Mechanical vs. informational components of price impact

We study the problem of what causes prices to change. It is well known
that trading impacts prices — orders to buy drive the price up, and orders
to sell drive it down. We introduce a means of decomposing the total
impact of trading into two components, defining the mechanical impact of a
trading order as the change in future prices in the absence of any future
changes in decision making, and the informational impact as the remainder
of the total impact once mechanical impact is removed. This decomposition
is performed using order book data from the London Stock Exchange. The
average mechanical impact of a market order decays to zero as a function
of time, at an asymptotic rate that is consistent with a power law with an
exponent of roughly 1.7. In contrast the average informational impact
builds to approach a constant value. Initially the impact is entirely
mechanical, and is about half as big as the asymptotic informational
impact. The size of the informational impact is positively correlated to
mechanical impact. For cases where the mechanical impact is zero for all
times, we find that the informational impact is negative, i.e. buy market
orders that have no mechanical impact at all generate strong negative
price responses. Copyright EDP Sciences/Società Italiana di Fisica/Springer-Verlag 2007