Saturday, September 28, 2013

Air Liquide and Praxair

A look at the numbers for the companies in this industry.


Revenues and Profits

Comparing the size of the four largest players, their revenue and profits (in same currency):

Linde's profits are for the Gas Division segment only (probably excluding admin costs), which is why they are so high.  Ignore them.

Air Liquide is by far the largest, in terms of revenue and profit.  

Praxair is far more profitable, with a 20% operating margin in 2012 vs a 14% margin for Air Liquide.  Praxair has also grown profits a lot faster over the 8 years: up 140% vs 87% for Air Liquide.

Their profits are not cyclical, profits barely dipped for Air Liquide and Praxair in 08/09.

A note that profit margins (i.e.: percentage of revenue) cannot be used to spot trends over time in this industry, due to indexing to fluctuating energy prices (e.g.: increased electricity price by $X --> increased end price by $X --> same profit, but reduced profit margin).


Geographic Regions

Praxair is mostly in America:

North America
(incl. Mexico)
South AmericaEuropeAsia
Revenue (%)50191312
Profit (%)58171010

Air Liquide is mostly in Europe:

(North + South)
AsiaPacMiddle East/AfricaOthers (non-gas)
Revenue (%)46202239
Profit (%)47271934

AirProducts only gave revenue, not profit breakdown:

US/CanadaEuropeAsiaLatin America
Revenue (%)4028275

If you think that US or Mexican manufacturing will boom, Praxair would be a good play on this.



The companies have similar numbers when broken down for delivery mechanism:

For Praxair:
  • Large Industries 25%, Merchant (liquid) 31%, Cylinder 29%, Others 15%.
  • Most Electronics business would be under "Large Industries".  Healthcare is probably under "Cylinder".
For Air Liquide:
  • Large Industries 33%, Industrial Merchant (both liquid and cylinder): 34% , Healthcare 16%, Electronics 8%, Other 9%
For Air Products:
  • Large Industries 33%, Industrial Merchant (both liquid and cylinder): 38%, Electronics 25%


Balance Sheet

This is a capital intensive business.  All 3 companies have quite high debt of 3-5 years earnings:

Company2012 Long 
Term liabilities
2012 PBTNumber of times 2012 earnings to pay off liabilities
Praxair8.7b USD2.3bUSD3.1 times
Air Liquide9b Euros 2.1b Euro4.2 times
Air Products7.2b USD 1.48b USD4.8 times

Long term debt is mostly financed by 3-10 year notes/bonds.  The repayments are spread out into the future: 10 years for Air Liquide (footnote 25.1) and Air Products (footnote 14), and at least 5 years for Praxair (footnote 11).  Given the stable nature of the industry, this is sustainable - debt is not a problem.

All companies have minimal operating leases and contingent liabilities (mostly guarantees, warranties or legal action).  Praxair is contesting a USD 830m fine from Brazil's CADE for anti-competitive behavior (5 companies were fined).

All 3 companies have underfunded pension plans.  This current underfunding is recognized (i.e.: will not have make provisions in future year's earnings).  However the companies differ widely in their assumptions.

Praxair's pension scheme is only slightly underfunded (~700m), but their assumptions seem aggressive:

  • Expected return on US assets - the majority of the plan - is 8.25%.  A lot higher than discount rate (4.7-5.4%).
  • 60+ percent of the investments are in stock (target 60-80%).
Air Liquide's pension scheme  is underfunded by 1.9b Euros, but they are more conservative:
  • Their expected return on European assets is 4.4% vs a discount rate of 3.2%.  In 2011, Eurpoean expected-ROA was actually less than the discount rate!  Mostly invested in bonds and real-estate.
  • They are still aggressive in the US (which forms about 30% of their obligations): 8% expected-return vs 3.8% discount rate.  53% is invested in shares.
No problems for Air Liquide.  I expect some provisions for Praxair's pensions in the future.


Cash Flows

For all 3 companies, working capital is minimal.  Cash flow from operations is almost equal income plus depreciation.

Air Liquide and Praxair both generate steadily increasing earnings and CFO, even during the 08/09 recession, as expected when for a business operating on long term 10-15 year contracts.  The green bar shows steadily increasing Cash Flow from Operations.  However, Cash Flow from Operations (i.e.: spending on Capex) varies a lot, year by year - see the yellow bar.  Sometimes it is due to large acquisitions (annotated).  We cannot tell how much of this spending is to replace old capacity and how is to add new capacity:

From the above, Air Liquide invests (spends) higher proportion of its CFO on capex.  In detail:

In the ten years from 2002 to 2012, Praxair has spent 80% of its CFO in CFI, while Air Liquide has spent 90%.  Combined with Praxair's greater operating margins, this makes them a better user of capital.  Praxair's earnings increasing more than 3X over this period, while Air Liquide increased less than 2X.


Capex and Acquisitions

Since capex is the only varying quantity, lets look at it in more detail.

