Trading at 21x LTM (pre-COVID) EV/EBITDA and 33x FCF, the market views Bright Horizons Family Solutions (BFAM), the child-care center operator, as a secular growth story largely untouched by the current maelstrom. Prima facie, this seems justified, as the business has put up high-single-digit revenue/low-double-digit EBIT growth CAGRs over the last decade; sports impressive returns on invested capital; and didn’t miss a beat during the financial crisis (EBITDA flattened out but never fell). But in my view the market has it all wrong. The reality is that BFAM’s margins haven’t grown in 5 years – despite continuing to raise tuition fees, improving utilization each year, and above-trend growth in higher-margin segments like back-up care. On closer inspection, BFAM has actually been a serial closer, not opener, of new centers – speaking to a much less attractive blue-sky runway than is currently perceived. Today, with enrollment at all-time highs heading into a recession (or depression), and with the balance sheet levered (2.5x funded/3.5x included rents), at the very least the ‘acquire, restructure and improve’ growth story is over. In the worst case, the company could need to be recapitalized (since it is burning $50-$100mm/month in the current status quo, is already too levered, and faces huge operational deleverage in a worse economic environment). When the dust settles, I think 2021 EPS could be ~$1.75/share, or 60% below current consensus, and assuming just modest multiple compression, the stock could fall 65% even if the more draconian balance sheet restructuring scenario does not come to pass.
After a business summary/quick history, the investment thesis goes something like this:
- Even pre-COVID, the growth story was running out of puff and misunderstood by the market;
- COVID itself creates meaningful risks for the balance sheet;
- Potential for large earnings deleverage in a normalized post-COVID environment;
- Perspectives on valuation post this crisis and key risks.
BFAM is a child-care center operator, maintaining 1084 centers across North America and Europe. Back in 2011, the company operated 760 centers; since then, growth has been mostly driven by rolling-up other providers. Child-care centers are generally good businesses, driven by long-term structural drivers (more urban density; more women/dual-parents in the workforce; more employer recognition of the benefits of company-sponsored child-care) as well as attractive underlying economics at scale (low initial and maintenance capital requirements, and high incremental margins). Partially offsetting these attractions, the market remains extremely fragmented, with the top 10 providers constituting <10% of the center-based market, and ~90% of the market operated by groups with <10 total centers.
The company has three segments: full-service, back-up, and educational. Full-service – 82% of LTM revenues/62% of LTM EBIT – encompasses, as you might expect, full-service child-care centers, offering care Monday-Friday, 7am-6pm, that vary widely in size and scope, but average around 126 children (at capacity) in the US and 82 children in Europe. About 30% of the centers in this segment are operated on a ‘cost-plus’ model, where a corporation pays for the build out and maintenance of a center (on behalf of its employees), and BFAM earns a management fee that supplements the fees it charges that company’s employees (which are subsidized by the employer). The other 70% of the centers are operated under a ‘profit and loss’ model, where BFAM undertakes all the costs of the operation, and earns all the rewards associated therein. The majority of P&L centers still have a corporate sponsor(s) associated with it, but these centers also serve third-party, unaffiliated families (often from locations at group office parks, etc). Personnel costs constitute the majority of operating costs in both models – around 70% in total blended, with a higher proportion for the cost-plus model and a lower proportion for the P&L model (where BFAM incurs a wide variety of other costs). Rent is generally the next highest discrete cost (at ~8-10% of COGS); with the other cost items (20-25% of COGS) comprising cleaning, maintenance, electricity, other overhead, etc.
BFAM guides that gross margins in the segment range from 20-25%, with the higher number coming from P&L centers, and the lower end from cost-plus centers. As you can imagine, most all the cost line-items are fixed and so incremental returns on higher child enrollment are high. Returns on invested capital in both models are also attractive: clearly BFAM is not committing much capital in the cost-plus centers; and even for the P&L centers, fit-out costs (for fixtures, etc) are very low relative to operating costs (basically all personnel, rent, and facilities maintenance). Low/mid-20s gross margins translate into ~10% EBIT margins on a segment basis, with a large portion of company SG&A (~11% of company-wide revs) – and basically all of amortization – allocated to this segment.
