• 沒有找到結果。

In this paper, the attempt is to find whether investor sentiment indicators can be used to explain market returns and market liquidity. The former has been tested by many other studies. The latter is a new issue which has seldom been mentioned in previous research. The market-based and survey-based investor sentiment indicators are simultaneously used in this study. The study hopes to discover the practical usefulness of investor sentiment indicators not only for forecasting the fluctuation of the market but also for providing a reference to the authorities, market makers and the related practitioners.

In our study, we amend the liquidity model of Chordia et al ―(2002) by adding investor sentiment indicators. The first hypothesis posits that the amended model can increase the explanatory power. Although the empirical result proves this hypothesis, the relationship between investor sentiment indicators and market returns seems to have room for further discussion.

In the second hypothesis, some factors are chosen to catch the variation of the market liquidity variables, particularly investor sentiment indicators. The second hypothesis postulates that market liquidity variables, percentage spread and OIBNUM, can be explained by investor sentiment indicators. The empirical result shows that investor sentiment indicators have little explanatory power in market liquidity variables despite the fact that the other independent variables in equation 2 and 3 seem to have better explanatory power on the market liquidity variables, especially on the percentage spread. These unanswered questions in the second hypothesis still need more thorough research in the future.

Overall, the research covering investor sentiment is relative new. There are unknown dimensions which are waiting to be discovered. This paper provides evidence to prove that investor sentiment can be used to predict market returns but maybe not be useful for predicting market liquidity. So far, the research into the relationship between investor sentiment and market liquidity is still incomplete. The field of investor sentiment requires more research to thoroughly explore the possibility.

References

Amihud, Yakov, 2002, Illiquidity and stock returns: cross-section and time-series effects, Journal of Financial Markets, 5, 31–56

Amihud, Yakov, and Haim Mendelson, 1986, Asset pricing and the bid–ask spread,

Journal of Financial Economics, 17, 223–249

Baker, Malcolm and Jeremy C. Stein, 2004, Market liquidity as a sentiment indicator,

Journal of Financial Markets 7, 271-299

Baker, Malcolm, and Jeffrey Wurgler, 2006, Investor sentiment and the cross-section of stock returns, Journal of Finance, 61, 1645-1680

Baker, Malcolm, and Jeffrey Wurgler, 2007, Investor sentiment in the stock market,

Journal of Economic perspectives 21, 129–151

Benston, George J., Robert L. Hagerman, 1974, Determinants of bid–asked spreads in the over-the-counter market, Journal of Financial Economics 1, 353–364 Blume, Marshall E., A. Craig MacKinlay, and Bruce Terker , 1989, Order imbalances

and stock price movements on October 19 and 20, 1987, Journal of Finance, 44, 827–848.

Brennan, M., and Avanidhar Subrahmanyam, 1996, Market microstructure and asset pricing: on the compensation for illiquidity in stock returns, Journal of

Financial Economics, 41, 441–464

Brown, Gregory and Michael Cliff, 1999, Sentiment and the stock market, Manuscript in preparation, University of North Carolina at Chapel Hill

Brown, Gregory W. and Michael T. Cliff, 2004, Investor sentiment and the near-term stock market, Journal of Empirical Finance 11, 1 –27

Brown, Philip, David Walsh, and Andrea Yuen, 1997, The interaction between order imbalance and stock price, Pacific-Basin Finance Journal 5, 539–557

Chan, Kalok, and Wai-Ming Fong, 2000, Trade size, order imbalance, and the volatility–volume relation, Journal of Financial Economics 57, 247–273

Chordia, Tarun, Richard Roll, and Avanidhar Subrahmanyam, 2000, Commonality in liquidity, Journal of Financial Economics 56, 3–28

Chordia, Tarun, Richard Roll, and Avanidhar Subrahmanyam, 2001, Market liquidity and trading activity, Journal of Finance 56, 501-530

Chordia, Tarun, Richard Roll, and Avanidhar Subrahmanyam, 2002, Order imbalance, liquidity, and market returns, Journal of Financial Economics 65, 111–130 Chung, Huimin, 2006, Investor protection and the liquidity of cross-listed securities:

Evidence from the ADR market, Journal of Banking & Finance 30, 1485–1505

Clarke, R. and M. Statman, 1998, Bullish or bearish? Financial Analysts Journal 4, 63–72

Datar, Vinay T., Narayan Y. Naik, and Robert Radcliffe, 1998, Liquidity and stock returns: an alternative test, Journal of Financial Markets 1, 203-219

D‘Avolio, Gene, 2002, The market for borrowing stock, Journal of Financial

Economics 66, 271-306

De Long, J. Bradford, Andrei Shleifer, Lawrence H. Summers, and Robert J.

