Network Analysis for Economics and Finance: An Application to Firm Ownership

In this chapter, we introduce network analysis as an approach to model data in economics and finance. First, we review the most recent empirical applications using network analysis in economics and finance. Second, we introduce the main network metrics that are useful to describe the overall network structure and characterize the position of a specific node in the network. Third, we model information on firm ownership as a network: firms are the nodes while ownership relationships are the linkages. Data are retrieved from Orbis including information of millions of firms and their shareholders at worldwide level. We describe the necessary steps to construct the highly complex international ownership network. We then analyze its structure and compute the main metrics. We find that it forms a giant component with a significant number of nodes connected to each other. Network statistics show that a limited number of shareholders control many firms, revealing a significant concentration of power. Finally, we show how these measures computed at different levels of granularity (i.e., sector of activity) can provide useful policy insights.

Classifying Counterparty Sector in EMIR Data

The data collected under the European Market Infrastructure Regulation (“EMIR data”) provide authorities with voluminous transaction-by-transaction details on derivatives but their use poses numerous challenges. To overcome one major challenge, this chapter draws from eight different data sources and develops a greedy algorithm to obtain a new counterparty sector classification. We classify counterparties’ sector for 96% of the notional value of outstanding contracts in the euro area derivatives market. Our classification is also detailed, comprehensive, and well suited for the analysis of the derivatives market, which we illustrate in four case studies. Overall, we show that our algorithm can become a key building block for a wide range of research- and policy-oriented studies with EMIR data.

Semi-supervised Text Mining for Monitoring the News About the ESG Performance of Companies

We present a general monitoring methodology to summarize news about predefined entities and topics into tractable time-varying indices. The approach embeds text mining techniques to transform news data into numerical data, which entails the querying and selection of relevant news articles and the construction of frequency- and sentiment-based indicators. Word embeddings are used to achieve maximally informative news selection and scoring. We apply the methodology from the viewpoint of a sustainable asset manager wanting to actively follow news covering environmental, social, and governance (ESG) aspects. In an empirical analysis, using a Dutch-written news corpus, we create news-based ESG signals for a large list of companies and compare these to scores from an external data provider. We find preliminary evidence of abnormal news dynamics leading up to downward score adjustments and of efficient portfolio screening.

Do the Hype of the Benefits from Using New Data Science Tools Extend to Forecasting Extremely Volatile Assets?

This chapter first provides an illustration of the benefits of using machine learning for forecasting relative to traditional econometric strategies. We consider the short-term volatility of the Bitcoin market by realized volatility observations. Our analysis highlights the importance of accounting for nonlinearities to explain the gains of machine learning algorithms and examines the robustness of our findings to the selection of hyperparameters. This provides an illustration of how different machine learning estimators improve the development of forecast models by relaxing the functional form assumptions that are made explicit when writing up an econometric model. Our second contribution is to illustrate how deep learning can be used to measure market-level sentiment from a 10% random sample of Twitter users. This sentiment variable significantly improves forecast accuracy for every econometric estimator and machine algorithm considered in our forecasting application. This provides an illustration of the benefits of new tools from the natural language processing literature at creating variables that can improve the accuracy of forecasting models.

Quantifying News Narratives to Predict Movements in Market Risk

The theory of Narrative Economics suggests that narratives present in media influence market participants and drive economic events. In this chapter, we investigate how financial news narratives relate to movements in the CBOE Volatility Index. To this end, we first introduce an uncharted dataset where news articles are described by a set of financial keywords. We then perform topic modeling to extract news themes, comparing the canonical latent Dirichlet analysis to a technique combining doc2vec and Gaussian mixture models. Finally, using the state-of-the-art XGBoost (Extreme Gradient Boosted Trees) machine learning algorithm, we show that the obtained news features outperform a simple baseline when predicting CBOE Volatility Index movements on different time horizons.

