IWG PLC IWG LN
March 12, 2023 - 5:08pm EST by
darthtrader
2023 2024
Price: 184.20 EPS 3.4 9
Shares Out. (in M): 1,007 P/E 54.1 20.4
Market Cap (in $M): 2,235 P/FCF 0 0
Net Debt (in $M): 720 EBIT 0 0
TEV (in $M): 2,955 TEV/EBIT 0 0

Sign up for free guest access to view investment idea with a 45 days delay.

Description

Brief Summary 

The bull case around IWG is a classic Hedge Fund-type story – a “misunderstood” story (due to complex accounting and a business transformation that’s being executed), tremendous valuation upside on “normalized” assumptions, numerous “free options” being given to you by the market, potential catalysts such as the sale of a business to “crystallize upside”, and “smart” holders such as an owner/operator and a hedge fund with what looks to be like a concentrated position figuring highly on the shareholder register and likely motivated to effect change. 

In terms of putting rough numbers onto that, the bare bones of it are that the digital business alone, which represents a negligible percentage of group sales and revenue currently, is potentially worth about 80% of the market cap according to recent press speculation around potential PE buyers. On top of this, shareholders would own a leading portfolio of flexible offices including about 3,400 locations, of which probably upwards of 2,400 are “mature”, meaning they’ve been in the possession of IWG for several years and are not in the “ramp up” stage. The mature estate will do about £2.1bn of revenue this year at a gross margin of around 17%; pre-covid, mature did about 22% gross margins (the trough was 13% last year and there should be some further recovery beyond 2022), while in the peak year of 2013 it was 29% gross margin. The company are currently in the early stages of trying to franchise the good name of IWG (they have been able to do a relatively small deals in Asia thus far), which, if successful, will transfer a significant part of the cost base onto a third party and leave IWG with a far higher margin royalty stream likely to be valued by the market on a more generous multiple than the one the company trades on. The franchising, if successful, would also transfer the currently enormous (relative to the market cap) operating lease liability off the company’s balance sheet, which currently looks to me like a significant cause of discussion among investors. 

Assuming that they are able to transfer the majority of the current business into franchised business by the end of the decade, then applying what I think is a reasonable multiple to that revenue stream, and including the valuation of The Instant Group touted in the press, the SOTP value, discount back to today, would be around 430p (+140%), I think, while I get to an even higher valuation on a DCF basis. I am aware of the dangers of valuing a business like this on the basis of earnings many years out – it’s not necessarily a valuation I’d be willing to lean into too much, I’m just illustrating the upside. 

I think that the key bear argument is simply that these valuations are a fantasy and just analysts with too much time on their hands playing around with spreadsheets without really understanding the business (hence the “magical thinking” title). The argument would be something along the lines of “they paid ~£250m for The Instant Group last year – how does combining it with their own modest assets and separating out the financials magically create £1.25bn of value?”. And with respect to the franchising model, my modelling of it suggests that the equity would be worth around £9bn by 2030 – I think that a bear would rightly question how franchising a pretty mediocre brand is going to magically create value out of thin air between now and the end of the decade. 

Some of the other bear arguments would include: 

·        Monotonically declining core profitability

·        Significant off balance sheet liabilities that the company continue to suggest don’t exist

·        Earnings disappointments that underscore the difficulty in forecasting the business 

Business Description 

They are the global leader in flexible working spaces with 4x the locations of their next-largest competitor (they have 3,314 locations as of 31st December 2021 and 8m+ people using their network) and 20 brands covering high end to low end. Brands include: 

·        Regus

·        Spaces

·        HQ

·        Signature 

Peers include WeWork, Servcorp, Breather and Knotel, however nearly 75% of the market as of 2020 was “Other”, i.e. small local/regional players that are probably not profitable (Deskmag suggest that only 42% of coworking spaces globally are profitable). The market has been growing at a rapid clip and the data from industry consultants is that the number of coworking spaces has grown from ~17k in 2018 to around 24k in 2021, with that number expected to rocket up to 42k by 2024, while Transparency Research argued in 2020 that the market could be worth $32bn by 2030. Investec wrote in a note that they expect 30% of the global office space market to be flexible space by 2030 (I believe the company said it was a low single digit percentage currently). 

