Research articles for the 2020-10-05

A Mean Field Game Approach to Equilibrium Pricing with Market Clearing Condition
Masaaki Fujii,Akihiko Takahashi

In this work, we study an equilibrium-based continuous asset pricing problem which seeks to form a price process endogenously by requiring it to balance the flow of sales-and-purchase orders in the exchange market, where a large number of agents are interacting through the market price. Adopting a mean field game (MFG) approach, we find a special form of forward-backward stochastic differential equations of McKean-Vlasov type with common noise whose solution provides a good approximate of the market price. We show the convergence of the net order flow to zero in the large N-limit and get the order of convergence in N under some conditions. We also extend the model to a setup with multiple populations where the agents within each population share the same cost and coefficient functions but they can be different population by population.

A Note on Quadratic Funding under Constrained Matching Funds
Ricardo A. Pasquini

In this note I show that quadratic funding achieves decentralized social efficiency in the extent there are enough (donor) matching funds to cover the quadratic funding objective. If individual backers internalize that matching funds will not be sufficient to reach the quadratic level, allocation will be biased towards the capitalist allocation, the more so, the less matching funds are available. This result emerges even when individual contributors are not required to finance the deficit (i.e., the difference between total contributions and available matching funds). I also show properties of the level of required matching fund, in order to better understand under which conditions social efficiency will most likely be compromised.

A Regime Switching Skew-Normal Model of Contagion in Major African Stock Markets
Abubakar, Jamaladeen
This study was carried out to examine the contagion and structural break for some selected African stock markets namely: Nigeria, Ghana, South Africa (SA), Tunisia, and the US stock market. To achieve this, two periods were considered, that is the crisis period, 31st December, 2013 to 31st December, 2017 and the non-crisis period, 1st January, 2018 to 31st December, 2019. The choice of these periods was to reflect the structural break caused by economic recession in Nigeria during the crisis period. The expectation is that there is capital flight from Nigerian equities to safety in the selected emerging economies in Africa and in the US economy. Following the work of Chan et al (2018), the study used Regime Switching Skew-Normal (RSSN, henceforth) model which is capable of measuring contagion and structural breaks between markets and across crisis and non-crisis periods. The RSSN model was estimated using Bayesian method. The study finds the existence of moderate contagion between Nigeria and SA and the absence of contagion with the rest of the economies, suggesting there is capital flight from equities in Nigeria to SA for safety during the 2016 economic recession. Also, the study confirms the presence of structural break between crisis and non-crisis periods as the probability of average return declines during crisis period. The absence of contagion among African stock markets suggests there is no significant economic cooperation and cross-border portfolio investment flow among the countries. This development further underpins the imperative of the full implementation of African Continental Free Trade Area (AfCFTA), which is to encourage economic activities and investment flow on the continent.

Alpha Discovery Neural Network based on Prior Knowledge
Jie Fang,Shutao Xia,Jianwu Lin,Zhikang Xia,Xiang Liu,Yong Jiang

Genetic programming (GP) is the state-of-the-art in financial automated feature construction task. It employs reverse polish expression to represent features and then conducts the evolution process. However, with the development of deep learning, more powerful feature extraction tools are available. This paper proposes Alpha Discovery Neural Network (ADNN), a tailored neural network structure which can automatically construct diversified financial technical indicators based on prior knowledge. We mainly made three contributions. First, we use domain knowledge in quantitative trading to design the sampling rules and object function. Second, pre-training and model pruning has been used to replace genetic programming, because it can conduct more efficient evolution process. Third, the feature extractors in ADNN can be replaced by different feature extractors and produce different functions. The experiment results show that ADNN can construct more informative and diversified features than GP, which can effectively enriches the current factor pool. The fully-connected network and recurrent network are better at extracting information from the financial time series than the convolution neural network. In real practice, features constructed by ADNN can always improve multi-factor strategies' revenue, sharpe ratio, and max draw-down, compared with the investment strategies without these factors.

Asset Price Volatility and Investment Horizons: An Experimental Investigation
Anufriev, Mikhail,Chernulich, Aleksei,Tuinstra, Jan
We study the effects of the investment horizon on asset price volatility using a Learning to Forecast experiment. We end that, for short investment horizons, participants coordinate on self-fulfilling trend extrapolating predictions. Price deviations are then reinforced and amplified, possibly leading to large bubbles and crashes in asset prices. For longer investment horizons such bubbles do not emerge and price volatility tends to be lower. This is due to the fact that, for longer horizons, there is more dispersion in participants' forecasts, and participants extrapolate trends in past prices to a lesser extent. We also show that, independent of the investment horizon, if the initial history of asset prices is already relatively stable before participants start their prediction task, price volatility remains small, with prices close to their fundamental values for the duration of the experiment.

Automatic Financial Feature Construction
Jie Fang,Shutao Xia,Jianwu Lin,Yong Jiang

In automatic financial feature construction task, the state-of-the-art technic leverages reverse polish expression to represent the features, then use genetic programming (GP) to conduct its evolution process. In this paper, we propose a new framework based on neural network, alpha discovery neural network (ADNN). In this work, we made several contributions. Firstly, in this task, we make full use of neural network overwhelming advantage in feature extraction to construct highly informative features. Secondly, we use domain knowledge to design the object function, batch size, and sampling rules. Thirdly, we use pre-training to replace the GP evolution process. According to neural network universal approximation theorem, pre-training can conduct a more effective and explainable evolution process. Experiment shows that ADNN can remarkably produce more diversified and higher informative features than GP. Besides, ADNN can serve as a data augmentation algorithm. It further improves the the performance of financial features constructed by GP.

