts with negotiated contracts have lower yields, but their volume fills out the density between what may be higher-revenue drops.

FedEx Freight is a carrier with an interesting strategy here, breaking shipments into priority and non-priority tiers. In their 2026 investor day, they claim that margins on each shipment are roughly the same, illustrating that old inescapable tradeoff between throughput and latency.

Data granularity and transparency

Moving freight is inherently complex and uncertain. Carriers can allay shipper concerns by providing a high degree of data access and integration.

Carriers can differentiate their digital experience on multiple levels. An obvious place to start is the customer interface on the company site (in fact, simplifying the UI by breaking it out of the main Fedex services was one of the reasons Fedex Freight cited for its spin). On one level down, carriers can also help their shippers and brokers automate their transport flows through APIs for rates and shipments tracking.

To simplify and standardize pallet pricing, all major terminals have pallet dimensioners (or dimensionalizers) that scan and weigh pallets. Some dimensioners take photos to capture the commodity’s physical state. All these measurements reduce re-rates and disputes, providing a positive degree of certainty that benefits shippers and carriers.

Once the freight starts moving, tools like RFID tracking provide transparency and reduce manual communication. Over time, customers can use their history of shipments, routes, and prices to optimize loads.

Market penetration and specialization

Certain markets require special handling, special trailers, and special understanding of a customer’s business environment. Grocery is one such market, whose commodities involve temperature controls, regulatory compliance, shelf fees, and shelf fines.

[I]n the retail industry, including grocery, there’s a penalty if this freight is not on the shelf on time and in full. They’re called fines. Many of our competitors, they can go out and talk about meeting those expectations with fancy marketing material and so forth. Until they can stop those fines in our customers’ pocketbooks, nothing’s going to change. We figured out how to do that many, many years ago, especially in the grocery industry. So we do not see anybody getting close to what we can offer from a service standpoint in the retail industry.

—Marty Freeman, CEO ODFL | Q2, 2025

Finally, network geography provides a hard boundary on a carrier’s TAM. Through density and customer relationships, regional carriers can perform at certain scales, but long-distance linehaul requires a national network. Some shippers have RFPs that require coverage of all 48 mainland states.

4. LTL costs and bottom line

In the previous sections, I outlined the macro and firm-level drivers that create revenue for LTL carriers. However, revenue alone is not what makes a superior carrier. If that were the case, Fedex Freight, which has the highest revenues, the most volume, and the largest number of terminals, would achieve the best OR through scale alone. In reality, ODFL consistently has an operating ratio in the 70s, and FedEx Freight, with ~60% more revenue, sits in the mid 80s.

In this section, I’m going to break the capital and operating expenditures that stand between a carrier’s revenue and its operating margin. The main lessons here are that capital expenditures are investments in capacity, but a firm can only realize operating leverage if its unit variable costs decrease with each new shipment.

Capital expenditures

LTL is a capital-intensive business. Firms typically spend 10-15% of their revenue a year on capex. Here are the major categories of how they do so.

Terminals

“The door capacity is very critical, and certainly could be a limiting factor to growth, and it’s why we stay or try to stay so far ahead of the demand curve there.”

— Adam Satterfield, CFO ODFL| Q3, 2020

Terminals are the primary capital expenditure through which carriers can expand capacity, increase density, and absorb demand in good years. At a more granular level, each terminal’s capacity depends on the number of doors it has. Carriers also expand existing terminals in strategic areas to add doors.

While terminals are the most important way to build capacity, too many terminals can dilute density. Fedex Freight, for example, rationalized and consolidated some terminals in the past years.

“Since 2023, we have consolidated 39 service centers, removing 1,000 doors, while at the same time adding nine new locations and 600 doors in strategic locations to expand our capabilities in the densest freight markets. The result is a network that is positioned where the freight is, with unrivaled scale and proximity in a business where scale and proximity matter.”

—Clint McCoy, COO FDXF | investor-day, 2026

In TFI’s 2022 investor day, its management described how terminal clustering was causing linehaul efficiencies.

“As we close facilities, it’s not really a P&D advantage. It’s a huge line haul advantage, right? So if you have two facilities nearby, and you’re running 1,000, 1,100 miles like we do, and you’re running two different line hauls that are half full, or if you’re running four and two of them are not full, it’s extremely expensive, and that’s our biggest cost.”

— Paul Hoelting, President TForce Freight | investor-day-2022

Carriers generally prefer to own terminals rather than lease them, so investment in door capacity is also an investment in real estate.

Dock infrastructure

Within the terminal, carriers also must invest in the infrastructure to keep freight moving between trailers. Dispatch equipment, forklifts, dimensioners, and so on.

Trucks

ODFL and Saia aim for an average fleet age of 3.5 years. To realize fuel efficiencies, newer trucks do longhaul and older trucks do city routes.

Unlike terminals, carriers avoid spending capex on trucks to anticipate future capacity. The depreciation cost of trucks is fairly high, as is the opportunity cost of fuel efficiency gains coming from new generations. Carriers try to keep their fleet ages young and size their fleet relatively close to current demand.

Trailers

Trailers have a longer depreciation life and can be used in terminals for staging. As such, trailers are sometimes considered “growth” investments in future capacity.

XPO has the most unusual strategy here, with an in-house trailer manufacturing plant in Arkansas.

Technology and investment

Investing in software and data processing is a significant part of route optimization. The precision and efficiency gains from technology are a significant factor in how the industry’s OR has steadily trended downward over the last ten years.

When the digital strategy is done well, carriers can also use their existing data infrastructure to improve the customer’s digital experience. So this category is a bit of a catch-all:

• On the customer-facing side, it includes data platforms and interfaces, price APIs, freight tracking, and billing.

• On the internal carrier side, IT investment includes optimization of routes, dock management, meteorological forecasting, and performance tracking.

While hard to estimate externally, IT is surely a significant expense for all carriers. In its investor day, FedEx Freight indicated that it allocates 25% of capex to technology. ODFL is more coy about their investments, but the video of their central command tower shows obvious investment and technical sophistication.

Operating expenses

The key relationship for an LTL carrier is variable cost as a percent of revenue.

—Robert L. Sullivan, The LTL Carrier’s Profitability Blueprint

If capex is how a carrier builds density capability, opex is how a carrier realizes it. The carrier’s 10-Ks all break out operating expenses into some set of these categories.

Salaries

Wages and salaries are by far the biggest operating expense for all carriers. This includes dock workers, dispatchers, and, of course, P&D and linehaul drivers. Historically, union carriers―such as ArcBest, TFI, and Yellow―have been disadvantaged, as labor agreements reduce flexibility around hours and assignments.

Linehaul truckers for LTL firms have essentially the same job as all TL truckers. Some firms use spot trucking rates to minimize costs when truck labor supply outpaces freight demand. So the labor market is another place where money passes between the truckload and LTL industries.

All the big carriers have driver-training programs to bring up dock workers. This provides a strategic way to grow capacity without looking to the external truckload labor market.

Purchased transport

Some linehaul gets outsourced to rail and third-party trucking. This is known as purchased transport (PT).

Carriers spending on PT differs considerably. In the 2025 annual reports, carriers reported the following ratios of purchased transport to revenue.

• 20%, XPO

• 20%, Saia

• 10.7%, ArcBest (asset-based segment)

• 9.2%, Fedex Freight

• 2% Old Dominion

The economics here are complicated, but generally, carriers try to reduce PT. XPO has an explicit goal of “insourcing” longhaul miles. When truckload capacity is tight, firms that rely less on purchased transport share less revenue. That said, reliance on purchased transport is also part of how XPO protected margins in the 2023-2025 freight recession (if times were bad in LTL, they were dire in truckload).

PT also leaves the carrier with little control of the linehaul service levels. TFI mentioned that removing reliance on rail is part of its strategy to improve service levels. FedEx Freight provides the counterexample, since it considers its use of rail a specific strategic advantage.

Fuel

Diesel fuels the trucks and dock vehicles.

Since fuel surcharges are not pass throughs, companies that can reduce fuel through superior equipment, route efficiency, and cheap rail transport can capture more contribution dollars.

Insurance

Not much to say here. Good safety records reduce insurance costs. On the “carrot” side of incentives, carriers pay their drivers annual safety bonuses. On the “stick” side, carriers have started to adopt inward-facing cameras to monitor driver focus and head movement. Forward-facing camaras have also been standard, and sometimes are a requirment lower insurance rates.

Some firms “self-insure” a portion of their risk, a practice that likely reduces overall costs at the obvious tradeoff of volatility.

5. Who wins the LTL game?

So if you talk to our sales guy, they say, “Well, you know, it’s tough for us to get more business because the service is maybe not up to par to some of our peers,” right? So when we talk to our operation guys, “Guys, you know, we got to fix the service. We have to improve the service.” Now, you say, “Yeah, well, if we improve the service, maybe there’s a cost to that.” No, no, no, no. You have to improve the service and reduce the cost at the same time. So this is quite a challenge, and this is where the talent of your management team comes to play.

— Alain Bedard, TFI CEO | Q3, 2024

The carrier that wins the LTL game is the one that has the best density, the greatest pricing power, and the best operating performance. However, these three key determinants are inseparable.

Price power ultimately goes to the carrier with best service. But superior service is only economical if the carrier can achieve density. Density means that pick-ups and deliveries are constantly happening, and freight is constantly moving along the network. Density also means saying no to shipments that go too far out of the network (or charging a lot to offset the density loss).

Density is not cheap. It requires enormous capital expenditure―this is why the industry is so consolidated. And achieving the best margins through density requires relentless execution and a long history of customer satisfaction―this is why Old Dominion is so far ahead.

In the next part of this series, I’m going to take a data-driven look at the history and financial performance of Old Dominion and its peers. And while ODFL has the best operation, it is far from the only well-run company in this niche industry. At the time of writing, it is also the most richly priced.

LTL is a complex, capital-intensive niche with high barriers to entry and significant potential for operating leverage. It is an industry that is deeply plugged in to a nation’s industrial activity on the whole, one whose fortunes rise and fall with the pace of inventory turnover, construction, and business demand. While its business model, moving heavy-ish things on a truck, seems prone to commodification, LTL is also an industry of surprising differentiation, one where the top player realizes far better margins and pricing power than lower-tier operators who do more and make less.

In this first post, I outline the basics of how the industry works. In later posts, I’ll go over the economics of the industry, analyze the competitive landscape, and finally look at one or more key firms and situations in detail.

Subscribe now

When freight doesn’t fill a truck

The name less-than-truckload is appropriate. LTL’s business is all about moving freight that is on a pallet but doesn’t fill a whole truck. Of the moving-things-on-a-truck modality, LTL sits in the middle, moving shipments that are larger than parcels but not so large that they fill a whole truck. Semantically, the differences between truckload and less than truckload sound matters of degree. But having multiple freight shipments in one truck profoundly alters how these two forms of truck transport operate.

To understand LTL, it’s instructive to compare it to its simpler cousin, TL. The first, category-defining difference is that a truckload carrier’s truck has one shipment and an LTL truck has many shipments.

The fact that the LTL truck delivers freight from many origins to many destinations adds a substantial amount of operational complexity. To simplify only a little, a truckload carrier’s existential purpose is to follow this sequence:

1. Load a truck.

2. Haul it from point A to point B.

3. Unload the truck.

LTL is not nearly so easy. Since the cargo is picked up from different places and delivered to different places, routing requires much coordination. LTL carriers realize efficiencies by consolidating freight that’s moving in roughly the same direction. To achieve this, LTL carriers must have a network of terminals to route, break down, and consolidate freight.

The end-to-end process goes something like this.

1. In the afternoon, a P&D driver picks up freight from different areas close together.

2. The P&D driver hauls the pick-ups to a regional service center.

3. Dock workers in the service center break down and consolidate the pick-ups that are heading in the same general direction.

4. Another driver hauls these shipments to another destination, likely a breakbulk terminal where they are further consolidated with other shipments from the same hub.

5. The dock workers load a truck for a long distance haul. This haul may involve other breakbulk terminals, or even outsourced rail shipping.

6. Freight from the last breakbulk terminal is dropped off at the different service centers that it serves.

7. In the service center, dock workers consolidate freight by delivery route and load the P&D truck.

8. After deliveries in the morning, the driver makes pick ups along the route in the afternoon. (Go back to step 1)

From the shipper’s perspective, LTL freight travels from Point A to Point B. But from the carrier’s perspective, it’s more like going from Point A.a.1 to Point A.a to Point A to point B to Point C to Point C.a to Point C.a.1.

In this hub-and-spoke distribution pattern, the potential for variation is endless. As one common example, the linehaul portion of the route is often outsourced to a truckload or rail carrier. This outsourcing, known as purchased transport, is a common operating expense for LTL carriers. Such variation adds much complexity, a topic worth its own section.

Variation and complexity

LTL operations are complex. The development of terminal networks is only one example of the added variability that comes from having multiple shipments to deliver in one truck.

TL shipping again provides a useful point of comparison. In a truckload delivery, the driver has one bill of lading that goes from one point to another. The route is direct, and the driver likely never touches the freight. LTL’s modality of re-routing and consolidating items adds substantial variability to the task. Some significant ways the task changes when freight is carried over hub-and-spoke include:

• The freight touches many hands. It moves from truck to service center to truck to hub to truck, and so on. Each touch increases the potential for damage and loss.
  
  Claims on damaged cargo therefore are a large concern in LTL. As a result, the claims ratio is an important metric of customer service in LTL. It even has its own index.

• Every drop is different. A truckload driver’s job ends when the freight is moved off the truck in one go. A P&D driver drops off many goods over the course of the day. And each drop is different: some have loading docks, some don’t, some may have special care instructions, and the weight, dimension, and number of pallets varies with each stop. All this variation adds a degree of complexity and unpredictability.

• The cargo isn’t uniform. An LTL carrier may have different classes of goods with different shapes and handling instructions in the same delivery. Loading this to optimize for maximum capacity requires load planning.

Speaking of load planning, LTL is a domain full of optimization problems. The different topologies, both of networks and of geography, bring many possible combinations of routes. And the spatial capacity of the trucks themselves means that there are physically and economically optimal ways to load a truck, which must account for pricing differences in freight classifications.

Furthermore, freight networks are subject to the same tradeoffs as computer networks: throughput and latency. If a carrier had infinite time and an infinite capacity of trucks to ship an infinite supply of commodities held in infinite storage, the optimal strategy would involve waiting as long as needed to achieve maximum route density. The finite world, unfortunately, comes with the annoying constraints of space, time, and customer service. So carriers must make decisions about how to optimize route density while meeting customer expectations. And that’s the name of the game.

Density is the name of the game

If you read the 10-Ks and earnings calls for any LTL carrier, you’re sure to find one word that appears with abnormal frequency: density.

Our primary revenue focus is to increase density, which is shipment and tonnage growth within our existing infrastructure.
- Old Dominion Freight Line, 10-K, 2025-12-31

Density is a concept that has two important implications on the industry. For one, shipments are valued based on their density. But I’ll speak about pricing in another post. For this overview, I want to talk about route density.

Route density describes how tightly stops cluster in one delivery. Achieving route density is the key concern for an LTL carrier. LTL carriers recognize revenue when they drop freight off. All time that a P&D driver spends outside of picking up or delivering freight is economically unproductive.

When efficient LTL carriers minimize variable costs, they do it through density. When carriers make their largest capex investments in new terminals, the goal is density. In other words, density is how carriers express operating leverage, the most important tool they have to capture massive profits when demand is high.

When carriers have poor density, the effect of operating leverage cuts the opposite way. For example, TFI is a Canadian multi-modal carrier with a struggling American LTL segment. While TFI’s LTL operations are third-tier, its CEO, Alain Bedard, ranks first in candor. Here he is discussing why the sales team needs to consider density when accepting orders.

