Introduction

AMN Healthcare (NYSE:AMN) is not a business that I’d typically look at given their healthcare exposure broadly, which I tend to categorically avoid (oftentimes for non-rational reasons). That said, I’ve looked at a few staffing agencies in other sectors, so I thought I’d dig in a little here.

AMN has a very interesting setup that could result in an attractive IRRs for investors over the coming years. There’s a lot of uncertainty inherent in the model, but the good fundamental news is that the business is rationally justifiable from a customer’s standpoint. I.e., It’s not a question of whether their present-day decline from travel nurse normalization is going to stop, it’s only a matter of when.

Where one determines they’ll stop fundamentally determines the implicit IRR. But there are some other attractive elements here like their language interpretation solutions segment providing some offsetting trends within their business. Ultimately, at a high-$50/share, while I personally am not totally comfortable with the risk/reward, I wouldn’t fault others who are.

Nurse & Allied (N&A): When Does It Stabilize?

AMN’s N&A segment posted a hefty year-over-year of 35% to ~$537M in Q4 sales. For context, most of their revenue here is generated from MSP contracts, where AMN handles all of the contract labor needs for that client. Over half of their nursing sales are MSP generated, and ~50% for Allied, implying that over half of N&A sales here are MSP fills. The rest of revenue is from either (1) filling positions for a client that either uses another MSP or a VMS model, or (2) subcontracting fees where they’ll subcontract some MSP vacancies to other vendors.

Unpacking the quarter, travelers on assignment declined ~22% to ~11.9K in Q4, and the average bill rate was down 12% (down 3% sequentially). Travel nursing sales were down 40% while Allied was only down 16%.

Behind this, demand didn’t decline because patient demand decreased. If we look at large hospitals like HCA Healthcare (HCA), for instance, we can see that their patient days were up 1% year-over-year (as we would expect), and Tenet (THC) posted 1% same-store admissions growth.

Part of what’s happening is that their clients are utilizing more full-time labor. HCA (HCA) reported contract labor down 20% year-over-year, implying volumes were down high-single-digits thereabouts. And Tenet (THC) reported a 62% reduction, reflecting a material reduction in volume.

Hospitals aren’t recognizing a newfound value proposition of FTEs such that they’re choosing to hire more of their applicants. Instead, hospitals are independently improving the value proposition to FTEs by increasing their compensation for nurses, thus improving the value proposition vis-a-vis temporary staffing jobs, resulting in more employees choosing the full-time route. Reflexively, as contract demand declines, so too are pay (compensation) rates, thus resulting in a worsening value proposition.

What also seems to be (less clearly) contributing to the volume declines are some MSP clients shifting to VMS. I say this because Cross Country Healthcare (CCRN) noted this trend in 2023 (suggesting the win mix is ⅔ VMS and ⅓ MSP), and provided some historical context – post-2008, the MSP model really took off, but post-COVID, it’s the VMS model that’s starting to gain more popularity. AMN doesn’t seem to be seeing the same trends, but per CCRN, it’s hard to think some isn’t happening. As such, as an MSP client of AMN’s shifts to VMS, they essentially go from filling more than half of those orders to filling likely less than 20%.

Stepping back, the big question then is where all of this ultimately stabilizes. That’s hard to figure, frankly, but here’s how I’m thinking about it. There’s ultimately a market for contract labor as hospitals naturally endure temporal supply constraints – e.g., employees on leave, employees sick, etc. – and temporal demand-driven constraints – i.e., seasonal patient demand increases. It’s more economically rational to fill those holes with temp labor than full-time labor.

To this end, why isn’t 2019 a reflection of a normalized labor mix? We can see that the market as a whole was stable-to-growing pre-COVID which would be consistent with that. And we can also see that hospitals are aiming to get back to 2019-like contract labor levels. With this in mind, AMN had ~11.9K nurses on assignment in Q4 versus ~10.5K in Q4 of 2019 (~13% higher today) and versus 9.93K in FY19.

Assuming, then, that AMN’s volumes stabilize at a level below ~10K would necessitate a few possible things. Patient demand would have to be lower, but we can see it isn’t per HCA and THC data. Or that they’ve lost market share in the interim, but this is unlikely given the industry supply (labor) constraints. Actually, I wouldn’t be surprised if they stabilize higher than 10K. For one, they’ve likely won new customers, net, since 2019. And for two, it’s not clear that industry supply (FTE plus contract) is at 2019 levels given all of the commentary on burn out, etc., increasing post-COVID, leaving a void to be backfilled.

On the bill rate side of the equation, it’s a little trickier, but we can contextualize it similarly. The year-over-year decline in bill rates was, of course, resulting from a decline in contract labor demand, but also from pricing pressure and mix shift (lower premium contracts). Throw in some operating deleverage from lower volumes and lower contribution dollars per fill, and you end with lower margins.

