Link Prediction in Social Networks

Via unCLog I came across a paper about predicting future links in social networks -- specifically academic co-authoring. Here's one of the conclusions of the paper:

By running our predictors on some other datasets, we have discovered that performance swells dramatically as the topical focus of the dataset widens. In a narrow field, almost anyone can collaborate with anyone else, and new collaborations are largely random. It would be interesting to make precise a sense in which such new collaborations are simply not predictable from the training data

This makes me wonder whether you could come up with a number that characterizes the amount of link predictability in a given network, and then use that number to make predictions about the future of that network. For example, if the link correlation is low but increasing, maybe that means the field is a small industry that's starting to grow. On the other hand, if the correlation is high and increasing, maybe that means it's a field that's ripe for specialization. Who knows what the implications might be for directing research funding and venture capital?