Marko Miettinen The Impact of Social Media on Investors’ Risk- Taking Vaasa 2025 School of Accounting and Finance Bachelor’s thesis Finance 2 UNIVERSITY OF VAASA School of Accounting and Finance Author: Marko Miettinen Title of the Thesis: The Impact of Social Media on Investors’ Risk-Taking Degree: Bachelor of Science in Economics and Business Administration Programme: Finance Supervisor: Kanying Xu Year: 2025 Number of pages: 40 ABSTRACT: This thesis examines how social media impacts investors’ risk-taking behavior. Thesis is based on previous finance literature. Theory consists of behavioral finance theories, social learning theory, and efficient market hypothesis. With the help of these we seek to understand the un- derlying phenomena. In social media, rapid information, algorithmic content personalization, and the influence of other investors are things that may increase investors’ tendency to take risks. The phenomenon is explained through herd behavior, confirmation bias, and the effects of narratives. Also, information overload may be a significant factor. The aim of the thesis is to answer the research question “How does social media influence investors’ risk-taking behavior” by presenting a hypothesis that social media increases investors’ risk-taking. Based on the dis- cussed studies, support has been found for the hypothesis, but some of the studies emphasize the context-dependent nature of the effects. The study shows that the effects of social media on investor behavior are variable and situational. The results highlight the need for further re- search that will clarify the significance of social media’s influence and more empirical evidence is needed. KEYWORDS: behavioural economics, risk-taking, social media, investors, narrative, infor- mation overload 3 Contents 1 Introduction 5 1.1 Research question 5 1.2 Structure of the paper 6 2 Theoretical background 7 2.1 Efficient market hypothesis 7 2.2 Behavioral finance theories 8 2.2.1 Prospect theory 8 2.2.2 Herd behavior 10 2.2.3 Confirmation bias 12 2.3 Social learning theory 12 2.4 Risk perception 14 3 Literature review 16 3.1 Social media as a new information environment 16 3.2 Information overload 18 3.3 Social media and behavioral finance biases 19 3.3.1 Herd behavior and social media environments 20 3.3.2 Confirmation bias and social media’s personalization 22 3.3.3 Risk perception under social media influence 25 3.4 Influencer-driven market dynamics 27 3.5 Narrative economics 29 4 Conclusion 33 5 Limitations and future research 35 References 36 4 Figures Figure 1. Prospect Theory: A hypothetical value function (Kahneman & Tversky, 1979) 9 Figure 2. Impact of a negative sentiment shock on prices (Tetlock, 2007). 17 Figure 3. GME volatility (solid) and WSB subreddit tone (dotted) data from December 2020 – March 2021 (Anand & Pathak, 2022). 21 Figure 4. Interrelationship between narratives, emotions, and market behavior (Taffler et al. 2024) 31 5 1 Introduction We have seen how social media has increasingly become an influential force shaping investor decisions and perceptions of financial markets. Platforms such as Reddit, X (for- merly Twitter), TikTok, and YouTube serve as spaces where financial information, market rumors, and investment advice spread rapidly. These social media platforms have ena- bled the transmission of information and collective action of investors with speed and scale, which may substantially affect the investors' risk-taking behavior. This phenomenon’s importance and relevance are emphasized especially when inves- tors' collective action in social media leads to significant market disorder, such as liquid- ity risks and a sudden increase in volatility. Concrete evidence of that was the 2021 GameStop saga, when Reddit’s WallStreetBets community organized a massive buying of the GameStop stock, which led to market shutdowns, strong price deviations, and significant losses to institutional investors. It showed that social media’s influence on the investor’s behavior is not only theoretical, but it may have noticeable real-world conse- quences to the individual investors and to the stability of the whole stock market. The primary purpose of this thesis is to critically review existing literature to understand how social media influences investors’ risk-taking behavior. This study integrates insights primarily from behavioral finance and aims to offer a comprehensive understanding of the mechanisms behind social media’s influence on investor decisions. 1.1 Research question The primary research question of this paper is: How does social media influence inves- tors’ risk-taking behavior? This question is motivated by existing literature indicating that social media platforms impact financial decision-making by amplifying behavioral fi- nance phenomena, such as herd behavior and confirmation bias. My hypothesis is as follows: H1: Social media increases investors’ risk-taking behavior. 6 In the context of this thesis, the term risk-taking behavior is understood as an investor’s behavior that can be examined based on investments they have made. This can be ex- amined, for example, based on their investment portfolio allocations to a high-risk asset class (such as stocks or cryptocurrencies) or higher trading frequency. These indicators typically reflect investors’ higher risk tolerance and their willingness to accept bigger losses in the hopes of achieving higher returns. 1.2 Structure of the paper The thesis is structured into four main chapters following this introduction. The first main chapter, “Theoretical background,” introduces relevant financial and behavioral theories that are essential for understanding investor behavior. The theories included are efficient market hypothesis, behavioral finance theories, social learning theory, and risk percep- tion. The next chapter, “Literature Review,” provides an analysis of existing research that ex- amines the connection between social media and investors’ risk-taking behavior. This chapter discusses empirical findings of the research. Third chapter, “Conclusion”, summarizes the main insights of the reviewed literature and evaluates whether existing research generally confirms or challenges the hypothesized influence of social media. Last chapter, “Limitations and future research”, discusses the main limitations of the the- sis, and main considerations for future research. 7 2 Theoretical background 2.1 Efficient market hypothesis Efficient market hypothesis is one of the most studied and discussed theories in finance. The hypothesis was presented by Eugene Fama (1970), and it assumes that all stock prices reflect all the available information with precision. According to the efficient mar- ket hypothesis, it is impossible to achieve continuous excess returns in relation to the markets because all the new information is transmitted fast to the prices of stocks. Fama (1970) split market efficiency into three categories. First is the weak form, where the prices of the stocks reflect all historical information. Because of that, person cannot achieve excess returns by analyzing historical trends. Second is the semi-strong form. Stock prices include all publicly available information. As news, financial statements, and public analyses are all included in the prices. And last, strong form. Stock prices include all information, including private information. In this case, even insider information is included in the prices. Efficient market hypothesis has attracted a lot of criticism, especially from behavioral finance. The central assumption of the hypothesis is that investors act rationally and use all available information optimally. Behavioral perspective shows that investors’ behav- ior is affected by many psychological factors, such as emotions and biases, which can cause deviations from efficiency (Shiller, 2003). The social media era has questioned the assumption of information’s