Description

Quick Investment Case

I would buy Crown Holdings with a target price of $67 (+50%). They have recently closed an acquisition which I think will be, at worst, mildly value destroying but probably about neutral, which put pressure on the shares as the investor base rotated, and subsequently disclosed they are subject to regulatory investigation in Germany over price fixing, which tanked the shares again. The likely fine was, in fact, not new information to anybody who read their last couple of 10-K’s. The selloff in the shares has provided an opportunity to buy into a fair business at a very attractive price. Using a 12x target P/E multiple, or alternatively a target EV free cash flow yield of 6.5%, I see the shares trading at $65 a couple of years from now once the market gets over its frustration at the equity story changing slightly. That would be about 50% upside. Should the stock regain its recent rating of ~13x earnings, ~5.5% EV free cash flow yield, I see upside to $80 (+80%).

They are the second-largest producer of steel/aluminium cans globally, with 20% of the market, behind Ball Corp (30%). The company is 60% DM, which is low growth (1%-2% volume, slightly negative price) but consolidated and rational (top three control 80%+ of the market and are committed to generating economic value, not growth), and 40% EM, which provides profitable revenue growth. Prior to an acquisition announced last year, the company was a levered capital return story. Market cap stood at about $8bn, with net debt at about $5.5bn, which was 4.2x EBITDA, reasonable for a leading company in a stable, mature industry with reasonable returns (company normalised ROIC averaged about 10% in the last 5 years, 11% longer-term, 9% last year). The company pre-Signode was generating about $500m of free cash flow, about $420m if you exclude minority dividends which they classify as financing cashflow. They bought back $340m of stock in 2017, and had previously averaged about $300m per annum from 2010-2014 (they did a $730m acquisition in Spain in 2014 and a $1.2bn acquisition in Mexico the following year). The story they sold to investors was stable sales and profitability, $400m-$500m of FCF available for capital returns per annum. Given where yields were at the time, most people would probably value that cashflow somewhere between $8bn-$12.5bn I guess, i.e. $60-$90/share.

Instead of continuing with the capital return story, the company instead decided to acquire Signode from Carlyle for a consideration of $3.9bn. Signode is a producer of packaging materials such as corner board, airbags, container liners, film, plastic straps, steel straps and so on. The business has sales of $2.3bn, EBITDA of $384m and EBITDA less capex of $349m. The EV/EBITDA multiple of 10.2x compares to their historical range of 5.5x-9.5x. No synergy targets were offered. I suspect that what the market dislikes about the deal is:

• No obvious business overlap between container packaging and steel cans

• Acquired business is described as “GDP plus”, which indicates that it is more cyclical than the current business

• The business was previously a division of ITW and numbers were disclosed up to 2012 – looking at revenues then vs now, clearly it has not grown at GDP plus

• Margins at the acquired business appear to be up substantially – EBITDA margin was 14.4% in 2012 at ITW; the adjusted EBITDA margin reported in 2017 was 16.7%. It is not clear how much of that is business quality improving, how much is currency impacts on top line not dropping through to EBITDA, how much is Carlyle sweating the assets

• The company have been consistent on their conference calls in recent years saying that valuations look toppy and they don’t want to participate in that. Then eight years into a mega bull market, they do a big deal at a substantial premium to their own multiple.

• Investor base had been envisioning as levered capital return story. From the conference call after the deal, the company were clear that they view this deal as “creating a platform”. I believe there has been a certain amount of frustration, and the investor base will have had to rotate

While the company did not provide details on synergies from the deal, they did provide guidance on pro forma free cash flow generation, which I think is achievable. They guide for about $590m, most of which should be this year (they get three quarters of contribution from Signode), which should rise to $725m in 2019. Briefly walking through that, and why it makes sense, last year they did $830m of PBT; tax at 20% would take out $165m, getting you to $665m, while D&A was about $245m, getting you to $910m. There was a WC inflow of $56m which I do not see them repeating this year based on recent commentary. Voluntary pension contributions of $278m last year will not repeat. The company got a deferred tax benefit last year which will not repeat, but I am calculating a normalised PBT above. So CFO on a normalised basis was $910m, less capex of $500m gets you to FCF of $410m. Signode did $165m of EBITDA less capex, interest, tax and WC outflows, of which you get 75% in 2018 so you end up at $535m without breaking a sweat, I think. In terms of where the rest comes from, there were some non-cash restructuring charges in the EBIT number that add about $20m, while there are also some miscellaneous gains from China becoming a smaller part of the mix in the still-growing Asia Pacific division ($15m-$20m), while last year’s capex of $500m was a peak year according to the company (was in the $350m range prior to some projects they did in Europe and Asia that are coming to an end), so  I think that $590m is easily doable – in fact I think they will get to $620m.

