Dart Group: Looking back & looking forward

When I looked at Dart Group last year it was undervalued: it could have easily generated unencumbered cash sufficient to pay off its liabilities to its creditors and to cash out its shareholders at the value of its market capitalization. In other words, it was a pretty good cigar butt at least and maybe more – one didn’t have to think too much beyond that. 

Since then, as everyone knows, the share price has almost quadrupled. A nice trade that I missed out on because I was thumb-sucking over the 10p that I’d have to pay for the airline (I valued Fowler Welch at ~60p and I wanted Jet2 for free).

At the current market cap, however, more scrutiny is required. Jet2 is a low cost airline. Or rather, Jet2 competes with low cost carriers. Now, in that particular market, what matters is unit cost: people will fly with you if you’re cheap and they won’t if you’re not.

 Jet2's route map. 50% of its traffic is to Spain.

To date, competition has been muted: Ryanair and Easyjet have been busy frying other fish. So Jet2 has been filling its planes, earning profits on ancillaries, and funding its growth via 0% interest bearing loans from its customers in the form of deferred income. This has added up to a high ROIC.

But a high ROIC is not evidence of competitive advantage. Sometimes a high ROIC is built 

into the nature of the sector (every advertising company and recruitment consultant has a high ROIC) and sometimes it is a function of the fact that the competition hasn’t come for you yet.

So, as I say, what matters in price competition is unit cost. And unit costs in the airline business are denominated in “cost per available seat kilometer” (CASK) in Europe, and “cost per available seat mile” (CASM) in the United States. 

This is how Jet2’s CASK compares with the two low cost carriers that are most likely to compete with it in the future:

Consider now that Jet2’s loss of the one quarter of the Royal Mail business will raise its CASK further. Consider also that Ryanair and Easyjet fly much shorter routes than Jet2 does (it costs a great deal more per kilometer to fly short routes than long ones).

Suddenly, there’s cause for concern: if Ryanair, say, decides that it wants to fly holidaymakers from Bradford to Alicante, it can price the flight much lower than Jet2, Jet2's bookings will fall, deferred income will fall, cost per passenger will go up, Jet2 ticket prices will have to go up,  and the negative spiral will be firmly  in place. 

What was a high ROIC operation frolicking in a Jacuzzi of cash can very quickly become a low ROIC business suffering from liquidity problems as the legacy airlines have long since discovered.

A note on the figures:
1. I have taken average stage length of each airline from a source I know to have been reliable:
http://centreforaviation.com/analysis/jet2com-strong-fy2013-profit-growth-hides-major-challenges-120030

2. Royal Mail flights are 10% of Jet2's departures and Royal Mail flights have numbered 16 per weeknight, implying that total annual departures for Jet2 = 16 x 5 x 52 x 10 = 41,600 flights per year.   

3. We know the composition of Jet2's fleet (and Ryanair's and easyjet's) so that we can use the maximum seat configuration to arrive at a figure for average number of seats per departure.

4. Together, these give us available seat kilometers (average stage length x annual departures x average number of seats per departure).

5. Short flights are more expensive per kilometer than long flights; they are more taxing on passenger, luggage, and aircraft handling, airport fees, etc. In order to adjust for this, industry convention uses the following formula: a multiplier  -- (average stage length of airline X/ average stage length of all airlines)^(1/2) -- is applied to the CASK of each competing airline to arrive at a cost per equivalent seat mile. If one wants to know how the cost structures of Jet2 and Ryanair would compare on the same route, the cost per equivalent seat mile is what one wants to know.

6. Applying the cost per equivalent seat mile to the distance between two airports (Leeds-Bradford ("Bradford") and Alicante results in a cost figure that can be used to compare the cost structure of two airlines. Small differences (10% or less) don't matter too much but large differences matter a great deal. In this case, it appears that Ryanair could price the flight at a 10% mark up and still charge half Jet2's fare.

7. That may explain why Jet2 is so keen to generate revenues from ancillaries: the more it earns from in-flight extras, the more competitive it can make its ticket prices. One can see that tickets are already priced at below cost. Can such a strategy survive the future attentions of Ryanair or easyjet? Each person must judge that for him/herself but I, for one, doubt it.

This post was corrected, amended & extended on August 21st

Disclosure: No position in Dart Group

Unitek Global Services

Unitek Global Services, listed on the Nasdaq, is a contractor with two business lines: (1) it installs and maintains home satellite and cable connections on behalf of DirecTV and Comcast (the “fulfillment” segment); and (2) it performs network engineering, design, construction and project management for the wireless and cable broadband industry (the “E&C” segment).

