We have observations on random variables X, Y and Z, and we are interested in the relation between Y and X only. If we regress Y on X, this model suffers from the omitted variable bias problem

a) if Z is correlated with Y but not necessarily with X

b) regardless of whether Z is correlated with X and/or Y

c) if Z is correlated with both X and Y

d) if Z is correlated with X but not necessarily with Y