Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Wage gains from foreign ownership: evidence from linked employer–employee data

Wage gains from foreign ownership: evidence from linked employer–employee data We compare the wages of skilled workers in multinational enterprises (MNEs) versus domestic firms, the earnings of domestic firm workers with past, future and no MNE experience, and estimate how the presence of ex-MNE peers affects the wages of domestic firm employees. The analysis relies on monthly panel data covering half of the Hungar - ian population and their employers in 2003–2011. We identify the returns to MNE experience from changes of owner- ship, wages paid by new firms of different ownership, and the movement of workers between enterprises. We find high contemporaneous and lagged returns to MNE experience and significant spillover effects. Foreign acquisition has a moderate wage impact, but there is a wide gap between new MNEs and domestic firms. The findings, taken together, suggest that MNE employees accumulate partly transferable knowledge, valued in the high-wage segment of the local economy that is connected with the MNEs via worker turnover. Keywords: Multinational enterprises, FDI, Wage differentials, Wage spillover, Hungary JEL Classification: F23, J24, J31, O33 1 Introduction Heyman et  al. 2007; Andrews et  al. 2007; Malchow- While policymakers in developing countries are often Moller et  al. 2007). An adverse competition effect often criticized for ‘selling out’ the country to foreigners, FDI offsets the positive direct impact of FDI on productiv - can actually bring valuable knowledge to a less developed ity and wages even in relatively undeveloped economies economy, spreading through labor mobility channels. (Aitken and Harrison 1999; Djankov and Hoekman 2000; Undeniably, corporate revenues can find their way back Konings 2001; Barry et  al. 2005). The positive spillovers home via profit repatriation and transfer pricing, and many are often restricted to specific sectors (Keller and Yeaple MNEs enjoy a generous initial tax holiday. However, MNE 2009; Suyanto and Bloch 2014; Fons-Rosen et  al. 2017). workers’ wage premium over similar domestic-sector Still, the existence of a vast MNE premium in the emerg- employees in comparable firms directly benefits society, ing and transition economies (Lipsey and Sjöholm 2004; especially if the underlying excess productivity is portable OECD 2008a; Chen et al. 2017), and the findings of posi - and exerts positive spillover effects. Unlike the returns to tive spillovers (Smarzynska-Javorcik 2004; Görg and capital investment and part of the profit, the wage surplus Strobl 2005; Kosová 2010; Poole 2013; Gorodnichenko predominantly remains and is spent in the host country. et  al. 2014) encourage us to seek evidence for a ‘knowl- The literature provides ample evidence to call into edge flows’ scenario. To assess the magnitude of the question the general validity of such an optimistic sce- potentially beneficial impact of FDI, we study the direct nario. The foreign-domestic wage gap is negligible in and indirect wage effects of work experience in multi - countries close to the productivity frontier (Balsvik 2011; national enterprises (MNEs) using linked employer– employee data on skilled workers in Hungary, 2003–2011. We contribute to the literature by empirically show- ing in a single study that (i) MNEs pay markedly higher *Correspondence: bozaistvan@gmail.com Central European University, Budapest, Hungary wages than similar domestic firms. (ii) MNE employees Full list of author information is available at the end of the article Section  2 provides a detailed introduction to the previous literature includ- ing the sources cited here. © The Author(s) 2021. This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/. 3 Page 2 of 21 J. Köllő et al. lose a part of their wage advantage upon leaving the the effect it identifies is unsuitable for out-of-sample pre - foreign-owned sector. (iii) Even so, they earn more than diction. Only 5.3 percent of the observed firms changed their colleagues in domestic enterprises. (iv) Domestic the majority owner during the observation period in our firm employees benefit from having ex-MNE peers. We sample. These companies paid significantly higher wages interpret the coincidence of the MNE premium, par- than ‘always domestic’ firms (when they were domes - tial wage loss from separation, lagged returns to MNE tic) and significantly lower wages than ‘always foreign’ experience, and wage spillovers as a signal of knowledge companies (when they were foreign-owned): this is how transfer from MNEs to domestic firms. While alternative the 2FE model arrives at a close-to-zero estimate of the explanations exist for each of the presented symptoms, ownership-specific wage gap. These firms’ experience can in the last section of the paper we argue that a ‘knowl- hardly predict how big MNEs like Mercedes-Benz or IBM edge flows’ scenario has the best chance to produce all of would pay their employees in the unlikely event of takeo- the four outcomes. ver by a local business person. It also tells nothing about Regarding methodology, we draw attention to the dif- the potential wage gains from greenfield investments, ficulties of identification coming from the non-random which played a significant role in the 1990s (Calderon selection of firms into foreign ownership and of differ - et al. 2004). We utilize a difference-in-difference estima - ently skilled workers into foreign enterprises. We find tion of wage gains from joining a new MNE over joining a trade-offs between model quality and unbiasedness of the new domestic firm to learn about the ownership-specific samples on which the first-best models can be estimated. wage gap between ‘always foreign’ and ‘always domestic’ The analysis is based on a big administrative panel companies. This approach suggests that the employees data set covering half of the Hungarian population and of new MNEs earn 15 percent more than their domestic their employers in 2003–2011. We restrict the analysis counterparts. to skilled workers for three reasons. First, the traces of Turning to the MNE premium’s portability, we have to knowledge transfer are easier to find in the skilled labor deal with endogeneity and ability biases, as worker mobil- market. Second, data discussed later suggest that a part ity is not random. If a worker is fired from an MNE, it of the MNE premium compensates unskilled workers for may be because her marginal product is lower than aver- non-wage disamenities like shift work, weekend work, age. If a domestic employer attracts a worker, it may be and a higher probability of becoming unemployed. The because she has a higher-than-average marginal product data does not indicate ownership-specific differences of irrespective of the sector of employment. To address the this kind among highly skilled workers. Third, repeat - first problem, we compare domestic firm employees with ing the estimations for middling and unskilled workers recent MNE experience to their peers who had outside would triple the statistics to be presented, with minimal experience in the domestic sector. We focus on workers added content. Estimation on a pooled sample would losing or leaving their jobs in times of mass dismissals only attenuate the relevant parameters. when separations are more likely to be exogenous to the We start by estimating the foreign-domestic wage individual worker’s productivity. The model controls for gap using panel regressions. By gradually removing the heterogeneity of the sending firms via observable con - effects of observed and unobserved worker and firm trols and use fixed effects for the receiving ones. We find characteristics, we get from a substantial raw gap of 0.75 that former MNE employees earn more by 13 percent log points to 0.24 points after controlling for worker fixed than similar workers coming from collapsing domestic effects and a mere 0.03 points’ pure ownership-specific enterprises. Workers separating from their employers for wage differential estimated with both worker and firm reasons other than mass dismissals acquire a significantly fixed effects (2FE henceforth). lower (5 percent) lagged MNE premium. While a 2FE model can answer how an existing firm’s Satisfactory model quality comes at the cost of distor- wage level changes in response to a change in ownership, tions in the sample and a significant loss of observations in this case, too. Only about 7 percent of the person- months in our data makes it to the estimation sample of a model in which work histories and characteristics MNEs may pay high wages to skim the cream of the labor force, buy loy- of the sending and receiving firms are adequately con - alty, contain turnover, stimulate work effort, or prevent information leakage. Workers’ wages may fall upon leaving the MNE sector for losing these wage trolled. The problem would be further aggravated by components and because employers perceive their dismissal as a negative sig- nal. Ex-MNE workers may have high wages in domestic firms because they have high reservation wages and belong to the lucky few to find a well-paying job in the domestic sector. Spillover effects may arise from the employer’s wish to keep within-job wage differentials within tolerable limits. Antalóczy and Sass (2001) estimate that the share of greenfield FDI in total We justify this choice and present some results on less skilled workers in inward FDI amounted to 25–30 percent in Hungary and other CEE countries Sect. 7. during the transition. Wage gains from foreign ownership: evidence from linked employer–employee data Page 3 of 21 3 the inclusion of worker fixed effects to reduce ability 2 Previous findings on the foreign‑domestic wage bias. To avoid this issue while utilizing more data and gap, lagged returns and spillovers still controlling for the potential bias, we rely on a less Estimates of the foreign-domestic wage gap vary widely, demanding ‘overlapping cohorts’ model that compares with the MNE premium found to be nearly negligible in domestic firm employees with future and past experience the most developed market economies. In Norway, the in foreign versus domestic firms. This model can utilize a OLS estimate by Balsvik (2011), controlled for worker much broader sample, as workers with only two observed and plant characteristics, amounts to 3 percent, which spells can contribute to the estimation if any of those is at falls to 0.3 percent once she includes worker fixed effects. a foreign-owned employer. The estimated return to prior An OLS estimate for Sweden by Heyman et  al. (2007) MNE experience amounts to 0.07 log points. is even lower at 2 percent. Andrews et  al. (2007) and Finally, we estimate spillover effects for incumbent Malchow-Moller et  al. (2007) detect positive gaps in the domestic firm employees, controlling for observed and range of 1 and 3 percent in Germany and Denmark. The unobserved worker and firm characteristics. We deviate OLS estimate of Martins (2004) for Portugal is higher (11 from a similar attempt by Poole (2013) in two ways. First, percent), but he finds that the MNE wage premium vir - we also study how skilled incumbents’ wages respond tually disappears after controlling for worker selection. to the presence of less qualified ex-MNE peers. Second, These figures compare to 32 percent (pooled OLS for and more importantly, we address the selection prob- all skill levels) and 13 percent (after adding worker fixed lem that arises when the analysis is restricted to incum- effects) in our sample. Workers moving from domestic to bents (domestic workers with no experience outside foreign-owned firms are estimated to gain 6 percent in their firms). Incumbents in our data account for only 22 Germany and 8 percent in Norway (Andrews et al. 2007; percent of the workers ever employed in the domestic Balsvik 2011), which compares to 53 percent in the Hun- sector. Their exposure to peers with MNE experience dif - garian sample for all skill levels. fers substantially from that of the average worker. In an The foreign-domestic gap is much broader in less alternative specification, we ensure the identification of developed countries: according to raw data presented in within-firm spillovers using a 2FE model. We find that Lipsey and Sjöholm (2004), in Indonesian manufactur- a one-standard-deviation difference in the share of high ing, the MNE premium amounts to 47 percent for blue skilled ex-MNE peers shifts peers’ wages with no MNE collars and 55 percent for white collars (41 and 73 per- past up by slightly more than one percent. Having quali- cent in Hungary). Chen et  al. (2017) report a gap of 40 fied peers with outside experience in the domestic sec - percent in Chinese manufacturing. An overview of data tor and having low-skilled peers with MNE experience do in OECD (2008a), based on the World Bank Enterprise not affect wages. Survey, indicates raw gaps of between 40 and 50 percent Section  2 discusses previous findings on the paper’s in Africa, Asia, the Middle East, and combining all these topic and prewarns the reader of our estimates. Section 3 regions and adding Central and Eastern Europe. introduces the data and the local context. Section  4 is A more detailed analysis of the sources of the gaps in devoted to the study of the foreign-domestic wage gap. Germany, Portugal, the UK, and Brazil (OECD 2008b) Sections  5 and 6 present the results on lagged returns finds that takeovers’ marginal effect on wages falls short and spillover effects, respectively. Section  7 briefly com - of 3 percent in all of these countries. Results from Hun- ments on differences by skill levels and industries. Sec - gary point to similar patterns. Csengődi et al. (2008) use a tion  8 sums up the results and argues that the empirical different data set from ours (the Wage Survey, a repeated findings, taken together, yield support to a ‘skills diffu - cross-section LEED which allows the linking of firms but sion’ scenario. not workers) and find that after adding firm fixed effects, the MNE wage premium falls to a mere 3 percent as it does in our case. Earle et  al. (2017) use the same data 5 Note that in the Norwegian case, workers moving from MNEs to domestic With the requirement of controlling for lagged size changes, we would need firms also acquire a gain of 7 percent, while in our sample they lose 11 per - workers with at least four employment spells in a nine-year-long period, with cent. The median loss amounts to 26 percent in the case of skilled workers. a specific pattern DDFD, where F and D stand for foreign-owned and domes - See Table 2. tic firms. Identification in this setting would come from comparing the sec - ond and fourth domestic job entries. The third, F spell is the treatment, and In the Czech Republic, Jurajda and Stančík (2012) detect sigificantly faster a first spell is required for the inclusion of firm sizes. Besides, in this setting wage bill growth in (and only in) manufacturing firms with a low export the ex-MNE spells would sistematically happen later on in worker’s career, so share. They cannot decompose the wage bill effect into wage and employ - life-cycle wage changes may be potentially captured by the parameter as well. ment effects. They also show that domestic firms subject to foreign acquisition pay Which is a lower bound as in this model, we do not control for employ- higher-than-average wages already before the takeover, hinting at a non- ment change in the sending firm. random selection to foreign buy-out. 3 Page 4 of 21 J. Köllő et al. source and detect a slightly higher premium of 7 percent At the same time, several studies have identified posi - that is still very far from the estimates they get with- tive spillovers. Using Lithuanian data, Smarzynska-Javor- out controlling for unobserved firm characteristics and cik (2004) detects positive productivity spillovers from firm-specific trends. The effects identified using data on MNEs to local suppliers. Similarly, Gorodnichenko et al. worker mobility by OECD (2008b) are more substantial: (2014) find that backward linkages positively affect the the estimates vary between 6 and 8 percent in Germany productivity of domestic firms (while horizontal and for - and the UK, more than 10 percent in Portugal, and 20 ward linkages show no consistent effect) in 17 transition percent in Brazil. The authors argue that the discrepancy countries. Using Czech data, Kosová (2010) demonstrates between the estimates based on takeovers versus worker that crowding out is short-term: after an initial shock, flows are explained by foreign firms’ propensity to share domestic firm growth accelerates, and survival rates their productivity advantage more extensively with new improve. Görg and Strobl (2005) show that entrepreneurs workers than with workers who do not change firms. We with MNE experience start more productive small busi- believe that the difference instead roots in the non-ran - nesses in Ghana. Bisztray (2016) found that new entrants’ dom selection of firms to acquisition, as will be discussed growth in productivity was significantly higher when in more detail later. located close to Audi and operated in a supplier industry. To our knowledge, Balsvik’s paper is the only one esti- Importantly, from this paper’s point of view, Poole mating the wage advantage of ex-MNE employees in (2013) estimates that the wages of incumbent domes- domestic firms. She identifies a premium of 6.9 percent tic firm employees in Brazil rise by about 0.6 percent if for workers with three or more years of tenure in an the share of ex-MNE employees increases by 10 percent, MNE. However, she also detects an advantage of 3.3 per- while the effect of outside experience in local firms is cent on the part of workers arriving from local firms, sug - about ten times weaker than that. While the effect she gesting a net benefit from MNE experience of 3.6 percent estimates is not particularly strong, it is statistically sig- (and smaller advantages in case of shorter completed nificant at conventional levels. tenure in the previous job). We find that domestic firm One can also find indirect evidence on spillovers, con - employees, who left an MNE because of mass dismissals, sidering that MNEs are more productive and more likely closure, or relocation earn more than their ex-domestic to export and engage in R&D. Stoyanov and Zubanov counterparts by 13 percent. (2012) show that (in Denmark) workers coming from The empirical evidence on wage and productivity spill - more productive firms experience productivity gains. overs are mixed. Starting with papers that depict a not Similar results are presented for Hungary by Csáfordi too rosy picture of how MNEs affect the rest of the econ - et  al. (2018). Mion and Opromolla (2013) show that omy, Aitken and Harrison (1999) and Djankov and Hoe- export experience implies higher export performance kman (2000) identify a positive direct effect of foreign and a sizable wage premium for Portuguese managers, ownership on productivity in Venezuela and the Czech who leave for non-exporters. In Finland, Maliranta et al. Republic, but negative spillovers. Konings (2001) sug- (2008) identify positive impact of hiring workers with gests that the adverse competition effect is stronger than previous R&D experience to non-R&D jobs. the positive direct productivity effect of FDI in Bulgaria, Romania, and Poland. Barry et  al. (2005) found that for- 3 Data and the local context eign presence in a sector hurts wages and productivity in 3.1 Data sources domestic exporting firms in the same industry (but does Our estimation samples have been drawn from a big lon- not affect wages in domestic non-exporters) in Ireland. gitudinal data set covering a randomly chosen 50 percent Fons-Rosen et  al. (2017) conclude that in six advanced of Hungary’s population aged 5–74 in January 2003. Each European countries, positive spillovers are restricted person in the sample is followed monthly, from Janu- to sectors where domestic enterprises are technologi- ary 2003 until December 2011, or exit from the registers cally close to MNEs. Suyanto and Bloch (2014) find the for death or permanent out-migration. The data collect opposite in Indonesia. Keller and Yeaple (2009) detect information from the Pension Directorate, the Tax Office, significant worker-level wage spillovers only in high-skill- the Health Insurance Fund, the Office of Education, and intensive industries in US manufacturing. By looking at the Public Employment Service. We use information existing firms in an Audi plant’s supplier industries in on the highest paying job of a given person in a given Hungary, Bisztray (2016) finds no positive effect on pro - month, days in work, and amounts earned in that job. ductivity. She also finds that firms with foreign owners Throughout the paper, we use daily wages (the monthly account for all the positive impact on sales and employ- value divided by days in work) normalized for the given ment, suggesting a foreign-to-foreign complementarity month’s national average. We have data on occupation, rather than a galvanizing effect on the domestic sector. type of employment relationship, registration at a labor Wage gains from foreign ownership: evidence from linked employer–employee data Page 5 of 21 3 office, receipt of transfers, and several proxies of the a yearly basis would impair the precise measurement of person’s state of health. We do not observe educational tenure and the time between two jobs—essential controls attainment—this is approximated with the person’s high- in the analysis of lagged returns. Third, higher observed est occupational status in 2003–2011. The data on firms mobility helps in identifying firm and person effects. The come from the annual tax reports of businesses obliged problem raised by inflating observations at the same firm to conduct double book-keeping. The firm-level variables is taken care of by the worker and firm-level clustering of are merged into the respective person-month observa- errors. tions. We regard a firm as MNE if foreigners’ share in subscribed capital exceeds 50 percent.3.2 MNEs in Hungary We restrict the analysis to skilled workers employed In the first decade after the start of the transition, Hun - at least once in a foreign or domestic private enterprise gary was the most successful country within the former the employment level of which exceeded the ten work- Soviet bloc in attracting foreign capital. By 2003, the ers limit at least once in 2003–2011. We have several beginning of our observation period, cumulative FDI reasons to set a size limit. First, foreign firms are nearly inflows exceeded 40 percent of the GDP, multination- absent in the small firm sector. Second, financial data als employed 15 percent of the labor force (including self- are not available for sole proprietorships and unincor- employment and the public sector into the denominator) porated small businesses. Third, the financial reports of and more than 30 percent of private-sector employees. incorporated small firms are often incomplete and erro - They produced 20 percent of the GDP and delivered over neous. Finally, the earnings data of small firms are flawed two-thirds of the exports (Balatoni and Pitz 2012). Large by paying “disguised” minimum wages. Small firms’ multinationals, including Audi, General Motors, and inclusion would also raise the risk of measurement error Suzuki, dominated the motor industry. Foreign presence in the analysis of spillover effects since the probability was already significant in the tobacco, leather, chemi - of not observing an ex-MNE employee in a 50-percent cal, rubber, and electronics industries, with employment sample is much higher in small establishments. We itera- shares of between 50 and 80 percent. tively removed workers and firms with less than two data Almost three-fourths of the cumulative FDI inflows points, zero wages, and missing covariates. have arrived in sectors outside of manufacturing. As After these steps of data cleaning, we are left with shown in column 4 of Table  1, nearly 60 percent of the a sample of 19,961,622 person-months belonging to skilled employees within the MNE sector worked in the 344,203 skilled workers and 119,580 firms. 