At the same time also, we’re saying to our commercial team, “Guys, let’s focus on freight that fits us.” I mean, don’t give me a customer where I have to run 70 mi to pick up the shipment, because this is not what I want. I want something that is closer to my terminal to improve my density, okay? I want more shipment per stop, okay? To improve my cost per shipment, et cetera, et cetera.

— Alain Bedard, TFII Q1 2026 earnings call

But I’m getting ahead of myself―in the next post of this series I’ll describe the economics of this complex, capital-intensive, and high margin industry.

One Foot Hedge

What drives the LTL top and bottom line?

This is Part 2 of my ongoing series on LTL freight. If you have no idea what LTL freight is, I recommend reading Part 1 first…

Read more

2 months ago · 3 likes · OneFootHedge

Subscribe now

Supplementary

• Freightonomics: What is LTL. A podcast where an industry expert explains LTL to a beginner.

• How Does Less-Than-Truckload-Shipping Work. Whiteboard video in which a freight broker explains how freight travels in LTL.

• Old Dominion Freight Line, 10-K. Old Dominion is the best class, and their 10-Ks do a good job explaining the key concerns and metrics.

• Old But Good Analysis of Old Dominion.

• The NMFC Book. Standard guidance for freight classification.

• US Trucking Primer 1. A good overview of broader US trucking landscape.

The math is simple, but like all things involving compound interest, its results took me by surprise. From experimenting with different examples, I found that, tax-sensitive investors should be particularly cautious to sell a winner. This is especially in the case of short-term gains.

Depending on the time frame, many determinants can increase the hurdle rate to replace a winning asset with another:

• The short-term tax rate (higher rate, higher hurdle)

• The time needed to convert to a more favorable tax rate (less time, higher hurdle)

• The proportion of the asset value that is made up of short-term taxable profits (more proportion of profits, higher hurdle)

• The duration in which the replacement asset will be held (if held for a year or less, higher hurdle).

Short-term gains have a particularly high rate, so I’ll provide an example of those first.

The high cost of selling almost long-term gains

Say you invest $10,000 in a stock and it pops 30% sometime in the first year. Somehow, you know that the price will not change anymore (perhaps a slow-moving but guaranteed acquisition), so that $3,000 profit is locked in. The short-term t-bill rate is 3.61%. When does it become logical to sell the profits and move the asset into treasuries? The answer depends on the time remaining until the gains are reclassified as long-term profits..

First, calculate the value of the asset after selling it at the short term and long term rates. Using our example of a 30% gain on $10,000, the math is as follows:

PostTaxProfitsShort =  3000 * (1.0 - 0.23) 
PostTaxProfitsLong  =  3000 * (1.0 - 0.15)
AssetValueShort     = 10000 + PostTaxProfitsShort
AssetValueLong      = 10000 + PostTaxProfitsLong

This gives us:

• Short-term value = $12310.00

• Long-term value = $12550.00

In other words, selling the stock into a short-term capital gains tax is only rational if the remaining $12,310 can be reinvested at rate that returns at least 1.95%―the increase needed to turn $12,310 into $12,550 ―by the end of the year.

With a short-term t-bill rate of 3.61%, this 1.95% hurdle is cleared only if the replacement happens relatively close to the time of original purchase. Using months, m, as the time interval, for the investment to make sense, the re-invested proceeds must compound for this many months at a rate that meets or exceeds $12,550:

Solving for m gives a compounding period of 6.55 months.

In this example, after 5.5 months, it’s better to wait the year out and sell at a favored tax rate instead of reinvesting in t-bills. This particular example is further complicated by the fact that short-term t-bill interest itself is subject to the short-term tax rate.

At a real rate of 2.78%, the reinvestment needs to happen within the first 3.5 months of purchase to be profitable.

Taking this example to a further extreme, if the original asset increased by 100%, it is always illogical to pay the short-term gains and reinvest in t-bills. This is because more tax eats into the profits.

In the real world, the original asset almost surely will increase or decrease in value by some amount during the waiting period. So the calculations will change based on growth assumptions. But the hurdle always gets higher as the short-term tax “cliff” approaches and as the percentage of profits increases.

Long-term tax

Long-term tax gains are simpler to reason about. Unless you plan to hold the asset until death, only two questions matter:

• Will the replacement asset perform better than the current asset? (if yes, the replacement must equal the original’s returns)

• Will I sell the replacement asset for short-term gains (if yes, the hurdle is higher)

If you are comparing two assets that have the same taxable gains, replacing the original asset is logical only if the new asset performs better in the same year-or-more time frame. If you think, that’s very obvious, you’re right.

But replacing the stock with an asset that performs equally well will cause a permanent loss of capital. Triggering the tax payment gives you less to reinvest. Unless the reinvestment earns at a better rate, you’ll never recover the loss.

For example, this graph shows the pre- and post-tax asset values of a stock that returns 10% a year and is replaced after year ten with a stock that performs equally well.

Replacing a long-held asset with another that performs just as well essentially rolls back the compounding clock by whatever amount was paid to the state. For example here’s the same graph, but the replacement happens at after twenty years instead of ten.

Real life quirks

These examples all assume that future asset values are stable and predictable, probably the most dangerous assumptions possible. Two obvious conditions change these calculations significantly:

• The original asset’s value might decline. Obviously this lowers the hurdle.

• Stock prices are volatile. Even compounding machines have prices that fluctuate over and above their intrinsic value. In cases of wide spreads in valuations, it might be better to replace one long-term asset with another, even if their annualized returns are equal.

Besides that, the preceding examples take tax-rates as having only one variable: short or long. The actual US tax code is slightly more complex. Some other considerations include:

• Tax brackets change with income. In a year with very low income, the hurdle for replacement assets is lower, and vice versa.

• The lag between sales and tax payment. If an asset is sold in January, 2030 and taxes from proceeds must be paid in April of the next year, the amount owed essentially constitutes an interest-free loan until the taxes are paid.

• Taxes can be paid with debt. Money paid to interest is tax-deductible, so using debt to finance tax payments lowers the tax burden. Obviously this also ratchets up the risk of market downturns.

  
  John Malone provides an case study for using leverage as a tax shield (and all forms of extreme tax sensitivity):

  

There’s another strategy that’s extremely efficient for the tax-sensitive investor. It is is sophisticated enough that I’ll break it into its own section.

The one life hack that revenue men hate

It’s not entirely true that both death and taxes are inevitable. The sequence matters. Yes: if you pay taxes, you will later die. But if you die first, you actually don’t have to pay capital-gains taxes, and neither do your heirs.

Under the current US tax policy, if you die and will your shares, your heirs receive the stocks at a stepped-up cost basis. This is different from gifting shares, which come with the “gift” of taxable gains.

An interesting application of this principle occurs in master-limited partnerships (MLPs), a type of entity usually found in US midstream companies. In these structures, some part of dividends are tax deferred, with the deferred portion going to reduce the asset’s cost basis (until the cost basis goes to zero). For people of a certain age, it’s entirely possible that the MLP could provide a constant stream of tax-advantaged income that never gets taxed at a full rate for the investor’s entire life. The time arbitrage has an afterlife, too, as their heirs get the beautiful gift of the tax-free units.

“Hold forever” is a common mantra of MLP investors, evidence that indeed, some Texans would rather die than pay taxes.

In this post, I’ll:

• Briefly introduce operating leverage as a concept

• Outline my data-selection methods and measures

• Present charts of firms with high operating leverage, grouped by topic

• Finish with an appendix that provides:

  • an Excel file of the 150 firms that have the most operating leverage by different measures

  • PDFs for all the charts of the top 100 firms by each measure.

Looking through these charts, it’s instructive to compare the biggest losers and winners by industry. What makes some firms, even the most capital-intensive firms like shippers, absorb periods of declining revenue so much better than their peers? What gives some businesses such attractive prospects for margin expansion that their charts turn into a wall of green? The answers to these questions are vital to operators and investors.

The perilous rewards of operating leverage

Imagine you are buying widget businesses in imagination land, a place where demands, costs, and prices always remain constant. Which widget business is more attractive?

• BizVar makes widgets for a cost of $10 and sells them for $20.

• BizFixe buys a widget-making patent for $100, and charges $10 dollars per licensed use.

The correct answer depends on sales volume. BizFixe doesn’t break even until 10 units are sold, while BizVar is immediately profitable. And because unit margins are the same, BizFixe’s total profits always trail BizVar’s by the price of its patent investment ($100). But the story in the margins reveals the advantages of BizFixe’s. Every additional unit that BizFixe sells has a greater portion of its revenue flow directly to the bottom line. This is the effect of operating leverage.

Taking this already contrived example to its absurd extreme, as sales approach infinity, the fixed costs of BizFixe are so negligible that revenue essentially equals profit.

The downside of operating leverage is that, like all kinds of leverage, it amplifies effects in both directions. To give this contrived example slightly more realism, assume that BizFixe has annual fixed costs of $100 and that widget demand changes year over year. In down years, the high fixed costs drag BizFixe into deep losses, while the more nimble BizVar remains profitable through the trough years.

The real world, of course, is far more complex. First of all, this simple model of operating leverage has only one variable, sales volume. If either BizVar or BizFixe has power to set prices, that alone complicates the story significantly. Besides, all firms have some costs they can control and some they can’t. Fixed costs can vary and variable costs can become fixed.

The real data will show more nuanced stories.

About this data and its calculations

All data here is sourced directly from annual data in SEC filings.

Each chart plots revenue against operating costs. The space between the two points represents the year’s operating income. The wider the gap between, costs and revenue, the wider the profit or loss.

Sometimes, sales and costs trend in opposite directions. This “scissors effect” may reveal dramatic changes of fortune, where rising costs and declining revenues result in nasty losses.

These scissors cut both ways, though; some charts show growing revenues against decreasing costs.

Also note that these charts use operating income, not net income or cash-based measures. Using operating income isolates the firm’s operations from the money it pays out to taxes and debtors, expenses which depend more on location and capital structure than the nature of the business itself. One corrolary here is that interest expenses are generally fixed costs of another category, so the firms with heavy levels of indebtedness may experience a further amplification of their operating leverage.

Selection criteria

The selection criteria samples from the highest ranking firms for two different measures of operating leverage. One measure calculates the degree of operating leverage by dividing operating income growth rate for a given period by revenue for that same period (operating income CAGR / sales CAGR). The other calculates the ordinary least squares for revenue and operating income. There is some mathematical nuance in the differences, which I describe in the appendix at the end of this post.

A mega-gallery of charts is also in the appendix.

Some summary categorizations

Looking at the top industry classifiers by sensitivity, the gas, mining, pharma, and shipping sectors clearly dominate. Interestingly, the degree-of-operating leverage sort doesn’t have any such overrepresentation by a particular industry.

Looking at the top 100 by market cap, the degree of leverage frequencies skew to the larger caps. My guess is that firms that steadily compound revenue and margins for many years naturally grow into new capitalization buckets. By measures of sensitivity, though, the capitalization frequency distributes more normally across capitalization. Still, the dominance of midcaps is probably telling―it takes some growth and size to realize leverage.

The small relative frequency of megacaps isn’t very informative, because not many megacaps exist. However, the small frequency of microcaps is telling. For comparison, microcaps make up about 45% of the firms in my database and only have 5-12% representation in these datasets.

The charts

The following sections group firms in a few ways:

• By common industry

• By belonging to the classification I term “boomer”

• By coming to disastrous ends

Shipping

The shipping business, like all endeavors on the sea, requires a strong stomach. The up-front fixed costs are enormous, and industry sales are highly sensitive to the prevailing charter rates of the day. This makes for boom and bust patterns, giving these firms among the most dramatic charts of the set.

Oil tankers

First, let’s look at the firms that specialize in transporting oil. In a later section, I also show charts for various oil and gas services. Comparing these two related industries, the dominant pattern is that:

• Oil firms had trough years in 2020

• Oil shipping firms peaks in 2020 and troughs in 2021.

For context about this lag, this 2022 article from Global Energy describes how, as oil demand cratered, tankers were needed as floating storage. This effect reversed when prices started to normalize, storage demand dropped, and production cuts created less oil to be moved or stored anywhere.

Euronav is a rare case of simultaneously declining costs and increasing profits. What’s the secret? The declining costs, I suppose, are the result of depreciation amortization.

International Seaways has kept fixed costs quite stable, and the margins in the good years are very good.

Nordic American Tankers is a small oil tanker company. Note how 2021 was a particularly bad “scissors” year, where costs spiked and sales plummeted. This unfortunate effect shows up in many charts in this article. Perhaps crisis years also coincide with high incidental costs? Perhaps companies use these years as dumping grounds for impairments and other “kitchen-sink” expense recognition.

DHT also transports oil. But not even 2021 could tip its margins into the negative, though they contracted to a suspicious level of 0.0%.

Scorpio Tankers, AKA STNG, got stung in 2021 like the rest. But its relatively low base of costs appears to give it more support to absorb the pain.

Other shipping firms

Here’s a grab bag of other shipping firms that showed up in the screen. A 2021 article from Bloomberg describes how the year was a disaster for tankers but a boom for other shippers.

Global Ship Lease is a Greek container shipping company. I’m not sure why my database doesn’t have data for 2019 and 2020. There might be an interesting and less profitable story than what appears here.

Dorian LPG transports liquified petroleum gas. It is the gas tanker here that didn’t have a bad 2021. Crude and LPG transports apparently exist in different economic cycles.

Star Bulk Carriers does dry bulk shipping of various heavy commodities.

Boomer stocks

Boomer stocks is not a real industry classifier (yet), but in this arbitrary group I’m putting firms that are extremely well known and mature. Achieving operating leverage at this scale is a different form of achievement, involving, I assume, intense discipline in product development, cost control, and acquisitions.

First, I’ll highlight Proctor & Gamble and CAT, both of which have diverse product lines and sprawling operations.

Proctor & Gamble. Two years stand out as periods to investigate: 2015 had declining costs and revenues and 2019 had rising revenues but sharply rising costs. Perhaps 2015 was a period of divestiture and 2019 was a period of investment. I should note that the high relative revenue in 2015 causes a distorted DoL calculation by the two CAGRs. If I’d started in 2016, PG might not have appeared in my sort. No matter the year, though, the trend shows steadily expanding margins while revenue grows.

Caterpillar is a similar diversified, well-known megacap. Compared PG and many others in this list, CAT’s expense line tracks much closer to its revenue line, indicating a higher relative degree of variable costs and better stability when revenue plunges. Nevertheless, over the years, more revenue has flowed to the bottom line.

Coca-Cola Consolidated. The chart here is not for KO, the beverage and brand giant, but COKE, the bottler and distributor of that beverage. So while it’s not quite as “boomer” as the other two boomer stocks, I include it to show the effect of thin, steadily widening margins. Variable cost control and economies of scale are probably the story for all these boomer giants.

Software compounders

Because its gross margins are so good, software businesses are often literal textbook examples of operating leverage. After initial R&D and infrastructure costs, marginal costs for each user are often extremely low. The existence of network effects and knowledge barriers can create durable moats. The industry also benefits from performance gains that come from innovation in the hyper-competitive hardware industry that it depends on. Finally, automation is the industry’s raison d’etre, so the variable costs of maintenance can be gradually automated away through smarter algorithmic processing and infrastructure deployment.

Netflix is an example of all of these effects.

Trip.com Group. I’d never heard of this company but I’ve surely visited one of the many travel aggregators that it owns. Margins contracted sharply during COVID, but have expanded after the bounce.

AppLovin. Something about the name AppLovin makes me want to puke, but this adtech firm has the most ridiculous chart from the whole set.

