As AlterEgo said, it is a very long report and it is difficult to check the facts without being provided with extra detail. The big "extra costs" of a privatised railway according to the report (on p.18) are:
The excess interest payments by NR were due to the fact that it was not classified as part of the public sector - I believe it is now and so this should not be an issue anymore.
No detail is given here about why this would be significantly cheaper in a nationalised (and presumably sectorised) system. They give a single example (delay attribution) which has been discussed on here extensively and most people agree that a nationalised system would still need a considerable amount of people involved in this process. The report however assumes that a nationalised system would have none of these costs at all.
The report assumes that the saving would be equal to the profit margin of the contractors and does not mention any possible benefit from outsourcing.
Similarly, the report assumes that nationalised TOCs would be just as efficient and therefore their profit margins would be recouped by the Treasury, without mentioning any benefit from the profit motive for the TOCs.
The report does not mention the cost of buying the assets of ROSCOs at all, which would probably wipe out the benefit of not paying out dividends
Debt
The debt issue has resolved itself going forward and Network Rail can, in theory, fund itself through Treasury debt at a negative interest rate of 10 year debt. The historical debt is still issued and valid, and until it has been redeemed and replaced with proper UK Government debt, not the slightly wishy-washy Network Rail Government backed debt, they'll be stuck paying the higher interest rate cost.
TOC vs NR costs
Delay Attribution is ripe for major overhaul - if you take the profit and loss motive out of the system and use it purely to focus on weaknesses within the railway system, you would need maybe a quarter of the staff and the end result isn't a money-go-round, but properly identified problems on the network and ways to resolve them which can reduce delay repay payments.
It might be the case that a certain class of train is more vulnerable to damage (using the bird example) and instead of money going round to cover the failure, the issue is identified and it's decided to modify the rolling stock instead. The type of small engineering modification which might not make sense when NR will cover the cost of the failure, but when the focus on solely on stopping it happening, and we're not hamstrung by short franchises, it's the type of thing we would see happening.
There's also things like station maintenance, differences in ticketing systems, and procurement costs. If you were to make Network Rail responsible for all of that, there's big procurement cost savings to be made. It also saves the eternal rebranding vinyl and paint game, and thinking about winter, when stations are managed by different TOCs who do things like salting platforms and pavements, two and three teams from different TOCs can be doing work within a short distance of each other. Duplication and triplication of effort.
Controllers and relationship managing between NR and TOCs is another area ripe for cutting - do you really think we need London Overground, Virgin Trains and London Midland all with their own teams liasing with Network Rail around the lower WCML and Euston ? They've all got their own staff at Euston too, their own ticket machines and passengers get passed from pillar to post if they have issues with another operators ticket machines.
Outsourcing
There's nothing wrong with outsourcing if it's a cheaper way to run a business or service, or they can do something you can't do in house. The core task of a railway is to run a railway, not paint stations, clean windows or fix computers - there's nothing wrong with outsourcing some roles to dedicated, specialist contractors, saving money and getting a better result for the £ you spend.
Profit
The profit motive is valid, but there's a lot of easy money to be made on the railway that's just too damn complicated and risky right now.
Just think, a proper joined up railway could have Gatwick Express or Great Northern Class 387 stock borrowed on a weekend by Western region for a rugby match at Cardiff, running along the GWML out of Paddington, instead of sat on a depot being unused. British Rail made a good profit just by focusing on moving as many passengers as they could as efficiently as they could.
I don't know if a nationalised railway would be as profitable per passenger, but I see no reason why it wouldn't be more profitable all in.
Rolling stock
Buying out the ROSCOs with cheap government debt and putting the dividends back into the system to cover the borrowing costs isn't an issue, just set it up as a not for profit company.
