Journal of Finance and Data Science最新文献

Machine learning algorithms dedicated to financial time series forecasting have gained a lot of interest. But choosing between several algorithms can be challenging, as their estimation accuracy may be unstable over time. Online aggregation of experts combine the forecasts of a finite set of models in a single approach without making any assumption about the models. In this paper, a Bernstein Online Aggregation (BOA) procedure is applied to the construction of long-short strategies built from individual stock return forecasts coming from different machine learning models. The online mixture of experts leads to attractive portfolio performances even in non-stationary environments. The inclusion of neural networks experts in the aggregation contributes to a better average return, while Ordinary Least Squares with Huber Loss experts contribute to lower risk. The aggregation outperforms individual algorithms, offering a higher portfolio Sharpe ratio, lower shortfall, with a similar turnover. Extensions to expert and aggregation specialisations are also proposed to improve the overall mixture on a family of portfolio evaluation metrics.

Stock pledged loans have become prevalent among large shareholders of listed firms in China. The largest shareholder pledges a greater fraction of her holdings as collateral for credit when the firm is in growth industries, less profitable, not state owned, and has higher leverage. Stock performance of highly pledged firms is indistinguishable from that of firms with low pledge ratios in 2017. During 2018, however, highly pledged firms have worse stock returns and operating performance, and experienced ‘contagion’ – the crash risk of one highly pledged stock spreading to others. Using a regulatory reform in 2013 that allowed securities companies to provide stock pledged loans, we find that obtaining these personal loans had no adverse effects on the firms when the pledge ratio was low. Overall, forced sales of pledged stocks and worsened agency conflict are responsible for the poor performance of highly pledged firms during the 2018 bear market.

Economic indicators are essential for policymaking and strategic decisions in both the public and private sectors. However, due to delays in the release of government indicators based on macroeconomic factors, there is a high demand for timely estimates or “nowcasting”. Many attempts have been made to overcome this challenge using macro indicators and key variables such as keywords from social networks and search queries, but with a reliance on human selection. We present a fully data-driven methodology using non-prescribed search engine query data (Search Big Data) to approximate economic variables in real time. We evaluate this model by estimating representative Japanese economic indicators and confirm its success in nowcasting prior to official announcements, even during the COVID-19 pandemic, unlike human-selected variable models that struggled. Our model shows consistent performance in nowcasting indices both before and under the pandemic before government announcements, adapting to unexpected circumstances and rapid economic fluctuations. An exhaustive analysis of key queries reveals the pivotal role of libidinal drives and the pursuit of entertainment in influencing economic indicators within the temporal and geographic context examined. This research exemplifies a novel approach to economic forecasting that utilizes contemporary data sources and transcends the limitations of existing methodologies.

Option is an important financial derivative. Accurate option pricing is essential to the development of financial markets. For option pricing, existing time series models and neural networks are difficult to extract multi-scale temporal features from option data, which greatly limits their performance. To solve this problem, we propose a novel deep learning model named as MRC-LSTM-CI. It contains three modules, including Multi-scale Residual CNN module (MRC), Long Short-Term Memory neural network module (LSTM) and confidence interval output module (CI). The proposed model can effectively extract multi-scale features from real market option data, and make interval prediction to provide more information to the decision maker. In addition, the proposed model is further improved using the residual prediction strategy, where the output value is chosen as the residual value between BS theory price and actual market price. Experimental results show that our model has better prediction accuracy than other deep learning models and achieves the state-of-the-art performance.

This study investigates the impact of financial content structure on consumer appreciation in Australian Statement of Advice (SOA) documents. SOAs are essential for regulatory adherence and consumer protection, but their complicated nature can hinder consumers' understanding. The research uses independent subcategory variables of comprehension, value, and trust to measure consumer appreciation. Data was collected from 164 financial planning consumers in regional Queensland over 12 months. The research methodology collected both quantitative and qualitative data which was analysed using Analysis of variances, Econometric modelling, and Thematic analysis techniques. Results indicate that financial content structure significantly affects consumer appreciation, with higher levels of appreciation recorded for the introduced financial content structure. Findings have implications for financial advisors and institutions in developing effective strategies for communicating financial information to consumers.

