Last year we published a short Constellation update when they released their 2022 results. As it has now been a bit over a year since our original Extensive Research Report, we wanted to provide a more comprehensive of an update to our Constellation Software report. While we will update various numbers and provide some color on their latest spin-off, the general thesis and business model of Constellation requires no updating. If you are new to Constellation Software you should definitely start by reading our Extensive Research Report, which despite being released in September 2022, is still the best place for investors to learn about the company.

Revenue.

Constellation’s revenues have increased from $5.1bn in 2021 to $7.9bn LTM. Top line revenue figures are increasingly misleading though as they consolidate revenue from both Topicus and Lumine, despite there being a meaningful portion of each company they don’t own (which is recorded as a non-controlling interest, or NCI). It gets complicated backing out the portion of revenue attributable to the NCI as Constellation’s ownership changes and the subsidiaries themselves embark on acquisitions which they may not wholly own.

Just for reference though, LTM revenues for Topicus and Lumine are $1.15bn and $425mn, respectively. Constellation has approximately a 60% interest in Topicus and roughly the same in Lumine if they exercise their preferred shares (we will dive more into the Lumine holding later). The NCI’s of just these two spin-offs represent about 8% of Constellation’s total reported consolidated revenue. Even backing that out though, we still see considerable revenue growth.

While topline revenue figures not only include consolidated revenue from partially owned subsidiaries, it also includes revenue growth from acquisitions. Constellation separately reports organic revenue growth though, which allows us to see how their existing businesses have been doing.

If you recall from the original report, the maintenance revenue figure is the most important. These are the fees that Constellation charges “post-delivery” and also includes SaaS revenues. This is the best single barometer of health for their VMS businesses because maintenance revenues are recurring and thus any decrease is generally due to churn. This metric also provides the best insight into customer churn since they stopped publishing those stats. As shown below, maintenance revenues have been increasing mid-single digits for the past several quarters. 

If you recall from our original report, we originally assumed 0 to 2% organic revenue growth in our reverse DCF. This is because organic revenue growth since 2016 has been just ~2%. One of the original benefits of the Topicus transaction was originally to see if they could learn from them in order to stoke their organic revenue growth. It looks like they could have had some success with organic revenue trending a bit higher than historically.

Free Cash Flow.

A more telling financial metric is Constellation Software’s “Free Cash Flow Available to Shareholders” metric (referred to as “FCFA2S” hereafter). Notably, they back out the FCF attributable to non-controlling interests in this metric, so it is a more accurate metric of FCF Constellation Shareholders own. Here is some background on this metric from our Extensive Research Report:

In 2018, Leonard decided to move to a new metric that he calls Free cash flow available to shareholders (FCFA2S), noting that he feels more comfortable starting with an IFRS metric (cash from operating activities) and reducing the amount than starting with net income and adding to it. Usually, we would meet management adjusted figures with trepidation, but Leonard’s calculations are largely a fair representation of owner’s earnings.

The one adjustment we would want to make, though, is adding back the revaluation change in the IRGA liability. We go into more detail as to the nature of this liability in the Topicus section, but essentially Constellation is obligated to buy a portion of Topicus at a price determined by the IRGA (Investor Rights and Governance Agreement) and the change in that price runs through the income statement. It is added back to get operating cash flow, but then Constellation backs it out again to get FCFA2S. They do this as it technically could be cash that Constellation will have to come up with at some point, and they adjust their NCI downward as if the transaction occurred.

We would want to add the revaluation charge back because it isn’t a real recurring operating cost (i.e. they would only have to adjust the liability once to continue to enjoy the higher earnings that are associated with it). If we are backing out the revaluation charge though, we would also have to adjust the NCI. Admittedly, this is all rather confusing and the ultimate impact won’t change the analysis, so for simplicity we have opted to just use the FCFA2S figure in our ROIC calculation below. If we had done this adjustment for 2021, then FCF would be about 10% higher. We do recalculate invested capital, though, at an amount higher than what Constellation shows (changes noted below).

The IRGA adjustment has continued to be about a ~10% drag on FCF. FCFA2S reached $1,125mn LTM, up from $883mn in 2021 and $989mn in 2020. Below we see how FCFA2S has trended overtime.

Constellation has grown FCFA2S at a 22% CAGR since 2009 and a 16% CAGR since 2018.

Above we see that CSU trades at 49x FCF, which for context compares to 39x when we published the original report in September 2022. Adjusting for the IRGA liability and LTM FCF would be closer to $1.25bn, or trading at about 44x FCF (compared to ~35x when we originally published).

