Description

Alfa SAB is a Mexican conglomerate that’s cheap on a SotP basis.  …Wait, where’s everybody going?  Guys, come back, it gets better I swear!  BTudela16 wrote this up a while back, and like many good ideas on VIC it didn’t get much attention.  I’ll try to add as much incremental info as I can.

(Note: all figures in USD millions unless indicated.  Most revenues come from outside of Mexico.) 

Alfa is a conglomerate in the final innings of selling/spinning off all the parts such that what remains will be just Sigma - their 100% owned CPG business.  The biggest other part is Alpek, a maker of plastics and polyester.  Both Sigma and Alpek are quality businesses.  Axtel is a telecom biz with issues, but it’s tiny.

I’m sure y’all can relate when I say that there’s often negative surprises with event driven situations, and it always takes longer than you thought.  One such surprise happened recently with the failed sale of Axtel.  But all that said, I think this is a question of when, not if:

• Mgmt is totally committed to eliminating the holdco discount.  They talk about their  “Unlocking Value” plan on pretty much every call.
  
• Both Sigma and Alfa generate a lot of cash.
  
• The multiples I’m counting on are really conservative.  The Mexican stock market has been dead as a doornail for some time now and there’s no froth that could vaporize on us.  The only business that Mexican I-bankers can scrounge these days is basically just pitching companies on de-listing.  
  
• No worries that a Leftist victory in the July’24 presidential election could derail the plan.  Mexico already has a Leftist president (Obrador).  Obviously I’d prefer someone more business-friendly, but the point is that Alfa has accomplished plenty in terms of unwinding this thing since Obrador came to power on 12/1/18.

The upside here is a double on a simple long position in Alfa shares.  The stub offers far more potential, but that upside depends on the path they take (more on this later). 

The gating factor is how fast they can pay down debt - likely at the holdco level, but Sigma’s debt is also a possibility.  Once Alpek is spun-off, the holdco will be collapsed into Sigma, and mgmt wants to keep its investment-grade ratings on all debt.  This means getting net debt at the holdo + Sigma down to at least 3.0x Sigma’s 2024e EBITDA, including the ~15m of holdco that should remain at that time.  Currently they’re at ~3.9x, so about 700m of debt reduction is needed.

Complicating things is mgmt’s dedication to maintaining the Alfa dividend (now 6.6%).  My guess is that many in the founding Garza family are living off the dividends, and the Garza’s still own a lot of shares.  The Chairman and President (ie CEO) are both members of the family.

The table below calculates the current shortfall in funding using only the stuff that mgmt has strongly hinted is likely to happen, namely a special dividend from Alpek in 2H22, and the sale of Alfa HQ land in Monterrey (BTudela16 values this at 500m but all other estimates I’ve seen are more like 250m.  

I figure they’ll pay down 200m of holdco bank debt first, then deal with their 500m bonds on their 3/25/24 maturity or potentially pre-pay them next year.  So there’s a 112m funding gap that mgmt has a number of ways to fill.  They’ve mentioned potential sales of some non-strategic assets in addition to the land, but haven’t given any details.  Another option is a special dividend from Sigma, which should do ~250m in FCF in excess of the dividends shown above in each of ‘23 and ‘24. 

A second special dividend from Alpek - perhaps in ‘24 - might also be an option.  As you’d expect, Alpek is cyclical, but there’s no new capacity that’ll come online in the near-term that could screw things up, and it’s arguably at the peak of the cycle now - just gushing cash.  Using mid-cycle earnings, I estimate they could add ~290m of debt to fund a special dividend in 4Q22 and still stay investment-grade (2.0x net debt/ EBITDA ex minorities).  That 238m you see above is the amount going to Alfa’s 82.1% stake.

(Detail on a fairly minor tax issue - you might want to skip ahead for now)

One last detail to keep in mind is that spinoffs in Mexico are considered asset sales and are therefore taxable.  Following the route taken for the spin of Nemak in 4Q20 and the upcoming spin of Axtel (likely 4Q22), the tax - if any - is paid by the holdco only.  There are other methods such as a dividend distribution or a capital reduction, but Alfa has avoided these as they could create a tax liability for the holders of the spinoff shares.  

With Nemak and Axtel, the tax is based on the difference between the spinoff’s share price on the date the plan is approved and the tax adjusted cost of Alfa’s stake.  Note that all these companies are already publicly traded - that is, the portion of each that Alfa doesn’t own.  Tax adjusted cost is based on book value, but with a complicated set of adjustments that seem to have evolved from dealing with high rates of inflation in LatAm countries.  Thus it can wind up being much higher or lower than book. 