Waldmann, 1990, Noise trader risk in financial markets, Journal of Political

Economy 98, 703-738

Domowitz, Ian, Jack Glen, and Ananth Madhavan, 2001, Liquidity, volatility and equity trading costs across countries and over time, International Finance 4, 221-255

Fung, Joseph K.W., 2007, Order imbalance and the pricing of index futures, Journal

of Futures Markets 27, 697–717

Fung, Joseph K.W. and Philip L.H. Yu, 2007, Order imbalance and the dynamics of index and futures prices, Journal of Futures Markets 27, 1129–1157

Fisher, Kenneth L. and Meir Statman, 1999, The sentiment of investors, large and small, working paper, Santa Clara University

Glushcov, Denys, 2005, Sentiment betas, working paper, University of Texas

Hasbrouck, Joel, Duane J. Seppi, 2001, Common factors in prices, order flows and liquidity, Journal of Financial Economics 59, 383–411

Huang, Roger D., and Hans R. Stoll, 1996, Dealer versus auction markets: A paired comparison of execution costs on NASDAQ and the NYSE, Journal of

Financial Economics 41, 313–357.

Jones, Charles and Owen Lamont, 2002, Short Sale Constraints and Stock Returns,

Journal of Financial Economics 66, 207-39

Kraus, Alan and Hans R. Stoll, 1972a, Parallel trading by institutional investors,

Journal of Financial and Quantitative Analysis 7, 2107–2138

Kraus, Alan and Hans R. Stoll, 1972b, Price impacts of block trading on the New York Stock Exchange, Journal of Finance 27, 569–588

Lauterbach, Beni, and Uri Ben-Zion, 1993, Stock market crashes and the performance of circuit breakers: empirical evidence, Journal of Finance 48, 1909–1925 Lee, Charles M.C., and Mark J. Ready, 1991, Inferring trade direction from intraday

data, Journal of Finance 46, 733-747

Lee, Wayne Y., Christine X. Jiang, and Daniel C. Intro, 2002, Stock market volatility, excess returns, and the role of investor sentiment, Journal of Banking and

Finance 26, 2277-2299

Rattray, Sandy and Devesh Shah, VIX Whitepaper, CBOE, 2003

Simon, David P., and Roy A. Wiggins III, 2001, S&P futures returns and contrary sentiment indicators, Journal of Futures Markets 21, 447–462

Shleifer, Andrei, and Robert Vishny, 1997, The limits of arbitrage, Journal of Finance 52, 35-55

Stoll, Hans R., 1978a, The supply if dealer services in securities markets, Journal of

Finance 33, 1133-1151

Stoll, Hans R., 1978b, The pricing of security dealer services: an empirical study of NASDAQ stocks, Journal of Finance 33, 1153–1172

Wang, Yaw-Huei, Aneel Keswani, and Stephen J. Taylor, 2006, The relationships between sentiment, returns and volatility, International Journal of

Forecasting 22, 109– 123

Appendix: Data process flow

1. Deleting the unnecessary and wrong data:

a. If the quote prices and quote volumes retrieved in the same second is identical to another set, then one set is deleted.

b. If the bid or ask price is equal to or less than zero, the datum is deleted.

c. If the bid or ask volume is equal to or less than zero, the datum is deleted.

d. If the traded price or volume is equal to or less than zero, the datum is deleted.

e. The quote datum with bid-ask spread <0 or bid-ask spread>4 is deleted.

f. Pt is trade price. If

P

t

P

t1

 / P

t1

 0 . 1

, the datum is deleted.