The insurance industry has changed rapidly over the last few decades. One factor in this change is the continuous growth of massive amounts of data that need to be processed properly to be optimally utilized. This has led to a strong wave of advanced processing technologies that can systematically manage big datasets, such as machine learning and artificial intelligence. This study analyzes the current state of research on big data and insurance. Bibliometric analysis and a systematic review were conducted to analyze the patterns and trends of the subject area, with the main focus on citations as a key measurement unit. This analysis is important to fill the existing gap in the examined area because no other bibliometric analysis has been conducted previously on the same subject; it will also help in establishing a scientific background for future research. The research findings verify that the United States is the most popular and cited country in the research area of big data and insurance at both the single authorship and co-authorship levels. Finally, the major impact of the relationship between big data and the insurance sector was marked by human-related aspects.

The construction of replication strategies for contingent claims in the presence of risk and market friction is a key problem of financial engineering. In real markets, continuous replication, such as in the model of Black, Scholes and Merton (BSM), is not only unrealistic but is also undesirable due to high transaction costs. A variety of methods have been proposed to balance between effective replication and losses in the incomplete market setting. With the rise of Artificial Intelligence (AI), AI-based hedgers have attracted considerable interest, where particular attention is given to Recurrent Neural Network systems and variations of the Q-learning algorithm. From a practical point of view, sufficient samples for training such an AI can only be obtained from a simulator of the market environment. Yet if an agent is trained solely on simulated data, the run-time performance will primarily reflect the accuracy of the simulation, which leads to the classical problem of model choice and calibration. In this article, the hedging problem is viewed as an instance of a risk-averse contextual k-armed bandit problem, which is motivated by the simplicity and sample-efficiency of the architecture, which allows for realistic online model updates from real-world data. We find that the k-armed bandit model naturally fits to the Profit and Loss formulation of hedging, providing for a more accurate and sample efficient approach than Q-learning and reducing to the Black-Scholes model in the absence of transaction costs and risks.

Startups are a key force driving economic development, and the success of these high-risk ventures can bring huge profits to venture capital firms. The ability to predict the success of startups is a major advantage for investors to outperform their competitors. In this study, we explore the potential of using publicly available LinkedIn profiles as an alternative and complementary data source to Crunchbase for predicting startup success. We provide a comprehensive review of the existing literature on the factors that influence startup success to create a large set of features for predictive modeling. We train two models for predicting startup success employing light gradient boosting that use LinkedIn data as a standalone and as a complementary data source, and compare them to baseline models based on Crunchbase data. We show that using LinkedIn as a complementary data source yields the best result with a mean area under the curve (AUC) value of 84%. We also provide a thorough analysis of what types of information contribute most to modeling startup success using the Shapley value method. Our models and analysis can be used to develop a decision support system to facilitate startup screening and the due diligence process for venture capital firms.

We study the relation between bond characteristics and corporate bond returns in China's two distinct and segmented bond markets—the interbank market and the exchange market—with a large cross-sectional dataset of 8318 corporate bonds from January 2010 to December 2022. Corporate bonds with large sizes, long maturities, old ages, poor credit ratings and large Amihud illiquidity earn high monthly returns in the interbank market. The return predictive patterns of bond size, time to maturity, and credit rating are similar in the exchange market, but bond age and Amihud illiquidity predict returns in the opposite direction, implying market segmentation. We construct two factors based on credit rating and Amihud illiquidity to represent the common risk of corporate bonds—credit risk and liquidity risk—and use the Hansen-Jagannathan distance to evaluate the performances of factors in explaining the returns of corporate bond portfolios. We find that the two characteristic-based factors help reduce the model specification errors of the five factors in Fama and French (1993).