Machine learning methods in finance: Recent applications and prospects

We study how researchers can apply machine learning (ML) methods in finance. We first establish that the two major categories of ML (supervised and unsupervised learning) address fundamentally different problems than traditional econometric approaches. Then, we review the current state of research on ML in finance and identify three archetypes of applications: (i) the construction of superior and novel measures, (ii) the reduction of prediction error, and (iii) the extension of the standard econometric toolset. With this taxonomy, we give an outlook on potential future directions for both researchers and practitioners. Our results suggest many benefits of ML methods compared to traditional approaches and indicate that ML holds great potential for future research in finance.     

通向2075之路 全球经济增长放缓,但将继续趋同发展

<正>高盛全球投资研究部宏观经济研究团队2023年发表关乎未来数十年全球增长预测的四个核心主题,即人口增速放缓拖累全球潜在经济增长率;在亚洲经济引擎的带领下,新兴市场的趋同趋势保持不变;美国经济独有的超预期表现在未来十年不太可能重演;全球不平等减少,国家内部不平等增加。同时文章揭示了全球增长面临的两大主要长期风险:保护主义和气候变化。     

A NOTE ON BIAS IN CAPITAL BUDGETING INTRODUCED BY STOCHASTIC LIFE

This note extends work by Young and Contreras and by Rosenthal on the present worth of cash flows under uncertain timing. In capital budgeting analysis, the use of the expected life of a project instead of the life distribution of the project biases the estimate of its expected net present value. In most situations the bias results in an overestimate of the expected net present value of the project. When the exact life distribution is unknown, the bias can be approximated by Taylor series expansion. The sensitivity of the bias to the discount rate, to cash flow patterns, and to income taxes is also investigated.     

Employee mobility,spin‐outs,and knowledge spill‐in: How incumbent firms can learn from new ventures

Research summary: We consider conditions in which incumbent firms are particularly poised to benefit from knowledge spilling in from new ventures that employ individuals previously employed by the focal incumbent firm. We distinguish between inventors who leave their incumbent employers to found spin‐outs and those who become non‐founding employees of existing new ventures. Using a sample of new ventures and incumbent firms in the U.S. information technology (IT) sector, we find that incumbents are more likely to benefit from patented knowledge that spills in from their spin‐outs than from new ventures that employ non‐founding inventors formerly employed by the respective incumbent. Any advantage that parent firms have in reaping such knowledge quickly dissipates, however, when these parents have a history of misappropriating the intellectual property of others. Managerial summary: It has long been acknowledged that new ventures can acquire valuable knowledge from their larger and more established counterparts by hiring away their talented employees. We consider the possibility of a reverse flow of knowledge where established firms learn from those new ventures that have poached employees from them. We find that established information technology (IT) firms are more likely to learn and build on the technology of their spin‐outs (i.e., new ventures founded by their former inventors) than from new ventures that simply employ non‐founding inventors formerly employed by the respective IT firm. Any advantage that these IT firms had in reaping technical know‐how from their spin‐outs quickly dissipated, however, when they had a history of misappropriating the intellectual property of others. Copyright © 2016 John Wiley & Sons, Ltd.     

Voting, Regulation, and the Railroad Industry: An Analysis of Private and Public Interest Voting Patterns

We examine the role of public and private interests in the passage of major legislative bills that have governed the rail industry since 1887. Our model of voting patterns in the House and Senate introduces novel measures of industry concentration, competition, and network characteristics. We find that both the level and concentration of rail infrastructure in a state, the presence of water competition, and the level of agricultural production each have a major influence over voting patterns not only for the inception of regulation but also in the major bills that deregulated the industry in the 1970s and in 1980.     

Changes in the value relevance of accounting information over time: Evidence from the emerging market of China

We investigate the changes in the value relevance of accounting information among Chinese firms over the past two decades, during which accounting reforms are launched to provide decision makers with increased disclosure and higher quality financial information. We also investigate the factors that differentiate firms showing significant value relevance improvement from firms showing little improvement. We find increases in the value relevance of some financial variables and decreases in others, which suggests that accounting numbers help to explain the pricing process of stock shares although at different levels. In addition, we find that value relevance improvements are more pronounced for smaller firms, firms with lower growth rates, and those with greater asset tangibility. We also document that value relevance improvements are generally lower in an exuberant stock market. These results have implications for a variety of information users and policy makers in emerging countries which are reforming their accounting systems.     

Backtesting Stochastic Mortality Models

Abstract

This study sets out a backtesting framework applicable to the multiperiod-ahead forecasts from stochastic mortality models and uses it to evaluate the forecasting performance of six different stochastic mortality models applied to English & Welsh male mortality data. The models considered are the following: Lee-Carter’s 1992 one-factor model; a version of Renshaw-Haberman’s 2006 extension of the Lee-Carter model to allow for a cohort effect; the age-period-cohort model, which is a simplified version of Renshaw-Haberman; Cairns, Blake, and Dowd’s 2006 two-factor model; and two generalized versions of the last named with an added cohort effect. For the data set used herein, the results from applying this methodology suggest that the models perform adequately by most backtests and that prediction intervals that incorporate parameter uncertainty are wider than those that do not. We also find little difference between the performances of five of the models, but the remaining model shows considerable forecast instability.     

The Impact of No-Fault Legislation on Automobile Insurance

Abstract

Since its inception, the effectiveness of no-fault legislation has been highly debated. Although some research suggests that no-fault laws are effective in reducing costs, other evidence suggests that the current no-fault systems may not meet the original objectives. This study provides a detailed assessment of the relation of no-fault laws and automobile insurance losses for the period 1994 to 2007. By examining total automobile insurance losses along with liability and personal injury protection losses, we are able to determine if and how specific provisions of the laws are related to claims costs. We find a negative relation between the presence of a no-fault law and total losses, which suggests that no-fault systems are associated with lower losses than the traditional tort system. In addition, an examination of no-fault-only states suggests that specific provisions of no-fault laws, such as thresholds and limitations on benefits, have some effect on losses. With the sunset of Colorado’s no-fault legislation in 2003, the recent passage of Personal Injury Protection Reform in Florida, and proposed federal choice legislation, the overall impact of no-fault as well as the specific components of the laws are of heightened importance to consumers, insurers, and lawmakers.     

Real Estate Risk Exposure of Equity Real Estate Investment Trusts

This study examines the linkage between equity real estate investment trust (REIT) returns and the private real estate factor. The results reveal a tighter connection between REIT and the private real estate market starting from 1993. In addition, large-cap REITs seem to behave more like real estate than do small-cap REITs. Overall, the results are consistent with three notions: (1) that institutional investors provide information-gathering services (Bradrinath et al., Rev. Financ. Stud., 8:401–430, 1995), (2) that a more sophisticated investor base improves information flow, and (3) that a high degree of participation from institutional investors strengthens the linkage between REIT returns and the underlying real estate factor (Ziering et al., The evolution of public and private market investing in the new real estate capital markets, Prudential Real Estate Investors, Parsippany, NJ, 1997).