Bitcoin and its impact on the economy
Merrick Wang

The purpose of this paper is to review the concept of cryptocurrencies in our economy. First, Bitcoin and alternative cryptocurrencies' histories are analyzed. We then study the implementation of Bitcoin in the airline and real estate industries. Our study finds that many Bitcoin companies partner with airlines in order to decrease processing times, to provide ease of access for spending in international airports, and to reduce fees on foreign exchanges for fuel expenses, maintenance, and flight operations. Bitcoin transactions have occurred in the real estate industry, but many businesses are concerned with Bitcoin's potential interference with the U.S. government and its high volatility. As Bitcoin's price has been growing rapidly, we assessed Bitcoin's real value; Bitcoin derives value from its scarcity, utility, and public trust. In the conclusion, we discuss Bitcoin's future and conclude that Bitcoin may change from a short-term profit investment to a more steady industry as we identify Bitcoin with the "greater fool theory", and as the number of available Bitcoins to be mined dwindles and technology becomes more expensive.

Borrowing Constraints, Own Labour and Homeownership: Does it Pay to Paint Your Walls?
Lindner, Peter,Mathä, Thomas Y.,Pulina, Giuseppe,Ziegelmeyer, Michael
Using a dedicated set of questions in the 2014 Luxembourg Household Finance and Consumption Survey (LU-HFCS), we show that a substantial share of households contributes their own labour to the acquisition of their main residence. These contributions help households faced with credit constraints, since they reduce the need for external financing. We develop a simple theoretical model and show that own labour contributions decrease with the level of financial resources available, while they increase with the mortgage interest rate. These theoretical results are supported by empirical analysis, which also shows that own labour contributions vary by household characteristics (age, gender, profession) and by type of dwelling (house, apartment).

Complexity and Risk-Taking
Pély, Désirée-Jessica
Literature examines significant but differing effects of prior outcomes on risky choice. In this study, the seemingly different behavior is reconciled with the introduction of a complexity effect. Subjects behave more in line with prospect theory predictions when decisions are less complex as it is easier to categorize their gain and loss preferences. Complexity is increased by providing ambiguous information in repeated experimental decisions. The complexity effect is robust to subjects' individual characteristics, self-selection biases, sampling errors, and to effects stemming from the house money. Further, the complexity effect diminishes when subjects exhibit high cognitive skills or an analytical-rational thinking-style.

Corporate Environmental Policy and Product Market Competition
Grinstein, Yaniv,Larkin, Yelena
Does product market competition affect corporate environmental policy? It is commonly believed that firms in competitive environments have stronger incentives to cut costs, which could lead them to neglect negative externalities. However, we find that cost cutting incentives could in fact be environmentally friendly. To arrive at this conclusion, we use a quasi-natural experiment of the restructuring of the utility industry in the US, which has opened the market to competition. We find that the restructuring has incentivized utilities to move to cheaper, but also less polluting, fossil fuels. Moreover, competition forces have smoothed out inefficient peak-capacity operation across competing plants, also contributing to reduction in pollution.

Deep Learning algorithms for solving high dimensional nonlinear Backward Stochastic Differential Equations
Lorenc Kapllani,Long Teng

We study deep learning-based schemes for solving high dimensional nonlinear backward stochastic differential equations (BSDEs). First we show how to improve the performances of the proposed scheme in [W. E and J. Han and A. Jentzen, Commun. Math. Stat., 5 (2017), pp.349-380] regarding computational time and stability of numerical convergence by using the advanced neural network architecture instead of the stacked deep neural networks. Furthermore, the proposed scheme in that work can be stuck in local minima, especially for a complex solution structure and longer terminal time. To solve this problem, we investigate to reformulate the problem by including local losses and exploit the Long Short Term Memory (LSTM) networks which are a type of recurrent neural networks (RNN). Finally, in order to study numerical convergence and thus illustrate the improved performances with the proposed methods, we provide numerical results for several 100-dimensional nonlinear BSDEs including a nonlinear pricing problem in finance.

Deep Learning for Digital Asset Limit Order Books
Rakshit Jha,Mattijs De Paepe,Samuel Holt,James West,Shaun Ng

This paper shows that temporal CNNs accurately predict bitcoin spot price movements from limit order book data. On a 2 second prediction time horizon we achieve 71\% walk-forward accuracy on the popular cryptocurrency exchange coinbase. Our model can be trained in less than a day on commodity GPUs which could be installed into colocation centers allowing for model sync with existing faster orderbook prediction models. We provide source code and data at

DoubleEnsemble: A New Ensemble Method Based on Sample Reweighting and Feature Selection for Financial Data Analysis
Chuheng Zhang,Yuanqi Li,Xi Chen,Yifei Jin,Pingzhong Tang,Jian Li

