We now examine whether message posting activity, as reflected in the number of messages, aggregate sentiment, and aggregate disagreement among investors, can (a) explain contemporaneous return, volatility, volume, and spread, and (b) predict these variables. Previous studies find a good explanatory and predictive power of the number of messages, but not of sentiment and disagreement indices. However, in contrast to previous studies that use nonweighted indices, we use creditweighted indices. The higher a poster's reputation, the greater is the weight given to the poster's sentiment. Also, we include messages with and without a selfreported sentiment, with the sentiment of messages without a selfreported sentiment assessed using a text classifier.
We follow Antweiler and Franks' (2004) contemporaneous and time sequencing leadlag regression models. We first discuss the contemporaneous models followed by the leadlag predictive models. For both these sets of regression models, we first calculate the various endogenous variables, including the daily raw return, volumeweighted volatility, proportion of volume induced by small trades, and bidask spread. We then test each of them as an endogenous variable in separate regression models.
(i) Contemporaneous Regressions
In the contemporaneous models, the endogenous variables mentioned above and the message board activity variables are on the sameday basis. The contemporaneous relation between posting activities and the stock's raw return is tested using the following model:^{16}
 (7)
where R_{t} is the stock raw return, Sentiment is the creditweighted sentiment index as shown in formula (2), Disagreement is the creditweighted disagreement index as per (4), NMessages is the log of the number of messages posted for the stock, Size is the log of the firm's market capital, MktRet is the return on the PowerShares QQQ fund, and R_{t}_{1} is the oneday lagged stock raw return in order to control for autocorrelation.^{17}
The contemporaneous effect on volatility is tested using the following model:
 (8)
We use intraday standardized volumeweighted intraday price volatility. In Model (8) and the next two models, we use the absolute value of the sentiment index (AbsSentiment) because both optimistic and pessimistic sentiments may induce more volatility, volume, and spread. Since trading volume is positively correlated with volatility (Jones et al., 1994), we add NTrades, the log of the number of trades of the stock, as a control variable. Similar to Antweiler and Frank (2004), we include Mkt, the log of PowerShares QQQ fund's price level, to control for market movements. is included to control for autocorrelation.
We next examine the effect of message board activities on the proportion of smallsized trading volume. The idea is that retail investors, who trade in smaller sizes, are the ones likely to be influenced by message board activities. Accordingly, we use the ratio of dollar trading volume in smallsized trades (trades of $3,750 or less) and total dollar trading volume as the dependent variable. The model is as follows:
 (9)
Because more small trades occur on NASDAQ, we include Listing, a dummy variable for NASDAQlisted stocks. Oneday lagged volume is included to control for autocorrelation.
The contemporaneous relation between posting activities and liquidity, with the bidask spread as a proxy for liquidity, is tested using the following model:
 (10)
Spread is the ratio of the stock's average bidask spread and mean price on day t. Because spread and trading volume are negatively related, we include Trade, the log of the number of trades of the stock, as a control variable. is included to control for autocorrelation.
Table 5 reports the results of the contemporaneous regressions. First, returns are significantly positively related with Sentiment (β= 0.006, tstat = 2.13) and oneday lagged value of returns (β= 0.305, tstat = 2.64), and significantly negatively with Size (β=−0.045, tstat =−2.66). Thus, the sameday stock returns are higher for stocks with higher sentiment, smaller size, and high previous day raw return. Significant oneday lagged return suggests a twoday price momentum, which is consistent with a preevent twoday pumping hypothesis. Different from previous research, the number of messages does not significantly explain contemporaneous returns when there is no material news about a stock. The disagreement index is also insignificant.
Table 5. Contemporaneous Effects of Posting Activities Dependent Variable  Raw Return_{t}  Volatility_{t}  SmallTrade Vol._{t}  Spread_{t} 

