The effect of alcohol consumption in GDP is not a trivial thing to study. AFAIK, research has shown only slight affects on GDP. Take for example this study on "The societal cost of alcohol consumption" in Sweden, which showed only around 1% change in GDP:

http://www.ncbi.nlm.nih.gov/pubmed/18043953

I rather suspect that correlating caffeine to GDP would be even more difficult. Either way, the article doesn't even try to support these assertions factually. Which is OK -- it's a well-written article on an interesting subject. It's just good to keep in mind that some of the material is highly speculative.

Not just GDP but also nation-altering events such as warfare. I recommend "A Distant Mirror" by Barbara Tuchman if you haven't already read it, as it's a wonderful book (medieval history, but it still applies here).

Tuchman mentions our modern bewilderment at the insanely rash and reactionary decisions that were often taken by nobility in those times - declarations of war, terrible combat strategy and so forth, and which affected so many lives (mostly those of the poor). She goes on to speculate that these decisions make sense in the context of near-permanent drunkenness.

And [IIRC she also notes that] many of them were hotheaded teenagers / twenty-somethings. With the life expectancy of those days one had to grow up fast. And yet many did not.

This study is looking at modern recreational alcohol consumption. It's perfectly believable that its effect would be minor.

This an entirely different scenario from entirely replacing water with alcohol and being at least slightly drunk all the time. Not "all the time" as in "every weekend" but every hour of the day, including work.

Yes, but this is potentially counterbalanced by the increased dependence of modern work on clear heads. It's quite possible that the more manual, unmechanized work of pre-industrial societies was more immune to the effects of alcohol than even menial jobs today. It's also possible that modern technology raises the epidemiological cost of drunkenness in some ways; they didn't, for example, have too many drunk-driving deaths before there were cars.

Like I said, it's a complex area of research. If you think you're able to draw any meaningful conclusions based on your own intuition, then I think you're being a tad simplistic.

Really? Not a Ballmer curve believer? Never heard anyone suggest that add/adhd/being unable to focus/caffeine jitters can decrease productivity? Seems like exactly the kind of assumption without evidence that would warrant studying.

Maybe a topic for MythBusters? :)