Ultimately, whether that proves to be reasonable or too expensive will be dictated by their ability to continue to deploy ever greater amounts of capital at acceptable hurdle rates. (We will parse out assumptions in our Reverse DCF later).

In the past few years they have spent not only all of their FCF on acquisitions, but with the use of some debt, they have actually deployed over 100% of FCF.

In our original piece we noted that:

“Acquisition spend as a % of FCF has recently exceeded 100%. With more large acquisitions and some debt this could become more common.”

Whereas in 2021, when they spent more than their FCF on acquisitions for the first time in recent history, it wasn’t clear how much of the elevated acquisition spend was a temporary bump from 2020 (when acquisitions as a % of FCF fell to 49%). However, after 2021 they napped a large acquisition in the Allscripts (or Altera) acquisition for ~$730mn, and the following year they went on to acquire Optimal Blue for $700mn. (Altera is a hospital software provider and Optimal Blue helps manage and originate mortgages.  Interestingly, Constellation Software had the opportunity to purchase Optimal Blue when ICE and Black Knight had to divest it in order to assuage antitrust fears so they could consummate their merger. Constellation also purchased Empower, a loan origination system, from Black Knight, but that was much smaller). This is in addition to the Lumine spin-off, in which part of that acquisition allowed them to deploy capital (via preferred shares) into the newly formed group. More on Lumine later.

As we see above, Constellation software is showing an ability to deploy ~$1.5bn of cash flow, whereas they were doing just a third of that just a few years prior. If you recall, they acquired around 100 VMS companies prior to Covid and given the average purchase price is around ~$5mn, there were concerns that acquisition scalability would be nearing a wall. Mark Leonard though has always maintained that they could do larger acquisitions, they just usually had to be willing to drop their hurdle rate a bit (adding some moderate leverage though helps offset the decreased hurdle rate). We now have a couple more years of evidence that Constellation can execute these larger acquisitions successfully. Of course, valuations could always slip away from them and there could be a period of reduced activity.

Here is a short excerpt from our report that provides more color on Mark Leonard’s thinking:

Since Constellation announced its $20 special dividend, many investors have gotten louder at their capital allocation policy, which sets a hurdle rate ~1000bps above what an individual investor is likely to get elsewhere. Leonard would defend keeping their rates high, noting that “hurdle rates are magnetic”. His fear is that you cannot just lower your ROIC for incremental acquisitions, because whatever you set your rate at is the rate that every acquisition will fall to. So, if they lower their rates, more capital is deployed, but it is not so clear that the investor is better off. The middle ground that Leonard has arrived at is that they will lower their hurdle rate only on the large VMS acquisitions and keep them the same on the small and medium ones. He notes that 40-70 large VMS companies are sold each year, but Constellation has purchased fewer than 3 of them in their history. By lowering the hurdle rate only on large VMS acquisitions, the hope is that they can effectively maintain the ROIC for all other investments and only drop it on incremental capital, obviating the “hurdle rates are magnetic” conundrum.

However, at current valuations, reinvestment rates need to be near 100% for some period of time with high hurdle rates in order for the implied discount rate to mirror the historical average stock market return. In our original report we ran a reverse DCF at the current market price to see explicitly what was priced in, sensitizing around % of FCF deployed and ROIC.

In the next section we update that analysis.

Reverse DCF.

Below we run through our Reverse DCF analysis. The excerpt below is how we explained it in our report, with updated figures.

For Constellation, we run our reverse DCF sensitizing for ROIC and % of FCF deployed. Constellation has a ~$56bn market cap or trades at 49x FCF and the idea is that an investor would only be willing to pay that because of the high rates of return that Constellation can earn on retained earnings. While of course this is true for all companies, it is especially explicit here (for other companies you can sensitize for units, stores, or customers, etc.).  These two sets of assumptions (% FCF deployed and ROIC) are used to inform Constellations cash flows, which we then discount. The reverse DCF solves for the discount rate needed to make the discounted sum of all cash flows equal to its current market price of ~$56bn or $2,625 per CNSFW (CNSFW is their unsponsored ADR that trades OTC in USD. CSU is their Toronto listed stock).