Modern machine learning models (such as deep neural networks and boosting decision tree models) have become increasingly popular in financial market prediction, due to their superior capacity to extract complex non-linear patterns. However, since financial datasets have very low signal-to-noise ratio and are non-stationary, complex models are often very prone to overfitting and suffer from instability issues. Moreover, as various machine learning and data mining tools become more widely used in quantitative trading, many trading firms have been producing an increasing number of features (aka factors). Therefore, how to automatically select effective features becomes an imminent problem. To address these issues, we propose DoubleEnsemble, an ensemble framework leveraging learning trajectory based sample reweighting and shuffling based feature selection. Specifically, we identify the key samples based on the training dynamics on each sample and elicit key features based on the ablation impact of each feature via shuffling. Our model is applicable to a wide range of base models, capable of extracting complex patterns, while mitigating the overfitting and instability issues for financial market prediction. We conduct extensive experiments, including price prediction for cryptocurrencies and stock trading, using both DNN and gradient boosting decision tree as base models. Our experiment results demonstrate that DoubleEnsemble achieves a superior performance compared with several baseline methods.

Evaluation of company investment value based on machine learning
Junfeng Hu,Xiaosa Li,Yuru Xu,Shaowu Wu,Bin Zheng

In this paper, company investment value evaluation models are established based on comprehensive company information. After data mining and extracting a set of 436 feature parameters, an optimal subset of features is obtained by dimension reduction through tree-based feature selection, followed by the 5-fold cross-validation using XGBoost and LightGBM models. The results show that the Root-Mean-Square Error (RMSE) reached 3.098 and 3.059, respectively. In order to further improve the stability and generalization capability, Bayesian Ridge Regression has been used to train a stacking model based on the XGBoost and LightGBM models. The corresponding RMSE is up to 3.047. Finally, the importance of different features to the LightGBM model is analysed.

Explicit option valuation in the exponential NIG model
Jean-Philippe Aguilar

We provide closed-form pricing formulas for a wide variety of path-independent options, in the exponential L\'evy model driven by the Normal inverse Gaussian process. The results are obtained in both the symmetric and asymmetric model, and take the form of simple and quickly convergent series, under some condition involving the log-forward moneyness and the maturity of instruments. Proofs are based on a factorized representation in the Mellin space for the price of an arbitrary path-independent payoff, and on tools from complex analysis. The validity of the results is assessed thanks to several comparisons with standard numerical methods (Fourier and Fast Fourier transforms, Monte-Carlo simulations) for realistic sets of parameters. Precise bounds for the convergence speed and the truncation error are also provided.

Financial Distress and the Role of Management in Micro and Small-Sized Firms
Alexandre, Fernando,Cruz, Sara,Portela, Miguel
In this paper, we focus on managerial characteristics of micro and small-sized firms. Using linked employer-employee data on the Portuguese economy for the 2010-2018 period, we estimate the impact of management teams' human capital on the probability of firms becoming financially distressed and on their subsequent recovery. Our estimates show that the relevance of management teams' formal education on the probability of firms becoming financially distressed depends on firms' size and the type of education. We show that management teams' formal education and tenure reduces the probability of micro and small-sized firms becoming financially distressed and increases the probability of their subsequent recovery.The estimates also suggest that those impacts are stronger for micro and small-sized firms. Additionally, our results show that functional experience previously acquired in other firms, namely in foreign-owned and in exporting firms and in the area of finance, may reduce the probability of micro firms becoming financially distressed. On the other hand, previous functional experience in other firms seems to have a strong and highly significant impact on increasing the odds of recovery of financially distressed firms. We conclude that policies that induce an improvement in the managerial human capital of micro and small-sized firms have significant scope to improve their financial condition, reducing the likelihood of firms entering a state of financial distress.

Financial Innovations & Sustainable Development: A Paper For The Business & Sustainable Development Commission
Mainelli, Michael,Mills, Simon
This report was commissioned from Z/Yen by the Business and Sustainable Development Commission (BSDC) as part of a research programme designed to examine how the financial services sector could support the Sustainable Development Goals (SDGs).The paper argues that financial systems are tools that can aid achieving the Sustainable Development Goals, yet also need to exhibit sustainability of their own. Some specific innovations that might provide disproportionately large benefits for sustainability include:Financial Measurement and Products:Confidence Accounting: Representing financial information in its true form as ranges would help highlight the value of reducing uncertainty and the importance of SDGs over the long-term;Mutual Distributed Ledgers (AKA Blockchains): Provide a solid underpinning to the registries society uses in all areas and enable communities to work together without creating ‘natural monopolies’ over information; monetary systems:Digital Fiat Currencies: Governments using digital currencies may gain much greater control over their monetary systems, reduce corruption, and have more effective tax systems;Common Tenders: Community monies of many forms could be a major boost to specific community values and encourage credit creation among trust groups; financial technology:Identity: Systems in use for financial services could transform much wider areas of government, health, and qualifications for approximately 2.4 billion people, especially in providing access to other property rights;Peer-To-Peer Lending & Insurance: These tools are engaging communities and providing better assessments of risk and reward from people ‘on the ground’;Mobile Money: This innovation is bringing access to banking and finance services to millions of the world’s poorest people, who until recently did not have access to these services.Financial Structures and Systems:Policy Performance Bonds: These mechanisms have great potential in aligning financial incentives with wider SDGs;land value taxation â€" in combination with extensive tax reform, i.e. not just another tax, could simplify tax systems and encourage sustainable development.Disaster Reinsurance: A combination of reinsurance and catastrophe bonds could help to ensure longer-term investment and returns-on-investment in risky areas.While, perhaps with the exception of digital fiat currencies, these innovations are already here, they are often ignored or unevenly implemented. This paper concludes that, of all the innovations discussed, providing identity systems for the 2.4 billion people without legal identity is a single tool that could transform all areas of sustainable development.

