How Disinflation Crosses a Border: Global Value Chains and Intermediate Inputs

The international question is now specific: when zombie firms keep low-productivity capacity running in emerging Asia, how does that distortion reach an advanced economy? This chapter focuses on the trade route rather than assuming that every cross-border effect must appear in bank balance sheets. It explains how exposure is measured, why output can respond before consumer prices, and why intermediate goods provide a more convincing bridge from Asian excess capacity to importer disinflation.

Turning trade exposure into a measurable shock

An advanced economy is not equally exposed to every Asian economy. The relevant exposure depends on how much it imports from each source and how prevalent zombie firms are in that source. The BIS study captures this with a GVC zombie-spillover index. For an advanced economy, the index weights the prior year’s asset-weighted zombie share in each of the ten Asian emerging economies by the source’s share of the advanced economy’s imports or trade.

The weights come from average bilateral trade in 2005–2007, before the estimation period. Fixing the weights matters because it reduces the risk that current economic conditions simultaneously change the trade pattern and the measured outcome. The zombie share is lagged one year because firms need time to produce and export goods; the lag also makes the exposure predetermined relative to the contemporaneous importer outcome. Annual zombie shares are interpolated to quarterly frequency for the panel. The resulting index is standardized, so a coefficient describes the change associated with a one-standard-deviation increase in exposure rather than with a one-percentage-point change in a country’s zombie share.

That construction is best understood as a portfolio exposure, not a shipment ledger. It does not identify which individual firm sold which component to which importer. It says that an advanced economy with a larger established trade connection to an Asian source is more exposed when that source’s measured zombie share rises. The fixed weights also mean that the index holds the pre-sample trade relationship constant while the zombie prevalence changes. This is useful for interpretation, but it cannot capture every later change in supply-chain geography.

The baseline panel covers 29 advanced economies from 2008Q3 through 2021Q4. A one-standard-deviation increase in trade-weighted exposure is associated with about 0.135 percentage points lower CPI inflation and 0.405 percentage points lower real GDP growth. Using import weights rather than total-trade weights yields coefficients of about −0.089 for CPI inflation and −0.424 for GDP growth. The consistency of the output estimate across the two weight definitions matters more than a small difference between the inflation coefficients: the result is not simply a product of one arbitrary way of measuring bilateral exposure.

The dynamic pattern adds information about the route. Local projections estimate the response at horizons from zero to eight quarters. GDP growth falls more quickly, with an impact response of about −0.40 percentage points and a trough near −0.97 percentage points around three quarters. The CPI response builds more gradually, reaching about −0.75 percentage points around five quarters before partially reverting. A plausible reading is that imported goods can displace or delay domestic production relatively quickly, while lower input costs pass through to downstream prices as firms adjust contracts, inventories, and pricing decisions.