When I started a PhD program I spent a lot of time dealing with citations. I know I’m not the only one who has struggled with this issue. We’ve had a couple of FASRI posts (e.g. this post asking about codification citations) asking about specific citation questions and I’ve had several friends ask what format to use.

A couple of things make accounting citations problematic. First, citation formats are not consistent across the various accounting journals. Second, to my knowledge most of the journals don’t publish formats.

The result has been that I end up copying a lot of citations from other papers, realizing there are gross inconsistencies in my own bibliographies, and so creating my own citation style on the fly. My citation style is very consistent within papers, but inconsistent between papers and completely inconsistent with any acceptable format for any one journal. So far none of my professors have had a problem with my formatting, but it seems to have taken much more effort than should be required. I’m not doing a PhD just to learn how to create bibliographies.

I did recently find the following link by Weber State University that explains the citation format for AAA publications. However, while the cite thoroughly explains citation details, it doesn’t name the specific citation format. Hence, I don’t know which of the 1500+ formats to use in a software program. Similarly, I don’t know the name of citation formats used by other journals.

Does anyone

1. Have links to/copies of citation guides by the various journals (e.g. TAR, JAR, etc.)?
2. Know what the names of the preferred citation formats are for each journal?

Also, what format do you usually start a paper in? Do you change the citation format as you submit papers to different journals?

**UPDATE: Many thanks to David Godsell who provided the following two files with guidelines for TAR and CAR:
TAR Editorial Policy & Style Information
CAR Style Guidelines

If you know of guidelines for other journals please pass them along. I’m happy to upload the files for you.

As I look back on the semester, I thought, why micro? Why is microeconomics so important for accounting researchers that almost every PhD student is required to take it in their first semester. What’s the big picture? Having spent a semester asking that very question, I thought I’d offer the insights I’ve gleaned. I’d love to hear your thoughts as well.

1. Proof Techniques. Most accounting PhD’s don’t do much in the way of analytical research. However, that in no way exempts them from understanding that methodology. Understanding how models are created and used can be an extremely valuable tool for archival and experimental researchers as they look for testable results of models.
2. Welfare Economics. The fundamental reason for accounting is to attempt to solve the principal-agent problem. Economics provides a set of tools to understand the welfare gains and losses from policy changes.
3. Producer and Consumer Theory. While accounting researchers primarily focus on accounting for transactions, it is beneficial as well to understand why the transaction occurred in the first place. How do firms make profits? How do they balance between capital and labor? How does risk aversion and insurance drive investment decisions? Understanding the “why” behind a transaction is a critical step in understanding the “how” to account for it. This is arguably even more important in a principle’s-based framework.

Please add to my list if you have ideas.

SLCapex is a stock exchange owned and operated by “residents” of the online virtual world Second Life. Despite its almost complete lack of regulation and legal protections against fraud or insider trading, issuers were able to raise approximately US$145,000 from investors, which grew to US$900,000 in market value before plummeting, resulting in overall investor returns of -71%. Investors in large issuances lost more than investors in small issuances, and small investors experienced more severe losses relative to large investors when more money was at stake, suggesting that the market did a poor job of protecting investors from issuers and of providing a level playing field for investors. Theories from financial economics can explain the markets’ poor performance in the absence of regulatory and legal institutions, but cannot easily explain why issuers were able to raise capital in such a setting.

Traditional economic theory tells us that firms that can’t provide credible financial reporting and other investor protections won’t be able to raise capital, because investors protect themselves from the obvious risks of fraud.  This reasoning is the bedrock of libertarian-style claims that market regulation is unnecessary.  But laboratory research shows repeatedly that people fail to protect themselves from risks of exploitation.  Sometimes people trade too much (violating the Milgrom-Stokey no-trade theorem by exposing themselves to others with better information).  My favorite example is a paper by Forsythe, Lundholm and Reitz (JF 1999) that was once called the “half-a-sucker” paper.  Why?  It showed that in the absence of credible standards against overreporting asset values, sellers would indeed overreport and buyers would believe them, to their detriment.  But since the buyers and the sellers were the same people–they alternated between acting as buyers and acting as sellers–the authors modified a famous quote, noting that half-a-sucker is born every minute.

