My core mission as an academic is to create and disseminate knowledge.  Usually, we do this by teaching and by publishing peer-reviewed articles.

However, two-and-a-half years into my program, my day-to-day work is far removed from teaching and publishing in top accounting journals.  I won’t teach my first class until this coming summer, and the publication process is long and slow.  As a result, it’s easy to feel like my day-to-day activities (reading, writing models, running code, reading, thinking about accounting, writing drafts, reading) aren’t relevant to my broader goal of creating and disseminating knowledge about accounting.

This blog is meant to counter those feelings.  Every week, I’ll try to post a thoughtful reflection about an accounting topic that’s been on my mind.  There will probably be some data analysis or some simple models, but nothing elaborate enough to form a full research paper.  Please reach out if you’d like to chat about any of these musings…I’m always happy to talk.

 

 

 

Text Analysis of JAR Articles

November 30th, 2018

In our Ph.D. program, we spend a lot of time discussing what does and does not fall within the scope of accounting research.  In the coursework phase, we read the mission statements of several accounting journals, as well as articles and think-pieces where senior faculty lay out their opinions on what is and is not accounting research.  Moving out of the coursework phase, we still frequently debate whether papers are accounting in our pre-workshop reading groups.   These debates normally center on whether and why a paper belongs in accounting as opposed to economics or finance.

Over my brief time in the Ph.D. program, I’ve attached myself to a number of working definitions for accounting research.  So far, all of them have proven unsatisfactory.  When I entered the program, I thought accounting research was just financial reporting research.  Of course, this simplistic view can immediately be written off because it excludes the entire fields of managerial and tax accounting.

Later, I toyed with the idea that accounting research is research about what accountants do.  However, this definition merely punts the issue to delimiting who is and is not an accountant.  Are only CPAs accountants?  Are CPAs accountants even if they don’t work as accountants?  Are they still accountants when they are at the dentist’s office?  And if so, does this mean accounting research includes studies about how CPAs interact with their dentists?

Or, perhaps CPAs are only accountants when they are doing accounting.  Then we can also call non-CPAs accountants when they are doing accounting.  And what does it mean to do accounting?  This line of questioning brought me to my next working definition: accounting is the systematic measurement, summary, and disclosure of financial information.  Then, whenever we are systematically measuring, summarizing, or disclosing financial information, we are doing accounting.  But, what is financial information?  Balance sheets and income statements disclose more than just the amounts held in financial instruments.  They list more than just the values of bank accounts, bonds, and securities.  They also include dollar amounts for (i.e., information about) equipment, real estate, and intangibles.  But does that mean that information about real estate and intangibles is financial information?  If so, it seems that we must accept that information about anything is financial information, as long as it can be described (however poorly or inaccurately) in dollar amounts.  But then, our definition of accounting research is reduced to the systematic measurement, summary, and disclosure of information.

But this definition is incredibly broad!  Isn’t the entire essence of all scientific fields the systematic measurement, summary, and disclosure of information.  Does this mean that Einstein and Newton were accountants?  They did systematically measure, summarize, and disclose information about the behavior of matter in space.  Do we also accept that all written language is accounting?  Is classical literature accounting?  Is mathematics accounting (I mean, the historical development of mathematics and accounting are closely intertwined…)?

My current working definition of accounting, inspired by a short article in the Abacus by Mary Barth, is the systematic measurement, summary, and disclosure of information to hold someone(s) accountable.  This definition, while still quite broad, distinguishes accounting from many of the science disciplines.  For example, Einstein’s Theory of Relativity is not an accounting work under this view.  However, this definition still faces several challenges.  For one, its scope is much broader than the current focus of most contemporary accounting researchers.  For example, research about the use of journalism as a tool for holding politicians accountable does not appear in the top scholarly accounting journals.  Nevertheless, political journalism easily falls within the systematic measurement, summary, and disclosure of information to hold others accountable.

Does this imply that my working definition is too broad, or that accounting researchers are missing out on some large chunks of accounting turf?  Ultimately, I’m not sure, but it is something I’m looking forward to thinking about over the next several years.

Wow, this has been a long introduction to what was meant to be a very short post.  At any rate, I decided at some point that it would be fun to do a quick and dirty text analysis of all the accounting research articles published in the last couple of years.  For now, I’ve limited myself to full-length JAR articles from 2018 (I found 35 of ’em).  Here are the most commonly used words after removing the usual stop words (plus some redundant words like “Journal” and “Chicago”):

top_words

The list is not super shocking.  Our discipline talks a lot about firms, earnings, and information.  It is curious to see how often words from my working definition appeared in the list: disclosure and information both appear relatively early. Measurement also makes a respectable showing towards the end.  Notably absent, however, are systematic, accountable, and summary.  Also notably absent is the term decision.  (Decision does not appear in my working definition, but I still expected it to appear more often since one of the main goals of accounting is to assist decision-makers).

Based on the most used words in JAR, a better working definition would probably include words like financial, non-GAAP, managers, market,  and bank.   Perhaps we should define accounting as the disclosure of corporate financial information, especially earnings, to the market.  I don’t find this definition of accounting to be nearly as satisfying as my current one, but it does appear to be more representative of published work.

Before making any hasty conclusions, let’s snoop around a bit more.  Single words are not all that informative, and we also don’t know if only one or two papers drove some of those common words.  Perhaps more useful would be a list of common phrases, along with the number of articles they appear in:

top_phrases

This list keeps only the bigrams and trigrams that were used at least 35 times across at least five articles.  Words like financial, information, and disclosure retain their top spots,  but bank and market are nowhere to be seen (of course neither are systematic, accountable, summary,  decision, or user).

So, I’m not quite sure what to make of this… I feel like I should come to one of three conclusions:

(1) Adopt a new working definition of accounting that better reflects published work

(2) Embrace my current working definition and start breaking ground on the vast frontier of accounting topics that remain largely unpublished (i.e., something about measurement, summary, and disclosure that isn’t about corporations or banks).

or

(3) Decide that defining accounting research using the most commonly used words from accounting publications is deeply flawed.  Perhaps the authors of JAR publications really are talking about the summary of information and the creation of accountability despite using different words.

I certainly see (3) as a possibility.  But I don’t think (3) is the 100% correct answer.  Probably, the right answer is some combination of the 3.  My current working definition is likely a little too broad; current accounting research is probably a little over-focused on corporate earnings, and the most used words in JAR probably don’t need to be the building blocks for a definition of accounting research.  Hopefully, the answer will become a little less hazy during my next few years of study.