52.6 percent tertiary sector. Therefore, we do not restrict the analy - of the workers had at least one spell of employment in the sis to manufacturing, as most papers do in the strand of foreign-owned sector, of which 21.5 percent worked only the literature we follow (see Barry et  al. 2005; Görg and in MNEs. We draw special sub-samples from this start- Strobl 2005; Lipsey and Sjöholm 2004; Smarzynska- ing population for the study of new firms, lagged returns Javorcik 2004; Balsvik 2011 as opposed to Poole 2013, and spillover effects. Descriptive statistics are presented whose study covers all sectors in Brazil). While FDI typi- in Table 11 of Appendix 1. cally boosts exports and generates demand for domestic Even though our firm-level variables are of annual fre - manufacturers producing intermediate goods, its contri- quency, we prefer to analyze the data at a monthly level bution to the quality of retail trade, banking and services for several reasons. First, the affiliation of a worker can - can be equally important, especially in the former state- not be precisely measured on a yearly basis. About 25 socialist countries, which started the transition with percent of the workers employed by an MNE for at least critically undeveloped non-tradable sectors. The foreign- one month in a given year also had one or more spells in owned and domestic parts of the economy are closely the domestic sector in the same year. Second, turning to connected via labor turnover. In the skilled labor market, 37.2 percent of the domestic firms, employing 69 percent of the domestic labor force, hired at least one ex-MNE worker in 2003–2011. See Appendix 2 for variable definitions. Setting the limit elsewhere does not affect the results, since 93 percent of the firms with nonzero foreign presence are majority foreign-owned. 3.3 Descriptive statistics on wages and wage change In 2014, MNEs had a 4.5 percent employment share in the 1–10 work- Table  2 presents raw statistics on wage levels across ers category (Authors’ calculation based on the 2014 Q4 wave of the Labor ownership categories and wage changes associated with Force Survey). skilled workers’ shifts between them. The data shows vast This term hints at the practice of paying workers the minimum wage (subject to taxation) and the rest of their remuneration in cash. Elek et  al. (2012) estimate that in 2006 the share of workers paid in this way amounted to 20 percent in firms employing 5–10 workers, 10 percent in slightly higher firms (11–20 workers) and less than 3 percent in larger enterprises. UNECE (2001), p. 190. 3 Page 6 of 21 J. Köllő et al. Table 1 Foreign ownership in Hungary, 2003 Fraction employed in MNEs (percent of all person- Industrial composition of MNEs (percent months in the given industry) of all person-months in the MNE sector) All workers Skilled workers All workers Skilled workers Agriculture 5.0 6.1 0.8 0.5 Manufacturing 46.5 48.4 59.9 40.5 Construction 7.7 10.6 1.5 1.9 Energy, water, gas 57.5 55.6 3.3 3.1 Wholesale and retail trade 25.9 34.5 16.3 31.5 Finance and insurance 52.7 80.0 11.4 11.5 Services 20.7 24.3 6.8 11.0 Average/total 34.8 37.6 100.0 100.0 The data are annual averages observed in the estimation sample in 2003. The number of person-months amount to 8,704,486 (all workers) and 2,068,556 (skilled workers) differences between workers in MNEs versus domestic dummy for exporters. Alternatively, we use indicators of firms, on the one hand, and domestic firm employees investment and productivity. We gradually move from an hired from MNEs versus workers coming from other OLS equation only controlled for s to fixed-effects mod - jt domestic enterprises, on the other. els with all the covariates except for the P variables. According to the raw data, MNE employees earn more When the equation is estimated with OLS, the δ than twice as much as domestic sector workers. Persons parameter captures the ownership effect, plus the moving from domestic firms to MNEs gain 64 percent - employment-duration weighted average residual worker age points on average, while individuals who move to the and firm effects given personal characteristics P and X other direction lose 57 points. Measured with the median (Abowd et  al. 2006). The person fixed effects absorb the rather than the mean, the gain and the loss amount to 39 unobserved time-invariant mean “qualities” of work- and − 26 percentage points, respectively. The bottom ers. However, the estimated gap is still affected by the block suggests a substantial raw premium for outside experience in foreign-owned enterprises. In the forth- coming sections, we try to disentangle a ‘pure’ owner- Table 2 Descriptive statistics: wage levels and  wage ship-specific effect from differences in composition. changes of skilled workers Mean St. dev. Observations 4 Foreign‑domestic wage gap 4.1 Benchmark model Wage levels Our first model estimates the foreign-domestic wage gap Employer = MNE 309 288 7,937,675 in the following way: Employer = domestic firm 143 161 12,023,947 Wage change upon leaving an MNE for a domestic firm ln w = δF + [ϕP ] + αX + βY + γ V ijt ijt i it ijt jt Mean − 57 146 42,479 (1) + v + f + s + ε , i j jt ijt Median − 26 42,479 Wage change upon leaving a domestic firm for an MNE where w is the daily average (relative) earnings of per- ijt Mean 64 126 46,590 son i at firm j and month t , F is a dummy for being Median 39 46,590 employed in a majority foreign-owned firm, P and X are i it Wages of domestic firm employees with recent outside experience fixed and time-varying individual attributes, Y stands ijt Previous employer = MNE 171 193 963,075 for job-specific variables (like occupation and tenure), V jt Previous employer = domestic firm 118 122 3,557,788 denotes time-varying firm-specific covariates, v and f i j Wage in month t relative to the national average wage in month t, per cent are worker and firm fixed effects, respectively, and ε is ijt Person-months observed in 2003–2011 an error term. We allow for unobserved shocks to pro- The figures relate to persons moving from MNEs to domestic firms and vice ductivity by including sector–year interactions s . The jt versa. Mean earnings in the receiving firm is compared to the same worker’s firm-level variables are size, the capital-labor ratio, and a mean earnings in the sending firm. Wages are deflated with the national average wage in the same month The figures relate the mean earnings of domestic firm employees with previous outside experience to the mean earnings of incumbent domestic firm See Appendix 1: Fig. 2 for a box-and-whiskers plot of wage changes. employees, percent Wage gains from foreign ownership: evidence from linked employer–employee data Page 7 of 21 3 Table 3 Estimates of the foreign‑ domestic wage gap for skilled workers, 2003–2011 Specifications: (1) (2) (3) (4) (5) (6) Model A Foreign-owned 0.745 (25.5) 0.763 (20.3) 0.718 (31.6) 0.437 (23.4) 0.236 (28.6) 0.026 (3.5) aR2/within R2 0.260 0.329 0.414 0.480 0.238 0.103 Model B Always foreign-owned 0.794 (25.7) 0.817 (29.5) 0.772 (31.3) 0.507 (23.3) 0.307 (26.8) Temporarily foreign-owned 0.569 (8.1) 0.574 (8.7) 0.564 (12.6) 0.334 (5.2) 0.209 (14.1) 0.026 (3.5) Temporarily domestic 0.408 (10.5) 0.408 (10.5) 0.462 (11.3) 0.269 (8.7) 0.157 (11.6) aR2/within R2 0.267 0.327 0.422 0.482 0.242 0.103 Controls Sector × year Yes Yes Yes Yes Yes Yes Person No Yes Yes Yes Yes Yes Job No No Yes Yes Yes Yes Firm No No No Yes Yes Yes Person FE No No No No Yes Yes Firm FE No No No No No Yes All coefficients are significant at 0.01 level, t-values in parentheses. The standard errors are adjusted for clustering by persons and firms. Sample: 19,961,622 person- months belonging to 344,203 skilled workers in 119,580 firms. Singleton observations are excluded from the panel regressions Dependent variable: log daily wage in the given month relative to the national mean. Reference categories: employed in a domestic firm (Model A), employed in an’ always domestic’ firm (Model B). Controls: person, job and firm characteristics plus sector–year interactions. See Appendix 1: Table 12 for variable definitions. Specifications 5 and 6 include only time- varying covariates and worker and firm fixed effects. Estimation: all models were estimated with Stata’s reghdfe models employment-duration weighted average of the firm inclusion of firm size, the capital-labor ratio, and exports effects for the firms in which the worker was employed. bring the estimated MNE premium down to 0.437, while When both person and firm fixed effects are included, δ adding worker fixed effects reduces it to 0.236. Adding captures a pure ownership effect identified from worker firm fixed effects results in a major drop to only 0.026. flows between ownership categories, on the one hand, Controlling the worker fixed-effect model for TFP or and changes in ownership, on the other. It shows the value-added per worker instead of the firm fixed effects wage advantage of a foreign firm employee over a domes - yield estimates of 0.218 and 0.206, respectively. Includ- tic worker with similar observable attributes, controlled ing TFP into the set of firm controls in specification (4) for their average wages in the entire period of observa- results in a coefficient of 0.209. Including investment as tion, and also controlled for average wages of the firms well, which controls for the potential coincidence of posi- where they worked during the period of observation. tive productivity shocks and the hiring of high-quality For our multiple fixed effect estimations, we use a model labor, produces an estimate of 0.216. By contrast, adding proposed by Correia (2017), and implemented in Stata as firm fixed effects to specification (4) without including reghdfe. worker fixed effects decreases the estimate from 0.437 In Model A of Table  3, which measures MNE employ- to 0.036, clearly indicating that selection to acquisition ees’ wage advantage relative to domestic firm employees, drives the result of the 2FE model. the estimate rises from 0.745 log points to 0.763 after In Model B of Table  3, the observed person-months controlling for observed worker characteristics. The are classified by the ownership histories of employers. ‘Always domestic’ (the reference category) and ‘always foreign’ denote enterprises that did not change owner in 2003–2011. ‘Temporarily foreign’ and ‘temporar- The only exception would be observations on firms that, at the same time as changing ownership, would change all of their employees. We do not have ily domestic’ indicate the current majority owner of the such cases in the data. workers’ employer, for firms which underwent acquisi - Several methods have been developed in the last ten years (following tion at least once in 2003–2011. The ‘temporary foreign’ the pioneering work of Abowd et  al. 1999) to deal with two or more high dimensional fixed effects. The iterative methods (Cornelissen 2008; Martins dummy, for instance, is set to one for a person-month and Opromolla 2009; Guimaraes and Portugal 2010; Carneiro et  al. 2012; spent in a foreign-owned enterprise, which operated Mittag 2016) solve the problem by shuffling between the estimation of the slope and the intercept parameters. Balázsi et al. (2018) yield an alternative, which presses more on memory but runs faster. Early drafts of this paper, like Balázsi (2017) experimented with this method. With the size of the final data, iterative approaches turned out to be more productive. 3 Page 8 of 21 J. Köllő et al. under domestic ownership in a part of the observed The double difference (β − β ) − (β − β ) removes the 3 4 1 2 period. gap in the quality of F and D workers as measured with The estimates suggest that firms involved in takeovers their pre-entry wages. Since assignment to the groups and currently operating under domestic ownership pay compared is person-specific, and the firms do not change more than incumbent domestic firms (by 0.157 log points owner, we estimate the equation with pooled OLS. A in specification 5 where worker quality is controlled for). large battery of controls guarantees that we compare Switching firms currently under foreign ownership pay workers and firms with similar characteristics. lower wages than always foreign-owned companies by Note that we base the definition of a ‘new firm’ on its 0.099 log points. The gap between the coefficients for employment dynamics rather than its date of registration employment spells under ‘temporarily foreign’ and ‘tem- since the latter is often associated with break-ups, merg- porarily domestic’ ownership (0.052 log points) is an ers and acquisitions, rather than the birth of a new eco- alternative measure of how ownership changes affect the nomic actor. We rely on the fact that a medium-sized or wage. The magnitudes make it clear that switching firms large firm’s creation typically begins with hiring a small substantially differ from any of the incumbent categories. group of managers who arrange the start-up. This pre - paratory stage is followed by a ‘big bang’ when the firm 4.2 Exploiting inf ormation on new firms hires rank-and-file employees. We speak of a big bang As much as 94.8 percent of the firms in our estimation when a firm’s staff jumps from an initial level of L ≤ 5 t−1 sample did not change majority owner in the period cov- to L ≥ 50, or, from L ≤ 50 to L ≥ 300 within a month. t t−1 t ered by the data: 7.3 percent was foreign-owned, and 87.5 We found 519 such firms with no subsequent change of percent was domestic throughout 2003–2011. Rather ownership. Combined employment in these enterprises than merely neglecting the huge wage difference between jumped from 13 to 253 thousand (see Appendix 1: Fig. 3). them (as does the 2FE model), we exploit information Finally, the sample consists of 471,489 person-months on newly established and subsequently incumbent for- belonging to 8225 skilled workers hired by and staying eign and domestic enterprises. The critical event under until December 2011 in 366 new domestic and 147 new examination here is not the takeover of an existing firm, foreign-owned firms. but the birth of an incumbent firm. The analysis relates The results in Table  4 indicate a wage gap of 0.391 log to firms established after 2003 and staying under major - points between skilled workers in new MNEs versus new ity foreign or domestic control until 2011. We compare domestic firms—this is reasonably close to the 0.437 log the earnings of incumbent workers in these firms to the points gap estimated with a fully controlled OLS for all wages they earned before their entry. Formally, we esti- firms in Table  3, specification 4. New foreign firms’ work - mate the following difference-in-difference model: ers also earned more than their domestic counterparts before they entered the new firms by 0.245 log points on 0 0 1 1 ln w = β F + β D + β F + β D + Zγ + ε . ijt 1 2 3 4 ijt ijt ijt ijt ijt average. After deducting this difference from the post- (2) entry gap, an ownership-specific wage differential of 0.146 log points remains between incumbent workers in F and D are the acronyms for foreign-owned and domes- 0 0 incumbent firms. This point estimate falls between the tic firms. F and D are set to one for person-months pre- individual only and the two fixed-effects parameters, sug - ceding the worker’s entry date to a newly established F or 1 1 gesting a significantly stronger pure ownership-specific D firm. F and D are set to one for the months of service effect than the 2FE model. in a newly established firm. For instance, for a worker hired by a new foreign-owned company in month t = 37, 0 1 5 Lagged returns F = 1 if t < 37 and F = 1 if t ≥ 37. Z denotes controls 5.1 Are ex-MNE workers paid more in the domestic sector? listed in the notes to Table 6. In Eq.  (2), we compare workers in domestic firms, who β − β is the estimated wage difference between future 1 2 arrived at their employer from MNEs versus other F and D employees, whereas β − β measures the wage 3 4 domestic firms. The estimates are controlled for personal difference between the employees of new F and D firms. characteristics, current and past job attributes, tenure in the last job, months between the two jobs, selected indi- cators of the sending and receiving firms, and sector-year interactions. We retain firms with at least one ex-MNE Model B with added firm effects (Specification 6) is identical to Model B, as and one ex-domestic employee and exclude firms under - the always foreign indicator is absorbed by the added firm effects. Hence only going acquisition. a parameter on being temporarily foreign-owned can be estimated, which is identified only from firms going through acquisitions or divestments and accordingly, coincides with the parameter on the F dummy of Model A. Wage gains from foreign ownership: evidence from linked employer–employee data Page 9 of 21 3 Table 4 Wages before and after entry to new MNEs and new domestic firms Coeff. t-test Person-months Workers of domestic start-ups, before their entry 0 115,443 Workers of foreign start-ups, before their entry 0.245*** 8.3 146,585 Workers of domestic start-ups, after their entry − 0.014 0.3 84,018 Workers of foreign start-ups, after their entry 0.391*** 8.9 125,247 Double difference (point estimate, F-test) 0.146** 9.5 Significant at the **0.05, ***0.01 level. The t-values are based on standard errors adjusted for clustering by persons and firms OLS regression with dummies standing for the four distinct groups. Dependent variable: log daily wage in the given month relative to the national mean. Sample: 471,489 person-months belonging to 8225 skilled workers hired by and staying until December 2011 in 519 newly established firms (366 domestic and 147 foreign- owned). We considered a firm newly established if its staff number jumped from less than 5 to more than 50, or, from less than 50 to more than 300 within a month. Workers employed by new firms before their ‘big bang’, workers leaving the new firms and firms changing owner after the big bang are excluded. Controls: person, job and firm characteristics and sector-year interactions. See Appendix 1: Table 12 for variable definitions The lower blocks of the table display estimates on two ln w = αX + β F_After + β dL ijt it 1 ijt 2 jt (3) sub-samples distinguished along dL . Former MNE work- + β F_After dL + f + s + ε . 3 ijt jt j jt ijt ers who lost or left their jobs during mass dismissals ( dL < 0.5) had substantially higher wage advantages over F_After is a dummy set to 1 for workers who arrived ijt their ex-domestic counterparts (0.134 log points) than from foreign firms and 0 for workers coming from did those ex-MNE workers, who arrived from slightly domestic companies. dL = L /L measures the jt j,t+1 j,t−1 contracting, stable or expanding firms (0.06). change of employment in the sending firm between year t−1 and t + 1 , with t denoting the year when the worker 5.2 An overlapping cohorts model of lagged returns left the firm. The coefficient β measures how wages to MNE experience vary with employment dynamics of the sending domes- The estimates presented in the preceding sub-section are tic firms while the parameter β of the interaction term potentially subject to ability bias: workers returning to the F_After × dL captures the impact of dL on workers ijt jt domestic sector can be more productive wherever they arriving from foreign employers. The wage advantage of work. As it was put forward in the Introduction, addressing workers coming from MNEs over workers arriving from this problem by adding worker x fi ed effects to model (2) is domestic firms, conditional on employment dynamics of not a feasible option. Therefore, we estimate an alternative the sending firm, is given by β + β dL . Alternatively, we 1 3 jt model that compares the wages of domestic firm employees estimate the equation for two groups distinguished based with past and future experience in MNEs versus domes- on dL (lower or higher than 0.5), without the size-change tic companies other than their current employer. This and interaction terms. approach is close in spirit to models that study the wage Since we are interested in the within-firm wage dif - effect of incarceration by comparing past and future con - ferences between ex-MNE and ex-domestic entrants victs (Grogger 1995; Pettit and Lyons 2009; LaLonde and (rather than how a worker’s wage changes upon entering Cho 2008; Czafit and Köllő 2015) under the assumption a domestic firm), we include firm fixed effects, but not that the date of incarceration (mutatis mutandis the dates worker fixed effects. of entry to and exit from MNEs) can be treated as random. The upper block of Table  5 shows the results of the We can reasonably assume that future MNE workers are first variant of the model. The wage advantage of an ex- closer to former MNE employees in terms of unobserved MNE employee arriving from a firm where staff numbers characteristics than any control person selected from the did not change around the year of the worker’s separa- general population based on observables. A further advan- tion ( dL = 1) amounts to 0.057 log points, while it is esti- tage of this choice is a gain in sample size: 3,841,561 per- mated to be 0.074 points in case the sending firm was son-months instead of 797,261 in Model (2). closed or relocated ( dL = 0). We added a dummy indi- cating if the worker had arrived from another domestic firm but previously had some experience in one or more Workers who leave well-paying jobs in the MNE sector individually can MNEs. These workers have an advantage of 0.064 log be either negatively or positively selected. On the one hand, MNE employ- ees fired individually are likely to be less productive than the average. On the points. Only a part of these gaps results from within-firm other hand, those who manage to find a well-paid domestic job are predict - premia, as suggested by the differences between the spec - ably over-represented among voluntary quitters. The comparison of group- ifications with and without firm fixed effects. level estimates suggests that the first effect dominates: workers separating from their firms for reasons other than mass dismissals earn a lower lagged MNE premium on average. 3 Page 10 of 21 J. Köllő et al. Table 5 The wage advantage of  ex‑MNE workers in  domestic firms over  coworkers having arrived from  other domestic firms—regression estimates Firm fixed effects: No Yes Model A: entire sample Sending firm is MNE (F_After) 0.091*** (4.4) 0.077*** (4.9) Change of employment in the sending firm (dL) 0.010* (1.8) 0.011*** (3.0) Interaction term (F_After × dL) − 0.028** (2.4) − 0.024** (2.3) MNE experience before entry to the sending firm (dummy) 0.059*** (6.0) 0.038*** (5.0) Number of observations 723,421 722.913 2 2 aR /within R 0.461 0.288 Model B: workers arriving from mass layoffs and all workers Employment change in the sending firm: L /L ≤ 0.5 t+1 t−1 Sending firm is MNE 0.134*** (3.3) 0.109** (2.4) MNE experience before entry to the sending firm (dummy) 0.068*** (2.9) 0.056*** (2.7) Number of observations 153,482 153,213 2 2 aR/within R 0.479 0.277 Entire sample Sending firm is MNE 0.060*** (4.1) 0.049*** (5.1) MNE experience before entry to the sending firm (dummy) 0.058*** (6.0) 0.037*** (5.0) Number of observations 723,421 722,913 2 2 aR/within R 0.461 0.288 Significant at the *0.1, **0.05, ***0.01 level. The standard errors are adjusted for clustering by persons and firms. Sample: 723,421 person-months belonging to 96,277 skilled workers in 19,449 domestic firms, who had arrived from MNEs versus other domestic firms. 508 singleton observations are excluded from the equation with firm fixed effects. Estimation: Stata reghdfe. Change of employment in the sending firm: L /L , where t is the year of the worker’s separation. Controls: person t+1 t−1 and job controls, contemporaneous and lagged firm-level controls, as listed in Table 12. Additional controls are completed tenure in the sending firm, dummy for unobserved tenure, months between exit from the sending firm and entry to the receiving firm, one-digit sectoral affiliation of the sending and receiving firms and year dummies We define a collectively exhaustive classification mak - We measure the effect of foreign sector experience with ing a distinction between domestic firm employees with the double difference ( β − β ) − (β − β ) or equivalently 1 2 3 4 past MNE experience in month t (PF), workers with (β − β ) − (β − β ). The model controls for unobserved dif - 1 3 2 4 future but no past MNE experience (FF), workers with ferentials in worker quality as long as workers’ wages with prior experience in other domestic firms and no MNE future outside experience can be treated as a counterfactual experience (PD) and workers with future domestic sec- for the wages of workers with prior experience. However, it tor experience and none of the types mentioned earlier cannot address the possibly endogenous selection of work- (FD). Incumbent workers who had no contact with other ers to separation from their previous employers. employers in 2003–2011 constitute the reference cate- The results in Table  6 show that workers with past gory. The sample we work with consists of domestic firm MNE experience earn more by 0.112 log points than their employees in companies employing at least one worker counterparts with outside domestic experience. This dif - belonging to the categories mentioned above and one ference overestimates the returns to foreign sector expe- incumbent worker. We restrict the analysis to 2005–2009 rience since those domestic workers who are on their way to have sufficient observations on past and future experi - to an MNE also earn more by 0.043 log points than those ences outside the workers’ current firms. about to leave for another domestic employer. Ex-MNE We regress log wages on the respective dummies and workers earn more than future MNE employees by 0.048 person, job, and firm-specific controls plus sector-year log points while those with outside domestic experience interactions. Choosing incumbents as the reference cat- earn less by 0.021 log points than their counterparts, egory and denoting the controls with Z, the estimated leaving for another domestic firm later. equation with or without firm fixed effects (v ) is: Using these estimates, we can approximate the return to MNE work experience as the double-difference equal ln w = β PF + β PD + β FF ijt 1 ijt 2 ijt 3 ijt to 0.069 log points. The two models’ main results aimed (4) + β FD + v + Zγ + ε . 