Sean Callahan

How far do fiscal shocks travel through a production network? I answer this question for the euro area by applying the network granular instrumental variable (N-GIV) approach of Chodorow-Reich et al. (2026) to FIGARO inter-country input-output tables covering 1,178 country-sectors over the period 2010–2019. The method identifies two reduced-form propagation parameters: the share of a shock transmitted upstream from customers to suppliers (the demand channel) and the share transmitted downstream from suppliers to customers (the supply channel). I develop a static production network model with monopolistic competition and constant returns, that predicts an asymmetric response: quantities respond to demand shocks only through the demand channel, while prices respond to cost shocks only through the supply channel. I then test the model empirically, confirming these results. Each link in the network passes on 41 percent of a demand shock to employment and 32 percent to gross output, while 40 percent of a cost shock passes through to relative prices; the cross-channel parameters are statistically indistinguishable from zero. The implied propagation multipliers show that 17 percent of the employment response and 12 percent of the gross output response to government spending arise from network transmission beyond the sector where the spending lands first. Intuitively, sectors that are further away through the network see almost no effect from the shock, implying that the sectoral composition of fiscal policy matters for its aggregate effects.