Earlier this month the Trump administration announced hefty 25% tariffs on Canadian and Mexican imports, along with an additional 10% tariffs on Chinese imports. Since then the Canadian and Mexican tariffs have been temporarily (?) paused, though the Chinese tariffs are still in place, and Trump has since announced additional tariffs on steel and aluminum.
These tariffs are consistent with Trump’s campaign promises, though it’s somewhat unclear if the administration favors tariffs as a tool of economic policy (to make domestic industries more competitive) or if they’re more of a negotiating tool with other countries. Regardless, the tariffs on Canada and Mexico were widely criticized on both the right and the left (the Wall Street Journal editorial board described it as “the dumbest trade war in history”).
Since trade and tariffs are likely to continue to be a focus for this administration, it’s worth understanding how dependent the building and construction industries are on imports.
To look at this, I used USA Trade Online, a service of the US Census which lists all US imports and exports by country and by category. I put together a list of every NAICS1 code for products used by the building and construction industries, and what the value of imports were in each category.
NAICS codes are hierarchical. At the top are 2-digit codes that describe high-level industries (such as agriculture, mining, or manufacturing), below that are 3-digit codes that are slightly more granular, below that are 4-digit codes that are more granular still, all the way down to 6-digit codes. Here, for instance, is part of the NAICS hierarchy for Iron and Steel products:
For this analysis I mostly used 5-digit NAICS codes, except for a few places where 5-digit codes weren’t granular enough and I used 6-digit ones, and one case where I used a 4-digit code.2 Overall I came up with 55 major categories of imports used by building and construction, which I’ve grouped into six major clusters: steel and metal, wood, other materials, building components, interior fit-out, and tools.
Steel and metal — This includes relatively “raw” steel slabs and ingots from the mill (code 33111), as well as steel that has been processed into tubes, beams, and other shapes (codes 33111, 33131, 33122, and 33231). It also includes aluminum (33131), along with other metals like copper, zinc, and tin (33141, 33142, 33149) that are heavily used by the construction industry.
Wood — This includes lumber and other products that come from the sawmill (32111), along with plywood and other manufactured wood products like LVL (32121). It also includes millwork (32191), which is things like wooden doors, crown molding, baseboards, and other wood trim items.
Other materials — A sort of catchall category that includes raw and finished stone (21231 and 32799), cement and concrete (32731 and 32733), sand (21232), clay tile (327120), glass (327211), paint (32551), and gypsum products like drywall (32742).
Building components — These are manufactured parts and equipment that go into a building, and includes things like pipes and plumbing fixtures (326191, 32612, 332913, 332996), screws and bolts (33272), hardware like door handles and hinges (33251), elevators and escalators (333921), electrical components and wiring (33421, 335311, 335313, 33592, 33593), and heating and cooling equipment (333415).
Interior fit-out — This includes things that are used to finish and furnish the interior of a building, and includes things like carpet (31411), curtains (31412), lighting (3351), appliances (33521, 33522), furniture (33712, 33721), and cabinets (33711).
Tools — This includes both power tools (333991) and construction machinery (33312), as well as some construction-specific factory equipment like sawmill machinery (333243).
Products in each of these categories won’t exclusively be used by the construction industry. Steel, aluminum, and copper obviously have many uses besides buildings and infrastructure, for instance. And the construction industry isn’t the only user of power tools. But each category is a major input into construction and building, and I suspect construction makes up a large share of each category in most cases. For metals for example, construction is apparently the largest user of aluminum, the largest consumer of copper, and is probably the largest consumer of zinc in the form of galvanized steel.
Using USA Trade Online, we can get the total value of imports in each category. To get a sense of the relative importance of imports, we can compare these values to the total value of shipments by US businesses in each category, which is provided by the US Economic Census. (Because the most recent economic census is from 2022, I’ve used 2022 data for both cases.)
For each category I created a simple “imports index,” that’s just the total value of imports in a category by the combined value of imports and shipments from US businesses. Note that this is NOT the fraction of sales that come from imports. Figuring that out would be extremely complex, and need to take into account things like whether a product is an intermediate input, how and whether a product is resold, how much imports are marked up, and so on.