The paper on SLCapex suggests that another famous quote could be modified.  Why would issuers try to raise capital in a market with no investor protections?  Because you can fool all of the people some of the time, and and some of the people all of the time–and that is more than enough to make it worth trying to do so.

And that, in turn, is enough to make it worth imposing regulations that wouldn’t be necessarily if the traditional assumptions of economic theory were valid.

Kachelmeier asks an almost accusatory question at the beginning of the paper: “What can we learn from experimentation that adds to the knowledge gleaned from the model itself?” It almost seems like a no-win scenario. Either the experiment confirms the results of the model . . . and we learn nothing new, or the experiment gives different results from those predicted by the model . . . and we’re left wondering who goofed, the experimenter or the modeler. Luckily Kachelmeier answers his question. Behavioral researchers should not attempt to confirm models, but rather examine reasons why a model might be wrong:

Why, behaviorally, might actual decisions depart from the model’s predictions. If the researcher has a good answer to this question, what was described earlier as a seemingly no-win situation could instead become a no-lose situation. That is, if results support the model, the researcher could point out that this support came in spite of ex ante behavioral considerations to the contrary. If the model’s predictions are not supported, the researcher would have a theoretical basis for explaining why they were not.

In many ways this sounds similar to behavioral economics. Behavioral economists who simply try to show a model to be right or wrong usually don’t get much in the way of publications. On the other hand, if they bring psychology theory into economics research they can show limitations, exceptions, and modifications that a modeler may want to consider. Behavioral researchers in accounting can do the same thing. Rather than simply confirm or reject models, researchers can examine potential changes and limitations to models in a meaningful way:

One way to sum up my comments would be to suggest that experimentalists in accounting gradually move from demonstrating the predictive ability of their models toward investigating possible behavioral challenges to those models. The difference is one of perspective. The demonstrator sets forth a rigorous model of strategic incentives and then endeavors to design an experiment that has the best chance to support the model’s predictions. The investigator, by contrast, adopts much more of a falsificationist approach, looking for likely environmental features and behavioral preferences that might add qualifications to the model.

Just over a year ago I took the GMAT. While I spent a good deal of time preparing for the GMAT, I think I spent at least as much time trying to figure out where I wanted my GMAT scores to be sent. That was one of the most difficult parts of the entire recruiting process: choosing my initial list of schools. So, I thought I’d offer my two cents for anyone who may be in a similar dilemma. I’ve grouped my thoughts around questions that I was desperately seeking answers to. On every one of these questions realize that this is what I learned from my experience. It doesn’t necessarily apply to you, and you should definitely seek other ideas.

Question: How many schools should I apply to?

Answer: The common philosophy is that more is better. Each additional application increases your likelihood of getting accepted. While I think that’s a good rule of thumb, let me add a caveat. Your resources are limited. Each and every application needs to be personal. In the job market, blindly sending out a million resumes on the web doesn’t work nearly as well as making  a concerted effort for a smaller set of openings. The same is true, if not amplified, for PhD programs. For example, letters of intent need to be customized (see Rob’s post on this topic). Additionally, be respectful of the people who are writing letters of recommendation. Asking for 65 letters within a two week period is a bit overboard.

Question: How do I decide which schools to apply to?

Answer: The first thing to do is decide what your interests are. Do you want to teach or to research? What subject areas are you interested in (e.g., financial, audit, tax, managerial, systems)? Are you more interested in archival research or more behavioral/experimental research? You don’t need to know exactly what you want to do, but it’s good to have at least some sort of idea. From there, looking at school rankings (for some ranking options see the U.T. Dallas and BYU websites) may be a good place to start. However, it is in no way an endpoint.