4 ijt j ijt at measuring lagged wage effects (Tables  5 and 6) are Wage gains from foreign ownership: evidence from linked employer–employee data Page 11 of 21 3 similar. The first model identified a 0.060 log points Table 6 Wage difference between  domestic workers with/ without outside work experience advantage on the part of the median worker coming from an MNE over a worker arriving from domestic company Dependent variable: log daily (Table 5 bottom block). wage While the main results are close to each other, some OLS Firm fixed effects details differ in the two models. The wage difference Coefficients (t-test) between workers arriving from foreign-owned versus Past MNE experience (PF) 0.060*** (4.0) 0.005 (1.0) other domestic firms appear to be more prominent here: Future MNE experience (FF) 0.012 (0.9) 0.001 (0.6) 0.112 points as opposed to 0.060 points in Table 5, model Past domestic experience (PD) − 0.052*** (5.6) − 0.030*** (7.3) B, the estimate for all workers. Second, when we rees- Future domestic experience (FD) − 0.031*** (3.5) − 0.016*** (3.6) timate the model by adding firm fixed effects (column 2 Differences by type of outside experience (F-test) of Table 6), the contrasts fade away: the within-firm wage Past MNE − past domestic 0.112*** (61.6) 0.035*** (53.4) differentials between the PF-FD groups are smaller, and Future MNE − future domestic 0.043*** (15.0) 0.017** (5.5) the double-difference drops to only 0.018 log points. Past MNE − future MNE 0.048** (6.3) 0.004 (0.4) Unlike our first model, the second one suggests that the Past domestic − future domestic − 0.021** (4.2) − 0.014*** (10.0) lagged MNE premium predominantly stems from past Double difference 0.069*** (15.0) 0.018*** (11.5) and future MNE employees’ crowding in high-wage aR2/within R2 0.453 0.342 domestic firms. Regression estimates. The reported coefficients are significant at the *0.1, **0.05, ***0.01 level. Unmarked coefficients are not significant at the 0.1 level. 6 Spillover effects Sample: 3,841,561 person-months belonging to 153,323 persons and 18,510 6.1 E ec ff t of ex -MNE peers on incumbent domestic firm firms. The sample covers domestic firm employees in firms employing at least one worker with past or future experience in foreign-owned or domestic firms, employees and one incumbent worker. The coefficients measure wage advantages relative We estimate the effect of ex-MNE peers on incumbent to incumbent workers. Observations for 2005–2009 are used. Estimation: workers’ wage, that is, for domestic firm employees who reghdfe without and with firm fixed effects. The standard errors are adjusted for clustering by persons and firms. Controls: person, job and firm controls, and did not leave their firm in the observed period. Their sector–year interactions wages are regressed on a set of controls and variables measuring the share of workers with previous outside experience within the worker’s company and skill cat- The fixed-effects panel equations summarized in egory. We deviate from Poole (2013) in that we also study Table  7 regress the log wages of incumbent skilled how skilled incumbents’ wages respond to the pres- domestic workers on the share of workers with out- MNE,uskilled ence of less skilled ex-MNE peers. Share , for side experience within the worker’s firm and skill jt instance, measures the ratio of unskilled employees with group. The estimated own effect for skilled workers recent MNE experience. in a medium-sized or large firm ( θ = 0.074) implies F3 MNE,middling MNE,skilled MNE,unskilled ln w = θ Share + θ Share + θ Share ijt F 3 F 2 F 1 jt jt jt domestic,middling domestic,skilled domestic,unskilled (5) + θ Share + θ Share + θ Share + αX D3 D2 D1 it jt jt jt + βY + γ V + v + s + ε . ijt jt i jt ijt The fact that the Hungarian administrative panel is only a 50% sample on We estimate the model including only worker fixed the individual level, has some unfortunate implications for the spillover esti- effects, which also absorb the firm effects since the esti - mates. We observe only around half of any given firm’s labour force—esti - mates relate to incumbent workers. The controls are mates instead of the actual shares. As not observing ex-MNE workers has the same, 50% probability as those with no such experience, in large firms we will identical to those used in Eq. 1. We restrict the time win- only experience extra noise in the share variables. This noise in our explana - dow to 2005–2011 to leave time for the accumulation of tory variable will attenuate the estimated θ parameters, biasing them towards es an ex-MNE stock. The equations are estimated separately zero. However, in firms with a small number of workers, if the average share of given type is also low, we may mistakenly not observe any variation in our for smaller (11–50) and larger (50+) firms, taking into variables of interest, while we should. If a firm which previously never had a consideration the higher risk of measurement error in foreign worker acquires a skilled manager with foreign experience, and we do small establishments. not observe the given person, observations at this firm will not have variation in the share of skilled ex-MNE workers, thus this firm will not contribute to the identification of our parameter of interest in our model with firm fixed The difference may stem from differences in the samples and the periods effects. Considering these two processes we not only keep solely the firms covered by the data as well as from the influence of experience in MNEs other with at least 10 employees, as in most of the paper, but also focus on larger than the sending firm. This effect is directly estimated in Table  5 but not in (50+ firms), where the (predicted) share variables are less volatile. Table 6. 3 Page 12 of 21 J. Köllő et al. Table 7 The effect of  coworkers with  recent outside  work experience on  the  wages of  skilled incumbents in  domestic firms 2005–2011 Share of coworkers with recent MNE experience within skill Share of coworkers with recent experience groups in other domestic firms within skill groups Unskilled Middling Skilled Unskilled Middling Skilled Notations in Eq. 3: θ θ θ θ θ θ F1 F2 F3 D1 D2 D3 All firms 0.012 (1.5) 0.003 (0.4) 0.042*** (3.9) 0.015** (3.1) 0.010 (1.5) − 0.031*** (-4.5) Firms employing > 50 workers 0.000 (0.0) 0.028 (1.2) 0.074*** (4.3) 0.005 (0.7) 0.042*** (2.8) − 0.027** (− 2.1) Significant at *0.1, **0.05, ***0.01 level. The t-values are based on standard errors adjusted for clustering by persons and firms. θ is significantly larger than θ , θ and F3 F1 F2 θ . Sample: 3,730,789 person-months in 116,204 firms in the full sample, 2,474,830 person-months in 77,411 firms in the 50+ sample. Dependent variable: log daily D3 wage in the given month relative to the national mean. Controls: person, job and firm characteristics, sector–year interactions, and worker fixed-effects Table 8 Mean within‑firm share of coworkers with past MNE experience (percent) Skilled incumbents in domestic firms Skilled domestic firm employees without MNE experience Share of coworkers with MNE Number of workers Share of coworkers with MNE Number of workers experience experience Unskilled 7.0 38,355 13.3 73,320 Medium skilled 9.3 53,896 15.4 103,871 Skilled 9.0 55,900 14.6 107,250 Incumbents are workers, who had only a single domestic-owned employer in 2003–2011. The mean within-firm shares are weighted with firm size and relate to 2003–2011 that a one-standard-deviation difference in the share of mean within-firm share of skilled MNE-experienced high skilled ex-MNE employees (0.18) shifts the wages peers amounts to 9 percent in the case incumbents of skilled incumbents up by 1.3 percent. Having more instead of 14.6 percent in the case of their non-incum- skilled peers with outside experience in the domestic sec- bent counterparts—a predictable pattern since incum- tor has no effect. bents are more likely to be found in firms with low In evaluating the cross effects, one should consider the labor turnover. relevant range in the share of ex-MNE workers. While a A higher share of ex-MNE peers increases the likeli- jump from zero to 50 or 100 percent in the share of ex- hood of personal contacts, thereby assisting the diffusion foreign workers within the unskilled or medium-skilled of MNE-based skills within the firm. At the same time, workforce is beyond the realm of reality, which ren- the typical incumbent worker spends more time with the ders the spillover effect to be weak, this can happen in firm, so she has a better chance to absorb the imported the high skilled category. Domestic firms employing 50 knowledge. Because of the potential bias in either direc- workers have 7 high skilled workers on average. Hiring tion, we reestimate the spillover model for all domestic two managers or professionals with foreign sector experi- workers, including firm fixed effects on top of the worker ence can increase the ex-MNE share from zero to almost fixed effects in the model to ensure that it identifies 30 percent overnight, which implies a 0.022 log points within-firm impacts. wage increase for skilled incumbents. The results for firms with more than 50 workers and all firms are presented in Table  9. Starting with the for- mer: the own effect (0.060) is slightly lower than the 6.2 R eestimating spillover effects for all domestic firm point estimate for incumbents (0.074 in Table  4). Less employees skilled ex-MNE workers exert a weak effect—the respec - Incumbents in our data account for only 22 percent of tive coefficients are only significant at the 5 percent level. the workers ever employed in the domestic sector and Having more skilled peers with recent outside experience 34 percent of the workers never employed outside the in domestic firms do not affect wages positively at all. The domestic sector. The estimates of spillover effects using estimates for all firms are much lower and insignificant at their sample may be biased because their exposure to 5 percent level. The inward bias is probably explained by peers with MNE experience differs substantially from that of the average worker. As shown in Table  8, the Wage gains from foreign ownership: evidence from linked employer–employee data Page 13 of 21 3 Table 9 The effect of  coworkers with  recent outside  work experience on  the  wages of  skilled workers in  domestic enterprises 2005–2011 Share of coworkers with recent MNE experience by their level Share of coworkers with recent experience of skill in other domestic firms by their level of skill Unskilled Middling Skilled Unskilled Middling Skilled Notations in Eq. 3: θ θ θ θ θ θ F1 F2 F3 D1 D2 D3 All domestic firms 0.007 (0.9) 0.013 (1.5) 0.020* (1.9) 0.016*** (3.2) 0.024*** (3.4) − 0.037*** (− 4.5) Domestic firms 0.006 (0.5) 0.057** (2.1) 0.060*** (3.4) 0.002 (0.2) 0.064*** (3.4) − 0.019 (− 1.3) employing > 50 workers Significant at *0.1, **0.05, ***0.01 level. The t-values are based on standard errors adjusted for clustering by persons and firms. θ is significantly larger than θ and F3 F1 θ , but not θ . θ is significantly larger than θ D3 F2 F2 F1 Sample: 3,731,548 person-months belonging to skilled workers in 116,249 firms in full sample, 2,474,843 person-months in 77,412 firms in the 50+ sample. Dependent variable: log daily wage in the given month relative to the national mean. Controls: person, job and firm characteristics, sector–year interactions, worker and firm fixed-effects the noisy measurement of the F and D ratios in smaller an effect on skilled wages yields further support to the enterprises. learning hypothesis. The estimated spillover effect might seem economi - cally insignificant, but it is actually stronger than those 7 Two notes on differences by skills and sectors we know from the literature. The study of Poole (2013)— Throughout this paper, we focused on skilled workers which is closest to ours concerning method, sample char- mainly because we are interested in possible knowledge acteristics, and industry coverage—estimated that at the flows from foreign-owned to domestic firms, the traces average wage for a typical domestic worker, a 10 percent- of which are easier to find in the skilled labor market. age points increase in the share of former MNE work- We nevertheless estimated all our models for less-skilled ers increased incumbents’ wages by $23 per year. This workers and found that the effects of interest are smaller amount could buy a little more than one Starbucks solo and, in many cases statistically insignificant. Appendix 1: espresso a month in Rio de Janeiro in 2015. The compara - Fig. 1 illustrates this point. The figure compares the esti - ble estimate for skilled incumbents in our sample is $139 mates of the wage gap model (Table 3, model A) to simi- a year, which could buy 5.2 cups of Starbucks espresso a lar ones for unskilled and medium-skilled workers. The month in Budapest at 2015 prices. latter are very close to each other and amount to about Learning from ex-MNE peers is only one explanation 0.4 log points in the uncontrolled model, less than 0.1 in for the effect we identify. A firm’s effort to maintain its the panel regression with worker FE and less than 0.02 in wage ladder after hiring a high-wage ex-MNE worker the 2FE model. could occasionally motivate a firm to increase the wages Data available in the Labor Force Survey (Tables  13, of other employees. Still, we do not find this explanation 14 of Appendix 1) furthermore suggest that a part of convincing when spillover is observed in tens of thou- the MNE premium compensates unskilled workers for sands of firms. Why would so many domestic firms hire non-wage disamenities. Overtime work and afternoon high-wage workers from MNEs if this decision implies and night shifts are about twice as likely to occur among further wage growth without an underlying rise in pro- low and medium-skilled MNE employees compared to ductivity? A positive selection of all workers to firms their domestic counterparts. There is a smaller but simi - hiring from MNEs can also raise the average wage of larly signed difference concerning work on Saturdays coworkers with no MNE experience. However, our find - and Sundays. Furthermore, low skilled workers have a ings controlled for worker fixed effects and/or relating higher probability of becoming unemployed in foreign- only to incumbents are free of this kind of bias. Last but owned than domestic firms. The data does not indicate not least, the finding that only skilled ex-MNE peers have The calculation is based on the estimated own effect (0.074), the mean monthly earnings of skilled domestic firm employees in 2011 (236,078 Ft) and Skilled workers account for 25 percent of the total population observed in an average exchange rate of 225 Ft/$ in 2011 (National Bank, http://mnbko the source file. 15 per cent is unskilled (never worked in an occupation requir - zepar folya m.hu/arfol yam-2011.html). We could find Starbucks solo espresso ing nay kind of qualification) and 60 percent is classified as middling (worked prices for 2015 on the websites of local shops in Rio and Budapest: $1.92 and in skilled jobs but not in ones requiring tertiary educational attainment). $1.43, respectively. 3 Page 14 of 21 J. Köllő et al. ownership-specific differences of this kind among highly (Bernard and Sjoholm 2003). The implications of skills skilled workers. accumulation versus efficiency wages for the foreign- Table  10 summarizes point estimates of the wage gap, domestic wage gap and the wage loss from separation are lagged returns, and spillover effects from our preferred observationally identical. However, efficiency wages in model specifications for manufacturing and all other sec - MNEs do not imply that ex-MNE employees earn a pre- tors labeled ‘services’. The foreign-domestic wage gap is mium over the receiving domestic firm’s going wage rate more substantial in services than manufacturing, and the and exert influence on the earnings of their peers. lagged returns are broadly similar or somewhat larger in Third, a set of findings like this is likely to emerge only if services. By contrast, the spillover effects are estimated to MNE workers accumulate both general and r fi m-specific be stronger in manufacturing. We do not go to the details knowledge. As outlined in Becker’s (1962) seminal paper, of the between-sector differences. We only note that the in the case of general skills acquired through on-the-job returns to MNE experience are not restricted to the man- training, productivity and wages move in tandem. Work- ufacturing sector heavily over-represented in the related ers accumulating a substantial stock of general skills in literature. one firm are expected to earn higher-than-average wages in other firms. As far as general skills develop through informal communication between coworkers, their pres- 8 Discussion ence also tends to have a spillover effect. However, we do We interpret the coincidence of an MNE premium, sub- not expect that separation from an MNE induces a wage stantial wage loss from separation, lagged returns to loss in this scenario. MNE experience, and wage spillover as a signal of knowl- If the acquired knowledge is purely firm-specific, and edge flows from FDI to domestic firms. In such a sce - the risk of voluntary separation (motivated by factors nario, workers acquiring general and firm-specific skills other than between-firm wage differentials) is zero, then in the modern environment of MNEs are expected to the firm pays the going market wage before, during and earn more than their domestic counterparts. The specific after the period of skills accumulation. These skills lose components in their skills imply that MNE workers lose a their value with separation without an impact on sala- part of their wage advantage in case of involuntary sepa- ries. Pre-separation and post-separation wages are equal, ration. The general component in their skills gives rise to post-separation wages do not exceed the host firm’s aver - wage advantages in their new, domestic firm and tends age level, and they do not exert influence on the earnings to influence their peers’ productivity. The simultaneity of coworkers. In the likely case of non-zero risk of volun- of these symptoms calls into question some alternative tary quits, the firm will share in the costs and benefits of explanations, of which we discuss three ones. training, which implies lower wages in the accumulation First, the finding of a contemporaneous MNE pre - phase and higher wages afterward as long as the worker mium even after controlling for worker fixed effects calls into question that the foreign-domestic wage gap is fully explained by the crowding of high productivity workers in foreign-owned firms. Similarly, in a comparison of Table 10 Selected estimates by sectors domestic and foreign-owned start-ups, we find a sizable Manufacturing Services MNE premium even after controlling for their workers’ pre-entry wages. Contemporaneous MNE premium Second, intense human capital accumulation is admit- All firms, worker FE 0.152 0.236 tedly not the only potential source of an MNE premium, New, incumbent firms, DiD 0.135 0.232 with the most important alternative being efficiency wage Lagged MNE premium in domestic firms setting. MNEs may try to prevent leakage of information Sending firm is MNE, dL < 0.5, OLS 0.135 0.133 through labor turnover by paying a premium above the Sending firm is MNE, dL < 0.5, firm FE 0.056 0.044 market level (Fosfuri et  al. 2001). Their limited knowl - Overlapping cohorts estimate, DiD 0.027 0.072 edge of the local labor market and capital-labor relations Spillover effect, firms L > 50 employees may urge them to pay high wages and share a part of their On incumbents 0.088 0.057 revenues with workers. Furthermore, they may try to On all workers with no MNE experience 0.069 0.050 compensate their employees for a higher labor demand All coefficients are significant at the 0.01 level. The coefficients were estimated volatility (Fabri et al. 2003) or a higher plant closure rate separately for the two sectors Wage gains from foreign ownership: evidence from linked employer–employee data Page 15 of 21 3 stays with her employer. In this case, post-training invol- foreign-domestic wage gap from acquisitions. Thanks to a untary separations imply a wage loss, but we continue rich and big data set, we could compare how workers are not to expect lagged returns and spillover effects. selected to new MNEs and domestic firms, and identify a The literature emanating from Becker’s benchmark substantial wage differential between them. In the analy - models has been trying to reconcile the theory of on- sis of lagged returns and spillovers, we drew attention to the-job training with a series of empirical observations trade-offs between model quality and unbiasedness of the inconsistent with the extreme scenarios. A series of samples on which the models can be estimated. empirical findings and ample everyday experience sug- As we find substantial wage effects attributed to for - gest that (i) most skills are general, or at least sector eign ownership both in the short-run and long-run, even rather than firm-specific (ii) enterprises are willing to after controlling for potential biases as much as possible, pay for general training, and (iii) involuntary separa- we believe that the presence and significance of knowl - tions typically imply a loss. Acemoglu and Pischke edge transfer from MNEs is beyond doubt. Therefore, (1998) demonstrate that in a variety of market settings we argue that FDI coming from more developed coun- such as a compressed wage structure, substantial hir- tries exert positive effects on the receiving countries’ ing costs, information asymmetry, and other labor labor markets both through direct, and indirect channels. market imperfections, general skills are rewarded as Exploring whether these gains outweigh the potential if they were partly specific. The “skill-weights” model drawbacks could be the focus of future research on the of Lazear (2009) hypothesizes that skills are predomi- topic. nantly general, but firms attach different weights to Acknowledgements their components. A worker who leaves a firm will We thank Tibor Czeglédi, Éva Czethoffer, Endre Szabó, Melinda Tir and Kitti have a difficult time finding another employer that Varadovics for constructing the database from which we have drawn our sam- ples and Péter Elek, Miklós Koren, Gábor Kőrösi, László Lőrincz, László Mátyás, can make use of all the skills he acquired at the send- Frank Neffke, Álmos Telegdy, two anonymous reviewers and participants of ing firm. The limits of transferability impose a cost on several seminars and conferences in Budapest, Braga and Thessaloniki for their mobile workers, so the workers are unwilling to bear helpful comments and advice. the full cost of training, and the costs and benefits will Authors’ contributions be shared. Such a setting is likely to produce all of the LB’s contribution was the calculation of the contemporaneous wage gap, four outcomes observed in our data. and the discussion of estimation issues related to multi-way fixed effects models, alongside related technical issues. IB handled a large part of data management, worked out part of the methodology, and calculated the spillover effects. JK managed the project, worked out the paper’s structure, the 9 Conclusions methods and the identification of the various effects, calculated the lagged wage effect, and wrote most of the text. All authors read and approved the We found that high skilled MNE workers earn substan- final manuscript. tially higher wages than their domestic counterparts. They lose a part of their wage advantage after leaving the Funding Boza and Köllő gratefully acknowledge the financial support of the’ foreign-owned sector but, even so, they earn more than Knowledgeflows’ project of the European Research Council coordinated their domestic sector colleagues with no MNE experi- by Miklós Koren at the Central European University, Budapest. Grant ID: ence. Their presence in domestic firms exerts a positive FP7-IDEAS-ERC#313164. effect on the wages of their peers, who had no contact Availability of data and materials with foreign-owned firms or had no recent outside work Due to the size and sensitivity of the data, access to it is provided exclusively experience at all. for academic purposes by the Databank of the Centre for Economic and Regional Studies, Hungarian Academy of Sciences. We authorized the Data- The direct and indirect wage returns to work experi - bank to make the estimation samples and program codes available to referees ence in MNEs are large in Hungary, similar to less devel- (while maintaining their anonymity) and researchers willing to replicate the oped countries analyzed in the literature. The positive results. Access can be initiated at adatkeres@krtk.mta.hu. wage effects are not restricted to the manufacturing sec - Competing interests tor, which is in the focus of attention in the research on The authors declare that they have no competing interests. FDI.The estimates suggest that the effect of MNE expe - Author details rience on domestic sector wages is strongly affected by Institute of Economics, Hungarian Academy of Sciences, Budapest, Hungary. between-firm variance, that is, the higher-than-average 2 Central European University, Budapest, Hungary. wages of domestic firms connected with the MNEs via labor turnover. Appendix 1: Figures and tables Finally, the results draw attention to the difficulties See Figs. 1, 2, 3 and Tables 11, 12, 13, 14 and 15. of identifying a ‘pure’ ownership effect. The non-ran - dom selection of firms flaws the identification of the 3 Page 16 of 21 J. Köllő et al. 1 2 3 4 5 6 Specifications 95% CI Skilled 95% CI Middling 95% CI Unskilled Fig. 1 Estimates of the foreign-domestic wage gap by skills. Specifications: (1) sector–year interactions; (2) + person controls; (3) + job controls; (4) + firm controls; (5) + worker fixed effects; (6) + firm fixed effects. The confidence intervals are based on standard errors adjusted for clustering by persons and firms. On the sets of controls and the definition of skill levels, see Appendix 2: “Data ” points, zero wages and missing covariates. 