Abstract We introduce a novel approach to uncovering causation by exploiting transition episodes in aggregate macroeconomic data. Using a panel of 132 countries over 1961–2017, we identify twenty-six productivity-growth and forty savings transitions, episodes characterized by sudden, persistent, and unpredictable increases in growth and savings. Event-study evidence shows that productivity-growth transitions lead to sustained increases in savings, whereas savings transitions do not generate sustained growth effects. We conduct extensive robustness checks addressing staggered transitions, heterogeneous effects, and unobserved time-varying confounders. Having established the direction of causality, we estimate the effects of productivity-growth and productivity shocks (estimated separately for each country using a neoclassical model) on savings using panel Vector Auto-Regressions and local projections. We find that a one standard deviation increase in productivity growth raises the savings ratio by approximately 0.5 percentage points, highlighting the importance of policies that promote productivity growth to increase savings, rather than relying solely on savings-focused interventions.
Kumar et al. (Mon,) studied this question.