In order to make our assumptions more dynamic, rather than just assuming one ROIC and one % of FCF deployed figure for the full 40-year projection period, we adjust them based on the amount of FCF constellation generates and the amount they spend on acquisitions. Below is the table that shows our assumptions. Just looking at the top half of the table, the “Variable Set 1: % of FCF Deployed”, we see that there are three scenarios: 1) 50% FCF deployed, 2), 80% FCF deployed, and 3) 100% FCF deployed. For each scenario, the 50%, 80%, and 100% is in reference to the starting amount of FCF deployed and it drops based on how much FCF Constellation generates (that is the y-axis that starts with <$5bn). The idea is that as they generate more FCF, it will be harder to deploy it, so we drop the % deployed. If you look at the bottom half of the table, “Variable Set 2: ROIC”, the three scenarios are: A) 15%, B) 20%, and C) 25%. Again, these figures are in reference to the starting ROIC and it drops based off of the amount spent annually on acquisitions.

Below we updated our Reverse DCF for the current market price of $2,625 per CNSFW.

We show below how these values have changed from when we originally published our report. An investor may be struck by how relatively little these values have changed despite the market price increasing 75% ($2,625 from $1,500). This makes sense though if you think about what the reverse DCF is doing: it is essentially amortizing that increase over the entire life of the business. The same way how a small increase in annual returns compounded for long periods of time can have gargantuan implications for your future value, a large change in your starting value has a relatively muted impact on your CAGR when the time period is long (The explicit forecast period of our DCF is 40 years).  

Either way though, if you believe the parameters of the Reverse DCF still hold, then returns have compressed. However, an investor may now be comfortable assuming a higher organic growth rate or perhaps accepting a lower return because they are more comfortable with assuming a higher capital deployment figure.

Ultimately, it is an investors decision which scenarios they are most comfortable assuming and what prospective return they would want to demand for the risk they see inherent in the investment (see the risk section from the original report).  

Spin-Offs.

The main focus of this update is Constellation Software, but since CSU has both a stake in the spin-offs, and since shareholders were distributed Topicus and Lumine shares, it felt appropriate to provide a short update of each.

Lumine.

Lumine was a collection of businesses within the Volaris Operating Group and was spun-out in conjunction with an acquisition of a relatively large company, WideOrbit. In contrast to the Topicus spin-off, Lumine is going to be focused on a specific vertical, the Global Communications and Media industry, whereas Topicus is focused on a geography (and has over 20 verticals). Similar to Constellation they focus on “mission critical” software solutions and also intend to be acquisitive in the future.

High-level the transaction is as follows: Volaris is spinning off a small portfolio of VMS companies which are all broadly in the communications and media industry, and together are called Lumine. Lumine is then acquiring WideOrbit for $490mn. The acquisition is funded by cash, Lumine stock, and debt. After the spin-off of Lumine and acquisition of WideOrbit, Constellation Software will have an economic interest in Lumine equivalent to ~61%, existing CSU shareholders will own 25%, and the “Rollover Shareholder Group” are WideOrbit shareholders who will receive the remaining 14%. Lumine trades on the Toronto exchange under LMN and OTC with a USD denominated ADR as LMGIF.  

As noted in the revenue section, Topicus is consolidated in CSU’s results, but Lumine is not. This is because their Lumine stake is through preferred equity that has the right to convert to Lumine subordinate voting shares at a ratio of 2.43 per preferred. The 63.5mn preferred shares Constellation owns when exercised is equivalent to a ~61% interest in Lumine on a fully diluted basis. When Lumine was spun-out of Constellation, CSU shareholders received ~3 Lumine shares per CSU share (which was ~25% of Lumine on a diluted basis). On a fully diluted basis this comes out to 253mn shares outstanding (many data services quote the basic shares outstanding of 64mn, which doesn’t account for Constellation’s highly dilutive preferreds and thus leads to a very different market cap calculation). 

They note that Lumine will have three operating groups with two of those focused on communications and one focused on media. Below is how they describe the business of their communications groups:

The media operating group, which will primarily be WideOrbit, referred to as “The Acquisition”, is described as follows:

WideOrbit serves more than 5,000 stations and cable networks, processing more than $35bn in advertising revenue globally. On the acquisition, the CEO of Lumine, David Nyland commented:

In their prospectus they note that Lumine has embarked on a total of 20 acquisitions, which averaged $11.7mn each.  This of course make the WideOrbit transaction a magnitude larger. They also break out Lumine Portfolio companies’ revenue vs WideOrbit for 2021: $228mn vs $167mn (9M22 is $187mn vs $124mn, respectively). Lumine currently lists on their website 28 portfolio companies.