rational treatment in efficient markets. Especially “meme-stocks” such as GameStop and AMC Entertain- ment have brought up situations where investors' collective behavior has caused signif- icant price changes in the markets, which are hard to justify with fundamentals. From that perspective, social media’s effect on investors' risk-taking challenges the ef- fective market hypothesis assumption of rational financial decision-making and perfect 8 information. Social media can cause instances where investors act irrationally by follow- ing emotion-based narratives. Hence, in the era of social media, it is justified to look critically at the assumptions of the efficient market hypothesis. Even though markets may be efficient in the long term, social media can create substantial and potentially risky deviations from efficient pricing in the short term. 2.2 Behavioral finance theories Traditional financial theory, such as efficient market hypothesis, assumes that investors are rational and markets reflect almost perfectly all available information. Observations and empirical evidence have however shown that real investor behavior may deviate from these assumptions. That is why behavioral finance has emerged as a key research direction that seeks to explain and understand these deviations through psychological factors. Behavioral finance connects economics and psychological observations to focus on how cognitive biases, emotions, and social interactions influence investors’ decision-making. Social media may reinforce these factors as investors are constantly exposed to opinions and stories about investments. In this environment, risk-taking is not only determined by individual analysis but is also influenced by social and emotional stimuli. 2.2.1 Prospect theory Kahneman & Tversky (1979) published their study “Prospect Theory: An Analysis of De- cision Under Risk”. Theory consisted of empirical evidence showing that people do not evaluate their choices based on absolute outcomes but rather in relation to a reference point, such as the current level of wealth or the previous investment price. The theory highlights that people experience losses more intensely than equivalent gains (see Figure 1); this is called loss aversion. For example, 200-dollar loss is experienced approximately two times stronger than equivalent gain. This can create irrational financial decisions as 9 investors might keep their loss-making stocks in their portfolio in hopes that it will bounce back, even if it is justified to move on to fundamentally better investments. Figure 1. Prospect Theory: A hypothetical value function (Kahneman & Tversky, 1979) Kahneman & Tversky (1979) discuss that the larger the gains or losses are, the less changes affect the investor’s experience. For example, 50 dollars additional loss does not feel so bad if the investor has already lost 500 dollars. This kind of behavior can lead to risk seeking after losses, although rational investors should evaluate every decision sep- arately, regardless of previous results. Kahneman & Tversky (1979) show that a person’s risk attitude changes depending on the context. People generally are risk-averse when it comes to winning but risk-seeking in the face of losses. This can encourage investors to withdraw small winnings quickly but keep their losing or stagnated investments in the portfolio to hope for a turnaround. This is called the disposition effect, and it has been studied by Shefrin & Statman (1985). Prospect theory gives us great insight into understanding investor behavior in today’s world. Social media platforms like Reddit and TikTok may affect investors' reference point, as users may compare their own performance to the others, where a person’s own 10 financial goals can change to unrealistic. Additionally, social media culture can underline success, which can enforce risk aversion or encourage risk seeking. Barberis & Huang (2001) showed results that investor behavior, which is based on Pro- spect theory models, can cause irrational reactions and price changes in the markets. This behavior may stand out in social media’s environment, where financial decisions are often made quickly and based on examples of peers and influencers. 2.2.2 Herd behavior It is difficult to define herding, but in its most general form, it could be defined as behav- ior patterns that correlate across individuals (Devenow & Welch, 1996). It does not par- ticularly mean herding if many investors are buying the same “hot” stock, as it could be from correlated information arrival to independently acting investors (Devenow & Welch, 1996). Instead, Devenow & Welch (1996) consider herding as something that can lead to systematic erroneous decisions made by entire populations. Avery & Zemsky (1998) studied the relationship between asset prices and herd behavior. When there are two dimensions of uncertainty – both the existence of a market shock and its impact on asset prices – herding behavior may occur as investors imitate others during uncertainty (Avery & Zemsky, 1998). However, Avery & Zemsky (1998) argued that this herding does not necessarily lead to significant price distortions, as markets tend to discount the reliability of information seen during such episodes. Herding can be divided into two sections, irrational herding and rational herding. Devenow & Welch (1996) describe irrational herding (or non-rational herding) as when investors with insufficient information and no proper rational analysis disregard their prior beliefs and follow each other blindly. 11 Rational herding models consist of one or more of the three main effects: the payoff externalities model, the principal-agent model, and the cascades model (Devenow & Welch, 1996). Payoff externalities occur when the outcomes or returns of an investor’s decision depend on what other investors also do. Devenow & Welch (1996) provided an example of bank runs, where investors withdraw funds because they see others doing the same thing, fearing that the bank will run out of money. This could be linked to modern-day phe- nomena: FoMO, or the Fear of Missing Out. Przybylski et al. (2013) defined FoMO as the widespread anxiety or concern that others may be enjoying rewarding experiences in which one is not participating. Principal-agent theory can be caused when managerial performance is evaluated based on peers rather than on absolute performance (Devenow & Welch, 1996). The model shows that managers tend to mimic the actions of their peers, because if their own ac- tions cause low performance results, they are likely to take the blame. According to Devenow & Welch (1996) it is better for managers to fail as a whole and blame it on the market, this theory suggests that if enough “bad managers” invest in a poor product, it can cause a herding effect to the “good managers”, who would not have invested in that product otherwise. Cascades can be described in situations where people make decisions based on observ- ing other people's decisions, rather than relying on their own private information. Devenow & Welch (1996) suggest that if an investor with negative private information about a certain stock sees other investors previously buying that stock, they may be per- suaded to disregard completely their own private information and buy the stock as well. 12 2.2.3 Confirmation bias The confirmation bias refers to people’s tendency to notice and prefer information that already supports their beliefs, while disregarding the information that contradicts their existing beliefs (Costa et al., 2017). In other words, people naturally focus on the evi- dence that supports their existing belief and overlooks evidence against them. We can also see this in the field of investing. Russo & Schoemaker (1992) highlighted that confirmation bias can lead investors to sys- tematically underestimate risks or overestimate expected returns, resulting in less opti- mal investment decisions. Also, Costa et al. (2017) argued that this phenomenon can cause people to convince themselves about everything they want to believe in. So, where does the confirmation bias come from and why? According to Peters (2022), it evolved because it helps individuals to influence others and shape social interactions to align with their existing beliefs or expectations. Peters (2022) argues that it does not necessarily have to be a bad thing, as confirmation bias can result in significant develop- ment for us and others. It helps us to stay connected to social reality, which makes it easier to navigate social structures (Peters, 2022). Confirmation bias may significantly influence investor behavior on social media plat- forms. If investors believe that some stocks will perform well in the future, they might unconsciously notice social media posts and reactions that align with their preconcep- tions. At the same time, they are more likely to ignore or downplay warnings about that stock. This could lead investors to take greater investment risks based on biased percep- tions rather than objective evaluation. 