Is Divestiture Effective as a Merger Remedy in the U.S. Beer Industry?

The U.S. Federal Trade Commission (FTC) is the other relevant merger enforcement agency. It conducted a comprehensive investigation of divestiture, and more details can be found in FTC (Federal Trade Commission, 1999). The FTC investigation relied on interviews and a review of case studies instead of an explicit empirical evaluation of merger effects.

Kwoka (2014) documents other types of remedies that resulted in significant price increases. Additional examples of failed remedies include the Hertz-Dollar Thrifty merger in 2012 and the Albertsons-Safeway merger in 2015.

Grupo Modelo’s other breweries were not included in the divestiture.

See Kim and Singal (1993), Focarelli and Panetta (2003), Hastings (2004), Ashenfelter and Hosken (2010), and Kwoka and White (2019).

The only exception was that CB had owned and operated a beer brewery in Stevens Point, WI, from 1993 to 2002.

Note that the same divested brands are sold by ABI in Mexico.

A two-month investigation into the proposed transaction and CB’s suitability as the buyer was conducted before the merger was cleared for approval on April 19, 2013.

According to CB’s press releases, the interim supply arrangement between ABI and CB was extended to June 2017, because of CB’s capacity constraint.

The price effects of the merger and acquisition from Pabst/Blatz in the U.S. beer industry are analyzed by Elzinga and Swisher (2005). Ashenfelter et al. (2015) showed in the Miller-Coors merger that the price increase from concentration was offset by a nearly equal-and-opposite efficiency effect.

Observations before and after the four defined merger windows are dropped from our DID analysis. We kept the merger windows balanced since beer sales are highly seasonal. We also tested the empirical results with the use of all of the available observations, and the principal numerical results and conclusions were relatively stable and robust.

Other major distribution channels such as grocery stores and liquor stores are not included because of our data sharing agreement. These absent channels combined accounted for approximately 54% of overall sales in 2015. Therefore caution is required in the interpretation of our results given the potential heterogeneity across distribution channels, and the fact that conclusions are focused on a subset of the market channels in which products are sold.

The omitted region, product, and month category relative to these three categorical fixed effects variables is California, Beck’s, and the first month of data observations in the sample. See the data section below for details with regard to the time span of observations that are used in the analysis.

The five ABI brands included were Stellar Artois, Beck’s, Labatt Blue, Labatt Blue Light, and Labatt Ice. The nine non-divested imported beer brands that were used in the comparison were Heineken, Heineken Premium Light, Molson Canadian, Tecate, Tecate Light, Red Stripe, Dos Equis, Foster’s, and Newcastle.

The ten ABI domestic beer brands were Budweiser, Bud Light, Bud Ice, Bud Light Lime, Bud Light Platinum, Busch, Busch Light, Michelob, Natural Ice, and Natural Light. The ten non-divested domestic brands that were used for comparison purposes were Coors Light, Icehouse, Keystone Light, Miller High Life, Miller Lite, Milwaukee’s Best Ice, Milwaukee’s Best Light, Pabst Blue Ribbon, Steel Reserve 211, and Yuengling Lager.

In our data for these two outlets, the convenience store channel represented 70% of total sales on average, while the restaurant and bar channel accounted for 30% of total sales.

A 288 oz. package is equivalent to a “case” of beer, which consists of two 12-bottle packages or four 6-bottle packages, where each bottle contains 12 oz. of beer.

The IRI regional fixed effects were tested and found to be jointly statistically significant. This suggests that there may be unobserved time-invariant factors that possibly were caused by regional consumer preferences, breweries’ pricing strategies, and related factors. The joint impact of the time fixed effects was also tested and found to be statistically significant, which suggests the existence of underlying time trending factors that affected the beer industry. Such factors might include changing consumer tastes, breweries’ pricing strategies, and varying technologies.

There was a sharp percentage decrease volume for Victoria. There are at least three possible explanations for this decrease: First, consumer preferences for Victoria declined. We documented such a trend with the use of data from a beer rating site (beeradvocate.com). During our study period, ratings for craft beers surged. Second, since Victoria’s sale volume is small compared to products in the control group, even modest changes can appear to be large in percentage terms. The average monthly sales for the Victoria is 2.7 (units are 1,000 288-oz. package, see Table 1), which is 2.7/20.12 = 13.4% of the sales of the other imported beers control group. Third, the sharp drop in volume is only significant in the short-run post-merger window but not in the longer-run. Unobservable factors or shocks during the transition period that followed the merger consummation may have also contributed to the sharp transient decline in volume.