Okay, got it. Your choice of method may well be underestimating the correlation, by degrading the raw data down to ordinal variables. Why not regress GDP in PPP dollars on the raw index scores? That's what they did in the original report. If you want to shoot down their findings for real, either use their own methodology, or give a reason why yours is superior. By using an inferior method that can artificially deflate the correlation, you are going to raise doubts among statistically literate readers about your own argument.
I say the above not to support the Heritage paper. The fact that they got the temporal precedence backwards in that table I linked shows me that they have some methodological problems of their own.
thomas