---

Welfare Reform Bill

I thank noble Lords for their support. At this stage we are taking legislative powers in order to be able to do this. How it is done is something that we will actively develop. I will tell noble Lords what I think we should be doing without necessarily locking down that that is to be the process, because we have not developed it. Universal credit is the most amazing social science laboratory that I suspect we have ever seen, and I wonder how many other people will see it. Under universal credit you can change different aspects of people’s support. As such, it needs a unit built in which is constantly looking at how to improve it and optimise it or to adapt it to different circumstances. I anticipate, in answer to the question from the noble Baroness, Lady Hayter, that we would have a series of real questions. Many of the questions raised by noble Lords in the Committee—should we have a second earner disregard; should we have a lower taper; what happens when you move disregards up or down?—are real, basic questions. They are all being put in the form of amendments, but here, we can have a series of tests of different aspects, or tests in combination, to find out what really optimises the system. Clearly, it is impossible to get it absolutely right first time. No one would claim to do that, but this is an architecture which would allow us to optimise it. Transparency is important. One provision in the amendment is that we go to Parliament to ask for permission to pass the regulations. That is a transparent process; it is not hidden. We would have to explain what we are doing and why. I cannot imagine that that information would not keep coming back to Parliament; I do not know in exactly what forum. To that extent, I am sure that there would be a real process. On the question raised by the noble Baroness, Lady Hayter, about ““replace and extend””, there is a catchphrase that the DWP has more pilots than Pan Am, it used to be; it is now probably Ryanair. One reason for that is that when you have run a pilot and then want to introduce it universally because it worked, you have to pretend, rather than stopping a pilot and starting again, that you were continuing to pilot it.