Social media impact on household investors and their stock markets participation

Posted on September 03, 2018

"Social media impact on household investors and their stock markets participation" is one of the short-term projects supported by the Think Forward Initiative.

Eric Tham, PhD candidate in Finance at EDHEC Business School, examines to what extent households trust social media when making financial investment decisions and how this trust influences their participation in the stock markets. Using investment sentiments information, this report shows that trust in social media has gradually increased since 2008, especially because of the interest in the guru-follower system. However, this growing trust in social media has not increased household participation in the stock market. Household investors do not seem to invest based on their trust only, but also according to prevailing news headlines and their risk aversion.

Find out more downloading the full report in the link below!


How does social media impact household investing?

In the last two decades, the use of social media has grown exponentially. This has changed the notion of social interaction amongst households. Hong (2004) found that households that are more socially interactive participate more in stock markets than those who are not.

Social media impacts the behaviour of household investors in two ways: through expedited information exchange via peer to peer transmission, and through financial literacy. Over the last two decades, social media investing has changed in character. In the late 1990s to early 2000s, social investing platforms such as Motley Fools formed a strong attraction. People loved online chatrooms and bulletin boards, busy with investors talking stock.

These websites were followed by pyramidal platforms from the early 2010s, including, and all about financial ‘gurus’ working their way up the pyramid. How? By getting followers to like their feeds, which leads to a great online reputation.

Household investors in charge

I – Eric Tham, PhD candidate in Finance at EDHEC Business School, – examine two pertinent questions on the impact of social media on household investing. Firstly, to what extent do household investors trust social media when investing, and how has this trust evolved over time? And secondly, how does this trust impact the households’ participation in the stock markets?

To determine the sentiment of social media posts, most researchers – Bollen (2011), Chen (2014), Heston (2017) – use Natural Language Processing (NLP). These studies all point to sentiment predating stock returns and trading volume, hinting of the importance of social media on the stock markets. But the decision-making process of the primary users of these social media – the household investors – remain to be examined.

Decision-making under risk

I attempt to answer this question through Prospect Theory, a seminal paper by Taversky (1979) on decision-making under risk. A key principle of this theory is that an investor derives value from investment losses or gains to a reference level. Prospect theory states that people go through two different stages when making a decision between choices that carry uncertainty.

First, there is the editing phase. Here people simplify the complicated decisions into simpler ones by just making a belief of gains versus losses per option. In the second valuation phase, people examine the edited options and act on the one they believe has the highest value.

In the first editing phase, investors scan the information horizon and formulate their beliefs. They reflect their beliefs (ie their trust on the stock market) by posting on social media – posts that in turn influence other readers’ trust.The second valuation phase is about deciding whether or not to act on their beliefs based on their risk preferences. They do not necessarily act on their beliefs; they may have spent a heap of money, or they may need to save money to see their kids through college.

Natural Language Processing and A.I.

To cover the first phase of the decision-making process, I use Thomson Reuters MarketPschy indices (TRMI), which uses NLP from artificial intelligence to determine the investors’ beliefs on social media. I also use the Baker-Wurgler market sentiment. Unlike TRMI, it involves trading activities, encompassing both beliefs and valuation phases. In other words, it considers whether investors actually put their money where their mouths are – unlike NLP, which is plain ‘talk’.

The commonality between these two sentiments is thence the beliefs phase, which can be extracted by a mathematical technique. These beliefs have been found to be pro-cyclical since 1998, reaching a peak in the early 2000s before decreasing. From 2008 to 2016, this trust in social media increased, as it still does. A regression analysis on this trust level shows it to correlate with stock returns through an established stock factor.

The puzzle of limited household stock participation

Does the recent increase in trust translate into an increase of households participating in the stock market? On the contrary: from 2008 to 2013, the triennial Survey of Consumer Finance indicated that direct household participation in the stock market has actually decreased till 2016. In the light of Prospect Theory, this isn’t surprising: investors do not actually invest based on their trust only, but also on their increased risk aversion. This is consistent with research by Guiso (2018), which shows a segment of Italian investors exhibiting increased risk aversion after the 2008 crisis.

Examining this puzzle of household participation further, I computed a dynamic correlation between the market and social media sentiment, taking out the effects from news headline sentiment. The news headline sentiment is a proxy for the economic and broader market fundamentals. Using a non-parametric ranked test,I found that this correlation has a significant monotonic relationship with the household participation rate from 1998 to 2013.

This suggests that households’ trust in social media is correlated with their stock market participation rate. This trust is not ‘blind’, but households still make decisions to invest according to prevailing news headlines.


  • Hong, H., Kubik, J. D., and Stein, J. C. (2004). Social interaction and stock market participation. Journal of Finance, I:137–163.
  • Kahneman, D. and Tversky, A. (1979). Prospect theory: An analysis of decision under risk. Econometrica, 47:263–291.
  • Bollena, J., Ma, H., and Zeng, X. (2011). Twitter mood predicts the stock market. Journal of Computational Science, 2, 1–8.
  • Chen, H., De, P., and Hu, Y. (2014). Wisdom of crowds: The value of stock opinions transmitted through social media. Review of Financial Studies, 27.
  • Heston, S. L. and Sinha, N. R. (2017). News vs. sentiment: Predicting stock returns from news stories. Financial Analysts’ Journal, 73.
  • Guiso, L., Sapienza, P., and Zingales, L. (2018). Time varying risk aversion. Journal of Financial Economics, 128:403–421.
  • Guiso, L., Haliassos, M., Jappelli, T., and Claessens, S. (2003). Household stockholding in Europe: Where do we stand and where do we go? Economic Policy, 18:123–170.