Coeff.  tstat  Coeff.  tstat  Coeff.  tstat  Coeff.  tstat 


Intercept  0.674**  2.19  0.014  0.01  0.194  0.89  0.038*  1.88 
Msg. board variables         
Sentiment_{t}  0.006**  2.13       
Abs. sentiment_{t}    0.015*  1.70  0.003**  2.49  0.001  0.92 
Disagreement_{t}  0.003  0.02  0.017  0.14  0.015  0.41  0.002  0.71 
NMessages_{t}  −0.004  −0.05  −0.005  −0.06  0.013*  1.68  −0.001  −0.52 
Control variables         
Size_{t}  −0.048**  −2.66  0.002  0.12  −0.020***  −2.71  −0.003***  −3.25 
MktRet_{t}  0.897  0.35       
Mkt_{t}    −0.002  −0.14  0.001  0.28  0.000  0.66 
NTrades_{t}    0.043***  2.85    −0.001**  −2.18 
Listing      0.051**  2.02   
Lagged variables         
Return_{t1}  0.305**  2.64       
Volatility_{t1}    0.521***  9.82     
Volume_{t1}      0.846***  17.37   
Spread_{t1}        0.435***  7.10 
Rsquared  0.11  0.68  0.89  0.84 
Second, the contemporaneous volatility test shows that intraday volatility is significantly negatively related with AbsSentiment (β=−0.015, tstat =−1.70). This suggests possible herding among online traders when there is a strong sentiment, either optimistic or pessimistic, which is consistent with Hirschey et al. (2000). The other two facets of message board activity, the number of messages and disagreement index, are insignificant. As expected, the number of trades is significantly positively correlated with volatility (β= 0.043, tstat = 2.85). One day lagged volatility is also significantly positive (β= 0.521, tstat = 9.82).
Third, the contemporaneous test of the proportion of trading volume in smallsized trades yields several interesting findings. The greater the absolute sentiment, the greater the proportion of trading volume in smallsized trades (β= 0.003, tstat = 2.49). Further, the proportion of trading volume in smallsized trades is greater for stocks with a higher number of messages (β= 0.013, tstat = 1.68); smaller stocks (β=−0.020, tstat =−2.71), and NASDAQlisted stocks (β= 0.051, tstat = 2.02). Finally, oneday lagged value of the dependent variable is significantly positive (β= 0.846, tstat = 17.37).
Last, the contemporaneous bidask spread regression shows that AbsSentiment is not significantly related to the bidask spread. However, the signs of absolute sentiment, disagreement, and the number of messages are all negative, suggesting that higher levels of these variables tend to reduce the spread (improve liquidity). As expected, the bidask spread is negatively related to the control variables firm size (β=−0.003, tstat =−3.25) and the number of trades (β=−0.001, tstat =−2.18) and positively related to the oneday lagged spread (β= 0.435, tstat = 7.10).
To conclude our discussion of contemporaneous tests, investors’ creditweighted sentiment dominates the other two message posting parameters (number of messages and disagreement) in terms of the relation with return, intraday volatility and smallsized trading volume. Higher sentiment is associated with higher returns while higher absolute sentiment is associated with less volatility and more smallsized trades.
(ii) Predictive Power
To most investors, the ability to predict subsequent changes is more important than contemporaneous relations. So, we now examine if sentiment, disagreement, and the number of messages on a given day can predict market activities on the next two days. Our time sequencing leadlag models follow Antweiler and Frank (2004). We test the oneday (twoday) predictive power of posting activities by leading the endogenous variables and the control variables by one day (two days) relative to the explanatory variables in the contemporaneous models examined earlier.
The oneday predictive results are included in Table 6. Similar to our contemporaneous results, we find that creditweighted sentiment, not the number of messages, predicts the return on the following day. Consistent with our event study results that suggest a twoday dumping pattern, a higher sentiment predicts a lower return on the following day (β=−0.022, tstat =−1.77). The next day significant drop could be due to dumping by the influential posters who had originally pumped up the stock and profit taking by them and their followers on the message board. Overall, based on our contemporaneous and predictive regression results, greater bullish opinion about a stock on event day t is associated with a price jump on day t but a price drop on day t+1. Theoretical foundation of such a pattern (overreaction on private information followed by a correction) is presented in Daniel et al. (1998) and Aggarwal and Wu (2006). Our findings are also consistent with prior studies such as Neumann and Kenny (2007) and Keasler and McNeil (2010) that document a significant stock price drop after media hype. We also find that a higher disagreement among posters on day t predicts a price drop on the next day (β=−0.155, tstat =−2.37).
Table 6. OneDay Predictive Power of Posting Activities Dependent Variable  Raw Return_{t+1}  Volatility_{t+1}  SmallTrade Vol._{t+1}  Spread_{t+1} 

