Konchitchki and Patatoukas Journal of Accounting and Economics 57 (1-2), 76–88, (2014a) show that aggregate accounting earnings growth predicts future nominal gross domestic product (GDP) growth and that professional macro forecasters do not fully incorporate the information contained in aggregate accounting earnings. Based on results from prior literature, which find that accounting earnings reflect bad economic news in a timelier manner than good news, we condition Konchitchki and Patatoukas’s GDP growth forecast model on the sign of earnings changes. We show that negative changes in aggregate earnings predict future GDP growth while positive changes in earnings do not. Furthermore, we show that professional macro forecasters underreact to the information contained in negative changes in aggregate earnings about future GDP growth. Additional tests suggest our findings are a result of conservative accruals in earnings.

We examine whether the change in earnings announcement textual tone, aggregated across individual publicly traded firms, helps predict gross domestic product (GDP) growth. The literature finds that changes in aggregate accounting earnings do help predict GDP growth, but only when aggregate earnings changes are negative. Because conservative accounting rules limit managers' ability to communicate positive news promptly, we examine the tone of quarterly corporate earnings announcements as a possible source of timely positive information provided by firms. We find that the change in aggregate tone in the earnings announcements from the same quarter in the previous year predicts one‐quarter‐ahead GDP growth, but only when the change is positive. Our study contributes to the literature by investigating the relation between aggregate corporate disclosure tone and macroeconomic outcomes.