Fiserv is an interesting case of having an almost lockstep pattern for many years before it suddenly unlocked a new period of leverage in the 2020s. Fiserv also recently had an earnings blow up that came after it acknowledged that it had unsustainably aggressive growth and income recognition. This may be a false positive.

Palantir’s chart shows a clean double helix centered at 2023, when costs and expenses finally intersect before diverging in opposite directions. This story, of course, is still playing out.

Microsoft is another textbook case of operating leverage, one that I’ve literally seen referenced multiple times in textbooks, and yet I’m still surprised by the relentless, decade-long expansion of margins, from 20% to 45% over the course of 11 years.

Verisign is not well-known to the public, but it sits deep in the stack of internet infrastructure―nameservers and DNS, the stuff even web developers prefer not to think about. By the numbers, VRSN doesn’t have the best degree of operating leverage or sensitivity, but flat costs and rising revenue show a firm awash in profit.

Commodities.

Commodity firms, like shipping firms, take their prices from the market. This condition brings fantastic years when commodity prices are high and disastrous years when they are low. Disciplined control of costs and balance sheets makes the firms that best survive the boom-and-bust cycles.

Upstream Oil

While the oil-shipping firms all showed terrible years in 2021, the oil services themselves suffered most when demand cratered in 2020.

Baker Hughes is in the oilfield services business. Literally off the charts scissors effect in 2020, I think this is from a giant goodwill impairment.

Magnolia is another Oil and Gas company with a huge cost spike in 2020. Since then, its costs have fallen well below revenues, though operating margins obviously heavily depend on the price of oil.

EOG also appeared in one of my searches for firms with the greatest fixity of cost. When demand for oil services is high, this fixity gives EOG a floor to capture a lot of margin. Fun fact, the E in EOG used to stand for Enron, but this company spun off as a “boring asset” before the firm famously went off the accounting rails.

Targa obviously has more variable costs than others in this list.

Texas Pacific Land Corp really doesn’t fit in this section, since its business is oil real-estate and rights rather than exploration, but the chart is just absurd. Not many businesses can look like this.

Other commodities

CVR partners makes nitrogen, ammonia, and urea products. This will be an interesting chart to revisit once the effects of the 2026 closure of Hormuz are better manifested in its income statements.

CF produces ammonia. Yes, another one to watch with the closure of Hormuz.

Timber! (I refer here to both the commodity and West Fraser’s revenue collapse).

Alamos Gold has a striking chart; costs have stayed flat while gold prices went parabolic.

ICL Group. I don’t have much to say about this company that produces Potash and Phosphate. Until recently I thought potash was a type of midwestern breakfast.

Southern Copper Corp. Whoa, this company seems to keep a tight leash on costs. I suppose this is what you want to see as an owner of a minerals producer (Alamos Gold, above, is another example. If copper prices climb with the data center build out, the margins may get obscene.

And a little bit of gore

Of course, leverage cuts both ways. I’d be remiss to omit some disaster charts.

NAAS. WTF is this? One of the worst scissors effects I found. Note that this company’s chart ends in 2023, when it went private to transition out of English-language training and into electric vehicles.

Abcellera shows one of the cleanest charts for the downside of operating leverage.

AMBA. Semi-conductors are a hot business, but obviously the fixed costs are punishing if demand for the product isn’t there. Now that it’s raining more than ever, I won’t be standing under this Ambarella.

What to do with all this information?

Operating leverage, like all forms of leverage, is a relationship of amplified effects. Just as physical levers amplify force and financial levers amplify investment, operating levers amplifies changes in revenue.

One thing that’s clear from these charts is that a business that manages to realize revenue growth while controlling fixed costs can enjoy enormous margin expansion. If you can reliably see revenue into the far future, a firm with low variable costs is the most profitable around. But predict wisely, because the leverage cuts both ways.

Appendix

I originally wanted to include galleries for all charts directly in Substack, but the “gallery” widget only permits 9 at a time. This post is already big enough, anyway. Instead I’ll provide an Excel file and PDF charts of the top firms by each measurement (calculation notes in the next section).

Oplev Top150

38KB ∙ XLSX file

Download

Download

Oplev Dol Top100

10.3MB ∙ PDF file

Download

Download

Oplev Sensitivity Top100

10.3MB ∙ PDF file

Download

Download

Calculation notes

This dataset calculates operating leverage over every firm in two ways. First, the simple method divides the cumulative annual growth rate (CAGR) of operating income for a given period against the revenue for that same period. This simple calculation provides an understandable view of the degree of operating leverage over a long run. The downside is that I needed to filter out starting years of operating losses, and the calculation is biased towards first and final years.

For another view of operating leverage, I also calculated the ordinary-least squares line for revenue and operating income. This calculation provides a better degree of the sensitivity of the margins given a fixed cost base: every dollar of revenue above the fixed base results in this some percentage of profit after variable costs. Higher scores mean more sensitivity. With good operating leverage, the farther that revenue climbs above the fixed cost base, the more margins expand, and vice versa.

One other technical note is that I calculated operating costs by subtracting the reported operating income from the reported revenue. The reason I did this was that these two measures are standard GaaP numbers, available in virtually all reports. The more granular measures, like costs of goods sold and SGA, would provide more detail, but there’s much variation over how they are reported. Trying to catch each case is difficult over a dataset of this size. Certainly though, it’s easy to do for an individual firm. If you’d like me to disaggregate one of the charts for any of these firms, let me know.

This portfolio gets marked to market on at the end of each quarter and the gains or losses are booked as unrealized operating income or loss.

Yesterday, on April 29, 2026, Markel reported its Q1 results. The EPS as reported shows a horrific reversal: $12.11 for the first three months of 2025; negative $18.90 for the first three months of 2026. Terrible!

Yes, terrible! And within 30 minutes of market opening, the stock fell almost 8%.

Yet, if you look at adjusted operating income, which adds back the unrealized losses, the story is not so dire.

The next question is how permanent are these “unrealized” losses. The portfolio was marked-to-market at March 31, 2026. Before getting into Markel’s specific investment portfolio, I’ll post this chart of the last six months of the S&P to provide a sense of the overall market dynamics. For those who remember this distant time, it’s been a bit volatile.

Interesting: March 31 is just one day beyond the absolute six-month trough. Good thing that March 30th wasn’t the day Markel had to mark to market: operating losses would have been much worse.

Of course, Markel’s portfolio is not the S&P. Fortunately though, we don’t need to guess what its holdings are. The entire equity portfolio is reported in the 13Fs filed with the SEC. Using the latest 13F, filed on February 6, 2026, I constructed a chart that plots the portfolio’s month-end values over the last 7 months.

Now a new story emerges: Markel would actually report an income gain if its portfolio were marked to market today.

This chart was hastily done and it surely does not capture the whole story. The actual reported loss was ~$720 million, not $560. This discrepancy is likely because:

• The latest 13F is a four months stale

• My script missed a few tickers

• during Q1, Markel bought $80M of new equity and sold $52M.

But, unless Tom Gaynor, Markel’s CEO, has spent the last 4 months going all in on BDCs and energy shorts, the actual portfolio movement surely tracks pretty closely to this. A question in the earnings call addresses this directly.

Tom Gaynor confirms the rebound in the earnings call:

Andrew Klingerman (Analyst at TG Cowan)

My first question is around book value per share. It sequentially in the quarter came down to 1553 from 1566 and it was in part due to a loss on the equity portfolio of about $58 million. So I’m wondering now with the S&P 500 year to date up, I think more than 9%, where would that book value be now? You know, assuming that equity gain, could you help out with that in the equity portfolio?

Tom Gaynor (Chief Executive Officer)

Yeah. Andrew, it’s Tom. It would be higher. We would have to do mid period calculations on that. But your math would be correct and the number would be higher. That first quarter equity market volatility is something we’ve had decades of experience, with. It’s this normal mark to market stuff. Those changes were unrealized gains and losses, not anything realized.

But what makes the price go up and down?

So, what made the stock slide 7-8% on earnings?

I have two working theories:

a) The market detected an inherent softness in its actual operating position and has deftly priced the declining cash flows coming from its insurance and subsidiary operations.

b) The market over-reacted to a “paper” loss in operating income, which, at the time it was reported already would amount to gain.

If the latest 10-Q actually reveals a new and less rosy picture of Markel’s future, then I am wrong. Feel free to tell me I’m wrong in the comments.

My bet, though, is that much of the sell off today is a response to the sharp decline of reported earnings per share. If that is the case, I further argue that this sell off was broadly “mechanical.” Trading programs react to EPS changes. Operating income is even more important, and Markel is somewhat unusual in that its investments are considered operating rather than financing income. My hypothesis is that the “dumbest” programs and investment processes blindly overreacted to these two superficially awful signals. I say dumb because actually calculating the current value of Markel’s portfolio is not a difficult task―I approximated it in about 40 minutes and I’m no genius.

So, if you are interested in Markel and believe, like me, that today’s sell off came from an overreaction to an already irrelevant signal, then today is a good time to buy Markel. Even if the market continues to undervalue the stock, the company in all likelihood will continue to compound its book year over year.

If the share price considerably trails its intrinsic value, buybacks will be highly accretive. Markel has plenty of cash; they can surely afford to bring some more shares into its treasury. At the end of the earnings call, Tom Gaynor mentions this option explicitly.

Andrew Klingerman (Analyst at TG Cowan)

Got it. And I guess lastly, stock looks like it’s trading off about 7% this morning. It looked like, you know, some pretty stellar property I’m sorry, property casualty results a little mixed elsewhere. What do you, what do you, what do you gentlemen think it takes to kind of get the stock moving up

Tom Gaynor (Chief Executive Officer)

North from here performance And I think we’ve demonstrated a few quarters of doing that and if the market disagrees with the performance that’s happening, we’ll continue to repurchase shares. And we are price sensitive in our repurchasing. So when things are, when the stock price is going down, we buy more.

If you agree with the logic of this post, today is a good day to buy Markel. If you don’t, comment! I welcome critique.

I got the idea for this screening criteria from the Intelligent Investor. In chapters for both defensive and enterprising investors, Ben Graham mentions that a long period of earnings stability is a foundational quality to screen for.

[The investor] should require an indication of at least reasonable stability of earnings over the past decade or more—i.e., no year of earnings deficit—plus sufficient size and financial strength to meet possible setbacks in the future.

Of the ~2,000 companies that filed with the SEC and reported earnings every year from 2015 to 2025, only 742 never had a losing year. I built a table of all these firms and their reported series of annual incomes. Then I ran some calculations on the year-over-year values to compute 10-year cumulative annual growth, overall variance, regression-fit growth, and more. I also ran the numbers against a stock API to get metrics about the current share prices―stuff like PE, EV, and so on―for each firm. There are many, many ways to filter the data down.

Holy smokes: there are some good leads here!

The results have a fascinating mix of capitalization and sectors. Some patterns are predictable, others surprising. We’ve got big companies, tiny companies, some companies you know and some you don’t, some companies whose income compounds with the smooth convexity of a spoon, others whose growth looks like the craggy profile of Mt. Fitzroy. Yet all 742 firms have one thing in common: durable earnings through periods of economic shocks.

The following 3000+ words of analysis and visualizations is free for all. Along the way, I give out quite a few tickers. However, there’s a paywall at the very end of this article, and right behind that is a dataset. Subscribers get access to every dataset I put out all year.

The dataset

The dataset has a few separate tables.

• Company-info. Reports basic facts about each firm in the list. Information like CIK, ticker, and Industry Classifier is canonically recorded here. For user convenience, this info is duplicated in other tables.

• Annual-time-series. Records net income values across ten years for each firm. All values across all years are positive: that’s the entry criterion. Besides net income, there are other measures, such as EPS, and shares outstanding, provide important context about the income story. I will be adding more GAAP criteria to enrich the data. These are all sourced directly from SEC filings.

• Growth-metrics. Provides summary statistics about income growth. The most obvious metric here is cumulative annual growth rate (CAGR). But CAGR doesn’t tell the whole story, so there also statistical measurements like r-squared (to capture growth smoothness), fitted growth (to smooth out anomalies that may have happened from unusual income in the first or last year), and so on.

• Spot-metrics. Provides up-to-date screening metrics sourced from Finnhub’s API. These metrics typically incorporate share price, e.g. price-to-earnings, market capitalization, EV, and so on. But it also includes metrics like FCF growth and others. Latest upload is from April 22, 2026.

Some highlights from the growth metrics

First, I’ll highlight some interesting firms and trends derived from the growth metrics table.

The companies that grew earnings for nine straight years

These twelve companies have grown their income every year for nine years. If we include the year 2015 in the data, though, the number of qualifying companies falls to just nine. Either way, these firms are of a truly rare breed―less than 1% of the ~2000 firms that filed for every year in the range.

I’m struck that every firm in this elite list occupies a different industrial classifier (though in less granular classifications, a pair exists in the banking and utilities industries). Only two firms here are in the small-cap bucket, which is probably predictable―constant growth eventually bumps you up a cap.

┌───────────────────────────────────┬─────────┬────────────────────────────────────────────────────┬─────────┐
│               Firm                │ ticker  │                         sic_desc                   │ mkt_cap │
│              varchar              │ varchar │                         varchar                    │ varchar │
├───────────────────────────────────┼─────────┼────────────────────────────────────────────────────┼─────────┤
│ ADDUS HOMECARE CORP               │ ADUS    │ Home Health Care Services                          │ $1.8B   │
├───────────────────────────────────┼─────────┼────────────────────────────────────────────────────┼─────────┤
│ AGREE REALTY CORP                 │ ADC     │ Real Estate Investment Trusts                      │ $9.3B   │
├───────────────────────────────────┼─────────┼────────────────────────────────────────────────────┼─────────┤
│ AMETEK INC/                       │ AME     │ Industrial Instruments for Measurement,            │ $53.7B  │
│                                   │         │ Display and Control                                │         │
├───────────────────────────────────┼─────────┼────────────────────────────────────────────────────┼─────────┤
│ CHIPOTLE MEXICAN GRILL INC        │ CMG     │ Eating Places                                      │ $46.1B  │
├───────────────────────────────────┼─────────┼────────────────────────────────────────────────────┼─────────┤
│ COSTCO WHOLESALE CORP /NEW        │ COST    │ Variety Stores                                     │ $433.6B │
├───────────────────────────────────┼─────────┼────────────────────────────────────────────────────┼─────────┤
│ IDACORP INC                       │ IDA     │ Electric Services                                  │ $7.9B   │
├───────────────────────────────────┼─────────┼────────────────────────────────────────────────────┼─────────┤
│ INSTALLED BUILDING PRODUCTS, INC. │ IBP     │ General Building Contractors-Residential Buildings │ $7.9B   │
├───────────────────────────────────┼─────────┼────────────────────────────────────────────────────┼─────────┤
│ O REILLY AUTOMOTIVE INC           │ ORLY    │ Auto and Home Supply Stores                        │ $78.2B  │
├───────────────────────────────────┼─────────┼────────────────────────────────────────────────────┼─────────┤
│ PATHWARD FINANCIAL, INC.          │ CASH    │ National Commercial Banks                          │ $2.1B   │
├───────────────────────────────────┼─────────┼────────────────────────────────────────────────────┼─────────┤
│ QCR HOLDINGS INC                  │ QCRH    │ State Commercial Banks                             │ $1.5B   │
├───────────────────────────────────┼─────────┼────────────────────────────────────────────────────┼─────────┤
│ XCEL ENERGY INC                   │ XEL     │ Electric and Other Services Combined               │ $51.4B  │
├───────────────────────────────────┼─────────┼────────────────────────────────────────────────────┼─────────┤
│ ZOETIS INC.                       │ ZTS     │ Pharmaceutical Preparations                        │ $50.9B  │
└───────────────────────────────────┴─────────┴────────────────────────────────────────────────────┴─────────┘

The CAGR superstars: an unlikely duo

In the growth-stats table, a composite metric assigns a score that rewards growth that is both high and stable (CAGR x r2). By this measure, the firms at the top two spots have compounded earnings at an astounding rate. Besides that, they could not be less alike.