Note: The t and t-1 refer two adjacent trades because the trades are viewed as continuous trades. The following g and h are the same. We only compare these three criteria in the trading time (9:30:00~16:01:00).

g. at is ask price. If

 

t

 

t1

 / 

t1

 0 . 1

, the datum is deleted.

h. bt is bid price. If

b

t

b

t1

 / b

t1

 0 . 1

, the datum is deleted.

i. All trades and quotes which are beyond the trading time are deleted. Here we adopt 9:30:00~16:01:00 as our trading time. We choose 16:01:00 because there are still many large trades occurring after 16:00:00, and we need to cover those trades.

j. If there are quoting data in one day but no trading data in the same day, the additional quoted datum is deleted.

2. Merging data:

a. If there are many trading data in the same second, only the first one is reserved.

b. According to the trading data, we merge the quoting data by matching the time of the quoting data before one second or the older time of the trading data.

c. If there are many quoting data in one second, the quote with the largest quantity,

which is the largest sum of the bid quantity and ask quantity, is chosen. If that still cannot choose the single quote, the one with the smallest bid-ask spread is selected.

d. If there are no one-second-old quoting data, the first trading datum with the last-one-second or older quoting datum of the trading day before this trading day is merged.

e. Attention. For alleviating the work of the next section of distinguishing the orders, we need to reserve the five-second-old or older quoting data (comparing the time of the trading data). As a result, some quoting data before the time of the trading (before 9:30:00) should be reserved.

f. We only reserve the quoting data which is the largest quantity in every point of time.

3. Distinguishing the buyer-initiated or seller-initiated orders:

In this section, we compare the traded price with the midpoint of the ask and bid prices at least five seconds earlier. It is possible that the two traded price match the same midpoint of the ask and bid prices occurring at least five-second-old.

a. If the traded price is higher than the middle of the quote price occurring at least five seconds earlier, this trade is classified as being buyer-initiated.

b. If the traded price is lower than the middle of quote price occurring at least five seconds earlier, this trade is classified as being seller-initiated.

c. If the traded price is equal to the middle of quote price occurring at least five seconds earlier, we do the tick test. This is used to compare the traded price at time t with the traded price at time t-1.

Tick test 1: Here P

t and Pt-1 are compared. Pt is the current traded price. Pt-1 is the previous traded price.

If Pt is higher than Pt-1, the trade is classified as buyer-initiated.

If Pt is lower than Pt-1, the trade is classified as seller-initiated.

If Pt is equal to Pt-1, the tick test 2 is done.

Tick Test 2: Here P

t and Pt-2 are compared. Pt-2 is traded price before the previous traded price.

If Pt is higher than Pt-2, this trade is classified as being buyer-initiated.

If Pt is lower than Pt-2, this trade is classified as being seller-initiated.

If Pt is equal to Pt-2, this trade is deleted.

4. Printing out liquidity variables (daily data):

a. Percentage spread: After deleting the unnecessary and wrong trades and quote from the data, the percentage spread is calculated. The formula is ask price − bid price ask price + bid price /2 . Every percentage spread is summarized in every second and then the average is calculated to be a daily data.

b.OIBNUMt (Net buying pressure): The buyer-initiated orders are assigned +1, and the seller-initiated orders are assigned -1. Then the trade volumes are multiplied and they are summarized on a daily basis. This is the second liquidity variable.

Table 1 Summary Statistics for Equation 1

This table provides the descriptive statistics and correlation coefficients for the variables in equation 1. The variables include returns, contemporaneous and lagged excess buy and sell orders, lagged positive and negative returns, and investor sentiment indicators. The data range covers from 1995 to 2003 for SPY, and from 2001 to 2003 for QQQQ. Panel A reports the description statistics for SPY, and Panel B for QQQQ. Panel C reports the correlation coefficients for SPY, and Panel D for QQQQ. In Panel C and D, p-values are in parentheses. *, **, and *** indicate that the coefficient estimates are statistically significant at the 0.1 level, 0.05 level, and 0.01 levels, respectively.