Back-up care – 14% of LTM revenues, but 30% of LTM EBIT – constitutes the provision of ‘last-minute’ outsourced care and support for children (and the elderly). The reason margins are so much higher in this segment (mid-20s EBIT margin vs 10% for full-service) is that BFAM is essentially a platform provider of third-party care providers (through its call center and online service), taking a fee both from employers to organize and maintain the service and also from parents on a per-use basis. In my view, back-up care is the real ‘jewel’ of the business – it is just much smaller in absolute earnings power than the full-service segment.
There is no denying that BFAM has executed impressively over the last 10+ years. In 2009, BFAM was doing 21% gross/8% EBIT margins on an $850mm revenue base; in 2019, BFAM printed 25.4% gross/13% EBIT margins on $2.06bn of total revenue. This represents a revenue/gross profit/EBIT CAGR of 8.4%/10.2%/13.6%, respectively. The growth algorithm has been quite simple:
- Raise fees: average tuition was $1435/month in 2012 (the earliest disclosure I could find); today it is $1825/month, representing a 3.5% CAGR in this time period. The company guides to ongoing tuition hikes of 3-4% annually;
- Increase enrollment: there is no doubt that enrollment/utilization was sub-scale a decade ago. Whilst disclosures are spotty, BFAM has hiked enrollment by 7%, 12%, 5%, and 4%, respectively, over the last four years, collectively amounting to 31% growth in enrollment over this period. The last time a negative enrollment period was disclosed was 2010 (-6.4%);
- Acquire, improve, and close: acquisitions form the core of the BFAM model. BFAM operates 1084 centers today, having grown from 760 at end-2011 – but in that time they have acquired a total of 510 centers. In other words, whilst total center growth has compounded at 4% in this period, net center growth has been negative1%, as the company systematically acquires, improves, and subsequently closes underperforming centers. Whilst this juices enrollment growth (consolidating children into fewer, better-performing centers), it also creates challenges going forward, as we shall discuss, given the company has enjoyed many years of optimization of its center base at this point;
- Growth in adjacent businesses: the education/advisory business has been a strong secular grower, from a non-existent base (revenue/EBIT went from $8mm/$0mm in 2009 to $82mm/$21mm in 2019). This has helped filip the broader entity as other segments have stagnated in earnings power, but for the purposes of this thesis still remain too small to move the needle (<8% of group EBIT today).
Running out of steam before COVID?
Despite the historical success of these various drivers, what’s most interesting is how the core segments (full-service and back-up) have stalled out – in earnings power terms, not revenue terms – in the last few years. That is, between 2009-2015, Full-Service EBIT margins expanded from 6.6% to 9.8%, as the top-line compounded at 7.3%. Similarly, Back-up EBIT margins grew in tandem, from 19% to 31.3%, alongside a revenue CAGR of 10.2%. But from 2015-19, margin growth has stuttered: Full-service margins in 2019 were 9.9% (despite ongoing revenue expansion of 6.4%); and Back-up margins have gone into reverse, falling 4+ pts to 27% – despite further topline growth CAGR of 10.2%. Why, then, did margin expansion stop despite ongoing revenue growth?
Firstly, Bain acquired the business at the margin low-point in 2009 (coming out of the GFC), and no doubt extracted significant operational and cost synergies during its ownership (the business was relisted in 2013 and Bain fully exited in 2016). Many of the most attractive and synergistic acquisition targets also appear to have been monetized during this period (underperforming chains in London and Holland, for example). The retrospective, cynical view on private-equity ownership suggests this pattern – where margins expand during the PE hold period and then a fall-off thereafter – is not uncommon. But there are other reasons as well: coming out of the recession labor was far more available and at lower rates; this made it easier to hire (and keep) personnel, juicing margins in both full-service and back-up. In recent years, however, a tight employment market has made controlling personnel costs much tougher.
