 South Africa is from the industry point of view. Same as in 1993, there's just more of it. The great economist Nobel Prize winning economist, Bob Solo, once said, economists use the word structure to indicate they're talking about something deep, but they're not quite sure what it is. We want to be a bit more precise about this, and we focus much more on structure as the technical relations between industries and how those have developed, rather than simply the shares of industries in production, which is the way it's normally looked at. There are two key things in economic development. One is shifting resources from less productive into more productive activities, and I think that's crucial and everyone recognizes that. But the second thing is you want those more productive activities to be well linked to the economy, so that any benefits they have spill over into the rest of the economy. So a large part of the structure we were looking at was that those linkages have those changed in ways that make the spillover stronger, and we find that they didn't really change as much as that. I'm a firm believer that we can't explore things until we've described them. It's perhaps an iterative process. We describe a bit and then we try to explore them and we realize we need a deeper description or some other aspects. When we change a policy or when there's a commodity boom, how does that feed into the economy? We know it has some beneficial effect, but which sectors is benefiting, the linkages between those sectors, how do they spread? We spend a lot of time building a consistent dataset so that we know when we're comparing the past with the present, we're comparing like with like. We spend a lot of time in that. So now we've got a dataset which allows us to describe what the changes have been. Now, the next step is not for us. There's a public dataset. It's available through WIDA. We can encourage other people to say, can we find out why these changes have happened? How can we change them? What can policy makers do to get the kinds of changes they want in the economy? And data is not something which falls out of the sky. Data is something that's constructed and it takes hard work to do. And that's what we've been doing.