Fear thy neighbor: Spillovers from economic policy uncertainty

Economic policy uncertainty rose sharply after 2008 – US debt-ceiling standoffs and threatened trade-agreement annulments, Brexit negotiations and European elections, China’s leadership transition and currency adjustments – and this paper asks whose uncertainty actually hurts whom. It begins with a conceptual point that shapes how its results should be read: higher policy uncertainty “is not necessarily bad news,” since a wider distribution of future shocks “includes shocks that would lead to better outcomes,” but the existing evidence is that agents dislike it and respond by re-weighting probabilities toward unfavorable events, “in essence confounding certain for good and uncertain for bad and taking a ‘wait-and-see’ approach.” Two estimators are then run in parallel on the newspaper-based Baker-Bloom-Davis index, available monthly for 19 advanced and emerging economies and aggregated to quarterly. The first is Pedroni’s (2013) heterogeneous structural panel VAR, whose distinctive feature is that it splits each shock into a common component – uncertainty originating anywhere else in the sample and spilling in – and a country-specific one. At the median, a ten-unit rise in the index is associated with a 0.1 percentage point fall in real GDP growth one quarter later, a 0.1 point fall in private consumption growth after two quarters, and a 0.3 point fall in private investment growth after two quarters, with the output response statistically significant for up to three years. The decomposition is the paper’s headline: the common shock “explains roughly two-thirds of the decline in GDP growth across all horizons,” and dominates for consumption and investment as well – so most of the damage policy uncertainty does to a country comes from abroad. The second estimator, Jorda (2005) local projections estimated separately for shocks originating in the United States, Europe and China (and excluding the originating countries from the response sample), locates and sizes those spillovers. A ten-unit rise in US policy uncertainty is associated with roughly a 0.2 percentage point decline in real GDP growth across 64 other economies three quarters later and about 0.6 points for private investment; scaled to the 53-unit jump in the US index between 2008Q2 and 2008Q3, that implies a 1.1 percentage point average decline in other economies’ GDP growth. European and Chinese shocks are similar to each other and smaller than US shocks, with effects lasting about two years. Regionally, the spillovers are significant for Asia and Pacific, Europe and the Western Hemisphere, and largest for the latter two – which the authors read as consistent with transmission between more closely integrated economies. Throughout, the language is associational rather than causal, the samples for consumption and especially investment are much smaller than for output (15 and 8 countries in the panel VAR), and the identification rests on a recursive ordering plus the exclusion of the shock’s origin country, both of which the paper tests rather than asserts.

Francesco Grigoli | Macro Paper Warehouse

Fear thy neighbor: Spillovers from economic policy uncertainty