What do we know about the origin of parts in vehicles sold in the U.S. market?
Today’s supply chains in motor vehicle production extend well across country borders. According to analysis presented by the Center for Automotive Research, 48% of all light vehicles sold in the U.S. in 2017 were assembled abroad, with more than half of the imported vehicles produced in Canada (11%) or Mexico (14%) (see Schultz et al., 2019). What do we know about the origin of a vehicle’s parts? This is an important question as nearly 75% of a vehicle’s value added is represented in the parts that are shipped to the vehicle assembly line (Klier and Rubenstein, 2008). The source of parts is hard to measure as a vehicle consists of thousands of parts, many of which cross borders several times in the process of manufacture (NYT, 2018). This blog summarizes the proposed changes to regional content requirements for vehicles and vehicle parts under the new North American trade agreement and asks how we might go about measuring the parts content in vehicles sold in the U.S. We look at several tabulations of parts content and find consistently that the majority of parts sourced for vehicles produced in North America originate in North America.
New North American trade agreement—USMCA
After months of negotiations, a new trade agreement (USMCA) was signed by the leaders of the U.S., Mexico, and Canada last fall. The agreement still has to be ratified by the three legislatures. If it is, it will succeed the 1994 North American Free Trade Agreement (NAFTA) and introduce different rules for trade within North America.1
A key element of the new agreement is a higher level of required North American content (75% vs. 62.5%) for light vehicles. For motor vehicle parts, the content requirement has also been raised; it is staggered across different parts categories, ranging from 65% to 75%, up from the current 60% hurdle. In addition, the new agreement for the first time introduces a requirement on North American wage content: 40% of a passenger car’s value needs to be produced by workers making at least $16/hour. For light trucks, that requirement is set to 45%.2
How can we measure parts content for vehicles produced in North America?
Current NAFTA rules state that a vehicle for sale in North America needs to have at least 62.5% North American content in order to avoid import duties. How many vehicles sold in the U.S. pass the current NAFTA content requirement? The answer to that question is known to government authorities, but the data are not available to the general public. It is our understanding that, for the purpose of demonstrating compliance with current NAFTA requirements, carmakers report the North American content of every vehicle model for sale in the U.S. to the office of Automotive & Aerospace at Customs and Border Patrol.
Since we don’t have access to the data, we rely on other information about the origin of parts for vehicles assembled in North America. To estimate foreign parts content in vehicles, we can use either macroeconomic or microeconomic data.
The most frequently used macro data are aggregate trade statistics. In an industry that is highly integrated across borders, adding up the nominal value of traded goods will lead to double counting if parts are shipped back and forth across borders in the process of manufacture. For example, a capacitor, a small electrical part that stores electrical energy, is one of the elements in a circuit board, which in turn is part of the electronic seat controls, which in turn are part of a modern automotive seat. In the process of aggregation, the capacitor crosses a North American border four times before the finished seat, which includes the seat controls, gets installed in a vehicle at an assembly line (see Black, Diamond, and Merrill, 2017).
An alternative approach is to combine the input-output tables of individual countries to get a better handle of the source of value added by industry. The most recent paper in that literature was written in 2017 by de Gortari. The paper is innovative as is it could distinguish among the supply chains of different vehicle producers in Mexico (the author had access to detailed Mexican trade data that are not publicly available). For example, de Gortari was able to distinguish U.S. parts content in vehicles assembled in Mexico by a U.S. producer for sale in the U.S. from that in vehicles assembled in Mexico by a European producer for sale in Europe. The paper reports that in 2014, U.S.-produced vehicle parts accounted for 38% of the value added in light vehicles assembled in Mexico and exported from there to the U.S. market.
Alternatively, we can try to infer the origin of the parts content from detailed industry data. We summarize two such approaches below. One relies on content data collected by a government program, and the other uses the source of engines and transmissions for vehicles assembled in North America.
American Automobile Labeling Act (AALA):
There are several domestic content requirements carmakers are subject to for vehicles to be sold in the U.S. (see Klier and Rubenstein, 2007). One of these, the American Automobile Labeling Act (AALA), requires carmakers to report for each vehicle model sold in the U.S. the percentage of the U.S./Canadian parts content, the country of final assembly of the vehicle, the country of origin of its engine and transmission, and any country that accounts for more than 15% of the vehicle model’s parts content. The data are collected by the National Highway and Traffic Safety Agency (NHTSA) and reported annually on their website. The AALA has been in effect since 1997.
The data are measured in percentages of a vehicle’s parts value, a smaller cost base than is required per NAFTA rules3 (Menk, Chen, and Cregger, 2012). The data are only reported for models sold in the U.S. One of the quirks of the data is that Canadian and U.S. parts content are reported jointly. The data also do not provide a comprehensive accounting of a vehicle’s parts costs, as countries accounting for parts costs of less than 15% don’t need to be reported. In addition, carmakers have the option of reporting data in aggregate fashion; for example, if a specific vehicle model for sale in the U.S. market is assembled in several countries, the parts sourcing data can be reported for all production locations combined. The bottom line is that the data are somewhat noisy.
