Tuesday, July 24, 2012

Hawaiian Holdings - Air Transport

Hawaiian Holdings, owner of Hawaiian Airlines (Hawaiian), is in the business of transporting people and cargo to, from, and between the Hawaiian Islands. 

There are no substitutes for Hawaiian’s value proposition: if you want to go to Hawaii, or if you want to get away from it, you either fly or you swim. Whether you choose to fly with Hawaiian rather than with one of its competitors is, in a commodified and competitive arena like air travel, going to depend on price, which itself is going to depend on cost. 

An investment in Hawaiian, therefore, can only be premised on the belief that Hawaiian is – and will remain – the lowest cost airline in its market, a proposition that the remainder of this post will defend.

“Economies of scale” exist when indivisible costs are an important share of total cost, and when the assets represented by those indivisible costs are specific to the market in which the business operates – an increased volume of business will reduce average cost. It doesn’t exist when they’re not. The concept of economies of scale is not so much about size as it is about density – a better term would be “economies of concentration”. 

In the airline business, indivisible costs are those that do not change when new routes are added.  In Hawaiian’s case, those costs are gates, terminals, and runways in Hawaii; management and other overhead expense; and the infrastructure for aircraft maintenance, also in Hawaii.

No other airline has such a concentration of fixed costs in Hawaii. Hawaiian has 47% market share – they call it “seat share” in the airline business – in and out of Honolulu and Kahului; by contrast, the next largest competitors, JAL and United have ~7.5% seat share each.   

The effect is even more pronounced in inter-island flights, in which Hawaiian has 87% seat share. 

Every additional flight, wherever the provenance or destination, has at least one leg in Hawaii, and therefore reduces Hawaiian’s fixed costs per unit – and average cost per unit – even further. 

One would therefore expect Hawaiian to be the lowest cost airline flying to Hawaii, and one of the lowest cost airlines in the United States, and it is:



(Note: Spirit doesn't operate in Hawaii; Allegiant flies to Hawaii but doesn't operate regular scheduled passenger service -- it is more tour operator than airline).

Hawaiian's competitive position appears even better when viewed from the perspective of the interplay between the three distinct markets in which it operates. 



The Interisland market is Hawaiian's stronghold. Demand largely consists of business travelers, on the one hand, and transfer passengers from outside Hawaii, on the other. Virtually every airline that ferries passengers to Hawaii feeds their traffic automatically into Hawaiian's interisland route network. They don't really have a choice.  Demand in this segment is therefore is inelastic, and when inelasticity in demand meets monopoly in supply, it is reasonable to suppose that economic profits are in the offing.

The West Coast segment is Hawaiian's firewall: it is surely  the most contested airline market in the United States. There is no fat in the fares on these routes and Hawaiian flies them at cost, earning a slight premium only in those few city-pairs where it has dominant market share. If Hawaiian is flying these routes at cost, it must be true that other airlines are operating them at a loss. It can be in no-one's interest to lower prices even further, and so prices in this segment have remained relatively stable.

Although this West Coast segment earns no economic profit, it nevertheless serves Hawaiian's strategic purposes in two ways:  first, it ensures that no other airline has an economic interest in establishing a substantial beachhead in Hawaii; and second, the low West Coast fares feed high numbers of passengers into Hawaiian's profitable interisland segment. 

These two segments, then, are interdependent: West Coast erects the barriers to entry and allows Interisland to profit; Interisland puts the fixed assets to heavy use and allows West Coast to be an effective barrier to entry. This interdependence coheres so well that it might even constitute a business model.


An ancillary benefit is that Hawaiian, facing an inelastic demand curve in its interisland market, is able to pass on increased fuel costs without discomfort, thereby improving its cost leadership in the West Coast market.




This has been the happy balance since 2008, when Aloha Airlines, the other Hawaiian airline, fell into bankruptcy,  liquidated, and saw its 30% market share immediately assumed by its old rival. 