Panel A: SPY

Variable N Mean Median Std Minimum Maximum

Return 2061 0.0003 0.0006 0.0129 -0.0752 0.0580

Excess buy orders(t) 2061 568255 55600 1080225 0 16571300

Excess sell orders(t) 2061 -507585 0 999907 -11302000 0

Excess buy orders(t-1) 2061 567214 54100 1080391 0 16571300

Excess sell orders(t-1) 2061 -509092 0 1000888 -11302000 0

Lagged positive returns 2061 0.00494 0.0005 0.0077 0 0.0580

Lagged negative returns 2061 -0.00463 0 0.0078 -0.0752 0

VIX 2061 23.0745 22.02 5.9359 10.36 45.74

pcratio 2061 0.6931 0.68 0.1588 0.3 1.36

AAII 2061 1.9797 1.74 1.1567 0.3636 8.3349

Panel B: QQQQ

Variable N Mean Median Std Minimum Maximum

Return 728 -0.0008 0.0007 0.0255 -0.0889 0.1016

Excess buy orders(t) 728 911882.47 0 1982138 0 21868500

Excess sell orders(t) 728 -1420806.35 -476200 1962749 -13450900 0

Excess buy orders(t-1) 728 909128.77 0 1982556 0 21868500

Excess sell orders(t-1) 728 -1431541.65 -480900 1972096 -13450900 0

Lagged positive returns 728 0.0094 0.0009 0.0149 0 0.1016

Lagged negative returns 728 -0.0102 0 0.0155 -0.0889 0

VXN 728 43.6440 44.07 11.4186 23.34 71.72

pcratio 728 0.7740 0.76 0.1521 0.46 1.36

AAII 728 1.8613 1.4708 1.2127 0.3636 8.3349

Panel C: SPY Return Excess buy

(0.0007)*** (0.081)* (<0.0001)*** (<0.0001)*** (<0.0001)*** (<0.0001)***

VIX -0.1297 0.0951 -0.2520 0.0945 -0.2415 0.1045 -0.3090 1.0000

(<0.0001)*** (<0.0001)*** (<0.0001)*** (<0.0001)*** (<0.0001)*** (<0.0001)*** (<0.0001)***

pcratio -0.2712 -0.0324 -0.1982 -0.0700 -0.1654 -0.1244 -0.2554 0.2099 1.0000

(<0.0001)*** (-0.141) (<0.0001)*** (0.0015)*** (<0.0001)*** (<0.0001)*** (<0.0001)*** (<0.0001)***

AAII -0.0015 -0.0101 0.0677*** -0.0006 0.0552** -0.0870*** 0.0502** -0.2722*** -0.1679*** 1.0000

(0.9475) (-0.6463) (0.0021) (0.7722) (0.0122) (<0.0001) (0.0227) (<0.0001) (<0.0001)

Panel D: QQQQ Return Excess buy

Table 2 Estimated Empirical Results of Equation 1

The dependent variable is the daily return on the SPY or QQQQ, donated Rt. Explanatory variables include contemporaneous and lagged positive and negative daily order imbalances measured in number of trades, lagged positive and negative ETF returns, and investor sentiment indicators. The model is divided by the model with sentiment and the model without sentiment. The figures in this table have been standardized. The data range covers 1995-2003 for SPY, and 2001-2003 for QQQQ.

The t statistics are in parentheses. *, **, and *** indicate that the coefficient estimates are statistically significant at the 0.1 level, 0.05 level, and 0.01 levels, respectively.

SPY SPY QQQQ QQQQ

Excess sell orders, 0.2409 0.1662 0.3372 0.2921

- |Min [0,OIBNUMt]| (10.98) *** (7.66) *** (9.38) *** (8.41) ***

Lagged excess buy orders, 0.0063 0.0189 -0.0156 -0.0269

Max[0,OIBNUMt-1] (0.28) (0.88) (-0.42) (-0.76)

Lagged excess sell orders, -0.0265 -0.0652 -0.0312 -0.0364 -|Min[0,OIBNUMt-1]| (-1.16) (-2.96) *** (-0.81) (-0.99) Lagged positive return, max[0,Rt-1] 0.0296 0.0469 0.043 0.027

(1.27) (2.06) ** (1.13) (0.69) Lagged negtive return, min[0,Rt-1] -0.0993 -0.1981 -0.0501 -0.1372

(-4.22) *** (-8.33) *** (-1.27) (-3.33)

Table 3 Descriptive Statistics for Equation 2 and 3

This table reports the descriptive statistics for the four models of equation 2 and 3.

