Trade Liberalization and the Supply Chain: US Worker Outcomes

 The sources indicate that, contrary to the well-documented negative impact of the "China Shock" on the manufacturing sector, US workers outside manufacturing (NM) exhibited relative earnings increases following trade liberalization with China. This growth is primarily driven by the beneficial effect of upstream exposure—specifically, increased competition from China in input markets—which lowered costs and more than offset the detrimental impacts of a worker's own-industry and downstream (customer) exposures.

The following key concepts provide context for these relative earnings gains among non-manufacturing workers:

1. The Role of Supply Chain Position

The sources emphasize that adopting an input-output perspective is critical for understanding worker outcomes.

• Upstream Gains: NM workers saw relative earnings increases because their industries benefited from lower input prices or productivity growth among suppliers following the granting of Permanent Normal Trade Relations (PNTR) to China. These gains take several years to materialize but become significant and large in magnitude over time.
• Downstream Losses: While NM workers may lose earnings if important customers exit or shrink due to import competition, these downstream losses are smaller for NM workers than for manufacturing (M) workers.
• Asymmetry Across Sectors: Manufacturing workers did not experience similar upstream gains. This may be due to "co-offshoring," where manufacturing supply chain links move offshore together, whereas NM sectors (like hospitals or hotels) must remain near their local customers.

2. Geographic vs. Industry Exposure

The analysis reveals that county-level exposure is more influential than industry-level exposure in determining earnings outcomes for NM workers.

• Spatial Frictions: Because of geographic frictions to labor mobility, what matters most for a worker's earnings is the condition of their local county.
• Spillover Effects: For NM workers, "own-county" exposure captures the negative spillover effects of being located near directly affected manufacturing workers, which can reduce local demand for services or increase competition for service jobs.
• Net Benefits: Despite these negative local spillovers, the positive impact of greater import competition in supplier industries (upstream exposure) resulted in net relative earnings gains for the vast majority of NM workers.

3. Worker Heterogeneity and Tenure

The magnitude of relative earnings gains for NM workers varies significantly based on individual characteristics and firm tenure:

• Tenure: Gains are substantially higher for "high-tenure" workers (those employed by the same firm for a prolonged pre-period). High-tenure NM workers saw average relative gains of 29 log points, while "mixed-tenure" workers saw only 3 log points.
• Labor Market Competition: Mixed-tenure NM workers are more susceptible to competition from the net inflow of 1.9 million former manufacturing workers who moved into the NM sector between 2000 and 2007. This influx of workers increased competition for service jobs, which likely constrained earnings growth for those with lower tenure.
• Demographics: Within the NM sector, relative earnings gains were notably higher among women, whites, and younger workers.
• Initial Earnings: Workers who had high initial earnings performed relatively better than those with lower initial earnings, suggesting they possessed human capital that made them more competitive in a crowded labor market.

4. Duration and Timing of the Shock

The sources suggest that the impact of PNTR evolved over a long time horizon. While the negative impact of own-county exposure was immediate, the positive upstream effects for NM workers peaked around 2010. These gains eventually began to fade as firms absorbed initial cost reductions and competitive pressures adjusted. By contrast, manufacturing workers saw their outcomes worsen after the Great Recession as downstream losses intensified.

While non-manufacturing (NM) workers often experienced relative earnings gains following the "China Shock," the sources state that manufacturing (M) workers experienced substantial relative earnings losses. These losses are attributed to a combination of high exposure to direct competition, a lack of beneficial "upstream" supply chain effects, and significant "downstream" customer losses.

The following key themes characterize the experience of manufacturing workers within this context:

1. Asymmetric Supply Chain Exposure

The sources identify a critical asymmetry in how supply chain position affected M versus NM workers:

• Lack of Upstream Gains: Unlike NM sectors, which benefited from lower-priced Chinese inputs, M workers saw no significant earnings boost from upstream exposure.
• "Co-Offshoring" Hypothesis: The sources suggest this lack of benefit may be due to supply chain "co-offshoring," where multiple links of a manufacturing supply chain move abroad together. Consequently, domestic manufacturing firms may be unable to benefit from cheaper foreign inputs if their suppliers or customers have already relocated.
• Severe Downstream Losses: M workers were particularly susceptible to downstream exposure—the loss of domestic customers to Chinese competition—which exerted a strong negative impact on their earnings.

2. Primacy of Geographic Exposure

A major finding in the sources is that county-level exposure was more influential than industry-level exposure for M workers' earnings.