2.3 Social learning theory Albert Bandura’s paper “Social learning theory” (1977) offers central theoretical frame to understand social learning. According to Bandura (1977), people learn new behavioral patterns by observing the actions of others and their consequences. This learning does 13 not require direct experience or validation, as it happens through observations and mod- eling. Bandura represents four key factors which are necessary in social learning: 1. Attention, learners must pay attention to the model’s behavior. If there are distractions, it will have a negative effect on observa- tional learning. 2. Retention, learners must remember the behavior of models. 3. Reproduction, learners should be able to repeat or mimic the be- havior of models. 4. Motivation, learners must have a wish or a cause to adopt and re- peat the behavior. These factors emphasize that learning is also a cognitive and social process where inter- nal factors of the individual and environmental factors interact. Bandura (1977) presents that people can indirectly learn through others without them experiencing the consequences of their behavior themselves. That is known as vicarious reinforcement, and it makes social learning especially strong in situations where behav- ior can be observed on a large scale, for example in social media. In the context of financial decision making, social learning theory can be used to identify how investors can imitate others, for example finfluencers (financial influencers) behav- ior on the internet. When investors notice positive attention gained by finfluencers, they might be inclined to adopt similar patterns and procedures. Long-term exposure to this kind of behavior can normalize certain forms of risk behavior, where individual learns passively through models. Also, social media platforms are designed so that they highlight prominent users and their content. This makes it easier for users to copy their own behavior to match those 14 that they find successful, confident and popular. Bandura’s (1977) social learning theory gives us tools to understand how publicity, social influence and perceived credibility on social media can affect investors that are affected by social media. 2.4 Risk perception Normally, the term “risk” is associated with a negative meaning – it is related to a possi- ble harm, loss, or an undesirable event (Ricciardi, 2004). However, in the case of invest- ing and finance, risk is associated with more of an opportunity to achieve or not to achieve (Ricciardi, 2004). When investors look to invest in a certain stock, they evaluate how much risk they are willing to take. In the stock markets, it is generally accepted that investors who take greater risks have the potential for greater returns. The rule applies in both directions: the more risks you take, the more you have the potential to lose. According to modern portfolio theorists, there is a great trade-off in investing, which is between risk and re- turn (Ricciardi, 2004). The main idea is that investments with a risk should have the po- tential for a greater return than risk-free investments (Ricciardi, 2004). According to Ricciardi (2004), the concept of perceived risk has been studied in con- sumer behavior since the 1960s. From the consumers' perspective, risk is often con- nected with uncertainty, and purchasing some products or services may give satisfaction and approval, or it may lead to potential disappointment (Ricciardi, 2004). This so-called perceived risk means that consumers assess the risk of purchases based on their own experience and observations. Consumer behavior research and behavioral finance are similar in this respect, as both consumers and investors make decisions through similar psychological mechanisms (Ricciardi, 2004). In the business field, the concept of perceived risk applies to many areas, such as con- sumer behavior, marketing, and investment (Ricciardi, 2004). In these areas, risk percep- tion refers to people’s feelings or perceptions of how great they think the risk is, 15 regardless of how great the actual or objective risk is (Ricciardi, 2004). Ricciardi (2004) continues that if a person does not have enough correct information, or they have false information, they might easily misjudge the risk, which may lead to incorrect decisions and assessments. 16 3 Literature review 3.1 Social media as a new information environment Social media’s rise has fundamentally changed how investors gather information, ana- lyze markets, and make financial decisions. Rather than traditional news outlets, social media platforms offer faster and user-based content, which may influence market senti- ments almost immediately. Financial information that is available to everyone brings its pros and cons: it enables broader market participation, while at the same time it may strengthen biases, speculation, and herd behavior. Many studies have argued that expressed emotions and views on social media may pre- dict stock market movements. According to research by Hailiang et al. (2014), sentiments expressed in articles published on social media platforms had predictive value regarding future stock returns and earnings surprises. One significant study that illustrates the influence of the media on investor behavior is Tetlock’s (2007) study, which analyzed the impact of the pessimistic tone of news on the stock market. Tetlock (2007) found that negative media sentiment forecasted low stock returns in the short term, but it corrected itself in the long term. This could point out that media may trigger emotional reactions that are not based on fundamentals. Figure 2 points out that if pessimism predicts future investor psychology, stock prices will tem- porarily decline, and if pessimism reflects past sentiment, the long-term recovery may be stronger. 17 Figure 2. Impact of a negative sentiment shock on prices (Tetlock, 2007). Similarly, Antweiler & Frank (2004) found that stock message board discussions contain useful predictive signals of market volatility. Alongside these observations it is also critically important to identify risks posed by social media regarding the reliability of information. Social media platforms do not only spread relevant information fast, but also spread misinformation, unfounded rumors and irra- tional optimism. Even though social media democratizes access to information, at the same time it may enable more risk for investors as they are exposed to misinformation and possible market manipulation. Findings suggest that social media users’ collective “wisdom” may enhance market fore- cast, but it also shows that social media represents new kind of information environment where traditional financial evaluation methods may not be relevant. This places new 18 kinds of requirements for investors' critical thinking and information literacy when mak- ing financial decisions. 