Coeff.  tstat  Coeff.  tstat  Coeff.  tstat  Coeff.  tstat 


Intercept  0.118  0.53  1.399**  2.25  0.017  0.09  0.032  1.32 
Lagged msg. board variables         
Sentiment_{t}  −0.022*  −1.77       
Abs. sentiment_{t}    −0.010**  −2.47  0.004*  1.75  −0.001  −1.12 
Disagreement_{t}  −0.155**  −2.37  −0.011  −0.12  0.024  0.85  −0.003  −0.89 
NMessages_{t}  −0.005  −0.10  −0.094  −1.36  0.014  0.80  −0.001  −0.38 
Control variables         
Size_{t+1}  0.010  0.82  0.014  0.68  −0.003  −0.48  −0.001  −0.77 
MktRet_{t+1}  0.195  0.13       
Mkt_{t+1}    −0.026**  −2.17  −0.000  −0.08  −0.000  −0.72 
NTrades_{t+1}    0.014  0.80    −0.001**  −2.33 
Listing      0.022*  1.70   
Lagged dependent variables         
Return_{t}  −0.004  −0.06       
Volatility_{t}    0.658***  11.57     
Volume_{t}      0.986***  28.86   
Spread_{t}        0.842***  8.32 
Rsquared  0.09  0.72  0.93  0.83 
Similar to our contemporaneous tests, we find that absolute sentiment predicts next day's intraday volatility and volume. A higher absolute sentiment is associated with a lower nextday volatility (β=−0.010, tstat =−2.47) and a greater proportion of smallsized trading volume (β= 0.004, tstat = 1.75).
We now examine the twoday predictive results included in Table 7. It comes as no surprise that the longer the window, the harder it is to predict. Neither the message board variables nor firm size and market return are effective in predicting the return two days later. However, the absolute sentiment is still effective in predicting intraday volatility (β=−0.001, tstat =−2.06) and the proportion of smalltrade volume (β= 0.007, tstat = 1.87). The number of messages continues to be insignificant in predicting tradingrelated activities.
Table 7. TwoDay Predictive Power of Posting Activities Dependent Variable  Raw Return_{t+2}  Volatility_{t+2}  SmallTrade Vol._{t+2}  Spread_{t+2} 

Coeff.  tstat  Coeff.  tstat  Coeff.  tstat  Coeff.  tstat 


Intercept  0.049  0.34  −0.517  −1.04  −0.018  −0.10  −0.024  −0.90 
Lagged msg. board variables         
Sentiment_{t}  −0.002  −0.13       
Abs. sentiment_{t}    −0.001**  −2.06  0.007*  1.87  −0.001  −0.88 
Disagreement_{t}  −0.022  −0.40  −0.006  −1.06  −0.031  −0.84  −0.004  −1.07 
NMessages_{t}  −0.037  −1.40  −0.004  −0.10  0.038  1.01  −0.004  −1.60 
Control variables         
Size_{t+2}  0.009  1.31  0.032*  1.76  −0.003  −0.58  −0.002**  −2.57 
MktRet_{t+2}  0.451  0.63       
Mkt_{t+2}    −0.015  −1.37  −0.002  −0.42  0.001  1.15 
NTrades_{t+2}    0.039**  2.79    −0.001  −1.52 
Listing      0.048***  3.08   
Lagged dependent variables         
Return_{t+1}  0.108  0.93       
Volatility_{t+1}    0.985***  7.69     
Volume_{t+1}      0.922***  27.78   
Spread_{t+1}        0.935***  9.95 
Rsquared  0.07  0.85  0.92  0.82 
To summarize, we find that when there is no material news about a stock, credit weighted sentiment index significantly negatively predicts the next day's return. Our findings are consistent with a pumpanddump trading scheme.^{18} The significant price drop during the dumping period suggests possible trading opportunities for certain traders such as arbitrageurs and contrarian traders who are willing to take opposite positions based on stock message board information. We also find that the absolute sentiment significantly negatively predicts the volatility on the next day as well as two days later and significantly positively predicts the smalltrade volume on those days. We conclude that message board sentiment is an important predictor of tradingrelated activities.
Overall, the results in this section support our hypothesis that online posting activities affect online intraday activities as message board sentiment is significantly related with same day's and next two days’ tradingrelated activities.