Labor Market Adjustment to Globalization with Heterogeneous Agents Carl Davidson1,2, Steven Matusz1,2 and Susan Zhu1 1Michigan State University 2GEP, University of Nottingham Large Literature on Firm and Plant Level Adjustment to Globalization Based on Bernard and Jensen 1999; Roberts and Tybout 1997 Only a fraction of firms export and those that do export only a faction of output Exporting firms are bigger, more capital intensive, pay higher wages There is “imperfect persistence” in the decision to export Openness and Productivity Openness enhances productivity in export markets -- could be due to “learning by exporting” or market share reallocations in favor of more productive firm Openness leads to exit of weakest firms in export markets Within-firm productivity gains in importcompeting markets Our Goal To take a similarly disaggregated look at labor market adjustment to globalization We emphasize: Firms are different by choice (adopt different technologies, employ different skill mixes of workers, pay different wages) Need heterogeneity on both sides of labor market Our Goal With imperfect labor markets, qualities of the firm-worker matches may depend on a variety of factors, including trade position Focus is on how globalization affects the performance of the labor market and the distribution of wages Some Reasons to Add Worker Heterogeneity to New Trade Models Firms that adopt different technologies hire workers (in terms of skill), pay different wages Impact of liberalization on dislocated workers varies greatly Underemployment a serious concern (especially lately with outsourcing of high-skill jobs) Framework High and low skill workers search for jobs Ex ante identical firms choose what type of technology to adopt In equilibrium, firms are different Low-tech firms pay low wages (can use both types of workers) High-tech firms use high-skill workers and pay high wages Key Feature of the Model If revenues generated by the two types of firms are not too different, high-skill workers accept low-tech jobs if they find them first (there is underemployment) Yields suitable framework to analyze how well the labor market sorts workers across firm The Model Labor market based on Mortensen and Pissarides (REStud 1994) Heterogeneity based on Yeaple (JIE 2005) and Albrecht and Vroman (IER 2002) Trade extension based on Davidson, Matusz and Shevchenko (JIE 2007) The Model Product market is perfectly comp. Labor market frictions Firms create vacancies until exp. profit from doing so = 0 Total measure of workers = 1 q = fraction with low-skills si = skill level of worker i Firms rent capital after filling vacancy The Model -- Technology Low-tech production process: k s f (k , s ) k s L j 1 L 1 M s s M if s s if s s j j L H L y f ( k , s ); y f ( k , s ) L L L M L H The Model -- Technology High-tech production process (note that HS workers better suited for HT production process): 0 f (k , s ) k s H s s H j M y f (k , s ) H H H 1 H if s s L if s s H j j The Model – Matching Matching Function: m(u,v) CTRS: m() with = v/u The arrival rate of vacancies for workers is m() The arrival rate of workers for firms is z()=m()/ We assume m’() >0; z’()<0 The Model -- Matching = fraction of vacancies that are low-tech = fraction of unemployed with low-skill Low-skill workers find jobs at rate m() High-skill workers find jobs at rate m() Low-tech firms fill vacancies at rate z() High-tech firms fill vacancies at rate (1-)z() Additional Assumptions Wages determined by GNBS (wages increasing in the size of the surplus and the worker’s outside option) High-skill workers accept low-tech jobs if surplus to be split is positive (takes into account outside options and search costs) Firms export if doing so increases the surplus to be split with the worker There is a fixed cost to exporting The Export Decision (Firm Adj.) Initially, all firms sell output at p in the closed economy Now, allow firms to export and get the world price p* > p There is a fixed cost of exporting The high-tech firms (the largest, most productive, most capital intensive firms) face the strongest incentive to export since they gain the most from exporting The Export Decision (Firm Adj.) Get entry by HT firms and as they leave to serve foreign market also get entry by LT firms to serve domestic market Relatively more entry by HT firms ( falls) In export markets industry-level productivity rises as market shares are reallocated towards high-tech firms But, there are no within-firm productivity changes The Export Decision (Firm Adj.) Can get imperfect persistence for low-tech firms Low-tech firms might change their export decision when the skill mix of its employee base changes if replacing a low-skill worker with a high-skill workers allows them to cover the fixed cost of exporting Results on Labor Market Adjustments Wage effects High-skill workers at high-tech firms gain (surplus higher, bargaining position improves since falls) High-skill workers at low-tech firms gain (outside opportunities better) Low-tech workers could gain or lose (higher surplus but bargaining position erodes) Wage Effects -- Predictions Falling trade costs should increase wages of the most highly skilled workers the most and may decrease the wages of the least skilled workers Increase in inequality within the firm This contrasts with Yeaple’s prediction (which assumes perfect competition in the labor market) in which middle level workers are harmed the most Wage Effects -- Evidence Exporting increases the wage gap (BJ 1997, Harrison and Hanson 1999) As markets have become more open, wage gap has increased (Baldwin and Cain 2000) Predictions on Labor Market Adjustment In terms of match quality, now get less underemployment of high-skill workers Labor market functions more efficiently (a new gain from trade) Can get complete labor market separation if revenues spread sufficiently so that low-tech firms cannot afford to pay high-skill workers enough Predictions on Productivity With separation, get within firm productivity losses in weakest firms in export markets Industry wide productivity still rises due to market share reallocations Import Competition If the model applies to an import-competing industry, openness lowers the price received by all firms This reduces the gap between the revenues earned the two types of firms If high-skill workers will not accept low-tech jobs in the closed economy, they may be willing to do so in an open economy Import Competition Can go from perfect labor market separation to equilibrium with underemployment Labor market is less efficient due to trade (a new loss from trade) Note that low-tech firms become more productive (can attract better workers) so there are within firm productivity gains for the weakest firms in the industry Results – Assortative Matching Big issue in literatures on marriage markets and imperfect labor markets – should the “good” types match with other “good” types and does this actually occur (Becker 1972)? Here, good workers are better suited for hightech employment so we want positive assortative matching Results – Assortative Matching Our model predicts that openness to trade affects what is likely to occur Increased openness makes positive assortative matching more likely in export-oriented markets but less likely in import-competing markets Openness affects the degree of assortative matching Match Quality -- Predictions Model yields several testable hypotheses about labor market adjustment to globalization One particular prediction worth highlighting: Openness may alter the types of job matches that we observe Openness may cause separation in exporting industries, mixing in import-competing industries Conclusion Acemoglu’s 1999 AER model has similar features (heterogeneous workers, initially identical firms, labor market frictions) Two types of equilibria (separating and pooling) Acemoglu presents evidence that middling jobs have been disappearing and have been replaced by the types of jobs that would be offered in a separating equilibrium Conclusion Our logic suggests that openness could cause this in exporting industries, and have the opposite effect in import-competing industries Acemoglu does not break his industries up into groups based on trade position nor does he control for openness – we should try this Conclusion Abowd and Kramarz (2004) test for positive assortative matching in France and the US using linked employer-employee data They find little or no evidence in favor of the theory that “good” workers match with “good” firms Conclusion Our theory implies that the trade position of an industry matters: positive assortative matching is more likely in export-oriented markets Abowd and Kramarz do not control for the effects of international competition; we hope to do so Conclusion Note also: Abowd and Kramarz are looking for the type of matching (good firms with good workers versus good firms with weak workers) With imperfect labor markets, get some mismatch We argue that trade affects the extent of mismatching – trade alters the degree of assortative matching