In my opinion the most important thing to consider is who you would work with. This is starkly different from an undergrad or master’s experience where the name of the school is the most important thing. In a PhD program, your focus is on quality and direction of training. After I assembled a large list of schools in my interest area I identified professors at each school who I might work with. Find out a little bit about them. Are they still actively researching? Do their recent papers look interesting to you? The BYU website helps with this as it gives authors who have done research in each area at each school.

Question: What should I put in my letter of intent?

Answer: First, make sure you answer everything they ask, in the format they ask, etc. Next, make your letters sincere. If you don’t have a good reason why you’re applying to a certain school, why are you applying there? Let schools know why you are interested in that particular school. Finally, sell yourself. Why do you want to do a PhD? What have you done to prepare for it?

Question: What else can I do to better prepare myself?

Answer: More math and econ will definitely help. I did a minor in econ as part of my undergrad and I still feel like I’m barely staying afloat in graduate microeconomics. Next, get into the literature. It’s never too early to start reading. Pick up a copy of an accounting journal (e.g., The Accounting Review, Journal of Accounting Research, Journal of Accounting and Economics, Contemporary Accounting Research, etc) and start reading. Most universities have online copies.

If you have other questions you would like answered, feel free to post them here or e-mail me.

One paper I assign is the Watts and Zimmerman excuses paper.  In case you are unfamiliar with the paper, I view it as presenting a demand-supply model of normative accounting research.  Accounting theorists could offer new insights about “better” accounting practices ex ante, and then auditors and preparers would adopt those practices.  However, the paper dismisses this model and instead offers a model whereby an increase in regulation leads special interests groups to demand “theory” that would support the accounting that they prefer.  The paper discusses some historical evidence that suggests theory lags regulation and practice, which is more supportive of the latter model.  Thus the premise of the paper – accounting theory arises from a market for excuses.

Here are the questions I posed in the seminar – the excuses paper was written when a lot of the journals were publishing normative work.  Now, most top journals are publishing positive research, or should I say, positive research with a paragraph or two of implications that might sound a bit normative.

1) Would the excuses theory predict that top researchers would shift from normative to positive?  Was this a shift in demand (or supply) for excuses?  Or was this a paradigm shift so that excuses are no longer a first order determinant of our research?

2) If the authors were to write the paper today, would the title be – “The Demand for and Supply of Empirical Accounting Research: The Market for Excuses”?  Or are the findings from empirical research sufficiently positive that they cannot be labeled as excuses?

I am sorry to say that neither the students nor the instructor in my seminar had very good answers to these questions.  Can anyone help me out?

After reading the paper I was was curious how a researcher could adjust for clustering across three dimensions. For example, a study on legal regimes usually has three sets of clusters: year, firm, and country. I contacted Dan Taylor to find out more. He told me that firm clustering is subsumed within country clustering. In other words, as long as you assume that a firm doesn’t change countries during the course of your study, cluster-adjusting by country automatically fixes the problem of lack of independence of observations within a firm.

I kept digging a bit and found a paper that deals with this idea (Gow et al. 2010 cite this paper): “Robust Inference with Multi-Way Clustering” by A. Colin Cameron, Jonah B. Gelbach and Douglas L. Miller, (c) 2006, an NBER working paper. To cite from their paper:

For two-way or multi-way clustering that is nested, one simply clusters at the highest level of aggregation. For example, with individual-level data and clustering on both household and state one should cluster on state.

So, long story short, if you have geographic locations in your study you should cluster on location and year and ignore firm.

While the paper contains a number of issues that would lead to interesting debate within the FASRI community, this post focuses on how the authors manipulate the relevance of fair value information regarding a building.  This task is harder than it might sound.  I was not impressed by their first attempt – one group of subjects was told the building was to be sold in the short term, while the other group was told the building was a long term investment.  The problem is, as a passive investment, fair value is relevant in both cases for assessing current period stock price.