The fulfillment segment is Unitek’s bread and butter. It earns its revenues under long-term, fixed-fee master service agreements and therefore enjoys high revenue and gross profit visibility and those modest competitive advantages that accrue to regional scale.  Small, independent contractors are losing business (and/or selling themselves) to larger regional players like Unitek and MasTec. Fulfillment’s revenue growth, therefore, has mostly come from acquired and organic market share gains. The benefits from DirecTV’s own market share gains and the benefits from increased subscription churn have been lesser tailwinds. 

The E&C segment, on the other hand, is Unitek’s growth business: increasing demand for broadband wireless has meant increasing infrastructure capex by the major wireless service providers, and Unitek’s E&C segment has benefitted from its association with AT&T in particular. These are trends and relationships that will more likely than not continue over the medium term although, in competing with the likes of General Dynamics, Bechtel, Quanta, and Dycom, its gross margins in this segment will necessarily remain modest.  

Regional scale economies and some modest customer stickiness ought to translate into a reasonably profitable business earning low-to-mid teen returns on capital and, after shedding itself of unprofitable geographies where it had no scale Unitek’s underlying performance conforms to those expectations: 

Unitek's footprint, by segment

The strategic use of the word “underlying”, however, and the fact that Unitek’s market cap is $30 million, lets us know that there is more to this story.

First, and least interesting, is that with amortizations of debt discounts and acquired (non-operating) intangibles, noncash impairments, transaction costs, and all the rest of it, the difference between GAAP earnings and underlying earnings is comically substantial:

Second, in taking on a substantial amount of debt to fund its acquisitions, Unitek was inviting a near-fatal accident. And, when it acquired Pinnacle Wireless in 2011, that invitation was accepted: a year and a half later, an investigation by the Board’s audit committee discovered fraudulent revenue recognition practices at Pinnacle; the CFO, Controller, and division president were terminated; and the company was unable to file its 10-K and, later, its 10-Q’s, on time. It received a notice of suspension from Nasdaq, of course, but more important for our purposes, its inability to file financial statements triggered a default under the terms of its borrowings.

At this point, its lenders could have taken it to Ch. 11, wiped out the equity and brought it back under their ownership. (Between December and April, the share price fell by a quarter and holders of the common fingered their worry beads; between April and June, the market cap was cut in half as even the devout fled).

That the lenders (and Cerberus, especially) did not do so ironically reflects the value of the business. (I think we can all agree that a business that generates $25 million  of free cash flows to equity is worth more than $30 or $60 or even $90 million). In a Ch. 11 process, transaction prices would have been high, the administrative fees as extravagant as usual, and the implied returns to the acquirer therefore low. And besides, DirecTV had served Unitek with a 180 day notice of termination which meant that time was tight.

The lenders’ alternative option – one that promised higher returns – was to add penalties (warrants for 20% of the equity, please!) and higher interest charges on the debt. And that, in the end, was the outcome.   

So, Unitek’s future, under highly conservative growth and margin assumptions, looks something like this:

Apply a conservative multiple to 2016 earnings, discount back to the present, subtract the inevitable cash costs related to the forensic audit, the refinancing, and any litigation costs that may arise from the alleged fraud at Pinnacle Wireless, and Unitek’s value is likely two, three or four times its current market cap, even with the 20% dilution. At the other extreme, as Unitek de-levers and grows, MasTec will be an imperfect, but good enough, comp and MasTec trades at 22x.

Disclosure: I am long Unitek Global Services  

Portfolio Update

I have taken a small position in Unitek Global Services. It's messy, this one. I'll write it up over the weekend.

Portfolio Update

I'm out of Precia with a 17.4% gain (8.1% from share price, 2.2% from the dividend and 4,9% from the exchange rate effect) which is well below the returns from the index (pick one, it doesn't matter much). 

It's still inexpensive, though (like GEA SA) it requires more patience than I am able to muster, especially with some interesting names reporting over the next few days.

I suspect that I'll circle back to it at some later date.

Portfolio Update

Reduced Hawaiian a little and picked up some more Northgate with the proceeds

Silverdell

This is what I saw when I looked at Silverdell a few months ago:

Reinvesting capital that costs 10% in order to earn 2.31% returns is an unhealthy practice.

Emeco's Downside

What will happen to Emeco’s valuation if there’s a material, permanent slowdown in mining activity?

