jrissman
commented
3 years ago I think it would be a good idea to break out construction as an industry as part of #89. This should address the issues you raised because policy effects applied to the "other industries" category would not be applying to the construction industry anymore - the policy would have to be intentionally applied to the "construction" industry. We wouldn't do that where it doesn't make sense - e.g. we would not have the construction industry utilize CCS themselves. (The construction industry already, in 3.0.0, gets revenue for installing CCS systems bought by others, so we don't have to do anything special there.)
Yes, I think construction would probably also need to be removed from the "Other Industries Category ISIC Code" sub-range. The other industries we break out as part of #89 might need this also. I'll resolve these technical implementation details as part of work on #89.
robbieorvis
We currently group all non-specifically broken out industries in the EPS into the "other industries" category, including when trying to allocate costs and changes to expenditures and revenue. As an example, when adding CCS in industry, we might see that the other industries category incurs very high costs to install and operate the equipment. When we then allocate costs to ISIC codes, we are assigning the total change in spending in the other industries category based on the BAU output of that group of industries. This makes sense for most industries in this bucket but not for construction.
Construction output is tied to spending in other sectors beyond the industry sector, and proportionally assigning other industry revenue and expense changes to construction based on its share of BAU output among other industries is not accurate. More importantly, when other industry expenses change, it is incorrect to group construction as part of this, as construction revenues would often be expected to increase when industry sector expenses increase to comply with a policy. For example, in the CCS scenario, we would expect many construction jobs to build the CCS equipment (which the model calculates correctly) but we wouldn't expect the construction industry itself to incur costs to comply with a CCS requirement (which the model is calculating). With the current structure, we see a large decrease in construction jobs because the net impact on the construction sector after proportionally assigning non-industry sector costs results in a net negative cash flow (testing this in the VA model).
The cleanest way (I believe) to do this is to remove construction from the other industries category in the ISIC code grouping and ultimately in the sector grouping as well. We have data from EIA on construction energy use so it may be the case that making this change ultimately pushes us to make construction its own industry (along with others, as previously discussed).