Massive Data Analytics for Macroeconomic Nowcasting

Nowcasting macroeconomic aggregates have proved extremely useful for policy-makers or financial investors, in order to get real-time, reliable information to monitor a given economy or sector. Recently, we have witnessed the arrival of new large databases of alternative data, stemming from the Internet, social media, satellites, fixed sensors, or texts. By correctly accounting for those data, especially by using appropriate statistical and econometric approaches, the empirical literature has shown evidence of some gain in nowcasting ability. In this chapter, we propose to review recent advances of the literature on the topic, and we put forward innovative alternative indicators to monitor the Chinese and US economies.

Extraction and Representation of Financial Entities from Text

In our modern society, almost all events, processes, and decisions in a corporation are documented by internal written communication, legal filings, or business and financial news. The valuable knowledge in such collections is not directly accessible by computers as they mostly consist of unstructured text. This chapter provides an overview of corpora commonly used in research and highlights related work and state-of-the-art approaches to extract and represent financial entities and relations.The second part of this chapter considers applications based on knowledge graphs of automatically extracted facts. Traditional information retrieval systems typically require the user to have prior knowledge of the data. Suitable visualization techniques can overcome this requirement and enable users to explore large sets of documents. Furthermore, data mining techniques can be used to enrich or filter knowledge graphs. This information can augment source documents and guide exploration processes. Systems for document exploration are tailored to specific tasks, such as investigative work in audits or legal discovery, monitoring compliance, or providing information in a retrieval system to support decisions.

Data Science Technologies in Economics and Finance: A Gentle Walk-In

This chapter is an introduction to the use of data science technologies in the fields of economics and finance. The recent explosion in computation and information technology in the past decade has made available vast amounts of data in various domains, which has been referred to as Big Data. In economics and finance, in particular, tapping into these data brings research and business closer together, as data generated in ordinary economic activity can be used towards effective and personalized models. In this context, the recent use of data science technologies for economics and finance provides mutual benefits to both scientists and professionals, improving forecasting and nowcasting for several kinds of applications. This chapter introduces the subject through underlying technical challenges such as data handling and protection, modeling, integration, and interpretation. It also outlines some of the common issues in economic modeling with data science technologies and surveys the relevant big data management and analytics solutions, motivating the use of data science methods in economics and finance.

Opening the Black Box: Machine Learning Interpretability and Inference Tools with an Application to Economic Forecasting

We present a comprehensive comparative case study for the use of machine learning models for macroeconomics forecasting. We find that machine learning models mostly outperform conventional econometric approaches in forecasting changes in US unemployment on a 1-year horizon. To address the black box critique of machine learning models, we apply and compare two variables attribution methods: permutation importance and Shapley values. While the aggregate information derived from both approaches is broadly in line, Shapley values offer several advantages, such as the discovery of unknown functional forms in the data generating process and the ability to perform statistical inference. The latter is achieved by the Shapley regression framework, which allows for the evaluation and communication of machine learning models akin to that of linear models.

Sharpening the Accuracy of Credit Scoring Models with Machine Learning Algorithms

The big data revolution and recent advancements in computing power have increased the interest in credit scoring techniques based on artificial intelligence. This has found easy leverage in the fact that the accuracy of credit scoring models has a crucial impact on the profitability of lending institutions. In this chapter, we survey the most popular supervised credit scoring classification methods (and their combinations through ensemble methods) in an attempt to identify a superior classification technique in the light of the applied literature. There are at least three key insights that emerge from surveying the literature. First, as far as individual classifiers are concerned, linear classification methods often display a performance that is at least as good as that of machine learning methods. Second, ensemble methods tend to outperform individual classifiers. However, a dominant ensemble method cannot be easily identified in the empirical literature. Third, despite the possibility that machine learning techniques could fail to outperform linear classification methods when standard accuracy measures are considered, in the end they lead to significant cost savings compared to the financial implications of using different scoring models.