RDP 8804: Pricing Behaviour in Australian Financial Futures Markets 3. Australian Futures Markets

Organised futures trading began in Australia with the opening of the Sydney Greasy Wool Futures Exchange (now the Sydney Futures Exchange, or SFE) in May 1960. After a relatively slow start, futures trading has grown rapidly, both in terms of turnover, and in the variety of contracts available for trade. The dramatic nature of this growth is illustrated in Table 1 on page 22, which shows total annual trading volumes for the main contracts.

As the table indicates, growth in futures trading has been particularly strong in the period since 1983, and the introduction of financial futures contracts has been largely responsible for this. The first financial futures contract, in 90 day bank bills, was introduced in October 1979; it was followed by contracts in the share price index (SPI) in February 1983, and in 10 year bonds (December 1984).^[1] These financial contracts have now largely replaced the older commodity futures as the main focus of trading, together accounting for over 90 per cent of the value of turnover by 1987. To some extent, this reflects overseas trends. In the United States, for example, the financial contracts have certainly been the main area of growth in futures trading; but there has been no corresponding decline in commodity futures. One reason for the relative demise of commodity futures in Australia may lie in the growing internationalisation of these markets. Most commodity contracts are now traded on the major exchanges in the U.S. which, being more liquid, are likely to be more attractive than the non-U.S. exchanges; this may underlie a tendency for the non-U.S. exchanges to specialise in local products, particularly local financial products, in which they may have a comparative advantage.

The close relationship between spot and futures prices in Australia is illustrated in figures 1, 2 and 3, showing weekly price observations on the three major contracts since December 1984. Broadly speaking, this close relationship is what the theory of storage would lead us to expect: when adjusted for net holding costs (which are probably small), the difference between futures and spot prices for an asset should be limited by the cost of spot-futures arbitrage. Consistent with this theory, the proportionate premium on SPI futures is generally larger and more variable than in the case of the bond contract. The 90-day bill market is less readily comparable to the other two at this level, because the cost of holding is related mainly to interest rate movements between the trading date and contract maturity date, which may at times be relatively large. Details of the three contracts are set out in the Appendix.

Based on casual inspection of the graphs, the predictive content of the futures prices in general does not appear to be large. A weak case can perhaps be made that the 10 year bond futures market anticipated the decline in bond yields which occurred in the early part of 1986 (though not the subsequent reversal of this trend). Generally, however, the two sets of prices move in parallel. An interesting exception concerns the period around the October share market crash. The premium on December SPI futures began to fall some three weeks prior to the crash, and was actually negative (a hitherto rare occurrence) for several days before hand. Immediately after the crash, the premium remained negative for some weeks, thus predicting further share price falls which subsequently occurred. Of course, this observation can be given more than one interpretation. Whether the futures market contributed causally to share price movements, or was merely forecasting them, is not easily determined from these data alone. However, if it is accepted that the initial impetus for lower share prices came from overseas, this would suggest that the main role of the futures markets lay in predicting, rather than instigating, the major share price movements, at least initially.

Figures 3 to 6 show the weekly trading volumes in the major futures contracts over the same sample period. Apart from their strong upward trend, the volume figures show little systematic behaviour in the pre-crash period, and appear to bear no obvious relation either to the level of prices, or to their volatility. In particular, there does not appear to have been any noticeable increase in the short-term variability of prices over the period while trading volumes were increasing.

In the week of the crash, volumes reached a high peak, due mainly to widespread closing out of positions in reaction to the drop in share prices. Subsequently, average trading volumes declined in all three markets, and have remained relatively low ever since. The decline has been particularly severe in SPI contracts, with the value of daily turnover so far in 1988 down to around 10 per cent of pre-crash levels. No doubt, many investors have become more cautious in view of the large losses incurred in October, particularly with respect to SPI futures. Turnover on the Sydney stock exchange is also reported to be considerably reduced. These observations suggest a model in which price volatility is mostly fairly constant, but may influence turnover in exceptional periods. There seems no clear evidence of a causal link in the other direction, from volume to volatility, but this proposition will be tested more rigorously in the work reported in section 6.

Footnote

A two year bond contract was also introduced (in 1982) but was not a success. More recently, a three year contract was introduced in April 1988, and has been quite heavily traded during its first few weeks of being available. [1]