In terms of how they get to the $725m the following year from my $620m in 2018 – an extra quarter of Signode FCF gives you $40m, delivering (should be about 0.5 turns of EBITDA per year) saves $30m, growth in Asia Pacific should add another $10m-$15m, while better contribution from a new plant in the US, together with continued growth in Mexico and Brazil should add about the same amount. Capex should continue to normalise in 2019, which gets you above $725m.

Prior to the deal, the company traded on an EV FCF yield of 5%. Since then, obviously yields have risen, as has the leverage in the company and the cyclicality, but if for the sake of argument, I add 1% to the target EV FCF yield to reflect all of this, then I get a conservative estimate of FV at $54. That is the $725m of equity FCF guided (actually I am higher at $790m to reflect strong volume growth across the portfolio and some cost savings from better cost absorption in Europe as well) plus the $190m of interest payments less the value of the tax shield for EV FCF of $915m. At a 6% target yield, that would be $15.2bn of EV, less 2019 YE net debt of $8.1bn gets me to $7.1bn of equity value or ~$54/share, 20% upside. The margin of safety there is that:

• It implies an equity FCF yield of >9% and a P/E of 9x – ok these are on a levered entity but most of the business is stable and they can de-lever quite efficiently – in a couple of years they will be back at the high end of the leverage range they have employed over time

• I am quite confident that the FCF they have guided for is conservative

• I believe that they can continue to grow FCF in the out years. Pricing will not be positive, but volumes can continue to grow, they have stated explicitly that Signode does not require WC investment to grow, while capex will still be elevated relative to D&A

I also struggle with the rationale behind the Signode deal, but at worst I think that the deal destroys maybe $1/share of value. If capex is a good proxy for Signode’s D&A, EBIT was about $350m last year, i.e. NOPAT of about $280m. At an acquisition price of $3.9bn, that is a ROIC of 7.1% - low, but of course the group employs a lot of leverage, hence I estimate their WACC to be about 7.4% (that is with a cost of equity of about 14%). The perpetuity value of that value destruction would be about $210m, or $1.50/share. The company guide for better top line growth and low to mid-single digit EBIT growth, but I do not really model that scenario. From the day of the announcement to now, the shares are down about $12. It is a bit much, I would say.

Company In Brief

They are a producer of cans for the beverage and food industry, as well as having a smaller machinery business. A recent acquisition of Signode has expanded their product range into packaging outside of the food and beverage can business. In 2017, prior to the Signode acquisition, which adds about $2.3bn of sales, the company had about $8.7bn of sales, of which $5.1bn were beverage cans, $2.3bn food cans, $1.3bn other products, which includes metal screw cap tops, and machines such as capping systems, conveyor belts, sealing machines, and so on. The sales pre-Signode were roughly 65% developed markets, with about 35% exposure to EM, the highest in the industry.

Industry Background

The beverage can industry is both very commoditised and competitive, but also quite consolidated. Volume trends in developed markets are not identical by region, with growth slightly higher in Europe, but the broad brushstrokes are similar. At the volume peak, which was 15 years or so ago, soft drink cans were about 70% of volumes, which is down to about 60% now. Soft drink volumes, particularly in the US, have been negative, driven by falling  soda consumption on the basis of the well-publicised health consequences – down about 20% from the peak

http://uk.businessinsider.com/americans-are-drinking-less-soda-2016-3?r=US&IR=T

Volume declines have recently moderated, driven by increasing demand for non-soda soft drinks taking up the slack from soda

(https://eu.usatoday.com/story/money/markets/2017/08/01/american-soda-consumption-plunges-31year-low/103953632/)