In terms of the reason the market is growing quickly, the company claim that there are a few main reasons: 

·        Covid has been a proof of concept that flexible workspaces can work as part of a flexible solution including WFH, working in a satellite office, and working in a (presumably smaller) head office

·        Workers want to do it due to the cost/commute time saving

·        Companies need to accommodate that and are willing to do so as it saves them costs and also bolsters their ESG credentials 

The company claim that employees surveyed at 90% of FTSE 100 and FTSE 250 companies would look at leaving for the competition if there wasn’t a flexible workspace option, while they reference another research report claiming that half of workers would resign if they weren’t allowed to WFH at least part of the time. While I have my doubts about this (it’s like everyone I know saying they never use facebook despite Meta having 3bn MAU’s), I know from personal experience that getting everyone to come back into the office five days per week is pretty much impossible, so I do think it’s plausible that hybrid working will continue to gain traction. 

The financial benefits for employees are obvious – lower travel costs, less need to pay high rents in city centres so they can be close to the office, and so on. 

In terms of the other side of the equation, employers seem happy to accommodate this. IWG claim that there’s a significant financial benefit to employers, pointing to a study from Global Workplace Analytics claiming that the average company saves $11k per employee from flexible workspace as they end up with a smaller real estate footprint and benefit from reduced absenteeism and increased productivity (I don’t agree on the increased productivity). Aside from the savings in lease costs, the cost benefits to companies include facilities management, security, reception staff and insurance. Colliers, the real estate company, estimated the saving to companies at 16% of the cost of traditional ownership, while WeWork rather optimistically put it at 57%. 

Employers also get the ESG bonus of hybrid working in that reduced commutes mean reduced carbon footprint, which is an important consideration currently. While a full-time WFH would realise the full benefits of these savings, there’s concern that company culture and employee development benefits are lost if people work entirely from home, thus some kind of hybrid solution with employees perhaps WFH a couple of days, working in a regional satellite office a couple of days, and spending a day in the head office seems like it could be the way things develop. Davy note that larger corporates, using the examples of Standard Chartered, NTT, Shell, Nestle and EY, have all entered into partnerships with IWG in recent years, and that these contracts are larger, more profitable, and stickier than the smaller deals. 

IWG estimate that about 2/3 of their locations and 50% of their revenues are already outside of major cities, which makes them well-positioned to benefit from the trend of people being less willing to endure long commutes but not keen to work from home with the distractions that entails. In terms of tenant mix, they used to be geared towards SMEs and startups when it was Regus, but per one of the other VIC writeups, in WeWork’s prospectus they said 43% of revenues were from larger corporations, and IWG have said their exposure is larger, so another bull argument is that this time around, if there’s a deep recession of the bursting of the tech bubble continues, they might be a bit less exposed now relative to history. 

The company is led by Mark Dixon, who founded Regus Group in Brussels in 1989. He is a bit of a Quixotic character, having left school at 16 (despite “being very academically gifted”, as he modestly claimed in an interview from a couple of years ago that I listened to) to set up a sandwich delivery business, before pivoting to encyclopedia sales and bar work in St Tropez among other things, before finally settling on offices in Belgium. While he comes across as slightly pompous in interviews, his net worth is significantly higher than mine, so I can’t be too critical of him - he still owns ~29% of the company, so interests of management and shareholders are arguably aligned. 

