International trade and the impact of globalization continues to be a hotly debated topic. See Brad DeLong’s thoughtful piece at Vox and Jared Bernstein’s rejoinder, for recent examples. They both highlight the fact that trade policy and individual trade deals are distinct and have different impacts, even as nuance gets lost in conversation.
While much of the current discussion tends to focus on a particular trade deal like NAFTA or trade with a particular country like China, in a world of global, just-in-time supply chains, it gets tricky and complicated really fast. Bill Conerly, economist and member of the Governor’s Council of Economic Advisors, made this point in a recent article. Bill noted that all major manufacturing subsectors rely upon imported parts and materials for some portion of their production, even if the bulk of that production was here in the U.S. It may be an appliance factory that imports a particular valve, or a paper plant that imports a particular machine used in the paper making process. As such it’s not just about direct trade or trade balances necessarily. In fact, the vast majority of states have very little direct trade exposure to any given country or region of the world. The real issue is about the global supply chain and any potential disruptions.
What follows is a high level look at imported supply chains, state industrial structures and then direct trade exposure by state to Latin America and China. Unfortunately the data used here from the BEA and Census is imperfect, as noted on their respective websites, but as good as we have available to us. The findings are suggestive of these trends and potential impact, even if specifics are sure to vary should any disruptions come to pass.
First, approximately 20% of the intermediate goods and commodities used in U.S. manufacturing are imported. That’s a sizable share overall but it does varies considerably by subsector. These figures are calculated using the BEA’s IO tables, their import matrix and use tables specifically. Overall the subsector patterns make intuitive sense. Food, Beverage, Tobacco and Wood Product manufacturing use a lot of domestic inputs (U.S. grown crops). Machinery, Transportation Equipment (cars, heavy trucks, airplanes), Computers and Electronics rely on imported iron, steel, wires and the like. The one item that really stands out, however, is Petroleum and Coal manufacturing. While the U.S. does import a lot of petroleum, I’m not entirely sure this import share perfectly applies here and then not by state either. This is where imperfect data comes in. The BEA compiles domestic vs imported sourcing for all types of commodities and intermediates. They then assume that each sector that uses a given commodity uses the same domestic vs import share. That said, the available data does confirm Bill’s argument of the reliance on imported inputs across manufacturing sectors.
Second, the industrial structure of regional economies can help us assess potential supply chain risk across the country. Not that these are the states that will be most impacted, just that they have a larger concentration of manufacturing in the sectors with higher import shares. This too relies on imperfect assumptions but is the best we can do; the results are suggestive of the general patterns seen in the data. As such, the energy states rise toward the top given their industrial structure. Locally, even as Oregon has a large high-tech manufacturing sector, the state has a lower implied share due to the size of Food and Beverage, and Wood Products.
Third, the work above speaks to global supply chains in general across industries and states. However much of the recent discussions have centered on trade negotiations, or renegotiations with specific countries. To what extent do regional economies trade or rely upon trade?
The first scatter plot shows state level trading patterns with Latin America overall. Mexico is the dominant trading partner here, with surprisingly little trade with the rest of the Caribbean, Central American and South American countries. To normalize the data across states, I show exports and imports as a share of state GDP. Note that the state import data is still relatively new in the data world and is known to be less complete and robust than the export data. That said, most states have relatively low levels of direct trade exposure to Mexico and Latin America more broadly. Clearly this is not the case for states like Michigan (auto and related imports) or Texas and Louisiana (large imports and exports).
The second scatter plot shows state level trading patterns with China. Again, there appears to be relatively small levels of direct trade exposure for most states. That said a few states stand out. On the export side, both Washington (aerospace) and Oregon (semiconductors) trade considerably with China. In Oregon’s case, however, we know some of that is within firm shipments and not necessarily selling products on the open market. To what extent within firm shipments, or vertically integrated companies would be impacted by worsening trade relations is an important question to ask. I don’t have the answer to that but it’s worth considering.
On the import side, California and Tennessee stand out. In both cases, a sizable share of these imports are either computer equipment or communications equipment (both in NAICS 334). For California it’s about 33% of all Chinese imports and for Tennessee it’s about 60%. One question I have would be does this represent direct trade or the fact that Tennessee and California act as distribution hubs. Are some of these goods just sent to warehouses from which they are then sold to customers all over the U.S.? This question brings us back to the start of the post.
Global supply chains mean that every major manufacturing sector relies in part on imported commodities, intermediates and the like. Assessing economic benefits, or costs, is considerably complicated. As the state trading patterns show, not that many states have large levels of direct exposure, however we know it’s not that simple. A disrupted supply chain, should it occur, will have knock-on effects throughout the economy. This is particularly the case during the initial adjustment phase where supply chains may need to be reworked through finding new suppliers, changing cost structure due to taxes or regulations and the like.