From small markets to big markets
Laurence Carassus,Miklos Rasonyi

We study the most famous example of a large financial market: the Arbitrage Pricing Model, where investors can trade in a one-period setting with countably many assets admitting a factor structure. We consider the problem of maximising expected utility in this setting. Besides establishing the existence of optimizers under weaker assumptions than previous papers, we go on studying the relationship between optimal investments in finite market segments and those in the whole market. We show that certain natural (but nontrivial) continuity rules hold: maximal satisfaction, reservation prices and (convex combinations of) optimizers computed in small markets converge to their respective counterparts in the big market.

Gold Standard Pairs Trading Rules: Are They Valid?
Miroslav Fil

Pairs trading is a strategy based on exploiting mean reversion in prices of securities. It has been shown to generate significant excess returns, but its profitability has dropped significantly in recent periods. We employ the most common distance and cointegration methods on US equities from 1990 to 2020 including the Covid-19 crisis. The strategy overall fails to outperform the market benchmark even with hyperparameter tuning, but it performs very strongly during bear markets. Furthermore, we demonstrate that market factors have a strong relationship with the optimal parametrization for the strategy, and adjustments are appropriate for modern market conditions.

Group cohesion under individual regulatory constraints
Delia Coculescu,Freddy Delbaen

We consider a group consisting of N business units. We suppose there are regulatory constraints for each unit, more precisely, the net worth of each business unit is required to belong to a set of acceptable risks, assumed to be a convex cone. Because of these requirements, there are less incentives to operate under a group structure, as creating one single business unit, or altering the liability repartition among units, may allow to reduce the required capital. We analyse the possibilities for the group to benefit from a diversification effect and economise on the cost of capital. We define and study the risk measures that allow for any group to achieve the minimal capital, as if it were a single unit, without altering the liability of business units, and despite the individual admissibility constraints. We call these risk measures cohesive risk measures.

Information thermodynamics of financial markets: the Glosten-Milgrom model
Léo Touzo,Matteo Marsili,Don Zagier

The Glosten-Milgrom model describes a single asset market, where informed traders interact with a market maker, in the presence of noise traders. We derive an analogy between this financial model and a Szil\'ard information engine by {\em i)} showing that the optimal work extraction protocol in the latter coincides with the pricing strategy of the market maker in the former and {\em ii)} defining a market analogue of the physical temperature from the analysis of the distribution of market orders. Then we show that the expected gain of informed traders is bounded above by the product of this market temperature with the amount of information that informed traders have, in exact analogy with the corresponding formula for the maximal expected amount of work that can be extracted from a cycle of the information engine. In this way, recent ideas from information thermodynamics may shed light on financial markets, and lead to generalised inequalities, in the spirit of the extended second law of thermodynamics.

Intermediated Credit and Local Resilience
Jiang, Erica Xuewei,Liu, Will Shuo,Seltzer, Lee
This paper demonstrates the importance of bank capital in improving local resilience and the complementarity of bank capital and government aid programs. We show that following the COVID-19 pandemic and shutdown, areas with more jobs supported by subsidized bank loans during normal times had more job losses and business closures, and more so if the local banking sector is less capitalized. Such losses were heavily borne by low-income workers. We also find that areas with a less capitalized banking sector received disproportionately less Paycheck Protection Program funding. Using a dynamic model of firm entry and exit with bank borrowing, we formulate the mechanism of how bank capital can mitigate the impact of adverse aggregate shocks on employment and firm exit. We calibrate the model to quantify effects of bank capital on resilience and the amount of government funding necessary for full resilience in various simulated scenarios of adverse shocks and bank capitalization.

Lbl - Lstm : Log Bilinear And Long Short Term Memory Based Efficient Stock Forecasting Model Considering External Fluctuating Factor
Gurav, Uma,Kotrappa, Dr. S.
Stock market prediction problem is considered to be NP-hard problem because of highly volatile nature of stock market. In this paper, effort has been made to design efficient stock forecasting model using log Bilinear and long short term memory (LBL-LSTM) considering external fluctuating factor such as varying Bank's lending rates. The external factor bank's lending rates affects stock market performance ,as it plays vital role for the purchase of stocks in case of financial crisis faced by various business enterprises. Proposed LBL-LSTM based model shows performance improvement over existing machine learning algorithms used for stock market prediction.

