The only force we could identify that can explain the observation of a negative correlation between recombination rate and FST is directional selection; that is, hitchhiking linked to positively selected alleles (sweeps) or background selection linked to negatively selected alleles.
http://plosgenetics.org/article/info:doi/10.1371/journal.pgen.1000886#article1.body1.sec3.p3

If I understand this statement correctly, while positive selection (as measured by Fst) may show inverse correlation to recombination rate, positive selection itself is not shown to be inversely correlated with recombination rate. Fst has an inverse correlation with recombination rate because positive selection in regions with lower LD has the effect of raising Fst values across a wider region encompassing more SNPs. However, positive selection events are not shown to vary with recombination rate. Positive selection is just more easily detected in regions with lower LD. Unfortunately, it is much harder to detect positive selection in regions with low LD, leading to the superficial appearance that positive selection occurs more often in regions with low recombination rate.