Using data from the Frankfurt Stock Exchange, this paper investigates
the impact of an increase in trading hours (from two to three on 15 Ja
nuary 1990) on the variance of stock returns. The results confirm thos
e of most earlier studies that report that trading time volatility is
significantly larger than non-trading time volatility. In addition, th
e results are consistent with the private and public information hypot
heses with regard to stock return volatility, but they do not support
the noise trading hypothesis.