3.2 Information overload Information overload means a situation where the amount of information received by an individual exceeds their cognitive abilities to process it effectively. With investing in mind, this topic is particularly relevant because social media, news and other digital plat- forms produce huge amounts of information every day. Nobel prize winner Herbert Si- mon famously said: “a wealth of information creates a poverty of attention”, referring that an individual’s decision-making abilities weaken when there is too much infor- mation available. Simon’s paper (1972) about bounded rationality helps even today to understand investor behavior. Simon (1972) discovered that people cannot make perfect rational decisions, because they have limited ability to process time, information and computing power. And because of this individual strive for a satisfactory decision rather than an optimal one, as they choose the first option that meets their minimum criteria. Bernales et al. (2023) study the information overload’s effect on the stock market by creating an index, that measures the news stream and its effect on investor behavior. Study shows that excessive amount of information forecasts higher market returns for an 18-month period, because investors demand a higher risk premium to fight against uncertainty (Bernales et al. 2023). Bernales et al. (2023) discuss that especially small, high beta, and unstable stocks react to the information overload more aggressively. Agnew & Szykman (2005) in turn, focus on how limited financial knowledge combined with complex and abundant information leads to investors’ tendency to choose the de- fault option in retirement savings. This might represent how information overload can passivate investors and reduce the quality of decisions, especially those with weaker 19 cognitive abilities. Agnew & Szykman (2005) study supports Simon’s view that individuals use simplifying rules known as heuristics to deal with complex choices. Engelberg et al. (2012) demonstrate that financial “news”, in this case Jim Cramer’s pop- ular “Mad Money” affects investors’ attention, which reflects on stock market prices and trading volume. Stock recommendations made by the show led to large overnight re- turns but were reversed over the next few months (Engelberg et al. 2012). Mad Money is known to be an intensive and fast-paced show that could alter investors’ ability to process all the information Cramer is representing. This could support the claim that information overload is not only an individual’s cognitive challenge but also a larger mar- ket factor, in this case short-term. Engelberg et al. (2012) study findings are somewhat aligned with Bernales et al.'s (2023) findings. They both support the claim that media and other digital platforms may create information overload, which affects individuals’ financial decisions. Bernales et al. (2023) strive to explain information overload effects through risk premi- ums while Agnew & Szykman (2005) emphasize the role of passivity and simplicity. In- formation overload can affect both risk-taking and decision-making differently. In social media, understanding the effects of information overload is especially im- portant. For example, platforms such as Reddit and Twitter create large quantities of financial information for us to observe, sometimes even harmful. This enables investors to make decisions based on incomplete or misunderstood information, which might make investors overreact or withdraw from the market. Information overload emerges as a phenomenon that not only challenges investors’ decision-making ability, but also the assumptions of the efficient market. 3.3 Social media and behavioral finance biases As social media platforms have become more common, they might create environment where investors are exposed regularly to decision-making biases identified in behavioral 20 finance. Behavioral finance theories suggest that investors do not always act rationally as psychological and social factors such as herd behavior, confirmation bias, and different heuristics may affect their decisions. Social media features like information spread, al- gorithmic personalization, and collective discussion may enhance biases and lead to in- creased risk-taking among investors. 3.3.1 Herd behavior and social media environments Social media may have contributed to the rise of herding behavior in the financial mar- kets. Social media platforms have attracted a lot of retail investors, who have been em- powered by other retail investors to act as a herd. This can create a strong price move- ment in the markets that do not have fundamentals backing it (Bikhchandani & Sharma, 2000). Barber et al. (2022) showed that investment platform Robinhood’s users did significantly more attention-induced trading than other investors. According to Barber et al. (2022), investors’ trading activity focused especially on stocks that were actively discussed on social media. The study also found that this attention-induced trading activity broad- casted negative returns in the short term, which may emphasize social media’s and dig- ital investing platforms' potential negative effects on investors' rational decision-making. A significant example of herding behavior comes from the 2021 GameStop phenomenon, where Reddit’s WallStreetBets community’s users coordinated large-scale stock pur- chases. This led institutional investors to close their “short” positions, which caused the stock to experience a sharp price increase. A study by Anand & Pathak (2022) showed that Reddit’s WallStreetBets community’s positive tone and sentiment displayed signifi- cant predictive associations with GameStop stock return, volatility, bid-ask spreads, and volumes (see Figure 3). Anand & Pathak (2022) also mention that a tiny minority of 462 most influential users had the most major impact on the GameStop stock’s fluctuation. Even though the whole 21 community had an impact on the markets, the biggest impactor was a small and influ- ential group of users and not the whole collective force of the community. This raises questions as how evenly social media’s impact distributes across investors and whether the dynamics of herd behavior are after all driven by a small influential group instead of large masses. Figure 3. GME volatility (solid) and WSB subreddit tone (dotted) data from December 2020 – March 2021 (Anand & Pathak, 2022). From the perspective of institutional investors, herd behavior has been studied by Frijns & Huynh (2018), who focused on analyst recommendations. They found that analysts’ tendency to herd varied significantly based on the tone of the media. Security analysts' herding showed that when there is high media coverage of the company, analysts tend to herd less towards consensus recommendations. However, when there is negative news sentiment, herding increases among analysts as they are unwilling to stand out from the crowd. And when the is a strong disagreement among media sources, analysts tend to follow the consensus more (Frijns & Huynh, 2018). There were also strong per- sonal incentives when and why analysts herd. According to Frijns & Huynh (2018), there 22 were more herding among experienced analysts after negative news, likely due to con- cerns about their reputation and business relationships as they have more to lose. Also, inexperienced analysts tend to deviate more from the consensus when the media cov- erage is strong or negative, this was likely due to desire for more publicity (Frijns & Huynh, 2018). Frijns & Huynh (2018) conclude the results suggesting that analysts’ herding behavior is driven by incentives and career concerns, and their herding is indeed rational, because they maximize their utility. Overall, the study suggests that media