It's exactly the kind of analysis you need to go through, because there are costs and there are benefits. There are benefits of them taking on the risks of cost overruns, of them taking on the risks of design flaws, of them taking on the risks of catastrophic issues or operating problems, all of those risks that in another way would be the government's. You have to do a risk evaluation of those kinds of things.
Against that you have to put the cost, because there will be a premium to the private sector finance. If they're going to take on risk, they're going to expect return. The question in a P3 is to allocate the risk to those who are best able to manage it. In that way, you have to do an evaluation. We do what we call a value for money evaluation on what is the right model on every single project that we would consider. We would not do it unless our analysis, which is very rigorous, concludes that it's the right thing. That's why I said that only in about 15% of the cases do you conclude that.
Are there rules of thumb? We will do a deep dive, but are there rules of thumb where we say that we wouldn't even do a deep analysis on one, but on another one, probably. You need to have something that has scale. It has to be $50 million to $100 million, depending on the sector. Why is that? You need to attract the private sector interest. If the private sector is going to bid on one of these things, that means they need to do design work. They have to invest millions of dollars in the bid process in order to think through how they will handle this, to get that innovation—