Switching costs in the wholesale distribution of cigarettes.

• Author: Elzinga, Kenneth G.

1. Introduction

   Consider the following scenario. A manufacturer introduces a new product and begins earning profit. The manufacturer's success attracts entrants who introduce similar products to vie for the incumbent manufacturer's customers. The entrants undercut the incumbent's price, even as the incumbent's price begins to fall. A price war ensues during which some of the incumbent's former customers switch to an entrant while others continue to buy the good from the incumbent. New customers also are drawn into the market because prices are lower than before. Once the entrants have established a clientele of regular customers, the free fall in prices stops and all firms raise their prices to a common, sustainable level. The post-price-war price of the good is less than the pre-entry price.(1)

   Klemperer (1989) and Elzinga and Mills (1998) use discrete-time dynamic models to explore the theory of price wars triggered by entry and show that customer switching costs can produce the sequence of events just described. Entrants offer temporary discounts from the prevailing price to attract customers who incur switching or set-up costs when they buy a manufacturer's product for the first time. Without such inducement, none of the incumbent's customers would switch to the entrants. In reaction, the incumbent cuts its price to stem the loss of its customers to the entrants. Prices rebound partially after the entrants lock in regular customers. The incumbent earns positive (discounted) profits in both of these models of price warfare because of its first-mover advantage, but free entry means entrants earn zero (discounted) profits.

   Assumptions about the size distribution of customer switching costs plays an important role in these two theories. In Klemperer's model, all customers are assumed to have the same switching cost. As a consequence, there is no switching; the only customers the entrants attract during the price war are new customers - those drawn into the market for the first time by the prospect of lower prices in the future. In Elzinga and Mills' (1998) model, customers are assumed to have heterogeneous switching costs. Whether a customer switches to an entrant or continues buying from the incumbent in this model depends on the size of the customer's switching costs. Individual customers are induced to switch suppliers if the temporary savings afforded by an entrant's lower prices during the price war exceed the cost of switching. Thus, customers who change suppliers are those with lower switching costs while those who remain with the incumbent have higher switching costs.

   In this paper, we study the prices charged and the pattern of switching that occurred in a price war that epitomizes Elzinga and Mills' theory. This episode is the generic cigarette price war of 1984-1985. Section 2 provides some background on the cigarette manufacturing industry and describes the events that transpired in 1984-1985. Section 3 discusses the nature and kinds of switching costs that occur generally in the wholesale distribution of consumer goods and in the wholesale cigarette market in particular. This section focuses on potential sources of heterogeneity in switching costs for a cross-section of customers and offers some empirically testable hypotheses.

   Section 4 describes data we assembled from the voluminous trial record in the antitrust litigation surrounding this episode of price warfare. From these data, we estimate a probit equation to explain the actual switching behavior of wholesale cigarette customers and to test the proposed hypotheses. In section 5, we infer the size distribution of switching costs from the parameter estimates obtained from the probit equation.

2. The Generic Cigarette Price War of 1984-1985

   At the time of the pricing episode analyzed here, the U.S. cigarette manufacturing industry was a six-firm oligopoly in which no major firm had entered or exited for decades. Besides these, there were a few very small producers of specialized cigarette products.(2) Cigarettes were then and are today differentiated in a variety of ways (e.g., by length, flavor, and tar delivery), and all major manufacturers sell many brands. The 1984 Maxwell Report lists more than 210 separate brands and styles sold in the U.S. by the major firms. The direct customers of the cigarette manufacturers are hundreds of grocery wholesalers, specialized distributors of candy and tobacco products, and direct-buying retail chain stores.

   Before 1980, cigarette manufacturers competed largely along nonprice dimensions: advertising,(3) packaging, and new product innovations. Manufacturers' list prices typically were uniform (with a small but uniform premium charged for the longer 100-mm cigarettes), although slotting allowances paid to customers who stocked new brands were de facto temporary discounts.

   The demand for cigarettes has been in decline since 1981 when consumption peaked at 638 billion units (i.e., cigarettes); in 1995, 490 billion units were consumed (Maxwell 1988, 1995). The decline in demand is attributed generally to health concerns and increasing restrictions placed on smokers. As cigarette sales declined throughout the 1980s, excess production capacity emerged in the industry and the firms became more active in price competition.

   During the relevant time period, Liggett and Myers (Liggett) was the smallest firm in the industry. Its market share had fallen from over 20% in the 1940s to less than 3% by 1980. In 1980, Topco, a large grocery wholesaler distributing a broad range of generic items, decided to test-market generic cigarettes. It sought out Liggett, who agreed to use its excess capacity to produce private label generic cigarettes for Topco. Liggett sold these cigarettes to Topco at prices markedly below its branded products, and Topco resold them at retail prices lower than branded cigarettes.

   Soon thereafter, Liggett expanded its discount cigarette operations by producing three kinds of generic cigarettes: private label, franchise label, and broker label generics. Private label cigarettes had customized packaging and were made for large customers, mostly retail chain stores. Franchise label generic cigarettes, with plain-Jane names like "Filter Lights," were sold to specific wholesale customers on an exclusive basis within the customers' geographic area and then resold to retailers. Typically, the franchise label cigarettes were awarded to the largest wholesale establishments specializing in the distribution of candy and tobacco products. The third kind of generic cigarettes were sold to grocery wholesalers and the remaining candy and tobacco distributors. These broker label generic cigarettes were marketed by a few large generic multiproduct brokers, and their packages bore the brokers' trademarks. Generics sold at wholesale prices 30-40% below the price of branded cigarettes. Liggett's generic sales grew so much that its share of the cigarette market reached 6% in 1984, notwithstanding a continuing decline in its brands.

   As in the scenario offered at the beginning of this paper, Liggett's success in selling generic cigarettes attracted competition. In mid-1984, R. J. Reynolds repositioned its failing Doral brand as a discount cigarette and Brown and Williamson (B&W) introduced a line of generic cigarettes using a marketing approach much like Liggett's. The arrival of these new products triggered a price war in the generic segment of the market. Prices fell precipitously in mid-1984 and remained low until the close of 1985.(4)

   B&W's generic cigarettes, especially, were a close substitute for Liggett's, and the rivalry between these two manufacturers was intense. In keeping with the theory of price wars triggered by entry, entrant B&W's prices were lower than incumbent Liggett's throughout the price war.(5) When the price war ended, both firms raised their prices to a level that was below Liggett's pre-entry prices.

   During the 18-month-long price war, B&W induced many of Liggett's regular generic cigarette customers to switch their generic cigarette purchases to B&W.(6) In 1984, the cigarette manufacturers had about 3500 wholesale customers. Liggett sold its generic cigarettes to approximately 2000 of these customers before B&W introduced its generics. During the price war, B&W acquired about 1500 generic cigarette customers, many of whom switched from Liggett. Others were new generic customers who did not begin to distribute generic...