It will reduce its inventory of equipment and shrink its working capital needs, thereby generating increased free cash. If revenue falls by 15% a year between now and 2016, the free cash flow schedule, at more or less 80% utilization and 36 days of working capital, will look like this:

Doesn’t this assume an orderly liquidation?

Yes, it does and for a number of reasons:

(1) Emeco is more exposed to mining operations than to mining capex (and therefore more exposed to commodity supply than to commodity prices);

(2) Under conditions of commodity price uncertainty, the benefits of leasing v. ownership are brought into sharper relief;

(3) Emeco is exposed to thermal coal (with its blue-chip customers themselves operating under long-term supply agreements with Japanese, Korean and Chinese buyers) as well as to gold, iron ore, oil, and civil construction;

(4) Emeco’s used equipment holds its value well, reselling at 15% above book value even in 2009;

(5) Pressure in the Australian rental equipment market will squeeze out smaller players first, thereby stabilizing rental rates somewhat;

(6) Oversupply of heavy equipment will shrink the supply of new equipment first, thereby cushioning the resale market from the effects of a sharp downturn in mining activity;

(7) Emeco’s fixed operating costs amount to ~$48 million, or 8.5% of current revenue, meaning that expectations of heavily negative operating leverage in case of a fall in revenues are misplaced;

(8) Emeco management’s incentive pay is structured around ROIC and share price which leads one to believe that the company will attempt to (a) reallocate existing equipment to more productive uses, (2) shed equipment at attractive prices, and/or (3) repurchase its shares at attractive prices;

(9) Three and six months ago, with the same information available to them then as we have now, the company was repurchasing a material number of shares at $0.53.  

Why does this opportunity exist?

My best guess is that the market has lost sight of Emeco’s place in the mining value chain, has overstated its fixed cost base, and has underestimated its cash flow levers. The Indonesian contract dispute has served up a profit warning that was perfectly timed to reinforce the market’s assumptions about its business.  

I am sometimes wrong but almost never in doubt so, as always, please do your own research.

Disclosure: I am long Emeco

Northgate's 2013 Results

Northgate  reported its FY 2013 results todays The company incurred a £54 million charge in order to reduce its interest expense from 7.1% to 2.8% . The lower interest charge translates into $20 million in additional annual after-tax cash flows to equity.

Here’s its free cash flow history:

Here’s what its valuation looks like:

Corrected:  July 9th

This is how I arrive at 700p:

Disclosure: I am long Northgate

More on Emeco Holdings

Some things it is helpful to know about Emeco:

“A large part of our sustaining capex is the replacement of machines at the end of their useful life. We can therefore naturally contract certain asset classes within our fleet where we see lower demand for them over the medium term and use the cash to retire debt. I refer to this as “right sizing” the balance sheet to the earnings of the business through the cycle.”

and

and

And a long presentation by the CEO here.

Disclosure: I am long Emeco

Portfolio Update

I’ve bought some shares in Emeco Holdings at ~AUD 0.310 (or <2x trough FCF). 

Epicentre Holdings

Epicentre Holdings is a Singapore-listed “Apple Premium Reseller”: it retails Apple products and third party accessories (iPad covers, speakers, etc) in Singapore and Malaysia.

Longer videos featuring the CEO are here and here

The business model is simple enough: pull people into one’s stores by offering Apple products at a slim margin and make one’s money on the accessories, the after-sales service, and so forth. Not complicated and there are probably only three ways to muck it up: (1) People no longer like Apple products; (2) Apple decides to sever its relationship with Epicentre; (3) the Singapore Dollar takes a nosedive.

The last of these is, for our purposes, the item of interest. The USD has appreciated against the Singapore dollar. Since Epicentre buys its Apple-branded inventory in USD,  its gross margins on Apple products have nosedived, turned negative, and made their effects felt all the way down the income statement. Plunging earnings, plunging share price.

Epicentre has now negotiated an arrangement with Apple whereby its Singapore operations (80+% of revenue) may now buy Apple-branded inventory in Singapore dollars. Its Malaysian operations (<20% of revenue) will continue to buy in USD.

The upshot is that gross margin should improve by a minimum of ~3.6% and that the earnings margin should therefore improve to a minimum ~4.0% .  A 4% net income margin on SGP $180 million translates into an earnings yield of ~41%. And, since Epicenter pays out 85%+ of its earnings, it seems reasonable to expect a forward dividend yield of 35%.