Hawaiian's financial performance in this period is a hymn to the blessings of muted rivalry:  





and, from the perspective of owner earnings:




Hawaiian's management, recognizing that this yogic balance between the West Coast and Interisland markets represents a solid foundation, is in the process of assembling a high yield, high demand international segment -- Japan and Korea, in particular, but also Australia and New Zealand -- to lay on top.


Source: my calculations from US DOT data


At an average margin contribution of eight cents per available seat mile, the additional 3.3 billion seat miles in capacity can be expected to contribute at least $100 million in additional annual after tax profit, or approximately $2 per share.


Turning now to the matter of valuation. Hawaiian's Interisland and West Coast markets -- the base -- can be counted on for $161 million in annual normalized operating profit: an average return of 11.5% on installed operating assets of $1,398 million. These two segments alone value Hawaiian's equity at $854 million, or $16.84 per share. (I have used an 11% cost of capital, even though Hawaiian's pretax cost of debt is 5.15%, because, let's face it, it's an airline).




If the international segment can be relied on for an additional $100 million in operating profit, and I think that is a low estimate, then the equity is worth $19 per share. This valuation implies a multiple of less than 4x on forward owner earnings which, even for an airline, is conservative.


Disclosure: I have a position in Hawaiian.

32 comments:

  1. Great post.

    If you haven't seen it, here's another presentation on Hawaiian Airlines (along with a short call on American) from two years ago:

    http://valuehunter.files.wordpress.com/2010/09/hawaiian_airlines.pdf

    ReplyDelete
  2. Thanks. Hadn't seen it. We emphasize different elements of the investment case but it's a very well thought out piece and was certainly proved right.

    ReplyDelete
  3. Very interesting business model and you wrote it up well, however on the financials I have a few questions:

    How did you get to ~$1.9bn of debt. From the latest 10-k I see $430 in interest bearing loans, $600m in operational leases and $200m odd in unfunded pension obligations. Is there something I'm missing?

    Also the tax shield; surely that only applies to the interest bearing portion of the debt? Which from the above would only be $172m. Adjusting your valuation for this would lead to the company actually being over valued at current prices - would be interested to hear your thoughts on this.

    I also question equating depreciation to maintenance capex in the calculation of owner earnings. At least I assume that's what you've done, owner earnings is a bit of a grey area for me.

    Look forward to your response.

    ReplyDelete
  4. Jason,

    Thanks for the questions. You're quite right: I've overstated the debt by a small amount. From latest 10Q and capitalizing operating leases at 8x we get:

    So:
    pension liability = 327.3
    straight debt = 687
    leases = 665
    Total = 1679.3

    Each of the above items (and not just straight debt) has an interest (or embedded interest) component. All three are treated like debt for tax purposes. I have added back the implied interest on pensions and on leases to arrive at "Adjusted EBITA".

    Do you have any particular reason for suspecting that Hawaiian is over/under-depreciating its aircraft & engines? I have already eliminated the $23m amortization of acquired intangibles. The depreciation itself, however, seems quite accurate to me: it uses 20 years, 15% residual value, straight line.

    ReplyDelete
  5. "All three are treated like debt for tax purposes"

    I wasn't aware that either operational leases or pension obligations provided a tax shield. Do you have a source?

    Also for operating leases I'm curious as to why you simply capitalise next year's obligations at 8x, rather than using the total they owe, which would be $600m as per page 70 of the latest 10-k (clearly not a big difference but just a matter of curiosity).

    No I have no reason to believe that - if anything they're overstating it as planes can last for more than 20 years. However, considering the fact that their capex seems to regularly exceed depreciation by a very large amount, I'm reluctant to believe that only the $30-$40m that is depreciated is required to maintain their competitive position (I understand this year's capex figure is distorted by the large Airbus order). But I like to be quite conservative with my figures, so maybe I'm being overly cautious.

    ReplyDelete
    Replies
    1. Easiest source = http://www.anderson.ucla.edu/faculty/dick.rumelt/Docs/Notes/num_101-restruct_fin_state.pdf

      Any corporate finance textbook will also have that info. For a more comprehensive treatment of valuation generally, I recommend Koller et al.

