نقدینگی بازار الماس، Q's و سهام مربوط به آنها

We investigate the market liquidity effects of the introduction of index-tracking stocks for the Dow Jones Industrial Average (DIAMONDS) and the NASDAQ 100 index (Q's). Our main finding is liquidity of the underlying DJIA 30 index stocks improves after the introduction of the exchange-traded fund, largely because of a decline in the cost of informed trading. Further, we find the DIAMONDS has significantly lower liquidity costs over the first 50 days of trading as compared to the portfolio of its component stocks, again primarily because of lower adverse selection costs. Finally, we find weaker but qualitatively similar results for the Q's.

How liquid are the markets for “basket” securities and how do they affect the market liquidity of underlying securities? There has been a virtual explosion in the introduction of index-based products known as exchange-traded funds (ETFs) developed by the Nasdaq-Amex Group. The first exchange-traded fund, Standard and Poor's Depository Receipts (i.e, SPDRs®), was introduced in 1993. Other products based on the Dow Jones Industrial Average (DIAMONDS®), NASDAQ 100 Index (a.k.a. Q's), S&P MidCap 400 Index (MidCap SPDRs®), major industry sectors (Select Sector SPDRs®), and the Morgan Stanley Capital International Indexes (iShares®) followed. ETFs are the recent innovations in basket securities. While the promoters of these new instruments highlight their tax efficiency, diversification benefits and transaction cost advantages, there is little empirical evidence on the liquidity of the new instruments nor on the effect of their introduction on the market liquidity of underlying stocks. Whether the recent proliferation of index-based products enhances or robs liquidity of trading from the underlying component stocks is of importance to investors and regulators alike. The introduction of DIAMONDS (DIA) on January 20, 1998 and the NASDAQ 100 Index-tracking stock (Q's) on March 10, 1999 and the availability of transaction level data allows us an opportunity to gain some insight into these issues. The DIA and Q's provide ownership in equity investment trusts that track the performance of their respective indices. Large investors and institutions create the index-tracking shares by depositing a portfolio of stocks that closely resembles the composition of the index and a specified amount of cash if necessary. The index-tracking stocks trade in essentially the same manner as individual equity securities on the floor of American Stock Exchange (AMEX). A notable difference between the baskets and their component stocks is that the two baskets accumulate dividends paid by underlying stocks and distribute them quarterly to their investors. Like SPDRs (see Ackert and Tian, 2000 and Ackert and Tian, 2001; Poterba and Shoven, 2002), both the DIA and Q's have proven especially popular. In the first 50 days from the introduction of the DIA, average daily trading volume was 826,000 shares and average daily dollar volume was $68.4 million. It ranks consistently as one of the 10 most active issues on the AMEX. Trading activity in the DIA security is comparable to the median trading volume in the underlying component stocks. For example, in March 1998, the median daily average NYSE trading for the 30 DJIA stocks was roughly 1.481 million shares worth $110 million, whereas the median DIA daily average trading volume is 592,000 shares worth $52 million. Similarly, the initial trading activity in Q's compares very favorably with the median trading volume in the NASDAQ 100 component stocks. In the first 20 days of May 1999, the average daily trading volume in Q's was 177% (312%) of the median trading volume in shares (dollars) for the underlying stocks. At the end of 2001, assets held by the DIA and Q's trusts stood at $3 and $22 billions, respectively, as compared with $30 billions held by the SPDR trust. The popularity of ETFs is attributed to the ease with which investors can obtain portfolio diversification benefits at low transaction costs as compared to trading a portfolio of underlying stocks. The relatively small price per share of ETF shares allows small investors to more easily take positions in the entire index as compared to buying and holding a portfolio of the component securities.2 Another advantage to trading in ETFs is that short selling is permitted on a downtick. These securities enjoy significant market liquidity over conventional index mutual funds, which cannot be sold short or bought on margin, and do not trade intraday. Also, the expense ratios for many large ETFs are substantially below those charged by equity mutual funds, even by index funds. For example, investors in DIA and Q's faced annual expense ratios of 12 and 18 basis points respectively in 2001, as compared with the asset-weighted average annual expense ratio of 24 basis points for domestic equity index funds in 1998 (Poterba and Shoven, 2002). By resorting to “redemption in kind” to reduce substantially their distributions of realized capital gains, ETFs have proven to be more tax efficient than many of their counterparts in the equity mutual fund industry. However, unlike shares in many mutual funds, which can be purchased from the sponsors, ETFs entail the added cost of bid–ask spreads and commissions. Moreover, for the more heavily weighted issues index-tracking securities provide another investment vehicle for firm-specific informed trading.3 Also, index-tracking stocks provide additional profit opportunities for factor informed trading (e.g., interest rate changes) as well as for trading on any information for which the returns of the component securities would be highly correlated (e.g., industry effects). ETFs can serve as a hedging instrument for firm-specific informed traders trying to hedge against market-wide movements. They can also be used in an arbitrage play with either the underlying component stocks and/or the respective index futures or option contracts. The available empirical evidence on the market liquidity effects of the introduction of basket securities is limited. Ackert and Tian, 2000 and Ackert and Tian, 2001 find the pricing of the SPDRs is efficient and the introduction of SPDRs results in an improvement in the pricing efficiency of the S&P 500 index options, but they report little evidence of enhanced arbitrage across markets. Using monthly returns and monthly bid–ask spread data, Jegadeesh and Subrahmanyam (1993) find weak evidence of decreased liquidity in underlying stocks after the advent of index futures. Edwards (1988) and Harris (1989) present mixed evidence on the effects of S&P 500 futures introduction on cash market volatility. Recently, Clarke and Shastri (2001) compare the liquidity of closed-end funds to their corresponding portfolios of underlying stocks and conclude that the closed-end funds have, on average, greater liquidity. Kurov and Lasser (2002) report the introduction of Q's has led to improved index futures pricing efficiency due to inter-market arbitrage. Our objectives are to present a comprehensive empirical analysis of the market liquidity of DIA and Q's and how the advent of trading in these baskets has affected the market liquidity of their underlying stocks. Our main finding is that, contrary to the potential for adverse liquidity effects noted by models that assume perfectly integrated markets (e.g. Subrahmanyam, 1991), the market liquidity of component stocks improves over the first 50 days of trading in both ETFs. This improvement in liquidity stems largely from a decline in the adverse selection cost of trading in the underlying stocks. Second, based on intraday data over the first 50 days of trading, we find significantly lower liquidity costs for both DIA and Q's relative to the corresponding portfolios of component stocks. Further, our sample evidence indicates that the higher market liquidity of the baskets is primarily attributable to lower adverse selection cost of trading in the composite securities. Finally, a brief review of derivatives trading shows that open interest and volume of trading in the DJIA and NASDAQ 100 index futures also increase significantly over the first 50 days of basket trading. Thus, our findings indicate that not only the markets in DIA and Q's are relatively more liquid, but also the advent of trading in these baskets has improved the market liquidity of corresponding derivative contracts and component stocks. The remainder of this paper is organized as follows. In Section 2 we review the relevant theoretical and empirical work on the advent of basket trading. Section 3 describes the sample and data we use to investigate the liquidity effects. In Section 4, we present empirical results for the DIA. Section 5 provides a brief summary of our analysis of Q's. Section 6 concludes with a summary of our findings.

