We analyze incomplete contracts to induce efficient investment. With e
xogenous switching costs, fixed-price contracts are efficient, generat
e some rigidity in prices, are renegotiated intermittently by possibly
small amounts, and when inflation is positive, generate asymmetric re
sponses to shocks, all consistent with evidence on prices and wages. W
ith two-sided specific investments, efficiency requires prices to have
sufficient escalator clauses to avoid renegotiation, as observed in m
any long-term contracts. A third case, with one-sided specific investm
ents, can generate ''take or pay'' contracts and explain why firms som
etimes pay for specific investments that appear to benefit employees d
irectly.