We combined the AALA data with vehicle-model level production data in order to be able to calculate production-weighted content averages. In the process we cleaned up several inconsistencies in the reported data. Our approach captures nearly all of light vehicle production in the U.S. (98%) and Canada (99%), and the vast majority of Mexico’s production (85%).
We find that in 2016, 27% of the parts value of a vehicle assembled in Mexico for sale in the U.S. market was represented by parts originating in the U.S. or Canada. That share is about half as large as it was in 1997 (63%) (see Table 1). The U.S./Canadian parts content for vehicles assembled in the U.S. or Canada also declined over the same period, but not as steeply. We also find that overall about three-quarters of a vehicle’s parts content assembled in any of the three North American countries originated in North America in 2016 (see Table 2).
1. U.S./Canadian parts content (in percent) in vehicles produced in:
Note: Data reported are production-weighted averages by vehicle model years for light vehicles sold in the U.S.
2. Source of parts content (in percent) for light vehicles produced in:
|Country of origin for parts:||2011||2016||2011||2016||2011||2016|
Note: Data reported are production-weighted averages by vehicle model years for light vehicles sold in the U.S.
This analysis draws on proprietary data acquired from IHS Markit for 2000 to 2016. The data comprehensively identify in great detail the sourcing of only two parts, engines and transmissions, for vehicles assembled in North America. Engine and transmission together, also referred to as a vehicle’s powertrain, represent about 21% of a vehicle’s cost (Menk, Chen, and Cregger, 2012). The data do not tell us where the various engine and transmission parts originate.
For the industry as a whole, we find a level of supply chain integration similar to what we observe in the AALA data: Most vehicles produced in North America include engines and transmissions produced in North America. In 2016, 86% of all engines (and 77% of all transmissions) sourced for light vehicles assembled in North America originated in North America (see Table 3). Within North America, the share of engines and transmissions sourced from Mexico rose over the period, in conjunction with the growth of light vehicle production in Mexico (Klier and Rubenstein, 2017).
3. Origin of powertrains used in light vehicles assembled in North America
|Same country that builds vehicle||57.6||55.1||49.4||49.0|
|Elsewhere w/in N. America||30.8||30.6||25.6||27.5|
|Sourced w/in N. America||88.4||85.7||75.0||76.5|
|Imported from elsewhere||11.6||14.3||25.0||23.5|
|All (units)||17.1 mio||17.8 mio||17.1 mio||17.8 mio|
Note: Percentages reported are production-weighted averages by calendar year
Furthermore, for fewer than one-third of the vehicles assembled in North America does vehicle assembly, engine production, and transmission production take place in the same country. In 2016, only 27% of light vehicles produced in North America were assembled in the U.S. and had both engine and transmission sourced from the U.S.; 4% of all light vehicles produced had all three functions performed in Mexico.
There is no perfect way to measure supply chain integration. Both the macro and micro approaches represent approximations. Macro data tend to miss important details due to aggregation, while micro data tend to miss due to lack of coverage.
We presented two different micro-data approaches. Both provide evidence for high levels of supply chain integration within North America’s auto industry. A comparison to findings based on the input-output table analysis is not straightforward, but the results seem broadly consistent.
1 See USITC (2019) for a report assessing the likely impact of USMCA.
2 At least 25% of the value for cars, and 30% for light trucks, need to originate in the production of materials or of the vehicle (USITC, 2019, p. 19).
3 NAFTA content accounting for vehicle assembly utilizes net costs, which include vehicle manufacturing costs as well as overhead, but not costs for selling, marketing, and shipping of the vehicle. AALA reports percentages of only the vehicle’s parts value (Freund, 2017, pp.2-3; and Canis et al., 2017, pp. 13-17).
Black, Thomas, Jeremy Scott Diamond, and Dave Merril. 2017. One Tiny Widget’s Dizzying Journey Shows Just How Critical Nafta Has Become. Bloomberg. February 2, available online.
De Gortari, Alonso. 2017. Disentangling global value chains. Mimeo, available online.
Klier, Thomas, and James Rubenstein. 2017. Mexico’s growing role in the auto industry under NAFTA: Who makes what and what goes where. Economic Perspectives (6). Federal Reserve Bank of Chicago, available online.
Klier, Thomas, and James Rubenstein. 2008. Who really made your car? The Upjohn Institute, available online.
Thomas Klier, and James Rubenstein. 2007. Whose part is it? Measuring domestic content of vehicles. Chicago Fed Letter. 243 (October). Federal Reserve Bank of Chicago, available online.
Menk, Debbie; Yen Chen, and Joshua Cregger. 2012. Methodology for creating a matrix to assess the domestic content of a vehicle by make and model. Center for Automotive Research, available online.
Reuters. 2019. Delays at U.S.-Mexico Border Crossing hits Autos, Trucks still lining up. April 8, available online.
Schultz, M., Dziczek, K., Chen, Y., and Swiecki, B. (2019). U.S. Consumer & Economic Impacts of U.S. Automotive Trade Policies. Center for Automotive Research, Ann Arbor, MI, available online.
U.S. International Trade Commission. 2017. U.S.-Mexico-Canada Trade Agreement: Likely Impact on the U.S. Economy and on Specific Industry Sectors, available online.