Tax losses at the spinoff-company level will tend to increase the tax adjusted cost basis.  Axtel has had a troubled history, resulting in NOLs equal to roughly 125% of book.  Thus we see that Axtel’s tax adj cost is over 4x higher than the book value of Alfa’s 53.9% stake, which should add ~230m to the Alfa holdco’s NOLs upon spinoff.  Nemak, on the other hand, had NOLs equal to just ~12% of book, and its tax adj cost was about 40% of the book value of Alfa’s stake.

Oddly enough, Alpek’s NOLs are ~53% of book, and so I doubt that its tax adj cost will be dramatically lower than the book value of Alfa’s stake.  That book value is MXN 41.3 billion vs Alfa’s stake’s market cap of MXN 46.9 billion.  Post the Axtel spin, Alfa holdco should have roughly 930m (MXN 18.6 billion) of NOLs, so I don’t envision any material tax bill.  But even if the tax adj cost was zero, Alfa would owe taxes equal to about $1.70 per Alfa share (at a 30% tax rate), so I’m not worried about this.

(I say “oddly” because from 2009 onward (the extent of my data), Alpek has been solidly profitable on a consolidated basis, posting a loss only in 2017 due to a writedown.  FCF has always been positive.  The vast majority of Alpek’s NOLs relate to their ops in Brazil, which are only ~10% of assets.  But Alpek’s full history dates back to 1975, so who knows.  

About Sigma

Sigma’s origins date back to 1980.  It’s done 29 acquisitions since 1997, and in the past 20 years has expanded into LatAm, then the US, then Europe.  Today about half of EBITDA comes from Mexico, 19ish% from the US, 24ish% from Europe, and the remainder from LatAm.  Sigma’s main category is packaged meats (mainly pork), followed by cheese and yogurt.

From 2003 onward, Sigma’s ROE has been 20% on average on 2.1x net debt/EBITDA.  Exposure to emerging market currencies hasn’t really been a problem.  The Mexican ops do import a lot of pork from the US, but they have a lot of ways to offset peso depreciation - e.g. renegotiating price, reformulating products, etc.  Mexico EBITDA margins have ranged between about 12-16% since 1997 with little correlation to the exchange rate.  LatAm margins have generally improved over time and have been in a range around 9-10% from 2015 onward.

Sigma’s Europe ops (Campofrio) has suffered from higher energy and input costs due to the war in Ukraine, and margins went from 6% in 1H21 to 2% in 1H22.  Lower exports of pork to China due to African Swine Fever have also had an impact.  Mgmt revised its ‘22e EBITDA guide from 760m to 700m.  A first round of price increases was felt in 2Q22, and a second round at the end of the quarter should help 3Q results.  Management is also in the process of optimizing its manufacturing footprint - 8 of 24 plants are slated to be sold.  I think they’ll get things back on track and my 2024e EBITDA assumes just 2% annual revenue growth with margins going from 10.0% this year to a more normal 11.0% in 2024.

About Alpek

Alpek makes plastics and polyester from oil and NGLs.  It’s a commodity business but not a bad business, with an average Operating ROA of 11.8% from 2010 onward, which translates to 8.6% after an effective tax rate of 27%.  Normalized ROE has been ~15% on 1.5x net debt/EBITDA.  Around 89% of Alpek's revenues comes from the consumer goods markets (food and beverage, consumer staples), which is more resilient than end markets such as construction and automotive.  Anytime you buy a drink in a plastic bottle in North America, there’s a 50% chance it was made with Alpek plastic.  Also, over 95% of revenues are in USD, while over 70% of assets are in emerging markets, so it’s very much short EM currencies.

(Current share price is $27.00.  Note that cumulative FCFE would be $23.24 if the cost of working capital was included.)

 Alpek closed the acquisition of OCTAL on 6/1/22 for a cost of 620m, which should add 120m EBITDA in ‘22 and ~100m in a normal year. 

There’s an ESG angle to Alpek, as plastic bottles have a much lower carbon footprint than other forms of packaging (glass, aluminum, etc.).  In the past, environmentalists have derided plastic due to sea turtles getting straws caught in their noses and a lack of recycling infrastructure.  But most of that sea trash comes from Asia.  The US is actually pretty good about waste disposal.  And the recycling infrastructure is getting built, and Alpek is a part of that.

