DOC PREVIEW
PSU ACCTG 597E - Analytic Modeling in Management Accounting

This preview shows page 1-2-3-4 out of 12 pages.

Save
View full document
View full document
Premium Document
Do you want full access? Go Premium and unlock all 12 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 12 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 12 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 12 pages.
Access to all documents
Download any document
Ad free experience
Premium Document
Do you want full access? Go Premium and unlock all 12 pages.
Access to all documents
Download any document
Ad free experience

Unformatted text preview:

Analytic Modeling in Management Accoun tingResearchbyJoel S. DemskiUniversity of FloridaabstractThis Chapter examines analytic or modeling based research, butwith an emphasis on the broader perspective of viewing researchas a portfolio of investment projects. I stress three keys to goodmodeling: primacy of the research question, preparation of themodel, and the Ralph test. I also identify dominant themes in therecent literature.prepared for The Handbo o k of Managem e nt Accounting ResearchAugust, 2005Analytic Modeling in Management Accoun tingResearchThis Chapter focuses on modeling in management accoun ting research. Inthis context, modeling refers to the representation of a concept or process, whileanalytical refers to the use of deductive logic. On the surface, this takes usinto the realm of “research method.” But, you will see, I step lightly on thissubject and concentrate on the more fundamental issues of interpretation andassimilation.In particular, this is not a tutorial or a survey.1Rather, it is an invitation toreflect, to put modeling in its proper perspective. Following some backgroundremarks aimed at research methods I discuss what I consider to be the threekeys to good modeling: primacy of the research question, proper preparation ofthe model, and what I call the "Ralph test." From there I turn to dominantthemes in the literature: hyper vs. muted rationality on the part of the presentand implied actors. Some concluding remarks round out the Chapter.1 background rem arksResearch refers to diligent, systematic inquiry. In its broadest sense accountingdeals with particular institutions, such as formalized measurement and report-ing inside a firm, an audit firm per se, care and feeding of financial informationaimed at an organized trading market, and so on. Accounting research, then,refers to diligent, systematic inquiry into institutional regularities. It is a socialscience exercise in which we use the window of accounting institutions to studybehavior, at both the organizational and individual levels. In broad terms westudy such things as (1) organizational arrangements, including divisionalizedstructures, alliances and allocation of decision rights; (2) decision methods andframes; (3) evaluation and compensation, including costing systems; (4) gover-nance structures; and (5) the comparative advantage of the accounting systemwith its elaborate, nested controls and professional management. Moreover, wedo this in a variety of settings, real and imagined, using a variety of methods.Regardless, the overriding concept is to focus, laser-like, on the issue at hand.This necessitates a focus on first order effects.2The more subtle nuances arepurged from the analysis. When studying an ABC implementation we do notidentify precisely the firm’s technology (e.g., via estimation of a translog modelusing industry data), nor do we delve deeply in to the implementation team’s1Th e C h ristensen and Felth am volumes are the sta rting p oint for anyone interested seri-ously in the topic. R ecent reviews commissioned by the Jouranl of Accounting & Economics(Volumes 31 and 32) should also be consulted, along with appropriate chapters in this Hand-b o ok. Christensen and Demski (2002) is a particular favorite.2Vie wing resea rch a s c on st ru c tin g or estima tin g a Taylor series app roximatio n is a us efulmetaphor. In turn, management accounting research is accounting research in which m anage-m e nt’s b e havi o r is a first order concern. Notice how we now merge into the realm of aud itingor financial reporting improprieties!1psychological profile. Similarly, when studying managerial compensation weabstract from an ov erwhelming array of information flow,tax,andimplicitfactors. Sims (1996, p 105) is particularly insightful when he states: “Advancesin the natural sciences are discoveries of ways to compress data concerning thenatural world — both data that already exists and potential data — with minimalloss of information.”3Successful examples of understanding this art form include option pricing,where transaction costs are ignored, the personal cost term in an agency model,where consumption at work, as in Stafford and Cohen (1974), is surrogated bya generic personal cost assumption, or an ABC model where a variety of costdrivers are used as a substitute for identifying the underlying commodity space,as in Debreu (1959) or Christensen and Demski (1997). Moreover, one shouldnot think reliance on first order effects is confined to modeling. Empiricalcompensation studies, such as Gibbons and Murphy (1992), or experimentalevaluation studies, such as Hackenbrack and Nelson (1996), come to mind.Two implications follow. First, no research exercise is perfect. Movingfrom the research exercise to the issue under study always focuses on first ordereffects and therefore always carries an error term. A model is not going to beperfect (though we certainly hope its logic is), just as the presumed controls inan experimental investigation are not going to be perfect. Errors are alwayspresent. Get used to it!Some errors are, of course, egregious. In good research, however, secondorder errors are tolerated because pushing them to the background helps usfocus on the issue at hand. This is Sims’s compression idea at work.The second implication is less comforting: we know very little about howto sort among potential error patterns. This is the art dimension to goodresearch. The ageless adage is appropriate: I’ll tell you when I see it! Studyof art history is essential for the budding fine artist, just as study of accountingresearch history is essential for the budding accounting researcher. Yet I fearwe give short shrift to the art of doing good research, including the importanceof extended study of our own "art history."4Of course, the Blac kwell Theorem has something to say here. We know (e.g.,Blackwell’s classic “Comparison of Experiments") that one research program isbetter than another if the errors in the second can be modeled as if they arestatistically equal to the errors of the first plus noise.5To paraphrase, supposethere is an uncertain state of the world or forthcoming event that will take on onefrom among a given list of possible events or states. Denote the possible events3Closely related is Ijiri’s (1971 ) trea tise on the theory o f agg regation. Likewise the Heisen-berg U n c er ta inty P rin c iple g uara nte es lim its to th e power of o bservation in


View Full Document

PSU ACCTG 597E - Analytic Modeling in Management Accounting

Download Analytic Modeling in Management Accounting
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view Analytic Modeling in Management Accounting and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view Analytic Modeling in Management Accounting 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?