jrissman
commented
3 years ago I think this is a great idea, but it's a little trickier than it might seem at first glance, because the best way to divide the policy cost/savings up varies by sector, and by cash flow type within the Industry sector.
In the Transportation sector, the policy costs/savings are based on which industries own/operate the most vehicles. (The costs/savings are on buying vehicles, buying fuel for vehicles, OM for vehicles, and buying EV chargers.) Since transportation fuel for trucks isn't sold by the utilities sector, I don't think the utilities sector spending is a good proxy for how much each industry uses vehicles. I don't think we can use spending on products from the refined petroleum fuels industry either, because a lot of industries buy petroleum products for purposes other than powering vehicles, such as for use as a chemical feedstock, or to power industrial machinery. So if we're looking for a proxy for the amount of vehicles owned/used, output is probably the best we've got, since it should be somewhat correlated with the total volume of input materials and products being moved around.
In the Electricity sector, the only time we allocate any costs among industries based on output is demand response service payments, and I think utility spending is superior to output for allocating these.
In the Buildings sector, the costs relate to buildings, and I agree utility spending is superior to Output for allocating these, since the changes that happen in the buildings sector relate to energy use (and even if the change isn't directly energy-related, like improving insulation, it is more likely to be undertaken by firms with higher utility spending, since they'd see greater energy cost savings).
In the Industry sector, we assign most costs and savings to specific ISIC codes using other input data tables (such as a table for who gets the revenue from process emissions abatement expenses), and we only use output-based allocation for certain specific things. We still use it for allocating (to ISIC codes within the "other industries" category):
- Revenue from carbon tax rebates due to industries engaging in CCS
- Revenue from changes in demand for the industries' products
- Spending on fuels
- Spending on labor or machinery directly caused by efficiency, fuel switching, or process emissions policies (not material efficiency or cost-driven production changes)
For (1) and (3), I think utility spending is decent for the U.S., since the utilities include natural gas distribution systems, and industries mostly use NG and electricity. But it's not perfect for the U.S., because some industries use coal and/or petroleum products. And it would be a pretty poor methodology for China, where industries predominantly use coal. We can probably handle this by using each ISIC code's share of spending on the sum of the products of the relevant energy-supplying industries.
For (2), I think output-based allocation is the closest fit. These industries are losing/gaining money based on changes in demand for their products, which is intermediated by the IO tables, and has little to do with how much energy those industries themselves use to produce the products.
For (4), it can thankfully be handled like 1 and 3, because spending on process emissions policies is entirely done by specifically-broken-out industries, not industries in the "other industries" category. So that leave us with the changes caused by energy efficiency or fuel switching policies, which is just like 1 and 3.
So my plan at this point is to:
- in the Transportation sector, leave output-based allocation in place
- In the Electricity sector, switch to utility spend allocation
- In the Buildings sector, switch to utility spend allocation
- In the Industry sector, leave output-based allocation in place for cash flow type (2)
- In the Industry sector, allocate cash flow types (1) (3) and (4) based on the share of spending on the sum of the: utilities, coal, and refined petroleum products sectors, rather than based on utility sector spending alone.
robbieorvis
From Robbie on 10/29/20:
One challenge I’m running into is the way in which we divvy up policy costs to the other industries sector in assigning them to them the ISIC codes that make up that sector.
For example, food, tobacco, and beverage is a big industry in Virginia, but as a source of emissions/energy use in Virginia I suspect it is quite small. Nationally, another example is paper in pulp, but in the opposite direction.
Currently, when we assign policy costs the other industries ISIC codes, we do so proportionally based on BAU output, but I’m wondering if we shouldn’t instead do this based on BAU amount spent on the utilities sector (ISIC 35T39) from each ISIC code. This isn’t perfect, but it does a better job capturing how energy/emissions policies might affect certain ISIC codes more than others, relating this to the amount they spend on utilities instead of what they produce. This would likely better represent how energy and climate policies might affect the specific other industries. This would fix an issue where a very low energy intensity industry might end up with a very high economic impact from a policy, even though in reality it uses far less energy than other emitting industries.
For example:
I believe you calculate this all internally in Vensim by multiplying the BAU output variable by the DLIM value from the utilities sector to each sector, then dividing by the total spend on the utilities sector to develop a fraction of other industries costs that should be assigned to each ISIC code.
What do you think?