98.5 percent of the workers belong to a single connected group. Spe- cial subsamples have been selected for the study of new Appendix 2: Data and key variables firms, lagged returns and spillovers. Data access: Data for the estimation samples and Stata Data do files are available on request. The original data set Starting sample: 50 percent random sample drawn from called Admin2 is also available via remote access to the Social Security Numbers (SSN, Hungarian TAJ) valid on Databank’s servers. Write to adatkeres@krtk.mta.hu for January 1, 2003. SSN holders aged 5–74 were retained. requesting access to the data. Note that the size of the Data held by the Pension Directorate (ONYF), the Tax original data ranges between 60 and 120 Gbytes, depend- Office (NAV), the Health Insurance Fund (OEP), the ing on the amount of information stored in special mod- Office of Education (OH), and the Public Employment ules that you want to merge to the base file. The files are Service (NMH) were merged and anonymized by the in Stata16 format. R and Python codes are allowed. National Information Service (NISZ). The original data consisted of payment records with start and end dates, a type-of-payment code and amounts received by the person. Employers were identified by ONYF and their annual financial data were provided by NAV. The data ‘When a group of persons and firms is connected, the group contains all the workers who ever worked for any of the firms in the group and all the firms was transformed to a fixed format monthly panel data at which any of the workers were ever employed. In contrast, when a group set by the Databank of the Institute of Economics of the of persons and firms is not connected to a second group, no firm in the first Hungarian Academy of Sciences. group has ever employed a person in the second group, nor has any person in the first group ever been employed by a firm in the second group. From Estimation sample: Workers employed with a labor an economic perspective, connected groups of workers and firms show the contract at least once in a foreign or domestic private realized mobility network in the economy. From a statistical perspective, con- enterprise the maximum employment level of which nected groups of workers and firms block-diagonalize the normal equations and permit the precise statement of identification restrictions on the person exceeded the 10 workers limit at least once in 2003–2011. and firm effects.’ Abowd et al. (2006). We removed workers and firms with less than two data log points Wage gains from foreign ownership: evidence from linked employer–employee data Page 17 of 21 3 -1 -2 FD DD FF DF excludes outside values Fig. 2 Shifts between sectors and wage change. The data relate to 307,874 shifts by skilled workers between ownership sectors in 2003–2011. F and D denote foreign-owned and domestic firms, respectively, in chronological order. The boxes display the interquartile ranges of log wage changes, with a horizontal line within the box indicating the median, and the whiskers showing the highest and lowest adjacent values. Heavy outliers are excluded. Wage change is measured as ln(w /w ), where w and w are average earnings (normalized for the national mean) in the job 1 0 1 0 spells after and before the shift, respectively Start-ups Other booming irms 150 800 -100 -50 0 50 100 -100 -50 0 50 100 Months before/after the big bang Months before/after the big bang Fig. 3 The mean size of firms classified as newly established. The data relate to 544 firms the size of which jumped from less than 5 to more than 50, or from less than 50 to more than 300 within a month (big bang). Firms changing majority owner after the big bang are excluded Mean size (number of employees) Log wage change Mean size (number of employees) 3 Page 18 of 21 J. Köllő et al. Table 11 Descriptive statistics Table 12 Pooled OLS results for Eq. 1, specification 4 Mean St. dev. Coefficient t-value Male 0.619Majority owner Age 37.9 10.4 Foreign 0.437 20.4 Log health expenditures/national average wage − 2.08 1.8 Person controls Receives disability pension/payment 0.006 Male 0.154 19.1 Receives care benefit 0.008 Age 0.032 15.2 Log regional unemployment rate − 2.66 0.386 Age squared/100 − 0.033 13.8 Central Hungary including Budapest 0.458 Months spent non-employed in 2003–2011 − 0.003 31.8 Tenure is unobserved 0.398 Receipt of disability payment − 0.373 23.3 Tenure (months) 13.44 19.0 Receipt of care allowance − 0.207 12.1 Top manager 0.051 Health expenditures (log) 0.002 7.3 Other manager 0.271Job controls Professional 0.299 Tenure if observed 0.001 4.3 Other white collar 0.112 Tenure is unobserved 0.138 12.2 Skilled blue collar 0.025 Spell lasting for 1 day 0.354 3.1 Assembler, machine operator 0.169 Top manager Ref. Elementary occupation 0.012 Other managers − 0.062 1.6 Agriculture 0.025 Professional − 0.016 0.4 Manufacturing 0.277 Other white collar − 0.298 9.0 Construction 0.061 Skilled blue collar − 0.607 28.8 Trade 0.278 Assembler, machine operator − 0.728 18.7 Finance 0.126 Laborer in elementary occupation − 0.821 18.8 Energy 0.018 Regional unemployment rate (log) − 0.063 5.2 Other services 0.215 Budapest 0.142 11.3 Foreign 0.397 Firm controls Domestic 0.603 Firm size (log) 0.086 7.3 Capital-labor ratio (log) 0.041 9.3 Firm size (log) 4.67 2.32 Fixed assets per worker (log) 7.92 1.81 Exporter 0.185 9.4 Exporter 0.521 Constant − 1.650 24.0 Adjusted R-squared 0.479 Skilled workers, estimation sample for the wage gap model (Eq. 1) Number of observations 19,961,622 Each variable covers 19,961,622 person months. The spells belong to workers employed at least once in a firm, the size of which exceeded the 10 workers limit Skilled workers, 2003–2011 at least once in 2003–2011. Public sector and state-owned firms are excluded. Dependent variable: log daily earnings relative to the national mean. For the Note that other samples used in the paper have been drawn from this source file exact definition of the variables see Appendix 2: "Data ". The coefficients of 63 sector–year dummies are not shown. The standard errors are adjusted for clustering by persons and firms. All coefficients are significant at 0.01 level except n) not significant at 0.1 level Key variables Wage: The daily wage figure used in the paper was calcu - lated as monthly earnings divided by the number of days by dividing them with the national average wage in the covered by pension insurance (‘working days’ henceforth) given month, as measured in the starting sample. Source: in the given month. Multiple payments made by the ONYF. same employer to the same person within a month were Foreign-owned firm, MNE: dummy variable set to 1 summed up. Working days belonging to these payments for firms majority owned by one or more foreign own - were also summed up but capped at 30 or 31 days. In the ers. Ownership shares are measured as fractions of sub- case of multiple job holders the wage figure belongs to scribed capital. Source: NAV. the highest paying job. We normalized the wage figures Wage gains from foreign ownership: evidence from linked employer–employee data Page 19 of 21 3 Table 13 Incidence of atypical work schedules in foreign and domestic enterprises Level of  education Domestic Foreign Domestic Foreign b b Shift workOvertime work Low 27.5 58.2 14.4 32.0 Middling 22.4 41.2 12.0 24.3 High 4.4 4.5 4.0 7.6 c c Work in the afternoon Work in the night Low 14.4 29.1 8.1 20.3 Middling 18.6 33.1 9.4 22.1 High 17.7 14.4 7.1 6.6 c c Work on Saturdays Work on Sundays Low 26.3 29.3 16.9 17.6 Middling 35.4 36.6 21.2 24.0 High 26.8 18.9 16.7 12.7 2003–2011, percent Low = primary school attainment, High = college or university, Middling = rest Source: Wage Surveys, 2003–2011, private sector. Firms are classified on the basis of their majority owners. The data indicate the percentage share of employees receiving shift pay and overtime pay, respectively. Authors’ calculation Source: Labor Force Surveys, 2003 Q1–2011 Q4, excluding public administration, education, health and social services. The data indicate the percentage share of employees working in the respective periods at least occasionally. Authors’ calculation classified as high skilled. Persons never employed outside Table 14 The effect of  ownership on  the  probability of becoming unemployed—logit odds ratios occupations 6 and 7 are classified as low skilled. Other persons are classified as medium skilled. Source: ONYF. Educational attainment Total time spent non-employed: The number of months Low Middling High out of employment in 2003–2011. Source: ONYF. Disability payment: dummy variable, with 1 standing Employer: MNE 1.199*** (2.57) 0.971 (0.50) 1.061 (0.89) for any kind of transfer (pension or allowance) received Female 0.916 (1.40) 1.029 (0.55) 1.149*** (2.44) on the basis of permanent disability (rokkantnyugdíj, rok- Age 1.012 (0.71) 0.941*** (3.85) 0.919*** (5.01) kantsági járadék). Monthly data. Source: ONYF. Age squared 0.999** (2.06) 1.000*** (3.08) 1.000*** (4.25) Care allowance: dummy variable, with 1 standing for Tenure (years) 0.894*** (9.27) 0.886*** (13.9) 0.895*** (10.0) any kind of benefit received by the observed person on Number of observations 82,638 205,597 227,074 the basis of raising children (tgyás, gyed, gyes, gyet) or Pseudo R2 0.076 0.067 0.068 taking care of disabled relatives (ápolási segély). Monthly Wald chi (51) 617.4*** 958.0*** 763.8*** data. Source: OEP, ONYF. Significant at the **0.5 and ***0.01 level Health expenditures: Expenditures and costs regis- Discrete time survival model, logit form, following (Jenkins 1995). Estimated tered by the National Health Insurance Fund (OEP). for the pool of 28 quarterly waves of the Labor Force Survey in 2003–2009. The estimation excludes the crisis period (2010 and 2011). Sample: employees. The items include total amount paid for OEP-sup - Dependent variable: 1 if the person was ILO-OECD unemployed in wave t + 1 ported medicine and the costs of OEP-supported ser- and 0 otherwise. The coefficients of 19 county dummies and 27 survey wave dummies are not shown vices/treatment provided by district doctors, specialists Low = primary school attainment, High = college or university, Middling = rest and hospitals. We normalized the nominal figures by dividing them with the national average wage in the given month, as measured in the starting sample. Zero Person controls expenditures were replaced with 1 Ft (0.3 Euro cents Gender, age: Source: ONYF. per annum). Annual data. Source: OEP. Skill levels: Skill levels are inferred from the ‘highest’ occupational status held by the person in 2003–2011. The classification is basedon one-digit occupational Job controls codes: 1 Top managers, 2 Other managers, 3 Profession- Tenure: Months elapsed since entry to the firm. Set to als, 4 Other white collars, 5 Skilled blue collars, 6 Assem- zero in the case of left-censored employment spells. blers and machine operators, 7 Elementary occupations. A dummy stands for observations from left-censored Persons employed in occupations 1–3 at least once are spells. 3 Page 20 of 21 J. Köllő et al. Table 15 On‑the ‑job training: fraction participating among MNE and domestic firm employees ForeignDomesticRatio 2003 0.102 0.0651.57 2004 0.100 0.0601.67 2005 0.043 0.0241.80 2006 0.045 0.0192.41 2007 0.033 0.0191.73 2008 0.024 0.0191.30 2009 0.020 0.0131.59 2010 0.025 0.0151.59 2011 0.021 0.0141.61 High skilled employees working at least one hour in the reference week = 1 Source: Authors’ calculation using waves 45–80 of the LFS. Sample: ILO-OECD employed with college or university background. Key variables: participates in training of any kind outside the school system; the employer is majority or minority foreign-owned Note that the question on participation changed in 2005. Figures above and below the dotted line are not directly comparable worker-firm data. GEP Research Paper, No. 2007/08, University of Not - Spell lasting for 1  day: Hungarian firms often pay to tingham (2007) individual subcontractors by formally employing them Antalóczy, K., Sass, M.: Greenfield investments in Hungary. Transnatl. Corp. 3, on the day of payment. This practice results in exorbi - 39–60 (2001) Balatoni, A., Pitz, M.: A működőtőkehatása a bruttónemzetijövedelemreMagya- tant ‘daily wages’ in some cases. rországon. KözgazdaságiSzemle 59(1), 1–30 (2012) Occupation: One-digit ISCO codes. Balazsi, L.: The econometrics of linear models for multi-dimensional panels. Regional unemployment rate: seasonally adjusted Doctoral thesis in Economics, Central European University, Budapest (2017) https ://www.ceu.edu/sites /defau lt/files /phdth esisb ookle tbala zsi. ILO-OECD unemployment rate in the given month and pdf NUTS-2 region. The worker’s region is identified on the Balazsi, L., Matyas, L., Wansbeek, T.: The estimation of multidimensional fixed basis of his/her zip code in 2003. Source: author’s cal- effects panel data models. Econom. Rev. 37, 212–227 (2018) Balsvik, R.: Is labor mobility a channel for spillovers from multinationals? culation using the Labor Force Survey. Evidence from Norwegian manufacturing. Rev. Econ. Stat. 93(1), 285–297 (2011) Barry, F., Gorg, H., Strobl, E.: Foreign direct investment and wages in domestic firms in Ireland: productivity spillovers versus labour-market crowding Firm controls out. Int. J. Econ. Bus. 12, 67–84 (2005) Firm size: average number of employees. Annual data. Becker, G.S.: Investment in human capital: a theoretical analysis. J. Political Source: NAV. Econ. 70(5, Part 2), 9–49 (1962) Bernard, A.B., Sjoholm F.: Foreign owners and plant survival. NBER Working Capital-labor ratio: net value of fixed assets per Paper No. 10039, October 2003 (2003) worker. Annual data. Source: NAV. Bisztray, M.: The effect of FDI on local suppliers: evidence from Audi in Hun- Exporter: non-zero exports revenues. Annual data. gary. Institute of Economics, Centre for Economic and Regional Studies, Hungarian Academy of Sciences Discussion papers MT-DP-2016/22, pp. Source: NAV. 46 (2016) Sector: NACE 2. Source: NAV. Calderon, C., Loayza, N., Serven, L.: Greenfield foreign direct investment and mergers and acquisitions: feedback and macroeconomic effects. Policy Research Working Paper Series No. 3192. The World Bank, Washington Received: 17 February 2020 Accepted: 27 January 2021 (2004) Carneiro, A., Guimarães, P., Portugal, P.: Real wages and the business cycle: accounting for worker, firm, and job title heterogeneity. Am. Econ. J. Macroecon. 4(2), 133–152 (2012) Chen, C., Zhao, H., Zhou Y.: Foreign direct investment and wage inequality: References evidence from the People’s Republic of China (ADBI Working Paper No. Abowd, J.M., Kramarz, F., Margolis, D.N.: High wage workers and 734). Asian Development Bank Institute, Tokyo (2017) high wage firms. Econometrica 67, 251–333 (1999). https ://doi. Cornelissen, T.: Thestata command Felsdvreg to fit a linear model with two org/10.1111/1468-0262.00020 high-dimensional fixed effects. Stata J. 8 (2), 170–189 (2008) Abowd, J.M., Kramarz, F., Roux, S.: Wages, mobility and firm performance: Correia, S.: reghdfe: stata module for linear and instrumental-variable/gmm advantages and insights from using matched worker-firm data. Econ. J. R. regression absorbing multiple levels of fixed effects. Statistical Soft - Econ. Soc. 116(512), F245–F285 (2006) ware Components s457874, Boston College Department of Economics. Acemoglu, D., Pischke, J.S.: Beyond Becker: training in imperfect labour mar-https ://ideas .repec .org/c/boc/bocod e/s4578 74.html (2017) kets. Econ. J. 109(453), F112–F142 (1998) Csáfordi, Z., Lőrincz, L., Lengyel, B., Kiss, K.M.: Productivity spillovers through Aitken, B.J., Harrison, A.E.: Do domestic firms benefit from direct foreign invest - labor flows: productivity gap, multinational experience and industry ment? Evidence from Venezuela. Am. Econ. Rev. 89(3), 605–618 (1999) relatedness. J. Technol. Transf (2018). https ://doi.org/10.1007/s1096 Andrews, M., Bellmann, L., Schank, T., Upward, R.: The takeover and selec- 1-018-9670-8 tion effects of foreign ownership in Germany: an analysis using linked Wage gains from foreign ownership: evidence from linked employer–employee data Page 21 of 21 3 Csengődi, S., Jungnickel, R., Urban, D.: Foreign takeovers and wages in Maliranta, M., Mohnen, P., Rouvinen, P.: Is inter-firm labor mobility a channel of Hungary. Rev. World Econ. 144, 55–82 (2008). https ://doi.org/10.1007/ knowledge spillovers? Evidence from a linked employer–employee panel. s1029 0-008-0137-0 ETLA Discussion Papers, The Research Institute of the Finnish Economy Czafit, B., Köllő, J.: Employment and wages before and after incarceration; (ETLA), No. 1116 (2008) evidence from Hungary. IZA J. Eur. Labor Stud. 4(1), 1–21 (2015). https Martins, P.: Do foreign firms really pay higher wages? Evidence from different ://doi.org/10.1186/s4017 4-015-0044-z estimators. IZA, Discussion Paper No. 1388 (2004) Djankov, S., Hoekman, B.: Foreign investment and productivity growth in Martins, P.S., Opromolla, L.D.: Exports, imports and wages: evidence from Czech enterprises. World Bank Econ. Rev. 14(1), 49–64 (2000) matched firm-worker-product panels. IZA Discussion Paper No. 4646 Earle, J.S., Telegdy, Á., Antal, G.: Foreign ownership and wages: evidence (2009) from Hungary, 1986–2008. Ind. Labor Relat. Rev. 71(2), 458–491 (2017) Mion, G., Opromolla, L.D.: Managers’ mobility, trade performance, and wages. Elek, P., Köllő, J., Reizer, B., Szabó, P.A.: Detecting wage under-reporting using European Central Bank Working Paper Series No. 1596, October (2013) a double hurdle model. In: Lehmann, H., Tatsiramos, K. (eds.) Research Mittag, N.: A simple method to estimate large fixed effects models applied to in Labor Economics. Informal Employment in Emerging and Transition wage determinants and matching, IZA DP No. 10447, December (2016) Economies, Ch 4, vol. 34, pp. 135–166. Emerald, Bingley (2012) OECD: Chapter 5: Do multinationals promote better pay and working condi- Fabri, F., Haskel, J.E., Slaughter, M.J.: Does nationality of ownership matter for tions? In: OECD Employment Outlook, pp. 263–331. OCED, Paris (2008a) labor demands? J. Eur. Econ. Assoc. Pap. Proc. 1(2–3), 698–707 (2003) OECD: The impact of foreign direct investment on wages and working condi- Fons-Rosen, C., Kalemli-Ozcan, S., Sorensen, B.E., Villegas-Sanchez, C., Volos- tions, OECD-ILO Conference on Corporate Social Responsibility. Employ- ovych, V.: Foreign investment and domestic productivity: identifying ment and industrial relations: promoting responsible business conduct in knowledge spillovers and competition effects. NBER Working Paper No. a Globalising economy 23–24 June 2008, OECD Conference Centre, Paris, 23643, August (2017) France (2008b) Fosfuri, A., Motta, M., Ronde, T.: Foreign direct investment and spillovers Pettit, B., Lyons, C.J.: Incarceration and the legitimate labor market: examin- through workers’ mobility. J. Int. Econ. 53(1), 205–222 (2001) ing age-graded effects on employment and wages. Law Soc. Rev. 43(4), Görg, H., Strobl, E.: Spillovers from foreign firms through worker mobility: an 725–756 (2009) empirical investigation. Scand. J. Econ. 107(4), 693–709 (2005) Poole, J.P.: Knowledge transfers from multinational to domestic firms: evidence Gorodnichenko, Y., Svejnar, J., Terrell, K.: When does FDI have positive spillo- from worker mobility. Rev. Econ. Stat. 95(2), 393–406 (2013) vers? Evidence from 17 transition market economies. J. Comp. Econ. Smarzynska-Javorcik, B.: Does foreign direct investment increase the pro- 42, 954–969 (2014) ductivity of domestic firms? In search of spillovers through backward Grogger, J.: The effect of arrests on the employment and earnings of young linkages. Am. Econ. Rev. 94(3), 605–627 (2004) men. Q. J. Econ. 110(1), 51–71 (1995) Stoyanov, A., Zubanov, N.: Productivity spillovers across firms through worker Guimaraes, P., Portugal, P.: A simple feasible alternative procedure to estimate mobility. Am. Econ. J. Appl. Econ. 4(April), 168–198 (2012). https ://doi. models with high-dimensional fixed effects. Stata J. 10(4), 628–649 (2010) org/10.1257/app.4.2.168 Heyman, F., Sjöholm, F., Tingvall, P.: Multinationals, cross-border acquisitions Suyanto, R.S., Bloch, H.: Which firms benefit from foreign direct investment? and wage dispersion. Can. J. Econ. 44(2), 627–650 (2007) Empirical evidence from Indonesian manufacturing. J. Asian Econ. 33, Jenkins, S.P.: Easy estimation methods for discrete-time duration models. 16–29 (2014) Oxford Bull. Econ. Stat. 57(1), 129–136 (1995) UNECE: Economic Growth and Foreign Direct Investment in the Transition Jurajda, Š, Stančík, J.: Foreign ownership and corporate performance: the Economies. Economic Survey of Europe, 2001 No. 1, United Nations Czech Republic at EU entry. Financ. úvěr-Czech J. Econ. Financ. 62(4), Economic Commission for Europe, Geneva (2001) 306–324 (2012) Keller, W., Yeaple, S.R.: Multinational enterprises, international trade, and tech- Publisher’s Note nology diffusion: a firm-level analysis of the productivity effects of foreign Springer Nature remains neutral with regard to jurisdictional claims in pub- competition in the United States. Rev. Econ. Stat. 91(4), 821–831 (2009) lished maps and institutional affiliations. Konings, J.: The effects of foreign direct investment on domestic firms. Econ. Transit. 9(3), 619–633 (2001) Kosová, R.: Do foreign firms crowd out domestic firms? Evidence from the Czech Republic. Rev. Econ. Stat. 92(4), 861–881 (2010) LaLonde, R.J., Cho, R.: The impact of incarceration in state prison on the employment prospects of women. J. Quant. Criminol. 24(3), 243–267 (2008) Lazear, E.P.: Firm-specific human capital: a skill-weights approach. J. Political Econ. 117(5), 914–940 (2009) Lipsey, R.E., Sjöholm, F.: FDI and wage spillovers in Indonesian manufacturing. Rev. World Econ. 140, 321–332 (2004) Malchow-Moller, N., Markusen, J., Schjerning B.: Foreign firms, domestic work - ers. NBER Working Paper, No. 13001, Cambridge, Massachusetts (2007) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal for Labour Market Research Springer Journals

Wage gains from foreign ownership: evidence from linked employer–employee data

Loading next page...