Firm #1 has a market cap of $238 million; firm #2 has a market cap of nearly $450 billion.

You probably know Firm 2. Besides its extraordinary market value, is also the unique member of the “Video Tape Rental” category in the entire dataset. The other firm, however, is one you almost certainly don’t know. It’s a weird microcap from the US midwest with a pair of businesses that have no synergy. While its growth is lumpier, it has compounded earnings for ten years at a rate of 100%!

Any guesses? (Hint: Firm #2 stopped delivering physical DVDs some time in the early 2000s).

┌─────────┬───────────────────────────────────┬──────────────┬────────┬────────┬─────────────────┐
│  firm   │            Description            │  market_cap  │  CAGR  │   r2   │ composite_score │
│         │                                   │  (millions)  │        │        │                 │
├─────────┼───────────────────────────────────┼──────────────┼────────┼────────┼─────────────────┤
│   #1    │ Amusement and Recreation Services │     $238.36  │ 100.47 │   0.66 │           66.13 │
│   #2    │ Video Tape Rental                 │  $449786.94  │  57.26 │   0.91 │           52.07 │
└─────────┴───────────────────────────────────┴──────────────┴────────┴────────┴─────────────────┘

Yes, firm #2 is Netflix. Over ten years, it has compounded earnings at an annualized rate of 57% and, as impressively, has done so very smoothly, with an r-squared of 0.91.

Yet, the growth gold medal goes to Butler National Corp. (BUKS). One part casino, one part small aircraft repair, this microcap centaur is an enigma. And yet, even adjusting for variations, it is by all measures of earnings growth―CAGR, fitted growth, and composite―the best of the entire dataset.

Smooth and Spiky compounders

These charts compare the log-adjusted net incomes for two sets of firms whose earnings compounded at at least 7.5% a year. The smooth set shows the five firms with the highest r-squared scores. The spiky set displays the firms with the lowest r-squared values from an arbitrary floor of 0.1 (as r-squared approaches zero, the scatter plot approaches purely random noise).

I must admit: having an r-squared of 0.1 seems stressful:

The tickers for the highest R2s are for CostCo, Addus Homecare, Agree Realty Corporation, Fair Isaac, and Intuit. The lowest are National Healthcare Corporation, ABM industries, Franklin BSP Realty, HealthStream, and Bankwell financial group.

While the prospect of smooth compounding certainly presents a more soothing investor experience, it’s also possible that the volatility of the high-variance group presents more possibilities for mispricings. Besides that, these smooth compounders are well-known and the market may put a premium on their perceived lack of risk. Fair Isaac and Costco immediately come to mind as great companies whose market cap nevertheless approaches an illogical level of optimism.

That said, it’s notable that two of the smoothest growers, Addus and Costco, also appear in the list of 14 firms with 9 years of consecutive earnings growth. There might be a lesson there.

Industry classifiers

The firms in the dataset are also quite diverse in terms of industry classification and market capitalization. Nevertheless, a few key industries have a relative concentration. These are typically the businesses that have very good innate economics―banks, market makers, insurance, and so on―the kind of businesses that turn money into more money.

The dataset’s firms most often land in the large- and mid-cap buckets. Here’s a visualization that represents the number of firms and types of industries in each capitalization band.

Note how the concentration of industries changes across capitalization:

• In the microcap bucket, banks dominate the dataset.

• As we move up capitalization buckets, REITs play an increasingly important role.

• In the megacaps, the regional banks and REITS all but disappear. Yet “prepackaged software” becomes a significant category.

• National banking is a significant player across all buckets.

Of course, as shown by the large swaths of gray, there’s a great diversity of “other” industries across all capitalizations. This catchall category includes some one-off oddballs, like video rentals and malt beverages. It’s also interesting that the other category becomes proportionately larger with capitalization size. On the other end, the microcap bucket has the most predictable set of industries.

One interpretation of this finding is that certain business types, notably regional banks, have outstanding earnings durability but cannot scale beyond a certain size. On the other end, some businesses are just sui generis and make profits out of any category―even a “Video Rental Service” might end up in the megacap bucket.

Industry classifier and median growth

Here are the numbers that represent the median CAGR and regression-fit growth rates for industries that appear in the dataset at least 8 times.

┌──────────────────────────────────────────┬───────┬────────────┬──────────────┐
|            IndustryClassifier            | count | MedianCAGR | MedianFitted |
|------------------------------------------|------:|------------|--------------|
| Prepackaged Software                     | 13    | 19.56%     | 21.70%       |
| Operative Builders                       | 10    | 15.30%     | 21.98%       |
| Services-Misc Business Services NEC      | 15    | 14.37%     | 13.77%       |
| State Commercial Banks                   | 106   | 11.51%     | 10.17%       |
| National Commercial Banks                | 55    | 11.08%     | 10.50%       |
| Fire, Marine and Casualty Insurance      | 11    | 10.68%     | 9.01%        |
| Semiconductors and Related Devices       | 9     | 9.57%      | 13.29%       |
| Electric Services                        | 10    | 7.31%      | 4.42%        |
| Electric and Other Services Combined     | 12    | 7.20%      | 7.16%        |
| Real Estate Investment Trusts            | 36    | 7.13%      | 6.53%        |
| Investment Advice                        | 17    | 5.71%      | 5.67%        |
| Motor Vehicle Parts and Accessories      | 9     | 4.23%      | 2.82%        |
| Savings Institution, Federally Chartered | 17    | 3.23%      | 6.20%        |
└──────────────────────────────────────────┴───────┴────────────┴──────────────┘ 

Software has been the bull story over a long bull run, so I’m not surprised to see it at the top spot. However, the fact that the regression-fit median is a bit higher than its overall median CAGR might indicate industry growth has slowed since the COVID boom. He

It’s also interesting to note the relative spread of cumulative annualized returns. Semiconductors have some powerful growers―and some shrinkers. Business services and banks all have fairly low spreads, with a respectable CAGR overall.

I’d also like to highlight the respectable CAGR of “State Commercial Banks,” the industry that has by far the largest sample in the dataset. Of those 106 firms, half have compounded at a better rate than 11% over ten years. My guess is that this sector also has the most hidden value for the investor willing to figure out the loan books.

PEs by industry classifier

Here are the Median PEs for the industries that appear at least 8 times in the data set:

┌──────────────────────────────────────────┬──────────┐
│            IndustryClassifier            │ MedianPE │
├──────────────────────────────────────────┼──────────┤
│ Semiconductors and Related Devices       │  40.8108 │
│ Real Estate Investment Trusts            │  29.1368 │
│ Prepackaged Software                     │   25.338 │
│ Motor Vehicle Parts and Accessories      │  17.2257 │
│ Services-Misc Business Services NEC      │  17.0047 │
│ Investment Advice                        │ 13.43845 │
│ Savings Institution, Federally Chartered │  12.8797 │
│ National Commercial Banks                │  12.5442 │
│ State Commercial Banks                   │  12.2532 │
│ Fire, Marine and Casualty Insurance      │  10.3524 │
│ Operative Builders                       │   9.8035 │
└──────────────────────────────────────────┴──────────┘ 

Given the enormous expectations the world has for the semiconductor industry, it’s no surprise that it has the highest median PE in the group. Finding REITs in the second spot number isn’t much help for value hunters, either, since PE isn’t a very useful measure in this depreciation-heavy industry.

Towards the bottom of the list, though, we might be getting into value territory―or we might not. Are the builders of single-family homes undervalued―or is the market already pricing in a cyclical downturn? And what’s the story with Fire, Marine, and Casualty insurance―have we entered a new period of diminished force majeure? I don’t know the answers to these questions. But if you use the industry classifiers as starting points, you may find some treasures among the individual firms.

On that note, let’s look a bit more at PEs from the level of the firm rather than the industry.

Firms and their price-to-earnings ratios.

For better or worse, the PE ratio is investing’s most famous metric. Since this dataset is based directly on earnings―that’s the E in PE―this writeup would be incomplete without some more PE stats from this data.

One interesting thing to note is that most firms in this dataset have PEs that are below their ten-year CAGR. Butler National and Stride ( ) stand out as particular anomalies.

One other note: the blue dots indicate lower r-squared statistics for their ten-year growth. When viewed as a whole, the obvious finding is that consistency of earnings strongly overlaps with high overall cumulative returns.

Highest and lowest PEs

First of all, let’s look at the highest and lowest PEs in the dataset. I’ll present the top five from each end.

• Lowest PEs. Are these firms super cheap deals, or simply traps in government conservatorship? (at least two of these are in government conservatorship)

┌──────────────────────────────────────────────────┬────────┬─────────────────────────────────────────────────┐
│                       Firm                       │ PE_TTM │                    Industry                     │
├──────────────────────────────────────────────────┼────────┼─────────────────────────────────────────────────┤
│ FEDERAL HOME LOAN MORTGAGE CORP                  │ 0.4319 │ Federal and Federally-Sponsored Credit Agencies │
│ CHS INC                                          │ 0.5544 │ Farm Product Raw Materials                      │
│ FEDERAL NATIONAL MORTGAGE ASSOCIATION FANNIE MAE │ 0.6619 │ Federal and Federally-Sponsored Credit Agencies │
│ COMCAST CORP                                     │  5.069 │ Cable and Other Pay Television Services         │
│ STAR GROUP, L.P.                                 │ 5.3837 │ Retail Stores, NEC                              │
└──────────────────────────────────────────────────┴────────┴─────────────────────────────────────────────────┘

• Highest PEs. Are these firms overvalued, or poised for explosive growth?

┌───────────────────────────────────────────┬──────────┬───────────────────────────────────────────────┐
│                   Firm                    │  PE_TTM  │                   Industry                    │               │
├───────────────────────────────────────────┼──────────┼───────────────────────────────────────────────┤
│ NORTHFIELD BANCORP, INC.                  │ 720.9011 │ Savings Institution, Federally Chartered      │
│ ON SEMICONDUCTOR CORP                     │ 234.8818 │ Semiconductors and Related Devices            │
│ KNIGHT-SWIFT TRANSPORTATION HOLDINGS INC. │ 152.6329 │ Trucking, Except Local                        │
│ DARLING INGREDIENTS INC.                  │ 150.6315 │ Fats and Oils                                 │
│ QUAINT OAK BANCORP, INC.                  │ 135.2486 │ Savings Institutions, Not Federally Chartered │
└───────────────────────────────────────────┴──────────┴───────────────────────────────────────────────┘

• I was not expecting to see a firm in the “Fats and Oils” classifier here―a market “darling” indeed.

PE vs Growth Rate

In some ways, the PE ratio is a crude proxy statistic for the market’s expectations of growth. Comparing a firm’s PE against its historic growth rate, then, is a crude way to relate past performance with current expectations. The PEG ratio, popularized by Peter Lynch, measures this by dividing the PE ratio by earnings-per-share growth.

A PEG ratio of less than 1 indicates that historic growth (the denominator) is larger than current PE (the numerator). The implication is that, if historic growth continues, the PE undervalues the stock.

Here are the firms with the ten lowest PEG ratios. For this calculation, I’m using only the last three years of EPS growth. The reason is more mechanical than analytical―stock splits make it hard to reliably track EPS growth for such a bulk of data.

┌───────────────────────────────┬────────┬────────────────┬────────┬─────────────────────────────────────────┐
│             name              │   pe   │ eps_growth_3yr │  peg   │                   sic                   │
│            varchar            │ double │     double     │ double │                 varchar                 │
├───────────────────────────────┼────────┼────────────────┼────────┼─────────────────────────────────────────┤
│ PROGRESSIVE CORP/OH/          │   10.2 │          153.1 │   0.07 │ Fire, Marine and Casualty Insurance     │
│ COMCAST CORP                  │    5.1 │           64.3 │   0.08 │ Cable and Other Pay Television Services │
│ GLOBAL PAYMENTS INC           │   13.5 │          141.7 │    0.1 │ Services-Misc Business Services NEC     │
│ ASSURED GUARANTY LTD          │    7.4 │           74.7 │    0.1 │ Surety Insurance                        │
│ HANOVER INSURANCE GROUP, INC. │    9.5 │           78.4 │   0.12 │ Fire, Marine and Casualty Insurance     │
│ LIFEVANTAGE CORP              │    6.9 │           49.4 │   0.14 │ Pharmaceutical Preparations             │
│ ARCH CAPITAL GROUP LTD.       │    7.8 │           44.8 │   0.17 │ Fire, Marine and Casualty Insurance     │
│ GREENBRIER COMPANIES INC      │   11.0 │           65.7 │   0.17 │ Railroad Equipment                      │
│ UNIVERSAL HEALTH SERVICES INC │    7.5 │           36.4 │   0.21 │ Hospitals                               │
│ RAYONIER INC                  │   13.8 │           61.4 │   0.22 │ Real Estate Investment Trusts           │
├───────────────────────────────┴────────┴────────────────┴────────┴─────────────────────────────────────────┤

So, what’s the story with these extremely low PEG ratios? As always, no single metric should ever be taken too seriously. The high incidence of insurers here suggests that 2022 might have been an unusually horrific year for underwriters. And the high growth rates across the board suggest very good subsequent years. In any case, a rapid rise from a terrible floor will distort EPS growth. So, tepid expectations may not be unwarranted.

Perhaps, a sweet spot exists in looking at firms that have higher PEs and less dramatic spreads. Here’s a random selection of firms whose PE is greater than 15 and 3-year PEG is between 0.5 and 0.7. There is no logic behind this specific filter: just another view through the data prism.

┌─────────────────────────────────────────────┬────────┬────────────────┬────────┬───────────────────────────────────────┐
│                    name                     │   pe   │ eps_growth_3yr │  peg   │                  sic                  │
│                   varchar                   │ double │     double     │ double │                varchar                │
├─────────────────────────────────────────────┼────────┼────────────────┼────────┼───────────────────────────────────────┤
│ CONSOLIDATED WATER CO. LTD.                 │   29.7 │           44.6 │   0.67 │ Water Supply                          │
│ META PLATFORMS, INC.                        │   27.8 │           40.6 │   0.69 │ Computer Programming, Data Processing │
│ EMCOR GROUP, INC.                           │   28.7 │           51.6 │   0.56 │ Electrical Work                       │
│ HA SUSTAINABLE INFRASTRUCTURE CAPITAL, INC. │   27.5 │           46.9 │   0.59 │ Investors, NEC                        │
│ BRIGHT HORIZONS FAMILY SOLUTIONS INC.       │   24.0 │           34.9 │   0.69 │ Child Day Care Services               │
└─────────────────────────────────────────────┴────────┴────────────────┴────────┴───────────────────────────────────────┘

EPS vs Net Income CAGR

In the previous section, I calculate the PEG ratio using EPS growth rather than net income growth. This raises an important point: income growth and EPS growth are not the same thing!

To view the general trend, I’ve plotted five years of EPS growth against five years of income growth. I also removed many of the largest anomalies to control, again, for the effect of splits, which are difficult to account for across such a large batch. If I can ensure accurate EPS measures across all years for all firms, I’ll likely make another post that focuses just on interesting findings from just this relationship. Because even with the present limitations, the results are interesting.

Two big points come to mind:

• Income growth and EPS growth have an obvious positive relationship (thank heavens).

• Capital structure can dramatically affect earnings for shareholders. Even controlling for anomalies, the chart displays a decent amount of dispersion.

Capital allocation considerations that are likely at play here include buybacks on the “better-EPS” side, stock-based compensation and other forms of dilution on the “better-net-income” side, and on both sides is the influence of M&A, whose blade cuts both ways.

