Friday, October 19, 2012

Portfolio Update

I have a new position in Cybergun SA at an average cost of 2.29. 

Thursday, October 18, 2012

FreightCar America -- Railcars

FreightCarAmerica (“RAIL”) makes and sells 80% of the coal cars in the United States. Its customers are railways, power companies, and leasing firms. 

RAIL has a couple of competitive strengths: 

  • many railcar purchasers prefer to maintain a standardized fleet of railcars, making RAIL, the incumbent, the preferred supplier of coal cars; and 
  • it has the capacity to manufacture 15,000 units a year in its plants at Danville and Roanoke, meaning that, in most years, it operates below capacity. 

Competition based on design is therefore as unlikely as competition based on cost.  And, given the high cost of transporting railcars across the ocean, competition from provenances with low cost labor is not a concern  

Looking back

The business has generated an average of ~$34 million per year in operating profit on average operating assets of $96 million, for an average return on invested capital of 35% -- a fair reflection of its competitive position

18% of this operating profit was used to close down a unionized plant, a nonrecurring expense, and 45% was paid out to owners. 15% of the operating profit was reinvested for growth.

The story in railcar units:

Looking forward

Core business – replacement coal cars

There are currently 270,000 coal cars in operation in the US and 3.7% of them – 9,950  – have needed to be to be replaced in an average year, 7,950 of which have been manufactured by RAIL.

At an average selling price $75,000 and an average gross margin of 11.75%, this recurring, though lumpy, business of replacing old coal cars has generated $590 million in gross profit since 2005. 

The coming business cycle calls for approx. 79,000 coal cars – those aged 30 or more years old today – to be replaced. Moreover, the coal cars to be replaced are made of steel and need to be exchanged for the more efficient aluminum cars which are RAIL’s specialty. The average replacement rate, therefore, is going to be comparable to the past eight years: 9,875 cars per year, perhaps 85% of which will be supplied by RAIL at an average price of $80,000 per unit. 

After overhead expenses and taxes (net of the NOL carryforward), the average annual operating profit from this core business is therefore going to approximate $36 million a year. 


In order, it says, to smooth revenues across the business cycle, RAIL entered into the railcar leasing business in 2008. Setting aside the long-term wisdom of competing with its customers, the minimum revenue schedule over the next few years from this segment is knowable – about $5 million per year. 


RAIL has also, via acquisitions, strengthened its commitment to providing railcar maintenance services in the busiest rail traffic corridors in the US. We can expect this segment to contribute another $4 million per year in after-tax profit.

RAIL’s total minimum earning capacity over the next business cycle, is therefore, about $45 million per year ($3.78 per share). Bearing in mind that RAIL pays out what it doesn’t reinvest for growth, we can project that, in the absence of the an incident like the closing of the Johnstown plant, it will pay out its excess cash and at least 75% of its operating profit, for an average of $3.10 per share per year, which is not bad for a stock priced at $18 and change. 

More conventionally:

It's a wildly cyclical stock, so everyone tries to time it. Some brokers even have a sell rating on it. The only real risk with this stock -- the only one that I can see --  is that it might be acquired at $25 or so by an outfit like Berkshire. The concerns over the near-term death of coal are founded more on hope than fact. 

Disclosure: No position 

Edited for typos on Oct 19th

Thursday, October 11, 2012

Portfolio Review

Three months and a few dozen page views into the life of this blog, the tracking portfolio fashionably underwater, it may well be time to do a little review of who’s where and what’s what.

In adopting the concept of look-through earnings , I’ve come up with this: 

If the portfolio were a business (a conglomerate, say), it would, in July of this year, have been a competitive firm earning  20% on its net operating assets, yielding 15%, growing at 13%, and getting better as it grows – that is, earning a greater return on every new dollar of investment than it has in the past.  

This set of circumstances is what I look for and what suits me best: moat + growth + incremental quality mean that intrinsic value is accelerating away from my purchase price, which in turn means that time is on my side. To use a familiar expression, the market can close down for a few years and I can be confident that I will have earned at least the yield, 15%, and probably a great deal more, by the time it reopens.