coverage can both reduce and increase herding among analysts, strongly depending on the tone of the me- dia and news, and from the analysts’ personal incentives. Previous studies suggest (Barber et al., 2022; Anand & Pathak, 2022; Frijns & Huynh, 2018) that herding appear among retail and institutional investors, however the reasons and dynamics behind them can differ significantly from each other. Institutional inves- tors, in this case analysts, herd primarily because of their career incentives. Instead of providing independent stock recommendations, analysts may conform to the consensus, which more reinforces market trends. If this conformation continues, it can lead to a collective downgrade and fire selling of the stock, without fundamentals backing it. The GameStop case showed how retail investors can coordinate a viral movement that disrupts institutional expectations and causes massive market volatility. Retail investors did not herd because of their career incentives, but for social validation and FoMO (fear of missing out). Both studies show how herding can cause market volatility (risk), which may not be easily identified. The thesis hypothesis aligns with the studies discussed re- sults. 3.3.2 Confirmation bias and social media’s personalization When investors are looking for information in financial decision making, they may look for things that reinforce their preconceptions, and at the same time ignore contradictory 23 evidence. This cognitive bias is called confirmation bias and has been widely studied in the academic field (Nickerson, 1998; Peters, 2022). Confirmation bias may have affected investors’ decision making previously (Rabin & Schrag, 1999), and this chapter discusses has social media’s rise intensified its effect. Social media platforms filter and tailor content based on algorithms. This predisposes investors to see news and conversations that reinforce their own thinking. It can cause investors who use social media actively to become overconfident in their financial deci- sions, reject alternative perspectives, and take unnecessary risks. This phenomenon does not only apply to social media but has also been observed more broadly in internet- based investment behavior. For example, Barber & Odean (2001) demonstrated that online investors traded more actively, received weaker returns, and were more overcon- fident about their own decisions. There were also signs of under diversification among the online investors. Under diversification enables more risk to the investor, which they might not notice. One of the contributions to this could be information selection on the “hot” or “booming” industries as they might selectively look for news and analyses that reinforce their existing beliefs, a key concept of confirmation bias. Study from Park et al. (2013) support this view with their research on online investment communities. Research consisted of surveys and a controlled field experiment on Korean investors’ message boards. Park et al. (2013) found that investors preferred systemati- cally messages that were aligned with their prior opinions and perceived expertise. This information selection led to overconfidence and to an increase in optimism, which in turn caused excessive trading and unrealistic expectations about returns to investors (Park et al., 2013). Study highlights that increased financial information in the hands of investors may not lead to better returns. As it might weaken investors’ rationality, be- cause they may become more confident about their own abilities. In turn, social media’s personalization may be a stronger explanatory factor than com- munity dynamics. Audrino et al. (2020) expanded the perspective by examining online 24 data, such as Google search volume and messages posted to “StockTwits”, effect on short-term volatility of the stock market. Audrino et al.'s (2020) findings suggest that investors’ attention, which was particularly visible in online discussions, was linked to increased short-term market volatility. Social media platforms offer its users content that matches their prior beliefs and interests, because of algorithms. The result is the emer- gence of rapidly strengthening collective opinions, which may explain the observed strong short-term market reactions. Also, Sprenger et al. (2014) reviewed how Twitter’s stock message sentiment and quan- tity influenced to the stock market returns and trading volume. They observed that pos- itive sentiment (bullish) correlated with the return of stocks and actively shared mes- sages anticipated the next day's trading volume growth. Furthermore, users whose stock advice was better than average gained more followers, and their messages were spread more. Results may suggest that the cause of the confirmation bias among the investors can be more of a technological (algorithms via sentiment) than a social (community structure). It can also be argued that although confirmation bias can increase investors’ confidence, it can also reinforce investors' caution if the person is risk averse. This may lead to more controlled risk assessments, such as portfolio diversification which reduces the risk taken by the investor. The studies presented above (Barber & Odean, 2001; Park et al., 2013; Audrino et al., 2020; Sprenger et al., 2014) show that social media platforms reinforce investors’ exist- ing beliefs. Even though the studies agreed that confirmation bias is visible in digital in- vesting, their explanatory models differ. Park et al. (2013) emphasize community dynam- ics, as Audrino et al. (2020) and Sprenger et al. (2014) refer to sentiment and attention as a main factor. This also raises questions: does confirmation bias originate in the digital age from technological, social structures, or from a combination of both? 25 Based on these studies, it can be stated that confirmation bias is a significant phenome- non in the context of the digital investing environment, but the underlying mechanisms and consequences need deeper and more comparative research. It can be stated that the hypothesis (H1) “social media increases investors' risk-taking behavior” is supported, but not unequivocally confirmed. 3.3.3 Risk perception under social media influence In recent years, social media platforms have become an essential part of the financial environment, acting as primary sources of information and discussion for some investors. This change may affect how investors perceive and manage risks. Social media’s influence on financial markets may be deeper than the general sentiment of investors. Social media can change rapidly investors’ perception of risk as information spreads very quickly. Karppi & Crawford (2015) described this phenomenon with a case from 2013, when the news organization Associated Press (AP) Twitter account was hacked. The account posted a false tweet, saying there was an explosion in the White House. This tweet caused an immediate 136,5 billion dollar decrease in the S&P 500 - index value (Karppi & Crawford, 2015). Karppi & Crawford (2015) discovered that the tweet spread rapidly, creating patterns of repetition, which confused even more human and non-human traders, as they could see the news from different sources. Study shows how automatic trading systems and human traders react immediately to unconfirmed information. The case shows two critical problems with the markets risk perception affected by social media. Firstly, usage of social media as a real-time news source may lead to overreaction, especially when investors hold the news as a reliable source before it’s confirmation. Secondly, social media can cause market volatility by amplifying the spread of misinfor- mation. These can lead to panic selling and rapid value changes, which may not be justi- fied considering actual market conditions. Case may explain that in the digital age inves- tors’ perceptions of risk may change rapidly as information