For fun, I’ve highlighted two firms on the edges, HCA and O . Even though O has compounded its earnings at almost double the rate of HCA, shareholders of HCA have seen far better EPS growth. I haven’t investigated these firms at all, but this chart may explain the phenomenon. HCA has steadily retired shares over this period while O has steadily issued them:

But what makes the price go up and down?

I’ve tried in this post to give a wide and useful sweep of the dataset, but I really only scratched the surface. The 742 companies are a diverse set, all with a long history of profitability. Some of these are “melting ice cubes”, no doubt, but surely others are undervalued gems. I believe that just limiting the search to the small and micro-cap buckets is enough for an individual to build a durable portfolio. Hell, it’s possible that there’s enough alpha in just blindly building an index of these firms―or at least those from certain industry classifiers.

Besides PE and market capitalization, I did not even touch on the 140 other measurements in the spot-metrics table―stuff like EV/FCF, beta, interest coverage, and much more. There are many ways to slice and dice this data to generate ideas for the investment style that makes sense to you―even if that style is selling short.

All that said, the most important truisms in investing are that past performance does not predict future performance and that prices and values are rarely equal. I’ve done no analysis on how the earnings from the past ten years correlated with price movements. While I’m sure there’s some relationship, returns to shareholders also depend on the expectations that were embedded in the price when the shareholder bought in. Chipotle, for example, has one of the best CAGRs by earnings in the entire dataset. However, while the ten-year return to its stock is far from bad, it’s also far from the 50% rate that its income compounded at. This is because high expectations for the burrito business have always been built into its share price. Even Netflix stock has ten-year returns that lag its earnings growth, though laggard returns such as Netflix’s are nonetheless so outstanding that I doubt any investors have complained. No doubt, using this dataset to test hypotheses about “what makes the price go up and down?” is a worthy area of future research.

Below is the dataset. Good luck!

The data

For some reason, Substack only supports XLSX uploads. In the files below, I’ve uploaded a spreadsheet where each table is a sheet. I’ve also linked a google drive where each of the tables exists as a CSV.

A truncated version of the model is represented as follows.

This PDF has some reference docs about the data tables and their columns. The excel file has a sheet with this reference, too.

Consistent Earners Data Model

269KB ∙ PDF file

Download

Download

And here’s the excel file and link to the zipped CSVs:

Read more

In short, warrants expire in less than one month and would yield a profit today of at least 11%. Insiders own about 40% of outstanding warrants. Yet, there is no information about an exercise window.

~Until I have an answer and decide on my own position, this one is “paywalled” to avoid crowding.~

UPDATE: May 11th, 2026 will be the first and last day left to exercise. Now that the uncertainty is gone, I expect to see very little trading left.

Follow IRSA’s latest filings with the CNV (Argentina’s version of the SEC). Announcements about these warrants should come here.

IRSA ( ) is an Argentine firm with a long history of investments in real estate and development. In 2021, in the trough of a global real-estate crisis and a local financial crisis, it issued warrants to purchase stock1. As the business improved, the warrant exercise price was updated to adjust for dividend payments that the warrant holders did not receive.

In 2025, IRSA’s liquidity situation was healthy enough that the board authorized a cashless exercise option. If used, the cashless option results in fewer shares but better return on invested capital.

By my calculations, at today’s warrant and stock prices, both the cashless and cash exercises offer a nice upside. But there’s one big problem: though they expire on May 12th, there’s no information about the last exercise window! So these are either a steal or worthless.

My suspicion is that the exercise window will be announced quickly and taken advantage of primarily by insiders.

Warrant valuation

The formula to value the warrants has two main components:

1. First, get the net value from the shares received against the costs to exercise the warrants.

2. Second, multiply that value by the risk that either:

   1. The share price falls below the breakeven value;

   2. The warrants are unexercisable.

Obviously, if either risk has a 100% probability, then we are multiplying by zero and the warrants are worthless. The warrants also represent a 5% dilution overhang that needs to be considered if holding for the long term. In my calculations below, I assume full dilution.

Cost vs Reward calculation

In a press release for the exercise window in February 26th, IRSA described the updated exercise costs and warrant-to-share ratio. Don’t worry if you find this passage inscrutable―I put a calculation below.

In accordance with the applicable terms and conditions, each warrant entitles its holder to receive 1.6367 common shares, each with a par value of ARS 10, at an exercise price of USD 0.2641 per share.

As from this exercise window, holders may elect to exercise their warrants under the “Net Exercise with Par Value Contribution” (cashless) modality, as approved by the Shareholders’ Meeting held on October 30, 2025. Under this alternative, holders will receive shares equivalent to the difference between the cash exercise price and the market value of the business day prior to the start of the exercise period, and will only be required to pay the par value of the shares to be issued (ARS 10 per share), plus the GDS issuance fee charged by BNY (USD 0.05 per GDS), in the event of conversion of common shares into GDS.

My calculations of how to determine value are this snippet of python code. Even if you’ve never coded, it should be readable ―these are just variables and arithmetic expressions.

GDS is the term for the ADR; One ADR traded on the NYSE is worth ten shares traded on the Buenos Aires Stock Exchange. Unless the US dollar sharply devalues compared to the Argentine Peso, using the cashless option results in a better return but provides less ADR shares.

    # CAsh exercise
    shares_received = num_warrants * 1.6367
    exercise_cost   = shares_received * 0.2641
    gds_fee      = (shares_received / 10) * 0.05
    warrant_cost = num_warrants * warrant_price
    total_cost   = warrant_cost + exercise_cost + gds_fee
    total_new_shares_cash = total_warrants_out * 1.6367
    diluted_gds_price     = (
        ((gds_price / 10) * shares_outstanding + 0.2641 * total_new_shares_cash) 
         / (shares_outstanding + total_new_shares_cash) * 10
                            )
    worth        = shares_received * (diluted_gds_price / 10)
    profit          = worth - total_cost
    breakeven_price = (total_cost / shares_received) * 10

    # CASHLESS
    cashless_share_price     = gds_price / 10
    cashless_shares_given_up = (shares_received * 0.2641) / cashless_share_price
    cashless_net_shares      = shares_received - cashless_shares_given_up
    par_value_cost           = (cashless_net_shares * 10) / ars_usd
    cashless_gds_fee         = (cashless_net_shares / 10) * 0.05
    cashless_total_cost      = warrant_cost + par_value_cost + cashless_gds_fee
    cashless_net_per_warrant = 1.6367 - (1.6367 * 0.2641) / cashless_share_price
    total_new_shares_cashless  = total_warrants_out * cashless_net_per_warrant
    diluted_gds_price_cashless = (
        (gds_price / 10) * shares_outstanding) \
         / (shares_outstanding + total_new_shares_cashless) * 10
    cashless_worth           = cashless_net_shares * (diluted_gds_price_cashless / 10)
    cashless_profit          = cashless_worth - cashless_total_cost
    cashless_breakeven_price = (cashless_total_cost / cashless_net_shares) * 10

 

Here’s the value that I get for a cash and cashless exercise for exercising 10 warrants and converting them into 1 ADR share. The values use April 10th’s closing market prices of $1.8 per warrant and $15.99 per ADR.

Assuming full dilution:

    CASH EXERCISE
    -------------
    At $1.79 cash breakeven GDS is 13.63.
    Dilution-adjusted current price: $15.31 (undiluted: $15.99)
    
    For 10 warrants purchased at $1.79,
    and a GDS price of $15.99, return is.

    gds_shares_received = 1.6367
    total_cost = 22.304359700000003
    worth = 25.06568962942707
    profit = 2.761329929427067
    return = 12.38%

    =====

    CASHLESS
    -----------
    At $1.79 cashless breakeven GDS is 13.22.
    Dilution-adjusted current share price: $15.31 (undiluted: $15.99)
    
    For 10 warrants at $1.79, GDS=$15.99

    net_shares GDS = 1.3663732520325205
    total_cost = 18.065916752032518
    worth = 20.91792383738316
    profit = 2.8520070853506425
    return = 15.79%
    

When I first wrote the article, the spread was more like $1.6 a warrant and $17 a share―20-30% upside!

Even as the spread has narrowed since the beginning of the week, the numbers still present a nice return for something that exercises shortly. For warrants at these trading prices to be worthless, the warrants either must be unusable or the ADR price must fall by about 25%. While a price collapse is certainly possible, I’ll comment only that I think the underlying stock itself trades at at least a fair price today (DYOR on IRSA).

So my focus is on the other risk: what if these cannot be exercised?

Are they exerciseable?

I believe so, but there are some reasons to worry that the exercise period is easy to miss:

• There is no information about the final window.

• The previous years’ May exercise windows always happened between May 17th and 27th, which would occur after these expire.

• In a February earnings call for Cresud, IRSA’s semi-controlling company, the CFO for both companies, Matias Gaivironsky, stated that February was the last exercise window. However, upon investigation, this statement appears to be about some related warrants from Cresud2 whose exercise period was in the same week and also had a cashless option and indirectly resulted in an ownership stake of IRSA. As you might have noticed from this bullet, there is some circularity in the governance of IRSA and Cresud, which would take a whole other post to untangle.

On the other hand:

• The press release for February says “penultimate” not “ultimate” window. I confirmed this in the Spanish release too3.

• In Form 4 filings from March, the insiders, particularly CEO Elzstain and CFO Gaivironsky, still report holding a massive amount of warrants4. It seems strange that they’d have let them expire rather than using the cashless option to cheaply gain more control.

• The warrants still trade on the open market, mostly very thinly―less than 100 a day―but there still are surges of higher volume. Who is buying and selling these if they are worthless (besides me)?

• They expire on May 12th, not yesterday! If they no longer were usable, the owners should have delisted the warrants.

Yes, I reached out to IR

After multiple attempts, I still haven’t heard from IR. I’ll continue to investigate.

I do own some warrants, but not enough to hurt much if they go to zero. If I get a confirmation of the exercise window, I’ll buy more at the right price.

Update April 30th, 20206: one day exercise window announced

May 11th is the first and final day to exercise these warrants. Since the time of writing, the spread did some weird things, closing for a bit as warrant prices rose as share prices fell, then actually spreading in reverse where the warrant prices traded well above their intrinsic value. Now that the uncertainty is gone, I imagine there will be very little trading at a meaningful spread.

To force change, management came up with an unusual solution: use debt to pay out a giant special dividend. The big shareholder payday wasn’t the point, though: the idea was that interest payments themselves would force efficiency and operational changes in a way that management couldn’t achieve on its own.

This article looks at the conditions and results of Sealed Air’s wager. Whether or not the results were successful, the case presents an interesting use of dividend recapitalization coming from internal actors rather than external acquirers (i.e. private equity). It also presents a view of the feedback system that exists between a company’s financial structure and its operations.

All information about the case comes from Karen Wruck’s 1994 paper, Financial policy, internal control, and performance Sealed Air Corporation’s leveraged special dividend.

Historical and theoretical context

The dividend recapitalization is a fairly common technique in modern private equity, but the Sealed Air case is unique both for its time and because the debt was initiated by internal members.

By the end of the 1980s, leverage was certainly not unique to Sealed Air. As I’ve written in another article, a boom in leveraged buyouts in the 1980s fueled a decade of takeovers. There also was a theoretical argument for debt as a control mechanism, championed by the economist Michael C. Jensen. Jensen’s argument is that free cash flow creates incentives for management to waste money through value-destroying, self-dealing behavior. Adding debt forces a constraint on the use of free cash flow, which in theory drives more efficient operational execution.

While Sealed Air didn’t seem to be the target of any takeover bids, the debt did have covenants to make change of control costly. However, the intent of the debt was not defense from external threat, but internally-driven offense: use debt as an internal control mechanism to drive change.

One Foot Hedge

Looking Back at the LBO Boom of the 1980s

In films and newsreels, the 1980’s are a decade of greed and excess. What makes this period of ten years so evocative of money in the American cultural memory…

Read more

5 months ago · 1 like · Dumas Money

Mechanism

On April 27, 1989, Sealed Air Corporation announced a one-time special dividend of $40 per share. In the months prior, the stock had traded between $44 and $45

, so the dividend represented nearly 100% of the value of equity.

Sealed Air was not a highly leveraged company before the deal, and it paid for the dividend with a stack of debt. Before the dividend, interest coverage was 14.7x. After the dividend, interest coverage fell to 2.1x. Debt as a percentage of total assets swelled from 13% to 135.8%, putting the firm at negative equity. This was a major recapitalization.

Plan of action

The recapitalization came with a new plan for improved operational performance. Each plant had to generate a monthly report on metrics about inventory turnover, quality, yield, changeover, and so on. Managers received new incentives to receive bonuses based on EBITDA targets. Lower and middle managers received equity awards. Frontline workers were supposedly encouraged to make changes and provide feedback to improve operations.

The debt covenants also forced very low spending on capital expenditures: whereas in the 4 years before the dividend Sealed Air spent $13 million a year, in the years following the dividend it could contractually spend no more than $7 million.

Results

I’ll summarize the results of Sealed Air’s experiment from operational and shareholder performance. Note that Wruck’s data covers only the 4 subsequent years, and I didn’t look into how the debt structure may have affected its long-term health.

Operational results

In the years that followed, the combination of debt and new incentive procedures seemed to have an outsized effect on operational efficiency. Wruck’s paper compares Sealed Air to other firms in the industry. At least in the reported time frame, Wruck’s data unambiguously finds better profitability with more efficient use of invested capital.

One surprise is that Sealed Air’s relative R&D expenditure remained relatively high, though less than its level in the 1980s. Capex was notably constrained. It would be interesting to investigate how that affected Sealed Air’s investment requirements and competitive advantage through the rest of the 1990s.

Shareholder results

Wruck compared Sealed Air’s performance against an industry average and found outsized returns in the post-dividend years. She also tried to control for effects from exogenous events and trends and found no serious confounder. This is all to say that Sealed Air investors were rewarded for sticking with the company for the four years that followed its special dividend, and it is likely that the debt and internal changes were mostly responsible for these shareholder rewards.

The type of shareholders changed following the special dividend. The percentage of employees that owned stock rose from 7% to 18%. The institutional investor base changed, as well. The pre-special dividend institutions were conservative funds who looked for reliable dividends and low debt. The new institutional investors―including funds like Tiger Management―were more aggressive, “willing to look beyond negative equity and focus on cash flow projections.”

What could have gone wrong

Wruck’s paper paints an ambiguously positive picture of the effects of the leveraged special dividend. Jensen himself, the priest of the use of debt as a control on waste, later cited this paper to support his theories. However, Wruck’s paper covers only four years and has little discussion of risks involved. While obviously all counterfactual thoughts are speculative, I’d like to take a little time looking at unresolved questions:

• What if the industry had a downturn? While Wruck’s paper compares Sealed Air’s performance to peers and shows positive effects, the comparisons also suggest that the industry as a whole did at least okay. What if a major downturn in the demand for packaging had happened in the early years of the experiment? Risk is symmetrical, and heavy debt in a downturn can punish a firm. If conditions had worsened, the experiment might have forced Sealed Air to lay off frontline workers, reduce capacity and shut down plants, or even enter bankruptcy and forced restructuring.

• Did the new policy of measurement-based performance incentivize gaming the metrics? Plant managers received bonuses based on EBITDA metrics. All organizations are vulnerable to becoming victims of the metrics that they hold most valuable. EBITDA, in particular, is a metric that encompasses a wide range of subjective aggregations, and there are many examples of it being the source of financial shenanigans. Sometimes these effects require a long time frame for their virulence to be revealed (for example, see the controversial legacy of Jack Welch’s time at GE). It would be interesting to investigate how this culture of metrics influenced the operation and whether it had any corrosive effects on performance and morale.