Learning Time Varying Risk Preferences from Investment Portfolios using Inverse Optimization with Applications on Mutual Funds
Shi Yu,Yuxin Chen,Chaosheng Dong

The fundamental principle in Modern Portfolio Theory (MPT) is based on the quantification of the portfolio's risk related to performance. Although MPT has made huge impacts on the investment world and prompted the success and prevalence of passive investing, it still has shortcomings in real-world applications. One of the main challenges is that the level of risk an investor can endure, known as \emph{risk-preference}, is a subjective choice that is tightly related to psychology and behavioral science in decision making. This paper presents a novel approach of measuring risk preference from existing portfolios using inverse optimization on the mean-variance portfolio allocation framework. Our approach allows the learner to continuously estimate real-time risk preferences using concurrent observed portfolios and market price data. We demonstrate our methods on real market data that consists of 20 years of asset pricing and 10 years of mutual fund portfolio holdings. Moreover, the quantified risk preference parameters are validated with two well-known risk measurements currently applied in the field. The proposed methods could lead to practical and fruitful innovations in automated/personalized portfolio management, such as Robo-advising, to augment financial advisors' decision intelligence in a long-term investment horizon.

M&A Motives and Home Country Institutions: Evidence from Asian-Pacific Acquisitions in Europe
Larsen, Matthew,Zamborsky, Peter
With cross-border acquisitions on the rise, especially by multinationals from the Asia-Pacific region, it is important to better understand their motives. Using the Marketline Advantage data on over 700 cross-border acquisitions of European firms by Asian-Pacific multinationals in 2007-2017, we analyse 11 types of rationales for these deals and how they relate to income levels and institutional development of the acquirers’ home countries. We find that the home country development mostly does not affect propensity to invest with a particular motive. However, the quality of the home country’s regulatory environment is significantly and positively related to M&A motives. We conclude with implications for the research on emerging market multinationals, international M&A strategy and outward FDI policies of emerging economies.

Market laws
Caglar Tuncay

More than one billion data sampled with different frequencies from several financial instruments were investigated with the aim of testing whether they involve power law. As a result, a known power law with the power exponent around -4 was detected in the empirical distributions of the relative returns. Moreover, a number of new power law behaviors with various power exponents were explored in the same data. Further on, a model based on finite sums over numerous Maxwell-Boltzmann type distribution functions with random (pseudorandom) multipliers in the exponent were proposed to deal with the empirical distributions involving power laws. The results indicate that the proposed model may be universal.

NetVIX - A Network Volatility Index of Financial Markets
Ahelegbey, Daniel Felix,Giudici, Paolo
We construct a network volatility index (NetVIX) via market interconnectedness and volatilities to measure global market turbulence. The NetVIX multiplicatively decomposes into an average volatility and a network amplifier index. It also additively decomposes into marginal volatility indices for measuring individual contribution to global turmoil. We apply our measure to study the relationship between the interconnectedness among 20 major stock markets and global market risks over the last two decades. The NetVIX is shown to be a novel approach to measuring global market risk, and an alternative to the VIX. The result shows that during crisis periods, particularly the tech-bubble, sub-prime, and COVID-19 pandemic, the interconnectedness of the markets amplifies average market risk more than 700 percent to cause a global meltdown. We find evidence that the highest risk-contributing markets to global meltdown are the US, Brazil, Hong Kong, France, and Germany.

Neural Jump Ordinary Differential Equation
Calypso Herrera,Florian Krach,Josef Teichmann

Combinations of neural ODEs with recurrent neural networks (RNN), like GRUODE-Bayes or ODE-RNN are well suited to model irregularly-sampled time series. While those models outperform existing discrete-time approaches, no theoretical guarantees for their predictive capabilities are available. Assuming that the irregularly-sampled time series data originates from a continuous stochastic processes, the optimal on-line prediction is the conditional expectation given the currently available information. We introduce the Neural Jump ODE (NJ-ODE) that provides a data-driven approach to learn, continuously in time, the conditional expectation of a stochastic process. Our approach models the conditional expectation between two observations with a neural ODE and jumps whenever a new observation is made. We define a novel training framework, which allows us to prove theoretical convergence guarantees for the first time. In particular, we demonstrate the predictive capabilities of our model by proving that, under some regularity assumptions, the output process converges to the conditional expectation process. We provide experiments showing that the theoretical results also hold empirically. Moreover, we experimentally show that our model outperforms one state of the art model in more complex learning tasks and give comparisons on a real-world dataset.

Neural Network-based Automatic Factor Construction
Jie Fang,Jianwu Lin,Shutao Xia,Yong Jiang,Zhikang Xia,Xiang Liu

Instead of conducting manual factor construction based on traditional and behavioural finance analysis, academic researchers and quantitative investment managers have leveraged Genetic Programming (GP) as an automatic feature construction tool in recent years, which builds reverse polish mathematical expressions from trading data into new factors. However, with the development of deep learning, more powerful feature extraction tools are available. This paper proposes Neural Network-based Automatic Factor Construction (NNAFC), a tailored neural network framework that can automatically construct diversified financial factors based on financial domain knowledge and a variety of neural network structures. The experiment results show that NNAFC can construct more informative and diversified factors than GP, to effectively enrich the current factor pool. For the current market, both fully connected and recurrent neural network structures are better at extracting information from financial time series than convolution neural network structures. Moreover, new factors constructed by NNAFC can always improve the return, Sharpe ratio, and the max draw-down of a multi-factor quantitative investment strategy due to their introducing more information and diversification to the existing factor pool.

