Does accounting details play an allocative role in predicting macroeconomic indicators? Evidence of Bayesian and classical econometrics in Iran

Abstract

Purpose The purpose of this study is to analyze the predictability of firm level data for determining macroeconomic indicators such as unemployment. Design/methodology/approach This study uses quarterly GDP and unemployment data manually collected from the Statistical Center of Iran (SCI). Accounting numbers are also collected from the Tehran Stock Exchange library for the 2004-2015 period. Dispersion of earnings growth provides related data about labour reallocation, unemployment change and finally aggregate output. To summarize, this study attempts to examine the effect of these variables using classical and Bayesian approaches. Findings At a firm level, our results suggest that sectoral shift in previous years is likely to increase labour reallocation in subsequent years. At the macro level, the results reveal that dispersion of earnings growth and labour reallocation has a negative and positive impact on unemployment changes, respectively. However, the study suggests no significant relationship between stock return and unemployment changes. Consequently, we determine that the real estimates of macroeconomic indicators have predictive power because nominal estimates are not statistically associated with firm-level details. Finally, the results obtained from classical and Bayesian approaches suggest similar findings, thus confirming the robustness of our conclusions. Note that, based on Bayesian approach, the nominal reallocation has predictive power in unemployment rate. Originality/value The study is the first conducted in a developing country and the results provide important insight into current line of accounting literature.