The variables include contemporaneous and one-term-lagged market liquidity, Ln(sigma), Ln(volume), and investor sentiment indicators. The data range covers from 1995 to 2003 for SPY, and from 2001 to 2003 for QQQQ. Panels A and B are for percentage spread models, and Panel C and D are for OIBNUM models. The difference of the number of observations is because of the quantity of data.

Panel A: SPY (Percentage Spread)

Variable N Mean Median Std Minimum Maximum

Percentage spreadt 2055 0.0015 0.0013 0.0017 0.0002 0.0159

Lnσ 2055 -4.6090 -4.5223 0.2477 -5.2628 -4.3640

Ln(volume) 2055 15.7268 15.8220 1.3650 11.0413 18.4886

Return 2055 0.0003 0.0006 0.0129 -0.0752 0.0580

Percentage spreadt-1 2055 0.0015 0.0013 0.0017 0.0002 0.0159

VIX 2055 23.1004 22.0300 5.9249 10.3600 45.7400

pcratio 2055 0.693 0.6800 0.1598 0.3000 1.5862

AAII 2055 1.9787 1.7350 1.1591 0.0005 8.3349

Panel B: QQQQ (Percentage Spread)

Variable N Mean Median Std Minimum Maximum

Percentage spreadt 727 0.0015 0.0013 0.0007 0.0005 0.0044

Lnσ 727 -3.2269 -3.231 0.0777 -3.36 -3.0941

Ln(volume) 727 18.1524 18.1704 0.3002 16.43 19.2076

Return 727 -0.0008 0.0009 0.0256 -0.0889 0.1016

Percentage spreadt-1 727 0.0015 0.0013 0.0007 0.0005 0.0044

VXN 727 43.6272 44.06 11.4174 23.34 71.72

pcratio 727 0.7771 0.76 0.1662 0.46 2.6001

AAII 727 1.8593 1.4708 1.2138 0.3636 8.3349

Panel C: SPY (OIBNUM)

Variable N Mean Median Std Minimum Maximum

OIBNUMt 2060 60672.4248 55350 56860.73 -11302000 16571300

Lnσ 2060 -4.6106 -4.5224 0.2494 -5.2628 -4.3640

Ln(volume) 2060 15.7204 15.8208 1.3695 11.0413 18.4886

Return 2060 0.0003 0.0006 0.0129 -0.0752 0.0580

OIBNUMt-1 2060 58174.0752 54450 57757.14 -11302000 16571300

VIX 2060 23.0794 22.0250 5.9331 10.3600 45.7400

pcratio 2060 0.6930 0.6800 0.1588 0.3000 1.3600

AAII 2060 1.9799 1.7350 1.1570 0.3636 8.3349

Panel D: QQQQ (OIBNUM)

Variable N Mean Median Std Minimum Maximum

OIBNUMt 727 -504960 -475600 3221625 -13450900 21868500

Lnσ 727 -3.2269 -3.2310 0.0777 -3.3600 -3.0952

Ln(volume) 727 18.152 18.1704 0.3010 16.4300 19.2076

Return 727 -0.0007 0.0009 0.0255 -0.0889 0.1016

OIBNUMt-1 727 -517877 -476800 3228861 -13450900 21868500

VXN 727 43.6285 44.0600 11.4188 23.3400 71.7200

pcratio 727 0.7743 0.7600 0.1521 0.4600 1.3600

AAII 727 1.8603 1.4708 1.2133 0.3636 8.3349

Table 4 Pearson Correlations Matrix for Equations 2 and 3

This table represents the correlation coefficients for the four models of equation 2 and 3. The variables include contemporaneous and one-term-lagged market liquidity, Ln(sigma), Ln(volume), and investor sentiment indicators. The data range covers from 1995 to 2003 for SPY, and from 2001 to 2003 for QQQQ. Panels A and B are for percentage spread models, and Panels C and D are for OIBNUM models. *, **, and

*** indicate that the coefficient estimates are statistically significant at the 0.1 level, 0.05 level, and 0.01 level, respectively. The p-values are in parentheses.

