"At 3.5% of 2009 GDP it is a cost that is big enough to raise serious questions about whether taxpayers could have been better protected."
You're just plain wrong mate. Taxpayers don't need to be protected from profitable ventures. I dunno what the hell definition for profit you're making up, but it's literally defined on the basis of cost in accounting and finance.
Protection means "structuring the bailout differently to reduce the downside risk".
Section 2.1.1 - "For a bailout cost measure to be economically meaningful, it has to be evaluated as of a fixed point in time. In most cases, the natural choice is the year the bailout is initiated, for instance, when new legislation is passed or administrative policy changes are announced or implemented, or shortly thereafter."
The author does this. She evaluates the cost of the bailout at a fixed point in time, the year the bailout is initiated, which is to say 2008. She arrives at a number of $500bn.
What happens after 2008 is irrelevant to her cost analysis. If you don't understand this, you don't understand anything about the paper at all. Look at every subsection in section 3 where the author considers the different elements of the bailout. When she's considering fair value cost, it's always in reference to contemporaneous reports (2008/2009 sources) or estimates on that basis.
Also, please stop accusing me of gaslighting: I'm not asking you to question your reality, I'm just asking you to question your understanding of an economics paper - there's no power dynamic here that would put you into a vulnerable position and allow me to bully you, even if that were my intent. We're just people on the internet.
Quoting the author:
In fact because it is most likely that a recession will be followed by a recovery, it is probable that the government will show a “profit.” However, bailouts are costly because of the possibility of relatively unlikely but very costly states of the world where recessions persist and recoveries are low.
As I think we can agree, it was not the case the 2008 recession was persistent with low recovery; yet it is the possibility this could have been the case that increases the cost.
Cost analysis is absolutely orthogonal to an actual accounting of profit and loss, which is why the author consistently uses quotes around the word "profit". The cost of doing a thing has nothing at all do with whether it turns out to be profitable! Profitability is exactly a matter of ex post cash accounting; there is just no other way to measure it.
I guarantee you have never quoted a section of the paper that says the bailouts were unprofitable to the taxpayer because the author never makes that claim: because it would be false. You will also never find a reference to "cost accounting" in the paper, again because those concepts are orthogonal within the author's framework. By all means double-check the paper on both those points.
If you're trying to answer the question "did the tax payers get back more money than they put in to the bailout", then the only way to do that is by ex post analysis of cash flows: the answer is "yes they did". If this were a business, that would be the definition of a profitable investment.
Until you understand that costs are nothing to do with profitability, you're doomed to misunderstand this paper.