No, it's better to look for the marketista-disproving explanation in what ThatBritGuy said. The rise in passenger numbers was generated by private railways that lowered fares and raised capacity (including a rather significant wave of new vehicle purchases) while spending nothing on the infrastructure, e.g. all the difficult stuff was externalised from their viewpoint, and this (alongside infrastructure policies that aggravated it) led directly to the Railtrack financial disaster.

Are you forgetting the RER?

No, it is accounted for in SNCF's statistics.

Signaling is the one thing I'd expect to not get cost overruns, assuming it's been tested

That's the point: it wasn't tested. They thought wireless moving-block signalling (which will become ERTMS Level 3) can be implemented easily. But even today, it is still far away. What's more, the wireless fixed-block ERTMS Level 2 system, despite tests, caused cost overruns in multiple countries due to failure to achieve reasonably trouble-free operation (two Swiss lines, the Madrid-Barcelona high-speed line, Italian high-speed lines).

Switzerland may be a freight transit country, but it is also the country with the highest per capita train travel, so your point actually points in the opposite direction: the Swiss railway infrastructure achieved great reliability for dense passenger traffic despite a heavy freight traffic load.

Separation of infrastructure and operations is 'easy to understand' for simpleton marketista politicians, but the basic fact is, railways are not easy to understand, they are among the most complex systems operated in the economy. Railways aren't like roads, each of the planning, maintenance, and network coordination of fixed infrastucture and rolling stock is closely interwinded. I refer you back to the British example. Also, new freight terminals has nothing to do with separating passenger and freight business.

Per capita rail infrastructure investment figures touted around are difficult to interpret: on one hand, they depend on what works are counted as investment and what 'only' as maintenance, on the other hand, they often reflect a few megaprojects. Plus, it's one thing to build new lines and another to make up for decades of lack of proper maintenance. But, off-hand I give a figure of currently just above €100 per capita for Sweden, after a strong upward tendency (Botniabanan, Malmö and now Stockholm city tunnels boosted it), somewhat higher for Britain and Italy, and significantly lower for France and Germany after downward trends. (Though note, I'm not sure about the level of non-RFF rail infrastructure investments, expecially by regions.)

Regarding EU open access policy, part of it is already in force, there were the initial laws on institutional separation (executed in France by the creation of RFF in 1997), then the first (2001) and second (2004) railway packages that liberalised international traffic. The third package, with the thorny question of open-access in passenger service and domestic services, is now in second reading, and back when the Council formed its common position, France & allies only blocked the domestic liberalisation part (though that's the most important). Just a few days ago, while most of the Third Package (including international passenger service liberalisation from 2010) passed the EP vote, the joint Commission-relevant EP Committee proposal on internal liberalisation failed (narrowly, due to absent MEPs...) to get qualified majority.

_{*Lunatic*, n.
One whose delusions are out of fashion.}