The authors’ second approach to manipulating relevance seems much more appropriate.  One group is told the building is being held as a passive investment, while the other is told the building will be used for production purposes.  During my year as FASB Research Fellow, I and others at the FASB observed that constituents tend to accept fair value accounting in the first case but tend to view it as not very relevant in the second.  Why?  My intuition is that as a passive investment, the cash flows generated for the company by the building will be determined by market forces, so a market price of the building is very relevant to assessing the value of the company.  But when used in production, the cash flows produced by an asset will depend on all of the potential firm-specific synergies that the asset brings to the company.  For example, while the market may value two buildings on different sides of a town (or a state) the same, the cash flows generated from using the building in production will be higher if the company acquires the building on its side of town (lower transportation costs).  My impression is that rather than trying to add up the fair value of each productive asset, most investors try to project the economic rents (income, cash flows, or other measures) generated by using the productive assets collectively, and apply some present value technique to the rents.

In this sense, the fair value of a single productive asset is not very relevant to assessing current stock price of the company because the market price for the single asset probably does not reflect all of the possible firm-specific synergies associated with that asset.  Kudos to Koonce and her co-authors for developing this second manipulation.  If any other experimental researchers are struggling to find a good manipulation for relevance, I urge you to contact Lisa.

Obviously there are a lot of factors that go into this decision, the first of which probably being what I enjoy most and the second being what I’m good at. However, I’ve thought another factor may be what the market wants right now.

So, I’d like your input, what seems to be the skill most lacking in accounting researchers? What skill set or area of expertise would you most want from a potential co-author?

Another way of phrasing this may be, what part of the research do you least enjoy or do you feel least qualified to do?

Using an experimental methodology, the authors randomly assigned participants to one of two groups who had to negotiate salary payments for three workers. Half of the participants were assigned to be a boss, essentially having the last say on how much workers were paid (a form of the dictatorship game). The other half of the participants were assigned to be a worker. Bosses were seated individually in a large office with windows and a big desk. Workers were seated in a small cubicle with no windows. After looking over the details of the job descriptions, etc., the worker had to walk into the boss’s office and discuss salaries with the boss, who was sitting behind the desk. After the discussion, the boss decided how much would be paid to all workers, including the participant-worker.

All of this first scenario set the stage for what was to happen next. Participants were told they were needed to participate in a second study. In this study, participants were led into a room and sat in front of a computer screen. They were told that within a minute, the computer screen would tell them either to take $100 inside a book on the desk or not to take that $100. People told to take the $100 needed to hide it somewhere out of sight. Participants were told that after five minutes, a person would enter the room who had no idea whether they had actually taken the $100. If participants could convince the person that they had not taken the $100, they could keep the $100, regardless of what the computer had told them to do. This meant that if participants followed the instructions about taking the $100, half of the participants would be telling the truth and half of the participants would be telling a lie. The incentive for all participants was to persuade the person that they had not taken the $100. If they did persuade the person, participants could keep the $100 and were entered into a random drawing for an additional $500.

The authors wanted to know whether people assigned to the POWER group in the first study would be better than those assigned to the WEAK group at deceiving the questionner in the second study. The authors hypothesized that being in a powerful setting makes it less difficult to lie because of the emotional and chemical changes in the body that power seems to cause. The authors measured physiological changes and asked many debriefing questions to get at the mental state of participants. Based on the deception attempts, coded body language indicative of lying (also called deceptive tells), and a saliva analysis of a hormone that often elevates after lying situations, the authors provide convincing data suggesting that power is likely to make it easier for a person to lie.

Of course this is way too brief of a summary, but I hope you get the idea (and want to read the paper yourself). After reading the paper, I wondered whether there were any testable hypotheses that would relate earnings restatements to the number of “deceptive tells” a CEO or CFO might display in a video earnings announcement. Or, is their a relationship between the number of syllables per second spoken by a CFO in relation to a pointed conference caller’s question and the likelihood of an earnings restatement. Or perhaps even better, does the market somehow pick up on these deceptive tells in conference calls and react? I also wonder about audit implications, but those sound to invasive to even imagine for right now. So, I’ll stop here. Take a look at this study and tell me what you think.