There are some other factors that justify more upside to the valuation – the announced exit from the loss-making presence in PRC, immature new stores in Malaysia, investments in “customer-centricity” that have been expensed – but these are immaterial to the investment case at the current share price.

The shares are illiquid and any buyers will be competing with the company itself.

Disclosure: I am long Epicentre and intend on buying more shares.

ConscienceFood Holding (Continued)

So, what’s going on?

The dirty shop window: It’s hard to see what’s on offer unless one actually reads CSF’s quarterly and annual reports. The financials, as reported, look like this (give or take some minor rounding errors):

The notes to the financials and management’s discussion and analysis add the following information:

First, CSF had tried to make a go of adding snack noodles to its repertoire. But the demand for snack noodles just wasn’t there. The gross margins were thin or negative, the opportunity cost of using the production lines for snack noodles rather than for instant noodles were therefore high, and CSF stopped making them.

Second, CSF invested in, and opened, a factory and warehouse complex 30km from Jakarta in order to serve western Java. Soon after, however, that complex was subject to a major fire, causing Rp. 27 billion in damage to property and inventory and slowing production to 30% of what had been its capacity. In addition, an additional Rp. 11.6 billion in G&A expenditure was tied to that plant and that spending was therefore mostly unproductive in 2012. The factory should be up and running at full tilt in the second half of 2013. 

Third, CSF’s marketing and distribution costs rose by 84% in 2012 partly for the purpose of establishing a distribution system to service Western Java and partly to promote the full launch, in 2013, of its line of beverages. Capital spending also rose. These increased expenses and capital investments had little or no corresponding revenues in 2012, and margins and asset turns were therefore crushed:

So, as I say, that’s part of it. No-one likes to see such a steep collapse in margins and turns and, if one hasn’t read the filings and if one is also of the view that big is beautiful everywhere and every time (that, for example, Coke will see off Irn-Bru) then one will read in these numbers cause for alarm: this is just the start and it’s going to get worse; “reversion to the mean” and all the rest of it.

What else?

The custom in the Singapore exchange is for listed companies to pay dividends. It is so much the custom that it is quite normal for companies go public and pay out dividends in their first quarter of being listed.  Following this norm, the declared dividend policy in CSF’s IPO prospectus was 20% of earnings.  CSF paid out dividends in 2010 and 2011. But in 2012 it amended its dividend policy: henceforth it would pay dividends in the absence of capex claims on cash. In 2012 it didn’t pay out a dividend.

And these two overall factors – dirty windows, dividend policy – are the long and short of it. I don’t think that they add up to an enterprise yield of 30% but the market apparently does.

There’s a lot more to say – about the cup noodles and beverages, about CSF’s prospects in the Jakarta market, etc – but I’ll leave it at that. If you’re reading this and are interested in researching it further, you should know that the market for the stock is not as illiquid as it may appear: there are days when no shares trade and there are days when 600,000 shares change hands; it wouldn’t take a month to buy $100,000 worth.

Disclosure: I am long Consciencefood Holding

ConscienceFood Holding

Consciencefood Holding (CSF) is a Singapore-listed Indonesian company that makes and sells  just under a billion packs of instant noodles annually in Aceh, North Sumatra, Riau, Jambi, West Sumatra and South Sumatra.

Its sales in Aceh and North Sumatra account for half of its revenues and its market share in these two northernmost provinces of Sumatra exceeds 50%. In Sumatra overall, its market share is 30%. (At 50 million, Sumatra is as populous as England and more populous than California).

Which is nice because instant noodles are an Indonesian staple: its population consumes more of it per person than any people but the Koreans. (As a point of comparison, Americans eat 46 slices of pizza per person per year).

Indonesian consumers often buy instant noodles by the carton, each carton containing 40 packs.

Whereas CSF's position in Sumatra (and Northern Sumatra, especially) is strong, it is nevertheless a midget in the overall Indonesian instant noodle marketplace:  it operates in the shadow of the two giants of the instant noodle business in the country – Indofood CBP Sukses Makmur and Wingsfood – whose market shares in Indonesia overall are ~75% and 15% respectively

Indofood CPB’s “Indomie” brand, in particular, is so popular that “Indomie” is synonymous with "instant noodles" in much of the country. In the diagram below, pulled from an analyst report on Indofood CBP, “Alhami”, CSF’s principal brand and accounting for half its sales, doesn’t even merit a mention. (CSF’s national market share is, in fact, 4.5%).

Instant noodle brand share in Indonesia.

