on domestic sales of a domestic mon

on domestic sales of a domestic monopolist but not on exports, rather than charge a lower tax on all production. These strategic benefits may exceed the welfare loss resulting f’rom reduced domestic sales.’7 Second, in sequential markets firms may produce unprofitably in one period to gain the strategic benefit of’ making competitors less aggressive in future periods. For example, if’ a Japanese firm has decreasing costs over time (as in an industry with a “learning curve”) and its American competitor’s product is a strategic substitute, .the firm may “dump” output in the early stages of a market’s develop- ment to encourage the competitor to either contract operations or withdraw from the market.
D. Holding Idle Capacity to Deter Entry
Building extra capacity converts marginal costs into fixed costs and so raises a firm’s output. Therefore, such actions may deter entry. How- ever, beyond a certain point additional capacity will not deter entry further. Capacity deters competitors only if they believe the capacity will be used after entry. Thus the most extra capacity a firm will build is the amount it will actually use if entry occurs. Dixit (1980) and Spulber (1981) (effectively assuming strategic substitutes) concluded that if it would be profitable to use all capacity after competition enters, then it surely would be profitable to use all capacity if no entry occurs. Hence no capacity would be built and subsequently left idle. With strategic complements and quantity competition, however, a firm will want to supply less if it remains a monopolist than ifcompeti- tors produce. Consequently, capacity can be built that the firm could credibly threaten to use in the event of entry but that would be left idle if entry was deterred. Firms anticipating price competition with strategic complements may also rationally install idle capacity to deter entry. We give further details in Bulow, Geanakoplos, and Klemperer (1985).
E. Is a Little Bit of Competition a Good Thing?
In this section we restrict ourselves to quantity competition in homo- geneous products. Consider an entrant into a monopoly market, where the entrant has constant marginal costsjust below the pre-entry