New Finance: In Search for Analytical Framework
Monkiewicz, Jan
Since the beginning of the century, modern financial systems have been experiencing a period of dynamic change. These changes have caused the uniform fabric on which the systems were based for a long time to become eroded and a multitude of alternative solutions to appear in their place. The existing systems are becoming increasingly heterogeneous and increasingly less transparent. In addition to the traditional financial system based on highly regulated financial intermediaries with a legal monopoly, an alternative sector is emerging. It is based on an often different market infrastructure that allows the direct allocation and acquisition of funds on a P2P or B2B basis as well as on other rules and intermediary entities. A special role is played in this process of change by current technological innovations, especially the ubiquitous digitization of financial systems. This state of affairs creates conceptual and terminological confusion,which hinders the process of academic communication and complicates the comparison of empirical research results and theoretical reflections. The purpose of this study is to systematize various concepts of new finance and analyse its main components in a wider systemic context.In the analysis which follows special attention is given to the changing architecture of digitalis-ed financial systems,changing nature of financial services providers and new challenges facing public supervisory systems.The basis of the analysis is a critical review of the literature on the subject, both domestic and foreign.

Optimal Order Execution in Intraday Markets: Minimizing Costs in Trade Trajectories
Christopher Kath,Florian Ziel

Optimal execution, i.e., the determination of the most cost-effective way to trade volumes in continuous trading sessions, has been a topic of interest in the equity trading world for years. Electricity intraday trading slowly follows this trend but is far from being well-researched. The underlying problem is a very complex one. Energy traders, producers, and electricity wholesale companies receive various position updates from customer businesses, renewable energy production, or plant outages and need to trade these positions in intraday markets. They have a variety of options when it comes to position sizing or timing. Is it better to trade all amounts at once? Should they split orders into smaller pieces? Taking the German continuous hourly intraday market as an example, this paper derives an appropriate model for electricity trading. We present our results from an out-of-sample study and differentiate between simple benchmark models and our more refined optimization approach that takes into account order book depth, time to delivery, and different trading regimes like XBID (Cross-Border Intraday Project) trading. Our paper is highly relevant as it contributes further insight into the academic discussion of algorithmic execution in continuous intraday markets and serves as an orientation for practitioners. Our initial results suggest that optimal execution strategies have a considerable monetary impact.

Quantifying the impact of COVID-19 on the US stock market: An analysis from multi-source information
Asim Kumer Dey,Toufiqul Haq,Kumer Das,Irina Panovska

We develop a novel temporal complex network approach to quantify the US county level spread dynamics of COVID-19. The objective is to study the effects of the local spread dynamics, COVID-19 cases and death, and Google search activities on the US stock market. We use both conventional econometric and Machine Learning (ML) models. The results suggest that COVID-19 cases and deaths, its local spread, and Google searches have impacts on abnormal stock prices between January 2020 to May 2020. In addition, incorporating information about local spread significantly improves the performance of forecasting models of the abnormal stock prices at longer forecasting horizons. On the other hand, although a few COVID-19 related variables, e.g., US total deaths and US new cases exhibit causal relationships on price volatility, COVID-19 cases and deaths, local spread of COVID-19, and Google search activities do not have impacts on price volatility.

Replicating Maximum Yield Annuities with U.S. Treasury Funds
Prendergast, Joseph
Many individuals rely on annuity purchases to provide a steady stream of income during their post-retirement years. Life insurance companies provide a variety of annuity contracts but often come with sizable fees that are extremely detrimental to the purchaser in a low interest rate environment. This paper develops a methodology that enables retail investors to structure annuities using commonly available U.S. Treasury Exchange Traded Funds (ETFs) or mutual funds. Only historical price or NAV data and Treasury yield data are required to implement the strategy. The funds comprise a dynamically managed portfolio requiring only an initial investment. Each period, the portfolio is rebalanced to match the modified duration of a reference annuity which will minimize interest rate risk from parallel shifts in future yield curves. The weights are optimized to provide the highest yield possible. The objective is for the portfolio to provide the required periodic annuity payments and have a zero balance at annuity maturity under a variety of interest rate scenarios, while providing the maximum possible yield. Scenario tests show that the strategy is effective under parallel yield curve shifts but may have shortfalls for curve steepening and gains for curve flattening. Investors may choose to add to their initial investment to reduce the risk of shortfalls if the curve steepens. The paper concludes with an implementation of the strategy using actual Treasury ETFs.

Responsibility Without Power? The Governance Of Mutual Distributed Ledgers (aka Blockchains)
Mills, Simon,McDowall, Bob
Mutual Distributed Ledger (MDL, aka blockchain) technology is in an emergent phase. New applications are under development; new uses are being researched; new consortia are being formed to explore MDL applications. Considering appropriate governance structures has had a lower priority so far, but trust in the increasingly popular systems will depend on their incorporating good governance principles. It was the aim of this study to identify those principles, in order to provide a roadmap for developers and users alike.An analysis of the material collected through desktop research, as well as several discussions with practitioners and stakeholders including a conference and webinar on the subject, has revealed that effective governance in MDL systems relies on people rather than software and rests on three pillars:Architecture: The role of the governance structure, its composition, remit, powers, responsibilities, and its relationship with users, is a critical component.Accountability: Effective governance of MDLs enhances trust. Trust is enhanced when a governance structure is accountable to its stakeholders, transparent in its decision-making, and subject to periodic audit and third party review.Action: The governance structure must develop strategic and risk management plans, which are delivered through effective performance management frameworks. Trust can be further enhanced through the use of the voluntary standards market to independently verify performance metrics and the systems established to compile them.

