Or maybe they're being deliberately misleading. Back in May, Sam wrote a barnstormer in The Telegraph refuting an IMF report that called it a 'subsidy' when the government didn't tax oil and gas as much as the IMF would have liked it to. This report was important because the IMF is taken seriously, especially by those who don't usually find themselves on its side, when it says stuff they like ('even the IMF' etc). Sam made the obvious point that a subsidy is a subsidy and not taxing something is never a subsidy, even if you think we should—even if your model says that it should—tax that thing more to increase social welfare. If it were a subsidy to tax something (like fossil fuels) below the optimal amount, then it would also be a tax not to subsidise something (like basic research) the optimal amount. But it's not; it's patently, blatantly, ridiculous and obfuscatory use of language.

We might have thought the issue was put to bed. But here we go again, this time with the Overseas Development Institute's new report, which tells us:

While other nations have responded to the drop in energy prices by reducing fossil fuel consumer subsidies, the UK has reduced taxes on fossil fuel production, increasing subsidies to fossil fuel producers.

The industry pays £30.1bn in total taxes, including downstream levies like fuel duties, but since this is not £35bn or £40bn, it's being 'subsidised' according to the IMF and ODI.

It gets onto actual things we might call 'subsidies' later, and they are as piffling as you might expect, given how they are buried:

The International Energy Agency (IEA) estimates for budgetary support for R&D to all fossil fuels in the UK were $76 million in 2013

Note the units.

In addition, the UK government has committed to providing significant direct support for the development of CCS. The largest commitment is for the $1.6 billion for the Commercialisation Programme, although this has not yet been disbursed.

OK, this is a real subsidy, and of a significant size (but nothing compared to oil and gas tax contributions!), but you might note that (a) it's not been paid yet; and (b) it's specifically tied to abating emissions!

So this is a non-story. The real story is 'the energy industry pays a large amount in tax, slightly lower this year, but probably still enough to cover its externalities, and even if it wasn't it's certainly not a "subsidy"'.

And the ODI are as stupid as the IMF.

UPDATE: Turns out the WTO uses the 'tax break = subsidy' definition too; it's still not clear this is a useful or meaningful definition—it requires an idea of the relevant baseline.