|
ABSTRACT:
The weak form of the efficient market hypothesis states that a speculator
can
not obtain profit using only past information on price variations. Despite
the
great amount of work, so far it has not been found a transaction strategy
contradicting this hypothesis. For instance, technical analysis is the
earliest attempt to make profit from the past price data formulating
rigorous
transaction rules which can be implemented as computer programs. The
trading
rules are based on the emission of buy and sell signals using the values
of
some indicators calculated with the past prices.
In this paper we have numerically simulated the stock market speculations
achieved using the technical indicator Rate of Change Crossover (ROCC). We
have used the transaction data of 1733 stocks randomly selected from those
listed at American stock markets between 1968 and 1998. For each stock we
have
built time series of 601 consecutive transaction days (300 days for the
stock
investment and 301 days for the indicator calculation at the investment
beginning) under the conditions that the data have no obvious recording
errors
and they do not contain daily price variations greater than 20\%. We have
obtained a sample of 12288 time series, large enough to allow significant
statistical estimations.
The numerical results are the same as those reported so far in the finance
literature. In accordance with the efficiency hypothesis, the profit of
ROCC
is smaller than that of the passive investment in which the shares bought
at
the beginning of the investment are preserved without any transaction
until
the end of the investment interval. However, ROCC contains useful
information:
in average the buy signals are followed by days with price variations
greater
than the average daily variation.
back
|