Bill rates will stabilize when demand stabilizes, and where that is precisely, again, who knows. They ended 2023 with bill rates at ~30% higher than 2019 (corroborated by CCRN). Per my travel volume assumptions, bill rates should decline further. However, my assumption is that it remains higher than 2019 levels as there’s been obvious inflation since. It wouldn’t make sense, for instance, to have full-time wages up above 2019 levels, but temp staffing bill rates remain at 2019 levels.

To this end, what I can find is that full-time nursing wages on average across the U.S. are about 5-6-7% higher than 2019 levels. However, BLS data here and here shows the median wage increasing to the tune of 10%, which is more in line with what I’d think is more reasonable. To this end then, with more wage growth coming in 2024, it seems to make sense to me that something like 15% higher than 2019 levels is reasonable for temp staffing bill rates.

What this means financially, if we think they’ll eventually get back to 2019 volumes at the lowest, but can sustain bill rates ~15% higher than pre-COVID territory, 2019 N&A sales of $1.42B would translate to $1.633B. This gives us wiggle room, too, as it’s also likely that volumes don’t go all the way back to 2019 levels – let’s call it $1.65B for simplicity.

Once we adjust for seasonality, they’re currently posting a margin rate close to 13% on a full-year basis (Q4 margins are seasonally lower for mix reasons) on ~$2.1B in annual run rate sales. I’m not worried about material pricing pressure given the fundamental supply limit, but bill rates declines will pressure margins via deleverage. Historically, decrementals have been in the low-double-digit range, but they should be lower than the 14.5% posted in Q4 which were exaggerated by pricing pressure and mix shifts. Net then, should sales decline $450M from lower volumes and bill rates at, say, 14% decrementals, we’d be looking at EBITDA of ~$210M thereabouts, or 12.7% margins.

They’re guiding for N&A to be down 37% in Q1 (sales of ~$520M). These are extreme seasonal declines, frankly. Q4 to Q1 is typically lower, and they expect high-single-digit sequential declines in Q2 (versus a 6% average) with most of that volume driven. Per that math, by Q2, they’d have ~10.7K nurses on assignment (~14% higher than 9.4K in Q2 2019) with bill rates about 26% higher than 2019 levels.

Ultimately, there is a range of different outcomes from here. Everyone can see that today’s volumes and bill rates are going to go lower, but it’s unclear just where they stabilize. Ultimately, however, they will stabilize – this is not a business that’s going away.

Physician & Leadership Solution (PLS): Future Looks Positive

Q4 PLS sales were up just marginally year-over-year to ~$168M. Within this, they have basically two revenue buckets. Their biggest is their “locum tenens” business, where they supply temporary physicians or specialists. In the other bucket, they have interim leadership (c-suite), permanent placement, and executive search services they provide too.

Locums revenue grew to $123M in Q4, while interim leadership was down 35% to $29M in sales, and Search revenue was down 20% to $15M. With respect to Locums, they’re filling ~10% more days than last year, implying either (1) they staffers were working longer hours, and/or (2) they had more staffers on assignment. And then we can see that their revenue per day was about 8-9% higher as well.

Part of the growth for Locums was M&A driven – in November 2023, they acquired MSDR for ~$293M, which specializes in locum tenens supply. This added ~$13M to Q4 sales (a month’s worth), so excluding it, locums posted organic sales growth of 7% with volumes down and bill rates up. Actually, they noted that throughout 2023, most of the locums growth was rate driven and less volume driven.

As for why volumes are down, it’s not clear. Again, it’s not a patient demand thing as evidenced by the aforementioned data from larger hospitals (HCA, Tenet). Yes, there was a time when things like surgeries were being deferred, but that should be behind us. My suspicion is that they’re losing some volume to a mix shift back to FTE. Consistent with this, assuming the 2019 mix was normalized, we can see that locums too grew post-COVID from similar supply constraints the nursing side experienced, so if there is indeed a shift back to those 2019 levels, that would be qualitatively consistent. And they also talked about how physician salaries are increasing.

Indeed, if we look at Q3 2023, which didn’t include much of MSDR, they posted ~46K days filled versus 42.7K in Q3 2019, or ~7.7% higher. And they’re noting that some practices are still 50% higher than pre-COVID periods. This would only be economically justifiable if patient demand concurrently increased, but since we know (per earlier) that it hasn’t versus 2019, we can implicitly infer that, versus 2019, contract labor has been put in place of full-time positions versus the mix in 2019. However, this argument would not be consistent with higher bill rates, which leads me to believe the mix shift is not happening, at least to any material extent.