About Axtel

Not much to say here.  Axtel does telecom almost entirely in Mexico.  Alfa has been trying to sell it for years, but declining performance over the past 2-3 years made that difficult and so they finally threw in the towel on 6/27/22 and decided to spinoff their stake.  This was a real disappointment as the proceeds from a sale would have gone a long way towards paying down holdco debt.  Axtel has a 440m bond due 11/24/24 which is now trading at 71 (Alfa has not guaranteed the debt).  They’re working on a refi using bank debt right now.  Alfa is financially savvy and I really doubt they’d spin it off vs accept a lowball bid if they didn’t think Axtel could refi, but it’s not a big deal either way.  If Axtel bit the dust, downside is just $0.47 per Alfa share.  If they can refi and stabilize the biz there’s upside of maybe $2.30 per Alfa share.

Relevant history

2014  After a good run since its founding in 2006, Alfa’s wholly-owned Newpek oil & gas subsidiary saw its fortunes turn in late-2014 when the oil & gas prices fell off a cliff.  Newpek had spent over 1 billion to buy shares in Pacific Rubiales, and raised an additional 1 billion of debt to further invest in this area.  Mexico had recently lifted its 75-year ban on foreign investment in the oil & gas sector.  This is where the two holdco bonds come from.  It’s also the source of the holdco NOLs, though some of these were used in the spinoff of Nemak.

2015  Alfa IPO’d a ~25% stake in Nemak - its former auto parts business - to raise money for investment in other areas.

2016  Alfa merged its Alestra - its telecom unit - with Axtel, another telecom that had gone public in 2005.  This gave them 51% of the merged entity, and their stake today is 53.9%.

2016  Around this time, Mexico did away with a law that had allowed holding companies to use losses at one opco to offset the taxes of another(s), thus making the conglomerate structure less attractive.

9/27/17    Alfa calls off its planned IPO of Sigma due to disagreements over price.  You can see in the chart below that the Sigma stub had been trading around 8.0x forward EBITDA in anticipation of the IPO, until this announcement.  For what it’s worth, 8.0x is the lowest multiple I’ve ever seen any analyst use to value Sigma.   

12/14/20  Alfa spins off its remaining ~75% stake in Nemak.

February 2021  Alfa kicked out of various indices due to a purely technical issue. Foreign-held shares were in something called a NAFINSA Trust, which was capped at 50% of total shares.  Foreign investors had been buying shares post Alfa’s announcement of its Unlocking Value plan and breached the limit, forcing index funds to sell.  The cap has been raised to 75% and there’s good breathing room on that issue currently.  Alfa is back in the indices.

2024(?)  Spinoff of Alpek, holdco collapsed into Sigma, name of surviving entity changed to “Sigma”.

2025(?)  ADR listed in the US.  Management hasn’t talked about doing this, but they’re well aware that the US markets would almost certainly ascribe a higher multiple to Sigma.  I think they’d easily get a 2.0x EBITDA valuation bump.  

The Stub

The Sigma stub is highly levered, owing to the net debt at both Sigma and the holdco, and also to the leverage on the market values of Alfa’s stakes in Alpek and Axtel, which are 83% of Alfa’s market cap.  No surprise, then, that the stub price has been highly volatile, usually due to the Alfa share price.  I can’t figure out what caused the stub to crater in June’18 or July’20, which I guess is kinda scary.  

On the bright side, short interest is de-minimus in both Alpek and Axtel, and borrow costs are 1% each at IB.  Aside from the market rout from Covid in early-2020, Alpek is trading at 10-year lows on either a P/B or EV/Fwd EBITDA basis.  But I see near-zero risk of retail investor crazies suddenly getting excited about plastics in Mexico.  

Aside from volatility, I see two main concerns with playing the stub.  First, as the Alpek share price rises, so too does the potential tax owed by the holdco.  I can’t be certain how much it would have to rise before holdco NOLs can no longer shield gains, as I don’t know Alfa’s tax adj cost.  You absolutely will not get this number from mgmt, and I suspect the main reason is that they don’t want to attract unwanted attention from the tax authorities.  The head of Mexico’s tax agency is known as “The Iron Lady” for her get-tough attitude towards corporations.  Unfortunately she’ll be around until at least 12/1/24.

The second issue is that playing the stub means paying those special dividends from Alpek.  Absent the special dividends, the stub has a slight positive carry (including borrow costs) due to dividends from Alfa.   But with a stub market cap of just 505m, a 290m special dividend changes things quite a bit.  If Alpek did a second special dividend of 125m to fully cover the shortfall, the MoM reduces to 4.8x.  My inclination would be to put this trade on after the special dividend is paid around 4Q22.  Then, if the remaining shortfall is covered by a special dividend(s) from Sigma or asset sales, you’d make a much higher return, other things equal.   

Catalyst

The passage of time.  These assets generate a lot of cash, thus paying down debt is a question of "when", not "if"

Asset sales

Special dividends from opco's