A matrix adjustment for more recent data year

[WesIngwersen]

The USEEIO and CEDA models still rely on benchmark Make and Use tables that reflect an older data year. From the benchmark Make and Use, following schema adjustments, the direct requirements matrix, A, is derived. The 2017 benchmark tables are the most recent published tables and still used by both
The question is how is the A matrix best included in a model that is intended to be used to represent a future year like 2024.

Here is what the models have done in the past:

1. USEEIO has used the A matrix as derived from the schema-adjusted Make and Use tables as is.
2. CEDA used the summary tables from the target year to create a A matrix for the target year (e.g. 2023). The summary A values were mapped to the corresponding detailed level A values (one-to-many mapping). Ratios were created of summary-derived 2023 A matrix values to the correspond detailed benchmark 2017 A matrix values. These ratios (in 2023 USD/2017 USD) were multiplied by the original A matrix values to both "scale and inflate" them to the 2023 target year.