Panel A: SPY (Percentage Spread)

Percentage spreadt Lnσ Ln(volume) Return Percentage spreadt-1 VIX pcratio AAII

Percentage spreadt 1.0000

Lnσ 0.0001 1.0000

(0.9974)

Ln(volume) -0.0526 0.8943 1.0000

(0.0171) ** (<.0001) ***

Return -0.0331 -0.0253 -0.0347 1.0000

(0.1334) (0.2521) (0.1155)

Percentage spreadt-1 0.6760 0.0016 -0.0609 0.0151 1.0000

( <.0001) *** (0.941) (0.0058) *** (0.4950)

VIX 0.1390 0.5290 0.5847 -0.1302 0.1373 1.0000

(<.0001) *** ( <.0001)*** ( <.0001) *** (<.0001 )*** ( <.0001) ***

pcratio -0.1253 -0.0103 0.2069 -0.2692 -0.1504 0.2096 1.0000

(<.0001) *** (0.6406) (<.0001)*** ( <.0001)*** (<.0001) *** (<.0001)***

AAII -0.0700 0.0307 -0.0412 -0.0012 -0.0650 -0.2717 -0.1722 1.0000

(0.0015) ** (0.1648) (0.0618) * (0.9579) (0.0032) *** (<0.0001)*** (<0.0001)***

Panel B: QQQQ (Percentage Spread)

Percentage spreadt Lnσ Ln(volume) Return Percentage spreadt-1 VXN pcratio AAII Percentage spreadt 1.0000

Lnσ 0.5294 1.0000

(<.0001) ***

Ln(volume) 0.0546 -0.1233 1.0000

(0.1413) (0.0009) ***

Return -0.0695 -0.0651 0.0622 1.0000

(0.0611) * (0.0796) * (0.094) *

Percentage spreadt-1 0.8760 0.5302 0.0264 0.0031 1.0000 (<.0001) *** (<.0001) *** (0.4777) (0.9340)

VXN 0.6197 0.8216 0.1215 -0.0815 0.6177 1.0000

(<.0001) *** (<.0001) *** (0.001) *** (0.0280)** (<.0001) ***

pcratio 0.0131 -0.1433 0.1600 -0.3222 -0.0140 0.0225 1.0000

(0.7243) (0.0001) *** (<.0001 )

*** (<.0001)*** (0.706) (0.5446)

AAII -0.3053 -0.2517 -0.0642 -0.0046 -0.3306 -0.4535 -0.0675 1.0000

(<.0001) *** ( <.0001)*** (0.0841) * (0.9009) (<.0001) *** (<.0001)*** (0.0691) *

Panel C: SPY (OIBNUM)

OIBNUMt Lnσ Ln(volume) Return OIBNUMt-1 VIX pcratio AAII

OIBNUMt 1.0000

Lnσ -0.0271 1.0000 (0.2192)

Ln(volume) -0.04284 0.8950 1.0000 (0.0519) * (<.0001)***

Return 0.3317 -0.0249 -0.0343 1.0000 (<.0001) *** (0.2585) (0.1198)

OIBNUMt-1 0.0628 -0.0277 -0.0738 -0.0096 1.0000 (0.0044) *** (0.2091) (0.0008)

*** (0.6649)

VIX -0.0902 0.5325 0.5873 -0.1299 -0.0844 1.0000

(<.0001) *** (<.0001)*** (<.0001)*** (<.0001) *** (0.0001) ***

pcratio -0.1408 -0.0112 0.2060 -0.2712 -0.1454 0.2105 1.0000 (<.0001) *** (0.6128) (<.0001)

*** ( <.0001)*** (<.0001)*** (<.0001)***

AAII 0.0342 0.0278 -0.0428 -0.0015 0.0292 -0.2727 -0.1678 1.0000 (0.1203) (0.2068) (0.0523) * (0.9471) (0.1859) (<.0001)

*** (<.0001) ***

Panel D: QQQQ (OIBNUM)

OIBNUMt Lnσ Ln(volume) Return OIBNUMt-1 VXN pcratio AAII

OIBNUMt 1.0000

Lnσ -0.0453 1.0000 (0.2222)

Ln(volume) -0.0355 -0.1251 1.0000 (0.3386) (0.0007) ***

Return 0.4009 -0.0601 0.0574 1.0000 (<.0001)*** (0.1054) (0.122)