It is also likely that BFAM spent the initial period coming out of the GFC back-filling the most attractive areas and locations – dense urban areas in prosperous areas of the richest cities in the US and Europe – that by 2016 or so were largely penetrated. Thus, growth in more recent years appears to have been in less dense, and perhaps less wealthy, locales. In addition, as previously mentioned, the child-care market is highly fragmented, and capital requirements for single locations are not high. It seems likely competition for labor (and desirable rent locations) has contributed to the margin stagnation: since 2015, COGS have outgrown full-service segment revenues (7% CAGR vs 6.3% CAGR) and other expenses (SG&A) have grown even faster still (8.3% CAGR).
Near-term COVID impact is about the balance sheet…
COVID changes the equation for BFAM in many ways, both near- and medium-term. The company has said little about the impact thus far, other than closing half its centers worldwide, pulling 2020 guidance, and announcing cost-cutting measures. I expect at a minimum, the balance sheet to come under severe strain over the next few months.
With half the centers closed and most of the world on lockdown, I think it’s fair to assume enrollment even at those centers that remain open will be extremely low for the extent of the initial wave. Within the full-service segment, I assume half of the centers generate zero revenue (since they are closed); and half see attendance -75%; back-up care revenues I have estimated fall 50% (since most everyone is working from home, last-minute child care needs must necessarily come down); meanwhile I assume the Education segment revenue cadence is unchanged (generous). Even so I get revenues on a monthly basis of ~$38mm (vs 2019 LTM runrate of $173mm company-wide). that I make no adjustment for the likelihood that many of BFAM’s most profitable centers – those in the NYC surroundings, for example – are quite likely to be the most affected, and thus most damaging to BFAM’s PnL, given the particular spread of the virus.
On the cost side, if BFAM can furlough 50% of its staff – ie, all those who work at the closed centers – at no cost, and if personnel runs to 65% of COGS in aggregate, then monthly labor cost falls to $42mm. I assume rent still gets paid – this is ~$13mm a month. Then for all other COGS I assume a blanket cut of 25%. This suggests total monthly COGS of $79mm – or -$40mm of gross profit per month. SG&A ex amortization likely comes down – I budget -25% – but even so the monthly operational burn is ~$55mm. Interest expense adds another ~$4-5mm per month.
Thus, I think its reasonable to conclude BFAM is currently burning $60mm cash per month in the current status quo. Note that I believe this is very much a minimum burn rate, because in practice it may be quite difficult to furlough/lay off staff in a frictionless manner for this kind of business. Hiring for childcare is not like hiring for McDonalds: rich families paying ~$1.5-2k per month to have their children taken care of often develop a personal relationship with their carers and expect a certain level of continuity, familiarity, and trust from their child-care provider. Even pre-COVID, safety and training regimens for new hires in child-care are rigorous and costly; it may not be simple, nor desirable, for BFAM to simply wield the axe now in the hope that business doesn’t suffer later. Similarly, I make no adjustment for the likelihood that many of BFAM’s most profitable centers – those in the NYC surroundings and in London, for example – are quite likely to be the most affected, and thus most damaging to BFAM’s PnL, given the particular shape of how the virus has spread thus far.
Beyond operational burn, working capital is a major problem. Parents normally pay for care monthly in advance, and at year end, BFAM had $191mm of deferred revenue (basically one month’s revenue in the Full-Service segment) and also ran negative working capital outside of deferred revenue due to favorable trading terms elsewhere in the business. Even if we assume suppliers exercise forbearance, the simple draining of deferred revenue from the balance sheet – which will be accomplished with most all the centers in lockdown for a couple of months – necessitates another $100-200mm of funding to be found, on top of the ongoing burn rate.
At this point it’s worth mentioning that BFAM is quite levered: they carried $1bn of funded net debt at year end (about 2.5x LTM EBITDA of $395mm); on a rent-adjusted basis this leverage is a bit more threatening at 3.5x. If we assume, then, the lockdown costs the business $120mm cash in operational burn (so two lost months at minimum burn rates) and another $130mm of lost deferred revenue, pro-forma leverage – against pre-COVID earnings levels – rises to 3.2x/4.1x (funded/including rents).
Clearly this isn’t levered to the gills, but it is enough to constrain future acquisitions for quite a while, at the very least. And, since the company only carries $27mm cash (and has $225mm only of revolver capacity available), this may prompt liquidity question marks (or worse) in a more bearish scenario.
