TEMPORAL HALF-HOURLY EFFECTS
The temporal dummies capture events (publicly announced
news releases, market openings and closings) whose timing,though not generally their exact scale, is known in advance.
Public new related to macroeconomic variables is simultaneously
announced to all traders, at a time known in advance
since the scheduled time of all economic related news
is predetermined, and reported on another part of the
Reuters system, the FXNB page. The stochastic element in
such cases is the actual announcement, not the timing of it.
In general, the majority of the US announcements are
around 13:30 hours British Summer Time (BST), and the
German ones around 10:00 hours BST. Consequently, the
relationship between the dummy variables and the characteristics
of interest to us in the market predominantly reßect
response of these variables to publicly known events. Per
contra, the relationship between these variables, after conditioning
on such temporal constants, will primarily reßect
private information to a somewhat greater extent.
Notice that the constant represents the last half hour of
the last Friday in the sample. During this half hour all the
main markets are closed and only a few traders, if any at all,
input quotations. Therefore, the constant in the estimation
reßects, on average, the smallest number of observations in
the sample, but not necessarily the lowest level of volatility
or the smallest average spread. Let us now concentrate on
these dummy e¤ects.
The estimated dummy coe¦cients, for both currencies
and per equation, are not presented here because of space
considerations.2 Let us consider the half hour dummies Þrst.
In graphs 1a to 3b in Figure 1 the values of the estimated
dummy coe¦cients for both currencies are presented. They
reveal an interesting feature. In the last part of the day BST
time, from about the closing time of the European exchanges
and until the closing time of the New York exchange,
volatility is unusually high. Notice that this takes
place in both currency markets.
During this period there are few, or no, economic (or
other public) announcements from Europe or Asia (considering
only Japan). Most US economic announcements are
made before the opening of the New York Stock Exchange,
at 13.30 BST. There is a small spike at the relevant half hour
(27), but this remains quite small compared with the higher
volatilities apparent later on in the US market day.
Hence, it seems that public news is not the explanation of
this volatility increase. Furthermore, this increase seems
even more di¦cult to explain in the light of the Admati and
Pßeiderer (1988) theory. During this period we certainly
have a reduction in the number of traders in the market, as
only the New York exchange is in operation, so this increase
can hardly be attributed to an increase in the number of
liquidity traders.
There is then an apparent decrease in volatility for both
currencies, during the early morning period between 1:30
and 3:30 (BST). Most of the economic-related news for the
Japanese economy is announced either early in the Japanese
morning, i.e. around 1:00 BST, or in the late Japanese
afternoon, i.e. 6:00 BST. The same time period is characterized
by high spread and screen activity. However, it appears
that Japanese economic-related news has no e¤ect on the
volatility of the JPY currency. Although in line with the
results of Ito and Rolley (1987), this remains peculiar. Furthermore,
the same is true for the Deutschemark in relation
to German economic announcements, which are mostly
released either around 9:30 or 14:00 BST. Hence, it seems
that only US economic news a¤ects the variability of DEM
and JPY exchange rates.
There is a further curiosity in the half-hourly dummies
which is worth mentioning. During the Tokyo lunch time
break (4:00Ð5:00 BST) there is a dramatic decrease of volatility
coupled with an increase in spread and a decrease in
the number of quotations in the Þrst half-hour period (between
4:00Ð4:30 BST), followed by an increase in volatility
coupled with a decrease in spread which cannot be explained
by public information theories. Perhaps traders who
come back early from lunch take ÔwildÕ positions to make
their early return worthwhile. On the other hand this volatility
increase could be a statistical artefact due to the small
number of quotations during that period; that is, a few
observations out of Ôequilibrium levelÕ can have a dramatic
increase in the sample variance of the rate.