OIBNUMt-1 0.0358 -0.0487 -0.0392 -0.0290 1.0000 (0.3346) (0.1899) (0.2908) (0.4347)

VXN -0.0888 0.8217 0.1196 -0.0788 -0.0929 1.0000

(0.0166) ** (<.0001)*** (0.0012)*** (0.0338) ** (0.0122) **

pcratio -0.1839 -0.1885 0.2060 -0.3169 -0.1389 0.0060 1.0000 (<.0001)*** ( <.0001)*** (<.0001)*** (<.0001)*** (0.0002)

*** (0.8714)

AAII 0.0160 -0.2497 -0.0660 -0.0046 -0.0025 -0.4521 -0.0687 1.0000 (0.6676) (<.0001) *** (0.0755) * (0.9008) (0.9474) (<.0001) *** (0.0642) *

Table 5 The Empirical Results for Equation 2 and 3

This table provides the empirical results for equation 2 and 3. The dependent variables are percentage spread for Panel A, and OIBNUM for Panel B. In each panel, it is separated two parts, SPY and QQQQ, respectively. *, **, and *** indicate that the coefficient estimates are statistically significant at the 0.1 level, 0.05 level, and 0.01 level, respectively. The t-statistics are in parentheses. All figures have been standardized.

Panel A: Percentage Spread

SPY SPY SPY QQQQ QQQQ QQQQ

Dependent variable: Percentage spreadt

Intercept 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000

(1.70) * (2.03) ** (0.98) (1.01) (1.01) (-0.55)

Lnσ 0.0479 0.0591 0.0119 0.0933 0.0924 0.0204

(1.31) (1.6) (0.29) (1.01) *** (4.39) *** (0.57)

Ln(volume) -0.0559 -0.0675 -0.0617 0.0486 0.0482 0.0283

(-1.53) (-1.82) * (-1.44) (4.46) *** (2.70) *** (1.46)

Return -0.044 -0.0441 -0.0471 -0.069 -0.0689 -0.0636

(-2.71) *** (-2.72) *** (-2.76) *** (-2.74) *** (-3.92) *** (-3.39) ***

Percentage spreadt-1 0.6732 0.6705 0.6542 0.8255 0.8236 0.8055

(2.71) *** (40.75) *** (38.8) *** (39.82) *** (38.58) *** (36.03) ***

VIX 0.0778

(3.58) ***

VXN 0.1034

(2.59) ***

pcratio -0.0459 0.00007

(-2.34) ** (0)

AAII -0.0311 -0.0172 -0.007 0.0146

(-1.88) * (-0.99) (-0.37) (0.72)

Adjusted R2 0.4584 0.4591 0.463 0.7787 0.7778 0.7793

Number of observations 2055 2055 2055 727 727 727

Panel B: OIBNUM

SPY SPY SPY QQQQ QQQQ QQQQ

Dependent variable: OIBNUMt

Intercept 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000

(0.98) (0.71) (0.08) (1.02) (1.03) (1.25)

Lnσ 0.0349 0.02495 -0.0058 -0.0265 -0.0244 0.034

(0.75) (0.53) (-0.11) (-0.77) (-0.68) (0.49)

Ln(volume) -0.0581 -0.04793 0.0037 -0.0603 -0.0595* -0.0336

(-1.24) (-1.01) (0.07) (-1.76) * (-1.72) (-0.98)

OIBNUMt-1 0.0626 0.0622 0.0574 0.0439 0.0441 0.0329

(3.00) *** (2.98) *** (2.73) *** (1.29) (1.29) (0.95)

Return 0.3312 0.3313 0.3182 0.4040 0.4041 0.3846

(15.95) *** (15.96) *** (14.51) *** (11.35) *** (11.85) (10.51) ***

VIX -0.0301

(-0.11)

VXN -0.0874

(-1.18)

pcratio -0.0375 -0.0449

(-1.49) (-1.15)

AAII 0.0302 0.0189 0.0079 -0.0185

(1.44) (0.85) (0.22) (-0.47)

Adjusted R2 0.1136 0.1141 0.1148 0.1622 0.1611 0.1626

Number of observations 2060 2060 2060 727 727 727

相關文件