This paper develops the basic analytics of moral hazard, for the two-outcome case where either a fixed damage accident occurs or it does not. The analysis focuses on the relationship between the insurance premium paid and the insurance benefits received in the event of an accident, and is conducted in benefit-premium space. The central message of the paper is that even when the underlying functions, the expected utility function and the function relating the accident probability to accident-prevention effort, are extremely well-behaved, the indifference curves and feasibility set (the set of insurance contracts which at least break even) are not-indifference curves need not be convex and feasibility sets never are; price-and income- consumption lines may be discontinuous; and effort is not in general a monotonic or continuous function of the parameters of the insurance policies provided. Part I of this paper establishes these results, while Part II discusses sane of their implications. The bad behavior of indifference curves and the feasibility set profoundly affects the nature and existence of a competitive equilibrium. We illustrate this, though we do not provide a thorough analysis. We also show that our canonical model of an insurance market with moral hazard can be reinterpreted to provide a model of loans with bankruptcy, or of work incentives.