Since I never* pay for growth, the tricky part has always been assessing whether that which I do pay for, earning power, has legs. And figuring that out, as the owners of Hewlett-Packard, Radioshack, French Connection, Carrefour, and countless others can testify, involves more than looking at historical returns on net operating assets.  

That’s why I like to focus first on the business model – why does this business exist? How and why does it earn its handsome returns on capital? How much would it cost to break it, or to relegate it to obsolescence? Who does it compete with? Where is it in the value chain? And so on. The business model usually tells me what I can expect to see in terms of numbers, and if I like the business model, and the numbers affirm what the business model says they should be, then I look at price.

The no doubt irritating habit I have of writing about the business first, dealing with the financials second, and turning to valuation last, is a mirror of the process that I follow in mulling over the stock in the first place. In fact, I “screen” for candidate stocks by looking laterally within business model types: like F.W. Thorpe and Gerard Lighting; HSN and CJO Shopping Co.; Dunn and Bradstreet and Nice Information Service; Matchtech and Harvey Nash; Electrocomponents and Arrow Electronics; Northbridge and Speedy Hire; Cegid and Cegedim; etc. I don’t screen by price or cheapness.

The margin of safety is in the strength of the business model of the stock that I'm looking at, not in the price tout simple. If the business model is, for example, to buy equipment and rent it out, in whole or in part – Norhbridge, Aircastle, Carnival, Silver Chef, Hawaiian Holdings, Marriott – there had better be a bloody good reason as to why anyone and everyone else couldn’t do exactly the same thing, driving returns down to the cost of capital and below. If there isn’t any good reason why not, thinking that the business is worth twice its operating assets is, in my view, a mistake, and not one that is remedied by buying it at 50% “margin of safety”. One may wait years for the stock to “mean revert” and in that time, the opportunity cost is being metered and charged.

It is this belief – that the margin of safety is in the business model – that leads to me the kinds of stocks that populate the portfolio in its current and future iterations. These are companies with easy-to-understand ways of turning a profit, operating in small, stable, well-defined markets, and earning returns appropriate to the strength of their (local) competitive advantage. Each of them is in a position to grow; the avenues for growth are well-defined, specified in advance; and growth is likely to be value-enhancing because their business models scale well, stretching fixed or sunk costs over a larger revenue base.

And, most importantly, I am not at an informational disadvantage. Small as these stocks are**, no one is going to spend an awful lot of money in an information arms war to gain a slight edge over other market participants. iSuppli doesn't cover Lectra, Tessi, Haynes, or James Halstead. If the price is low, it’s most likely just low, not priced for risk.

There are, I know, retail investors who like taking the other side of the trade on well-watched stocks like Apple, and opaque stocks like Citi, situations in which they know that they are likely at a severe informational disadvantage (in terms of both the known unknowns and the unknown unknowns) armed with the belief (perhaps justified, perhaps true, but nevertheless suffering from a Gettier problem) that professional investors are stupid or myopic or whatever else. I’m not one of them and that kind of trade is not my bag. I like to know at least as much about a stock as any non-insider. Otherwise, and this is a personal thing – necessarily so – I’d feel as silly as if I opened a neighborhood grocery store in the shadow of a Walmart.

In any case, back to the tracking portfolio. It now looks like this:

It’s gotten, on average, both cheaper and better, now earning  22% on its net operating assets, still growing at 13%, and getting even better as it grows. The difference is that it is now yielding almost twice what it was -- 29% rather than 15%.

The area for concern is CEGID. It is now both a bit more expensive and a bit lower quality than it was then.  It's also more complicated than the others, with several different product lines, markets, demand dynamics, and so on. If WEB shorted it, I might wobble, which is something I wouldn't say of the other components of the tracking portfolio. On the other hand, I still think that the investment thesis is intact, that it is worth 28 or so, and that, cyclicality aside, its underlying fundamentals are improving. But I now wonder if there aren't easier ways to earn the same returns. That I'm a little overexposed to France fans the flames of that concern; that I've held it for just shy of a year, only half as long my preferred holding period, dampens them. I dunno. I'm thinking out loud.