It appears they’re losing some market share, however. CCRN, for instance, posted a 26% organic growth rate in their physician business in Q4. This doesn’t include the offsetting interim leadership and search headwind, but it’s still materially higher than the 7% growth AMN posted. Also, CCRN’s Mint acquisition was completed in October 2022, so that rate for them was organic. Mix is likely not perfectly comparable, but that’s a material gap in performance.

Arguing, however, that these are long-term structural competitive issues appears weak. Looking at AMN’s locum history, you’ll actually see that sales were declining between 2017 and 2019. However, it wasn’t really market share-related in a structural sense. Some of this was clear industry related declines in that the emergency room landscape plus things like hospice were changing. And then some of it was from an ERP issue – as they were converting over, their recruiters were becoming less efficient filling orders.

CCRN was posting locum sales declines as well from 2017 to 2019, although they were suggesting that their issues were internal, not market driven. But also, PLS grew from 2015 to 2017, and then from 2019 onwards (yes, COVID contributed), two data points consistent with this hypothesis.

Regarding the remainder of this segment (leadership, perm placement, and search) it’s relatively easy to explain. They’re all tied to economic activity – when times are good, businesses are hiring for growth; when times are bad, everybody is trying to hang on to their jobs while positions are cut. As hospitals have become increasingly more cost conscious throughout 2023, this has resulted in less revenue here.

Margin-wise, they posted segment gross/EBITDA margins of 33.3%/13% compared to 35%/16.7% in the prior year. Contribution margins didn’t decline, really – obviously, locums bill rates are higher and while they didn’t discuss pricing for the remainder of the segment, I don’t sense any material changes here (for search, they’re paid for the process/retained search, not the fill).

There was some operating deleverage, however. Yes, revenues are flat, but organically, they’re down given the leadership and search revenue declines. Sales are only flat because they added MSDR, but this also added costs. So, net, sales are flat and costs are higher. And then there were mix headwinds. They don’t MSDR margins, but my sense is that they’re going to be similar to their segment average, if not a smidge lower from less scale. But anyways, locums has lower margins than search and perm placement, so the mix shift to locums was a mix headwind.

Zooming out, the future here is a little tough to pin down because the data is less clear, I think. I suppose it’s possible that today’s results are supply constrained and they’re not losing market share, but another argument is that they are losing market share. And one could also argue that clients are shifting back to FTEs. Of these possible trends, it seems conservative to me to think that it’s unlikely there are any structural competitive issues, or that demand should go below 2019 levels.

Assuming days filled go back to 2019 levels, that’d imply the $123M in Locums sales would be ~$115M assuming flat bill rates. Considering that perm/search and leadership are likely macro affected today, it’s understandable why sales are lower than 2019, but I don’t see why they couldn’t get back to 2019 levels as the macro normalizes. Plus, they acquired Connectics post-2019, which increased sales here. Net, then, there’s some seasonality throughout the year – typically growing from Q1 to Q3 – but it’s not huge. So, that $115M would translate into ~$460M in sales for Locums. Right now, they’re posting interim/perm/search sales of ~$180M annually, amounting to total sales of $640M assuming this doesn’t get much worse. They posted sales of ~$240M in 2019, so if the interim and search business recovers, that’d put sales at $698M down the line.

They’re guiding for PLS sales to be ~13% higher in Q1 (~$187M) as MSDR contributes a full quarter. Implicitly, this isn’t a huge sequential decline. It’s not the locum business that they’re expecting to soften – they’re expecting growth – but their search and interim business.

Should sales finish the year around $640M on a run rate basis (versus the run rate of ~$740M implied by the Q1 guide), there’s obvious deleverage. Mix-wise, this’ll be locums behind this decline, so that should result in margins not declining as much, but assuming ~12% decrementals (consistent with the past few years), they’d go from doing $168M in sales and ~$22M in EBITDA (13% margin) today to $160M in sales and ~$20M in EBITDA (~12.5% margins). Or ~$640M in annual sales and ~$81M in EBITDA.

Technology & Workforce Solution (TWS): Hidden Asset

Q4 TWS sales were down ~16% year-over-year to ~$112M, but this needs to be broken out. Part of their business here (about 60%) consists of their language interpretation solution where hospitals can speak with their interpreters – sales here were up 18%. Another ~30% of their business is their VMS offering – software hospitals can use to manage temp labor themselves – which declined 45%.

The VMS side is fairly easy to contextualize. AMN is paid a fee for every time one of their VMS clients fulfills an order for a contract. Considering then, just how much industry volumes and bill rates have declined, they’re understandably earning materially less revenue today. So, by and large though, those trends aren’t abnormal.