Resurrecting the Value Premium
Blitz, David,Hanauer, Matthias X.
The prolonged poor performance of the value factor has led to doubts about whether the value premium still exists. Some have noted that the observed returns still fall within statistical confidence intervals, but such arguments do not restore full confidence in the value premium. This paper adds to the literature by showing that the academic value factor, HML, has not only suffered setbacks in recent years but has, in fact, been weak for decades already. However, we show that the value premium can be resurrected using insights that are well documented in the literature or common knowledge among practitioners. In particular, we include more powerful value metrics, apply some basic risk management, and make more effective use of the breadth of the liquid universe of stocks. Although our enhanced value strategy also suffers in recent years, it has a solid long term track record that does not warrant existential concerns. We conclude that a healthy value premium is still clearly present in the cross-section of stock returns.

Sequential hypothesis testing in machine learning driven crude oil jump detection
Michael Roberts,Indranil SenGupta

In this paper we present a sequential hypothesis test for the detection of general jump size distrubution. Infinitesimal generators for the corresponding log-likelihood ratios are presented and analyzed. Bounds for infinitesimal generators in terms of super-solutions and sub-solutions are computed. This is shown to be implementable in relation to various classification problems for a crude oil price data set. Machine and deep learning algorithms are implemented to extract a specific deterministic component from the crude oil data set, and the deterministic component is implemented to improve the Barndorff-Nielsen and Shephard model, a commonly used stochastic model for derivative and commodity market analysis.

Shifting Influences on Corporate Governance: Capital Market Completeness and Policy Channeling
Gilson, Ronald J.,Milhaupt, Curtis J.
Corporate governance scholarship is typically portrayed as driven by single factor models, for example, shareholder value maximization, stakeholder theory, or director primacy. These governance models are Copernican; one factor is or should be the center of the corporate governance solar system. In this essay, we argue that, as with binary stars, the shape of the governance system is at any time the result of the interaction of two central influences, which we refer to as capital market completeness and policy channeling. In contrast to single factor models, which reflect a stable normative statement of what should drive corporate governance, in our account the relation between these two governance influences is dynamic. Motivated by Albert Hirschman’s Shifting Involvements, we posit that all corporate governance systems undergo repeated shifts in the relative weights of the two influences on the system. Capital market completeness determines the corporate ownership structure and privileges shareholder governance and value maximization by increasing the capacity to slice risk, return, and control into different equity instruments. The capability to specify shareholder control rights makes the capital market more complete, tailoring the character of influence associated with holding particular equity securities and its reciprocal, the exposure of management to capital market oversight. Policy channeling, the real government’s instrumental use of the corporation for distributional or social ends, pushes the corporate governance gravitational center toward purposes other than maximizing shareholder value. We show that this pattern is not limited to a particular country, and illustrate our argument by tracing the cyclical reframing of Berle and Means’ thesis in the U.S., Japan’s sluggish shift from policy channeling in its postwar heyday toward capital market completeness under the Abenomics reforms, and the distinctive case of China, where capital market completeness has itself been used as a policy channeling instrument under the pervasive influence of the Chinese Communist Party, creating the world’s most stakeholder-oriented system of corporate governance.We close by examining the means through which the current shift toward policy channeling in U.S. and U.K. corporate governance is taking place â€" the “stewardship” movement and the debate over “corporate purpose.” We view both as a reaction to the reduced managerial discretion caused by the reconcentration of ownership in the hands of institutional investors, and analyze factors suggesting that this reform movement, like others before it, is likely destined to result in a disappointment-driven shift in the opposite direction.

Stock2Vec: A Hybrid Deep Learning Framework for Stock Market Prediction with Representation Learning and Temporal Convolutional Network
Xing Wang,Yijun Wang,Bin Weng,Aleksandr Vinel

We have proposed to develop a global hybrid deep learning framework to predict the daily prices in the stock market. With representation learning, we derived an embedding called Stock2Vec, which gives us insight for the relationship among different stocks, while the temporal convolutional layers are used for automatically capturing effective temporal patterns both within and across series. Evaluated on S&P 500, our hybrid framework integrates both advantages and achieves better performance on the stock price prediction task than several popular benchmarked models.

Strikingly Suspicious Overnight and Intraday Returns
Bruce Knuteson

The world's stock markets display a strikingly suspicious pattern of overnight and intraday returns. Overnight returns to major stock market indices over the past few decades have been wildly positive, while intraday returns have been disturbingly negative. The cause of these astonishingly consistent return patterns is unknown. We highlight the features of these extraordinary patterns that have hindered the construction of any plausible innocuous explanation. We then use those same features to deduce the only plausible explanation so far advanced for these strikingly suspicious returns.

