Using long, low frequency data on money and output over 1884-1996 for Argen
tina and over 1912-1995 for Brazil, it is found that money is long-run neut
ral but not long-run superneutral with regard to real output. A rise in mon
ey growth is associated with a decline in output - the opposite of the Tobi
n effect. The introduction of dummy variables for 1930s or to capture recen
t periods of financial disruption associated with bank insolvencies does no
t restore long-run superneutrality for either country. However, results ind
icate that bank insolvency episodes have a distinct and negative influence
on output. (C) 2000 Elsevier Science B.V. All rights reserved. JEL classifi
cation: E31; E51.