The purpose of this paper is to locate the specific items from the financial statements that are responsible for the dirty surplus accounting flows and how important they are in its explanation.
It is generally accepted that some country accounting rules allow some operations that can generate dirty surplus in the annual statements. Working on this basis, it is necessary to consider information at the same time across firms and across time, using panel data econometric techniques. A static panel data estimated by generalized least squares can be used to correct correlations between firms and account numbers or a dynamic panel data estimated by GMM‐SYS with instrumental variables to avoid endogeneity.
Results show that in a static panel data model, the income statement items have a lower explicative power of balance sheet items variations, having higher explicative power a dynamic one (AR(1)). Results show that, specifically, financial assets, debts and book value capture the dirty accounting flows.
Working in differences reduces the explicative power of the income statement and working in levels could be inconsistent if it is impossible to contrast, first, stationary in data due to their shortage. It is suggested that future works increase the frequency of the observed data, and contrast the cointegration as a way to check the accounting relationships.
It is important to evaluate whether the income statement can (or cannot) explain the financial position of a firm. Also it is important to know where dirty surplus accounting flows are located can be useful for firms' valuation.
The econometric technique proposed in the paper deals with the main limitation in accounting research: information is bigger in cross‐section (number of firms) than in time series (economic periods).