Strategy 

The focus is on growing the network in a capital-light way, for instance through franchising, partnering, and so on. They like this as it’s capital light, involving a fee structure, no capex spend, and no lease liabilities. Presently conventional leases are about 65% of the portfolio (that was at 2021 year end and target was to get it to 50% by the end of 2022) but they think that the optimal percentage over the longer term is 10%. Over 2021 they added 146 new centres at a cost of £142.5m. They think it’s important to add scale quickly and efficiently as customers are increasingly looking for a global network of spaces, so that will be their moat. They also believe that the lower upfront cost allows them to add higher quality locations and keeps the sellers/franchise partners with skin in the game. In terms of the franchising, Davy points out that prior to their note, the company had sold about 7% of their sales for well over £400m, at multiples of 3x sales and 14x-16x EV/EBITDA. In the case of Japan, they sold 130 centres to TKP with £94m of revenues and £21m of EBITDA for £320m, i.e. 3.4x sales, 16x EBITDA (though I think that these assets were subsequently sold on at a lower valuation!). 

The other focus is then on developing and deploying their digital assets. Aside from the various hybrid working brands, they then have various digital assets that they transferred into a business to be merged with and operated by The Instant Group. Will be led by TIG CEO Tim Rodber and comprises a net cash investment of £270m to acquire TIG’s shares and provide capital for growth, with TIG management investing an additional £50m. Creates a leading platform for booking, inventory management and so on – it is the largest digital platform accessing 30,000 buildings in 175 countries. Next step will be a formal separation and listing, either in the UK or US. 

Finally, there will also be a great deal of focus on costs, with management suggesting there will be significant savings, and Davy suggest that quite a lot of that will come from rent, with landlords willing to cede ground to long-term operators like IWG, especially with large corporates pulling back office space spend (HSBC announced 40% cut in office space, Lloyds 20%, KMPG and Standard Chartered similar). They embarked on a cost optimization program at the onset of COVID which they claim has delivered run rate cost savings of £324m, o/w £129m was reinvested into new centre investment. 

Financials 

Income Statement 

At a high level, the company did £2.2bn of sales last year at a gross margin of about 10.8% for gross profit of £241m; SG&A was then £295m, and some other accounting adjustments drove a statutory operating income of -£87m and a net loss of £205m, with the large delta mainly being the imputed interest on the operating leases, which stood at £6.1bn in 2021. The financials were heavily impacted by covid, but frankly the financials are confusing, which is probably one of the key sources of opportunity for the bulls. 

If we look at the reported numbers for 2019, the last year before covid impact, revenues were ~£2.7bn, gross profit £415m, SG&A was £281m and operating income was £288m. This is on an IFRS 16 basis, not the IAS 17 basis, which seems to be the basis on which the company talk and the basis on which consensus numbers are compiled, and explains the vast difference between my forecasts and what you’ll see on Bloomberg. I have just done the forecasts on an IFRS 16 basis as I’m not an accountant and IFRS 16 (i.e. accounting for the operating leases) is the standard that the experts recommend.

Looking at 2022, I am expecting ~£2.5bn of revenues, £372m gross profit and an operating profit of £62m. Again, my numbers are different from consensus, which I think is mainly as a result the IAS/IFRS discrepancy, but also the way I model the franchising business, which I will go into in more detail below.

  

I have pasted some historical numbers here for context, just to give a sense of what a mean reversion-type scenario could look like. I also have those numbers going back just over 10 years to give more historical context, if anybody wants them. You will also notice that there is a “divisional” split – the company break out revenues and gross profit according to newly-opened centres, ones that they are closing down or selling for various reasons, and then finally the rump, which is the mature business, and the part that the company thinks gives a true sense of the economics. I’ll make a couple of observations on the divisional split. 

First, the economics of the newly-opened centres are poor. The reason for this is, I think, to do with fixed cost absorption. I’ve have attempted to estimate the unit economics of the mature and newly-opened centres. For the mature centres, the company give you occupancy rates, contribution margin, and they also provide details in the footnotes about rental costs. Using some assumptions on the split of the estate between new and mature centres, it’s possible to estimate rough unit economics for a centre with sales of 100:

  

From here, If I then assume that the newly-opened centres are open for exactly half of the year and thus incur half of the revenue and costs, it’s possible to make a tentative estimate as to what occupancy must be for the part of the year the centres are open, with occupancy a plug figure to make the derived contribution margin equal to the stated margin:

  

The other observation that I’d make is that, in the mature business, gross margins have been steadily declining. This is not merely a hangover from covid, it was happening well before then. This is, I think, partly as a result of declining occupancy:

  

Here I include my estimate of operating margin for the mature estate as well, where I pro forma the SG&A to the mature estate according to my estimate of the share of mature locations in the overall estate. These numbers aren’t given by management, it’s just my best estimate using the disclosure that they give.

