This issue has been studied extensively. There are no significant pharmacokinetic interactions between BZD and ethanol, so drinking doesn't increase BZD blood levels (source: see the citations below)
The really interesting (an potentially counterintuitive) thing is that BZDs and ethanol produce only additive effects and they do not act synergistically or potentiate each other. Experiments have consistently shown that BZD drugs and ethanol DO NOT produce supra-additive effects (i.e., 2+2=5). Many people are under the impression that they are supra-additive because you can get very intoxicated by using BZDs and ethanol in combination, but it turns out that it is just the sum of the individual effects of the BZD and ethanol. Here are a few studies (Pubmed ID#s):
6140699
6138055
8799523
7870919
2201724
6142906
There are certainly other studies that have examined this issue, but I don't remember seeing any other reports that reached the opposite conclusion.
It may be that when someone is drunk, it is much easier to perceive the effects of benzos because there is already alcohol-induced impairment.