The others are fine, their fundamentals behaving more or less as advertised, although Value Investing France, in an excellent blog post, sees GEA differently than I do. 

Disclosure: I have positions in XPO, Hawaiian, Cegid, Gea and Precia.

* I do, however, seem to have paid, or at any rate come close to paying, for some of XPO's future growth. A rush of blood to the head that was a mistakeand that hopefully won't recur.

 ** The sharp-eyed reader will have noted that Hawaiian is hardly small; it is a special case in that every bit of useful information is made available via the DOT/BTS. It is possible, with a bit of work, to reconstruct the unit economics of each route, even each flight, using publicly available information. I have done that work.

Wednesday, October 10, 2012

Instem Plc

Instem provides software applications to the early development drug and chemical R&D market. Its clients use these applications to collect, analyze and report complex scientific data; comply with regulatory reporting requirements; improve quality, consistency and efficiency of information reporting; and to reduce the time of critical path R&D activities.
Its principal product is Provantis, a suites of modules for managing and recording Early Development Safety Assessment (EDSA)  studies, from receipt of the compound through to the automated assembly of statistical analyses and final reports, allowing scientists to collect, analyze and share data, whether at one central site or remotely. Provantis promises to make their work easier, faster, and less prone to error. Remote deployment of Provantis is made possible by secure centralized data centers in Shanghai and the US;  Instem has partners that maintain the system 24 hours a day, 365 days a year.

The EDSA market is a modestly-sized, and Instem has been, for more than ten years now, the market leading provider of information solutions for that market. Half of the world's pre-clinical drug safety data has been collected over the last 20 years via its software; it has over 9,500 users of its solutions and supplies over 130 customers, including sixteen of the world’s top twenty pharmaceutical and biopharmaceutical companies. It has two competitors – Xybion Medical Systems and Pathology Data Systems in the European and North American markets – but Instem’s user base is approximately double the size of these two combined.It has an average client relationship exceeding 10 years and a 95% customer retention rate. Approximately 65% of its  revenues are recurring.  

In any case, Instem has also been increasing its penetration in the established Japanese market as well as in emerging other-Asian markets, particularly China. Instem reports that it won the overwhelming majority of new EDSA business placed worldwide in 2011 and that it has secured a record evel of new customers for Provantis. The most recent upgrade of the Provantis offering, in 2012, places Instem in a position to capture an even greater share of the EDSA market and to expand into other, closely related, markets such as safety pharmacology studies, drug metabolism and pharmacokinetics studies.

Its other main product line is Centrus an application which grabs legacy scientific research data, transforms it into an easy-to-use format, facilitates communication between the "discovery" & "clinical" functions in a drug development value-chain; and produces SEND datasets in a few simple steps (SEND is the regulatory standard for the transmission of research data). The Centrus offering was strengthened in 2011 by Instem's acquisition of  BioWisdom, an app for extracting intelligence from R&D related healthcare data. As it stands, Centrus is like CRM software, only for raw scientific data.

Centrus is making its market -- there were no competing products when it was launched some 5 years or so ago,  and there are none today. Instem "spotted a gap in the market", as Alan Sugar would say, and they have developed the product and the market from scratch.

In any case, one can, from the above, more or less divine what the numbers are going to look like: 

As laboratory working practices change -- the growing prominence of multi-site working and outsourcing to contract research organizations and so on, as well as the adoption of new standards – Provantis and Centrus are likely to benefit.

Instem is a good business with a 7% CAGR since its AIM listing. It shouldn't be yielding 15%. Still, there is some cyclicality to this business, especially in the awarding of new business, and a conservative 9% hurdle rate is probably called for, easily placing the value of its shares at more than 2x the current price.

Disclosure: No position

PS: The trick to valuing this stock is to normalize properly. It is a cyclical business, so taking the average NOPAT margin of 17.84% and applying it to this year's revenue will come close to specifying the current earning capacity of the business, £1.93 million. Discounting at 9% (or applying an 11x multiple) gets you to a no growth earning power value of £21.39  million. Screens can't easily capture this: to them it looks like Instem is currently trading at 11x earnings, as it should be.