  

While occupancy has been a headwind, it seems as if pricing has also been a factor – in the table below I take the mature revenue and divide by the number of centres to try to give a rough “ARPU”, and then strip out the effect of changes in occupancy and fx impacts to get to what underlying pricing plus other (with other, for instance, being mix effect) must be:

  

Again, we see that pricing looks like it might have been a negative, with declining occupancy also having an impact. In terms of what has driven that, Berenberg speculate that they’ve spent most of the last decade competing in the market with a lot of competition from irrational business models (WeWork, for instance), which has impacted occupancy, pricing, and so on. One potential bull argument from here could be that with rates higher and almost free capital now in the rear view mirror, that should present a tailwind for them. In my model, I only estimate gross margins in the mature business ending up at 22%, which is higher than the 13.7% they did last year and the 17% they’ll  probably do this year, but well below the 29% they did in the peak year of 2012. 

Balance Sheet, Cash Flow & Returns 

The company has historically generated limited free cash flow, which I think in the early part of the decade was driven by the heavy investment in estate expansion, which generated a capex burden and then represented a headwind to CFO due to the dilutive effect of a large number of new centres continually entering the mix (I have estimates of the split of the estate by nature and new, if anybody’s interested); when this effect began to fade, occupancy rates then began to weaken and the company rarely generated attractive FCF on a levered basis, though it did on an unlevered basis (which is another topic):

 

 

The reason for the large discrepancy is the existence of operating leases. When calculating the unlevered FCF, I’m adding back the rental expenses which are disclosed in the footnotes to the various annual reports, and then I’m subtracting out the value of the tax shield that the rental payments provide. The company obviously only had to begin disclosing the PV of the operating leases from the 2019 FY accounts onwards, but I have gone back and made the adjustment for the years 2011-2018 using the disclosure that they used to give on non-cancellable lease commitments. The proper accounting treatment would then be to go back and adjust the income statement and cash flow statement, but  it is a rather involved process, so to make life easy for myself, I just added back the amounts to the income statement to give a true sense of the returns.  

Even after adjusting for the larger balance sheet, the company has generated decent returns on capital over time:

  

Of course everyone has their own calculation for ROIC and mine it not especially robust, but I think that this is probably one of the interesting bull points on the company for a patient investor. The discussion that I read in the sell side reports is quite focused on monthly EBITDA, whether the IAS 17 net debt will be £600m or £700m at the FY stage, and so on. While this is obviously important given the debt facility that they will need to refinance in the not-too-distant future, It’s quite interesting to note that the company has historically beaten its cost of capital by about 500bps, on average (excluding occupancy getting torpedoed over covid), and yet the company trades on “just” 1.2x EV/IC, so there is some element of time arbitrage if one can believe that they will get the earnings trajectory back on the track they were on pre-covid. 

The other observation on FCF/capex/invested capital that I’d make is that they did pursue a very aggressive expansion strategy. In the table below, I lay out the expansion and closure schedule, together with my estimates of the split of estate between mature and new:

  

The reason I wanted to highlight this is simply to demonstrate the magnitude of the estate expansion (1,200  locations to 3,300 locations in a decade) and to lay the table for a discussion of one of the key value drivers from here, which will be the attempt to move to a franchise-based model. 

