ABSTRACT
Recent studies in macro accounting suggest that changes in aggregate accounting earnings can predict quarterly GDP growth in the USA. Our objective in this study is to test the robustness of the association between aggregate earnings and real GDP growth across multiple countries and for different definitions of aggregate earnings. We test whether aggregate earnings changes predict future economic growth in eight countries – Australia, Canada, China, India, Japan, South Korea, the UK, and the USA. We find positive evidence regarding the generalisability of the aggregate earnings – real GDP relation in an international context. In additional tests, we find that economic forecasters appear to underreact to aggregate earnings information. Our results show that aggregate earnings lead economic growth and forecasts of real GDP growth can be improved by incorporating aggregate earnings information. Further, we also test if negative earnings changes contain more information than positive earnings changes and find only modest evidence in favour of this hypothesis. The results are robust to alternative statistical methodology and the removal of the USA data from the sample.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes
3 CLI is a composite leading indicator that includes multiple components. These components vary according to economic characteristic of each country. For example, India’s CLI consists of Industrial production of durable consumer goods, non-metallic minerals, car sales, Monetary aggregate M1, Share prices and call money rate. The OECD report on CLI components has no mention of explicit inclusion of corporate earnings. Therefore, we have sufficient reason to believe that aggregate earnings are orthogonal to the CLI series.
4 Fiscal year end for Australia in 30-June of every year. For India and Japan, this figure is 31-March every year. Remaining countries follow the calendar year as their fiscal years. This approach is similar to Konchitchki and Patatoukas (Citation2014b) who take only the firms with December fiscal year ends for the USA. For any country, we have chosen the month end when a majority of the firms report their annual earnings as their fiscal year.
5 We discuss the results from the PCSE estimation later in the robustness tests section but would still like to mention that the inferences are essentially identical which suggests that our results are quite robust to a change in the methodology.
6 In unreported results, we run regressions of future GDP growth on aggregate earnings and lagged GDP growth only and find that lagged GDP growth is not a very significant predictor of current GDP growth and the exclusion of the CLI leads to higher significance of the aggregate earnings coefficients.
7 This notation is in line with IMF’s usage. They release two World Economic Outlook reports every year in spring and fall.