Limits...
Decrypting Financial Markets through E-Joint Attention Efforts: On-Line Adaptive Networks of Investors in Periods of Market Uncertainty.

Casnici N, Dondio P, Casarin R, Squazzoni F - PLoS ONE (2015)

Bottom Line: By measuring the investors' expertise, we found that their behaviour could help predict changes in daily stock returns.We also found that expert investors were more influential in communication processes during high volatility market phases, whereas they had less influence on the real-time forum's reaction after bad news.Our findings confirm the crucial role of e-communication platforms.

View Article: PubMed Central - PubMed

Affiliation: Department of Clinical and Experimental Sciences, University of Brescia, Brescia, Italy.

ABSTRACT
This paper looks at 800,000 messages on the Unicredit stock, exchanged by 7,500 investors in the Finanzaonline.com forum, between 2005 and 2012 and measured collective interpretations of stock market trends. We examined the correlation patterns between market uncertainty, bad news and investors' network structure by measuring the investors' communication patterns. Our results showed that the investors' network reacted to market trends in different ways: While less turbulent market phases implied less communication, higher market volatility generated more complex communication patterns. While the information content of messages was less technical in situations of uncertainty, bad news caused more informative messages only when market volatility was lower. This meant that bad news had a different impact on network behaviour, depending on market uncertainty. By measuring the investors' expertise, we found that their behaviour could help predict changes in daily stock returns. We also found that expert investors were more influential in communication processes during high volatility market phases, whereas they had less influence on the real-time forum's reaction after bad news. Our findings confirm the crucial role of e-communication platforms. However, they also show the need to reconsider the fragility of these collective intelligence systems when under external shocks.

No MeSH data available.


Related in: MedlinePlus

Unicredit stock log-return, rt, series (blue line, left axis) and the filtered regime of volatility st/t (red line, right axis).
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4526688&req=5

pone.0133712.g002: Unicredit stock log-return, rt, series (blue line, left axis) and the filtered regime of volatility st/t (red line, right axis).

Mentions: Tab. 2 shows the estimates (posterior means) of the autoregressive coefficients in the three regimes. Finally, a key quantity in the estimation, which was also a useful output of the inference procedure, was the probability to be in a low-volatility regime condition on the information available up to time t. This was called a filtered probability and is denoted with ξkt+t = P(st = k/r1,…,rt,z1,…zt) (e.g., [31] and [28]). The filtered probabilities ξkt+t, k = 1,2,3, were obtained by the Hamilton-filter recursions. The filtered hidden state was obtained as st/t = arg max{ξkt+t,k = 1,2,3}. Fig 2 shows the stock return series (blue line, left axis) and the filtered regime of volatility st/t (red line, right axis).


Decrypting Financial Markets through E-Joint Attention Efforts: On-Line Adaptive Networks of Investors in Periods of Market Uncertainty.

Casnici N, Dondio P, Casarin R, Squazzoni F - PLoS ONE (2015)

Unicredit stock log-return, rt, series (blue line, left axis) and the filtered regime of volatility st/t (red line, right axis).
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4526688&req=5

pone.0133712.g002: Unicredit stock log-return, rt, series (blue line, left axis) and the filtered regime of volatility st/t (red line, right axis).
Mentions: Tab. 2 shows the estimates (posterior means) of the autoregressive coefficients in the three regimes. Finally, a key quantity in the estimation, which was also a useful output of the inference procedure, was the probability to be in a low-volatility regime condition on the information available up to time t. This was called a filtered probability and is denoted with ξkt+t = P(st = k/r1,…,rt,z1,…zt) (e.g., [31] and [28]). The filtered probabilities ξkt+t, k = 1,2,3, were obtained by the Hamilton-filter recursions. The filtered hidden state was obtained as st/t = arg max{ξkt+t,k = 1,2,3}. Fig 2 shows the stock return series (blue line, left axis) and the filtered regime of volatility st/t (red line, right axis).

Bottom Line: By measuring the investors' expertise, we found that their behaviour could help predict changes in daily stock returns.We also found that expert investors were more influential in communication processes during high volatility market phases, whereas they had less influence on the real-time forum's reaction after bad news.Our findings confirm the crucial role of e-communication platforms.

View Article: PubMed Central - PubMed

Affiliation: Department of Clinical and Experimental Sciences, University of Brescia, Brescia, Italy.

ABSTRACT
This paper looks at 800,000 messages on the Unicredit stock, exchanged by 7,500 investors in the Finanzaonline.com forum, between 2005 and 2012 and measured collective interpretations of stock market trends. We examined the correlation patterns between market uncertainty, bad news and investors' network structure by measuring the investors' communication patterns. Our results showed that the investors' network reacted to market trends in different ways: While less turbulent market phases implied less communication, higher market volatility generated more complex communication patterns. While the information content of messages was less technical in situations of uncertainty, bad news caused more informative messages only when market volatility was lower. This meant that bad news had a different impact on network behaviour, depending on market uncertainty. By measuring the investors' expertise, we found that their behaviour could help predict changes in daily stock returns. We also found that expert investors were more influential in communication processes during high volatility market phases, whereas they had less influence on the real-time forum's reaction after bad news. Our findings confirm the crucial role of e-communication platforms. However, they also show the need to reconsider the fragility of these collective intelligence systems when under external shocks.

No MeSH data available.


Related in: MedlinePlus