This paper first assesses the costs of the US-Japan auto VER in a general equilibrium constant returns to scale (CRTS) model at about $10 billion. It then sequentially introduces important features of the auto VER: endogenous rent premium determination, wage distortions in autos, the US capturing some of the rents of the VER, US monopsony power in autos, increasing returns to scale, pure profits and entry, foreign direct investment, and endogenous conjectures. In the preferred monopolistic competition, initial profit model, the estimated costs are about 10% less than under the assumption of CRTS, but costs remain high at over $200 000 per job protected in autos. Compared with exogenous rent determination, endogenous rent determination results in significantly lower estimated costs of the VER because domestic entry reduces the rent premium. Foreign direct investment with initial profits is shown to lower the costs of the VER if, and only if, the rent premium is endogenous.