155 / 322
臺大管理論叢
第
26
卷第
3
期
155
lead to better liquidity.
Liquidity is defined as the price of immediacy (O’Hara, 1995), the price impact from
large trade (Amihud, 2002), or the ability to convert assets into cash instantaneously
(Lippman and McCall, 1986) at the lowest transaction costs (Harris, 1990). It is easier to
define but far more difficult to measure. Most of the literature views volume, market depth,
bid-ask spreads and price impact as proxies for market liquidity. Since there is no consensus
about the best measure, we employ four common liquidity proxies: trading volume, depth,
the Amihud illiquidity measure, and the Amivest liquidity ratio.
Empirical studies have reported that open interest is positively related to trading
volume, a convenient proxy of liquidity. For instance, Martell and Wolf (1987), in
investigating the determinants of trading volume in metal futures, find that open interest has
a significantly positive explanation for trading volume. Moser (1994) reports that the
coefficient of elasticity between open interest and volume of S&P 500 index futures is
positive and the coefficient increases substantially after the Black Monday crash of October
19, 1987.
3
Market depth is usually viewed as the proxy of liquidity in empirical evidence, this can
be seen in Lee, Mucklow, and Ready (1993) and Brockman and Chung (1999). In a deeper
market, market orders, which seek for immediate transaction, are matched with higher
probability, fast speed, and smaller market impact costs. Therefore, market depth should be a
proxy of liquidity and participation.
Goyenko, Holden, and Trzcinka (2009) suggest that the illiquidity ratio provided by
Amihud (2002) does well in measuring price impact thus is a high quality liquidity proxy
based on daily data. Furthermore, the Amivest liquidity ratio is also a daily measure of price
impact long adopted by academics and practitioners. Hence, we utilize four liquidity metrics:
volume, depths, the Amihud illiquidity ratio and the Amivest liquidity ratio.
4
3 Open interest may co-move with futures trading volume because of the two variables share the same
cyclical pattern of futures contracts. That is, open interest and volume of a contract increase as the
contract about to become the nearby contract. Both decrease sharply when the futures contract approaches
it maturity (Kolb and Overdahl, 2006). To investigate the association between open interest and volume
independently to the cyclical pattern, this study remove the cyclical pattern in open interest (and volume)
by aggregating the open interest in the two near-month contracts. See section 3.1 for detail.
4 It should be noted, the quoted bid-ask spread may not be an appropriate liquidity proxy for the TAIEX
futures contracts. This is because that the TAIEX index futures is so liquid that the bid-ask spreads are
constantly as low as the minimum tick. As a result, the bid-ask spreads are insensitive to the changes in
liquidity.