Post-COVID earnings power looks structurally lower
But the main reason I like this short is NOT the near-term balance sheet impact of the temporary lockdown; it’s because I think the mid/longer-term earnings power of the business will be significantly impaired from current levels, even if the longer-term reaction to COVID from a behavioral perspective is fairly muted. That is to say, even if we all go back to work in 12 months, and even if we can all agree that child-care is a necessary and valuable service, I can’t see how pre-COVID enrollment/utilization levels (let alone fees) are maintained through a recession (or, heaven forbid, a depression). And the PnL looks supremely vulnerable to even small changes in both of these items. (Note: this of course pre-supposes social distancing is done and dusted and normal child-care centers can be used similarly to before the crisis – perhaps a generous assumption).
The fact that this recession is different, as regards to its impact on BFAM, is something the market appears to be missing, precisely because last recession the business seemed so bulletproof. In 2008-2010, revenues kept growing (albeit at a lower rate), and EBITDA didn’t decline – not bad at all for the GFC! No doubt the market is today sanguine about the rebound potential here because the business was so resilient back then. But the market is missing that the starting point for the earnings power of the business today – mid-teens EBIT margins, 19% EBITDA margins – is a full 5+pts higher now than back then, and that this higher baseline level of earnings speaks to much higher – and thus more vulnerable – operating leverage in the model (that is, 10-15pts higher enrollment, and 35%+ higher fees in nominal terms). Furthermore, it ignores the very specific, COVID-related residual risks that are likely to crimp earnings potential even in a fully-normalized, post-COVID world.
Consider the unit economics today. Based on 10-K disclosures, the average Full-Service center does $1.65mm of annual revenue (blending US and Europe), and generates an average 23.5% gross margin (using blended averages for P&L/cost-plus models). This equates to ~$388k in gross profit per center – or, at company-wide segment EBIT margins (10% in 2019), around $39k in EBIT per average center. Since the average center – again, blending US and Europe – can hold 111 children at capacity, and since a simple average of fees on a monthly basis is $1825/month/child, at those average revenue levels the implied utilization rate is around 70%. Actually, we know utilization is higher than this since younger children who garner fatter fees – infants and toddlers – should be a smaller component of total enrollment. Adjusting this assumed utilization rate slightly higher to account for this, to 75% (probably still too low based upon external research) suggests that gross profit per child per month is around $390. Keep in mind that average fees in 2012 were $1435 per child per month; they are now $1825 per child per month. That is to say – the entirety of the Full-Service center gross profit being earned today is simply the sum of fee increases pushed through during the last seven years.
That aside, what could a post-COVID environment look like? Well, I think it’s reasonable to assume at least some parents will be permanently scarred by COVID such that they fear for their childrens’ health in a child-care center. There will also be at least some structural trend towards work from home – likely alleviating some of the demand for out-of-home child-care at the margin. Of course, it also seems pretty clear we will go through either a recession or depression – none of which is good for household incomes and expensive line-items like $2k/month child-care. Finally, BFAM is over-indexed to large urban agglomerations where the virus has been particularly virulent – NYC area, SF, and London – and where the majority of its’ highest-performing centers are located (judging by average fees and density) – again, likely a significant negative during the new normal to come. Add it all up and I think a 10% decline in enrollment, along with some fee compression, is more than generous for a normalized base case, say from 2021.
Assuming a 10pt decline in utilization to 65%, then, and say a 5% decline in average fees, and I think average unit revenue falls to ~$1.5mm (vs $1.65mm today). On the cost side, personnel costs are probably largely fixed in numeric terms (you can’t necessarily fire 10% of your employees if you have 10% less kids), but maybe you can cut salaries in line with lower fees, so let’s lower the personnel cost by 5%; meanwhile other COGS – rent, facilities – are basically fixed if not higher (since we will definitely see stricter maintenance and cleaning requirements for all public spaces post-COVID). In total, I have gross profits per unit dropping to $291k, and gross margins fall to 19.3% – or around 420bps below where they are today. And in reality I think enrollment and fees could well fall further than this…
Whilst no doubt SG&A costs can be cut somewhat in this kind of reversal, the very fact that historically gross margin improvement was mostly responsible for EBIT margin improvement suggests in reverse the same should be true (looking at the group PnL, gross margins went from 21% in 2009 to 25.4% last year as EBIT margins went from 8% to 13%). Thus, a 400bps+ contraction in gross margins – based off just a 10pt enrollment drop and giving back not even one year’s worth of price hikes (since average tuition falls just to $1733/month vs $1715/month in 2018) suggests to me that Full-Service segment EBIT margins could realistically fall from current levels near 10% back near levels seen in 2010-11, or 6-7%.