comes to us faster than ever. 26 Although the case is exceptional, it brings up social media’s ability to influence investor’s risk perception suddenly and extensively. It also points that market reactions do not come from singular investors’ actions but are the joint result of many factors, including automated systems. Zhao & Li (2024) in turn brings new perspective to the risk perception. They examine how social media usage affects financial decisions of families. The study analyzed infor- mation from 75 different countries from the years 2000-2022 and their research suggests that increase in use of social media is connected to changes in family risk perception, which in turn influences the investment decisions and strategies they make. Also, social media may act as a central channel, where families get their financial information. This may affect their financial literacy and how they approach risk management (Zhao & Li, 2024). As Karppi & Crawford (2015) and Zhao & Li (2024) examine different levels of risk per- ception, they both refer that social media can affect risk perception fast, extensively and undirect. The main difference is that in the case of Karppi & Crawford, the reaction is based on the quick spread of false information and collective panic, as in the case of Zhao & Li, they show long-term impact on the decisions of the families. However, the studies leave several questions open. Firstly, to what extent are changes in investors' and families' risk perceptions a result of algorithmically generated content, and to what extent is it an active information-seeking behavior? Also, is there a change in risk itself or is it only how the risk is felt and reacted? And lastly, how permanent these changes are, is it only an instantaneous reaction or a long-lasting transition in financial behavior? Social media’s impact on risk perception is a multidimensional phenomenon, where technological, cognitive, and social factors collide. While studies point out the 27 connection between social media and risk perception, there is a need for more compar- ative research. 3.4 Influencer-driven market dynamics Social media influencers have become more visible in today’s financial markets, as they have been seen to direct investor interest and create rapidly spreading interest in specific investment targets. These “finfluencers” (financial influencers) utilize their massive amounts of followers and influence, which can result in individual investment targets temporarily rising in value. A recent study by Keasey et al. (2024) analyzes how influencers affect the financial mar- ket performance of firms. Keasey et al. (2024) study shows that especially social media’s big influencers, so-called mega-influencers (influencers with over 1 million followers), can quickly direct investor attention and increase the trading volume of shares. The study found that influencers' effect on increased stock returns was insignificant, and they were only temporary. This may suggest that while influencers may not have an impact on per- manent firm value, however they may be able to generate quick and strong market re- actions and increase volatility simply through their social media visibility. The results of the study also highlight how investors' risk perception can be distorted by influencers' posts. When investors react to influencers' posts, market volatility and transactions may increase, which may lead to financial decisions that are based on emotions and social media trends rather than traditional financial analysis. We have also seen this in the case of the “GameStop short squeeze”. As Keasey et al. (2024) focused on influencers' external market impact, Hayes & Ben- Shmuel (2024) discuss how influencers build their influence in the social media environ- ment. According to Hayes & Ben-Shmuel (2024), finfluencers do not act as institutional professionals, but their credibility is formed through the communicative strategies fa- vored by algorithms. This means that the influencer’s influence does not come from the quality of their advice but rather from their ability to produce content which resonates 28 with their followers. It is more about personal brand, relatability and narrative than an- alytical proficiency. Hayes & Ben-Shmuel’s (2024) study points out that finfluencers participate in four forms of communication: building narratives, information, discussion and promotion. They can tell personal stories about prosperity, a debt-free lifestyle, share tips to investments and create community spirit through the comment section and livestreams. Through these, they not only pass on information but also modify their followers' understanding of in- vesting and financial management. Hayes & Ben-Shmuel (2024) point out that social media’s usage, especially among the youth, is part of their daily life, and financial content merges easily as a part of enter- tainment content. This normalizes investing, but at the same time it personalizes risk, as financial decisions are made more and more from the perspective of identity, status, and peer recognition. This development is contradictory. Even if it could add financial participation and make investing more attractive, it can further add misleading content impact and lead to deci- sions that are not based on a realistic risk assessment. Social media platform’s algorith- mic structure, preference for short-form content, and the lack of transparency in com- mercial partnerships can all weaken investors’ ability to distinguish between information, entertainment, and advertising. It can be stated that influencer-based investment trends are not only consequences of recommendations given by finfluencers, but their effect depends on many psychological factors, like the quantity of followers and perceived authenticity. Hence, finfluencers ef- fect on the markets may not be linear, as great visibility can both accelerate and dilute market reactions depending on context and how the audience receives it. 29 3.5 Narrative economics Narrative economics explores economic narratives in market dynamics and individuals’ financial decisions. According to Robert Shiller (2017), narratives can spread like an epi- demic and significantly influence economic situations such as market bubbles and reces- sions. Shiller (2017) says that narratives influence is based on their ability to create com- mon understanding and expectations, which control people’s actions. Shiller’s (2017) ap- proach is macro-level, as he focuses on how narratives become established as societal beliefs and create new market dynamics. Johnson et al. (2023) complement Shiller’s views by taking along Conviction Narrative Theory (CNT), which means that narratives not only influence by spreading collectively, but they also shape individuals’ inner psychological structures in uncertain decision-mak- ing situations. CNT highlights that narratives bring conviction to the decision-making pro- cess, which helps individuals to make choices in uncertain situations. CNT differs from Shiller’s approach in that it focuses more on individual-level processes, while Shiller (2017) views narratives spread as collective and cultural phenomena. Although the term “CNT” was used by Johnson et al. (2023), Chong & Tuckett (2014) introduced the term earlier. Chong & Tuckett (2014) studied how fund managers use conviction narrative to further help their decision-making process in times of uncertainty. Fund managers need to sell themselves as an expert to their clients and to themselves, in which conviction helps them (Chong & Tuckett, 2014). Roos & Reccius (2024) evaluates narrative economic approaches critically as they em- phasize it’s theoretical and methodological challenges. According to Roos & Reccius (2024) narratives measurement and definitions are often unclear, which makes it difficult to conduct empirical research. They suggest more precise conceptual