Price index - industry or commodity-based?

[WesIngwersen]

Adjusting variables between dollar years in the models requires annual price index data that are sector specific and that closely correspond with the valuation used in the model.
USEEIO and CEDA have both derived price indices used for adjustments from the same data series of detailed industry gross output data provided by BEA.
In CEDA they are used as is (after schema adjustment) to make price adjustments to the A matrices.
In USEEIO they are used as is (after schema adjustment) when applied to industry-based values, but they are transformed into a commodity-form before used with adjustments to final results.

[WesIngwersen]

To simplify this it's a matter of getting the units correct. Industries and commodities are intentionally separate entities in the Make/Supply Use model.

The price adjustments are ratios of price indices that are specific to industries or commodities. BEA's gross industry output chain-type price indices are for industries. The ratio developed for e.g. from those data is

$$ \frac{\Pi_{i,335110,2023}}{\Pi_{i,335110,2017}}$$

We cannot multiply that directly by a commodity-based value, or we get a unit error, for instance in this equation were we are inflating final demand from 2017 to 2023 for commodity 335110:

$$ y_{c,335110,2023} = y_{c,335110,2017} * \frac{\Pi_{i,335110,2023}}{\Pi_{i,335110,2017}}$$

To do this we need to estimate those price indices for commodities and derive the ratio based on commodity price indices.

$$ y_{c,335110,2023} = y_{c,335110,2017} * \frac{\Pi_{c,335110,2023}}{\Pi_{c,335110,2017}}$$