Anyway, had a lot of fun working on this. Very grateful to all who helped, and to the authors whose work I'm building off of. And to the BEA for having great data and a good API! My code and data are available on Github, linked here.(11/11)
github.com/jacobgosseli...

Lot of questions left. Note 2: (1) ES/SP govern how sectors respond to input price changes, so my estimates are relevant to anything that propagates through prices (distortions, price rigidities). (2) ES swings are big; why are sectors more/less sensitive to price changes in different years?(10/11)

SP shifts raise [lower] the GDP effect of shocks to sectors as they become more[less] central suppliers: e.g. between 1997 and 2023 GDP effect of “Paper products” shocks fell and GDP effect of “Computer and electronic products” shocks rose due to changes in their importance as input suppliers(9/11)

Temp. het. raises[lowers] the GDP effect of shocks in years where sectors are less[more] able to substitute: e.g. generating GDP distribution from realistic sector-level TFP shocks, diff btwn using 1st and 3rd qtile of sector-level ES across years is same as diff btwn 3rd qtile and CD(8/11)

I quantify the effect of heterogeneity of ES and year-to-year changes of SP in a GE model of multi-sector production. Find that sec. het. of ES raises[lowers] effect of shocks to sectors whose customers are less[more] able to substitute: e.g. "chemical products" shocks 10% worse.(7/11)

Having SP shifts lets me test ext. validity. Sum abs. value of shifts by industry-year: magnitude of changes in sectoral production which shifted relative demand for inputs. Show that this measure is dynamically correlated with weighted patents; reassuring that I'm not just picking up noise.(6/11)