NEC (not elsewhere categorized) items in t-FBS

[realxinzhao]

This is a question raised by Willem-Jan (@wjvanzeist) on NEC items.
Adding some details here as more work on this will be needed.

Indeed, NEC is the place we should further explore. Some of them are more processed than others, and they are not really small in calories. Globally, about 5%, but about 14% (552/3921 kcal per ca per day in 2022) in the US (more developed regions have a higher portion of processed food/beverage consumption). Currently, we simply aggregate them into NEC, and we can easily separate them as individual items in the mapping (setting tiers as 0 and source item to the same as sink item).

I have quickly pulled some numbers from the data. About 44% of the US NEC (552 kcal per ca per day) is “Margarine, short,” and most NEC items are kind of ultra-processed. I guess Margarine can be mapped to vegetable oil (or other fats) processing in PCE if we have good data. Similarly for the sugar products, we could probably do more if we had better data. For alcohol and beverage or other cases with multiple processing (e.g., the share of wheat that was used for processing into beverages; we could likely use the “output-specific-extraction-rate” to implement them in the existing framework), we will need to know how much source item was processed into them. But you are certainly right; we didn’t have any information regarding the processing (relationship and extraction rate), so we left them out.