      On the 2nd q: it's a back to basics issue, really. A company can report whatever number it likes in the minimum lease obligations; what it will actually use, however, will approximate 7x or 8x this year's rent.

      In the case of Hawaiian, I'm pretty sure that the number of planes that Hawaiian needs in order to service its current routes will not diminish over time. And those planes are either going to be owned or leased. Given that there's a tax benefit to leasing, it doesn't make sense that they would convert their leased planes into owned planes, etc.

      Also, but far less important than that logic, 8x is standard practice in the airline business.

      You can see why management might want to report a lower lease commitment figure, though: it encourages people use NPV instead and raises perceived ROIC while lowering perceived indebtedness.

      Capex: HA is growing. They need new planes for new routes, and bigger planes to carry more pax and cargo on old routes. The capex figure you're seeing, not just this year but since 2010, is almost entirely growth capex.

      "I'm reluctant to believe that only the $30-$40m that is depreciated is required to maintain their competitive position".

      Okay, but I guess I'm not sure I understand why. Their competitive position doesn't really have anything to do with their planes. HA just got lucky -- Aloha went bust and it inherited a monopoly that it has since used to build a moat. That's pretty much it. If anything, maintenance capex is slightly lower than D: planes are becoming cheaper over time.

      Delete
    2. Ok, thanks for the above info. I will look to get that textbook.

      For the last point, if you're willing to accept that it's mostly growth capex, then surely that means the incremental return on that capex is very low? I note that they have pretty hefty capex requirements going forwards. You must be of the opinion that operating profits are going to see a fairly significant upturn, otherwise I assume you'd prefer a buyback?

      Apologies for the basic nature of some of my questions - I'm trying to get past the low level analysis that's generally accepted where I work and you seem like a good person to learn from.

      Delete
    3. No worries. They're all good questions.

      I have incremental ROIC at about 17% between 2006 and 2012. That's above ROIC, so I'm comfortable with it.

      Bear in mind that aicraft additions are a bit lumpy and sometimes get added at the end of the fiscal year, so best thing is to use (profit in year 2)/(average IC in years 1 & 2) for ROIC; and a similar kind of treatment for RONIC.

      They DO have hefty capex reqmts over the next 5 years at least: they're replacing their Boeings with bigger Airbuses and they're adding to their fleet in order to accommodate new routes. So, no FCF for a while.

      And no FCF is what I want to see until they've exhausted all the low hanging fruit in the Asian and Western European markets.

      Delete
    4. Interesting. Thanks for the reply. You'll no doubt see me on your other posts at some point.

      Delete
  6. What are your thoughts after the recent results announcement and the loss in the 4th quarter 2012? I've only just started looking into Hawaiian but can see its not disastrous as there's little change in the fundamentals, possibly a good entry point but would be interested to hear your take.

    Also, how did you calculate your 2011 figures? The ones I have seen in the result announcement referencing 2011 are different in some key areas, e.g. operating income is 20.3m - possibly I've misunderstood the accounting used, not entirely au fait yet :)

    Regards,
    Kevin

    ReplyDelete
    Replies
    1. Everything is as it should be, Kevin. Strip out the $7.9m in one-time, noncash accounting charge and HA's earned ~11% on net operating assets, and $1.26 in earnings.

      When the balance sheet is published in the 10-Q in the next day or so, I'll post make available my updated spreadsheet.

      For the moment, the thing to remember is that 23.5 million in annual amortization expense is not a reserve for any future expenditure; and that adjusting for lease interest expense and the non-operating portion of the net periodic expense associated with pension & postretirement benefits reveals what is very likely the most profitable and most consistently profitable airline in the United States.

      That it is trading at the lowest multiples of any US airline reveals the opportunity, and that it can probably defend its turf without getting seriously hurt adds conviction to my understanding of the opportunity.

      Between now and 2015, the amortization expense will disappear altogether and capex will slow down dramatically. At which point, underlying earnings will become free cash flow. That's my take on it in a nutshell.