definition and more diverse research methods to better study the effects of narratives. Roos & Reccius (2024) also recognizes that empirical confirmation of narratives effect faces challenges, 30 especially in demonstrating cause and effect relationships, which more limits the effec- tiveness of current research. Previously, Tuckett et al. (2014) conducted empirical research by analyzing narratives in unstructured data sets. Tuckett et al. (2014) reported that narratives can cause changes in investor behavior especially in situations where traditional financial analysts and fun- damentals cannot explain market movements. Tuckett et al.’s study is aligned with Shiller’s (2017) thoughts that narratives can inflate markets. Taffler et al. (2024) studies narrative emotions and market crises. Study shows that emo- tional content of narratives can trigger collective behavior in times of crises. Taffler et al. (2024) highlights the effect of storytelling, especially when institutional trust is missing. Shiller (2017) represents that economic narratives can influence peoples’ behavior and through that to the economy. Shiller remarks that these stories do not always directly bear from economic events, but can be creative and surprising, and still have a great impact if they spread widely (Taffler et al. 2024). Taffler et al. (2024) continues that sto- ries effect on the investors just because they evoke emotions. When the story touches the individual emotionally, they might react quickly and intuitively, not with a sense. Figure 4 presents a diagram based on Taffler’s et al. (2024) study that illustrates the in- teraction between narratives, investor sentiment and market behavior. Taffler et al. (2024) illustrates that economical narratives are built from events, speculation and ru- mors, which shape investors emotional response. This emotional response influences investment decisions, market prices and volatility. At the same time, market behavior is reflected in new narratives and further influences investors and the media (Taffler et al. 2024). 31 Figure 4. Interrelationship between narratives, emotions, and market behavior (Taffler et al. 2024) Wisniewski & Yekini (2015) address the narratives effect in corporate communications. Wisniewski & Yekini (2015) show that the tone and structure of corporations’ annual reports can have an impact on stock market reactions. The study found that active and realistic language in the annual reports predicted higher stock returns. Wisniewski & Yekini (2015) also found that higher “activity” and “realism” values were not connected to higher risk that the investor might face. This might suggest that the predicted returns are not just risk premium, but potential deviations from rational pricing. This challenges the concept of rational markets and Fama’s (1970) concept of semi-strong markets in Effective Market Hypothesis, as information from annual reports is public information. The results need to be taken with a grain of salt, because the correlation coefficient was low. 32 The study of Wisniewski & Yekini (2015) represents the technical language approach and its effects in a one-year period in stock markets. This differs from Taffler et al. (2024) study how investors’ emotions such as fear and hope create narratives. These studies have common however, is that they show how the impact of narratives extends beyond just information, as they influence how information is experienced and interpreted. Between the studies there are clear differences in how narratives are studied and ob- served. Shiller (2017) and Roos & Reccius (2024) examine macro-level narratives on how they are born and spread in society. Johnson et al. (2023), Chong & Tuckett (2014) and Taffler et al. (2024) in return discuss the psychology of narratives and emotions. Wisniewski & Yekini (2015) show practical examples on how we can possibly measure narratives and use them in our financial decisions. Narrative economics offers a great and important framework to understand investors’ course of action in times of turmoil. Even though studies approach the topic from differ- ent perspectives, they all support the idea that narratives can influence financial deci- sions significantly. 33 4 Conclusion This thesis examines the social media’s impact on investors’ risk-taking behavior. The aim of the study is to understand what kind of behavior social media may activate and how that can lead to increased risk-taking. The thesis theoretical framework is built around behavioral finance theories, efficient market hypothesis and social learning theory. Literature review is divided between dif- ferent psychological and social phenomena such as, herd behavior, confirmation bias, narratives, financial influencers and information overload. The research question which was: “How does social media influence investors’ risk-tak- ing behavior?” was addressed through the hypothesis: H1: “Social media increases investors' risk-taking behavior.” Based on the thesis, we can see that social media affects investors’ decision-making in ways in which many of them support the hypothesis. Herding behavior was particularly evident in individual phenomena, such as the GameStop case. Confirmation bias led in- dividuals to seek out information that supported their preconceptions, thus increasing their self-confidence and willingness to take risks. This was seen as investors preferred higher volatility asset classes such as stocks. Narratives, like the story-based explanations of economic events raised by Shiller and Tuckett, influence investors’ emotional state and can thus increase impulsivity and devi- ation from rational thinking. This is especially emphasized when investors must make decisions in the middle of uncertainty and information overload, which social media can add. Even though empirical evidence somewhat supports the hypothesis, research results were not always unambiguous. Some of the research point out that social media can also increase investor cautiousness or act as tool for learning. Also differences between 34 individuals, such as experience, educational level and cognitive capacity may affect how sensitively they respond to social media stimuli. It can be stated that social media offers complex frames where investor behavior can alter based on social, cognitive and emotional influences. Social media can increase in- vestors’ risk-taking in situations where emotions, biases and collective narratives control decision-making. Thus, the hypothesis is supported, but the effect is not unambiguous. 35 5 Limitations and future research Thesis is based on previous finance literature and research, and no empirical evidence has been produced to support it. This limits the possibility to make strong conclusions about how social media affects investors’ risk-taking behavior. Also, most of the dis- cussed studies focus on specific countries, in this case mostly the USA. For future re- search it could be useful to utilize findings from broader geographical contexts and dif- ferent social media platforms, which would allow for more comprehensive view of the global impact of social media on investor behavior. This thesis also did not analyze differences between different investor groups. Future research could investigate, for example, how differently young and more experienced investors differ in their reactions to social media content. Surveys and studies focusing on demographic factors could provide additional information on how social media may affect investors’ risk perceptions and decision-making processes of different age groups. Social media’s content and dynamics itself offers a great opportunity for further research. Finfluencer’s growing influence over the investors raises concerns about the infor- mation’s reliability and transparency, especially in the cryptocurrency world. Experi- mental studies could help us to better understand how social media’s narratives and emotional storytelling may affect investors’ risk-taking. In addition, future research could examine how investors identify, evaluate and process potentially misleading information when they are making financial decisions. 