The Fed in the Corporate Bond Market in 2020
McCauley, Robert N.
The Federal Reserve interventions in private securities markets in the spring of 2020 extended its 2008 playbook from buying high quality short-term paper to bonds, and departed from it by buying junk bonds. In March 2020, the Fed reprised its last-resort lending to primary dealers, accepting private securities as collateral, and its last-resort underwriting and buying of commercial paper. Given the reliance of non-financial firms on corporate bonds, some were not surprised when the Fed then extended last-resort underwriting and buying to corporate bonds. In April, however, the Fed departed from its playbook with its announcement that it would buy junk bond exchange-traded funds (ETFs): it set no minimum quality criterion for its credit extension. The Fed’s announced intervention in corporate bond markets succeeded before the buying even started. It raised prices of corporate bonds, narrowed both trading and fund valuation spreads, reversed investor runs and encouraged record-setting corporate bond issuance. ETF prices jumped on announcement, flipping a flashing market “billboard” from sell to buy, and underlying bond prices, spreads and flows subsequently improved across a broad range of dollar credit markets.This paper raises two policy questions. First, could the Fed have reduced the conflict between buying junk bonds and its previous efforts to reduce supervised banks’ involvement in leveraged loans? The Fed could have bought only junk bond funds holding a smaller weight of the lowest quality bonds issued by firms that private equity deals had leveraged up. Second, should the Congress authorize the Fed to do open market operations in corporate bonds? Such authority could avoid the legal awkwardness of using emergency lending powers to buy corporate bonds and could allow the Fed to develop operational capacity in this important market. Similar issues of role conflict and legal powers arise in any market, including emerging markets, when the central bank buys private securities.

The energy distance for ensemble and scenario reduction
Florian Ziel

Scenario reduction techniques are widely applied for solving sophisticated dynamic and stochastic programs, especially in energy and power systems, but also used in probabilistic forecasting, clustering and estimating generative adversarial networks (GANs). We propose a new method for ensemble and scenario reduction based on the energy distance which is a special case of the maximum mean discrepancy (MMD). We discuss the choice of energy distance in detail, especially in comparison to the popular Wasserstein distance which is dominating the scenario reduction literature. The energy distance is a metric between probability measures that allows for powerful tests for equality of arbitrary multivariate distributions or independence. Thanks to the latter, it is a suitable candidate for ensemble and scenario reduction problems. The theoretical properties and considered examples indicate clearly that the reduced scenario sets tend to exhibit better statistical properties for the energy distance than a corresponding reduction with respect to the Wasserstein distance. We show applications to a Bernoulli random walk and two real data based examples for electricity demand profiles and day-ahead electricity prices.

Use Cases of Quantum Optimization for Finance
Samuel Mugel,Enrique Lizaso,Roman Orus

In this paper we briefly review two recent use-cases of quantum optimization algorithms applied to hard problems in finance and economy. Specifically, we discuss the prediction of financial crashes as well as dynamic portfolio optimization. We comment on the different types of quantum strategies to carry on these optimizations, such as those based on quantum annealers, universal gate-based quantum processors, and quantum-inspired Tensor Networks.

Using Soccer Games as an Instrument to Forecast the Spread of COVID-19 in Europe
Gomez, Juan-Pedro,Mironov, Maxim
We provide strong empirical support for the contribution of soccer matches held in Europe during the first quarter of 2020 to the spread of the COVID-19 virus. We analyze more than 1,000 games across 194 regions from 10 European countries. Daily cases of COVID-19 grow significantly faster in regions where at least one soccer game took place two weeks earlier, consistent with the existence of an incubation period and the lack or massive testing in the early phases of the pandemic. Game attendance and venue capacity show an analogous pattern. These results weaken as we include stadiums with smaller capacity, consistent with a minimum agglomeration needed for the virus transmission to be detectible. We discuss the relevance of these variables as instruments for the identification of the causal effect of COVID-19 on firms, the economy, and financial markets.

Wealth and Poverty: The Effect of Poverty on Communities
Merrick Wang,Robert Johnston

This paper analyzes the differences in poverty in high wealth communities and low wealth communities. We first discuss methods of measuring poverty and analyze the causes of individual poverty and poverty in the Bay Area. Three cases are considered regarding relative poverty. The first two cases involve neighborhoods in the Bay Area while the third case evaluates two neighborhoods within the city of San Jose, CA. We find that low wealth communities have more crime, more teen births, and more cost-burdened renters because of high concentrations of temporary and seasonal workers, extensive regulations on greenhouse gas emissions, minimum wage laws, and limited housing supply. In the conclusion, we review past attempts to alleviate the effects of poverty and give suggestions on how future policy can be influenced to eventually create a future free of poverty.

Why Do Directors Join Poorly Performing Firms?
Dou, Ying,Zhang, Emma Jincheng
Prior research has suggested that sitting on the board of a poorly performing firm can be undesirable to directors. Yet, almost 60% of these firms are able to appoint new directors following director departures. Contrary to a quality matching explanation, we do not find that only poorly performing directors join these firms. Upon joining poorly performing firms, directors are more likely to fill the leadership positions, without necessarily receiving higher pay. These directors subsequently receive career benefits, especially those who are relatively junior in the pool. As such, the evidence is consistent with the leadership positions providing a certification effect.