When forward-looking individuals face binding borrowing constraints, c
onsumption will exhibit excess sensitivity'' to lagged variables which
are correlated. with income. However, the predictive capacity of such
variables for consumption growth should. diminish when contemporaneou
s and future income terms are used as regressors. U.S. data confirm th
ese expectations. But the data from three other industrial countries a
re not consistent with either borrowing constraints or other leading P
ermanent Income Hypothesis (PIH) variants. The evidence from all sampl
e countries suggests that variables correlated with future income migh
t be used to re-examine the inconclusive evidence for the PIH from hou
sehold data.