• Earnings vs. Employment: While highly exposed manufacturing industries saw massive employment declines, a worker’s subsequent earnings depended more on their local county conditions.
• Mobility Frictions: Workers in less-exposed counties could transition from manufacturing to services with less of an earnings penalty, whereas those in highly exposed counties faced increased labor market competition and falling local demand, leading to sharper earnings declines.

3. Worker Transitions and Earnings Declines

The China Shock induced a massive shift in the labor force, with a net flow of 1.9 million workers moving from the manufacturing sector to the non-manufacturing sector between 2000 and 2007.

• Sectoral Shifts: Former M workers primarily transitioned into sectors like healthcare, construction, and wholesale.
• The "Earnings Penalty": Transitions to low-skill service industries—such as staffing agencies or food services—often resulted in nominal wage declines of up to 20 percent. The sources note these findings align with the narrative that well-paid manufacturing workers face significant income drops when moving to the service sector.

4. Heterogeneity in Outcomes

The impact on manufacturing workers varied significantly based on firm and worker characteristics:

• Firm Size and Diversification: M workers at small firms fared relatively better than those at large firms, possibly because small firms produce customized goods that are less substitutable by imports. Conversely, workers at "diversified" firms (those spanning both M and NM activities) actually had worse outcomes than those at firms focused solely on manufacturing.
• Tenure and Initial Earnings: "High-tenure" M workers—those with the same firm for years—experienced slightly larger relative earnings losses than "mixed-tenure" workers, primarily due to lower earnings conditional on remaining employed. However, M workers with higher initial earnings generally performed better, suggesting they possessed more transferable skills.
• Trading Status: M workers at firms that engaged in importing or exporting saw relatively better earnings conditional on employment compared to those at non-trading firms.

5. Timing and the Great Recession

The negative impacts on M workers evolved over time. While the initial negative impact was driven by "own-county" exposure immediately after trade liberalization, downstream losses intensified around the time of the Great Recession. This suggests that manufacturing firms that initially survived direct competition were eventually unable to withstand the loss of their customers during the subsequent economic downturn.

The sources provide several key analytical insights regarding the impact of the "China Shock"—triggered by the granting of Permanent Normal Trade Relations (PNTR) in 2000—on the relative earnings of US workers. While previous research emphasized the "China Syndrome" of manufacturing decline, these sources reveal a more nuanced picture characterized by relative earnings gains for non-manufacturing (NM) workers and the critical role of supply chain positioning.

The following analytical insights are central to understanding these labor market outcomes:

1. The Primacy of Upstream Gains for NM Workers

The most significant finding is that non-manufacturing workers exhibited relative earnings increases because the benefits of upstream exposure (increased competition from China in input markets) outweighed the costs of direct and downstream exposure.

• Cost Reductions: NM industries, such as hospitals or software publishers, benefited from lower prices or increased productivity among their suppliers due to trade liberalization.
• Magnitude of Gains: After accounting for supply chain linkages, NM workers in the majority of county-industry pairs saw relative gains ranging from 3 to 29 log points, depending on their tenure.
• Analytical Necessity: The sources argue that failing to account for these input-output (IO) linkages leads to an underestimation of NM gains and an underestimation of manufacturing (M) losses.

2. Spatial vs. Industry Exposure

A major analytical contribution of the sources is the discovery that county-level exposure is more influential than industry-level exposure for both NM and M workers.

• Labor Mobility Frictions: Because workers face geographic frictions when moving between regions, their earnings are more affected by the economic health of their "own-county" than by the specific exposure of their industry.
• Local Spillovers: For NM workers, own-county exposure captures negative spillovers, such as displaced manufacturing workers reducing demand for local services or increasing competition for service-sector jobs.

3. Asymmetric Impact and the "Co-Offshoring" Hypothesis

The sources highlight a stark asymmetry between the manufacturing and non-manufacturing sectors.

• M Worker Losses: Unlike NM workers, manufacturing workers saw no significant upstream gains and faced stronger downstream losses (the loss of domestic customers to Chinese competition).
• Co-Offshoring: The sources suggest manufacturing sectors may experience "co-offshoring," where multiple links of a supply chain move abroad together to maintain proximity. NM sectors, such as hotels or healthcare, cannot co-offshore because they must remain near their domestic customers, allowing them to capture the benefits of cheaper inputs while maintaining their domestic market.

4. The Impact of Worker Tenure and Heterogeneity

Worker outcomes were highly dependent on individual and firm characteristics:

• The Tenure Gap: High-tenure NM workers (those at the same firm during the pre-period) saw the largest relative gains. Conversely, "mixed-tenure" NM workers saw smaller gains, likely because they faced more direct competition from the 1.9 million manufacturing workers who transitioned into the service sector between 2000 and 2007.
• Initial Earnings: Workers with high initial earnings in both sectors performed relatively better, suggesting they possessed transferable human capital or financial cushions that allowed them to be more selective in finding new positions.
• Firm Size: In the manufacturing sector, workers at small firms fared better than those at large firms, possibly because small firms produce specialized, "customized" goods that are less easily replaced by Chinese imports.