For example, if a business imports a widget for $10, repackages it, and resells it for $100. The imports index would show imports as less than 10% of the value of US and import shipments combined, even though US shipments were totally dependent on them: $10/($100 + $10)≈ $9. But this can still give us a rough idea of how important imports are to different categories of goods.
This treemap shows each category of import: the larger the box, the greater the dollar value of imports. The darker and more solid the box’s color, the larger the import index, and the higher that category’s imports are in proportion to US company shipments.
Overall I counted $469 billion worth of construction-related imports in 2022, or just under 15% of the $3.2 trillion worth of goods the US imported in 20223, though as we’ve noted not all these imports will be used by the construction industry.
The largest category was steel and metal, with imports totaling $163 billion. But while steel is the largest individual subcategory (iron and steel product imports total $48 billion), overall non-steel metal imports are higher. Aluminum and nonferrous metals like copper and tin in various forms combined total $98B, or more than 60% of the import value of this category.
Steel also has a lower import index than non-steel metals. Steel products direct from the mill have a 27% import index, rolled and drawn steel products are at 12%, structural plate products (which we can assume are mostly heavy steel beams and columns) are at 9%, and tubes and pipes are at 8%. With non-ferrous metals, on the other hand, aluminum is at a 34% import index, other non-ferrous metals are at 78%, and rolled and extruded non-ferrous metals other than copper are at 60%.
Overall, the US imports a lot of metal used in the construction and building industries. These are mostly metals other than steel, which dominate both in dollar terms and in proportion to domestic shipments. There’s also a tendency for imports to be less finished products: the three largest categories in this group, totalling $117 billion worth of imports, include products directly from the mill that need to undergo further processing (being rolled, drawn, extruded into different shapes, and so on). This sort of processing seems to be more frequently done in the US: metal products produced by processing purchased metal tend to have lower import indexes.
The next largest category is interior fit-out — things like furniture, lighting, and appliances — totalling $129 billion in imports. Lighting, curtains, and small appliances aren’t surprising to me as major import categories, but I didn’t expect to see furniture have such high dollar values and import indexes. Household & institutional furniture, office furniture, and cabinets and countertops combined are $49 billion worth of imports, and household & institutional furniture has a 58% import index. Apparently the low dollar density of furniture doesn’t stop it from being shipped very long distances.
The other notable bit of information in this category (though not exactly surprising) is the very high import indexes for small appliances and curtains — 84% and 80% respectively. Apparently very few of these products are produced in the US.
Building components is the next largest category, with $84 billion worth of imports. The biggest individual subcategory here is heating and air conditioning equipment at $18 billion — though if you add together the various categories of electrical equipment (switchgear, wiring and cable, transformers, telephone equipment), it’s larger at $38 billion. On the other hand, pipes and plumbing equipment make up just $9 billion across several different categories.
Tools are up next with $46 billion worth of imports. The largest subcategory here, making up more than 50% of the total, is construction machinery at $28 billion, followed by power tools at $7 billion. Power tools is another subcategory with a very high import index — something we apparently produce very little of in the US.
Wood is next at $31 billion. While wood imports often get a lot of media attention, we can see that they’re a relatively small fraction of construction-related imports. Wood imports are less than 20% the value of steel and metal imports, close to 1/3rd the value of building components, and only slightly higher than the value of imported construction machinery. Wood products also have relatively low import indexes, indicating that most of our wood is produced domestically.
And last, we come to the catchall category of other materials at $17 billion. The largest subcategory here is “other nonmetallic mineral products” at $8 billion which includes things like cut stone and mineral wool, followed by clay building materials (ceramic and clay tiles for roofs and interiors) at $3 billion, and cement at $2 billion. In general these subcategories are small, and most of them have very small import indexes: outside of clay tile and “other materials”, no item in this category has an import index higher than 16%.
One item in this group is drywall (under the heading “gypsum products”), and is another area where the actual numbers don’t seem to quite align with how much attention it gets. Groups like the National Association of Homebuilders are very concerned about the impact of tariffs on drywall imports, but not only are drywall imports very small (less than $500 million, or less than 3% the value of dimensional lumber imports), only a tiny fraction of drywall is imported. The import index is only 3.5%, suggesting the overwhelming majority of drywall is made in the US (the large number of factories US Gypsum operates in the US seems to confirm this.) So it’s not clear to me why drywall imports should merit much attention.
