genetics - effect of background selection on promoter regions compared to distant enhancers?

Has anyone looked at the effect of background selection on the levels of conservation of promoter regions compared to distant enhancers? Do promoter regions have a higher conservation due to background selection from nearby genes compared to more distant enhancer regions?

My attempts to google for this only gave me a 2004 Drosophila paper on a few genes survey:

http://mbe.oxfordjournals.org/content/21/2/374.full

I am wondering if this has been looked at before or if there are any other relevant bits of genetics one should consider first.

EDIT: a bit more of explanation into the hypothesis below, and a reference that could be useful:

Let's start supposing a region near a coding gene that can gain regulatory potential via a new point mutation or small translocation. This new regulatory potential needs to be positively selected after it appears, or it will gradually accumulate more mutations and disappear. But if the region is close enough to the coding gene, recombination rate between the region and the nearby gene body will be low, proportional to their distance. If the gene body is assumed under selective constraint, there will be a hitchhiking effect of the nearby regulatory region by the gene nearby. The background levels of mutation on the regulatory region are then going to be lower (lower background mutation) than in a region that is not near a gene under strong selective constraint. So one way of explaining why close promoter/enhancer regions are more conserved than distant enhancer regions would simply be due to linkage.

A second part of the hypothesis would be about the presence or absence of recombination hotspots in upstream regions of a coding region: if nearby promoter/enhancers are kept via linkage to the closest gene, unbiased measures of cross-over (e.g. PRDM9 experiments in mouse testis) should indicate if there is an increased linkage of functional genomic blocks, like enhancer/promoter/gene blocks (Hi-C, 4C, 3C data), whereas if recombination doesn't play a role, regulatory blocks wouldn't necessarily be kept in the same haplotype block.