Switching Costs

The title of Klemperer’s 1986 PhD at Stanford was “Markets with Consumer Switching Costs”, which gave rise to a series of publications on switching costs in different contexts (these and other papers he reviewed in Klemperer 1995a).

Ex ante homogeneous products may, after the purchase of one of them, be ex post differentiated by switching costs including learning costs, transaction costs or artificial costs imposed by firms, such as repeat-purchase discounts. The non-cooperative equilibrium in an oligopoly with switching costs may be the same as the collusive outcome in an otherwise identical market without switching costs. However, the prospect of future collusive profits leads to vig­orous competition for market share in the early stages of a market’s develop­ment. The model thus explains the emphasis placed on market share as a goal of corporate strategy (Klemperer 1987a: 375).

In his paper “The Competitiveness of Markets with Switching Costs” (Klemperer 1987b), Klemperer modelled a two-period model with two firms. Switching costs can arise in the second period if consumers decide to switch from one seller to the other. Customers are modelled in a locational setting with the sellers at either end of the street. Customer demand in the first period is given by prices and linear transport costs. In the second period, there is an additional switching cost. The prices in the first period are influenced not only by maximising profit in that period but also by building up a customer base for the second period. It is in the second period that the firms can exploit the monopoly power created by the switching costs.

“Entry Deterrence in Markets with Customer Switching Costs” (Klemperer 1987c) also looked at a two-stage game, but where the only mover in the first period was an incumbent, who could influence the second stage entry game. The key here was the ability of the incumbent to build up a captive customer base (captive due to switching costs); modelling was of Cournot oligopoly with firms choosing quantities. Switching costs were captured by making sec­ond period purchases depend on the first period (positively). In the first period, the incumbent may over-invest in output to lock in customers and so deter any entry (or reduce the entrant's output) in the second period. However, for some parameter values, the incumbent prefers to deter entry by under­investing in output, thus committing to competing aggressively for new cus­tomers with any entrant in the second period (Klemperer named this phenomenon limit over-pricing).

“Welfare Effects of Entry into Markets with Switching Costs” (Klemperer 1988) showed that the presence of switching costs can lead to excessive entry that reduces social welfare. This is even possible if new entrants have lower costs than incumbents. The mechanism is that ‘a large amount of social sur­plus is dissipated by the consumers' cost of switching to the new competitor' (ibid.: 164). This phenomenon can even occur if industry output increases, resulting in lower prices.

In “Price Wars Caused by Switching Costs”, Klemperer (1989) extended the switching cost model to four periods. In the first two pre-entry periods, the monopolist sets up shop. In period three, a fringe of competitors enters (treating the incumbent's output as given) and there is the threat of additional fringe entry in period four. The path of prices over time is then traced. In the final period, with the switching costs, prices tend to rise to a higher level. Firms have just one period left to exploit all of their monopoly power. In period three, entry drives a price war. New entrants and the incumbent lower prices to build up or maintain their customer base to exploit in the final period. The effect of entry is to lower prices in period three relative to the pre­entry periods (and the incumbent may also lower prices in period two, imme­diately prior to the new entry). ‘In our model of a market in which consumers have switching costs, the entry of new firms leads to a price war, and we have argued that this conclusion is robust' (ibid.: 415).

In a paper with his then-student, Alan Beggs (Beggs and Klemperer 1992), “Multi-Period Competition with Switching Costs”, the switching costs approach was generalised to an infinite-period version of the duopoly model found in Klemperer (1987b). Firms have a discount rate and customers turn over, with some dying to be replaced with new-born customers. The paper established the conditions for existence of a symmetric steady-state equilib­rium and showed that equilibrium prices will be higher when there are switch­ing costs than when they are absent. Indeed, the net present values of firms' profits are usually larger—even for a new entrant which has no customer base—when there are switching costs.

In a paper with Kenneth Froot titled “Exchange Rate Pass-Through When Market Share Matters” (Froot and Klemperer 1989) took the basic idea of switching costs that market share in one period can lock in demand to subse­quent periods, and applied it to the classic issue of exchange rate pass-through.

Foreign firms may either raise or lower their dollar export prices when the dollar appreciates temporarily (i.e., the pass-through may be perverse) and import prices may be more sensitive to expected future than to current exchange rates. We explore whether expected future exchange rates provide a clue to the puz­zling recent behavior of US import prices (ibid.: 637).

Indeed, the paper had an empirical dimension that was not usual in the other papers on switching costs, in that it looked at data on import and export prices between the major economies in the 1980s.

In his 1992 paper “Equilibrium Product Lines: Competing Head-to-Head May Be Less Competitive”, Klemperer took a model where consumers face shopping costs, which mean that consumers prefer to shop at fewer outlets. In this case, Klemperer showed that in equilibrium, firms may prefer to offer identical product ranges so that customers will only choose one shop to visit. This head-to-head competition can lead to higher prices. If they are only shopping at one outlet, it requires a bigger price incentive for a consumer to switch outlets or shop at both outlets. If firms offer distinct product ranges, consumers shop at both outlets and can respond to quite small price differen­tials. This line of thought went against some then-established economic mod­els where firms maximise product differentiation in order to increase monopolistic markups. However, Klemperer argued that his model fitted with the marketing literature which identified umbrella branding (selling goods in related markets under the same brand name) and brand extension (selling new brands under an established brand name).

The issue of product lines, variety and welfare was explored in a paper with another of his students at the time, Jorge Padilla (see Klemperer and Padilla 1997), titled “Do Firms’ Product Lines Include Too Many Varieties?” A firm that offers an additional product can capture business from rival firms for other products when consumers prefer to concentrate their purchases at a single supplier.

Gilbert and Klemperer (2000) developed an equilibrium theory of ration­ing. In a simple model of pricing, a monopolist would never want to set a price in which all consumer demand was not met (a price with rationing). The argument is simple: the monopolist can make more money by raising the price and selling the same output. However, Gilbert and Klemperer consid­ered a two-stage set-up where consumers have to make an upfront investment to enter the market. In this case, the monopolist might want to precommit to keep the price low in order to encourage more consumers to enter. Although ex post the rationing is inefficient, it can yield higher profits if the firm is able to commit to it. This paper was later developed in the context of auction theory in Bulow and Klemperer (2002).

Klemperer's interest in oligopoly theory was not just theoretical. He wanted to apply his theoretical ideas to policy and his opinion was sought after. He was an adviser to the US Federal Trade Commission 1999-2001 (adviser on merger and competition cases and policy) and also a member of the UK Competition Commission 2001-2005 (and later adviser 2006-2014). He also undertook an applied analysis of the major economic issues raised by the 1997 Tobacco Resolution and the ensuing proposed legislation that were intended to settle tobacco litigation in the United States (see Bulow and Klemperer 1998). In addition, Klemperer wrote about the appropriate breadth of patent or copyright protection (see Klemperer 1990), stimulated by a debate about whether Japan's very narrowly defined patents took unfair advantage of American technology, and he has contributed to subsequent public policy debate in this area (see Klemperer 2004a).^[226]'

3