• Did the lowered capex leave the firm vulnerable to disruption? Investments are forward-looking, and capex is an investing use of cash. It is possible that reducing capex by ~45% for many years ended up creating a “debt” of ancient, noncompetitive equipment.

All of these questions require investigation and I don’t know the answer to any of them. But nothing is ever as clear-cut as it seems.

What’s to learn from this?

The Sealed Air special dividend is an interesting case study for over-capitalized companies everywhere.

Constraints reduce choices, which forces a more focused sense of direction. Furthermore, a company’s capital structure influences operations―despite the many financial models that divorce the two. If an order backlog can cause a bankruptcy and there’s no spare money for extra material and parts, a company has no choice but to do more with less. Note that this effect can be negative, too. An obsessive cost-cutting and metrics-driven micromanagement can destroy frontline morale, product quality, and future competitive position. Nevertheless, efficiency and tightness can be a source of innovation―this is one half of the formula for the success of Japanese manufacturing in the 1980s (the other half is the recognition that frontline workers are also the only people with the knowledge to enact continuous bottom-up improvement).

This is all to say that a leveraged special dividend might be a good way for over-capitalized companies to both unlock shareholder value and impose tighter internal controls on inefficient operations. Even companies that enjoy relative monopolies might use such a strategy to simulate competition where none currently exists. Many companies have used debt for buybacks, but buybacks add value only when the stock is undervalued. If the share trades in a fair range, a special dividend presents an interesting allocation alternative. Activist investors also may push for dividend recaps as a way of shaking up complacent companies. But the change doesn’t have to come from outside. Any company that can get a loan can use debt to pay a special dividend. They just might look a little crazy.

In films and newsreels, the 1980’s are a decade of greed and excess. What makes this period of ten years so evocative of money in the American cultural memory?

One might say it that it was the political climate of widespread deregulation and consolidating American geopolitical power. Others might say it was the triumph of shameless consumerism over any other more contemplative form of life. I say it was the LBO.

LBO stands for leveraged buyout. And while neon is no longer fashionable and the future of America’s status as global hegemon is anybody’s guess, the leveraged buyout continues to reconfigure financial markets and society. This was the period that gave birth to the modern private-equity industry, the distressed-debt industry, and the “activist” investor strategy. Understanding the crude and rough LBO takeovers of the 1980s helps us understand the more sophisticated plays of today.

WTF is an LBO?

I’m going to use the phrases “LBO” and “tender offer” a lot in this article, so I’ll provide a quick explanation.

To keep it simple, an LBO is when a purchaser uses a large amount of debt (that’s the leverage) to finance the purchase of a company (that’s the buyout). This works especially well when the target is expected to generate a large and stable amount of free cash flow. By saddling the company with debt, the wisdom goes, the acquirer imposes discipline on a company whose own success has given rise to bloat and waste.

In most examples here, the acquirer purchases a public company through a tender offer, that is, an offer to purchase shares for some premium of the current market price. Since the company is public, the acquirer doesn’t need to collaborate with the board of directors to make a purchase. When the directors resist the acquirer, the takeover is called hostile.

After the purchase, the LBO endgame has a few options. The acquirer may:

• Hold the company perpetually

• Resell it at a higher multiple

• Get exit liquidity through an IPO.

Many places on the fin-ternet detail how to model the viability of an LBO using spreadsheets and EBITDA projections. This is not the place for that.

Chronology of a levered decade

To make sense of the 1980s, one needs to look at the financial conditions that came before it. From 1966 to 1982, the US equity market was flat and choppy, and its chart looks like Bart Simpson’s hair:

With stagflation, oil crises, and the collapse of the “nifty fifty,” equity investors suffered throughout the 1970s. With my recent memory of how much people fretted over the “bear market” of 2022, I can’t imagine what the general sentiment must have been like after 16 years of sideways movement. But we can glimpse the feeling of the times through a 1979 cover of Business Week, which famously called uncle on stocks forever.

However, under this choppy surface of share prices was pent-up value. Warren Buffet, value investor, commented in 1974 that with the cheapness of stocks he felt like an “oversexed man in a harem.” The historical chart of the Shiller CAPE ratio provides quantitative evidence of broad undervaluation, too―evidence that doesn’t require us to imagine Warren Buffet in a harem. Note how 1982 marks a deep trough of relative earnings.

This is to say that investor outlook was irrationally depressed after a decade of economic malaise and two decades of poor corporate governance and haphazard conglomeration. A spring was coiling.

Towards the end of the 1970s, Kohlberg, Kravis, and Roberts (KKR) had emerged as a specialist shop that used debt to buy out companies that were either too small to go public or that were public but trading at a value that was less than the sum of its parts. In 1979, KKR worked with the management of Houdaille, a US manufacturer, to tender the firm’s shares at 2.8 times its market capitalization. This transaction is historic: it was the first time that a public company had been taken private in the US and the largest buyout in history. However, in the light of what was to come, the Houdaille transaction now seems quaint.

The LBO mania of the 1980s is widely considered to have started with Wesray’s purchase of Gibson Greetings, a greeting cards company. Sometimes the first instance of a trend is also its most perfect example. Many literary critics consider Don Quixote to be both the first modern novel and the all-time best novel. I consider Wesray’s ‘80s LBO to be the best ‘80s LBO. Headed by former treasury secretary William Simon, Wesray used a medley of debt financing to purchase Gibson Greetings for $80 million, of which Simon himself staked only $1 million. Sixteen months later, the company had an IPO for $290 million, and Wesray’s founders pocketed $70 million dollars each1. I don’t know of any subsequent LBO that was so profitable and effortless. But others took note. Blood was in the water.

Enter the raider

KKR’s acquisition of Houdaille and Wesray’s acquisition of Gibson Greetings were both done in conjunction with management. But if a company is public, an acquirer doesn’t necessarily need to work with directors and management to purchase the firm. In fact, getting rid of existing management and directors may be the reason to acquire the company in the first place. With enough money and a pugnacious temperament, an acquirer might fight for control. This was the role of the “raider.”

Raiders acquired large positions in firms they wanted to purchase and made tender offers to buy out other shareholders. In the early days of the LBO boom, sometimes just trying to buy a company was a profitable strategy. T. Boone Pickens’ Mesa, for example, tendered for Cities Services and General American Oil in 1982. In each attempt, Boone did not successfully buy the company but profited nonetheless by reselling his shares at a greater price. Extra profits came from exiting the position in General American, whose directors paid Pickens $15 million to withdraw his tender.2 Raiders, like people of other industries, were some times “paid to leave.” This practice of greenmail was a source of early raider profits. Carl Icahn is another well-known benefactor of the greenmail play.

What is this junk?

As the ‘80s raged on, the pace of LBOs was accelerated by a closely intertwined phenomenon: the rise of the junk-bond financing. In the 1970s, companies whose bonds were not deemed “investment grade” floundered in an illiquid neverworld. Michael Milken, whom I’d call the chief financial architect of the 1980s, built an empire at Drexel Burnham based on the sale of these out-of-favor bonds. This is a massive story, one so big that I’m going to skip over almost completely. Connie Bruck’s The Predator’s Ball is a good entry for the curious:

As the 1980s progressed, junk bonds underwritten by Drexel Burnam were a major source of takeover liquidity. Pickens’ bid for Unocal, for example, had a tender in two tiers: the first shareholders who tendered received cash, and the latter received the value of the tender in junk bonds. Sometimes, the bidder did not even need to secure the funds: Drexel Burnham would simply issue its famous “highly confident” letter stating that the bank could secure the funds to underwrite the transaction when the time was necessary.

Junk-bond-financed bidding became the vehicle for “pawns to capture kings” (paraphrasing Connie Bruck). With easy access to large amounts of debt, small companies could use leverage to swallow larger ones. The heavy merger activity also spurred a lively legal industry of takeover offense and defense. Directors adopted increasingly sophisticated “poison pills,” mechanisms designed for shareholders to make takeovers not impossible but highly financially unpleasant.

Towards the end of the 1980s, buyouts got bigger, liquidity got looser, and the fees and fame won by financiers fomented a debt-deal frenzy. The culmination of the era was the massive KKR bid for RJR Nabisco, famously chronicled in the book Barbarians at the Gate. By that point, the deal-making process itself had become a massive apparatus that fed itself on fees and dubious valuations. Martin Lipton, author of the poison pill and key M&A attorney of the decade, wrote of a bubble looming in 1988. Lipton’s critique frames the issue in broader societal terms:

As with tulip bulbs, South Sea bubbles, pyramid investment trusts, Floridaland, REITs, LDC loans, Texas banks and all the other financial market frenzies of the past, the denouement will be a crash. We and our children will pay a gigantic price for allowing abusive takeover tactics and boot-strap, junk-bond takeovers. We are overleveraging American companies and forcing them to focus on short-term stock market results. Research, new product development and capital investment are no longer the keys to business success. To the contrary they have become the invitation to a junk-bond, bust-up party. While the rest of the industralized world is investing for the future, we are squandering our assets in a speculative binge of junk bonds, financial futures, program trading, put and call options and the other games of today’s financial market

By the end of the decade, the boom did in fact bust. While Milken and Drexel Burnham were embroiled in legal issues (to put it mildly), the deals themselves became less and less profitable. As economists Steven Kaplan and Jeremy Stein put it, the decade concluded with “too much financing chasing too few good deals.” In one decade, buyout volume rose from $1 billion in 1980 to $60 billion in 1988 to $4 billion in 1990.3 The party was over.

Perspectives

What to make of all this? There are two basic competing perspectives of the ‘80s takeover boom. One argues that the debt-fueled takeovers constituted a healthy financial disruption of a complacent corporate America, with the end results being unlocked shareholder value and a leaner, more efficient business environment. The negative view holds that the buyouts fueled a short-term focus on flipping and asset stripping, a practice that generated profits for financiers but whose end results were socially harmful, leaving a slew of ruined companies and livelihoods in its wake.

Economists from the University of Chicago liked LBOs

Though it was published in the middle of the LBO boom in 1986, the economist Michael Jensen’s “Agency Costs of Free Cash Flows” has served as the de-facto intellectual argument for the entire era.

The debt created in a hostile takeover (or takeover defense) of a firm suffering severe agency costs of free cash flow is often not permanent. In these situations, levering the firm so highly that it cannot continue to exist in its old form generates benefits. It creates the crisis to motivate cuts in expansion programs and the sale of those divisions which are more valuable outside the firm. The pro- ceeds are used to reduce debt to a more normal or permanent level. This process results in a complete rethinking of the organi- zation’s strategy and its structure. When successful a much leaner and competitive organization results. 4

Jensen’s paper argues that having large amounts of free cash flow increases agency cost, where the firm’s agents (managers) have a incentives to use the its excess cash in a way that does not align with the interests of shareholders. An obvious way for management to misuse free cash flow is to pay itself too much money. More pernicious ways are to bloat the company beyond its optimal size. Manager prestige is associated with the size of the managed organization, so managers who want to seek prestige are served by adding layers of middle management (managing managers is the ultimate managerial flex) or pursuing value-destroying acquisitions.

Debt, according to Jensen, serves as a perfect control mechanism on the agency cost of free cash flow. An optimal level of leverage limits managerial bloat by directing excess free cash flows to paying down tax-shielded interest payments. Jensen’s paper also floats the idea of substituting dividends, which are non-enforceable, with debt, which are enforceable in bankruptcy courts. (Though he never explains how equity holders would become creditors―I guess by issuing a convertible bond or other such recapitalization. If anyone knows, I’m interested in examples of this “bond dividend”).

As an example of his theory, Jensen mentions the positive effect that debt had on the oil conglomerates. The oil industry in general had a counter-cyclical boom in the 1970s, which led to a high degree of poorly managed ventures beyond the firm’s realm of competence. Jensen claims that takeover bids from Pickens and Mesa helped limit wasteful expenses on exploration and development. Even when the takeover failed, as with Unocal, the defensive actions themselves were sufficient to spur shareholder-friendly changes.

In “The Takeover Wave of the 198Os” economists Andrei Shleifer and Robert Vishny survey the results of the takeovers from the vantage of 1990 and come to similarly positive conclusions5. Their chief argument focuses not on the cost of free cash flow specifically but of conglomeration. As firms “de-worsified” into various areas, they dulled the ability to extract value through specialization. These conglomerates ended up with large centralized “HQs” that poorly managed its sprawling subsidiaries. If an LBO takeover broke up its target and resold the units, the result is superior economic efficiency through greater specialization.

Boot-strap, bust-up, short-termism

Hostile takeovers do not occur for the benefit of the target. The goal of the acquirer is to make a profitable return on their purchase, whether the company persists or not. For insiders to a company and its ecosystem, though, companies are not only shares of paper and projections of cash flow―companies generate income, products, local taxes, a sense of community and so on. While leverage can generate outsized returns quickly, the other side of that symmetry is that disaster comes more quickly, too. Throwing debt on the balance sheet also puts a company at more risk in economic downturns. Furthermore, the “control mechanism” of debt that Jensen champions also naturally limits a company’s ability to invest in R&D and other growth assets. This is why venture capitalists trade money for equity, not debt―it would be foolish to demand interest payments on unproven ideas with no cash flow.

All of these arguments circle around the core contemporary critiques of the 80s merger mania:

• as a corporate strategy, highly leveraged takeovers sacrifice longevity for short term gains;

• in the broader context of society, hostile LBOs often result in the top-down, outsider-made decisions to eliminate a company’s departments and workforce.

The argument that mergers resulted in a speculative bubble is easy to make: takeovers did result in substantially more defaults and bankruptcies as the decade rode on. Kaplan and Stein’s research in 1993 found “Of the 41 deals in our sample put together between 1980 and 1984, we have found only one that defaulted on its debt; a default rate of just 2 percent. In stark contrast, 22 of 83 deals put together between 1985 and 1989 had (by August 1991) defaulted: a rate of almost 27 percent.” 6

Given the profitability of early deals, the access to liquidity that was free upfront and expensive to pay down, and a general mania about mergers in the markets and the newspapers, it’s likely expected that the boom ended in a number of bust outs. Warren Buffett, who unsurprisingly was never directly involved with any of the hostile takeovers of the 1980s, was thus in an unscathed position in 1990 to use its reckoning as a lesson on the margin of safety:

The disciples of debt assured us that this collapse wouldn’t happen: Huge debt, we were told, would cause operating managers to focus their efforts as never before, much as a dagger mounted on the steering wheel of a car could be expected to make its driver proceed with intensified care. We’ll acknowledge that such an attention-getter would produce a very alert driver. But another certain consequence would be a deadly - and unnecessary - accident if the car hit even the tiniest pothole or sliver of ice. The roads of business are riddled with potholes; a plan that requires dodging them all is a plan for disaster.

—Warren Buffett, 1990 Berkshire Shareholder’s Letter

Are raiders company killers?

Bubbles and financial irresponsibility are arguably an inescapable periodic feature of capitalist markets. And neither is it controversial to suggest that debt limits spending on growth. But I’ll end this discussion on the perspectives of the boom with a much trickier question: were the raiders socially harmful? Much of the answer depends on how much you, the reader, put a premium on the value of creative destruction in capital markets, and how much of that premium is afforded to financial recapitalization over basic competition and innovation. Different readers will come to different answers. It is too big a subject for me at this time.