Both soft drink cans and beer cans have been supported by a continued substitution effect from glass to metal containers, which tend to be cheaper, keep products fresh, and are environmentally friendly (https://earth911.com/living-well-being/recycled-beverage-containers/) , and are a bit easier to transport. Specialty products such as craft beer have also supported volume growth in North America as well. Craft beer tended to be more focused on glass containers when it first became popular, but there’s been substitution here in recent years, too. About 15%-20% of volumes are now specialty. The substitution effect has been stronger in Europe, which accounts for the difference in volume growth (about a 100bps differential). Canned food has generally seen weaker volume trends – flat to small down – not rocket science to understand why – people prefer fresher foods, canned often does not fit the bill. Volume trends are stronger in emerging markets, with the usual drivers being behind that – higher GDP per capita growth driving higher consumption of branded products (Coke, Pepsi and so on). South America has grown at mid single digits in recent years, Asia, Middle East and Africa around high single digit combined.

The market is led by Ball Corp, with about 30% of the market, with Crown in second with 20%, and Ardagh (the parts spun out of Rexam when it was bought by Ball) at just under 20%. The tail includes such companies as Silgan, Amcor, and several smaller regional or unlisted players. In the more developed markets, it is important to note that markets are quite a bit more consolidated - in North America (largest single market - North and Central Amerca are 30% of global volumes), the top three own 80% of the market, while Europe is even more consolidated, with the top three owning nearly 90% of the market. I would also add that two of the top three in both of those markets have a commitment to generating EVA, with Crown partially compensating management on this metric, which I think reduces (though not eliminates) the risk of irrational price competition. Utilisation rates in North America and Europe are in the high 80’s low 90’s at the moment, with the rates comparably high in most of the ASEAN markets they are in, but lower in the Middle East. Competition is currently irrational in China, with competitors bidding for contracts at a loss – they are scaling back here.

Crown Holdings

The company will report four segments going forward – Americas Beverage, European Beverage, European Food, Asia Pacific, and Signode (the business they recently acquired). Aside from that, there is a fairly large Other division, which includes the machinery business and, most significantly, the North America Food business, which was, until 2018, reported as a separate segment. I believe this unit would potentially be up for sale if they got an attractive offer. I believe it would attract a price of about $700m – pro forma for the reduced earnings, that would take about 0.3 turns off the leverage.

The company have chopped and changed their reporting structure over the years. The European specialty packaging business used to be reported separately until being moved to “Other” in 2011. At the same time, Asia Pacific was broken out into a separate division in the segment reporting.

Organic revenue growth has been just under 1% annualised since 2009, by my estimates. The company does not have pricing power, and all the organic growth has come from volumes, mainly in the emerging markets part of the business, according to my work. The company has called out the fact that they have invested quite a bit in capex to scale up the specialty lines, which add value to the customers and which they think they should be better compensated for. So far that has not happened and I do not expect it to in the future – I think that the customers are just too powerful. I do not think one needs to believe in pricing power to buy into the stock here – that would be incremental upside.

The beverages businesses are generally better margin than the food businesses, I think because of the volume trends being less of a headwind. In Europe specifically, the company spent so time investing in their facilities in France in particular to move from tin to aluminium, which depressed margins. As those investments have wound down, that has been a tailwind for margins. Asia Pacific margins are somewhere between DM beverages and DM food. I would break Asia Pacific into China and “The Rest” (Singapore, Malaysia, Indonesia, Myanmar and so on). I would guess from company commentary that China is breakeven at best, as they have continually pointed to irrational price competition, so that would imply contribution margins of The Rest at the highest level within the group (probably about comparable to Mexico, which is included in Americas Beverage).

In terms of why margins have improved at the group level from around 10% on average from 2006-2015 to 12% currently, the key factors there have been:

• Improvement in gross margin by a couple of hundred bps, which I think may partially be driven by more specialties in the mix

• Steady volume growth of 1.5% or so, which has more than offset the more modest pricing pressure. Utilisation rates are currently high, so I think there has been better drop through than one would normally expect from volume improvements, as they have benefitted from better absorption of fixed costs

• Mix trends in Asia, with China (much less profitable) mechanically becoming a smaller part of the mix as they scale back from irrational price competition

I lay out my forecasts by division in the table below. There is quite a bit of information to digest in the table, but on the revenue side I have Europe showing limited growth – that is 1% or so volume growth being offset by pricing pressure. In Americas Beverage, I am going for low single digit growth – again pricing pressure across all countries in the segment, low volume growth in the US, but Brazil/Mexico’s mid-single digit volume growth providing a boost. I have Asia Pacific growing at about 3%, which will most likely wind up being conservative. Signode I have growing at 2.5% - historically I believe it has been less than that (2017 revenues are slightly down from the level they were at in 2012 when segment information was last given by ITW, however the dollar has strengthened 15% on a trade weighted basis since then). Crown guide that growth will be GDP plus going forward, so I believe 2.5% factors in a degree of conservatism.