But what about the Back-up segment? Whilst clearly harder to forecast as all the drivers are not readily observable, it further stands to reason that in a much weaker employment/recessionary environment, and one in which some portion of the workforce is out of work and/or working from home, there would be less demand for high-priced, last-minute child-care services. Still, it is hard to handicap how much lower demand translates into lower revenue, so I am simply assuming a minimal 10% decline here in nominal revenues, along with slightly higher decremental margins, for my 2021 estimates. I am not touching the Education segment.
What’s BFAM worth in the new normal?
Putting it all together, in even this somewhat mild downside scenario, I think group EBIT could fall to ~$180mm (vs $268mm in FY19) and EBITDA is closer to $300mm (vs $395mm in FY19); keeping tax and interest expense relatively constant implies FY21E EPS of $1.75/share – 60% below the current consensus of ~$4.25 (which is basically flat vs realized FY19 and may be stale at this point). At the same time, pro-forma leverage jumps to 3.8x funded/4.7x including rents (assuming deferred revenue normalizes back in the company’s favor and just the $120mm of cash burnt during two months of lockdown). Remember again this assumes no further burn/deterioration in the cap structure over the rest of 2020, and (I believe) a mild recessionary downside scenario in 2021…so I think reality could certainly look worse than this.
It is tough to estimate what the market may capitalize this kind of lower earnings profile at going forward. No doubt there will be some residual support for the name, as well as assumed bounce-back potential if/when the economy returns to growth (perhaps late 2021 or early 2022). Still, I find it hard to believe the name would not at least partially re-rate, especially as in this scenario the company would be breaking the historical pattern (namely, seeing strongly negative comps in this crisis versus the last), would probably trip covenants, and may at least have questions posed about its leverage burden. Thus, even a high residual multiple of say 20x EV/EBIT would imply a much lower stock price around $42, or 65% downside. Alternately, even at 20x EV/EBITDA – basically the current multiple, and where Bain paid to take out BFAM back in 2009 – the stock would be worth only $80 in this scenario, still good for substantial downside before considering derating potential. And to be clear I fully expect the balance sheet to be in much worse shape by the time we get to 2021, adding further pressure to residual equity value.
Upside case, other risks
Nothing in life or investing is without risk, and of course that applies here too. There is a chance BFAM could sail through this recession – perhaps the economy bounces back rapidly, and completely; perhaps BFAM is able to offset lower enrollments and/or lower fees with aggressive cost cuts and other efficiencies. More likely though is perhaps a higher multiple applied to lower or similar FY19 earnings (thanks, Jerome Powell!) and – in a fast recovery scenario – an unchanged perception of the business as a strong secular growth story. In that scenario I could see the stock push back up to recent valuation highs around 28x EV/EBITDA and 45x FCF – which, even with the added cash burn through COVID, would imply a stock north of $160 and up near the recent highs, so ~33% downside from here.
The other major risk is another PE-led takeout (given the history here). However, I am less concerned about this risk since Bain paid ‘only’ 20x EV/EBITDA based on a much less profitable and less-scaled business – ie, around the current pre-COVID multiple – and the absolute ticket size on a takeout is much larger now than then. Also, the business is now levered significantly (again a legacy of the original buyout) so PE can’t rerun the same levered playbook here even if the equity component derates significantly. Mostly, though, COVID, social distancing, and work from home have the potential to fundamentally change the calculus for a business like this over the long-term, such that a buyout near-term – before the dust settles and we can at least glimpse what the new normal looks like – seems remote.
Disclosure: short BFAM