I will however, take a brief look at some near-contemporary claims of the LBO’s virtues and vices. First, I’ll return to Shleifer and Vishny’s 1990 paper, which dismisses the most common objections to the LBO. Regarding the claim that LBOs reduce R&D, they acknowledge that the final effect is unclear, but reiterate Jensen’s claim that takeover targets need little R&D to begin with. Regarding the claim that takeovers reduce layoffs, they acknowledge that this does happen, but the impacted workers were typically white-collar folks who could find other jobs quickly. This second point is perhaps an interesting view of the times. Whether the claim is even true, I don’t find it a consoling argument that LBOs resulted in more white-collar layoffs than blue-collar ones. Perhaps 30 years ago white-collar layoffs were considered less harmful, but I don’t feel that’s the case today. And while it’s amusing to imagine the likes of Carl Icahn terrorizing fratboy CEOs7 who overpay themselves, surely many of the people in the 12 floors of people he fired were just normal workers clustered around the median income doing their best to support their families.

The other persistently popular claim is that takeover artists are asset strippers who ruin companies to unlock temporary value. The story of Houdaille, as reported by Max Holland’s When the Machine Stopped, is an illustrative example.8 Houdaille, if you remember, was the target of a friendly acquisition in one of KKR’s early levered successes. Holland reports that to service the debt, Houdaille expended its $35 million cash balance and liquidated unprofitable segments. In a year or two, enough extra cash was around that the company actually acquired a high-value maker of machine tools. However, Houdaille’s fortune reversed as a recession in 1981-1982 coincided with new competition from Japanese imports. With the combination of debt, recession, and high competition, Houdaille recapitalized and restructured, splitting off its seven divisions. At the end of the decade, Tube Investments repurchased an emaciated Houdaille with the intention to keep just one department in operation. For investors, the case is a success: interests were paid and equity holders made great return. But 2100 high-skilled manufacturing jobs were lost and a company was destroyed. Would the company have survived as a rare American manufacturer if its management hadn’t cashed out through KKR’s LBO? Who can say?

The ‘80s live on

Times have changed, and what was gauche is now debonair. Junk bond investors are now distressed debt investors, among the most sophisticated financial practitioners around. Nobody calls anyone a “raider” anymore, but the practice of acquiring shares and filing 13Ds lives on as “shareholder activism.” Almost no one looks at a corporation with sentimental community feeling, anymore, either. So takeovers feel less like raids than they do mere re-shufflings of wealthy self-interest. Private equity is a massive industry, and LBOs remain a vehicle of choice, especially when interest-rates are low. Are consumers better off? Is society? Would the world work better if more companies went public again? All these threads run off into their own stories, each worthy of a volume of books. The ‘80s live on.

Ten years ago Jan Loeb, a veteran money manager, came across Acorn Energy, a penny-stock company that was burning cash on the way to bankruptcy. In an unusual switch from investor to operator, Jan Loeb loaned Acorn money, became its CEO, stripped it down to one essential business, and methodically built the company back into a profitable, high-margin nanocap trading on the NASDAQ.

But I must emphasize nano: last year Acorn Energy had about $2 million dollars of operating income on $11 million of revenue. However, Acorn is now well-positioned for a new phase. Its business of monitoring generators and oil pipelines has a lot of tailwinds as demand for energy grows and concerns about grid instability increase. It also has a new exclusive partnership to expand services in a deal that has a lot of potential upside with very low necessary investment.

I found a lot to like about this company, which I’ll detail in this write up. Nonetheless, I find its current price―$22.5 a share and a $55 million dollar market cap―too rich even assuming many years of linear top-line growth and decent margin expansion. I bought some Acorn at ~17.50 while researching this article and the price spiked1. I won’t buy more now, but I’m not selling. And unless something terrible happens in the annual earnings report coming this week2, I’ll buy more in one of two situations: 1) the company persists in its current state and the price swings down to around $15; 2) I see signs that the company is going realize some of its latent options to boost free cash flows.

About the company and its finances

In this section I’ll give some context about what Acorn does and its present financial situation.

Monitoring stuff that can’t go down

Acorn Energy is a holding company that holds just one company, Omnimetrix. Omnimetrix makes monitoring systems for backup generators and natural gas pipelines. For the rest of this post, when I say “Acorn Energy,” I also mean Omnimetrix.

The main source of revenue and growth is monitoring emergency generators in the industrial space. Basically, when the main source of power goes out, the backup generator comes on. But what if the backup generator fails or never starts? How can power-needy business be sure their back up plan will work? By monitoring the device. That’s the Omnimetrix business: a failsafe for a failsafe.

Omnimetrix has a lot of devices, but they all work roughly the same way: users hook the device up to the control plane of what they need to monitor, and Omnimetrix grabs the data stream for remote monitoring and analysis. Besides obvious failure prevention, remote monitoring reduces the need to send technicians to manually check physical sites. One of its products in the oil-pipeline segment, RAD, goes beyond passive monitoring and also provides a mechanism to manually disconnect.

Omnimetrix has been around for a long time and despite its microsize, it has a lot of installations and repeat customers with major industries. I assume the products are good. If you want to learn more about the product, they have a lot of tutorials on Youtube.

The financial story

I’ll give a few words on how Acorn’s turnaround has played out financially. After the divestiture scramble, the company has expanded its sales team and developed new products. These costs show up in the SGA and R&D segments in the chart below.

Also note how revenue is growing faster than costs. For the last few years, the plan has finally started to pay off and operating leverage is being realized. Along with generally increasing hardware and monitoring revenue, the company built out some partner networks and closed out some bigger deals, most recently with a large telco. Acorn also has a large tax shield from the massive operating losses that happened in the pre-Loeb and turnaround years.

The business ends up with a lot of upfront cash that gets deferred, which brings a pattern of cash flow bumps preceding improvements in operating income as the deferred revenue is amortized.

A good business and competent people

There are also some qualitative factors of Acorn Energy’s business that are attractive.

Structural advantages of the business

Acorn Energy’s revenue comes in two phases. The first revenue comes from the one-time upfront payment of the hardware. Then customers pay for monitoring on a recurring basis. Gross margins are 50% for the hardware and 90% for the monitoring.

Acorn’s products target critical systems, away from areas that anyone wants to fiddle with. Once the devices are installed, they are installed. Some customers have also committed to a program to donate excess capacity in times of grid instability. Acorn would capture a bit of that income whenever such events happen.

Given that the monitored systems are analog devices, I imagine that product development requires electronic-engineering expertise, making the knowledge barrier to entry fairly high.

Capital allocators at the Helm

Besides the business structure, the other qualitative elements that I find attractive about Acorn are its three key managers: Jan Loeb, the CEO; Tracy Clifford, the CFO and COO; Harold M. Jarrett Jr., the Omnimetrix CTO. In total, insider ownership is about 35%.

Jan Loeb owns at least 21% of Acorn’s stock3, and, as a highly experienced financier, he thinks in terms of return to capital. I am reminded of Thorndike’s The Outsiders, which provides 8 case studies of how CEOs delivered outstanding longterm returns through intense focus on capital allocation. Like the outsiders from the book, Jan is a thrifty investor-owner. In earnings calls, he is realistic and laconic, and appears at all times to be focused on profitable use of debt and equity. He’s also frugal, always a good sign. For example, Acorn sub-lets the unused part of its main office for an extra $27,000 a year.

Aside: In terms of transparency, I find Acorn unusually investor friendly. Its IR is not flashy but the earnings transcripts and presentations have a lot of good information.

Tracy Clifford came on in 2018, and she has brought years of operational execution. From what I gather, besides managing the balance sheet, she has been the force that works with customers, builds out the sales team, and has ensures customer satisfaction through new product development and increasingly large customer rollouts.

Harold Jarrett founded Omnimetrix in 1998 and oversees product development. I don’t have much more info about Jarrett, only that he is obviously the key technical leader.

The inflection point

Good business, good people, etc. But the company is tiny and the price is already by my estimates too high4. If there’s not a new growth story, I don’t think Acorn Energy is a great deal for shareholders. That leads me to the current state of the energy world and the new partnership deal.

Have you heard? LLMs use a lot of energy. And the capex buildout is going to use even more. Electric cars and high weather events put more pressure on the grid. Taxing the grid causes more need for backup generators. Grid instability causes more need for generator monitoring. While grid instability is the main growth story I’ve seen discussed, natural gas pipelines are also being built across the US and it’s possible Acorn captures some revenue there too.

Besides that, Acorn just signed a big partnership with AIO, a company that provides remote monitoring for other non-generator assets. AIO operates outside the US and has 110,000 installations. In 1H of 2026, Acorn is going to find some integrators and sales people to market and implement the product within the US. AIO and Acorn will share 50% of the revenue from hardware and sales. If AIO penetrates the US as well as it did other parts of the world, Acorn will capture a huge upside.

Risks

I’m surprised I’ve devoted so much time to a company with less than 30 people and ~$10 million in revenue, but I like the business and think its growth is more likely than not. However, there are plenty of risks:

• Bus factor. 3 key people in a small company means there could be a big loss of knowledge if one or more of them becomes “unavailable.”

• A tiny company with competitors selling to massive industries is vulnerable. I wonder how much power Acorn has to protect itself from aggressive buyers who want to squash its margins. Besides a mention of an 8-10% price increase from post-Covid inflation, I don’t recall seeing any mention of sales prices in the reports or earnings transcripts.

• Reversals of economic situation could cause a large revision in power demand and buildout, shrinking the new-opportunity set by a large degree.

• Generator OEMs and monitoring vendors may introduce rival products that carve into their potential customer base. This has already happened in the residential space, where Generac has locked out Omnimetrix with its in-house monitoring5. Who knows? An open-source solution might make customers all prefer DiY. That said, this risk has existed since Omnimetrix existed, and I’m guessing that working with bespoke analog systems isn’t quite the same as making a connector for data emitted over some software protocol.

• Management may fail to meet new growth opportunities. The current story has been of disciplined, incremental execution. While management has shown itself quite capable there, a different set of skills may be required to capture rapidly expanding opportunity. There is new operational complexity that comes with AIO’s expanded product line.

Monitoring the situation

As I mentioned in the intro, the current share price of 22.5 is too high for me to buy now, but I’m not mad that I bought some a bit cheaper. I’ll be following Acorn and might invest more if I think the growth inflection is plausibly being realized.

The company

Bloomia Holding Co. a.k.a Lendway a.k.a Insignia systems is a company with a lot of history and no history. Through the nineties and early two-thousands, it was part of an ads-in-grocery-stores oligopoly, trundling along in this niche all the way until 2020, when activist investors from Air T, a freight company, gained controlling interest, sold its ads business to a rival, and briefly pivoted to specialty agricultural financing before it ousted two CEOs and made a levered purchase of Bloomia, a European company whose entire business is devoted to selling a single flower. Yes, I’m pitching literal Dutch tulips.

What’s the interest in a highly levered, borderline profitable, low-margin, historic meme business? Through the fog of restructurings, recapitalizations, and calendar-year resets, there’s a business whose books could get cleaned up rapidly. And the rights offering provides a way to cheaply bet on the upside.

Rights offering

On January 23rd, 2026 Bloomia filed a preliminary prospectus announcing a rights offering to raise $15.5 million to take advantage of favorable debt payment windows. On February 6, 2026, Bloomia released a press release with more details: each owner of 1 share receives the right to purchase 2.16 shares at a price of $4.05. Today, at market close on Feb 10, 2026, the price is at $4.49. The offering closes on February 16th, and the rights won’t trade on the secondary, so Bloomia gave shareholders less than a month to get in on the deal.

The offering is expected to expire on March 28th, “subject to early termination.” The stock is illiquid and volatile, so there’s absolutely no guarantee that the rights will be in the money on exercise day. However, rights holders lose nothing by holding out to see. And if the price does look favorable and the exerciser wants to sell right away, the risk comes in the anxious period between exercising the offering and receiving the results. If the price does not look favorable, management should subscribe anyway, and the company has a good chance of rerating from its low price.

It’s not exactly shooting fish in a barrel, but there are some favorable elements of both the rights offering and the business that could cause the market to boost its expectations:

• Outsized debt reduction. In the original sale, the sellers agreed to defer some of the pay as a bridge loan of $12,750,275. Bloomia has negotiated to pre-pay the loan at the discount of $7,330,000. The rest of the proceeds go to pay off $6,600,000 of related-notes payable, which as far as I can tell, were given by Air T related Air T persons. I’ll say a bit more on the self-dealing, incestuous nature of the notes-payable portion in the next bullet, but paying ~$13.9 million to reduce $19.35 million of debt and free up $2.2 million in annual interest payments is probably efficient capital allocation.

• Management is going to participate. Co-CEO Dan Philp said in a Jan 23 press release “The immediate impact of the use of proceeds from the rights offering is such a glowing positive for the Company’s future that the Company’s largest stockholders, plus the corporate Executive team, will plan to enthusiastically participate in the offering, including in portions of oversubscription opportunities that may be available.” The notes payable parties also may participate, though really they’d just convert their debt to equity1―how generous!

• If the spread is favorable, rights holders have a good chance to oversubscribe for extra profit. This ship of Theseus company has been public since 19912. I cannot imagine that all shareholders actively follow the latest developments. I can imagine that some of these offerings are going to estate proceeding.

• Bloomia’s public financial history makes the company likely look worse than it is. To be clear: the reports are a mess. Earnings have been devoured by expenses related to the acquisition, and even though it’s been the sole focus of its parent since 2024, we’ve never gotten a proper year of financial reporting―they filed a 10-KT reporting 6 months of earnings so that they could align their fiscal year with the tulip season. I find that annoying, but most tulip sales happen from February through June. Through boosted sales and debt reduction, the top and bottom-line likely both go up in the next few quarters. As Joel Greenblatt likes to point out, high debt can magnify small gains in value. Unfortunately, the 10-Q that will come during the rights offering won’t include the big sale season, so it will likely have a negative or neutral effect on the market price around exercise time (see update). But over a horizon of a year or more, high risk may bring high reward.

Maybe in a later post, I’ll get into Bloomia’s valuation, but my current strategy depends more on observation, bet sizing, timing. For now I’ll just say that inflation-level growth, a high discount rate, and liquidating at acquisition price comes pretty close to my assessment of market expectations.

Strategy till June

1. Wait to receive my rights offerings.

2. Consider selling original shares to free-roll the offering. See update #2

3. Assess rights depending on spread:

   1. If the spread is nice, exercise and oversubscribe up to 1.5% of portfolio. Hope the price doesn’t crater while the rights exercise settles.

   2. If the spread is bad, monitor the situation. If owners backstop the offering and debt does reduce, the price might rerate nicely in spring as the 10-Q reports the results of its busy season in conjunction with the reduced debt.

      
      I do expect Bloomia’s situation to improve by the end of the year. But sales in the cut-tulip business depend on a few make-or-break days. If things don’t improve, it’ll be ugly.

This is not a fat pitch. I’m betting on asymmetric probabilities and a situation that is too small for normal investors to pay attention to. I’m also learning as I go. Chance of loss must be accounted for in position sizing.

Updates

A running log of updates.

Update 1, 10-Q from Feb 13, 2026

As all significant sales happen between Feb 14 and June, I expected the February 10-Q to be bad. In fact, it is worse than I expected: to meet current financing obligations, Bloomia had to use its $10 million dollar credit revolver. Paid-in-kind interest is swelling, and it received a waiver for breaching the liquidity covenants of its loan.

Inventories are piling up. Bloomia claims that it has “strategically accelerated the growing of stems to meet spring demand, which led to less stems available for sale in the beginning of the year to cover fixed costs such as rent, which reduced margin year over year.”

I now consider it unlikely that Bloomia’s stock price rises to a point above 4.05 to make worthwhile to exercise or over-subscribe. However, if management follows its stated goals and participates and oversubscribes, a lot of financial tension might be reduced in the coming quarter as it simultaneously de-levers and passes through its busy sales season. We’ll see how strategic the accelerated production was.

Update 2, 2026-02-26

Received warrants and sold existing shares over two weeks for minuscule profit (~1-2%).

