Maya Forstater

God (who sounds a bit like @ali_readhead ) would tell us that all depends on commodity prices, costs etc… Lets assume there is profit shifting. Say the management fee should be cut in half and interest reduced by 3 percentage points. (Disclaimer: I just making these numbers up)

If we make these adjustments it raises pre tax profit by 70% over the lifecycle of the mine (i.e. if OO base case represents what the company declared and the adjusted case represents the true ALP it means 60% of profits were being shifted) (6/)

Which is why I don’t think its all that useful a question – because it turns the answer that the Transfer Pricing Deity is 100% certain that the company is shifting 60% of its profits away into the statement that there is 4% chance that any given loss year is not a real loss year

(.... obviously these calculations will change if we change the commodity price for the OT model, or if we look at a different mine with different costs of production....) (/9)

What I think it shows is that the intuition that loss years are a good indicator sign of profit shifting does not hold up (although as @hselftax says several years of losses demand explanation) (10/10)

Great answer! So counting binaries misses ways of deflating profits whilst holding the pattern of 0 or 1s over the years fixed. I'm totally convinced counting false or true loss years is a poor measure and I hadn't thought of that, so thanks

However, the tweet I initially responded to said (something like) 'I agree that when a mine reports a loss it's not always fake' so the question how often fake remains, even if a poor way of assessing the overall extent of profit shifting, as you have explained.