Praxair had high capex in 2011 and 2012, mostly buying small packaged gas companies - hard to tell as they do not give details and financial terms are not disclosed in their press releases.  In 2013, they spent 1.1b buying NuCo2, a provider of beverage carbonation.  At 9.6X EBITA, management admitted the acquisition is "dilutive to our return on capital", but expect that to lessen over time. From their September Investor Day Conference, 2012 was the peak of the capex cycle:

"CapEx spend. We think going forward we'll be around $1.8 billion, plus or minus. That's about 13% of sales versus the, say, 20% of sales that we had in 2012, where we were at the top of the CapEx cycle.
I've included 1% for acquisitions. We have a strategy to consolidate the packaged gas industry in the United States. So we're going to be spending some money towards that, call it $100 million to $200 million per year. Nothing on the order of NuCO2. I think it could be a decade before we do anything of that size. But this is our plan."

Air Liquide expects more than 2bn capex in 2013.  They will probably spend 2bn (plus acquisitions) for the next 3 years to 2015.



Praxair would be my first choice - with higher profit margin, it has been able to use less capital to grow faster.     Would still consider buying Air Liquide - they have a stable business with barriers to entry, and would provide some diversification as the operate in different parts of the world.

Even though revenues and profits for these companies are not cyclical, the stock price is.  Wait for a correction/crash to buy.  Anyway, valuations are too high to buy now.

Sunday, September 15, 2013

Bulk Industrial Gases

This industry is interesting due to its natural barriers to entry.  Because gas is expensive to transport relative to its value - basically they are selling air! - operations are regional.  Once a region is served by an incumbent, it is not economical for a newcomer to make the investment required to enter.  In addition, there are only a handful of global players.

Most of this comes from a Morningstar article, and Air Liquide's 2012 Annual report.

Industry Segments

Because bulk gases are more expensive to deliver than to produce, we segment the industry  by both delivery mechanism and end usage.

For the Large Industries segment: The company builds a plant next to its anchor customers and supplies them through a pipeline.  A plant usually has a regional monopoly within a 250km area.  Anchor customers sign a 15 year take-or-pay contract, indexed to energy costs.  The end users are metal industries (Oxygen), Chemicals (Oxygen, Hydrogen, CO), and refiners (Hydrogen).

For Air-gases (Oxygen, Nitrogen) the anchor customer takes up only about 70% of the output; the remaining 30% can be cross sold to smaller industries (e.g.: glass, automotive) in the area and distributed by truck (liquid) or cylinder (gas) (Industrial Merchant segment).  This includes product  delivered in liquid form (by truck) or gas (cylinder).

Hydrogen has less cross-selling opportunities, as only oil refiners use it.

Another segment is the Electronics segment, which is divided into:
  • Bulk Carrier Gases: Similar to the Large Industries segment, delivering a steady supply of these gases (hydrogen, nitrogen, oxygen) requires building production nearby the customer. We also have long term steady contracts: 10 years, indexed to energy, take-or-pay.
  • Electronics Specialty Gases are sold in small amounts, so distribution activity is worldwide, not localized. Demand varies with the electronics goods consumption and production cycle. These gases are high purity products requiring high tech expertise.
Lastly, the Healthcare segment involves delivering gases (usually Oxygen) by truck or cylinder to Hospitals and Homes.  This is usually also within a 250km radius, and the keys are coverage density (many customers in same area), quality of support services and efficiency.

Air Liquide illustrates this nicely their 2012 AR:


For Large Industries: Worldwide, most bulk gas (80% of hydrogen production and 65% of oxygen production) is produced in-house.  The remainder is produced by suppliers.  It is a concentrated market: the 4 largest suppliers (Air Liquide, Linde Group, Air Products and Praxair) had 60-80% of the U$74 billion global market.

In advanced economies: most oxygen production is already outsourced, while the supply of hydrogen is largely in-house.

In developing economies, outsourcing is relatively new, but accelerating.   In China, there are 3 local Chinese player: Yingde Gases (listed), Hangzhou Oxygen Plant Group Co (private), and Hang Yang Group.

The Industrial Merchant segment is also localised, but highly fragmented with many players.  So by itself, it offers lower returns.

Healthcare is also localized and fragmented.  Air Liquide expects growth through acquisitions in this sector.

In Electronics, three companies play a major role: Air Liquide, Air Products and Taiyo Nippon Sanso.



Bulk Industrial Gases is an attractive industry because:
  • An established supplier has a sustainable competitive advantage.  After an anchor plant is built, all Large Scale and Merchant Gas customers are served from this plant.  It would not be feasible for a competitor to build another plant in the vicinity. This applies to all segments except Electronics Specialty Gases (which are not bulk).
  • The industry is mostly non-cyclical due to long term contracts.  Electronics Specialty Gases and the Industrial Merchant segment are cyclical.
  • There are only a handful of large global players for Large Scale and Electronics.  Industrial Merchant and Healthcare are fragmented, with many local smaller players.