36 References Agnew, J. R., & Szykman, L. R. (2005). Asset allocation and information overload: The influence of information display, asset choice, and investor experience. The Journal of Behavioral Finance, 6(2), 57-70. https://doi.org/10.1207/s15427579jpfm0602_2 Anand, A., & Pathak, J. (2022). The role of Reddit in the GameStop short squeeze. Economics Letters, 211, 110249. https://doi.org/10.1016/j.econlet.2021.110249 Antweiler, W. and Frank, M.Z. (2004), Is All That Talk Just Noise? The Information Content of Internet Stock Message Boards. The Journal of Finance, 59: 1259-1294. https://doi- org.proxy.uwasa.fi/10.1111/j.1540-6261.2004.00662.x Audrino, F., Sigrist, F., & Ballinari, D. (2020). The impact of sentiment and attention measures on stock market volatility. International Journal of Forecasting, 36(2), 334-357. https://doi.org/10.1016/j.ijforecast.2019.05.010 Avery, C., & Zemsky, P. (1998). Multidimensional uncertainty and herd behavior in financial mar- kets. American economic review, 724-748. http://www.jstor.org/stable/117003 Bandura, A. (1977). Social learning theory (Vol. 1, pp. 1-46). Englewood Cliffs, NJ: Pren- tice hall. Barber, B. M., & Odean, T. (2001). The Internet and the Investor. The Journal of Economic Per- spectives, 15(1), 41–54. http://www.jstor.org/stable/2696538 Barber, B. M., Huang, X., Odean, T., & Schwarz, C. (2022). Attention-induced trading and returns: Evidence from Robinhood users. The Journal of Finance, 77(6), 3141-3190. https://doi.org/10.1111/jofi.13183 37 Barberis, N., & Huang, M. (2001). Mental accounting, loss aversion, and individual stock re- turns. the Journal of Finance, 56(4), 1247-1292. https://doi.org/10.1111/0022- 1082.00367 Bernales, A., Valenzuela M., & Zer, I. (2023). “Effects of Information Overload on Financial Mar- kets: How Much Is Too Much?” International Finance Discussion Papers 1372. Washing- ton: Board of Governors of the Federal Reserve System. https://doi.org/10.17016/IFDP.2023.1372 Bikhchandani, S., Sharma, S. (2000). Herd Behavior in Financial Markets. IMF Econ Rev 47, 279– 310. https://doi.org/10.2307/3867650 Chong, K., & Tuckett, D. (2014). Constructing conviction through action and narrative: how money managers manage uncertainty and the consequence for financial market func- tioning. Socio-Economic Review, 13(2), 309-330. https://doi.org/10.1093/ser/mwu020 Costa, D.F., de Melo Carvalho, F., de Melo Moreira, B.C. (2017). Bibliometric analysis on the as- sociation between behavioral finance and decision making with cognitive biases such as overconfidence, anchoring effect and confirmation bias. Scientometrics 111, 1775–1799. https://doi.org/10.1007/s11192-017-2371-5 Devenow, A., & Welch, I. (1996). Rational herding in financial economics. European economic review, 40(3-5), 603-615. https://doi.org/10.1016/0014-2921(95)00073-9 Engelberg, J., Sasseville, C., & Williams, J. (2012). Market madness? The case of mad money. Management Science, 58(2), 351-364. https://doi.org/10.1287/mnsc.1100.1290 Fama, E. F. (1970). Efficient Capital Markets: A Review of Theory and Empirical Work. The Journal of Finance, 25(2), 383–417. https://doi.org/10.2307/2325486 38 Frijns, B., & Huynh, T. D. (2018). Herding in analysts’ recommendations: The role of media. Jour- nal of Banking & Finance, 91, 1-18. https://doi.org/10.1016/j.jbankfin.2018.03.010 Hailiang, C., Prabuddha, D., Yu, U., Byoung-Hyoun, H., (2014). Wisdom of Crowds: The Value of Stock Opinions Transmitted Through Social Media, The Review of Financial Studies, Vol- ume 27, Issue 5, Pages 1367–1403. https://doi.org/10.1093/rfs/hhu001 Hayes, A. S., & Ben-Shmuel, A. T. (2024). Under the finfluence: Financial influencers, economic meaning-making and the financialization of digital life. Economy and Society, 53(3), 478- 503. https://doi.org/10.1080/03085147.2024.2381980 Johnson, S. G., Bilovich, A., & Tuckett, D. (2023). Conviction narrative theory: A theory of choice under radical uncertainty. Behavioral and Brain Sciences, 46, e82. https://doi.org/10.1017/S0140525X22001157 Karppi, T., & Crawford, K. (2015). Social Media, Financial Algorithms and the Hack Crash. Theory, Culture & Society, 33(1), 73-92. https://doi.org/10.1177/0263276415583139 Keasey, K., Lambrinoudakis, C., Mascia, D. V., & Zhang, Z. (2024). The impact of social media influencers on the financial market performance of firms. European Financial Manage- ment. https://doi.org/10.1111/eufm.12513 Matthew Rabin, Joel L. Schrag, (1999). First Impressions Matter: A Model of Confirmatory Bias, The Quarterly Journal of Economics, Volume 114, Issue 1, Pages 37– 82, https://doi.org/10.1162/003355399555945 Nickerson, R. S. (1998). Confirmation bias: A ubiquitous phenomenon in many guises. Review of general psychology, 2(2), 175-220. https://doi.org/10.1037/1089-2680.2.2.175 39 Park, J., Konana, P., Gu, B., Kumar, A., & Raghunathan, R. (2013). Information valuation and con- firmation bias in virtual communities: Evidence from stock message boards. Information Systems Research, 24(4), 1050-1067. https://doi.org/10.1287/isre.2013.0492 Peters, U. (2022). What Is the Function of Confirmation Bias?. Erkenn 87, 1351–1376. https://doi.org/10.1007/s10670-020-00252-1 Przybylski, A. K., Murayama, K., DeHaan, C. R., & Gladwell, V. (2013). Motivational, emotional, and behavioral correlates of fear of missing out. Computers in human behavior, 29(4), 1841-1848. https://doi.org/10.1016/j.chb.2013.02.014 Ricciardi, V., (2004). A Risk Perception Primer: A Narrative Research Review of the Risk Percep- tion Literature in Behavioral Accounting and Behavioral Finance. http://dx.doi.org/10.2139/ssrn.566802 Roos, M., & Reccius, M. (2024). Narratives in economics. Journal of Economic Surveys, 38, 303– 341. https://doi.org/10.1111/joes.12576 Russo, J. E., & Schoemaker, P. J. (1992). Managing overconfidence. Sloan management re- view, 33(2), 7-17. Shefrin, H., & Statman, M. (1985). The disposition to sell winners too early and ride losers too long: Theory and evidence. The Journal of finance, 40(3), 777-790. https://doi.org/10.1111/j.1540-6261.1985.tb05002.x Shiller, R. J. (2003). From efficient markets theory to behavioral finance. Journal of economic perspectives, 17(1), 83-104. https://doi.org/10.1257/089533003321164967 Shiller, Robert J.. (2017). "Narrative Economics." American Economic Review, 107 (4): 967–1004. https://doi.org/10.1257/aer.107.4.967 40 Simon, H. A. (1972). Theories of bounded rationality. Decision and organization, 1(1), 161-176. Sprenger, T. O., Tumasjan, A., Sandner, P. G., & Welpe, I. M. (2014). Tweets and trades: The in- formation content of stock microblogs. European Financial Management, 20(5), 926-957. https://doi.org/10.1111/j.1468-036X.2013.12007.x Taffler, R. J., Agarwal, V., & Obring, M. (2024). Narrative Emotions and Market Crises. Journal of Behavioral Finance, 1-21. https://doi.org/10.1080/15427560.2024.2365723 Tetlock, P. C. (2007). Giving content to investor sentiment: The role of media in the stock mar- ket. The Journal of finance, 62(3), 1139-1168. https://doi.org/10.1111/j.1540- 6261.2007.01232.x Tuckett, D., Smith, R. E., & Nyman, R. (2014). Tracking phantastic objects: A computer algorith- mic investigation of narrative evolution in unstructured data sources. Social Net- works, 38, 121-133. https://doi.org/10.1016/j.socnet.2014.03.001 Wisniewski, T. P., & Yekini, L. S. (2015). Stock market returns and the content of annual report narratives. In Accounting Forum (Vol. 39, No. 4, pp. 281-294). https://doi.org/10.1016/j.accfor.2015.09.001 Zhao, Z., & Li, X. (2024). Social media and family investment behavior. Finance Research Let- ters, 61, 104945. https://doi.org/10.1016/j.frl.2023.104945