What concerns me is market share. We know CCRN, who’s not a small player, started rolling out their own VMS called Intellify in 2022, so this is effectively a new entrant in the market. And furthermore, to the extent that CCRN is right in that the mix is shifting to VMS such that every year, clients who are up for contract renewal are seeing more interest in VMS, I would expect AMN to be outperforming N&A which is more levered to MSP, but that’s not the case today.

So, Intellify is probably taking some market share – I don’t think there’s a directional question about that. But I wouldn’t expect this to continue forever. Pre-COVID when competitive conditions were seemingly more normalized, AMN was growing their VMS business – see 2018 and 2019, which would disconfirm that they were structurally behind in any way.

The language business is interesting because it’s like a sneaky grower hidden behind the travel nurse noise. I don’t think their growth today is share related, but mostly market driven. And this growth came despite the cost conscious nature of hospitals today.

The reason for the sales growth is likely driven via increased exposure. And given the number of relationships they have across the U.S., this is a relatively easier engagement for them versus a pure play. Couple that with the fact that the industry is probably underpenetrated – still utilizing legacy systems of using in-house interpreters – and the growth makes sense. Indeed, they noted that they’ve only sold language services to 15% of their MSP clients, for instance. And they also talked about how there’s a good amount of per customer growth where they’ll introduce the offering in location X, then scale it out to the rest of their business.

Assuming we’re right about the travel nurse market, their VMS business will decline further, although Language should grow. They expect Q1 sales to decline 19% (to ~$110M). Assuming VMS declines ~25% per what we said earlier for N&A, that amounts to ~$6M less quarterly. In the meantime, Language can likely continue growing by ~10% annually. So, if I scale these assumptions out, quarterly VMS sales would go from $30M to $24M, and $67M of Language sales would grow to ~$89M by the end of 2026. Add in another $15M of sales for their other segments (no growth), and they’d be doing total TWS sales of $128M per quarter, or ~$510M thereabouts (versus the $440M run rate today).

They’re going to naturally pick up some operating leverage. However, there will be a negative mix shift with VMS sales declining and Language sales increasing. Stratus – their Language business – carries ~15% EBITDA margins, whereas VMS has ~70% EBITDA margins. Net, we can largely think of the EBITDA growth from VMS fully offsetting the EBITDA growth from Language, implying that today’s annual EBITDA run rate – ~$164M – is a reasonable assumption.

Valuation: Interesting

We haven’t discussed capital allocation, which is something that shouldn’t go overlooked. They’re suggesting that a run rate of ~$80M is appropriate given their pipeline of investments they want to keep making to digitize the customer/contractor experience. That makes sense, but there is some concern around M&A, which they spend a lot of money on. Based on the math for MSDR, for instance, it’s not clear to me that shareholders are getting much value. Although, to be fair, their historical acquisitions have seemingly come at tolerable IRRs.

Added all up, there are various scenarios that can play out. Per what I laid out earlier, I tend to think volumes revert back closer to 2019 levels, but there are some puts and takes we discussed to that. Assuming we do trek that path, by the end of 2026 per today’s growth rates, N&A would be posting sales/EBITDA of $1.65B/$210M, PLS would be posting $640M/$81M, and TWS would be posting ~$496M/~$164M, or ~$2.79B/~$455M on a consolidated basis.

Minus corporate expense of ~$89M, SBC of $15M (Q1 is seasonally higher), D&A of $173M, interest expense of $64M (adjusted for MSDR), and taxes at 25%, I get a net income of ~$85M. Adding back D&A and subtracting capex of $80M, that implies FCF of ~$178M.

What’s that worth? Well, that would implicitly be their trough if we’re right. Considering then that over time, patient demand will continue growing, we wouldn’t be at some elevated contract labor mix, they’re not losing material market share, and their language business will continue growing, something like 3% sales growth plus some operating leverage would support 15x FCF.

At 15x plus, call it, $250M in interim FCF between 2024 to 2025, that gets me to an adjusted market cap of ~$2.92B. Discounted back 3 years implies a fair value per share of ~$58/share ($2.19B) – that’s about even with today’s share price.

Thinking about what this assumes, however, is a bearish scenario – i.e., that a lot of today’s volumes go back to 2019 levels. It’s not clear that this is the likely outcome, in my opinion. We’re also not assuming that their leadership and search business revert to higher sales, which it historically has. And finally, we’re not assuming either that their margins improve back to pre-COVID levels. So, there is a case to be made that their present value should be higher.

Conclusion

Ultimately, considering the fact that the market isn’t going away and that it’s unlikely they’re a market share loser in any way, this provides some comfort that eventually, the market will find a stable point. It’s just a question of when.

Personally, I’d be more interested in a lower price considering potential risks around M&A, insurance-related impacts to demand, and idiosyncratic market risks – e.g., ER and hospice in 2018. However, I wouldn’t fault anyone for being bullish at these levels.