This is my first post here, and I hope to find an answer to my question.

I have found a paper that examines the relationship between earnings management (absolute value of the abnormal working capital accrual) and IFRS.
I have noticed that the author divided the earnings management by the beginning total assets. Further, the control variables include leverage which is the total debt divided by the current year's total assets, CFO which is the cash flow from operating activities divided by the beginning total assets, and ROA which is the operating profit divided by the beginning total assets. Firm size is the log of total assets for the current year.
For other papers, I have found that most of the past paper didn’t show a consistent approach, I mean some of them divided the earnings management and CFO on the current year of total assets, while divided the ROA on the previous year of total assets and used the log of total assets of the previous year.

So, I am wondering why, I mean is there any statistical issue behind that?
Can I scale the variables on the beginning and current total assets randomly in a way that serves the results, or do you think there is a benchmark for this?


Scaling should be "reasonable" and perhaps "logical."

The issue here is that the firms being compared vary widely in size. Hence, the data is scaled.

Prior year or current assets would be reasonable to scale earnings, debt leverage, and CFO. Assets are always changing, so assuming they don't change too fast, either is "reasonable." It would likely be better to use the average of prior and current year, but the gain to be had is probably not worth the complexity.

However, what if assets do change quickly? If a firm made a huge acquisition, then its assets would suddenly jump much higher. How to handle that? The average mentioned above would be one way. Or, use the ending assets to scale! So if the firms being analyzed do a lot of mergers and acquisitions, then scaling by ending assets would be reasonable, as it prevents misleading jumps in the scaled data.

ROA I would scale by prior year assets, as it is the growth of the assets themselves. Scale by the starting point, not the ending point. If there was an acquisition, used the combined assets from both firms for the prior year.

  • $\begingroup$ Thanks very much for the answer. Much appreciated. but, I think it is really hard to decide if the assets for all firms is jumping too fast or not, do you mean to observe that by my eyes?? $\endgroup$ – user323133 May 29 at 17:58

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.