Does dividend policy lead the economy?

I investigate the predictive role of the aggregate dividend–payout ratio (de) for future economic activity. A vector-autoregression-based variance de-composition shows that the main driving force of de is long-run predictability of earnings growth, with dividend growth predictability assuming a secondary role. Consistent with this result, long-horizon regressions indicate
that de is a significant predictor, especially at intermediate and long forecasting horizons, of future aggregate business conditions. Critically, de outperforms several popular equity and bond predictors from the literature. The predictive ability of de remains robust in an out-of-sample forecasting analysis. Overall, de conveys important information about the economy.