Capital-Light Transformation 

After covid put a stop to the company’s searing pace of estate expansion (they argue that flexible office space is a game of scale winning – for instance if Unliever has a British sales executive based in South-East Asia who’s travelling for work frequently, it’s important for them to be able to pick an office space that’s suitable for them regardless of whether they’re going to be in Bangkok, Bali or Bolton), Mark Dixon decided that they way forward for the company will be to “franchise out” their currently-leased space. The franchising model has a few advantages: 

·        Costs such as rent, maintenance, reception staff and so on leave their income statement (at least partially)

·        A fully cost-encumbered leased revenue stream is replaced by a higher margin franchise fee

·        Operating leases associated with the estate will presumably leave the balance sheet when the franchise agreement is made

·        The market has traditionally been willing to pay higher multiples for franchise income streams. Mark Dixon rather optimistically referred in one of the older annual reports to McDonald’s as a franchising model he’d like to emulate. While I’m not sure Regus quite has the brand cache that McDonald’s enjoys, I daresay he is looking at the 22x EBIT multiple the combined McDonald’s Group enjoys (and that is not all franchised revenue) and dreaming of being (even) rich(er) 

While the disclosure is currently a little confusing, to me at least, they told us in the 2021 annual report that presently conventional leases are about 65% of the portfolio (with the aim to get that to 50% by the end of 2022) but they think that the optimal percentage over the longer term is 10%. They also believe that the lower upfront cost allows them to add higher quality locations and keeps the sellers/franchise partners with skin in the game. In terms of the franchising, Davy point out that prior to their note, the company had sold about 7% of their sales for well over £400m, at multiples of 3x sales and 14x-16x EV/EBITDA. In the case of Japan, they sold 130 centres to TKP with £94m of revenues and £21m of EBITDA for £320m, i.e. 3.4x sales, 16x EBITDA. 

The disclosure is a bit confusing, as the company reported “System-Wide Revenue” in 2021 of ~£2.5bn but revenue from their estate of ~£2.2bn. They have given disclosure in the past that the franchise fee would be about 10%, so to make the numbers here work, I think that “owned” revenues must be about 88% of that System-Wide Revenue (SWR), with the balance being franchised revenues where they take a 9% fee. I am possibly overlooking something here, so this is a topic I will look to clear up with the management. 

The company have roughly guided that upfront sales price for a centre would be about 2x revenue, with a 9% franchise fee taken and IWG shouldering about 50% of the SG&A, which would be about 8.6% of the sales of a mature centre according to my estimates (see notes in cell N182 of the drive sheet). My estimate of the economics for a franchisee are in my model, which unfortunately I can’t include here, but I include a screenshot here:

 

  

 

The long and short of it is that I think it works for a franchisee at a 2x sales multiple – I can get to a double digit ROIC and IRR at that multiple, so this is what I use. Yes, they have done a few deals at higher multiples, but it could have been them cherry-picking good assets, and anyway a broker correctly points out an agency problem in that buyers would prefer to buy out of city centre locations with lower revenues and hence a lower fee. Whereas IWG would prefer the latter. Eventually IWG will probably need to concede some ground on price, particularly given there are probably landlords with half-empty buildings at the moment who are likely to be aggressive on pricing just to get occupancy up and spread their fixed costs. 

I can get to the franchise model being attractive if they can execute it. I’ve done some scenario analysis in my model (again, can’t seem to post it as an attachment, unfortunately), but  I’ve contemplated a scenario where they are successful in franchising, and my 88% plug figure for implied % of revenues from owned centres (so 12% currently from franchised) goes down to 50% franchised by 2030 and 75% franchised by 2035 (calculations in rows 76-77), while in a “Failed Franchising” scenario, I assume there that the percentage franchised doesn’t change. It’s a really simplified example, as the amount they are able to franchise will hugely impact the debt levels, and in this scenario analysis I have simply assumed that they grow the estate at the same rate regardless, with the cash needed if they cannot successfully franchise coming from raising debt – in reality I doubt they would do that – more likely they would just slow the pace of expansion, which is a refinement I will eventually make to my model.

 

If I assume that the franchising is a success, at a high level I get to the following economics split up by activity:  

This is driven by the following estate expansion trajectory:

  

I then lay out the development of the franchise business below, and I then include the impact on the group in terms of the change in operating lease liability: