\documentclass[12pt]{article} \usepackage{amssymb} \usepackage{graphicx} \usepackage[dcucite]{harvard} \usepackage{amsmath} \usepackage{color}\usepackage{setspace} \usepackage{booktabs} \usepackage{hyperref}\hypersetup{colorlinks=true, linkcolor=black,citecolor=black,filecolor=black,urlcolor=black} \usepackage[T1]{fontenc} \usepackage{threeparttable} \usepackage{array} \usepackage{longtable} \usepackage{subcaption} \usepackage{caption} %\usepackage{natbib} \setcounter{MaxMatrixCols}{10} \newtheorem{theorem}{Theorem} \newtheorem{acknowledgement}[theorem]{Acknowledgement} \newtheorem{algorithm}[theorem]{Algorithm} \newtheorem{axiom}[theorem]{Axiom} \newtheorem{case}[theorem]{Case} \newtheorem{claim}[theorem]{Claim} \newtheorem{conclusion}[theorem]{Conclusion} \newtheorem{condition}[theorem]{Condition} \newtheorem{conjecture}[theorem]{Conjecture} \newtheorem{corollary}[theorem]{Corollary} \newtheorem{criterion}[theorem]{Criterion} \newtheorem{definition}[theorem]{Definition} \newtheorem{example}[theorem]{Example} \newtheorem{exercise}[theorem]{Exercise} \newtheorem{lemma}[theorem]{Lemma} \newtheorem{notation}[theorem]{Notation} \newtheorem{problem}[theorem]{Problem} \newtheorem{proposition}[theorem]{Proposition} \newtheorem{remark}[theorem]{Remark} \newtheorem{solution}[theorem]{Solution} \newtheorem{summary}[theorem]{Summary} \newenvironment{proof}[1][Proof]{\textbf{#1.} }{\ \rule{0.5em}{0.5em}} \setlength{\topmargin}{0.1in} \setlength{\headheight}{0in} \setlength{\headsep}{0in} \setlength{\topskip}{0in} \setlength{\textheight}{8.5in} \setlength{\oddsidemargin}{0in} \setlength{\evensidemargin}{0in} \setlength{\textwidth}{6.5in} \makeatletter \setlength{\@fptop}{0pt} \makeatother \begin{document} \title{Trade Policy, Human Capital, and Costly Worker Transitions} \author{Ross Jestrab and David Riker} \date{\vspace{1.5in}% \today} \thispagestyle{empty} { % set font to helvetica (arial) to make it 508-compliant \fontfamily{phv}\selectfont \begin{center} {\Large \textbf{TRADE POLICY, HUMAN CAPITAL,}} \\ \vspace{0.25in} {\Large \textbf{AND COSTLY WORKER TRANSITIONS}} \\ \vspace{0.75in} {\Large Ross Jestrab and David Riker} \\ \vspace{0.75in} {\large ECONOMICS WORKING PAPER SERIES}\\ Working Paper 2023--10--A\\ \vspace{0.5in} U.S. INTERNATIONAL TRADE COMMISSION \\ 500 E Street SW \\ Washington, DC 20436 \\ \vspace{0.5in} October 2023 \\ \end{center} \vspace{0.5in} \vfill \noindent The authors thank Saad Ahmad and Peter Herman for helpful comments and suggestions on an earlier draft. Office of Economics working papers are the result of ongoing professional research of USITC Staff and are solely meant to represent the opinions and professional research of individual authors. These papers are not meant to represent in any way the views of the U.S. International Trade Commission or any of its individual Commissioners. \newpage \thispagestyle{empty} % remove headers, footers, and page numbers from %cover page \begin{flushleft} Trade Policy, Human Capital, and Costly Worker Transitions, \\ Ross Jestrab and David Riker \\ Economics Working Paper 2023--10--A \\ October 2023 \\~\\ \end{flushleft} \vfill \begin{abstract} \noindent Accumulation of human capital, through formal education and on-the-job experience, shapes workers' adjustment to job displacements after tariff reductions -- whether they switch industries, become unemployed, or leave the labor force. We develop an economic model that uses individual-level longitudinal CPS data on the employment transitions of U.S. manufacturing workers to estimate prospective labor adjustments in response to hypothetical reductions in tariff rates on U.S. manufacturing imports. The simulation suggests that a worker's observable human capital characteristics are important determinants of labor market outcomes following a trade policy change. \end{abstract} \vfill \begin{flushleft} Ross Jestrab \\ Research Division, Office of Economics \\ ross.jestrab@usitc.gov \\ \vspace{0.5in} David Riker \\ Research Division, Office of Economics \\ david.riker@usitc.gov \\ \vspace{0.5in} \end{flushleft} } % end of helvetica (arial) font \clearpage \newpage \doublespacing \setcounter{page}{1} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \section{Introduction \label{sec: section1}} The impact of international trade on the labor market outcomes of workers, such as employment and wages, is an important topic in the international economics literature.\footnote{For example, \citeasnoun{autor2013china} study the impact of rising import competition from China on local labor markets in the United States. Another example includes \citeasnoun{autor2014trade}, who use worker-level data from the U.S. Social Security Administration to study the impact of rising import competition from China on earnings and employment of U.S. workers.} Changes in trade policy, like tariff reductions on U.S. imports of manufactured goods, affect the domestic industry by reducing labor demand. The resulting effects on employment are distributed unevenly across groups of workers that vary in their human capital characteristics, specifically their education and age. \citeasnoun{Riker2023a} estimates prospective modeled-based measures of trade exposure that vary by worker demographic groups using an employment-weighted average of trade intensities.\footnote{The trade exposure measure builds on the work of \citeasnoun{ebenstein2014estimating}, who conduct retrospective analysis on potential wage impacts across occupations.} Based on data on U.S. manufacturing industries in 2021 and a hypothetical ten-percent increase in trade costs, the author finds that workers' race and education have larger effects on the trade exposure measures than the workers' gender, ethnicity, or occupation. In an extension, \citeasnoun{Riker2023b} reapplies the model with demographic groups defined by race, sex, age, and educational attainment in a retrospective analysis of rising trade costs on U.S. manufacturing imports between 2021 and 2022. \citeasnoun{Riker2023b} also considers changes in trade costs on imported intermediate goods, in addition to changes in trade costs on imported final goods. This paper presents an economic model to estimate labor market transitions by worker characteristics following a hypothetical trade policy change. The model is consistent with economic theory and industry data, transparent and reproducible, and detailed but practical in its data requirements. In particular, the model builds on the employment-weighted average approach from \citeasnoun{Riker2023a} and incorporates data on recently displaced workers and the transition of displaced workers using worker-level data from the Current Population Survey (CPS). We focus on trade in manufactured goods and wages are assumed to be downward-rigid, so reductions in labor demand translate into unemployment of some U.S. workers. The model estimates the percentage of displaced workers who are initially employed in manufacturing and become unemployed, exit the labor force, switch to another manufacturing industry, or switch to a non-manufacturing industry. These employment transitions are estimated for worker groups defined by age and education, which serve as proxies for the human capital of the workers. The model can be extended to include other worker groups. As an alternative, we group workers by age and gender rather than age and education. We simulate a hypothetical reduction in U.S. tariffs on imported manufactured goods from 2022 rates to 2017 rates for industries that experienced an average tariff increase between 2017 and 2022.\footnote{For two NAICS manufacturing industries, we set the tariff change to zero for the simulations as these industries experienced an average tariff decrease between 2017 and 2022.} The percentage of initial employed manufacturing workers who are estimated to experience a transition range between 0.005 and 0.186 percent across the worker groups of age and education. We find that workers who are 40 years or older and are not college graduates are more likely to exit the labor force, switch to another manufacturing industry, or switch out of manufacturing. Younger workers (less than 40 years old) who are not college graduates are the most likely to exit to unemployment, but only slightly more than older workers who are also not college graduates. For each of the worker groups, switching to employment outside of the manufacturing sector is the most likely transition. The rest of the paper is organized into five parts. Section \ref{sec: section2} derives the modeling framework. Section \ref{sec: section3} describes the data sources, aggregations, and definitions. Section \ref{sec: section4} reports a simulation of labor adjustments following a reduction in U.S. tariffs on manufactured imports, with U.S. tariffs reduced to 2017 tariff rates. Section \ref{sec: section5} describes a spreadsheet version of the model that replicates the simulation and can be easily modified to consider alternative policy scenarios. Section \ref{sec: section6} discusses the estimates and provides concluding remarks. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \section{Modeling Framework \label{sec: section2}} \noindent This section derives an equation for estimating the effect of changes in tariffs on labor demand and employment based on an industry-specific partial equilibrium model of international trade. The model uses conventional functional forms and its derivation follows \citeasnoun{Riker2023a}. The equation for domestic employment effects is abbreviated in this case, compared to \citeasnoun{Riker2023a}, because we focus on exposure to import shocks and we do not include exposure to export shocks. We extend the model to include estimates for the employment transition paths of different worker groups and estimates of the average wage effects for these worker groups. In the model derivation, we assume there are two countries (domestic and foreign) and domestic imports from the foreign country are subject to import tariffs.\footnote{For the simulations, the domestic country is the United States and the foreign country is an aggregate rest of the world. It is straightforward to increase the number of countries in the model.} The value of domestic shipments of industry $i$ has the constant elasticity of substitution (CES) form in equation (\ref{eq:1}). \begin{equation}\label{eq:1} V_i = \theta_i \ E \ (P_i) \ ^{\sigma_i \ - \ 1} \ (p_i) \ ^{1 \ - \ \sigma_i} \end{equation} \noindent $V_i$ is the value of shipments of domestic producers in the industry, and $p_i$ is the producers' price. $E$ is aggregate expenditure in the domestic market, and $\theta_i$ is the expenditure share of products in industry $i$. $P_i$ is the CES industry price index in equation (\ref{eq:2}), and $\sigma_i$ is the elasticity of substitution between domestic and imported varieties. \begin{equation}\label{eq:2} P_i = \left[ \ n_i \ (p_i) \ ^{1 \ - \ \sigma_i} \ + n_i^* \ (p_i^* \ \tau_i) \ ^{1 \ - \ \sigma_i} \ \right]^{\frac{1}{1 \ - \ \sigma_i} \ } \end{equation} \noindent $n_i$ and $n_i^*$ are the numbers of domestic and foreign varieties in the industry, $p_i^*$ is the price of foreign varieties, and $\tau_i$ is a tariff factor on imports that is increasing in tariff rates. $\tau_i$ is equal to one plus the tariff rates in industry $i$, also known as the power of the tariff. We assume that production in industry $i$ has a Leontief technology that combines labor and materials in fixed proportions. $w$ is the marginal cost of labor, and $c$ is the marginal cost of materials. $a_{wi}$ and $a_{ci}$ are industry-specific unit factor requirements. In addition, we assume that there is monopolistic competition in the domestic market.\footnote{The models of monopolistic competition and trade in differentiated products in \citeasnoun{Krugman}, \citeasnoun{Melitz}, \citeasnoun{Chaney}, \citeasnoun{HMR}, and subsequent studies also assume that consumers have CES preferences.} There is a continuum of firms, and each has monopoly power in the unique variety that it produces. Each firm perceives that the own-price elasticity of demand for its variety is a constant, so its price is a constant mark-up over its marginal cost of production, as presented in equation (\ref{eq:3}). \begin{equation}\label{eq:3} p_i = \left( \frac{\sigma_i}{\sigma_i \ - \ 1} \right) \ \left( a_{wi} \ w \ + \ a_{ci} \ c \right) \end{equation} \noindent Equation (\ref{eq:3}) implies equation (\ref{eq:4}), where the second equality is obtained by multiplying the numerator and denominator by the quantity of shipments of domestic producers. \begin{equation}\label{eq:4} \sigma_i = \frac{p_i}{p_i - a_{wi} \ w - a_{ci} \ c} = \frac{V_i}{V_i - W_i - C_i} \end{equation} \noindent $W_i$ is the industry's total domestic wage payments, and $C_i$ is the industry's total cost of materials. Equation (\ref{eq:5}) is the percent change in industry labor demand (and equivalently employment) resulting from changes in the tariff factor in industry $i$.\footnote{Alternative assumptions about labor supply would break the equivalence between the percent changes in revenues and employment in the domestic industry.} It is a first-order log-linear approximation evaluated at the initial equilibrium in the domestic market. \begin{equation}\label{eq:5} \hat L_i = \left( \frac{L_i \ - \ L_i^*}{L_i} \right) \hat V_i \ - \ \hat p_i \end{equation} \noindent $L_i$ is total employment of domestic manufacturers in industry $i$, and $L_i^*$ is their employment associated with exports, so $\frac{L_i^*}{L_i}$ is equal to the industry's exports as a share of its total shipments, $\chi_i$. $\hat L_i$ is the proportional (or percent) change in employment in the domestic industry, $\frac{dL_i}{L_i}$, and $\hat \tau_i$ is the proportional change in the tariff factor. $L_i$ and $V_i$ are endogenous variables that change in response to changes in trade costs. $p_i$ is an endogenous variable that does not change, because it is fixed according to equation (\ref{eq:3}). Equation (\ref{eq:6}) is the percent change in the total value of domestic shipments resulting from changes in the tariff factor in industry $i$. \begin{equation}\label{eq:6} \hat V_i = \left(\sigma_i \ - \ 1 \right) \ \left(\hat P_i \ - \ \hat p_i \right) = \left(\sigma_i \ - \ 1 \right) \mu_i \ \hat \tau_i \end{equation} \noindent $\mu_i$ is the industry's import penetration rate, $\frac{M_i}{V_i - E_i + M_i}$. $M_i$ is the value of imports in industry $i$. To simplify these expressions for percent changes, we have held $E$, $w$, $c$, $a_{wi}$, $a_{ci}$, $\theta_i$, $\sigma_i$, and $p_i^*$ constant. The delivered price in the domestic market ($\tau_i \ p_i^*$) is not constant when $\tau_i$ changes. We assume that $n_i$ and $n_i^*$ remain fixed in the short run. Equation (\ref{eq:7}) is the reduced-form expression for the percent change in domestic industry employment in response to a change in tariff rates. It is derived by substituting equation (\ref{eq:6}) into equation (\ref{eq:5}). \begin{equation}\label{eq:7} \hat L_i = \left( 1 \ - \chi_i \right) \left(\sigma_i \ - \ 1 \right) \mu_i \ \hat \tau_i \end{equation} \noindent This is the labor demand and employment effect of changes in the tariff factor on industry imports. Equation (\ref{eq:8}) is the percentage of workers in group $g$ that were initially employed in manufacturing industry $i$ who are displaced and experience transition $k$ (switch within the manufacturing sector, switch to another sector, become unemployed, or leave the labor force) due to changes in tariff rates. \begin{equation}\label{eq:8} \phi^k_{gi} = \beta^k_{gi} \ \gamma_{gi} \ (-\hat L_{i}) \end{equation} \noindent $\gamma_{gi}$ is the share of workers displaced from industry $i$ that are in group $g$. $\beta^k_{gi}$ is the share of workers in group $g$ transitioning from industry $i$ who experience transition $k$. To aggregate across the manufacturing industries indexed by $i$, equation (\ref{eq:9}) takes an employment-weighted average of equation (\ref{eq:8}). Equation (\ref{eq:9}) is the percentage of workers in group $g$ that were initially employed in manufacturing who are displaced and experience transition $k$ due to changes in tariff rates. \begin{equation}\label{eq:9} \phi^k_{g} = \sum_i \lambda_{gi} \ \phi^k_{gi} \end{equation} \noindent The share of group $g$ workers who are employed in industry $i$, $\lambda_{gi}$, is equal to $\frac{L_{gi}}{L_{g}}$. $L_g = \sum_i L_{gi}$ is total domestic employment of workers in group $g$ across all manufacturing industries indexed by $i$, and $L_{gi}$ is domestic employment of group $g$ workers in industry $i$. According to equations (\ref{eq:8}) and (\ref{eq:9}), the labor adjustments of each group of workers in response to a change in tariff rates depends on the industry's import penetration rate ($\mu_i$), the industry's export share ($\chi_i$), the extent of substitution between its domestic and imported varieties ($\sigma_i$), the change in the tariff rate ($\hat \tau_i$), the group's work transition rates ($\beta_{gi}^k$), the share of workers displaced from industry $i$ who are in group $g$ ($\gamma_{gi}$), and the group's labor share of each industry ($\lambda_{gi}$). Each industry employs experienced incumbent workers and can also hire additional inexperienced and less productive new workers. We assume that the supply of inexperienced labor is perfectly elastic at exogenous wage $w$. Workers are paid the value marginal product of their labor. Experienced incumbent workers in an industry receive a wage premium $\omega \geq w$ reflecting their higher productivity in the specific industry. If displaced from the industry, they lose some or all of their wage premium. We assume that the wages of incumbent workers are downward-rigid, and they cannot avoid displacements by making wage concessions. Equation (\ref{eq:10}) is the average change in the wage of workers in group $g$ that were initially employed in the manufacturing industry and were displaced due to changes in tariff rates, $\Delta \ \omega_{g}$. \begin{equation}\label{eq:10} \Delta \ \omega_{g} = \sum_k \sum_i \ \lambda_{gi} \ \phi^k_{gi} \ \Delta \ \omega_{gi}^k \end{equation} \noindent $\Delta \ \omega_{g}$ is an average over all worker in group $g$. It is the ex-ante expected value for workers in group $g$ before displacement. It is substantially smaller than $\Delta \ \omega_{gi}^k$, which is the change in wage income of workers who are displaced and are making costly transition $k$. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \section{Data Sources, Aggregations, and Definitions \label{sec: section3}} The model uses individual-level micro-data on U.S. workers initially employed in a manufacturing industry. The data on the industry of the workers, as well as their age and educational attainment, are from public use micro-data files of the Annual Social and Economic (ASEC) supplement of the CPS.\footnote{These data are publicly available at \url{https://cps.ipums.org/cps/}.} This CPS longitudinal data link an individual's survey response in March of one year to the individual's response in March of the next year. The dataset reports the worker's industry, employment status, wages, household income, and demographic characteristics in both years. We categorize each worker as either not transitioning during the one-year period and staying in the same industry or transitioning by switching to another industry within the manufacturing sector, switching to another industry outside of the manufacturing sector, becoming unemployed, or leaving the labor force. When calculating the share of workers who experience transition $k$, $\beta^k_{gi}$, we focus on workers who are initially employed in a U.S. manufacturing industry and experienced a decline in wages over the one-year period, since these workers are most likely to have experienced involuntary displacement due to negative labor demand shocks. The estimates of $\beta^k_{gi}$ also do not include workers who left the labor force and are 60 years or older, since they are likely retirements rather than involuntary displacements. In addition to this worker-level information, the model uses industry-level data on shipments, payrolls, and costs of materials by NAICS three-digit U.S. manufacturing industry in 2021 from the U.S. Census Bureau's Annual Survey of Manufactures (ASM).\footnote{These data are publicly available at \url{https://www.census.gov/programs-surveys/asm/data/tables.html}.} The model also uses data on the 2021 and 2022 free alongside value of domestic exports and the landed duty-paid value and customs value of imports for consumption by NAICS three-digit industry from the U.S. International Trade Commission's Trade Dataweb.\footnote{These data are publicly available at \url{https://dataweb.usitc.gov/}.} Finally, the model calibrates the elasticity of substitution parameter for each manufacturing industry using the approach in \citeasnoun{AhmadRiker}. Their formula for calibrating $\sigma_i$ is presented in equation (\ref{eq:4}). It uses data on the value of shipments, wage payments, and costs of materials of domestic producers in the industry in the 2021 ASM. Table \ref{tab: industry-level model inputs} provides a summary of industry-level inputs of the model. The elasticity of substitution values range from 1.81 for the beverage and tobacco products industry to 4.86 for the petroleum and coal products industry. The industries' import penetration rates range from 8.1 percent for the printing and related products industry to 95.3 percent for the leather and allied products industry. The industries' export share range from 4.9 percent for the beverage and tobacco products industry to 51.3 percent for the leather and allied products industry. Industry average tariff rates rose in all manufacturing industries from 2017 to 2021 (and would \textit{decline} if tariffs were returned to 2017 rates) except in the beverages and tobacco products and petroleum and coal products industries. These industries experienced a 0.02 percent and 0.05 percent decrease, respectively, between 2017 and 2022. For the simulations, the change in the tariff rate is set to zero for these two industries (i.e., $\hat \tau_i = 0$ if $i$ is NAICS code 312 or 324). \vspace{0.5cm} \begin{table}[tbph] \centering \footnotesize \begin{threeparttable} \caption{Industry-Level Model Inputs} \label{tab: industry-level model inputs} \begin{tabular}{p{7cm} r r r r r} Manufacturing Industry & Elasticity & Import & Export & Tariff & Tariff \\ Name and & of & Penetration & Share & in 2017 & in 2022 \\ NAICS Code & Substitution & (\%) & (\%) & (\%) & (\%) \\ \toprule Food manufacturing (311) & 3.33 & 10.1 & 8.4 & 1.34 & 2.21 \\ Beverage and tobacco products (312) & 1.81 & 16.9 & 4.9 & 0.41 & 0.40 \\ Textile mills (313) & 3.54 & 37.3 & 29.5 & 4.48 & 6.75 \\ Textile product mills (314) & 3.24 & 62.8 & 11.7 & 6.21 & 8.71 \\ Apparel (315) & 3.19 & 94.2 & 31.9 & 13.41 & 14.92 \\ Leather and allied products (316) & 3.44 & 95.3 & 51.3 & 10.80 & 12.60 \\ Wood products (321) & 2.74 & 20.0 & 5.0 & 0.98 & 2.25 \\ Paper manufacturing (322) & 2.96 & 12.4 & 12.1 & 0.24 & 2.38 \\ Printing and related products (323) & 2.91 & 8.1 & 6.0 & 0.01 & 2.40 \\ Petroleum and coal products (324) & 4.86 & 12.6 & 15.0 & 0.21 & 0.16 \\ Chemical manufacturing (325) & 2.08 & 35.0 & 28.1 & 0.79 & 1.45 \\ Plastic and rubber products (326) & 2.88 & 25.7 & 11.8 & 2.51 & 5.78 \\ Nonmetallic mineral products (327) & 2.56 & 19.8 & 7.7 & 3.11 & 6.03 \\ Primary metal products (331) & 2.90 & 39.0 & 21.2 & 0.37 & 2.43 \\ Fabricated metal products (332) & 2.97 & 21.7 & 10.1 & 2.14 & 6.77 \\ Machinery manufacturing (333) & 3.19 & 44.9 & 31.6 & 0.71 & 3.16 \\ Computers and electronics (334) & 2.88 & 69.5 & 35.9 & 0.18 & 1.17 \\ Electrical equipment et al. (335) & 2.92 & 62.8 & 30.4 & 1.65 & 5.37 \\ Transportation equipment (336) & 4.54 & 36.0 & 22.7 & 0.96 & 2.02 \\ Furniture (337) & 3.10 & 46.9 & 5.6 & 0.20 & 6.52 \\ Miscellaneous manufacturing (339) & 2.39 & 60.3 & 27.4 & 1.00 & 1.56 \\ \bottomrule \end{tabular} \end{threeparttable} \end{table} \newpage Table \ref{tab: loss of wage income} reports the substantial average changes in the wage and salary income of displaced workers experiencing each of the four types of transitions after displacement.\footnote{As a reminder, we focus on workers who are initially employed in a U.S. manufacturing industry and experienced a decline in wages over the one-year period, since these workers are most likely to have experienced involuntary displacement due to negative labor demand shocks.} Exiting the labor force resulted in the greatest wage loss for all four groups, ranging from 73.4 to 89.2 percent reductions. Switching to another industry within the manufacturing sector resulted in the smallest wage loss for all four groups, ranging from 32.1 to 44.3 percent reductions. \vspace{0.5cm} \begin{table}[tbph] \centering \begin{threeparttable} \caption{Loss in Annual Wage and Salary Income Associated with the Transitions} \label{tab: loss of wage income} \begin{tabular}{p{5cm} r r r r} \ & \ & \ & Switches & Switches \\ \ & Exits to & Exits the & within & Out of \\ \ & Unemployment & Labor Force & Sector & Sector \\ Age and Education & \$ (\%) & \$ (\%) & \$ (\%) & \$ (\%) \\ \toprule Younger than 40 & \ & \ & \ & \ \\ \ & \ & \ & \ & \ \\ \hspace{0.5cm}Not a College Graduate & 15,119 & 20,292 & 21,612 & 19,324 \\ \ & (49.3) & (82.9) & (44.3) & (46.3) \\ \ & \ & \ & \ & \ \\ \hspace{0.5cm}College Graduate & 39,644 & 61,938 & 28,476 & 45,363 \\ \ & (60.9) & (89.2) & (32.1) & (44.4) \\ \ & \ & \ & \ & \ \\ \midrule 40 or Older & \ & \ & \ & \ \\ \ & \ & \ & \ & \ \\ \hspace{0.5cm}Not a College Graduate & 23,177 & 37,758 & 23,206 & 34,708 \\ \ & (52.8) & (77.3) & (35.9) & (50.4) \\ \ & \ & \ & \ & \ \\ \hspace{0.5cm}College Graduate & 48,385 & 65,784 & 58,149 & 74,482 \\ \ & (49.9) & (73.4) & (39.8) & (49.8) \\ \ & \ & \ & \ & \ \\ \bottomrule \end{tabular} \end{threeparttable} \end{table} Table \ref{tab: loss of total personal income} reports the average change in the total personal income of displaced workers experiencing each of the four types of transitions after displacement. These are smaller in percentage terms, since there are parts of income that are unaffected and in some cases income replacement. \begin{table}[h!] \centering \begin{threeparttable} \caption{Loss in Annual Total Personal Income Associated with the Transitions} \label{tab: loss of total personal income} \begin{tabular}{p{5cm} r r r r} \ & \ & \ & Switches & Switches \\ \ & Exits to & Exits the & within & Out of \\ \ & Unemployment & Labor Force & Sector & Sector \\ Age and Education & \$ (\%) & \$ (\%) & \$ (\%) & \$ (\%) \\ \toprule Younger than 40 & \ & \ & \ & \ \\ \ & \ & \ & \ & \ \\ \hspace{0.5cm}Not a College Graduate & 13,767 & 18,408 & 22,348 & 17,266 \\ \ & (42.4) & (71.9) & (44.0) & (39.8) \\ \ & \ & \ & \ & \ \\ \hspace{0.5cm}College Graduate & 40,092 & 53,515 & 27,813 & 41,882 \\ \ & (56.7) & (75.1) & (30.3) & (40.1) \\ \ & \ & \ & \ & \ \\ \midrule 40 or Older & \ & \ & \ & \ \\ \ & \ & \ & \ & \ \\ \hspace{0.5cm}Not a College Graduate & 18,495 & 28,076 & 22,369 & 33,859 \\ \ & (37.3) & (50.3) & (32.4) & (44.7) \\ \ & \ & \ & \ & \ \\ \hspace{0.5cm}College Graduate & 48,655 & 44,700 & 59,139 & 73,411 \\ \ & (44.5) & (42.9) & (36.6) & (44.9) \\ \ & \ & \ & \ & \ \\ \bottomrule \end{tabular} \end{threeparttable} \end{table} %\newpage There are limitations to the CPS longitudinal public use micro sample. First, workers might have been displaced and switched to another company but within the same initial manufacturing industry, and this shows up as no transition (staying). Second, the data only provide information on each worker at two time periods in consecutive years. This yields limited information on the length or severity of unemployment spells: for example, we cannot tell whether a worker lost a job in February and was unemployed one month in March, lost a job in April and was unemployed eleven months or more, or switched industries but was unemployed for a while between March and March. Third, the number of worker-level observations for calculating employment and transition shares can get limited when including more detailed worker demographic groups. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \section{Estimated Effects on U.S. Manufacturing Workers \label{sec: section4}} We simulate a hypothetical reduction in tariff rates on U.S. imports of manufactures from 2022 tariff rates to 2017 tariff rates. \autoref{tab: transitions of displaced} covers four different groups of workers defined by their age and educational attainment. The table reports the estimated percentage of initial group $g$ employment in manufacturing experiencing labor adjustment $k$, denoted as $\phi^k_{g}$ in equation (\ref{eq:9}). The percentage of workers initially employed in manufacturing and that are estimated to experience a transition ranges between 0.005 and 0.186 percent across the groups of age and education. Exiting to unemployment is much more likely for workers who are not college graduates. Age does not appear to make much difference. Exiting the labor force is sensitive to age, and is especially high for workers who did not graduate from college. Switching to another manufacturing industry or to a non-manufacturing industry is most likely for older workers who are not college graduates. For each of the worker groups, switching to employment outside of the manufacturing sector is the most likely transition. \autoref{fig:figone} provides a graphical illustration of these group-specific effects. \vspace{0.5cm} \begin{table}[tbph] \centering \begin{threeparttable} \caption{Transitions of Displaced U.S. Manufacturing Workers} \label{tab: transitions of displaced} \begin{tabular}{p{5cm} r r r r} \ & \ & \ & Switches & Switches \\ \ & Exits to & Exits the & within & Out of \\ \ & Unemployment & Labor Force & Sector & Sector \\ Age and Education & (\%) & (\%) & (\%) & (\%) \\ \toprule Younger than 40 & \ & \ & \ & \ \\ \hspace{0.5cm}Not a College Graduate & 0.028 & 0.052 & 0.084 & 0.104 \\ \hspace{0.5cm}College Graduate & 0.005 & 0.006 & 0.023 & 0.041 \\ \midrule 40 or Older & \ & \ & \ & \ \\ \hspace{0.5cm}Not a College Graduate & 0.027 & 0.130 & 0.177 & 0.186 \\ \hspace{0.5cm}College Graduate & 0.007 & 0.040 & 0.058 & 0.096 \\ \bottomrule \end{tabular} \end{threeparttable} \end{table} \newpage \begin{figure}[h!] \centering \caption{Graph of Worker Transitions from \autoref{tab: transitions of displaced}} \includegraphics[width=0.9\textwidth]{figone.png} \\ %\floatfoot{Notes: To be added.} \label{fig:figone} \end{figure} %Alt text: Bar graph of the results from Table 4 that plots the percentage of initial employment that experience one of the following transitions: exit to unemployment, exit labor force, switch within manufacturing, or switch to non-manufacturing sector. Results are presented for the worker groups: less than 40, no college; less than 40, college graduate; 40 or older, no college; and 40 or older, college graduate. \autoref{tab: income of displaced} reports the estimated average reduction in income of workers in the four groups. The largest average reductions in wage and total income are for workers who are 40 or older and not college graduates. The smallest average reductions in wage and total income are for workers who are younger than 40 and college graduates. \begin{table}[tbph] \centering \begin{threeparttable} \caption{Reduction in Income of Displaced Workers} \label{tab: income of displaced} \begin{tabular}{p{5cm} r r} \ & \ & \ \\ \ & Reduction in & Reduction in \\ \ & Wage Income & Total Income \\ \ & \ & \ \\ Age and Education & Annual \$ (\%) & Annual \$ (\%) \\ \toprule Younger than 40 & \ & \ \\ \hspace{0.5cm}Not a College Graduate & 53 (0.12) & 50 (0.11) \\ \hspace{0.5cm}College Graduate & 31 (0.03) & 29 (0.03) \\ \midrule 40 or Older & \ & \ \\ \hspace{0.5cm}Not a College Graduate & 161 (0.26) & 144 (0.21) \\ \hspace{0.5cm}College Graduate & 135 (0.10) & 126 (0.09) \\ \bottomrule \end{tabular} \end{threeparttable} \end{table} \newpage Next, to highlight the flexibility of this approach for estimating transitions for different groups of workers, we extend the model to categorize workers into two gender groups rather than two education groups. \autoref{tab: transitions with gender groups} report the estimated transitions for the four new groups of workers. Older and male workers are much more likely to be displaced. \vspace{0.5cm} \begin{table}[tbph] \centering \begin{threeparttable} \caption{Transitions of Displaced U.S. Manufacturing Workers} \label{tab: transitions with gender groups} \begin{tabular}{p{5cm} r r r r} \ & \ & \ & Switches & Switches \\ \ & Exits to & Exits the & within & Out of \\ \ & Unemployment & Labor Force & Sector & Sector \\ Age and Gender & (\%) & (\%) & (\%) & (\%) \\ \toprule Younger than 40 & \ & \ & \ & \ \\ \hspace{0.5cm}Female & 0.005 & 0.023 & 0.016 & 0.035 \\ \hspace{0.5cm}Male & 0.021 & 0.030 & 0.080 & 0.102 \\ \midrule 40 or Older & \ & \ & \ & \ \\ \hspace{0.5cm}Female & 0.008 & 0.054 & 0.054 & 0.069 \\ \hspace{0.5cm}Male & 0.023 & 0.107 & 0.179 & 0.205 \\ \bottomrule \end{tabular} \end{threeparttable} \end{table} \autoref{tab: income of displaced2} reports the estimated average reduction in income of workers in four groups. Older and male workers also experience the largest average reductions in wage and total income. \begin{table}[tbph] \centering \begin{threeparttable} \caption{Reduction in Income of Displaced Workers} \label{tab: income of displaced2} \begin{tabular}{p{5cm} r r} \ & \ & \ \\ \ & Reduction in & Reduction in \\ \ & Wage Income & Total Income \\ \ & \ & \ \\ Age and Gender & Annual \$ (\%) & Annual \$ (\%) \\ \toprule Younger than 40 & \ & \ \\ \hspace{0.5cm}Female & 13 (0.03) & 13 (0.03) \\ \hspace{0.5cm}Male & 66 (0.07) & 62 (0.07) \\ \midrule 40 or Older & \ & \ \\ \hspace{0.5cm}Female & 54 (0.09) & 52 (0.08) \\ \hspace{0.5cm}Male & 242 (0.18) & 218 (0.15) \\ \bottomrule \end{tabular} \end{threeparttable} \end{table} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \section{Analysis of Alternative Policy Scenarios \label{sec: section5}} As a supplement to this paper, we built an easy-to-use Excel spreadsheet version of the simulation model. Using the inputs for each manufacturing industry from \autoref{tab: industry-level model inputs}, the spreadsheet replicates the model estimates reported in \autoref{tab: transitions of displaced}. The spreadsheet can also be used to estimate the effects of different tariff changes, simply by altering the values of $\hat \tau_i$ in Column C in the Industry Inputs tab of the Excel file.