Beyond knowing what sort of goods are being imported, the next question is: Where do they come from?” The chart below is another treemap showing construction-related imports, except this time imports are broken down by country. The color-coding has been kept the same from the first chart, so you can see what sorts of imports are coming from different countries.
The largest single country by dollar value is, unsurprisingly, China, with $89 billion worth of construction-related imports in 2022. But it’s interesting to see what these imports consist of. It’s almost entirely building components and products related to interior fit-out, which combined are $74 billion (more than 80% of all Chinese construction-related imports). China is also the largest individual supplier of most of our power tools, 40% of the total. Raw materials, on the other hand, are a very small fraction of construction imports from China. China is the source of just 3.3% of our steel and metal imports, 3.8% of our wood imports, and 10% of our “other material” imports.
With Canada, the second largest country for construction-related imports, this pattern is largely reversed. Materials (metal, wood, and other) combined make up more than 75% of Canada’s construction-related imports. More than 20% of our metal imports and 50% of our wood imports come from Canada.
Mexico, our third largest trading partner, has the advantages of both close proximity to the US and low labor costs, and we see quite a mix of imports. There are bulky, low dollar density things which benefit from short shipping distances like metals, large appliances, furniture, and AC equipment. But there’s also a fair amount of (presumably labor intensive) manufactured goods like small appliances, electrical equipment, and power tools.
Among German imports, construction machinery and metals dominate (construction machinery doesn’t surprise me, but I am surprised to see such a large volume of metal imports). Among Vietnamese imports, furniture dominates, making up almost 50% of construction-related imports. Among Japanese imports, construction machinery is by far the largest category (more than 40% of all our Japanese construction-related imports). And among South Korean imports, metal and major appliances are the two largest categories.
While these are the US’ largest individual trading partners, together they still make up less than 2/3rds of US building and construction imports. The rest come from everywhere else: Taiwan, Switzerland, Italy, Brazil, and dozens of other countries around the world. Occasionally one of these countries will be responsible for a very large fraction of imports in some particular category (India supplies around 36% percent of our lime mortar imports; Italy supplies around 24% of metal plumbing fixture imports), but the US typically imports a small amount of each category from many different countries.
Conclusion
One takeaway is that construction-related imports are very large, hundreds of billions of dollars, and make up a substantial fraction of all US imports — though it’s hard to know how much of these imports are used by the construction industry specifically. And this analysis is far from a full accounting of the various imports used by the construction process. It tracks many of the direct inputs (materials, equipment, and labor), but not the inputs to those inputs, or the inputs to those inputs, and so on. It considers paints and coatings, but not the pigments that go into those paints, or the various chemicals used to manufacture those pigments. It tracks pipes and other plastic parts, but not the molding machines used to manufacture them, or the machines used to manufacture those machines, and so on. In today’s globalized world, the production of anything requires a huge web of inputs and manufacturing processes that spans the globe.
Another takeaway is that the proportion of imports to total shipments in a given category is often very high — even for large, bulky items that I might have expected to be difficult to ship long distance cost effectively. The high import index of things like small appliances and power tools doesn’t surprise me, but we see substantial import indexes in almost every single category. Of the 55 import categories I looked at, 32 of them (~60%) had an import index of 20% or higher. If instead we weigh the total value of imports, that fraction rises to 85%.
Imports are also highly dispersed. Out of the 55 categories examined, only in 12 of them does a single country make up 50% or more of inputs, and no country is responsible for more than 20% of construction related inputs. The world economy is knit together by enormous and complex patterns of trade, and the construction industry is no exception.
North American Industry Classification System
For glass, for instance, you need to go down to a 6-digit code to get to the sort of flat glass used in buildings.)
2024 imports were similar.
One has to assume that this bodes poorly for the people in Pacific Palisades whose houses burned down.
So these imports will hurt our friends more than our adversaries. That’s on brand for Trump.