This paper provides:
(i) a broad overview of the Theory of Futures Markets and Market Efficiency
(ii) (ii) a more detailed literature review of empirical tests of the semi-strong form market efficiency hypothesis using the event study method, and
(iii) (iii) methodology and results of an empirical test of the semi-strong form market efficiency hypothesis using the event study method. The empirical study examines daily prices for the Share Price Index (SPI) and 90 day Bank
Accepted Bill (BAB) futures contracts on the Sydney Futures Exchange (SFE) for a 10 year period. The SPI sample period is 1983 to 1994, the BAB from 1980 to 1990.
The study generates and analyses the cumulative average abnormal returns around four macroeconomic announcements of the Australian Bureau of Statistics (ABS). The impact of the ABS Current Account (monthly and quarterly announcements), Consumer Price Index and Gross Domestic Product announcements are examined.
Four distinct models of normal returns are utilised, labelled NRI - 4 (The model parameters are detailed at page 76). Two of these, NRI and NR2 are found to produce reasonably useful and consistent results throughout (NRl, a CAPM model and NR2, a model constructed for this study). The other two models, a regression or market model and a market adjusted model, produce results of normal and abnormal returns that appear somehow systematically biased and are inconsistent with NRI and NR2, and with each other - consequently the results they produce are excluded from the analysis and reporting of empirical findings at Chapter 7.
Two price series data sets are constructed from the futures contracts, one using all contracts (closing in March, June, September and December), and the other using just the June and December closing contracts. Results from the two differently constructed data sets are consistent throughout the empirical study.
The results support the conclusion that during the period of the sample the SPI market was reasonably efficient in responding to all announcements except the GDP release. Similarly the BAB market is found to be reasonably efficient in the context of the monthly current account announcement.
The study produced some evidence that is inconsistent with the semi-strong efficient market hypothesis utilised in this study (outlined at Chapter 6, section 6.2.1), particularly in relation to the GDP announcement for both the SPI and BAB markets under study. However, the tradability of such 'inefficiency' identified is highly unlikely - given the small quantum of the delayed total percentage CAAR responses. The largest delayed CAAR response was 1.5% - approximately half the amount required to cover transaction costs alone.
In addition, the CIA (quarterly) and CPI announcements are found to have no relevant information components for the 90 day BAB market.