“…For example, Bennett et al (2020) exploited data in the national accounts of the United States to show that some important types of basic infrastructure have barely, or have not, kept pace with depreciation and population growth in recent decades. Many studies have built the analysis of infrastructure using GFCF data sourced from countries' national accounts, constructing measures of infrastructure investment using different GFCF sub-components, and therefore benefitting, when available, from GFCF data cross-classified by detailed asset type, and/or industry and institutional sector (Revoltella et al 2016; Asian Development Bank, 2017;Bennett et al 2020). As such, a lack of detail in GFCF sub-components by type of asset (such as residential and non-residential buildings, civil engineering works, and machinery and equipment) and the institutional sector/industry in which the asset is used makes it difficult to readily distinguish, and therefore evaluate, investments in infrastructure.…”