      Delete
  7. red, I'm looking into HA and I can't seem to reconcile the CASM data that you used with CASM data I found online. Sources below.

    http://web.mit.edu/airlinedata/www/2012%2012%20Month%20Documents/Expense%20Related/Total/System%20Total%20Expense%20per%20Equivalent%20Seat%20Mile%20(CESM).htm

    http://web.mit.edu/airlinedata/www/2012%2012%20Month%20Documents/Expense%20Related/Total/System%20Total%20Expense%20(Excluding%20Fuel)%20per%20Equivalent%20Seat%20Mile%20(CESM%20ex%20fuel).htm

    Where did you get your stage-length adjusted CASM data?

    Thanks.

    ReplyDelete
  8. Peter,

    You'll have to remove the annual intangible amortization cost to arrive at my CASM figures. (Also, remove the extraordinary charges in 2012 & 2011 and the extraordinary credit in 2008.

    As for the stage-length adjustment, I used the formula in your link:
    ((observed length of haul/base length of haul [15 carrier average])^.5) times respective carrier unit cost

    Here's a full spreadsheet:
    https://docs.google.com/spreadsheet/ccc?key=0AoCLhoPnjQDgdHZhQ3BzNzBGemdOalJvN1ZFV2Noa2c&usp=sharing

    ReplyDelete
  9. Sorry for the confusion, I was referencing the first graph with all the major airlines' CASMs.

    Thanks for the quick reply.

    ReplyDelete
  10. Ok, similar process. Take out/add back the exceptional items (making sure that they are indeed exceptional) from each airline's operating results and use the formula to arrive at the stage length adjusted CASM.

    One important adjustment is to capitalize aircraft rent and add back the imputed interest embedded in that rent. That way you get an apples-to-apples comparison of the airlines' underlying economics.

    I also take out the interest charge on net pension liabilities.

    ReplyDelete
  11. Can you explain your short on HA May$14 calls? And it was such a good one!

    ReplyDelete
  12. Just protecting some of my gains in the $9 calls. Decided to sell short-dated calls instead of buying long- dated puts.

    And Q1 is slow for Hawaiian -- most of their business is done in Q3 and Q4 and the company is not yet so well known that the market will take a near-certain Q1 GAAP loss in its stride.

    ReplyDelete
  13. Hi Red,

    How do you get the EV? 7xEBITDAR? And if so, where do you calculate EBITDAR in the spreadsheet?

    Thanks.

    ReplyDelete
  14. Hi Red,

    I´d like you to ask for your calls in Hawaiian Holdings:

    For example in your buy in 23th July of Hawaiian Holdings (Oct $12 calls) the price was 2.30. What I don´t understand is what does mean Oct $12 calls. I suposse October is the expiration date, but what is $12?

    Sorry for my ignorance.

    ReplyDelete
  15. It is an abbreviation for options to buy Hawaiian stock at $12 at or before the expiration date in October
    http://www.theoptionsguide.com/strike-price.aspx

    ReplyDelete
  16. But in july when you buy the options the stock was at 15 dolars. Now is 15,6 and the value of the option is worth more. Why?

    ReplyDelete
  17. "short" means "sell", David. I sold the 12s

    ReplyDelete
  18. Actually, now that I think about it some more, I'm not sure what you're asking exactly.

    I sold the 15s(not the 12s).

    Is this a general question about options? And are you asking about the $12 calls or the $15 calls?

    If it's a general question about how/why option values go up/down there are lots of introductory books on the subject.

    i should say that options are not at all appropriate if you're starting to learn about investing, though, and it would bother me if I didn't stress that in the strongest terms.

    ReplyDelete
  19. Thanks Red,

    Don´t worry, the link is very clear, I get it.

    By the way I read the Jason Rivera´s book. It´s a good book but for calculating the EV he uses always 4times EBIT,7 times EBIT, 11 times EBIT, and then like for art of magic he decides to take one of this multiples, and the reasonof the election is not very clear.

    The multiples used for calculating EV is a little confuse for me. But I´ll learn it.

    Thanks for your blog, its fantastic!!

