We use machine learning for relative valuation and peer firm selection. In out‐of‐sample tests, our machine learning models substantially outperform traditional models in valuation accuracy. This outperformance persists over time and holds across different types of firms. The valuations produced by machine learning models behave like fundamental values. Overvalued stocks decrease in price and undervalued stocks increase in price in the following month. Determinants of valuation multiples identified by machine learning models are consistent with theoretical predictions derived from a discounted cash flow approach. Profitability ratios, growth measures, and efficiency ratios are the most important value drivers throughout our sample period. We derive a novel method to express valuation multiples predicted by our machine learning models as weighted averages of peer firm multiples. These weights are a measure of peer–firm comparability and can be used for selecting peer‐groups

This study uses machine‐learning methods to predict next‐period change in profitability based on a model proposed by Penman and Zhang (2004, Working paper, Columbia University and University of California, Berkeley; “PZ”). We find that new machine‐learning methods predict out of sample substantially better than traditional regression methods and provide richer interpretations about the role and impact of different predictor variables through their nonlinear relationships and interaction effects. For example, our results contrast with previous research by showing that both components of the DuPont decomposition (change in profit margin and change in asset turnover) are informative of next‐period changes in profitability. Our results are robust across different performance metrics, alternative machine‐learning models, and software. Furthermore, an unconstrained machine‐learning model using a larger feature space could not significantly improve the performance of the PZ model. PZ variables alone accounted for most of the explanatory power of the unconstrained model, suggesting the PZ model is both well specified (in terms of feature selection) and robust in higher dimensional settings. With respect to the economic significance of this information, we find mixed results. The market appears to adjust its expectations more in line with the machine‐learning predictions relative to the PZ model but the portfolio returns are not significantly different

This paper uses a comprehensive set of predictor variables from the five largest Eurozone countries to compare the performance of simple univariate and machine learning-based multivariate models in forecasting stock market crashes. In terms of statistical predictive performance, a support vector machine-based crash prediction model outperforms a random classifier and is superior to the average univariate benchmark as well as a multivariate logistic regression model. Incorporating nonlinear and interactive effects is both imperative and foundation for the outperformance of support vector machines. Their ability to forecast stock market crashes out-of-sample translates into substantial value-added to active investors. From a policy perspective, the use of machine learning-based crash prediction models can help activate macroprudential tools in time

Practitioners allocate substantial resources to technical analysis whereas academic theories of market efficiency rule out technical trading profitability. We study this long-standing puzzle by applying a diverse set of machine learning algorithms. The results show that an investor can find profitable technical trading rules using past prices, and that this out-of-sample profitability decreases through time, showing that markets have become more efficient over time. In addition, we find that the evolutionary genetic algorithm’s attitude in not shying away from erroneous predictions gives it an edge in building profitable strategies compared to the strict loss-minimization-focused machine learning algorithms

Machine-learning methods exploit fund characteristics to select tradable long-only portfolios of mutual funds that earn significant out-of-sample annual alphas of 2.4% net of all costs. The methods unveil interactions in the relation between fund characteristics and future performance. For instance, past performance is a particularly strong predictor of future performance for more active funds. Machine learning identifies managers whose skill is not sufficiently offset by diseconomies of scale, consistent with informational frictions preventing investors from identifying the outperforming funds. Our findings demonstrate that investors can benefit from active management, but only if they have access to sophisticated prediction methods

Natural language is a key form of business communication. Textual analysis is the application of natural language processing (NLP) to textual data for automated information extraction or measurement. We survey publications in top accounting journals and describe the trend and current state of textual analysis in accounting. We organize available NLP methods in a unified framework. Accounting researchers have often used textual analysis to measure disclosure sentiment, readability, and disclosure quantity; to compare disclosures to determine similarities or differences; to identify forward‐looking information; and to detect themes. For each of these tasks, we explain the conventional approach and newer approaches, which are based on machine learning, especially deep learning. We discuss how to establish the construct validity of text‐based measures and the typical decisions researchers face in implementing NLP models. Finally, we discuss opportunities for future research. We conclude that (i) textual analysis has grown as an important research method and (ii) accounting researchers should increase their knowledge and use of machine learning, especially deep learning, for textual analysis

We investigate (1) how well a machine learning algorithm can predict one-year ahead effective tax rates (ETRs) and (2) which items in the financial statements and notes are most useful for these predictions. We compare our machine-generated ETR predictions with those from ETRs implied by analysts’ earnings forecasts and find the algorithm’s predictions are less biased, more precise, and explain more of the variance in future ETRs. We then use Explainable AI (based on Shapley values) to measure the usefulness of each disclosure item in the algorithm’s predictions. We find that while some tax-related items are useful, others offer minimal value. Using the machine learning algorithm’s use of information as a benchmark, we then further use Shapley values to examine which information is underweighted or overweighted by analysts. Overall, our results help inform standard setters on the relevance of certain tax disclosures in achieving the objective of predicting future ETRs

Major customers in strong bargaining position can exert pressure on dependent suppliers and adversely affect their financing conditions. Consistent with this prediction, our analysis shows that the concentration of customer bases can enhance the bargaining power of downstream customers in supplier-customer interactions, leading to the significant deterioration of financing constraints for upstream firms. We then introduce a novel machine-learning approach to analyze the heterogeneous effect of customer concentration. This allows us to identify approximately 15 % of the firms, especially those small non-SOEs, as most vulnerable to customer concentration as the bargaining effect dominates

We show, using machine learning, that fund characteristics can consistently differentiate high from low-performing mutual funds, before and after fees. The outperformance persists for more than three years. Fund momentum and fund flow are the most important predictors of future risk-adjusted fund performance, while characteristics of the stocks that funds hold are not predictive. Returns of predictive long-short portfolios are higher following a period of high sentiment. Our estimation with neural networks enables us to uncover novel and substantial interaction effects between sentiment and both fund flow and fund momentum

Building on administrative data and machine-learning models, I develop a firm-specific measure of regulatory intensity: cost of compliance with all federal paperwork regulations. Regulatory intensity increases the cost of goods sold and overhead spending (SGA). It also incentivizes companies to reduce capital investment, hire fewer employees, and lobby more. The effects are particularly strong among financially constrained firms and those with irreversible investment opportunities, suggesting that regulation affects companies through budgetary pressures and heightened uncertainty. The findings highlight the real effects of regulation and the underlying mechanisms