Update 3, 2026-03-26

The rights offering expires today below the strike price. Management put out an Investor Presentation a week ago, with 4/6 slides being about the rights offering. I doubt many investors will buy it.

That said, management is in an interesting situation here. The rights offering was to prepay a significant loan at a large discount. The terms of the loan are that it must be paid by April 15, 2026.

If management over-subscribes for the rights offering and buys all shares, they’ll:

• Retire debt at a discount

• Avoid vertiginous piling of PIK interest onto an already large debt stack

• Dilute all other shareholders by about 66%

• Be in the best position enjoy a large re-rating if the company sells a lot of tulips in the period from Valentine’s Day to Mother’s Day.

Basically all the company’s sales happen this quarter and next. The company’s outlook for the entire year is bound to a pretty narrow set of parameters (sales and delevering).

In the prospectus, management and inside creditors stated that they intend to participate but are not commited to participating. If they aren’t willing to pay a $0.10 premium to subscribe for all shares, what is the fate of this company?

Update 4, 2026-03-27

Bloomia extended the offering to April 1st.

Update 5, 2026-04-03

In a Press Release, Bloomia announced the offering raised $12.1 million, about $5 million from equity purchases and, as expected, $7 million in converted debt. The note about the debt conversion indicates Air-T insiders participated.

I expect to see a flurry of filings about insider purchases soon. Whether or not insiders feel confident about this critical sales quarter, they at least indicated cleaning up the balance sheet was necessary enough to pay a slight premium on shares to dilute others.

Next earnings should be reported in Mid-May. This won’t include Mother’s Day sales, so there’s at least one more high-volume sales day that won’t be reported until around August.

Current price is $3.98. I’m taking a small position when market opens in three days.

Update 6, insider purchases and less debt

I’ve taken a look at the Form 4s reporting new insider purchases and the form 13s reporting insider purchases. 4 insiders purchased, but besides director Nick Swensen, whom I’ll discuss more in a bit, only Co-CEO Jundt’s exercise was meaningful. He purchased nearly 24,000 shares, which gives him 3% of the company for the price of more than half an annual salary.

Nick Swensen’s case is more interesting. He is a director at Bloomia and the CEO of Air T. Swensen purchased $2.8 million dollars worth of shares to give him 19%. However, the form 13D that reports holdings of a group that includes Swensen, Air T, and his affiliated investing firms puts the total group ownership at 60.1%―up from 39.1% before. Besides that, a 13G for BCCM advisors gives this group another 10%. While that group is apparently passive, it’s interesting to note that Swensen is on the board of the directors.

In summary:

• Insider ownership is very concentrated

• All insider participated, though Air T really converted a bond for a company it already owned 40% of.

• At least $13 million of debt has been retired, saving $2 million a year in interest payments.

The float of actual non-insider shares is tiny, about $1.5 million.

Current heavy sales season is underway. The next two 10-Qs make all the difference.

2025-05-13

Valetines Day leaves equity owners heartbroken. Latest 10-Q shows an underwater operation, more debt added from revolver. I sold my small position in pre-market, I expect the price to crater. Essentially no money was won or lost, but the time cost was too much for a dubious business. I’m tired of tulips.

At the moment, The LGL Group is barely a company at all—it’s a pile of a cash, a board of directors, and a tiny precision-instruments manufacturer. To put the imbalance between its operational scale and cash balance in perspective, the latest 10-Q reports nine-months’ cash from operating activities of $443 thousand and a net cash balance of $41.5 million.

Nevertheless, I’ve been following LGL as a “net-net” with a “special situation.” At the end of 2025 I built up a position of LGL warrants and I exercised them with over-subscription privileges. I found a favorable margin of safety at the exercise price, which was less than the current share price, which was less than the net cash per share. I also trusted that management would, at the very least, not simply set the cash on fire.

My original plan was to sell once the share price approached net cash per share. Following the warrant exercises, I calculate LGL Group’s current cash holdings to be about $7.23 a share (waiting for the 10-K to know the exact number). Its share price has jumped from $5.80 to $7.05 in a few weeks. However, as the share price nears my original sale target, LGL hired a new CEO with an unusual compensation structure.

I’ve decided to hang on and see how things progress.

Update, 2026-03-31. LGL released its annual report, and its current share price of $7.01 trades at about a 4% discount of its net current assets, which are worth $7.23 /share (essentially all cash). The compensation structure for the new CEO, Jason Lamb, has changed, from a $60k/year with bonuses for deal-making to $160k/year + a capped $60k bonus for action. A press release from March 24th indicates the company has a new partner, Legion Capital, which will look for defense investment and partnerships.

This company’s value is basically the bag of cash plus the optionality of whatever action Jason Lamb brings. The initial CEO comp structure was unusually weighted for deal-making; the new compensation is less punishing of executive inaction. Lamb seems qualified (former US Seal and defense private equity), but I’m considering taking profits from the warrant exercise and oversubscription and moving on to something else.

Interesting developments since the warrant

The exercise of the warrants came with a flurry of interesting new activity.

First, they appointed a new CEO, Jason Lamb. Lamb has a long background in the naval field, first as a US Navy Seal and then in the defense industry. The old CEO, Marc Gabelli, remains chairman of the board. Just on its own, I consider this a favorable development. Marc Gabelli has his hand in many pots. His father is the billionaire value investor Mario Gabelli, and Marc is involved in various businesses in the sprawling Gabelli family operation. Since LGL has been on my radar, I questioned the focus that Marc could bring to the micro-stakes realm of LGL. A dedicated CEO brings new attention to deploy capital.

More interesting is Jason Lamb’s compensation structure, laid out in an 8-K/A from January 5th, 2026. I expect that Jason Lamb could make more than 60 thousand dollars a year elsewhere, so all his incentive comes from deal-making and performance:

The material terms of Mr. Lamb’s compensation are summarized below:

(i) an annual draw of $60,000 against incentive compensation (the “Base Draw”);

(ii) eligibility to receive incentive compensation for identifying, analyzing and managing investment opportunities according to standard payout rates then in effect to the extent it exceeds the Base Draw plus expenses; and

(iii) eligibility to receive a discretionary annual bonus based on individual performance and performance of the Company.

Given Lamb’s background and the Gabelli group’s long interest in defense and manufacturing, the LGL Group’s obvious focus is going to be on M&A in the naval space. I’m not inferring anything: at the Sidoti Micro Cap conference on January 22, 2026, Gabelli and Lamb specifically mentioned that they’d be looking for defense-related deals in the mid-Atlantic region.

Besides the appointment of Lamb, there’s been insider buying. Other members of management―in particular Gabelli― all exercised their over-subscription privileges as well. Managements interests and shareholder interests thus overlap.

Finally, the company paused the only deal it’s had in the works for the last year, and I consider that a good thing. The company had planned to acquire the Morgan Holding Group, another Gabelli—affilliated financial services firm which, despite having over a billion in AUM, had a net income of (yet again) less than 1 million dollars. I was not enthused by this prospect or its purported synergies and thought its finalization would constitute a rather pathetic catalyst. Now that the deal has been indefinitely paused, LGL can focus on deploying capital in areas that likely bring better returns with less opacity.

Risks

The most severe longterm risk is that Lamb and LGL make terrible purchases and set fire to all their assets. While I don’t think that’s likely―for all its byzantine schemes, the Gabelli group has a long history of positive capital returns in this space—bad deals certainly can happen. But even in such a case I’d expect upward price movement before the results of a bad deal manifest in a few years. Given the bellicose timbre of current world affairs, the market probably presently puts a premium on any news related to US defense. If a deal happens and the price re-rates, I’ll re-evaluate. If my whole thesis of the “defense premium” falls apart, perhaps a more peaceful world could be my minor solace for a loss of shareholder value.

Outsiders also suffer from information asymmetry about LGL’s deal flow in general and Jason Lamb’s compensation structure in particular. The 8-K mentions compensation and bonuses could come from “identifying” and “analyzing”, verbs that encompass a wide range of activities that don’t necessarily result in increased cash flows. The good news is that management also benefits from shareholder value, so they probably won’t write a fat check just for reading Wikipedia.

A less severe risk is that LGL group reverts to inactivity, plodding along as it pays its board and collects T-Bill interest while the share price stumbles down. But I think action is more probable than inaction this year, and even if nothing happens, the share price is buttressed by the cash position.

The market price undervalues management incentives

I believe the recent increases in share price mainly reflect the results of the warrant exercises. There is no longer uncertainty about its diluting effects, and it was accompanied by relatively large insider purchases. However, I also believe the price still undervalues the upside potential of Jason Lamb at the helm. I do not think I have any special contrarian insight, either; the company is just too small and weird to be under heavy market scrutiny.

If Lamb closes one or more acquisitions, I expect the market will actively re-price LGL. If nothing happens, the company still sits narrowly below its net current asset value. This is to say I weight its upside as less than its downside and am holding.

Yet, the stock is trading at 3 times earnings and 1.4 times free-cash flow. Greek-owned shipping companies are not exactly market darlings, and I wondered whether the low price was exaggerated.

Mr. Market evaluating Greek Shipping companies:

On further analysis, I decided that, while the expected value I derive is positive, the industry and shareholder arrangement is too complex for me. I don’t see a sufficient upside that warrants me looking further into this.

Financial history

This small multiple of charts tell the basic financial history of Dynagas. In 2014 and 2015, the company expanded its fleet from 3 to 6 ships. This major capital expenditure walloped free cash flow, which has otherwise gently trended upward. These ship purchases were financed by mostly debt. While EBITDA margins remained high in a market downturn of 2018 and 2019, earnings to common turned negative as debt was serviced. Since that low point, the company has steadily de-levered but the stock price has been flat and earnings have been flat and wobbly, no doubt depressed in part by the major dilution that happened in 2016.

You can inspect each chart individually in the appendix.

In the Shareholder’s favor

Dynagas has spent years paying down debts and freeing up cash flow. It pays out a dividend and has authorized 10 Million dollars for buybacks (at today’s price, this could reduce shares outstanding by ~8%).

The company seems competently operated: management has consistently secured longterm deals while navigating much geopolitical and economic uncertainty. Countries are likely to continue importing natural gas, and LNG shipping is apparently one of the least price-sensitive components of the natural gas industry. Jerry Kalogiratos, CEO of a different Greek LNG shipping company, recently called the industry a “floating pipeline.”

Most of the fleet is equipped for Ice transport, and if Russia ends its war with Ukraine, Dynagas may have an edge for artic contracts.

Besides all that, it’s just a cheap, ignored stock. Even with very pessimistic assumptions, it’s not hard build a model that prices it at 150 to 180 percent of its current value. Nevertheless, I decided there were too many risks in the story.

Risks

While debt is being paid down, the ships are aging. Replacing them surely would require more debt or more share issues. LNG scrapping is at a record high, which I assume reduces the price Dynagas can get for scrapping its old steam-powered ships.

As old ships get scrapped, new ships are getting built at a record rate. There are 324 orders for LNG ships through the rest of this decade. Dynagas in particular has some unlucky timing: in 2028 it will need to negotiate new charters right when the shipping supply glut is going to pressure charter rates. And debt payments will be due.

There are other risks inherent to the industry. The geopolitical temperature of the world is rising and can have unpredictable effects on the shipping and energy trades. Many other outside risks exists. On the extreme end, a ship could blow up. Even all six could blow up!

In a recent report, management mentioned they might pursue other growth opportunities. That signals that they themselves don’t feel great about going all in on LNG shipping.

What I didn’t investigate

I decided that it wasn’t worth it for me to untangle these questions. A more enterprising investor would want to investigate:

• The company’s ownership. It is a layercake of common and preferred stock. Earnings to common units is less than I expected.

• The structure and refinancing possibilities of its debt

• The complex engineering and operational questions that come with running and staffing an LNG shipping company.

Not swinging at this pitch

I am too ignorant for this stock. The business has too many variables that I don’t understand. While I think it’s likely undervalued at current prices, all the upsides that I imagine are gentle and all the downsides I imagine are severe.

I’m sure being unfashionable and complicated means that the LNG shipping stocks probably have some great deals, though. I may look into other companies.

Appendix

• A much more enthusiastic report from the industry (September 2025)

• A recent chat with LNG Shipping CEOs (not Dynagas)

Click to pop up the graphs:

For the past 5 years, Doximity has consistently generated a lot of cash at high margins. Yet though it consistently beat earnings estimates in 2025, the stock rose and fell wildly through the year. Today, on January 1st, 2026, at a price of $44.28, its shares sit near their 52-week low. Nevertheless, that’s still too expensive for me.

In the short term, the price bakes in a rate of growth that I find improbable. Longer term, the business structure does have attractive features (niche adtech money geyser), but its modus operandi carries a touch of the unsavory.

What’s to like

Over the course of the year, the Doximity share price took a few dramatic tumbles. Dramatic movements often have an element of overreaction. And on the surface, I found some enticing arguments for a variant view.

On the measurable side, it has:

• Increased free cash flow for every year for the last 5 years, with nearly linear accompanying EPS growth.

• Fat margins

• 5x current assets to liabilities

On the qualitative side, it has a decade of experience building a network in a valuable niche, and management (led by its CEO–founder) has succeeded to print money in a competitive field. Its engineering culture is apparently good, with small autonomous teams and a clear sense of direction.

On its users

The business claims that it counts 80% of US physicians as members, and it has been operating for more than a decade. What exactly that 80% means is ambiguous, but they must have a lot of data to mine for a valuable marketing demographic.

Surely far fewer than 80% of its verified members are active. Of those active members, I suspect only a minority heavily engage its flagship “LinkedIn for doctors” service. But I also suspect that these minority powerusers are a very monetizable group of MDs. Scuttlebutt says the comment section is full of conspiracies and cultural declamations―that’s a good sign. Surely, being a credulous, emotional, and paranoid member of an exclusively licensed group lands a person in a valuable marketing demographic.

From what I can tell, the majority of active users use the app only to mask their number (to make it look like they’re calling from the hospital while using a non-hospital phone). It probably has gained some traction with its AI scribe and HIPPA-compliant ChatGPT wrapper, too.

Maybe just having the app installed on a physician’s phone is a sufficiently valuable source of user data on its own. I don’t know.

What’s opaque

More on what I don’t know.

First of all, the financial statements themselves leave a lot for the reader to guess about where revenue comes from. They’re also sitting on a pile of liquid assets and it’s unclear how they’ll allocate them to meet the high bar of growth expectations.

From its implementation side, an outsider has no way of knowing how well Doximity uses the data it has. Adtech is a business full of murky measurements and potentially spurious event correlations. The skill of the data engineering and analytics teams make a big difference here.

The Doximity customers likely have their own measures of how effective the ads have been; customer retention is a proxy stat there, but they’ve changed the grain of how they measure this across reports, making direct comparison impossible. Lauren Balik has a detailed article this and other metrics, Doximity (NYSE: DOCS): Growth is Malignant, Not Benign.

What’s against them

In the short-term, pharmaceutical manufacturers are pulling back on their marketing budget. They have a big competitor in OpenEvidence, and the rivalry has already become embarrassingly lurid and litigious1. Generally the market is less enthused about tele-health than it used to be, though I don’t know if that’s a rational reason for the price drag.

In the long term, their market share can get chipped away from many angles.

There is a hard, government-licensed limit to their userbase. I don’t think that’s that big a problem if they can keep people active, but there’s a ceiling.

Finally the business itself has some dubious tactics in how it games engagement, membership, and reported revenue. Are dubious tactics a sign of strategic vitality in the adtech world? Maybe, but I’ll sit out.

Conclusion

I would not buy unless the price gets pummeled or the business shows itself to be more than what it currently is. I also wouldn’t be surprised if the price doubles. How nice that I don’t have to do anything.

Subscribe now