Technology Complexity and the Information Frontier
Year: 2026
Type: preprint
Abstract: Productive opportunity often depends on combinations of capabilitiesrather than on the abundance of individual capabilities. This creates anaggregation and identification problem: when only low-order aggregateinformation about the joint distribution of capabilities is observed, whenis economy-wide productive opportunity identified? We develop a for-mal framework that separates the technological complexity of individualproductive completions from the informational complexity of their ag-gregate opportunity. Representing productive technologies as a family ofcapability completions, we define aggregate productive opportunity as thepopulation prevalence of configurations satisfying at least one productivecompletion and characterize the minimum order of aggregate joint infor-mation required to identify it. Our main theorem shows that this infor-mation order equals the degree of the associated multilinear opportunityfunction. Consequently, the order of an individual technology need notdetermine the order of information required for aggregation.We establish a sharp separation result. For every even number of ca-pabilities n, there exists a technology portfolio in which every individualproductive technology requires only two capabilities, while exact iden-tification of aggregate productive opportunity requires n-way aggregateinformation:τ (Hn) = 2, ρ(Hn) = n.We further characterize the resulting partial-identification problem. In thecanonical pairwise economy, first-order aggregate information can leaveproductive opportunity maximally unidentified, with identification width∆1(Hn) = 1.These results distinguish technological complexity from information com-plexity. More precise measurement of low-order aggregates does not re-solve an identification problem generated by missing higher-order jointinformation. The framework therefore provides a criterion for assessingwhether an aggregate capability statistic contains the information requiredto identify the productive opportunity it is meant to measure
Author: Kim, HyeongWoo
Institution: Hanyang University
Language: English
Subfield: Economics and Econometrics
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7204655273