Multilayer networks are attracting growing attention in many fields, including finance. In this paper, we develop a new tractable procedure for multilayer aggregation based on statistical validation, which apply to investor networks. Moreover, propose two other improvements their analysis: transaction bootstrapping and categorization. The can be used integrate security-wise time-wise information about trading networks, but it is not limited fact, different applications, such as gene,...

Recent studies using data on social media and stock markets have mainly focused predicting returns. Instead of price movements, we examine the relation between Facebook investors' decision making in with a unique transactions Nokia. We find that decisions to buy versus sell are associated especially for passive households also nonprofit organizations. At same time, it seems more sophisticated investors---financial insurance institutions---are behaving independently from activities.

In recent years, methods from network science have been rapidly gaining traction in economics and finance. One reason for this is that, a globalized world, it crucial that we understand the interconnections between economic financial entities; networks provide natural framework representing studying such systems. paper, survey use of network-based to study economy- related questions. We start with brief overview graph theory some basic definitions. Then, discuss descriptive measures complex-...

The starting point of this paper is that neighboring investors may talk to each other sharing information about their transactions in stock markets, leading similar trading behavior. We find pairwise trade timing similarities between investor pairs are negatively associated geographical distance corresponding pairs. This suggests local transfer channels individual used decision making. also observe differences age and language moderate association. analysis conducted using level data from...

The complex networks approach has been gaining popularity in analysing investor behaviour and stock markets, but within this approach, initial public offerings (IPO) have barely explored. We fill gap the literature by clusters first two years after IPO filing Helsinki Stock Exchange using a statistically validated network method to infer links based on co-occurrences of investors' trade timing for 69 stocks. Our findings show that rather large part similar structures form different...

We introduce a novel method to identify information networks in stock markets, which explicitly accounts for the impact of public on investor trading decisions. show that has clear effect empirical networks' topology. Most importantly, our strengthens identified relationship between investors' network centrality and returns. Furthermore, when less significant links are removed, association returns becomes statistically economically stronger. Findings suggest approach leads more precise...

Despite the success of machine learning models, literature lacks their applications to identify exploitation non-public information. We address this gap by developing a tool, which ranks investors based on suspiciousness. achieve predicting future trading decisions retail social connections in an insider network. Particularly, high predictability investor's behavior with data her/his neighborhood can indicate that she/he takes advantage and trades Our system captures complex cyclical...

Abstract Previous studies suggest that individuals sharing similar characteristics establish stronger social relationships. This motivates us to examine what combinations of socioeconomic investor attributes are more likely be associated with joint trading behavior. We use a unique data set on actual ties between investors and find similarities in investors’ age, geographical location, or length the co-employment can affect trade synchronization under certain circumstances. Our findings have...

In this paper, we statistically analyze how investors distribute their trading intensity to different securities in stock markets. We find that trade allocation distributions are distinctive between investors. More importantly, the patterns of across surprisingly persistent for each investor, even when there is turnover portfolios. This suggests if have security-specific strategies, can be replaced by others while strategies remain. also time constraints do not appear limit investors'...

The starting point of this paper is that neighboring investors may talk to each other sharing information about their transactions in stock markets, leading similar trading behavior. We find pairwise trade timing similarities between investor pairs are negatively associated geographical distance corresponding pairs. This suggests local transfer channels individual used decision making. also observe differences age and language moderate association. analysis conducted using level data from...

In this paper, we ask whether the structure of investor networks, estimated using shareholder registration data, is abnormal during a financial crises. We answer question by analyzing networks through several most prominent global network features. The are from data on marketplace transactions all publicly traded securities executed in Helsinki Stock Exchange Finnish stock shareholders between 1995 and 2016. observe that feature distributions were 2008–2009 crisis, with statistical...

Within network analysis, the analytical maximum entropy framework has been very successful for different tasks as reconstruction and filtering. In a recent paper, same was used link-prediction monopartite networks: link probabilities all unobserved links in graph are provided most probable selected. Here we propose extension of such an approach to bipartite graphs. We test our method on two real world networks with topological characteristics. Our performances compared state-of-the-art...

We introduce the formation of a network money flows between assets in stock markets, which captures directed relations terms how investors have re-allocated exchange. Our approach is based on identifying link, or flow, that occurs when an investor funds purchase asset by selling another asset(s). extract investor-level flow networks daily basis from shareholder registration data, are then aggregated for both financial institutional and retail investors. Overall, we time series 877 2006 to...

Many of the real-world data sets can be portrayed as bipartite networks. Since connections between nodes same type are lacking, they need to inferred. The standard way do this is by converting networks their monopartite projection. However, simple approach renders an incomplete representation all information in original network. To end, we propose a new statistical method identify most critical links network Our takes into account heterogeneity node connections. Moreover, it handle...

The complex networks approach has been gaining popularity in analysing investor behaviour and stock markets, but within this approach, initial public offerings (IPO) have barely explored. We fill gap the literature by clusters first two years after IPO filing Helsinki Stock Exchange using a statistically validated network method to infer links based on co-occurrences of investors' trade timing for 69 stocks. Our findings show that rather large part similar structures form different...

We introduce a novel method to identify information networks in stock markets, which explicitly accounts for the impact of public on investor trading decisions. show that has clear effect empirical networks' topology. Most importantly, our strengthens identified relationship between investors' network centrality and returns. Furthermore, when less significant links are removed, association returns becomes statistically economically stronger. Findings suggest approach leads more precise...

We find that investors' future trading decisions are driven by the patterns of their social neighborhood and activity therein. Moreover, we provide evidence investors weigh connections differently in terms information transfer. Methodologically, tackle complex, cyclical investor networks graph neural networks, which allow us to propose a sophisticated way predict behavior with data on connections. Our analysis is based unique observed links through director (insider) positions same companies...

Investors in stock markets face a huge amount of financial information. For that reason, they must decide how to distribute their trading effort across different securities. We propose new measure investor trade allocation between securities, called the signature . This measure, which also serves as proxy for investor’s attention trade, allows us statistically analyze whether there are heterogeneous, persistent, and characteristic patterns way investors Building on large shareholder...

We introduce the formation a network of money flows between assets in stock markets, which captures directed relations terms how investors have re-allocated exchange. Our approach is based on identifying link, or flow, that occurs when an investor funds purchase asset by selling another asset(s). extract investor-level flow networks daily bases from shareholder registration data, are then aggregated for both financial institutional and retail investors. Overall, we time series 877 2006 to...