For the last quarter report, 3Q23, Lumine generated $131mn, which was +1% q/q. While there is possibly some seasonality in their cash collections, acquisitions and limited quarterly data of the combined entity make it hard to parse out how much. As noted in the prospectus, their business does “not generally experience significant seasonality”, so we are fine annualizing revenue, will have more to say on cash flow momentarily. Annualizing their latest revenue quarterly revenue figure suggest that the combined group is run-rating around $525mn. 

They also disclosed a similar FCFA2S calculation as Constellation. For the 3Q23 they generated $40mn versus $69mn YTD.  Calculating annualized run-rate FCF though presents significant difficulties as we do not have a long enough history to look at cash flow seasonality. While we can compare to Constellation Software, the difference between Constellation’s and Topicu’s cash flow cycles are significant enough to make us hesitant to do so (Topicus collects over >85% of their cash in 1Q versus Constellation at ~38%).

On a 9-month basis the lumpiness tends to wash out as Constellation collects 72% of annual cash flow in the first 9 months and Topicus collects 76%. Annualizing on a 9-month basis would suggest FCFA2S of ~$92mn. However, this understates the full impact of WideOrbit as it wasn’t consolidated until the middle of 1Q23, and wouldn’t include the full impact of subsequent acquisitions in 2023.

We are tempted to annualize 3Q23 for simplicity, but it could overstate free cash flows. However, looking at Constellation, on average 18% of annual cash flows are registered in 3Q, which gives us some confidence that annualizing 3Q23 (at an equal 25% per quarter weighting) could be fair. Doing so would yield us around $160mn in run-rate FCFA2S. While this could be a bit high, especially since it implies a FCF margin of 30%, our opinion is this figure is closer to true annualized revenue on a go forward basis than the $92mn figure.

The current fully diluted market cap is $5.6bn at today’s market price of $22 per LMGIF (Net debt is just $15mn now). This suggests a run-rate FCF multiple of 35x on our $160mn figure. This is a prospectus FCF multiple versus the LTM stated for Constellations and it is also an estimate. A more conservative estimate that applied a 25% FCF margin to their $525mn in revenues would suggest they trade at 42x (uses $130mn in FCF.

Topicus.

Topicus LTM revenues are ~$1.1bn and in the past 9 months, revenues grew +25% y/y. This is compares to organic revenue growth of +7% y/y. YTD they have deployed $110mn in acquisitions, which is down y/y from $150mn. We cannot read much into such a short time frame though.

FCFA2S is $155mn YTD, however, they have a highly seasonal cash collection cycle with 86% and 90% of cash from operations occurring in 1Q for 2022 and 2021, respectively. However, we have more of a quarterly history for Topicus and can see that on a 9-month basis most of this cash flow lumpiness evens-out. Applying last years seasonality (first 9 months are 76% of revenues), we get a FCFA2S of $204mn.

Above we see the NCI that reduces Topicus’s FCF from $155mn to $95mn. For those that don’t remember, this is because a large shareholder (the Joday Group) owns about 30% of Topicus and another investor, Ijssel B.V. owns a further ~7%. In short though, their ownership is not ordinary common shares but rather units which can under certain circumstances be converted to Topicus shares. Because of this, the basic shares outstanding amount is significantly less than the diluted shares (shown below).

Thus, instead of focusing on the basic shares and the FCF after the NCI is backed, we look at the fully diluted share count and total FCF to Topicus. So as long as we are looking at the total diluted share count, it is okay to not back out the NCI. (It is possible some of the NCI pertains to other acquisitions unrelated to the Topicus Coop units, but it is unlikely material).

Topicus’s USD-denominated ADR, TOITF, currently trades around $73, which on a fully diluted basis implies a market cap of $9.5bn. Given our FCFA2S valuation of $205mn, this comes out to a 46x FCF multiple which compares to CSU at 49x FCF (or 44x if you back out the IRGA liability). In the past, investors had generally put a higher multiple on Topicus under the assumption they would enjoy higher organic growth and that their smaller size means smaller acquisitions will be more impactful to the topline. Perhaps Constellation’s higher organic growth and larger recent acquisitions has negated this difference.

Conclusion.

This concludes our Constellation Software update. We will continue to cover any material updates Constellation may have in the future. Please feel free to reach out at info@speedwellresearch.com or DM us at Speedwell_LLC on Twitter for any questions, concerns, or comments. We will also cover Constellation Software in the near future on our Podcast, The Synopsis, so be sure to follow us on Apple, Spotify).

*At the time of this writing, one or more contributors to this report has a position in Constellation Software. Furthermore, accounts one or more contributors advise on may also have a position in Constellation Software. This may change without notice.