5. Temporal Dynamics of the Shock

The "annual" analysis of the data reveals that the China Shock was not a static event but evolved over time:

• Immediate vs. Delayed Effects: The negative impact of own-county exposure was felt immediately after PNTR, while the positive upstream effects for NM workers and the negative downstream effects for M workers became significant several years later.
• Peak and Fade: The positive upstream benefits for NM workers peaked around 2010 before gradually fading as cost reductions were absorbed and competitive pressures adjusted. For manufacturing workers, downstream losses intensified during the Great Recession, suggesting that firms that initially survived the shock were eventually pushed out by the loss of their domestic customers during the economic downturn.

To find the relative earnings gains after the China Shock, the authors utilize a rich infrastructure of matched employer-employee data from the US Census Bureau’s Longitudinal Employer-Household Dynamics (LEHD) program. This data source is central to the analysis because it tracks the earnings of nearly all workers—both in manufacturing (M) and non-manufacturing (NM)—permitting a longitudinal study across different sectors and counties.

The sources describe the data sources and their analytical value as follows:

1. The LEHD Infrastructure

The LEHD program is created via a federal-state partnership, drawing its primary data from state unemployment insurance (UI) records and the Quarterly Census of Employment and Wages (QCEW).

• UI Records: These provide the earnings for approximately 96 percent of all private-sector employees, including gross wages, salaries, bonuses, and tips.
• QCEW: This provides information regarding firm locations and their specific industries of activity.
• Worker Characteristics: The LEHD matches these records to the Individual Characteristics File (ICF), which includes demographics such as age, gender, race, birth country, and educational attainment, often derived from the Decennial Census.

2. Firm and Trade-Level Linkages

To go beyond basic earnings, the authors link the LEHD to other high-quality Census and trade datasets:

• Longitudinal Business Database (LBD): This tracks attributes like firm size, age, and multi-unit status. Crucially, the LBD provides employment counts for all industries and counties, which allows the authors to calculate county-level exposure without the data suppression issues found in publicly available sources like the County Business Patterns (CBP).
• Longitudinal Foreign Trade Transactions Database (LFTTD): This provides information on a firm's direct exposure to international trade.
• Unit-to-Worker (U2W) File: This file is used to assign workers to specific establishments within a firm, ensuring accurate tracking of their specific industry and county of employment.

3. Measuring Exposure via Tariffs and Supply Chains

The analysis of the China Shock—specifically the granting of Permanent Normal Trade Relations (PNTR)—requires specific trade data:

• US Tariff Schedule: Industry exposure is derived from the rise in US tariffs on Chinese goods (the "NTR gap") as of 1999.
• Trade Data (Feenstra et al. 2002): The authors use Harmonized System (HS) level ad valorem equivalent tariff rates to compute these gaps for NAICS industries.
• BEA Input-Output (IO) Tables: To assess exposure via the supply chain, the authors use total requirements data from the 1997 Bureau of Economic Analysis (BEA) tables. This allows them to compute "upstream" exposure (competition in a worker's supplier markets) and "downstream" exposure (competition in a worker's customer markets).

4. Sample Coverage and Scope

The sources note that the availability of data in the LEHD varies by state and time:

• The 19-State Sample: For the core regression analysis covering the full pre- and post-PNTR period (1993–2014), the authors utilize data from 19 states that represent 47 percent of total US employment.
• The 46-State Sample: For descriptive results, such as tracking the net flow of 1.9 million workers from manufacturing to services between 2000 and 2007, the authors use a broader 46-state sample representing 96 percent of US employment.
• Stratified Sampling: For computational efficiency, the authors use 5 percent stratified draws from the population, ensuring all workers from small counties (those in the first population decile) are included to maintain geographic representation.

5. Supplemental Data

The analysis also incorporates other policy controls and descriptive data:

• MFA Quotas: The authors account for the elimination of textile and clothing quotas under the Multi-Fiber Arrangement.
• Chinese Policy Data: They utilize data from Brandt et al. (2017) for Chinese import tariffs and the Annual Report of Industrial Enterprise Statistics for production subsidies in China.
• Public BLS Data: Publicly available Bureau of Labor Statistics (BLS) data on hourly wages is used as a complement to confirm broader sectoral wage trends.