At first glance, this situation looks ominous: what chance does CSF have in defending its market share against a goliath like Indofood CBP and at what cost? And yet, the history of the instant noodle market in Indonesia points to an altogether more optimistic outlook for CSF.

Indofood’s market share in 1999 was actually 20% higher than it is today. You need wheat flour to make instant noodles, and Indofood had a lock on wheat flour imports and wheat flour milling; no wheat flour for anyone else, no competition. Deregulation in the aftermath of the “Asian crisis” of that period, however, meant the end for Indofood’s control of wheat flour and this, in turn, allowed other firms to contest the instant noodle market.

No wheat flour, no noodles.

Now, it seems to me that there are two very different ways to enter the market. The first way, available to deep-pocketed entrants, is to launch in Java and wage a very expensive price war with Indofood, imposing serious losses on Indofood and oneself, taking market share, and then calling a truce. This is the approach that Wings Group took and the four year (2004-2008) price war between Indofood’s “Indomie” and Wings Group’s “Mie Sedaap” – a Javanese cockfight, if you like – is a celebrated episode in recent Indonesian business history. The unofficial terms of the truce? Indofoods sets the prices and Wings Food matches it. It's been that way since late 2008.

The other way, of course, is to focus on a niche: pick your geography and demographic; develop, test, market, and launch products for that demographic; build a dense distribution system (Makro, Carrefour and Hypermart; mini-marts, wet markets, and neighborhood grocery stores); and let local economies of scale and brand habit do the rest 

An analyst checking out the Carrefour in Median: the shelf space occupied by Alhami was 60 to 70 meters across. 

CSF’s instant noodle offerings are positioned as healthy and certified halal (hence “Consciencefood”), and are flavored to correspond to the range of popular north Sumatran dishes. In addition, all of its fixed costs – the production lines, the warehouses, the trucks, the TV/radio/promotional spending – are all concentrated on sales within a 300 km radius of Medan. 

This latter approach, therefore, takes market share without imposing losses on oneself since one’s average unit cost is lower than Indofood’s. In fact, one can price the noodles at 10% below Indofoods and make a fatter profit. 

To see why, one can run the numbers in full or one can take a shortcut. Consider the box of 40 packs of noodles pictured above. That box of instant noodles retails for about ~US$5.50 in local currency. How many boxes fit in a truck that can navigate Indonesian roads? How much would it cost for a truck to transport those boxes from Java to Medan and back? Multiply that problem by a 25 million boxes a year and one can appreciate the scale of the problem.

Local scale is important enough that the retail price of imported (Japanese, Korean, etc) brands of instant noodles is 10x to 30x that of Alhami, CSF’s premium product. 

Between 1999 and 2010, then, CSF had taken more than half the north Sumatran market away from Indofood, was growing at 18.6% a year, had maxed out its production capacity and, in order to extend its reach into Java, had hired a third party manufacturer to make its noodles in Jakarta.

At the same time, it had big plans for the future: it wanted to take control of manufacturing in Java; and it wanted to leverage its brand by entering the markets for cup noodles and health drinks. Cup noodles and soft drinks sell for 5x and 3x the price of a pack of instant noodles so the margin prospects are that much more attractive.  

So CSF went public in 2010, with the owner-operator Djoesianto Law selling ~45% of his stake.

Let’s pause here for a moment and consider the value of CSF’s instant noodle business.

Note: The marketing, distribution and G&A expenses are unallocated. Total corporate marketing expense in 2012 was 32.6 billion Rupiah, an 83% increase over 2011: this increase in very large part represents groundwork for the launch of CSF’s line of cup noodles and soft drinks in 2013.

Instant noodles are a cheap, fast, convenient alternative to rice. Consider Indonesia's long term per capita GDP growth rate (6%), its rate of urbanization (44%, growing at 1.7% per year), and the participation of women in its labor force (51%). Next, consider Medan's relative proximity to Kuala Lumpur and the cultural affinities between Indonesian Malays and Malaysians. Now, what's the correct multiple for this business? I'd feel lucky to buy the whole business at 8x earnings. 

Now consider how the market is pricing it: 

A P/E of less than 1.5x for a growing, branded consumer staple with some moat-like properties? What’s going on? I’ll turn to that in my next post.

Amended: June 8 2013

Disclosure: I am long Consciencefood Holdings.

Portfolio Update

I have sold out of my position in Cybergun's 8% 2016 bonds. They still offer a 19% YTM at this price.

Portfolio Update

I've reduced my position in Northgate by 40%.