A fast and scalable ensemble of global models with long memory and data partitioning for the M5 forecasting competition

This work presents key insights on the model development strategies used in our cross-learning-based retail demand forecast framework. The proposed framework outperforms state-of-the-art univariate models in the time series forecasting literature. It has achieved 17th position in the accuracy track of the M5 forecasting competition, which is among the top 1% of solutions.     

Toward a new business model of retail industry: The role of brand experience and brand authenticity

While the customer-to-manufacturer (C2M) business model has received increasing attention as a new business model for e-commerce and retail industry, little is still known about it and the effect of its approach. This study aims to understand how brand-related stimuli in C2M environments affect customer responses as the worldwide COVID-19 pandemic. The outcomes reveal that the Sensory, affective, and intellectual aspects of brand experience positively influence brand authenticity. Brand authenticity has a positive effect on behavioral intention, such as reuse intention and word-of-mouth. Additionally, this research finds that social presence moderates the association between the sensory aspect of brand experience. Thus, this study can suggest a C2M business model as a means of sustainable operation of the retail industry to both researchers and practitioners in relation to the retail industry.     

Do goodwill impairments affect audit opinions? Evidence from China

There has been a steady growth of goodwill impairments in the Chinese stock market since the adoption of the impairment approach in accounting. The influence of goodwill impairments on a firm’s financial position and profitability give reason to doubt its current and future performance. We examine whether auditors, as a crucial external monitor, identify the information risks of goodwill impairments and express their concerns about financial reporting quality in their audit opinions. Using a sample of firms listed on China’s A-share market from 2007 to 2017, we test the association between goodwill impairments and the type of audit opinion received in the same financial period. Our findings are as follows. First, the probability of receiving a modified opinion increases with the amount of goodwill impairments. Second, the positive association between goodwill impairments and modified audit opinions is driven primarily by earnings management risks. Third, this positive association is more salient when auditors are industry experts and there is no auditor–client mismatch. Fourth, auditors are more sensitive to the amount of goodwill impairments than to their mere existence. Overall, we document that auditors perceive goodwill impairments as a signal of information risks and communicate their concerns to investors to avoid litigation.     

M5 competition uncertainty: Overdispersion,distributional forecasting,GAMLSS, and beyond

The M5 competition uncertainty track aims for probabilistic forecasting of sales of thousands of Walmart retail goods. We show that the M5 competition data face strong overdispersion and sporadic demand, especially zero demand. We discuss modeling issues concerning adequate probabilistic forecasting of such count data processes. Unfortunately, the majority of popular prediction methods used in the M5 competition (e.g. lightgbm and xgboost GBMs) fail to address the data characteristics, due to the considered objective functions. Distributional forecasting provides a suitable modeling approach to overcome those problems. The GAMLSS framework allows for flexible probabilistic forecasting using low-dimensional distributions. We illustrate how the GAMLSS approach can be applied to M5 competition data by modeling the location and scale parameters of various distributions, e.g. the negative binomial distribution. Finally, we discuss software packages for distributional modeling and their drawbacks, like the R package gamlss with its package extensions, and (deep) distributional forecasting libraries such as TensorFlow Probability.     

Understanding machine learning-based forecasting methods: A decomposition framework and research opportunities

Machine learning (ML) methods are gaining popularity in the forecasting field, as they have shown strong empirical performance in the recent M4 and M5 competitions, as well as in several Kaggle competitions. However, understanding why and how these methods work well for forecasting is still at a very early stage, partly due to their complexity. In this paper, I present a framework for regression-based ML that provides researchers with a common language and abstraction to aid in their study. To demonstrate the utility of the framework, I show how it can be used to map and compare ML methods used in the M5 Uncertainty competition. I then describe how the framework can be used together with ablation testing to systematically study their performance. Lastly, I use the framework to provide an overview of the solution space in regression-based ML forecasting, identifying areas for further research.     

Exploring the representativeness of the M5 competition data

The main objective of the M5 competition, which focused on forecasting the hierarchical unit sales of Walmart, was to evaluate the accuracy and uncertainty of forecasting methods in the field to identify best practices and highlight their practical implications. However, can the findings of the M5 competition be generalized and exploited by retail firms to better support their decisions and operation? This depends on the extent to which M5 data is sufficiently similar to unit sales data of retailers operating in different regions selling different product types and considering different marketing strategies. To answer this question, we analyze the characteristics of the M5 time series and compare them with those of two grocery retailers, namely Corporación Favorita and a major Greek supermarket chain, using feature spaces. Our results suggest only minor discrepancies between the examined data sets, supporting the representativeness of the M5 data.     

Are USDA reports still news to changing crop markets?

This study investigates whether major USDA reports still provide important news to changing crop markets. The news component of each report, or market “surprise,” is measured as a difference between the USDA estimate and its private expectation in corn, soybeans, and wheat markets. Changes in the relevance of USDA information are assessed by examining changes in the magnitude of market surprises and shifts in the futures price reaction to these surprises, which isolates the impact of each report. The stable size of market surprises over time suggests that competition from alternative data sources has not reduced the news component of USDA crop reports. Increasing price reaction to most reports, including those facing competition from alternative information sources, suggests that value of public information may be enhanced in uncertain markets affected by structural changes.     

U.S. food aid and American exports to recipient countries during the Cold War

In this article, we test to what extent the food aid granted by the U.S. during the Cold War was strategically motivated and used to promote U.S. interests. Using the data for the period 1971–1990, we investigate whether U.S. wheat aid had an effect on recipient countries’ total import of American products. The evidence we provide suggests a positive and robust relationship and thus it is consistent with the argument that U.S. food aid helped to create larger markets for U.S. producers.     

Commemorating the 50‐Year Anniversary of Ball and Brown (1968): The Evolution of Capital Market Research over the Past 50 Years

We commemorate the 50th anniversary of Ball and Brown [1968] by chronicling its impact on capital market research in accounting. We trace the evolution of various research paths that post–Ball and Brown [1968] researchers took as they sought to build on the foundation laid by Ball and Brown [1968] to create a body of research on the usefulness, timeliness, and other properties of accounting numbers. We discuss how those paths often link back to the groundwork laid and questions originally posed in Ball and Brown [1968].     

FREE VERSUS SECOND‐BEST ENTRY IN COURNOT OLIGOPOLIES WITH FIRM‐SPECIFIC NETWORKS

This paper endogenizes the number of firms in an industry with positive network effects, complete incompatibility, and firms that compete in quantity. To this end, we compare two possibilities: free entry and second‐best number of firms (the one that maximizes social welfare). We show that with business‐stealing competition, free entry yields, in general, more firms than the socially optimal solution. In addition, we find that by the nature of the industry with firm‐specific networks, total production may be greater or lower under free entry than with a regulator; moreover, some industries attain their maximum social welfare with a monopoly.