Theoretical Issues

Most of the problems that should be tested for the purposes of this study have been addressed previously, but for the sake of clarity they will be repeated here. The first problem is that there is a controversy regarding whether the best age-related variable to test would be number of 25-34 year olds or median age, so this paper elects to test both. There is a perception that craft beer, more so than other types of beer, is a commodity enjoyed mostly by white people. However it has been suggested that it is the higher cost of craft beer, and the concentration of wealth in the hands of whites, that causes this to be the case. Thus, the effect of the number of white non-Hispanics in a county on the presence of craft breweries will be tested with and without controlling for wealth. Regressions will be run without the Midwestern states not covered by the Brewing News publication as it is not clear whether or not it is fair to assume that, because they are not included in the Brewing News maps, there are no craft breweries in those states.

One concern that deserves addressing is that breweries are not uniform in output. An area with a very high demand for craft beer could be just as easily supplied by one regional craft brewery supplying 150,000 barrels of beer annually as twenty microbreweries each supplying 7,500 barrels. One reason that some may not view this as a concern is that craft beer consumers are considered to be nowhere near as loyal to a particular brand as drinkers of more conventional American beers. The classical argument between the “Bud man” and a “Coors man” has no real craft beer parallel, since that sort of branding has yet to really catch hold in the craft beer community. Tom McCormick, writing for Probrewer.com, a beer industry resource site, states that “Most craft beer consumers have a range of 5-10 beer brands that they buy on a consistent and repetitive basis. Two or three of these brands are favorites, which typically capture about 50- 60% of their beer purchases.” The assumption, then, is that if a community had a large demand for beer and only one, very large craft brewery, it would probably be getting a good deal of the beer it consumes from other areas, and exporting a good deal of the beer it produces to other counties, and there would be incentive for other breweries to move in. A possible counterargument to this would be that large breweries may be able to diversify the number of different beers produced, thus keeping consumers from losing interest too quickly. Whether or not that is the case remains to be seen. However, in order to capture the possibility that large breweries could crowd out smaller ones in certain areas, another set of regressions will be run, with every regional brewery inflated to count for more than one brewery. Specifically, every one regional (larger) brewery will be assigned the weight of 4 microbreweries. The number 4 was based on the ratio between the output of Dogfish Head Brewery, as reported in a Senate press release (U.S. Senator Chris Coons of Delaware), estimated at 125,000 barrels, divided by 15,000, the minimum amount of beer that a craft brewery can produce while still retaining the “microbrewery” label (8.3) and then rounding down and dividing by two, assuming that less than half of most beer produced by a regional craft brewery is consumed in that county. Dogfish Head was chosen because of the availability of the data and their size, which falls somewhere between the very largest Regional Breweries, such as Sam Adams, and the smaller ones like Smuttynose, which up until recently was classified as a Microbrewery. This weighting system is obviously far from perfect, but it serves to test the possibility that results will appear different if larger craft breweries are weighted for their additional output. If the data were available, an ideal way to test this would be to actually weight every individual brewery by its output. Since those data were not available, this weaker method will be used instead.

Finally, it is fair to note that county level data may not be the most ideal geographic estimator for this purpose, particularly in certain areas of the Northeast, where there are many small, densely populated counties whose residents may move between counties on a regular basis. In Massachusetts, for instance, Greater Boston and the Pioneer Valley would probably be much more useful geographies than Suffolk and Hampshire counties, respectively. The use of Metropolitan Statistical Areas was ruled out on the basis that it would specifically exclude from the sample any brewery located in a rural area.