 
/lp/springer-journals/wage-gains-from-foreign-ownership-evidence-from-linked-employer-zaOJAwgirD
Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2021
ISSN
1614-3485
eISSN
2510-5027
DOI
10.1186/s12651-021-00286-0
Publisher site
See Article on Publisher Site

Abstract

We compare the wages of skilled workers in multinational enterprises (MNEs) versus domestic firms, the earnings of domestic firm workers with past, future and no MNE experience, and estimate how the presence of ex-MNE peers affects the wages of domestic firm employees. The analysis relies on monthly panel data covering half of the Hungar - ian population and their employers in 2003–2011. We identify the returns to MNE experience from changes of owner- ship, wages paid by new firms of different ownership, and the movement of workers between enterprises. We find high contemporaneous and lagged returns to MNE experience and significant spillover effects. Foreign acquisition has a moderate wage impact, but there is a wide gap between new MNEs and domestic firms. The findings, taken together, suggest that MNE employees accumulate partly transferable knowledge, valued in the high-wage segment of the local economy that is connected with the MNEs via worker turnover. Keywords: Multinational enterprises, FDI, Wage differentials, Wage spillover, Hungary JEL Classification: F23, J24, J31, O33 1 Introduction Heyman et  al. 2007; Andrews et  al. 2007; Malchow- While policymakers in developing countries are often Moller et  al. 2007). An adverse competition effect often criticized for ‘selling out’ the country to foreigners, FDI offsets the positive direct impact of FDI on productiv - can actually bring valuable knowledge to a less developed ity and wages even in relatively undeveloped economies economy, spreading through labor mobility channels. (Aitken and Harrison 1999; Djankov and Hoekman 2000; Undeniably, corporate revenues can find their way back Konings 2001; Barry et  al. 2005). The positive spillovers home via profit repatriation and transfer pricing, and many are often restricted to specific sectors (Keller and Yeaple MNEs enjoy a generous initial tax holiday. However, MNE 2009; Suyanto and Bloch 2014; Fons-Rosen et  al. 2017). workers’ wage premium over similar domestic-sector Still, the existence of a vast MNE premium in the emerg- employees in comparable firms directly benefits society, ing and transition economies (Lipsey and Sjöholm 2004; especially if the underlying excess productivity is portable OECD 2008a; Chen et al. 2017), and the findings of posi - and exerts positive spillover effects. Unlike the returns to tive spillovers (Smarzynska-Javorcik 2004; Görg and capital investment and part of the profit, the wage surplus Strobl 2005; Kosová 2010; Poole 2013; Gorodnichenko predominantly remains and is spent in the host country. et  al. 2014) encourage us to seek evidence for a ‘knowl- The literature provides ample evidence to call into edge flows’ scenario. To assess the magnitude of the question the general validity of such an optimistic sce- potentially beneficial impact of FDI, we study the direct nario. The foreign-domestic wage gap is negligible in and indirect wage effects of work experience in multi - countries close to the productivity frontier (Balsvik 2011; national enterprises (MNEs) using linked employer– employee data on skilled workers in Hungary, 2003–2011. We contribute to the literature by empirically show- ing in a single study that (i) MNEs pay markedly higher *Correspondence: bozaistvan@gmail.com Central European University, Budapest, Hungary wages than similar domestic firms. (ii) MNE employees Full list of author information is available at the end of the article Section  2 provides a detailed introduction to the previous literature includ- ing the sources cited here. © The Author(s) 2021. This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/. 3 Page 2 of 21 J. Köllő et al. lose a part of their wage advantage upon leaving the the effect it identifies is unsuitable for out-of-sample pre - foreign-owned sector. (iii) Even so, they earn more than diction. Only 5.3 percent of the observed firms changed their colleagues in domestic enterprises. (iv) Domestic the majority owner during the observation period in our firm employees benefit from having ex-MNE peers. We sample. These companies paid significantly higher wages interpret the coincidence of the MNE premium, par- than ‘always domestic’ firms (when they were domes - tial wage loss from separation, lagged returns to MNE tic) and significantly lower wages than ‘always foreign’ experience, and wage spillovers as a signal of knowledge companies (when they were foreign-owned): this is how transfer from MNEs to domestic firms. While alternative the 2FE model arrives at a close-to-zero estimate of the explanations exist for each of the presented symptoms, ownership-specific wage gap. These firms’ experience can in the last section of the paper we argue that a ‘knowl- hardly predict how big MNEs like Mercedes-Benz or IBM edge flows’ scenario has the best chance to produce all of would pay their employees in the unlikely event of takeo- the four outcomes. ver by a local business person. It also tells nothing about Regarding methodology, we draw attention to the dif- the potential wage gains from greenfield investments, ficulties of identification coming from the non-random which played a significant role in the 1990s (Calderon selection of firms into foreign ownership and of differ - et al. 2004). We utilize a difference-in-difference estima - ently skilled workers into foreign enterprises. We find tion of wage gains from joining a new MNE over joining a trade-offs between model quality and unbiasedness of the new domestic firm to learn about the ownership-specific samples on which the first-best models can be estimated. wage gap between ‘always foreign’ and ‘always domestic’ The analysis is based on a big administrative panel companies. This approach suggests that the employees data set covering half of the Hungarian population and of new MNEs earn 15 percent more than their domestic their employers in 2003–2011. We restrict the analysis counterparts. to skilled workers for three reasons. First, the traces of Turning to the MNE premium’s portability, we have to knowledge transfer are easier to find in the skilled labor deal with endogeneity and ability biases, as worker mobil- market. Second, data discussed later suggest that a part ity is not random. If a worker is fired from an MNE, it of the MNE premium compensates unskilled workers for may be because her marginal product is lower than aver- non-wage disamenities like shift work, weekend work, age. If a domestic employer attracts a worker, it may be and a higher probability of becoming unemployed. The because she has a higher-than-average marginal product data does not indicate ownership-specific differences of irrespective of the sector of employment. To address the this kind among highly skilled workers. Third, repeat - first problem, we compare domestic firm employees with ing the estimations for middling and unskilled workers recent MNE experience to their peers who had outside would triple the statistics to be presented, with minimal experience in the domestic sector. We focus on workers added content. Estimation on a pooled sample would losing or leaving their jobs in times of mass dismissals only attenuate the relevant parameters. when separations are more likely to be exogenous to the We start by estimating the foreign-domestic wage individual worker’s productivity. The model controls for gap using panel regressions. By gradually removing the heterogeneity of the sending firms via observable con - effects of observed and unobserved worker and firm trols and use fixed effects for the receiving ones. We find characteristics, we get from a substantial raw gap of 0.75 that former MNE employees earn more by 13 percent log points to 0.24 points after controlling for worker fixed than similar workers coming from collapsing domestic effects and a mere 0.03 points’ pure ownership-specific enterprises. Workers separating from their employers for wage differential estimated with both worker and firm reasons other than mass dismissals acquire a significantly fixed effects (2FE henceforth). lower (5 percent) lagged MNE premium. While a 2FE model can answer how an existing firm’s Satisfactory model quality comes at the cost of distor- wage level changes in response to a change in ownership, tions in the sample and a significant loss of observations in this case, too. Only about 7 percent of the person- months in our data makes it to the estimation sample of a model in which work histories and characteristics MNEs may pay high wages to skim the cream of the labor force, buy loy- of the sending and receiving firms are adequately con - alty, contain turnover, stimulate work effort, or prevent information leakage. Workers’ wages may fall upon leaving the MNE sector for losing these wage trolled. The problem would be further aggravated by components and because employers perceive their dismissal as a negative sig- nal. Ex-MNE workers may have high wages in domestic firms because they have high reservation wages and belong to the lucky few to find a well-paying job in the domestic sector. Spillover effects may arise from the employer’s wish to keep within-job wage differentials within tolerable limits. Antalóczy and Sass (2001) estimate that the share of greenfield FDI in total We justify this choice and present some results on less skilled workers in inward FDI amounted to 25–30 percent in Hungary and other CEE countries Sect. 7. during the transition. Wage gains from foreign ownership: evidence from linked employer–employee data Page 3 of 21 3 the inclusion of worker fixed effects to reduce ability 2 Previous findings on the foreign‑domestic wage bias. To avoid this issue while utilizing more data and gap, lagged returns and spillovers still controlling for the potential bias, we rely on a less Estimates of the foreign-domestic wage gap vary widely, demanding ‘overlapping cohorts’ model that compares with the MNE premium found to be nearly negligible in domestic firm employees with future and past experience the most developed market economies. In Norway, the in foreign versus domestic firms. This model can utilize a OLS estimate by Balsvik (2011), controlled for worker much broader sample, as workers with only two observed and plant characteristics, amounts to 3 percent, which spells can contribute to the estimation if any of those is at falls to 0.3 percent once she includes worker fixed effects. a foreign-owned employer. The estimated return to prior An OLS estimate for Sweden by Heyman et  al. (2007) MNE experience amounts to 0.07 log points. is even lower at 2 percent. Andrews et  al. (2007) and Finally, we estimate spillover effects for incumbent Malchow-Moller et  al. (2007) detect positive gaps in the domestic firm employees, controlling for observed and range of 1 and 3 percent in Germany and Denmark. The unobserved worker and firm characteristics. We deviate OLS estimate of Martins (2004) for Portugal is higher (11 from a similar attempt by Poole (2013) in two ways. First, percent), but he finds that the MNE wage premium vir - we also study how skilled incumbents’ wages respond tually disappears after controlling for worker selection. to the presence of less qualified ex-MNE peers. Second, These figures compare to 32 percent (pooled OLS for and more importantly, we address the selection prob- all skill levels) and 13 percent (after adding worker fixed lem that arises when the analysis is restricted to incum- effects) in our sample. Workers moving from domestic to bents (domestic workers with no experience outside foreign-owned firms are estimated to gain 6 percent in their firms). Incumbents in our data account for only 22 Germany and 8 percent in Norway (Andrews et al. 2007; percent of the workers ever employed in the domestic Balsvik 2011), which compares to 53 percent in the Hun- sector. Their exposure to peers with MNE experience dif - garian sample for all skill levels. fers substantially from that of the average worker. In an The foreign-domestic gap is much broader in less alternative specification, we ensure the identification of developed countries: according to raw data presented in within-firm spillovers using a 2FE model. We find that Lipsey and Sjöholm (2004), in Indonesian manufactur- a one-standard-deviation difference in the share of high ing, the MNE premium amounts to 47 percent for blue skilled ex-MNE peers shifts peers’ wages with no MNE collars and 55 percent for white collars (41 and 73 per- past up by slightly more than one percent. Having quali- cent in Hungary). Chen et  al. (2017) report a gap of 40 fied peers with outside experience in the domestic sec - percent in Chinese manufacturing. An overview of data tor and having low-skilled peers with MNE experience do in OECD (2008a), based on the World Bank Enterprise not affect wages. Survey, indicates raw gaps of between 40 and 50 percent Section  2 discusses previous findings on the paper’s in Africa, Asia, the Middle East, and combining all these topic and prewarns the reader of our estimates. Section 3 regions and adding Central and Eastern Europe. introduces the data and the local context. Section  4 is A more detailed analysis of the sources of the gaps in devoted to the study of the foreign-domestic wage gap. Germany, Portugal, the UK, and Brazil (OECD 2008b) Sections  5 and 6 present the results on lagged returns finds that takeovers’ marginal effect on wages falls short and spillover effects, respectively. Section  7 briefly com - of 3 percent in all of these countries. Results from Hun- ments on differences by skill levels and industries. Sec - gary point to similar patterns. Csengődi et al. (2008) use a tion  8 sums up the results and argues that the empirical different data set from ours (the Wage Survey, a repeated findings, taken together, yield support to a ‘skills diffu - cross-section LEED which allows the linking of firms but sion’ scenario. not workers) and find that after adding firm fixed effects, the MNE wage premium falls to a mere 3 percent as it does in our case. Earle et  al. (2017) use the same data 5 Note that in the Norwegian case, workers moving from MNEs to domestic With the requirement of controlling for lagged size changes, we would need firms also acquire a gain of 7 percent, while in our sample they lose 11 per - workers with at least four employment spells in a nine-year-long period, with cent. The median loss amounts to 26 percent in the case of skilled workers. a specific pattern DDFD, where F and D stand for foreign-owned and domes - See Table 2. tic firms. Identification in this setting would come from comparing the sec - ond and fourth domestic job entries. The third, F spell is the treatment, and In the Czech Republic, Jurajda and Stančík (2012) detect sigificantly faster a first spell is required for the inclusion of firm sizes. Besides, in this setting wage bill growth in (and only in) manufacturing firms with a low export the ex-MNE spells would sistematically happen later on in worker’s career, so share. They cannot decompose the wage bill effect into wage and employ - life-cycle wage changes may be potentially captured by the parameter as well. ment effects. They also show that domestic firms subject to foreign acquisition pay Which is a lower bound as in this model, we do not control for employ- higher-than-average wages already before the takeover, hinting at a non- ment change in the sending firm. random selection to foreign buy-out. 3 Page 4 of 21 J. Köllő et al. source and detect a slightly higher premium of 7 percent At the same time, several studies have identified posi - that is still very far from the estimates they get with- tive spillovers. Using Lithuanian data, Smarzynska-Javor- out controlling for unobserved firm characteristics and cik (2004) detects positive productivity spillovers from firm-specific trends. The effects identified using data on MNEs to local suppliers. Similarly, Gorodnichenko et al. worker mobility by OECD (2008b) are more substantial: (2014) find that backward linkages positively affect the the estimates vary between 6 and 8 percent in Germany productivity of domestic firms (while horizontal and for - and the UK, more than 10 percent in Portugal, and 20 ward linkages show no consistent effect) in 17 transition percent in Brazil. The authors argue that the discrepancy countries. Using Czech data, Kosová (2010) demonstrates between the estimates based on takeovers versus worker that crowding out is short-term: after an initial shock, flows are explained by foreign firms’ propensity to share domestic firm growth accelerates, and survival rates their productivity advantage more extensively with new improve. Görg and Strobl (2005) show that entrepreneurs workers than with workers who do not change firms. We with MNE experience start more productive small busi- believe that the difference instead roots in the non-ran - nesses in Ghana. Bisztray (2016) found that new entrants’ dom selection of firms to acquisition, as will be discussed growth in productivity was significantly higher when in more detail later. located close to Audi and operated in a supplier industry. To our knowledge, Balsvik’s paper is the only one esti- Importantly, from this paper’s point of view, Poole mating the wage advantage of ex-MNE employees in (2013) estimates that the wages of incumbent domes- domestic firms. She identifies a premium of 6.9 percent tic firm employees in Brazil rise by about 0.6 percent if for workers with three or more years of tenure in an the share of ex-MNE employees increases by 10 percent, MNE. However, she also detects an advantage of 3.3 per- while the effect of outside experience in local firms is cent on the part of workers arriving from local firms, sug - about ten times weaker than that. While the effect she gesting a net benefit from MNE experience of 3.6 percent estimates is not particularly strong, it is statistically sig- (and smaller advantages in case of shorter completed nificant at conventional levels. tenure in the previous job). We find that domestic firm One can also find indirect evidence on spillovers, con - employees, who left an MNE because of mass dismissals, sidering that MNEs are more productive and more likely closure, or relocation earn more than their ex-domestic to export and engage in R&D. Stoyanov and Zubanov counterparts by 13 percent. (2012) show that (in Denmark) workers coming from The empirical evidence on wage and productivity spill - more productive firms experience productivity gains. overs are mixed. Starting with papers that depict a not Similar results are presented for Hungary by Csáfordi too rosy picture of how MNEs affect the rest of the econ - et  al. (2018). Mion and Opromolla (2013) show that omy, Aitken and Harrison (1999) and Djankov and Hoe- export experience implies higher export performance kman (2000) identify a positive direct effect of foreign and a sizable wage premium for Portuguese managers, ownership on productivity in Venezuela and the Czech who leave for non-exporters. In Finland, Maliranta et al. Republic, but negative spillovers. Konings (2001) sug- (2008) identify positive impact of hiring workers with gests that the adverse competition effect is stronger than previous R&D experience to non-R&D jobs. the positive direct productivity effect of FDI in Bulgaria, Romania, and Poland. Barry et  al. (2005) found that for- 3 Data and the local context eign presence in a sector hurts wages and productivity in 3.1 Data sources domestic exporting firms in the same industry (but does Our estimation samples have been drawn from a big lon- not affect wages in domestic non-exporters) in Ireland. gitudinal data set covering a randomly chosen 50 percent Fons-Rosen et  al. (2017) conclude that in six advanced of Hungary’s population aged 5–74 in January 2003. Each European countries, positive spillovers are restricted person in the sample is followed monthly, from Janu- to sectors where domestic enterprises are technologi- ary 2003 until December 2011, or exit from the registers cally close to MNEs. Suyanto and Bloch (2014) find the for death or permanent out-migration. The data collect opposite in Indonesia. Keller and Yeaple (2009) detect information from the Pension Directorate, the Tax Office, significant worker-level wage spillovers only in high-skill- the Health Insurance Fund, the Office of Education, and intensive industries in US manufacturing. By looking at the Public Employment Service. We use information existing firms in an Audi plant’s supplier industries in on the highest paying job of a given person in a given Hungary, Bisztray (2016) finds no positive effect on pro - month, days in work, and amounts earned in that job. ductivity. She also finds that firms with foreign owners Throughout the paper, we use daily wages (the monthly account for all the positive impact on sales and employ- value divided by days in work) normalized for the given ment, suggesting a foreign-to-foreign complementarity month’s national average. We have data on occupation, rather than a galvanizing effect on the domestic sector. type of employment relationship, registration at a labor Wage gains from foreign ownership: evidence from linked employer–employee data Page 5 of 21 3 office, receipt of transfers, and several proxies of the a yearly basis would impair the precise measurement of person’s state of health. We do not observe educational tenure and the time between two jobs—essential controls attainment—this is approximated with the person’s high- in the analysis of lagged returns. Third, higher observed est occupational status in 2003–2011. The data on firms mobility helps in identifying firm and person effects. The come from the annual tax reports of businesses obliged problem raised by inflating observations at the same firm to conduct double book-keeping. The firm-level variables is taken care of by the worker and firm-level clustering of are merged into the respective person-month observa- errors. tions. We regard a firm as MNE if foreigners’ share in subscribed capital exceeds 50 percent.3.2 MNEs in Hungary We restrict the analysis to skilled workers employed In the first decade after the start of the transition, Hun - at least once in a foreign or domestic private enterprise gary was the most successful country within the former the employment level of which exceeded the ten work- Soviet bloc in attracting foreign capital. By 2003, the ers limit at least once in 2003–2011. We have several beginning of our observation period, cumulative FDI reasons to set a size limit. First, foreign firms are nearly inflows exceeded 40 percent of the GDP, multination- absent in the small firm sector. Second, financial data als employed 15 percent of the labor force (including self- are not available for sole proprietorships and unincor- employment and the public sector into the denominator) porated small businesses. Third, the financial reports of and more than 30 percent of private-sector employees. incorporated small firms are often incomplete and erro - They produced 20 percent of the GDP and delivered over neous. Finally, the earnings data of small firms are flawed two-thirds of the exports (Balatoni and Pitz 2012). Large by paying “disguised” minimum wages. Small firms’ multinationals, including Audi, General Motors, and inclusion would also raise the risk of measurement error Suzuki, dominated the motor industry. Foreign presence in the analysis of spillover effects since the probability was already significant in the tobacco, leather, chemi - of not observing an ex-MNE employee in a 50-percent cal, rubber, and electronics industries, with employment sample is much higher in small establishments. We itera- shares of between 50 and 80 percent. tively removed workers and firms with less than two data Almost three-fourths of the cumulative FDI inflows points, zero wages, and missing covariates. have arrived in sectors outside of manufacturing. As After these steps of data cleaning, we are left with shown in column 4 of Table  1, nearly 60 percent of the a sample of 19,961,622 person-months belonging to skilled employees within the MNE sector worked in the 344,203 skilled workers and 119,580 firms. 52.6 percent tertiary sector. Therefore, we do not restrict the analy - of the workers had at least one spell of employment in the sis to manufacturing, as most papers do in the strand of foreign-owned sector, of which 21.5 percent worked only the literature we follow (see Barry et  al. 2005; Görg and in MNEs. We draw special sub-samples from this start- Strobl 2005; Lipsey and Sjöholm 2004; Smarzynska- ing population for the study of new firms, lagged returns Javorcik 2004; Balsvik 2011 as opposed to Poole 2013, and spillover effects. Descriptive statistics are presented whose study covers all sectors in Brazil). While FDI typi- in Table 11 of Appendix 1. cally boosts exports and generates demand for domestic Even though our firm-level variables are of annual fre - manufacturers producing intermediate goods, its contri- quency, we prefer to analyze the data at a monthly level bution to the quality of retail trade, banking and services for several reasons. First, the affiliation of a worker can - can be equally important, especially in the former state- not be precisely measured on a yearly basis. About 25 socialist countries, which started the transition with percent of the workers employed by an MNE for at least critically undeveloped non-tradable sectors. The foreign- one month in a given year also had one or more spells in owned and domestic parts of the economy are closely the domestic sector in the same year. Second, turning to connected via labor turnover. In the skilled labor market, 37.2 percent of the domestic firms, employing 69 percent of the domestic labor force, hired at least one ex-MNE worker in 2003–2011. See Appendix 2 for variable definitions. Setting the limit elsewhere does not affect the results, since 93 percent of the firms with nonzero foreign presence are majority foreign-owned. 3.3 Descriptive statistics on wages and wage change In 2014, MNEs had a 4.5 percent employment share in the 1–10 work- Table  2 presents raw statistics on wage levels across ers category (Authors’ calculation based on the 2014 Q4 wave of the Labor ownership categories and wage changes associated with Force Survey). skilled workers’ shifts between them. The data shows vast This term hints at the practice of paying workers the minimum wage (subject to taxation) and the rest of their remuneration in cash. Elek et  al. (2012) estimate that in 2006 the share of workers paid in this way amounted to 20 percent in firms employing 5–10 workers, 10 percent in slightly higher firms (11–20 workers) and less than 3 percent in larger enterprises. UNECE (2001), p. 190. 3 Page 6 of 21 J. Köllő et al. Table 1 Foreign ownership in Hungary, 2003 Fraction employed in MNEs (percent of all person- Industrial composition of MNEs (percent months in the given industry) of all person-months in the MNE sector) All workers Skilled workers All workers Skilled workers Agriculture 5.0 6.1 0.8 0.5 Manufacturing 46.5 48.4 59.9 40.5 Construction 7.7 10.6 1.5 1.9 Energy, water, gas 57.5 55.6 3.3 3.1 Wholesale and retail trade 25.9 34.5 16.3 31.5 Finance and insurance 52.7 80.0 11.4 11.5 Services 20.7 24.3 6.8 11.0 Average/total 34.8 37.6 100.0 100.0 The data are annual averages observed in the estimation sample in 2003. The number of person-months amount to 8,704,486 (all workers) and 2,068,556 (skilled workers) differences between workers in MNEs versus domestic dummy for exporters. Alternatively, we use indicators of firms, on the one hand, and domestic firm employees investment and productivity. We gradually move from an hired from MNEs versus workers coming from other OLS equation only controlled for s to fixed-effects mod - jt domestic enterprises, on the other. els with all the covariates except for the P variables. According to the raw data, MNE employees earn more When the equation is estimated with OLS, the δ than twice as much as domestic sector workers. Persons parameter captures the ownership effect, plus the moving from domestic firms to MNEs gain 64 percent - employment-duration weighted average residual worker age points on average, while individuals who move to the and firm effects given personal characteristics P and X other direction lose 57 points. Measured with the median (Abowd et  al. 2006). The person fixed effects absorb the rather than the mean, the gain and the loss amount to 39 unobserved time-invariant mean “qualities” of work- and − 26 percentage points, respectively. The bottom ers. However, the estimated gap is still affected by the block suggests a substantial raw premium for outside experience in foreign-owned enterprises. In the forth- coming sections, we try to disentangle a ‘pure’ owner- Table 2 Descriptive statistics: wage levels and  wage ship-specific effect from differences in composition. changes of skilled workers Mean St. dev. Observations 4 Foreign‑domestic wage gap 4.1 Benchmark model Wage levels Our first model estimates the foreign-domestic wage gap Employer = MNE 309 288 7,937,675 in the following way: Employer = domestic firm 143 161 12,023,947 Wage change upon leaving an MNE for a domestic firm ln w = δF + [ϕP ] + αX + βY + γ V ijt ijt i it ijt jt Mean − 57 146 42,479 (1) + v + f + s + ε , i j jt ijt Median − 26 42,479 Wage change upon leaving a domestic firm for an MNE where w is the daily average (relative) earnings of per- ijt Mean 64 126 46,590 son i at firm j and month t , F is a dummy for being Median 39 46,590 employed in a majority foreign-owned firm, P and X are i it Wages of domestic firm employees with recent outside experience fixed and time-varying individual attributes, Y stands ijt Previous employer = MNE 171 193 963,075 for job-specific variables (like occupation and tenure), V jt Previous employer = domestic firm 118 122 3,557,788 denotes time-varying firm-specific covariates, v and f i j Wage in month t relative to the national average wage in month t, per cent are worker and firm fixed effects, respectively, and ε is ijt Person-months observed in 2003–2011 an error term. We allow for unobserved shocks to pro- The figures relate to persons moving from MNEs to domestic firms and vice ductivity by including sector–year interactions s . The jt versa. Mean earnings in the receiving firm is compared to the same worker’s firm-level variables are size, the capital-labor ratio, and a mean earnings in the sending firm. Wages are deflated with the national average wage in the same month The figures relate the mean earnings of domestic firm employees with previous outside experience to the mean earnings of incumbent domestic firm See Appendix 1: Fig. 2 for a box-and-whiskers plot of wage changes. employees, percent Wage gains from foreign ownership: evidence from linked employer–employee data Page 7 of 21 3 Table 3 Estimates of the foreign‑ domestic wage gap for skilled workers, 2003–2011 Specifications: (1) (2) (3) (4) (5) (6) Model A Foreign-owned 0.745 (25.5) 0.763 (20.3) 0.718 (31.6) 0.437 (23.4) 0.236 (28.6) 0.026 (3.5) aR2/within R2 0.260 0.329 0.414 0.480 0.238 0.103 Model B Always foreign-owned 0.794 (25.7) 0.817 (29.5) 0.772 (31.3) 0.507 (23.3) 0.307 (26.8) Temporarily foreign-owned 0.569 (8.1) 0.574 (8.7) 0.564 (12.6) 0.334 (5.2) 0.209 (14.1) 0.026 (3.5) Temporarily domestic 0.408 (10.5) 0.408 (10.5) 0.462 (11.3) 0.269 (8.7) 0.157 (11.6) aR2/within R2 0.267 0.327 0.422 0.482 0.242 0.103 Controls Sector × year Yes Yes Yes Yes Yes Yes Person No Yes Yes Yes Yes Yes Job No No Yes Yes Yes Yes Firm No No No Yes Yes Yes Person FE No No No No Yes Yes Firm FE No No No No No Yes All coefficients are significant at 0.01 level, t-values in parentheses. The standard errors are adjusted for clustering by persons and firms. Sample: 19,961,622 person- months belonging to 344,203 skilled workers in 119,580 firms. Singleton observations are excluded from the panel regressions Dependent variable: log daily wage in the given month relative to the national mean. Reference categories: employed in a domestic firm (Model A), employed in an’ always domestic’ firm (Model B). Controls: person, job and firm characteristics plus sector–year interactions. See Appendix 1: Table 12 for variable definitions. Specifications 5 and 6 include only time- varying covariates and worker and firm fixed effects. Estimation: all models were estimated with Stata’s reghdfe models employment-duration weighted average of the firm inclusion of firm size, the capital-labor ratio, and exports effects for the firms in which the worker was employed. bring the estimated MNE premium down to 0.437, while When both person and firm fixed effects are included, δ adding worker fixed effects reduces it to 0.236. Adding captures a pure ownership effect identified from worker firm fixed effects results in a major drop to only 0.026. flows between ownership categories, on the one hand, Controlling the worker fixed-effect model for TFP or and changes in ownership, on the other. It shows the value-added per worker instead of the firm fixed effects wage advantage of a foreign firm employee over a domes - yield estimates of 0.218 and 0.206, respectively. Includ- tic worker with similar observable attributes, controlled ing TFP into the set of firm controls in specification (4) for their average wages in the entire period of observa- results in a coefficient of 0.209. Including investment as tion, and also controlled for average wages of the firms well, which controls for the potential coincidence of posi- where they worked during the period of observation. tive productivity shocks and the hiring of high-quality For our multiple fixed effect estimations, we use a model labor, produces an estimate of 0.216. By contrast, adding proposed by Correia (2017), and implemented in Stata as firm fixed effects to specification (4) without including reghdfe. worker fixed effects decreases the estimate from 0.437 In Model A of Table  3, which measures MNE employ- to 0.036, clearly indicating that selection to acquisition ees’ wage advantage relative to domestic firm employees, drives the result of the 2FE model. the estimate rises from 0.745 log points to 0.763 after In Model B of Table  3, the observed person-months controlling for observed worker characteristics. The are classified by the ownership histories of employers. ‘Always domestic’ (the reference category) and ‘always foreign’ denote enterprises that did not change owner in 2003–2011. ‘Temporarily foreign’ and ‘temporar- The only exception would be observations on firms that, at the same time as changing ownership, would change all of their employees. We do not have ily domestic’ indicate the current majority owner of the such cases in the data. workers’ employer, for firms which underwent acquisi - Several methods have been developed in the last ten years (following tion at least once in 2003–2011. The ‘temporary foreign’ the pioneering work of Abowd et  al. 1999) to deal with two or more high dimensional fixed effects. The iterative methods (Cornelissen 2008; Martins dummy, for instance, is set to one for a person-month and Opromolla 2009; Guimaraes and Portugal 2010; Carneiro et  al. 2012; spent in a foreign-owned enterprise, which operated Mittag 2016) solve the problem by shuffling between the estimation of the slope and the intercept parameters. Balázsi et al. (2018) yield an alternative, which presses more on memory but runs faster. Early drafts of this paper, like Balázsi (2017) experimented with this method. With the size of the final data, iterative approaches turned out to be more productive. 3 Page 8 of 21 J. Köllő et al. under domestic ownership in a part of the observed The double difference (β − β ) − (β − β ) removes the 3 4 1 2 period. gap in the quality of F and D workers as measured with The estimates suggest that firms involved in takeovers their pre-entry wages. Since assignment to the groups and currently operating under domestic ownership pay compared is person-specific, and the firms do not change more than incumbent domestic firms (by 0.157 log points owner, we estimate the equation with pooled OLS. A in specification 5 where worker quality is controlled for). large battery of controls guarantees that we compare Switching firms currently under foreign ownership pay workers and firms with similar characteristics. lower wages than always foreign-owned companies by Note that we base the definition of a ‘new firm’ on its 0.099 log points. The gap between the coefficients for employment dynamics rather than its date of registration employment spells under ‘temporarily foreign’ and ‘tem- since the latter is often associated with break-ups, merg- porarily domestic’ ownership (0.052 log points) is an ers and acquisitions, rather than the birth of a new eco- alternative measure of how ownership changes affect the nomic actor. We rely on the fact that a medium-sized or wage. The magnitudes make it clear that switching firms large firm’s creation typically begins with hiring a small substantially differ from any of the incumbent categories. group of managers who arrange the start-up. This pre - paratory stage is followed by a ‘big bang’ when the firm 4.2 Exploiting inf ormation on new firms hires rank-and-file employees. We speak of a big bang As much as 94.8 percent of the firms in our estimation when a firm’s staff jumps from an initial level of L ≤ 5 t−1 sample did not change majority owner in the period cov- to L ≥ 50, or, from L ≤ 50 to L ≥ 300 within a month. t t−1 t ered by the data: 7.3 percent was foreign-owned, and 87.5 We found 519 such firms with no subsequent change of percent was domestic throughout 2003–2011. Rather ownership. Combined employment in these enterprises than merely neglecting the huge wage difference between jumped from 13 to 253 thousand (see Appendix 1: Fig. 3). them (as does the 2FE model), we exploit information Finally, the sample consists of 471,489 person-months on newly established and subsequently incumbent for- belonging to 8225 skilled workers hired by and staying eign and domestic enterprises. The critical event under until December 2011 in 366 new domestic and 147 new examination here is not the takeover of an existing firm, foreign-owned firms. but the birth of an incumbent firm. The analysis relates The results in Table  4 indicate a wage gap of 0.391 log to firms established after 2003 and staying under major - points between skilled workers in new MNEs versus new ity foreign or domestic control until 2011. We compare domestic firms—this is reasonably close to the 0.437 log the earnings of incumbent workers in these firms to the points gap estimated with a fully controlled OLS for all wages they earned before their entry. Formally, we esti- firms in Table  3, specification 4. New foreign firms’ work - mate the following difference-in-difference model: ers also earned more than their domestic counterparts before they entered the new firms by 0.245 log points on 0 0 1 1 ln w = β F + β D + β F + β D + Zγ + ε . ijt 1 2 3 4 ijt ijt ijt ijt ijt average. After deducting this difference from the post- (2) entry gap, an ownership-specific wage differential of 0.146 log points remains between incumbent workers in F and D are the acronyms for foreign-owned and domes- 0 0 incumbent firms. This point estimate falls between the tic firms. F and D are set to one for person-months pre- individual only and the two fixed-effects parameters, sug - ceding the worker’s entry date to a newly established F or 1 1 gesting a significantly stronger pure ownership-specific D firm. F and D are set to one for the months of service effect than the 2FE model. in a newly established firm. For instance, for a worker hired by a new foreign-owned company in month t = 37, 0 1 5 Lagged returns F = 1 if t < 37 and F = 1 if t ≥ 37. Z denotes controls 5.1 Are ex-MNE workers paid more in the domestic sector? listed in the notes to Table 6. In Eq.  (2), we compare workers in domestic firms, who β − β is the estimated wage difference between future 1 2 arrived at their employer from MNEs versus other F and D employees, whereas β − β measures the wage 3 4 domestic firms. The estimates are controlled for personal difference between the employees of new F and D firms. characteristics, current and past job attributes, tenure in the last job, months between the two jobs, selected indi- cators of the sending and receiving firms, and sector-year interactions. We retain firms with at least one ex-MNE Model B with added firm effects (Specification 6) is identical to Model B, as and one ex-domestic employee and exclude firms under - the always foreign indicator is absorbed by the added firm effects. Hence only going acquisition. a parameter on being temporarily foreign-owned can be estimated, which is identified only from firms going through acquisitions or divestments and accordingly, coincides with the parameter on the F dummy of Model A. Wage gains from foreign ownership: evidence from linked employer–employee data Page 9 of 21 3 Table 4 Wages before and after entry to new MNEs and new domestic firms Coeff. t-test Person-months Workers of domestic start-ups, before their entry 0 115,443 Workers of foreign start-ups, before their entry 0.245*** 8.3 146,585 Workers of domestic start-ups, after their entry − 0.014 0.3 84,018 Workers of foreign start-ups, after their entry 0.391*** 8.9 125,247 Double difference (point estimate, F-test) 0.146** 9.5 Significant at the **0.05, ***0.01 level. The t-values are based on standard errors adjusted for clustering by persons and firms OLS regression with dummies standing for the four distinct groups. Dependent variable: log daily wage in the given month relative to the national mean. Sample: 471,489 person-months belonging to 8225 skilled workers hired by and staying until December 2011 in 519 newly established firms (366 domestic and 147 foreign- owned). We considered a firm newly established if its staff number jumped from less than 5 to more than 50, or, from less than 50 to more than 300 within a month. Workers employed by new firms before their ‘big bang’, workers leaving the new firms and firms changing owner after the big bang are excluded. Controls: person, job and firm characteristics and sector-year interactions. See Appendix 1: Table 12 for variable definitions The lower blocks of the table display estimates on two ln w = αX + β F_After + β dL ijt it 1 ijt 2 jt (3) sub-samples distinguished along dL . Former MNE work- + β F_After dL + f + s + ε . 3 ijt jt j jt ijt ers who lost or left their jobs during mass dismissals ( dL < 0.5) had substantially higher wage advantages over F_After is a dummy set to 1 for workers who arrived ijt their ex-domestic counterparts (0.134 log points) than from foreign firms and 0 for workers coming from did those ex-MNE workers, who arrived from slightly domestic companies. dL = L /L measures the jt j,t+1 j,t−1 contracting, stable or expanding firms (0.06). change of employment in the sending firm between year t−1 and t + 1 , with t denoting the year when the worker 5.2 An overlapping cohorts model of lagged returns left the firm. The coefficient β measures how wages to MNE experience vary with employment dynamics of the sending domes- The estimates presented in the preceding sub-section are tic firms while the parameter β of the interaction term potentially subject to ability bias: workers returning to the F_After × dL captures the impact of dL on workers ijt jt domestic sector can be more productive wherever they arriving from foreign employers. The wage advantage of work. As it was put forward in the Introduction, addressing workers coming from MNEs over workers arriving from this problem by adding worker x fi ed effects to model (2) is domestic firms, conditional on employment dynamics of not a feasible option. Therefore, we estimate an alternative the sending firm, is given by β + β dL . Alternatively, we 1 3 jt model that compares the wages of domestic firm employees estimate the equation for two groups distinguished based with past and future experience in MNEs versus domes- on dL (lower or higher than 0.5), without the size-change tic companies other than their current employer. This and interaction terms. approach is close in spirit to models that study the wage Since we are interested in the within-firm wage dif - effect of incarceration by comparing past and future con - ferences between ex-MNE and ex-domestic entrants victs (Grogger 1995; Pettit and Lyons 2009; LaLonde and (rather than how a worker’s wage changes upon entering Cho 2008; Czafit and Köllő 2015) under the assumption a domestic firm), we include firm fixed effects, but not that the date of incarceration (mutatis mutandis the dates worker fixed effects. of entry to and exit from MNEs) can be treated as random. The upper block of Table  5 shows the results of the We can reasonably assume that future MNE workers are first variant of the model. The wage advantage of an ex- closer to former MNE employees in terms of unobserved MNE employee arriving from a firm where staff numbers characteristics than any control person selected from the did not change around the year of the worker’s separa- general population based on observables. A further advan- tion ( dL = 1) amounts to 0.057 log points, while it is esti- tage of this choice is a gain in sample size: 3,841,561 per- mated to be 0.074 points in case the sending firm was son-months instead of 797,261 in Model (2). closed or relocated ( dL = 0). We added a dummy indi- cating if the worker had arrived from another domestic firm but previously had some experience in one or more Workers who leave well-paying jobs in the MNE sector individually can MNEs. These workers have an advantage of 0.064 log be either negatively or positively selected. On the one hand, MNE employ- ees fired individually are likely to be less productive than the average. On the points. Only a part of these gaps results from within-firm other hand, those who manage to find a well-paid domestic job are predict - premia, as suggested by the differences between the spec - ably over-represented among voluntary quitters. The comparison of group- ifications with and without firm fixed effects. level estimates suggests that the first effect dominates: workers separating from their firms for reasons other than mass dismissals earn a lower lagged MNE premium on average. 3 Page 10 of 21 J. Köllő et al. Table 5 The wage advantage of  ex‑MNE workers in  domestic firms over  coworkers having arrived from  other domestic firms—regression estimates Firm fixed effects: No Yes Model A: entire sample Sending firm is MNE (F_After) 0.091*** (4.4) 0.077*** (4.9) Change of employment in the sending firm (dL) 0.010* (1.8) 0.011*** (3.0) Interaction term (F_After × dL) − 0.028** (2.4) − 0.024** (2.3) MNE experience before entry to the sending firm (dummy) 0.059*** (6.0) 0.038*** (5.0) Number of observations 723,421 722.913 2 2 aR /within R 0.461 0.288 Model B: workers arriving from mass layoffs and all workers Employment change in the sending firm: L /L ≤ 0.5 t+1 t−1 Sending firm is MNE 0.134*** (3.3) 0.109** (2.4) MNE experience before entry to the sending firm (dummy) 0.068*** (2.9) 0.056*** (2.7) Number of observations 153,482 153,213 2 2 aR/within R 0.479 0.277 Entire sample Sending firm is MNE 0.060*** (4.1) 0.049*** (5.1) MNE experience before entry to the sending firm (dummy) 0.058*** (6.0) 0.037*** (5.0) Number of observations 723,421 722,913 2 2 aR/within R 0.461 0.288 Significant at the *0.1, **0.05, ***0.01 level. The standard errors are adjusted for clustering by persons and firms. Sample: 723,421 person-months belonging to 96,277 skilled workers in 19,449 domestic firms, who had arrived from MNEs versus other domestic firms. 