\footnote{The effects of the trade policy changes could also be re-estimated for different assumptions about the value of the elasticity of substitution in each industry, by altering the values of $\sigma_i$ in Column D in the Industry Inputs tab of the spreadsheet file.} \autoref{tab: transitions with alternative policy} reports estimated effects when tariffs on manufacturing imports are eliminated, using the version of the model with two age groups and two education groups in \autoref{sec: section4}, \textit{except} there is no change in tariff rates for NAICS 331 (primary metals) and 332 (fabricated metals). They are maintained at 2021 rates in this alternative simulation. Specifically, we change the values in cells C18 and C19 in the Industry Inputs tab to zero, and the simulated effects update in the Estimated Impact by Group tab. \begin{table}[tbph] \centering \begin{threeparttable} \caption{Transitions with Alternative Policy Scenario} \label{tab: transitions with alternative policy} \begin{tabular}{p{5cm} r r r r} \ & \ & \ & Switches & Switches \\ \ & Exits to & Exits the & within & Out of \\ \ & Unemployment & Labor Force & Sector & Sector \\ Age and Education & (\%) & (\%) & (\%) & (\%) \\ \toprule Younger than 40 & \ & \ & \ & \ \\ \hspace{0.5cm}Not a College Graduate & 0.025 & 0.046 & 0.069 & 0.085 \\ \hspace{0.5cm}College Graduate & 0.005 & 0.006 & 0.022 & 0.040 \\ \midrule 40 or Older & \ & \ & \ & \ \\ \hspace{0.5cm}Not a College Graduate & 0.021 & 0.102 & 0.128 & 0.146 \\ \hspace{0.5cm}College Graduate & 0.006 & 0.038 & 0.048 & 0.091 \\ \bottomrule \end{tabular} \end{threeparttable} \end{table} \newpage To facilitate comparison of the two simulations, \autoref{tab: ratio of the estimated effects} reports the magnitude of the effects in \autoref{tab: transitions with alternative policy} (without tariff reductions in NAICS 331 and 332) relative to the magnitude of the effects in \autoref{tab: transitions of displaced} (with tariff reductions in all manufacturing industries). The ratios are less than one for all four transitions for all groups, except for younger workers who are not college graduates exiting to unemployment or exiting the labor force. \vspace{0.5cm} \begin{table}[tbph] \centering \begin{threeparttable} \caption{Ratio of the Estimated Effects, \autoref{tab: transitions with alternative policy} to \autoref{tab: transitions of displaced}} \label{tab: ratio of the estimated effects} \begin{tabular}{p{5cm} r r r r} \ & \ & \ & Switches & Switches \\ \ & Exits to & Exits the & within & Out of \\ Age and Education & Unemployment & Labor Force & Sector & Sector \\ \toprule Younger than 40 & \ & \ & \ & \ \\ \hspace{0.5cm}Not a College Graduate & 0.89 & 0.88 & 0.82 & 0.82 \\ \hspace{0.5cm}College Graduate & 1.00 & 1.00 & 0.96 & 0.98 \\ \midrule 40 or Older & \ & \ & \ & \ \\ \hspace{0.5cm}Not a College Graduate & 0.78 & 0.78 & 0.72 & 0.78 \\ \hspace{0.5cm}College Graduate & 0.86 & 0.95 & 0.83 & 0.94 \\ \bottomrule \end{tabular} \end{threeparttable} \end{table} \newpage %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \section{Discussion and Conclusions \label{sec: section6}} We have developed and applied a model of employment and wage effects due to changes in trade policy with heterogeneity by worker characteristics. The model has practical data requirements and can be used in retrospective analysis, as well as the hypothetical prospective simulation in this paper. The simulation suggests that a worker's observable human capital characteristics are important determinants of labor market outcomes following a trade policy change. The model has several limitations and these suggest potential extensions to the analysis. First, it may be possible to improve the estimated transition rates, $\beta^k_{gi}$ and $\gamma_{gi}$, using more detailed longitudinal micro data. Second, the data can be used to calculate a full distribution of the reduction in wage income within each group, rather than average reductions. Third, the analysis could include greater disaggregation of worker types, both within age and education and to include other worker characteristics, like in \citeasnoun{Riker2023a}. Finally, it could include indirect inter-industry effects through input-output links, like in \citeasnoun{Riker2023b}. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \singlespace \bibliographystyle{dcu} \bibliography{biblio} %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% \end{document}