I've been meaning to write this for a while, but I just got blocked on Twitter by the editor of Rail magazine for pointing it out (!), so now seems like a good time. If there is some reason I am laughably wrong, now's the time to point it out.
Fairly often, when renationalisation of the railways is discussed, a neat little pie chart turns up showing some small percentage of income goes on TOC profits (here is an example: http://www.nationalrail.co.uk/static/images/structure/css/fact-about-fare-2014.jpg
- this one discusses fare income, but as far as I can make out from http://www.orr.gov.uk/__data/assets/pdf_file/0020/24149/uk-rail-industry-financial-information-2015-16.pdf
today's figure of 1.9% does reflect the distribution of all income. I don't know why Network Rail can't replace their pie chart with one based on more recent figures...)
As far as I know this is true, but what pops up next is the assertion that only that small percentage is to be saved by renationalising the railways. That seems to be totally untrue, as a bit of a peek at the other slices of the pie chart will reveal.
First of all, there's a much bigger chunk (11% in 2014, 7% now) marked "leasing trains". Do the rolling stock companies (ROSCOs), which were of course created out of British Rail, make a profit? You bet they do. Their surplus is about 20%, so there's another 1.4% right there.
Secondly, there's "interest payments and other costs". There was a bit here about how the TOCs are probably hiding some profits via (say) borrowing money from associated companies in countries with less corporation tax, but as far as I can make out all the interest payments are made by Network Rail. There is a pretence that Network Rail is not just a bit of the government, and that compels it to borrow money at a higher interest rate than the government would.
(However, the ROSCOs may well be posting an artificially low surplus, either through such tax avoidance or via the private equity practice of buying an asset with a loan secured on that asset. That would represent yet more profit that doesn't show up on the pie chart.)
Then we have staffing costs (25% of the pie chart). Fragmenting the railway has added untold layers of bureaucracy; the ROSCOs have staff to deal with leasing the trains to the TOCs and the TOCs have staff to deal with leasing the trains from the ROSCOs. The TOCs have staff to deal with Network Rail and Network Rail has staff to deal with the TOCs - a lot, because a train cannot simply be delayed now without a careful apportioning of the costs arising from that delay. A vast management tree is essentially duplicated across 20-odd TOCs (yes, it would be a bit bigger in a company the size of BR, but there wouldn't be 20 of it). It's hard to obtain any decent estimate of this (I would be intrigued to see figures on the relative number of officebound staff employed by BR and the current system, but I suspect they are well hidden) but it's hard to suppose it's too small a proportion of that 25% to show up.
So I think two things are true; the proportion of the railways' income that is lost to the structures of privatisation certainly is not 1.9% - it must be at least as high as 3.3% if we add the ROSCOs' profits in - and there is every reason to suppose it is considerably higher, even if it is hard to know exactly how much.