Exchange traded funds have become popular and important investment vehicles. Whiletheoreticalworkintheareaoftradingincompositesecuritiesandtheireffects on market liquidity have greatly enhanced our understanding of how these new se- curitieschangetheincentivesofinformedanduninformedtraders,theempiricalpre- dictions emerging from these studies are somewhat ambiguous. In this study we provide a comprehensive empirical analysis of the market liquidity of the DIA and a summary analysis of the Qs as well as their underlying stocks. Consistent with the arbitrage hypothesis , our results demonstrate that the market liquidity of the underlying DJIA 30 index stocks improves over the first 50 days of trading in the DIA, primarily because of a decline in the asymmetric information cost of trading. We also find an increase in the volume and open interest of DJIA and NASDAQ 100 futures contracts after the advent of trading in the respective ETFs. Thus, we find the advent of portfolio trading generally improves the market liquidity of all assets. The results are less compelling for Q s, as we find weak evi- dence of increased liquidity in the underlying stocks following Q s introductionWe use fundamentally different market microstructure models to estimate several measures of market liquidity and find broad support for the recent theoretical ana- lyses of basket trading. Consistent with the general conclusions of these models, we find that over the first 50 days of trading, the DIA is more liquid as compared to a portfolio of its component stocks. Their superior liquidity stems largely from lower adverse selection cost of trading, presumably due to the diversification of firm- specific (and potentially factor) information in the basket. We also find weak evi- dencethattheQ shavelowertransactioncostsascomparedtoaportfoliooftheun- derlying NASDAQ 100 stocks. DIA and Q s are among the most popular ETFs introduced in recent years. It is possible that our key finding that the introduction of basket trading improves mar- ket liquidity all around is specific to the DIA or similar ETFs and cannot be gener- alized to all index securities. The extension of our analysis to other ETFs is an interesting avenue for future research