We don't have evidence on that. That's a simple and clear answer: we don't know. I've seen no empirics that suggest that.
That said, I think what we would want to think about, as a research matter and thus as a policy matter, would be how to distill a pathway to violent radicalization that doesn't presume what is thought about in the literature as a conveyor belt theory, as though somehow A causes B causes C and that this leads you on a conveyor belt to violent radicalization. We have not seen good evidence of that. The Aaron Driver case is one example. We have not seen evidence of predictive power around it.
Coming to the question of how we can predict and whether we can predict based on past experience, I haven't seen that evidence.
I would say that this is about determining what the vulnerability points are and acting on those vulnerability points. I would say that if we thought of it that way, I would want us to think about two things. The first would be to work at a primary level face to face, to work at a secondary level within communities, and to work at tertiary levels with law enforcement and other organizations of the state and others.
I also think you would need to do this in ways that are individual and that also attend to people who have had no contact with that risk, people before the fact, people who may have had some contact with that risk, and people who have in fact already been either radicalized or, as we say, radicalized to violence.
We have to figure out where that threshold is. I don't think the green paper tells us that, and that's going to be a judgment call. I think a lot of attention needs to be placed there. We have to think about those things.