Per Olsson, Professor for Accounting and director of the Center for Financial Reporting and Auditing at ESMT Berlin, is a principal investigator in project B05 “Transparency and the Equity Market”. In his project he examines how financial reporting transparency affects the information asymmetry on equity markets.
Accounting: getting into the nitty gritty details
I can still vividly remember how my fascination for empirical accounting research started. There was this one seminar that completely changed my life: a PhD seminar by Katherine Schipper at the University of Chicago. At that time, I was a visiting PhD student, having studied managerial economics in Sweden where I was born and raised. The way Katherine Schipper conveyed the contents of the course and the seriousness of her teaching made a positive and lasting impression. As did the topic itself. The workload was high though. We went through six to ten hours of preparation per paper. But all for a good, and even fun, cause: Getting into the nitty gritty little details of accounting and dealing with the big thoughts of the researchers of this field was great – and it led me to where I am today.
Based on my own experience, I always advise my students to have the courage to go to one of those really large places.
Research: crossing borders
My scientific career has been significantly shaped by my research stay in the US. It was crucial for my professional development and my self-concept as a researcher. When I was a doctoral student, empirical accounting research in the United States was conducted with greater seriousness than in many European countries. Now this has changed of course, applied empirical research on accounting has been recognized on a global level. Based on my own experience, I always advise my students to have the courage to go to one of those really large places. To go to those institutes where there is a lot of serious research going on. You will have a much easier time afterwards, even if you only spend a semester or a year there. Especially the network I was able to build – and consolidate during the 20 years I spent there – helped me advance my research decisively.
Working with different and young researchers equals a healthy injection of new ideas and perspectives.
TRR 266: new networks, new perspectives
Having a good, extensive network is very important for research as a whole. This is also a great advantage of the TRR 266. Our Collaborative Research Centre certainly offers new and exciting networks. When you reach my age, you could compare it to a vitamin shot. Working with different and young researchers equals a healthy injection of new ideas and perspectives. It positively affects your own research.
Choosing the right method
Good empirical research has to be conducted with rigor. I deem choosing the right method and a fitting research design crucial. I attach special importance to this in my research. Unfortunately, we can sometimes observe false null findings in science that can simply be attributed to a non-optimal and low-powered research design. This should of course be prevented if possible. It’s such a shame if you don’t find anything even though it is there.
At the same time, it is important to me to pass this knowledge on to my students. It is very satisfying to observe the moment when young people begin to understand and internalize what you teach them and make a success out of themselves. Especially students who initially came to my courses with rather low expectations and then realize how fascinating accounting can be. I am still in regular contact with some of my former students. It’s really nice to follow their careers and their development.
We provide empirical evidence that accounting information is useful for equity investors, especially from a risk assessment perspective.
Empirical evidence in current discussions
Within the TRR 266 I’m affiliated with the project B05. Together with Sönke Sievers, I am currently investigating the role accounting information can play in assessing risks for equity investors. As accounting information reflects the past, many people assume that it is not really suitable for modeling risks. But our research shows that the opposite can be true. If accounting information is understood and modeled correctly, it can actually help to assess risks. So, our B05 research can provide empirical evidence in current discussions. This is one of our primary tasks as researchers.
For example, we investigate the co-movement of stock prices and intrinsic value estimates focusing on the estimation of risk. We apply risk measurements based on accounting and market data, respectively. We find that when using an accounting-based risk measurement, in contrast to the market-based risk measurement, price and value co-move on an index-level. Without going into details, this result provides empirical evidence that accounting information is useful for equity investors, especially from a risk assessment perspective. (see our Working Paper No. 45)
This is another example of how research can inform on important issues with bearing on practice and policy – even if it deals with extreme methodological issues.
Research with bearing on practice and policy
In a recent , just accepted publication in the Review of Accounting Studies (jointly with some of my colleagues from Duke University and Frankfurt School), we deal with a pervasive issue in accounting research. We investigate how data requirements, often encountered in archival accounting research, can produce a data-restricted, non-random sample. A sample that is a non-random selection of observations from the reference sample to which the researcher wishes to generalize results. In plain English: when we require data on for example analyst-following, positive earnings or a time series to calculate or estimate some measure, we often end up with large, stable, well-analyzed firms.
This means that there is relatively little variation among firms, making it hard to detect effects regarding, for example, firm risk – simply because the riskier firms (with losses, relatively young, etc.) are not included in the estimation sample. We develop and validate a resampling approach that uses only observations from the data-restricted estimation sample to construct distribution-matched samples that approximate randomly-drawn samples from the reference sample (the sample of all firms). This is another example of how research can inform on important issues with bearing on practice and policy – even if it deals with extreme methodological issues.