February 9, 2016
Throughout discussions on the Department of Liquor Control (DLC), there have been many assertions about DLC’s prices. Some claim they are too high, others claim they are lower than private competitors, but so far all of these claims have been anecdotal. Fortunately, this is an empirical question that can be tested. To provide all of us with an objective answer to this question, I asked my staff to undertake such a study. Through much painstaking work, Tommy Heyboer on my staff has completed a wholesale price study comparing the prices DLC charges Montgomery County licensees with the prices of the same products charged by private wholesalers elsewhere in Maryland. We compared DLC prices to private sector prices for 771 randomly selected products, enough to state the average price differences with 95% confidence.
The study finds that DLC’s wholesale prices on stock products are essentially equal with private sector prices elsewhere in Maryland, while DLC wholesale prices on special order products are about 10% more expensive than those products would be elsewhere in Maryland.
The Council’s Office of Legislative Oversight and staff from the County Executive’s office, CountyStat, and DLC have reviewed the study and provided extensive feedback. I want to thank them for their numerous comments, which greatly strengthened the narrative and conclusions.
DLC vs. Private Sector: A Wholesale Price Comparison1
Based on a careful analysis of Department of Liquor Control (DLC) and private sector pricing data, DLC’s wholesale stock prices are roughly equivalent to private sector prices in Maryland while DLC’s wholesale special order prices are about 10% more expensive than private sector prices. This empirical analysis a) confirms anecdotal evidence from restauranteurs and retailers and b) follows logically from the fact that DLC is layering an additional markup to special orders when purchasing them from other distributors while DLC buys most stock items directly from manufacturers. The Council’s Office of Legislative Oversight (OLO), CountyStat, and DLC reviewed this analysis and provided comments that were incorporated into the study.
The following tables summarize the results.2 3
|Wholesale Special Order Price Difference: DLC vs. Private Sector|
|Post Markup Reduction
|Special Order Beer||47||21.53%||12.15%|
|Special Order Keg||18||39.02%||39.02%|
|Special Order Liquor||36||-1.99%||-1.99%|
|Special Order Wine||315||8.73%||8.73%|
|Total / Avg||416||10.56%||9.50%|
|Note: The individual product categories (special order beer, kegs etc..) price differences are not statistically valid in themselves because they do not have sufficient N. They are shown only to provide context.|
|Wholesale Stock Price Difference: DLC vs. Private Sector|
|Total / Avg||355||0.04%|
|Note: The individual product categories (stock beer, kegs etc..) price differences are not statistically valid in themselves because they do not have sufficient N. They are shown only to provide context.|
The data indicate that:
- Stock prices are essentially equal between the DLC and the private sector
- DLC special order prices are about 10% more expensive than the private sector
- DLC stock and special order liquor prices are less than the private sector
Research Design and Methodology
DLC provided the wholesale prices for every product in its inventory. Private sector wholesale pricing data came from the Maryland Beverage Journal4 and directly from several distributors5. DLC’s inventory was separated into stock and special orders and then each list was randomized6 separately. To achieve a 95% confidence level with a 5% confidence interval, each sample had, at least, an N of 3557. Further, the samples’ split of wine, beer, and liquor closely aligns with the actual proportions of DLC’s inventory; if beer products comprised X% of the inventory, they comprised X% of the sample. This study uses frontline private sector prices from May 2015 and compares them to the DLC prices. To determine the overall price differences, this study used price differences in percentage terms of each item (where a positive percent indicates DLC being more expensive by that percent and a negative percent vice versa) and then averaged the differences for each sample.
It is important to note that DLC did decrease the markup on special order beer8 over the summer after the study was underway. To determine whether the markup reduction had a significant effect on the price difference, the study shows both the sample with May 2015 special order beer prices and the same sample with the markup-reduced prices on special order beer. The markup reduction on special order beer reduced the price difference from 21.53% to 12.15% on those products. However, it only reduced the overall special order price difference by 1%. Therefore, the markup reduction did not produce a significant effect on the overall analysis. The table above shows both results.
Since the table above shows that special order beer is between 12% or 21% more expensive and special order kegs are nearly 40% more expensive, one could argue those product types are driving the bulk of the overall price difference. While it is true that they are part of the sample, they do not “punch above their weight,” so to speak. First, special order beer and kegs comprise about 15% of the sample, which, not coincidentally, is the same percentage of special order beer and kegs in the entire inventory. Special order wine, however, comprises over 75% of the sample.
Previous Price Comparisons
While both OLO and CountyStat have performed separate price comparisons, these analyses lack the inferential power provided by a statistically significant sample. Rather, they focused on the most popular DLC products. Those analyses say some interesting things about price differences on these specific products. However, they cannot say whether the trends identified apply to moderately popular and less-popular products or more importantly, to stock and special order products more broadly. In other words, the real question, “are DLC prices more expensive?” was still largely unanswered.
Executive Branch staff asked whether the analysis should use weighted averages based on the volume of a product sold by DLC. This analysis does not follow this methodology, because it is designed to answer a very specific question: “are DLC wholesale prices higher or lower than the Maryland private sector?” A random sample of the entire DLC inventory captures the price differences over the full breadth of products.
CountyStat’s and OLO’s price analyses focused on the impact on licensees of the prices of the most popular, most high-volume items, which answers a slightly different question. It is best to view this analysis as supplementing the results of OLO and CountyStat.
If licensees are price-sensitive when making purchasing decisions (and they presumably are), relatively higher prices on special order products will likely depress sales of these products compared to lower-priced items. DLC’s sales numbers cannot reflect the absence of sales, so using weighted averages based on sales volume could skew the results. Testimony from licensees to the Ad Hoc Committee on Liquor Control supports the assertion that licensees take price into consideration when deciding which products to order.
Second and more importantly, this analysis calculates overall price difference by averaging the price differences of each individual product, not by simply averaging the average of each product category. Therefore, if one calculates the weighted average of each product category in the sample, they would find that it too equals 10.5%. That said, the price difference in each product category is not statistically significant, given the insufficient amount of N for each product category. Those numbers simply provide additional context.
It is important to note once again, that the price studies done by the Office of Legislative Oversight and CountyStat shed some insight on the question of, “What effect do price differences of the most popular items have on licensees?”. In discussions before the Ad Hoc Committee, representatives from the DLC and CountyStat sought to focus attention on price comparisons of popular items, such as the top 50 products sold in stock or special order. One key challenge for this methodology is that it does not allow us to draw strong conclusions because the sample sizes are small.
The goal was to go beyond this limited view of the pricing and draw stronger inferences on DLC pricing overall. As discussed above, this why a statistically significant sample is more appropriate.
1 Prepared by Tommy Heyboer, Deputy Chief of Staff to Councilmember Hans Riemer. Special thanks to the County Executive’s staff and the Council Office of Legislative Oversight for insightful comments and suggestions.
2 The average price differences of each product category (beer, kegs, liquor, and wine) are shown for context, but given the relatively small N for each individual product category, they do not constitute statistically significant results on their own.
3 Weighted averages must be used for the individual product category price differences to add up to the total average.
4 As of May 2015
5 Distributors include Bond Distributing, Legends Ltd., and the Country Vintner
6 Once separated into stock and special order lists, a random number (using the RAND() function) was assigned to each product and then the lists were ordered from small to large. The items included in each sample are, at least, the first 355 items with DLC and private sector prices.
7 An N of 355 for the sample is based on an inventory of ~4,700 stock products and ~27,000 special order products
8 The DLC reduced their special order beer markup by 28.5%. It went from 35% to 25%. See the August 2015 DLC Newslink here.