In this paper we develop a differential technique for investigating th
e welfare effects of financial innovation in incomplete markets. Utili
zing this technique, and after parametrizing the standard competitive,
pure-exchange economy by both endowments and utility functions, we es
tablish the following (weakly) generic property: Let S be the number o
f states, I be the number of assets and H be the number of households,
and consider a particular financial equilibrium. Then, provided that
the degree of market incompleteness is sufficiently larger than the ex
tent of household heterogeneity, S - I greater than or equal to 2H -1
[resp. S - I greater than or equal to H + 1], there is an open set of
single assets [resp. pairs of assets] whose introduction can make ever
y household better off land, symmetrically, an open set of single asse
ts [resp. pairs of assets] whose introduction can make them all worse
off). We also devise a very simple nonparametric procedure for reducin
g extensive household heterogeneity to manageable size, a procedure wh
ich not only makes our restrictions on market incompleteness more pala
table, but could also prove to be quite useful in other applications i
nvolving smooth analysis.