Paying for performance with altruistic or motivated providers

We present a model of optimal contracting between a purchaser (a principal) and a provider (an agent). We assume that: a) providers differ in efficiency and there are two types of provider; b) efficiency is private information (adverse selection); c) providers are partially altruistic or intrinsically motivated; d) they have limited liability. Four types of separating equilibrium can emerge, depending on the degree of altruism, characterised as very low, low, high and very high. i) For very low altruism the quantity of the efficient and inefficient types is distorted upwards and downwards respectively; the efficient type makes a positive profit. ii) For low altruism the quantity of the efficient and inefficient types is also distorted respectively upwards and downwards, but profits are zero for both types. iii) For high altruism the first best is attained: no distortions on quantities and zero profits. iv) For very high altruism the quantity of the inefficient type is distorted upwards, and the quantity of the efficient type is distorted either upwards or downwards. The inefficient type might have a positive profit. The quantity of the efficient type is higher than that of the inefficient type in all four possible equilibria. The transfer for the efficient type can be higher or lower than the inefficient one, unless altruism tends to zero in which case the transfer for the efficient type is higher. The utility of the efficient type is higher than that of the inefficient one when altruism is very low, low or high, though not necessarily when altruism is very high.