The Brewers Association recently completed an updated demographic audit of brewery owners with the top line results presented in my State of the Industry presentation at the Craft Brewers Conference. Before we get to those results, it will be helpful to outline a new methodology we used and why. For previous demographic data, we’ve simply opened our response to breweries (members or not). While this approach gathered breweries that were broadly distributed around the United States, any non-random survey, particularly one on something like race/ethnicity or gender, has the potential to carry with it a response bias. Response bias — a type of sampling bias — is when some group of respondents is more likely to respond than other groups. Given the nature of American conversations about race/ethnicity and gender, I’ve been worried about the potential for this type of bias in our results. By randomly selecting the breweries before we send out the survey, and filling in the results for those who don’t respond, we can control for that bias.
Of the 500 randomly selected breweries, 136 responded to our survey and 364 we coded. That also allowed us to compare the responses to the coded data and look to see whether we are getting a response bias for future iterations. Note: not every brewery who responded provided full data, and not every brewery has “human” owners. Four of the coded breweries had “institutional” owners that we left blank (two colleges, a monastery, and a co-op). As I’ll outline below, there is evidence we were getting a response bias in the past, meaning that this data should not be compared to the 2019 results.
The data below is for individual owners, not breweries. So a brewery with two white owners and an Asian owner would have each coded and aggregated separately. An example of this difference is that 92.2% of breweries have solely white ownership, so the percentage of breweries with some BIPOC ownership is slightly higher than the percentage of BIPOC owners.
For race/ethnicity, here are the owner results:
Race/Ethnicity | Percentage of Owners |
---|---|
White (non hispanic) | 93.5% |
Black (non hispanic) | 0.4% |
Asian | 2.0% |
Hispanic, Latina -o, or of Spanish Origin | 2.2% |
American Indian or Alaska Native | 0.4% |
Native Hawaiian or Other Pacific Islander | 0.0% |
Other | 0.5% |
Prefer not to answer | 1.1% |
Due to evidence of response bias, you shouldn’t compare these benchmarking results to the 2019 benchmarks, so we can’t assess change over the last two years.
And Gender of Owners:
Gender | Percentage of Owners |
---|---|
Female | 23.7% |
Male | 75.6% |
Non-Binary/Third Gender | 0.2% |
Prefer not to answer | 0.6% |
Similar to race/ethnicity, the percentage of breweries with some female/non-binary ownership is higher than the percentage of female owners. I also created the following table showing the distribution of female ownership by brewery. While we didn’t explicitly track, it’s likely that a high percentage of the 50/50 ownership breweries are husband/wife ownership teams. As evidence, 91.3% of the 50/50 ownership breweries have two owners.
Female Ownership | Percentage of Breweries |
---|---|
None | 58.6% |
At least 1 female owner | 41.4% |
Gender Split (male/female) | Percentage of Breweries |
50/50 | 28.3% |
Some other non-50/50 and non 0/100 or 100/0 ratio | 10.2% |
100% female | 2.9% |
That female-male breakdown is pretty similar to our previous benchmarking (conducted in 2019), but the race/ethnicity numbers have shifted fairly significantly. Now, some of that could be due to changes in the brewery owner population over the last two years (we did have 1,700+ brewery openings in 2019-20). That said, by comparing the responses to the coded data, it’s pretty clear that the race/ethnicity results last round were skewed by response bias on race/ethnicity on open surveys.
Digging into the data, within the 500, those that responded were 86.6% white, versus 97.0% for the people we coded. If you add in people outside the 500 who responded (people may have found the link or the page), our total percentage of white owners for responses was 88.5% — a very similar percentage to the last time we benchmarked (also by free response). I won’t show you the statistical tests, but it’s very unlikely we’d get a gap between responses and coded data that large without response bias. Basically it appears that people more invested in these efforts are more likely to respond, and those people are a higher percentage BIPOC.
Importantly, that means you shouldn’t compare these benchmarking results to the work we did last time. Despite the respondent group looking similar to last time, I would be wary in even assuming that the response bias is similar now as it was two years ago. A lot has happened in the intervening two years related to race in the United States, so separating changes in population from changes in response bias isn’t really feasible. We’ll benchmark again in two years — again using this random sample approach — and those data points will be comparable to these.
As for the results, a reminder that even with a strong methodology, this is still a sample, and so there is going to be random error, particularly in sub-groups with small percentages and so small sample sizes. Without a full census, I’d be wary of extrapolating from these to total numbers (for example, multiplying 2% by the total breweries then giving a precise number of Asian brewery owners).
In lieu of a longer results writeup, I’ll briefly summarize the comments I made from the stage at CBC. The American beverage alcohol consumer is increasingly BIPOC and female. If you want data that shows this, read this Rabobank report. That shift is ongoing and will likely continue going forward. For example, female drinkers under 25 now outnumber male drinkers under 25. So for craft to continue growing and moving more in the larger beer and beverage alcohol consumer market, it will need to connect better with that diverse customer base. While there’s nothing that says white and male owned businesses can’t connect with that more diverse customer base, it’s going to require additional work in building diverse organizations and shoring up blind spots. The BA is working to build those resources, but our industry has to see the value and want to use them. These results should underline why they are needed.