I have modelled operating margins gradually fading back to the long-term average over time, mainly a function of the slightly negative pricing that I am envisioning – utilisation is fairly high, so it is just a reflection of the view that there is only a finite amount that they can get from better cost absorption, efficiency gains and so on.

I have also included a line called “Assumed Acquired Businesses”. From the conference calls they have given since the acquisition, it seems to me that the see Signode as a platform from which to grow, as they see limited opportunities in the can market. My assumption is that they spend 2-3 years delivering themselves to appease the market a bit, resume the buyback modestly (starting off with $100m per annum) in 2021, and then start doing deals to build on the Signode platform from 2022. It is obviously fun with spreadsheets, but by 2022, I think FCF (including minority dividends) will be about $900m or so; if they pay out $150m via buybacks that year, up from $100m in 2021, there should be more than enough cash available to start building on Signode. I tentatively model $500m in acquisition spending. I assume they basically overpay for businesses, paying 15x EBIT for businesses doing 11% margins (around their group average). That would add about $300m of pro forma sales, $33m of pro forma EBIT. Assuming they do the deals midway through the year, on average, it is ~$150m of sales and ~$17m of EBIT - $150m would add about 1.2% inorganically to sales. As group level free cash flow increases in the out years, I model steady outflows for deals:

I should note that the estimates above imply value destruction – the deals, should they happen at the multiples I am modelling, would generate a ROIC of 5.3%. I value the company according to a number of methodologies, and the DCF will of course reflect any long-term value destruction from overpriced deals.

The divisional income statement driven is, then, as follows:

This would then generate net income as per the income statement below. The company use adjusted EPS, with the main adjustments going forward being adding back intangible amortisation (about $60m), and adding back restructuring costs ($20m), less the tax shield. Historically, there’s been a litany of addbacks and deductions, with the main things being asbestos provisions, gains and losses on early retirement of debt at prices away from par, inventory gains and losses, and so on. I just use the GAAP numbers in my valuation work.

The historical reconciliation between GAAP and adjusted EPS is as follows:

I think that the Signode deal will take the company to a net debt position of about $8.8bn by 2018 year end, which would be about 5.4x reported EBITDA. It is high, but they have been close to that number in 2014 and 2015 after digesting the deals they did in Spain and Mexico. By 2019 year end, they should be back down to 4.5x, thanks to a full year of earnings contribution from Signode boosting the denominator, and a year of FCF generation reducing the numerator. From 2019 onwards, they would be able to delever by about half a turn of EBITDA per year. Ball Corp, their nearest competitor, operates with about 3.3x net debt/EBITDA, by comparison.

The following two table set out the historical leverage of the business, my estimates for how that leverage will look going forward, and finally the cash flow statement.

Valuation

In terms of valuing the business, 12x the 2020 GAAP EPS of $6.44 gets me to ~$77. The historical 1y forward P/E of the business has ranged from 9.5x-15x, around 13x pre-Signode:

By 2020, the company should be back around the historical leverage range, so the average P/E when net debt levels have been “normal” is probably not a bad market.

A target EV FCF yield of 6% on 2020 EV FCF of $1.06bn gets me to about $65. The EV FCF yield range historically has been about 10.5% (but that was during the financial crisis) to 5% (but that was with interest rates far lower). The range over the last 4-5 years has been 5%-7%, so I have gone for a yield right in the middle of that range:

A DCF gets me to $78, while a target EV/EBITDA of 9x (towards the higher end of thhe historical range, but a 10% discount to peers) would get me to $51. The average of all of the techniques I use to value the company gets me to ~$67, 47% upside, which I think is attractive.

Catalyst

No hard catalyst in mind