508 singleton observations are excluded from the equation with firm fixed effects. Estimation: Stata reghdfe. Change of employment in the sending firm: L /L , where t is the year of the worker’s separation. Controls: person t+1 t−1 and job controls, contemporaneous and lagged firm-level controls, as listed in Table 12. Additional controls are completed tenure in the sending firm, dummy for unobserved tenure, months between exit from the sending firm and entry to the receiving firm, one-digit sectoral affiliation of the sending and receiving firms and year dummies We define a collectively exhaustive classification mak - We measure the effect of foreign sector experience with ing a distinction between domestic firm employees with the double difference ( β − β ) − (β − β ) or equivalently 1 2 3 4 past MNE experience in month t (PF), workers with (β − β ) − (β − β ). The model controls for unobserved dif - 1 3 2 4 future but no past MNE experience (FF), workers with ferentials in worker quality as long as workers’ wages with prior experience in other domestic firms and no MNE future outside experience can be treated as a counterfactual experience (PD) and workers with future domestic sec- for the wages of workers with prior experience. However, it tor experience and none of the types mentioned earlier cannot address the possibly endogenous selection of work- (FD). Incumbent workers who had no contact with other ers to separation from their previous employers. employers in 2003–2011 constitute the reference cate- The results in Table  6 show that workers with past gory. The sample we work with consists of domestic firm MNE experience earn more by 0.112 log points than their employees in companies employing at least one worker counterparts with outside domestic experience. This dif - belonging to the categories mentioned above and one ference overestimates the returns to foreign sector expe- incumbent worker. We restrict the analysis to 2005–2009 rience since those domestic workers who are on their way to have sufficient observations on past and future experi - to an MNE also earn more by 0.043 log points than those ences outside the workers’ current firms. about to leave for another domestic employer. Ex-MNE We regress log wages on the respective dummies and workers earn more than future MNE employees by 0.048 person, job, and firm-specific controls plus sector-year log points while those with outside domestic experience interactions. Choosing incumbents as the reference cat- earn less by 0.021 log points than their counterparts, egory and denoting the controls with Z, the estimated leaving for another domestic firm later. equation with or without firm fixed effects (v ) is: Using these estimates, we can approximate the return to MNE work experience as the double-difference equal ln w = β PF + β PD + β FF ijt 1 ijt 2 ijt 3 ijt to 0.069 log points. The two models’ main results aimed (4) + β FD + v + Zγ + ε . 4 ijt j ijt at measuring lagged wage effects (Tables  5 and 6) are Wage gains from foreign ownership: evidence from linked employer–employee data Page 11 of 21 3 similar. The first model identified a 0.060 log points Table 6 Wage difference between  domestic workers with/ without outside work experience advantage on the part of the median worker coming from an MNE over a worker arriving from domestic company Dependent variable: log daily (Table 5 bottom block). wage While the main results are close to each other, some OLS Firm fixed effects details differ in the two models. The wage difference Coefficients (t-test) between workers arriving from foreign-owned versus Past MNE experience (PF) 0.060*** (4.0) 0.005 (1.0) other domestic firms appear to be more prominent here: Future MNE experience (FF) 0.012 (0.9) 0.001 (0.6) 0.112 points as opposed to 0.060 points in Table 5, model Past domestic experience (PD) − 0.052*** (5.6) − 0.030*** (7.3) B, the estimate for all workers. Second, when we rees- Future domestic experience (FD) − 0.031*** (3.5) − 0.016*** (3.6) timate the model by adding firm fixed effects (column 2 Differences by type of outside experience (F-test) of Table 6), the contrasts fade away: the within-firm wage Past MNE − past domestic 0.112*** (61.6) 0.035*** (53.4) differentials between the PF-FD groups are smaller, and Future MNE − future domestic 0.043*** (15.0) 0.017** (5.5) the double-difference drops to only 0.018 log points. Past MNE − future MNE 0.048** (6.3) 0.004 (0.4) Unlike our first model, the second one suggests that the Past domestic − future domestic − 0.021** (4.2) − 0.014*** (10.0) lagged MNE premium predominantly stems from past Double difference 0.069*** (15.0) 0.018*** (11.5) and future MNE employees’ crowding in high-wage aR2/within R2 0.453 0.342 domestic firms. Regression estimates. The reported coefficients are significant at the *0.1, **0.05, ***0.01 level. Unmarked coefficients are not significant at the 0.1 level. 6 Spillover effects Sample: 3,841,561 person-months belonging to 153,323 persons and 18,510 6.1 E ec ff t of ex -MNE peers on incumbent domestic firm firms. The sample covers domestic firm employees in firms employing at least one worker with past or future experience in foreign-owned or domestic firms, employees and one incumbent worker. The coefficients measure wage advantages relative We estimate the effect of ex-MNE peers on incumbent to incumbent workers. Observations for 2005–2009 are used. Estimation: workers’ wage, that is, for domestic firm employees who reghdfe without and with firm fixed effects. The standard errors are adjusted for clustering by persons and firms. Controls: person, job and firm controls, and did not leave their firm in the observed period. Their sector–year interactions wages are regressed on a set of controls and variables measuring the share of workers with previous outside experience within the worker’s company and skill cat- The fixed-effects panel equations summarized in egory. We deviate from Poole (2013) in that we also study Table  7 regress the log wages of incumbent skilled how skilled incumbents’ wages respond to the pres- domestic workers on the share of workers with out- MNE,uskilled ence of less skilled ex-MNE peers. Share , for side experience within the worker’s firm and skill jt instance, measures the ratio of unskilled employees with group. The estimated own effect for skilled workers recent MNE experience. in a medium-sized or large firm ( θ = 0.074) implies F3 MNE,middling MNE,skilled MNE,unskilled ln w = θ Share + θ Share + θ Share ijt F 3 F 2 F 1 jt jt jt domestic,middling domestic,skilled domestic,unskilled (5) + θ Share + θ Share + θ Share + αX D3 D2 D1 it jt jt jt + βY + γ V + v + s + ε . ijt jt i jt ijt The fact that the Hungarian administrative panel is only a 50% sample on We estimate the model including only worker fixed the individual level, has some unfortunate implications for the spillover esti- effects, which also absorb the firm effects since the esti - mates. We observe only around half of any given firm’s labour force—esti - mates relate to incumbent workers. The controls are mates instead of the actual shares. As not observing ex-MNE workers has the same, 50% probability as those with no such experience, in large firms we will identical to those used in Eq. 1. We restrict the time win- only experience extra noise in the share variables. This noise in our explana - dow to 2005–2011 to leave time for the accumulation of tory variable will attenuate the estimated θ parameters, biasing them towards es an ex-MNE stock. The equations are estimated separately zero. However, in firms with a small number of workers, if the average share of given type is also low, we may mistakenly not observe any variation in our for smaller (11–50) and larger (50+) firms, taking into variables of interest, while we should. If a firm which previously never had a consideration the higher risk of measurement error in foreign worker acquires a skilled manager with foreign experience, and we do small establishments. not observe the given person, observations at this firm will not have variation in the share of skilled ex-MNE workers, thus this firm will not contribute to the identification of our parameter of interest in our model with firm fixed The difference may stem from differences in the samples and the periods effects. Considering these two processes we not only keep solely the firms covered by the data as well as from the influence of experience in MNEs other with at least 10 employees, as in most of the paper, but also focus on larger than the sending firm. This effect is directly estimated in Table  5 but not in (50+ firms), where the (predicted) share variables are less volatile. Table 6. 3 Page 12 of 21 J. Köllő et al. Table 7 The effect of  coworkers with  recent outside  work experience on  the  wages of  skilled incumbents in  domestic firms 2005–2011 Share of coworkers with recent MNE experience within skill Share of coworkers with recent experience groups in other domestic firms within skill groups Unskilled Middling Skilled Unskilled Middling Skilled Notations in Eq. 3: θ θ θ θ θ θ F1 F2 F3 D1 D2 D3 All firms 0.012 (1.5) 0.003 (0.4) 0.042*** (3.9) 0.015** (3.1) 0.010 (1.5) − 0.031*** (-4.5) Firms employing > 50 workers 0.000 (0.0) 0.028 (1.2) 0.074*** (4.3) 0.005 (0.7) 0.042*** (2.8) − 0.027** (− 2.1) Significant at *0.1, **0.05, ***0.01 level. The t-values are based on standard errors adjusted for clustering by persons and firms. θ is significantly larger than θ , θ and F3 F1 F2 θ . Sample: 3,730,789 person-months in 116,204 firms in the full sample, 2,474,830 person-months in 77,411 firms in the 50+ sample. Dependent variable: log daily D3 wage in the given month relative to the national mean. Controls: person, job and firm characteristics, sector–year interactions, and worker fixed-effects Table 8 Mean within‑firm share of coworkers with past MNE experience (percent) Skilled incumbents in domestic firms Skilled domestic firm employees without MNE experience Share of coworkers with MNE Number of workers Share of coworkers with MNE Number of workers experience experience Unskilled 7.0 38,355 13.3 73,320 Medium skilled 9.3 53,896 15.4 103,871 Skilled 9.0 55,900 14.6 107,250 Incumbents are workers, who had only a single domestic-owned employer in 2003–2011. The mean within-firm shares are weighted with firm size and relate to 2003–2011 that a one-standard-deviation difference in the share of mean within-firm share of skilled MNE-experienced high skilled ex-MNE employees (0.18) shifts the wages peers amounts to 9 percent in the case incumbents of skilled incumbents up by 1.3 percent. Having more instead of 14.6 percent in the case of their non-incum- skilled peers with outside experience in the domestic sec- bent counterparts—a predictable pattern since incum- tor has no effect. bents are more likely to be found in firms with low In evaluating the cross effects, one should consider the labor turnover. relevant range in the share of ex-MNE workers. While a A higher share of ex-MNE peers increases the likeli- jump from zero to 50 or 100 percent in the share of ex- hood of personal contacts, thereby assisting the diffusion foreign workers within the unskilled or medium-skilled of MNE-based skills within the firm. At the same time, workforce is beyond the realm of reality, which ren- the typical incumbent worker spends more time with the ders the spillover effect to be weak, this can happen in firm, so she has a better chance to absorb the imported the high skilled category. Domestic firms employing 50 knowledge. Because of the potential bias in either direc- workers have 7 high skilled workers on average. Hiring tion, we reestimate the spillover model for all domestic two managers or professionals with foreign sector experi- workers, including firm fixed effects on top of the worker ence can increase the ex-MNE share from zero to almost fixed effects in the model to ensure that it identifies 30 percent overnight, which implies a 0.022 log points within-firm impacts. wage increase for skilled incumbents. The results for firms with more than 50 workers and all firms are presented in Table  9. Starting with the for- mer: the own effect (0.060) is slightly lower than the 6.2 R eestimating spillover effects for all domestic firm point estimate for incumbents (0.074 in Table  4). Less employees skilled ex-MNE workers exert a weak effect—the respec - Incumbents in our data account for only 22 percent of tive coefficients are only significant at the 5 percent level. the workers ever employed in the domestic sector and Having more skilled peers with recent outside experience 34 percent of the workers never employed outside the in domestic firms do not affect wages positively at all. The domestic sector. The estimates of spillover effects using estimates for all firms are much lower and insignificant at their sample may be biased because their exposure to 5 percent level. The inward bias is probably explained by peers with MNE experience differs substantially from that of the average worker. As shown in Table  8, the Wage gains from foreign ownership: evidence from linked employer–employee data Page 13 of 21 3 Table 9 The effect of  coworkers with  recent outside  work experience on  the  wages of  skilled workers in  domestic enterprises 2005–2011 Share of coworkers with recent MNE experience by their level Share of coworkers with recent experience of skill in other domestic firms by their level of skill Unskilled Middling Skilled Unskilled Middling Skilled Notations in Eq. 3: θ θ θ θ θ θ F1 F2 F3 D1 D2 D3 All domestic firms 0.007 (0.9) 0.013 (1.5) 0.020* (1.9) 0.016*** (3.2) 0.024*** (3.4) − 0.037*** (− 4.5) Domestic firms 0.006 (0.5) 0.057** (2.1) 0.060*** (3.4) 0.002 (0.2) 0.064*** (3.4) − 0.019 (− 1.3) employing > 50 workers Significant at *0.1, **0.05, ***0.01 level. The t-values are based on standard errors adjusted for clustering by persons and firms. θ is significantly larger than θ and F3 F1 θ , but not θ . θ is significantly larger than θ D3 F2 F2 F1 Sample: 3,731,548 person-months belonging to skilled workers in 116,249 firms in full sample, 2,474,843 person-months in 77,412 firms in the 50+ sample. Dependent variable: log daily wage in the given month relative to the national mean. Controls: person, job and firm characteristics, sector–year interactions, worker and firm fixed-effects the noisy measurement of the F and D ratios in smaller an effect on skilled wages yields further support to the enterprises. learning hypothesis. The estimated spillover effect might seem economi - cally insignificant, but it is actually stronger than those 7 Two notes on differences by skills and sectors we know from the literature. The study of Poole (2013)— Throughout this paper, we focused on skilled workers which is closest to ours concerning method, sample char- mainly because we are interested in possible knowledge acteristics, and industry coverage—estimated that at the flows from foreign-owned to domestic firms, the traces average wage for a typical domestic worker, a 10 percent- of which are easier to find in the skilled labor market. age points increase in the share of former MNE work- We nevertheless estimated all our models for less-skilled ers increased incumbents’ wages by $23 per year. This workers and found that the effects of interest are smaller amount could buy a little more than one Starbucks solo and, in many cases statistically insignificant. Appendix 1: espresso a month in Rio de Janeiro in 2015. The compara - Fig. 1 illustrates this point. The figure compares the esti - ble estimate for skilled incumbents in our sample is $139 mates of the wage gap model (Table 3, model A) to simi- a year, which could buy 5.2 cups of Starbucks espresso a lar ones for unskilled and medium-skilled workers. The month in Budapest at 2015 prices. latter are very close to each other and amount to about Learning from ex-MNE peers is only one explanation 0.4 log points in the uncontrolled model, less than 0.1 in for the effect we identify. A firm’s effort to maintain its the panel regression with worker FE and less than 0.02 in wage ladder after hiring a high-wage ex-MNE worker the 2FE model. could occasionally motivate a firm to increase the wages Data available in the Labor Force Survey (Tables  13, of other employees. Still, we do not find this explanation 14 of Appendix 1) furthermore suggest that a part of convincing when spillover is observed in tens of thou- the MNE premium compensates unskilled workers for sands of firms. Why would so many domestic firms hire non-wage disamenities. Overtime work and afternoon high-wage workers from MNEs if this decision implies and night shifts are about twice as likely to occur among further wage growth without an underlying rise in pro- low and medium-skilled MNE employees compared to ductivity? A positive selection of all workers to firms their domestic counterparts. There is a smaller but simi - hiring from MNEs can also raise the average wage of larly signed difference concerning work on Saturdays coworkers with no MNE experience. However, our find - and Sundays. Furthermore, low skilled workers have a ings controlled for worker fixed effects and/or relating higher probability of becoming unemployed in foreign- only to incumbents are free of this kind of bias. Last but owned than domestic firms. The data does not indicate not least, the finding that only skilled ex-MNE peers have The calculation is based on the estimated own effect (0.074), the mean monthly earnings of skilled domestic firm employees in 2011 (236,078 Ft) and Skilled workers account for 25 percent of the total population observed in an average exchange rate of 225 Ft/$ in 2011 (National Bank, http://mnbko the source file. 15 per cent is unskilled (never worked in an occupation requir - zepar folya m.hu/arfol yam-2011.html). We could find Starbucks solo espresso ing nay kind of qualification) and 60 percent is classified as middling (worked prices for 2015 on the websites of local shops in Rio and Budapest: $1.92 and in skilled jobs but not in ones requiring tertiary educational attainment). $1.43, respectively. 3 Page 14 of 21 J. Köllő et al. ownership-specific differences of this kind among highly (Bernard and Sjoholm 2003). The implications of skills skilled workers. accumulation versus efficiency wages for the foreign- Table  10 summarizes point estimates of the wage gap, domestic wage gap and the wage loss from separation are lagged returns, and spillover effects from our preferred observationally identical. However, efficiency wages in model specifications for manufacturing and all other sec - MNEs do not imply that ex-MNE employees earn a pre- tors labeled ‘services’. The foreign-domestic wage gap is mium over the receiving domestic firm’s going wage rate more substantial in services than manufacturing, and the and exert influence on the earnings of their peers. lagged returns are broadly similar or somewhat larger in Third, a set of findings like this is likely to emerge only if services. By contrast, the spillover effects are estimated to MNE workers accumulate both general and r fi m-specific be stronger in manufacturing. We do not go to the details knowledge. As outlined in Becker’s (1962) seminal paper, of the between-sector differences. We only note that the in the case of general skills acquired through on-the-job returns to MNE experience are not restricted to the man- training, productivity and wages move in tandem. Work- ufacturing sector heavily over-represented in the related ers accumulating a substantial stock of general skills in literature. one firm are expected to earn higher-than-average wages in other firms. As far as general skills develop through informal communication between coworkers, their pres- 8 Discussion ence also tends to have a spillover effect. However, we do We interpret the coincidence of an MNE premium, sub- not expect that separation from an MNE induces a wage stantial wage loss from separation, lagged returns to loss in this scenario. MNE experience, and wage spillover as a signal of knowl- If the acquired knowledge is purely firm-specific, and edge flows from FDI to domestic firms. In such a sce - the risk of voluntary separation (motivated by factors nario, workers acquiring general and firm-specific skills other than between-firm wage differentials) is zero, then in the modern environment of MNEs are expected to the firm pays the going market wage before, during and earn more than their domestic counterparts. The specific after the period of skills accumulation. These skills lose components in their skills imply that MNE workers lose a their value with separation without an impact on sala- part of their wage advantage in case of involuntary sepa- ries. Pre-separation and post-separation wages are equal, ration. The general component in their skills gives rise to post-separation wages do not exceed the host firm’s aver - wage advantages in their new, domestic firm and tends age level, and they do not exert influence on the earnings to influence their peers’ productivity. The simultaneity of coworkers. In the likely case of non-zero risk of volun- of these symptoms calls into question some alternative tary quits, the firm will share in the costs and benefits of explanations, of which we discuss three ones. training, which implies lower wages in the accumulation First, the finding of a contemporaneous MNE pre - phase and higher wages afterward as long as the worker mium even after controlling for worker fixed effects calls into question that the foreign-domestic wage gap is fully explained by the crowding of high productivity workers in foreign-owned firms. Similarly, in a comparison of Table 10 Selected estimates by sectors domestic and foreign-owned start-ups, we find a sizable Manufacturing Services MNE premium even after controlling for their workers’ pre-entry wages. Contemporaneous MNE premium Second, intense human capital accumulation is admit- All firms, worker FE 0.152 0.236 tedly not the only potential source of an MNE premium, New, incumbent firms, DiD 0.135 0.232 with the most important alternative being efficiency wage Lagged MNE premium in domestic firms setting. MNEs may try to prevent leakage of information Sending firm is MNE, dL < 0.5, OLS 0.135 0.133 through labor turnover by paying a premium above the Sending firm is MNE, dL < 0.5, firm FE 0.056 0.044 market level (Fosfuri et  al. 2001). Their limited knowl - Overlapping cohorts estimate, DiD 0.027 0.072 edge of the local labor market and capital-labor relations Spillover effect, firms L > 50 employees may urge them to pay high wages and share a part of their On incumbents 0.088 0.057 revenues with workers. Furthermore, they may try to On all workers with no MNE experience 0.069 0.050 compensate their employees for a higher labor demand All coefficients are significant at the 0.01 level. The coefficients were estimated volatility (Fabri et al. 2003) or a higher plant closure rate separately for the two sectors Wage gains from foreign ownership: evidence from linked employer–employee data Page 15 of 21 3 stays with her employer. In this case, post-training invol- foreign-domestic wage gap from acquisitions. Thanks to a untary separations imply a wage loss, but we continue rich and big data set, we could compare how workers are not to expect lagged returns and spillover effects. selected to new MNEs and domestic firms, and identify a The literature emanating from Becker’s benchmark substantial wage differential between them. In the analy - models has been trying to reconcile the theory of on- sis of lagged returns and spillovers, we drew attention to the-job training