    ReplyDelete
    Replies
    1. There is a way to connect EBIT multiples to DCF that one can reason about systematically. If one assumes a typical tax rate and WACC, then it's possible to use the two-stage continuing value formula from Keller's Valuation to chart EBIT multiple vs ROIC and growth rate. I once reverse-engineered such a chart from that book: https://docs.google.com/spreadsheets/d/1JhIGndo1NUoehP9_YMcsi5t3fYpDMTkwGyORcTGFEoc/edit#gid=2004453836. That helped me understand multiples.

      Delete
  20. Hi Red,

    Quick question on arriving at ROIC by NOPLAT/Invested Capital as noted in the Anderson pdf. If you exclude the operating and capital leases from the NOPLAT, then how is the real cost of renting equipment or space accounted for? I realize that if you capitalize operating and capital leases into invested capital it will lower the overall ROIC number. Also is ROIC just to compare against cost of capital to see if investments are value enhancing or can it be used to compare firms? Thanks!

    ReplyDelete
  21. Ah well that's the problem with excluding rent capitalization -- it makes a nonsense of ROIC and makes it impossible to compare one airline with another. Other wise, if I ran an airline and leased all my planes I'd show up as the highest quality airline in the world, spinning gold out of thin air.

    So you'd have to capitalize rent to get a sensible result and you'd have to use TEV/EBITDAR to assess the price you're being asked to pay for the business, and by implication, for the stock.

    Cost of capital: most of it is going to be debt and its equivalents, so I'd look up the actual interest charge on the debt components, and I'd calculate the implicit interest rate on the operating leases.

    You can find the second part in the 2nd table here:
    http://www.sec.gov/Archives/edgar/data/1172222/000117222215000012/exhibit122014.htm

    so 48m or aircraft rent is interest and 60m is depreciation. Rent capitalized at 8x = 800 so the implied interest rate = 48/800 = 6%

    Cost of equity ought to be 4% greater than highest cost debt instrument (i.e. the leases in this case), so cost of equity = 10% or so.

    Then weight varieties of debt and equity according to their size in the EV calculation and you have your cost of capital.

    It'll probably come in at 7% to 8%.

    Then you'd apply your common sense to it and say, well, this is maybe an abnormal interest rate environment and, in the long term, the cost of every kind of capital is likely to be bumped up by 3% or 4% as the Fed unwinds QE.

    So you'd maybe settle on 11% as your LT cost of capital estimate. I think 11% is fair.

    ReplyDelete
  22. ps

    you wouldn't be excluding all of the operating lease from your NOPLAT calculation -- only the interest portion of the lease expense.

    So, in this case, you'd exclude the 48m because that is a financing expense and not an operating expense.

    Imagine you buy a house for 500K and finance it via a mortgage. The sum of mortgage payments will equal the value of the house and the cost of borrowing. Same with operating leases.

    Since your cost of borrowing shouldn't affect the value of the house, a similar argument says that HA's cost of borrowing doesn't affect the value of the aircraft used in operating its business.

    ReplyDelete
  23. Hi Red,

    I'm trying to reconcile the operating margin information for interisland vs domestic and international, and can't seem to find the interisland info - would you mind explaining how you calculated that, or pointing me in the right direction?

    Thanks!

    ReplyDelete
  24. Hi Red,

    I'm trying to reconcile the operating margin information for interisland vs domestic and international, and can't seem to find the interisland info - would you mind explaining how you calculated that, or pointing me in the right direction?

    Thanks!

    ReplyDelete
  25. Sure thing

    Cost data is here
    http://web.mit.edu/airlinedata/www/Hawaiian.html

    Narrowbodies = interisland

    and

    Segment revenue figures are in the Annual Reports (reported as % of total revenue)

    ReplyDelete
  26. I found your story convincing back in 2012 and invested at around 6 dollars back then. I sold them last weekend at around 48 dollar each, the current rise makes me worrisome about an expected big drop or long lasting plateau fase. Usually I find airline investments too unstable and too unpredictable to invest into. Much appreciated anyway.

    ReplyDelete

Note: Only a member of this blog may post a comment.