Machine learning improves accounting: discussion, implementation and research opportunities
- PDF / 1,523,076 Bytes
- 21 Pages / 439.642 x 666.49 pts Page_size
- 118 Downloads / 198 Views
Machine learning improves accounting: discussion, implementation and research opportunities Jeremy Bertomeu1
© Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract Machine learning has been growing in importance in empirical accounting research. In this opinion piece, I review the unique challenges of going beyond prediction and leveraging these tools into generalizable conceptual insights. Taking as springboard “Machine learning improves accounting estimates” presented at the 2019 Conference of the Review of Accounting Studies, I propose a conceptual framework with various testable implications. I also develop implementation considerations panels with accounting data, such as colinearities between accounting numbers or suitable choices of validation and test samples to mitigate between-sample correlations. Lastly, I offer a personal viewpoint toward embracing the many low-hanging opportunities to bring the methodology into major unanswered accounting questions. Keywords Machine learning · Accounting · Estimates · Modelling JEL Classification C4 · C5 · G3 · M2 · M4 In their new book, The End of Accounting and the Path Forward For Inventors and Managers, Baruch Lev and Feng Gu provocatively argue that accounting has not kept pace with secular changes in economic structures. With the decline in the importance of controlling physical means of production, the value of businesses is increasingly driven by intangibles assets - a knowledge economy in which know-hows, customers, brands, and networks explain investor value. Their analysis further takes stock of the growing disconnect between markets where antiquated procedures focus on minutia of historical events of no interest to investors, bury relevant information into aggregated reports, and are often contaminated by managerial judgment. Baruch Lev is a co-author of this year’s must-read paper at the Review of Accounting Studies 2019 conference (Ding et al. 2019) and that we should see similar themes Jeremy Bertomeu
[email protected] 1
Olin School of Business, Washington University, St. Louis, MO, USA
J. Bertomeu
comes as no surprise. This begs the question: Can machine learning help provide better high-quality forward-looking information? Indeed, interests in the accounting community have been growing to integrate machine learning as a set of reporting tools to predict, diagnose, and improve reporting quality, with various new studies showing the quality of machine learning to predict errors and irregularities (Perols 2011; Perols et al. 2017; Bertomeu et al. 2019; Bao et al. 2019), measuring information content (Li 2010; Barth et al. 2019), analyzing financial statements (Binz et al. 2020) or improving audit procedures (Gerakos et al. 2016; Sun 2019), among many others. In this essay, my objective is twofold. First, leveraging on the insights from Ding et al. (2019), I will describe a new research paradigm that is slowly emerging from the application of machine learning to accounting research. I will argue that, while the tools of machi
Data Loading...
