Is (expensive) data that is 'soft' of any more use than no data? Honest question. My answer would be that it most likely isn't.
Fern
You're begging the question of whether the data is too "soft" to be of use for policy reasons. The methodology used is rather abstract. It doesn't try to define a "job" as green or not. Rather, it defines industries as green or not, and tracks jobs in those industries. The idea of parsing an infinite array of different jobs as green or not green is definitionally problematic.
This method obviously has the shortcoming of over-counting in some cases and under counting in others. For example, a floor sweeper at a solar company is a green job. An environmental compliance manager at an oil refinery is not. The theory is that the over and under counting roughly cancel each other out. That is the theory behind any statistical counting method, actually.
We don't need 100% accuracy for such a system to be of use for policy reasons. A reasonable approximation will probably due. If you used this method repeatedly over time, it will show a proper vector (trend line) because whatever error exists in the method will be consistent over time. If there's a more accurate system that is cost-efficient, then that's going to be preferred. I just don't know that there is. I also don't really know how close an approximation this system provides, and neither do you.
Another honest question: What is the purpose of developing this data? What is it used for?
If it's to be used for some political purpose such as claiming that Green Jobs have seen great growth under Obama then it's bogus.
Ostensibly, the purpose was for developing policy. I say ostensibly, because I can't rule out a political purpose either in addition to, or instead or, a policy purpose.
If it's to develop national policy then I think it's rather useless as it's currently operated. I think we should narrow the focus, a lot. If we're tracking green industry then I'd recommend we develop a list of the specific industries we're actually interested in, such as PV and other solar powered electrical generation, bio-fuels etc. Right now I think we're looking at GIGO.
Except the system does, in fact, track these individual industries. And I'm willing to bet that this data is even available to the public. It's how the BLS operates. For example, everyone talks about the total month job numbers, but I guess most people don't realize that there is a report every month which contains those numbers by sector. With "green jobs," they in fact are tracking them by first identifying particular industries and applying codes to them, then compiling a total number from all of this data. If we want to know how many jobs we have in solar, for example, I'll bet you can find that data in the same report as the one offering the "total green jobs" number.
It may well be, BTW, that the "total green jobs" number is essentially meaningless except as a political data point, whereas the useful information is in the specific sub-categories. Nonetheless, the total number is compiled from industry sub-totals.
- wolf