RAIL makes specialized aluminum railroad cars, mainly used to transport coal, and has done since 1901. How picturesque, pre-internet and manly can you get? There is little competition, partly because all of its competitors have died or merged and partly because the cars are too heavy to import cheaply.

The major issue in valuation, and what makes this an interesting case is the cyclicity of demand. We might even use this as a proxy case before tackling something similar but more meaty, like the house builders. The 10-Ksets out clearly the drivers but the easiest number to latch onto is the new rail-car backlog. At the end of 2005 it was over 20,000 and right now it is over 12,000. The worst year in the most recent cycle was 2002 when it was 1,000.

Like William James's universe, Liberty Capital turns out to be a turtle resting on an infinite array of other turtles. The bits of Liberty it tracks are various, if not quite infinite. We start at the bottom, its impressive holdings in other traded securities. These are valued at $18.3 billion and include big chunks of Time Warner (TWX), Sprint Nextel (S) and Motorola (MOT).

This glob of Liberty was spun off in July 2005, although Malone owns 5.5% of the outstanding shares [regrettably controlling 31% of the votes], and the two elements have wide-ranging service and tax sharing agreements.

Discovery wholly owns two business that make no money and $203m cash. It is also in the fortunate position of controlling 50% of Discovery Communications. This media titan runs the Discovery Channel, TLC, Animal Planet, BBC America and many other properties around the world, although three quarters of revenues are earned in the US. The Discovery channels earn over three quarter of a billion dollars a year in operating cash flow. Hurray! Boo! There is the pesky nuisance of depreciation [with capital expenditures more than keeping pace], interest payments [on $2.6 billion indebtedness at the associate level] and tax. Figuring all this out, and dividing by two since Discovery Holding owns only half, leaves at best $128m attributable to this slice.

Liberty Global (LBTYA) spun off from the rest of the group in June 2004 but, as we indicated last time, it is under a great degree of common management and ownership still. It operates cable, broadband and programming business in seventeen countries, principally in Europe but also in far flung Japan, Chile and elsewhere. It is a so called triple play supplier but has so far avoided mobile - the potential next move?

Liberty Global (LBTYA)
Discovery Holding (DISCA)
Liberty Media Capital (LCAPA)
Liberty Media Interactive (LINTA)

Sadly, it hasn't panned out that way. CTAC has been hampered by a lack of product, as the major lens manufacturers are not keen to supply it and annoy the retail opticians who are their main customers (and CTAC's high cost competitors). The main manufacturers are Johnson & Johnson (JNJ), Novartis (NVS), Bausch & Lomb (BOL) and Cooper (COO). There are also new legal restrictions that make it more difficult to supply lenses without a prescribing relationship. Moreover, CTAC's results have been dragged down by huge losses at its ClearLab manufacturing subsidiary.

The up news is that CTAC has decided to part company with ClearLab in a way that has not yet been decided but which will presumably stem the red ink. It is therefore possible to see if there is value in a sum-of-the-parts exercise. Best case is that they get rid of ClearLab at no cost, (which is optimistic). The remaining operations have an operating profit of $22m, forecast by the company to grow to $23-26m next year. Deducting interest and tax nets the common shareholders c.$14.6m or or $1.09 eps. So the best case p/e is 14.8 on next year's clean earnings.

Here's a useful table:

Microsoft makes all of its money from its operating system and office tools (actually over 100% since its entertainment biz is not yet profitable). These both benefit from the celebrated network effect: for every product sold all existing customers benefit - the more Windows PCs there are, the more programs will be developed for it.

Google makes all of its money from search related advertising (actually over 100% since it gives away lots of software, including the hosting of many blogs, in the hope of revenues to come). But search is not a beneficiary of the network effect and neither are the advertising programs Google inserts. Your search is not easier because other people have searched before you. The amount advertisers pay per keyword is, if anything, higher because of Google's dominance. The service is used because Google built a better mousetrap. But another, better algorithm is just around the corner. Ask.com (IACI) could be to Google as Google was to Yahoo (YHOO), for example. And as Yahoo was to AOL (TWX). Google has not developed sticky applications or given up so much margin that it has economies of scale. The reverse is true. Google's margins are so high, it is easy to see how smaller rivals could come in to its markets. And we would all switch again. Why wouldn't we?

1. How big is the world advertising market? This should have been the easiest question to answer but, of course, all estimates vary. Our number is $650 billion, a blend of what the largest advertising agencies and investment banks say. Fortunately, this is the least sensitive number in the analysis since we are more interested in market share. Our forecast leaves this number constant, since there is no inflation and we are not assuming advertising itself grows as a sector relative to others;

2. What is the internet share of this market? The blended guess at this point is that internet advertising is around $23 billion or 3.5%. Our assumption is that this could peak at 25%. The argument is a familiar one, but convincing - advertising is a hit-and-miss business which could do with targetting and the internet is excellent and virtually unprecedented in its ability to deliver this (the advertiser knows far more about the specifics of its viewers than they would on TV, radio or even direct mail).

Much recent comment has focussed on the Netflix rivalry but this really misses the point, which is this: there will be no physical home video rental market in ten years time in all of Blockbuster's principal markets. These are 100% digital media; in other words, this is bit shifting and this will all done via tubes or wirelessly more cheaply and faster and without getting off your couch. No need for Blockbuster (and bad news for Netflix fans too, we think).

RHD is one of the largest Yellow Pages publishers in the US. It is a fantastically high margin (44% cash operating profit), high repeat media format.

USG was a victim of the great asbestos clambake of the last decades. Plaintiff lawyers across the States, claiming en masse for peritoneal mesothelioma and the like, shook the company into bankruptcy. It emerged this past June owned by assorted parties but with a contingent damages liability capped at $3.05 billion and a friend in Warren Buffett, who underwrote a massive rights issue.

Annualized EBIT is about c.$1.2 billion and the net earnings per share will depend on whether the contingent liability is called. This itself depends on legislation passing in the newly Democratic Congress.

I will use Equity Residential (EQR) as an example only. A quick look at other REITs suggests a similar story. EQR is big and widely held. Fidelity, for example, owns 7.8% and Vanguard 5.7% so its probably well represented in a portfolio near you.

The first surprise is the yield - it is, against expectations, only 3.5%. Even grossed up for tax that is not automatically compelling. But the real problem needs some deep digging. Let's have a look at the latest 10-K, for the year to December 2005.