with a series of empirical observations trade-offs between model quality and unbiasedness of the inconsistent with the extreme scenarios. A series of samples on which the models can be estimated. empirical findings and ample everyday experience sug- As we find substantial wage effects attributed to for - gest that (i) most skills are general, or at least sector eign ownership both in the short-run and long-run, even rather than firm-specific (ii) enterprises are willing to after controlling for potential biases as much as possible, pay for general training, and (iii) involuntary separa- we believe that the presence and significance of knowl - tions typically imply a loss. Acemoglu and Pischke edge transfer from MNEs is beyond doubt. Therefore, (1998) demonstrate that in a variety of market settings we argue that FDI coming from more developed coun- such as a compressed wage structure, substantial hir- tries exert positive effects on the receiving countries’ ing costs, information asymmetry, and other labor labor markets both through direct, and indirect channels. market imperfections, general skills are rewarded as Exploring whether these gains outweigh the potential if they were partly specific. The “skill-weights” model drawbacks could be the focus of future research on the of Lazear (2009) hypothesizes that skills are predomi- topic. nantly general, but firms attach different weights to Acknowledgements their components. A worker who leaves a firm will We thank Tibor Czeglédi, Éva Czethoffer, Endre Szabó, Melinda Tir and Kitti have a difficult time finding another employer that Varadovics for constructing the database from which we have drawn our sam- ples and Péter Elek, Miklós Koren, Gábor Kőrösi, László Lőrincz, László Mátyás, can make use of all the skills he acquired at the send- Frank Neffke, Álmos Telegdy, two anonymous reviewers and participants of ing firm. The limits of transferability impose a cost on several seminars and conferences in Budapest, Braga and Thessaloniki for their mobile workers, so the workers are unwilling to bear helpful comments and advice. the full cost of training, and the costs and benefits will Authors’ contributions be shared. Such a setting is likely to produce all of the LB’s contribution was the calculation of the contemporaneous wage gap, four outcomes observed in our data. and the discussion of estimation issues related to multi-way fixed effects models, alongside related technical issues. IB handled a large part of data management, worked out part of the methodology, and calculated the spillover effects. JK managed the project, worked out the paper’s structure, the 9 Conclusions methods and the identification of the various effects, calculated the lagged wage effect, and wrote most of the text. All authors read and approved the We found that high skilled MNE workers earn substan- final manuscript. tially higher wages than their domestic counterparts. They lose a part of their wage advantage after leaving the Funding Boza and Köllő gratefully acknowledge the financial support of the’ foreign-owned sector but, even so, they earn more than Knowledgeflows’ project of the European Research Council coordinated their domestic sector colleagues with no MNE experi- by Miklós Koren at the Central European University, Budapest. Grant ID: ence. Their presence in domestic firms exerts a positive FP7-IDEAS-ERC#313164. effect on the wages of their peers, who had no contact Availability of data and materials with foreign-owned firms or had no recent outside work Due to the size and sensitivity of the data, access to it is provided exclusively experience at all. for academic purposes by the Databank of the Centre for Economic and Regional Studies, Hungarian Academy of Sciences. We authorized the Data- The direct and indirect wage returns to work experi - bank to make the estimation samples and program codes available to referees ence in MNEs are large in Hungary, similar to less devel- (while maintaining their anonymity) and researchers willing to replicate the oped countries analyzed in the literature. The positive results. Access can be initiated at adatkeres@krtk.mta.hu. wage effects are not restricted to the manufacturing sec - Competing interests tor, which is in the focus of attention in the research on The authors declare that they have no competing interests. FDI.The estimates suggest that the effect of MNE expe - Author details rience on domestic sector wages is strongly affected by Institute of Economics, Hungarian Academy of Sciences, Budapest, Hungary. between-firm variance, that is, the higher-than-average 2 Central European University, Budapest, Hungary. wages of domestic firms connected with the MNEs via labor turnover. Appendix 1: Figures and tables Finally, the results draw attention to the difficulties See Figs. 1, 2, 3 and Tables 11, 12, 13, 14 and 15. of identifying a ‘pure’ ownership effect. The non-ran - dom selection of firms flaws the identification of the 3 Page 16 of 21 J. Köllő et al. 1 2 3 4 5 6 Specifications 95% CI Skilled 95% CI Middling 95% CI Unskilled Fig. 1 Estimates of the foreign-domestic wage gap by skills. Specifications: (1) sector–year interactions; (2) + person controls; (3) + job controls; (4) + firm controls; (5) + worker fixed effects; (6) + firm fixed effects. The confidence intervals are based on standard errors adjusted for clustering by persons and firms. On the sets of controls and the definition of skill levels, see Appendix 2: “Data ” points, zero wages and missing covariates. 98.5 percent of the workers belong to a single connected group. Spe- cial subsamples have been selected for the study of new Appendix 2: Data and key variables firms, lagged returns and spillovers. Data access: Data for the estimation samples and Stata Data do files are available on request. The original data set Starting sample: 50 percent random sample drawn from called Admin2 is also available via remote access to the Social Security Numbers (SSN, Hungarian TAJ) valid on Databank’s servers. Write to adatkeres@krtk.mta.hu for January 1, 2003. SSN holders aged 5–74 were retained. requesting access to the data. Note that the size of the Data held by the Pension Directorate (ONYF), the Tax original data ranges between 60 and 120 Gbytes, depend- Office (NAV), the Health Insurance Fund (OEP), the ing on the amount of information stored in special mod- Office of Education (OH), and the Public Employment ules that you want to merge to the base file. The files are Service (NMH) were merged and anonymized by the in Stata16 format. R and Python codes are allowed. National Information Service (NISZ). The original data consisted of payment records with start and end dates, a type-of-payment code and amounts received by the person. Employers were identified by ONYF and their annual financial data were provided by NAV. The data ‘When a group of persons and firms is connected, the group contains all the workers who ever worked for any of the firms in the group and all the firms was transformed to a fixed format monthly panel data at which any of the workers were ever employed. In contrast, when a group set by the Databank of the Institute of Economics of the of persons and firms is not connected to a second group, no firm in the first Hungarian Academy of Sciences. group has ever employed a person in the second group, nor has any person in the first group ever been employed by a firm in the second group. From Estimation sample: Workers employed with a labor an economic perspective, connected groups of workers and firms show the contract at least once in a foreign or domestic private realized mobility network in the economy. From a statistical perspective, con- enterprise the maximum employment level of which nected groups of workers and firms block-diagonalize the normal equations and permit the precise statement of identification restrictions on the person exceeded the 10 workers limit at least once in 2003–2011. and firm effects.’ Abowd et al. (2006). We removed workers and firms with less than two data log points Wage gains from foreign ownership: evidence from linked employer–employee data Page 17 of 21 3 -1 -2 FD DD FF DF excludes outside values Fig. 2 Shifts between sectors and wage change. The data relate to 307,874 shifts by skilled workers between ownership sectors in 2003–2011. F and D denote foreign-owned and domestic firms, respectively, in chronological order. The boxes display the interquartile ranges of log wage changes, with a horizontal line within the box indicating the median, and the whiskers showing the highest and lowest adjacent values. Heavy outliers are excluded. Wage change is measured as ln(w /w ), where w and w are average earnings (normalized for the national mean) in the job 1 0 1 0 spells after and before the shift, respectively Start-ups Other booming irms 150 800 -100 -50 0 50 100 -100 -50 0 50 100 Months before/after the big bang Months before/after the big bang Fig. 3 The mean size of firms classified as newly established. The data relate to 544 firms the size of which jumped from less than 5 to more than 50, or from less than 50 to more than 300 within a month (big bang). Firms changing majority owner after the big bang are excluded Mean size (number of employees) Log wage change Mean size (number of employees) 3 Page 18 of 21 J. Köllő et al. Table 11 Descriptive statistics Table 12 Pooled OLS results for Eq. 1, specification 4 Mean St. dev. Coefficient t-value Male 0.619Majority owner Age 37.9 10.4 Foreign 0.437 20.4 Log health expenditures/national average wage − 2.08 1.8 Person controls Receives disability pension/payment 0.006 Male 0.154 19.1 Receives care benefit 0.008 Age 0.032 15.2 Log regional unemployment rate − 2.66 0.386 Age squared/100 − 0.033 13.8 Central Hungary including Budapest 0.458 Months spent non-employed in 2003–2011 − 0.003 31.8 Tenure is unobserved 0.398 Receipt of disability payment − 0.373 23.3 Tenure (months) 13.44 19.0 Receipt of care allowance − 0.207 12.1 Top manager 0.051 Health expenditures (log) 0.002 7.3 Other manager 0.271Job controls Professional 0.299 Tenure if observed 0.001 4.3 Other white collar 0.112 Tenure is unobserved 0.138 12.2 Skilled blue collar 0.025 Spell lasting for 1 day 0.354 3.1 Assembler, machine operator 0.169 Top manager Ref. Elementary occupation 0.012 Other managers − 0.062 1.6 Agriculture 0.025 Professional − 0.016 0.4 Manufacturing 0.277 Other white collar − 0.298 9.0 Construction 0.061 Skilled blue collar − 0.607 28.8 Trade 0.278 Assembler, machine operator − 0.728 18.7 Finance 0.126 Laborer in elementary occupation − 0.821 18.8 Energy 0.018 Regional unemployment rate (log) − 0.063 5.2 Other services 0.215 Budapest 0.142 11.3 Foreign 0.397 Firm controls Domestic 0.603 Firm size (log) 0.086 7.3 Capital-labor ratio (log) 0.041 9.3 Firm size (log) 4.67 2.32 Fixed assets per worker (log) 7.92 1.81 Exporter 0.185 9.4 Exporter 0.521 Constant − 1.650 24.0 Adjusted R-squared 0.479 Skilled workers, estimation sample for the wage gap model (Eq. 1) Number of observations 19,961,622 Each variable covers 19,961,622 person months. The spells belong to workers employed at least once in a firm, the size of which exceeded the 10 workers limit Skilled workers, 2003–2011 at least once in 2003–2011. Public sector and state-owned firms are excluded. Dependent variable: log daily earnings relative to the national mean. For the Note that other samples used in the paper have been drawn from this source file exact definition of the variables see Appendix 2: "Data ". The coefficients of 63 sector–year dummies are not shown. The standard errors are adjusted for clustering by persons and firms. All coefficients are significant at 0.01 level except n) not significant at 0.1 level Key variables Wage: The daily wage figure used in the paper was calcu - lated as monthly earnings divided by the number of days by dividing them with the national average wage in the covered by pension insurance (‘working days’ henceforth) given month, as measured in the starting sample. Source: in the given month. Multiple payments made by the ONYF. same employer to the same person within a month were Foreign-owned firm, MNE: dummy variable set to 1 summed up. Working days belonging to these payments for firms majority owned by one or more foreign own - were also summed up but capped at 30 or 31 days. In the ers. Ownership shares are measured as fractions of sub- case of multiple job holders the wage figure belongs to scribed capital. Source: NAV. the highest paying job. We normalized the wage figures Wage gains from foreign ownership: evidence from linked employer–employee data Page 19 of 21 3 Table 13 Incidence of atypical work schedules in foreign and domestic enterprises Level of  education Domestic Foreign Domestic Foreign b b Shift workOvertime work Low 27.5 58.2 14.4 32.0 Middling 22.4 41.2 12.0 24.3 High 4.4 4.5 4.0 7.6 c c Work in the afternoon Work in the night Low 14.4 29.1 8.1 20.3 Middling 18.6 33.1 9.4 22.1 High 17.7 14.4 7.1 6.6 c c Work on Saturdays Work on Sundays Low 26.3 29.3 16.9 17.6 Middling 35.4 36.6 21.2 24.0 High 26.8 18.9 16.7 12.7 2003–2011, percent Low = primary school attainment, High = college or university, Middling = rest Source: Wage Surveys, 2003–2011, private sector. Firms are classified on the basis of their majority owners. The data indicate the percentage share of employees receiving shift pay and overtime pay, respectively. Authors’ calculation Source: Labor Force Surveys, 2003 Q1–2011 Q4, excluding public administration, education, health and social services. The data indicate the percentage share of employees working in the respective periods at least occasionally. Authors’ calculation classified as high skilled. Persons never employed outside Table 14 The effect of  ownership on  the  probability of becoming unemployed—logit odds ratios occupations 6 and 7 are classified as low skilled. Other persons are classified as medium skilled. Source: ONYF. Educational attainment Total time spent non-employed: The number of months Low Middling High out of employment in 2003–2011. Source: ONYF. Disability payment: dummy variable, with 1 standing Employer: MNE 1.199*** (2.57) 0.971 (0.50) 1.061 (0.89) for any kind of transfer (pension or allowance) received Female 0.916 (1.40) 1.029 (0.55) 1.149*** (2.44) on the basis of permanent disability (rokkantnyugdíj, rok- Age 1.012 (0.71) 0.941*** (3.85) 0.919*** (5.01) kantsági járadék). Monthly data. Source: ONYF. Age squared 0.999** (2.06) 1.000*** (3.08) 1.000*** (4.25) Care allowance: dummy variable, with 1 standing for Tenure (years) 0.894*** (9.27) 0.886*** (13.9) 0.895*** (10.0) any kind of benefit received by the observed person on Number of observations 82,638 205,597 227,074 the basis of raising children (tgyás, gyed, gyes, gyet) or Pseudo R2 0.076 0.067 0.068 taking care of disabled relatives (ápolási segély). Monthly Wald chi (51) 617.4*** 958.0*** 763.8*** data. Source: OEP, ONYF. Significant at the **0.5 and ***0.01 level Health expenditures: Expenditures and costs regis- Discrete time survival model, logit form, following (Jenkins 1995). Estimated tered by the National Health Insurance Fund (OEP). for the pool of 28 quarterly waves of the Labor Force Survey in 2003–2009. The estimation excludes the crisis period (2010 and 2011). Sample: employees. The items include total amount paid for OEP-sup - Dependent variable: 1 if the person was ILO-OECD unemployed in wave t + 1 ported medicine and the costs of OEP-supported ser- and 0 otherwise. The coefficients of 19 county dummies and 27 survey wave dummies are not shown vices/treatment provided by district doctors, specialists Low = primary school attainment, High = college or university, Middling = rest and hospitals. We normalized the nominal figures by dividing them with the national average wage in the given month, as measured in the starting sample. Zero Person controls expenditures were replaced with 1 Ft (0.3 Euro cents Gender, age: Source: ONYF. per annum). Annual data. Source: OEP. Skill levels: Skill levels are inferred from the ‘highest’ occupational status held by the person in 2003–2011. The classification is basedon one-digit occupational Job controls codes: 1 Top managers, 2 Other managers, 3 Profession- Tenure: Months elapsed since entry to the firm. Set to als, 4 Other white collars, 5 Skilled blue collars, 6 Assem- zero in the case of left-censored employment spells. blers and machine operators, 7 Elementary occupations. A dummy stands for observations from left-censored Persons employed in occupations 1–3 at least once are spells. 3 Page 20 of 21 J. Köllő et al. Table 15 On‑the ‑job training: fraction participating among MNE and domestic firm employees ForeignDomesticRatio 2003 0.102 0.0651.57 2004 0.100 0.0601.67 2005 0.043 0.0241.80 2006 0.045 0.0192.41 2007 0.033 0.0191.73 2008 0.024 0.0191.30 2009 0.020 0.0131.59 2010 0.025 0.0151.59 2011 0.021 0.0141.61 High skilled employees working at least one hour in the reference week = 1 Source: Authors’ calculation using waves 45–80 of the LFS. Sample: ILO-OECD employed with college or university background. Key variables: participates in training of any kind outside the school system; the employer is majority or minority foreign-owned Note that the question on participation changed in 2005. Figures above and below the dotted line are not directly comparable worker-firm data. GEP Research Paper, No. 2007/08, University of Not - Spell lasting for 1  day: Hungarian firms often pay to tingham (2007) individual subcontractors by formally employing them Antalóczy, K., Sass, M.: Greenfield investments in Hungary. Transnatl. Corp. 3, on the day of payment. This practice results in exorbi - 39–60 (2001) Balatoni, A., Pitz, M.: A működőtőkehatása a bruttónemzetijövedelemreMagya- tant ‘daily wages’ in some cases. rországon. KözgazdaságiSzemle 59(1), 1–30 (2012) Occupation: One-digit ISCO codes. Balazsi, L.: The econometrics of linear models for multi-dimensional panels. Regional unemployment rate: seasonally adjusted Doctoral thesis in Economics, Central European University, Budapest (2017) https ://www.ceu.edu/sites /defau lt/files /phdth esisb ookle tbala zsi. ILO-OECD unemployment rate in the given month and pdf NUTS-2 region. The worker’s region is identified on the Balazsi, L., Matyas, L., Wansbeek, T.: The estimation of multidimensional fixed basis of his/her zip code in 2003. Source: author’s cal- effects panel data models. Econom. Rev. 37, 212–227 (2018) Balsvik, R.: Is labor mobility a channel for spillovers from multinationals? culation using the Labor Force Survey. Evidence from Norwegian manufacturing. Rev. Econ. Stat. 93(1), 285–297 (2011) Barry, F., Gorg, H., Strobl, E.: Foreign direct investment and wages in domestic firms in Ireland: productivity spillovers versus labour-market crowding Firm controls out. Int. J. Econ. Bus. 12, 67–84 (2005) Firm size: average number of employees. Annual data. Becker, G.S.: Investment in human capital: a theoretical analysis. J. Political Source: NAV. Econ. 70(5, Part 2), 9–49 (1962) Bernard, A.B., Sjoholm F.: Foreign owners and plant survival. NBER Working Capital-labor ratio: net value of fixed assets per Paper No. 10039, October 2003 (2003) worker. Annual data. Source: NAV. Bisztray, M.: The effect of FDI on local suppliers: evidence from Audi in Hun- Exporter: non-zero exports revenues. Annual data. gary. Institute of Economics, Centre for Economic and Regional Studies, Hungarian Academy of Sciences Discussion papers MT-DP-2016/22, pp. Source: NAV. 46 (2016) Sector: NACE 2. Source: NAV. Calderon, C., Loayza, N., Serven, L.: Greenfield foreign direct investment and mergers and acquisitions: feedback and macroeconomic effects. Policy Research Working Paper Series No. 3192. The World Bank, Washington Received: 17 February 2020 Accepted: 27 January 2021 (2004) Carneiro, A., Guimarães, P., Portugal, P.: Real wages and the business cycle: accounting for worker, firm, and job title heterogeneity. Am. Econ. J. Macroecon. 4(2), 133–152 (2012) Chen, C., Zhao, H., Zhou Y.: Foreign direct investment and wage inequality: References evidence from the People’s Republic of China (ADBI Working Paper No. Abowd, J.M., Kramarz, F., Margolis, D.N.: High wage workers and 734). Asian Development Bank Institute, Tokyo (2017) high wage firms. Econometrica 67, 251–333 (1999). https ://doi. Cornelissen, T.: Thestata command Felsdvreg to fit a linear model with two org/10.1111/1468-0262.00020 high-dimensional fixed effects. Stata J. 8 (2), 170–189 (2008) Abowd, J.M., Kramarz, F., Roux, S.: Wages, mobility and firm performance: Correia, S.: reghdfe: stata module for linear and instrumental-variable/gmm advantages and insights from using matched worker-firm data. Econ. J. R. regression absorbing multiple levels of fixed effects. Statistical Soft - Econ. Soc. 116(512), F245–F285 (2006) ware Components s457874, Boston College Department of Economics. Acemoglu, D., Pischke, J.S.: Beyond Becker: training in imperfect labour mar-https ://ideas .repec .org/c/boc/bocod e/s4578 74.html (2017) kets. Econ. J. 109(453), F112–F142 (1998) Csáfordi, Z., Lőrincz, L., Lengyel, B., Kiss, K.M.: Productivity spillovers through Aitken, B.J., Harrison, A.E.: Do domestic firms benefit from direct foreign invest - labor flows: productivity gap, multinational experience and industry ment? Evidence from Venezuela. Am. Econ. Rev. 89(3), 605–618 (1999) relatedness. J. Technol. Transf (2018). https ://doi.org/10.1007/s1096 Andrews, M., Bellmann, L., Schank, T., Upward, R.: The takeover and selec- 1-018-9670-8 tion effects of foreign ownership in Germany: an analysis using linked Wage gains from foreign ownership: evidence from linked employer–employee data Page 21 of 21 3 Csengődi, S., Jungnickel, R., Urban, D.: Foreign takeovers and wages in Maliranta, M., Mohnen, P., Rouvinen, P.: Is inter-firm labor mobility a channel of Hungary. Rev. World Econ. 144, 55–82 (2008). https ://doi.org/10.1007/ knowledge spillovers? Evidence from a linked employer–employee panel. s1029 0-008-0137-0 ETLA Discussion Papers, The Research Institute of the Finnish Economy Czafit, B., Köllő, J.: Employment and wages before and after incarceration; (ETLA), No. 1116 (2008) evidence from Hungary. IZA J. Eur. Labor Stud. 4(1), 1–21 (2015). https Martins, P.: Do foreign firms really pay higher wages? Evidence from different ://doi.org/10.1186/s4017 4-015-0044-z estimators. IZA, Discussion Paper No. 1388 (2004) Djankov, S., Hoekman, B.: Foreign investment and productivity growth in Martins, P.S., Opromolla, L.D.: Exports, imports and wages: evidence from Czech enterprises. World Bank Econ. Rev. 14(1), 49–64 (2000) matched firm-worker-product panels. IZA Discussion Paper No. 4646 Earle, J.S., Telegdy, Á., Antal, G.: Foreign ownership and wages: evidence (2009) from Hungary, 1986–2008. Ind. Labor Relat. Rev. 71(2), 458–491 (2017) Mion, G., Opromolla, L.D.: Managers’ mobility, trade performance, and wages. Elek, P., Köllő, J., Reizer, B., Szabó, P.A.: Detecting wage under-reporting using European Central Bank Working Paper Series No. 1596, October (2013) a double hurdle model. In: Lehmann, H., Tatsiramos, K. (eds.) Research Mittag, N.: A simple method to estimate large fixed effects models applied to in Labor Economics. Informal Employment in Emerging and Transition wage determinants and matching, IZA DP No. 10447, December (2016) Economies, Ch 4, vol. 34, pp. 135–166. Emerald, Bingley (2012) OECD: Chapter 5: Do multinationals promote better pay and working condi- Fabri, F., Haskel, J.E., Slaughter, M.J.: Does nationality of ownership matter for tions? In: OECD Employment Outlook, pp. 263–331. OCED, Paris (2008a) labor demands? J. Eur. Econ. Assoc. Pap. Proc. 1(2–3), 698–707 (2003) OECD: The impact of foreign direct investment on wages and working condi- Fons-Rosen, C., Kalemli-Ozcan, S., Sorensen, B.E., Villegas-Sanchez, C., Volos- tions, OECD-ILO Conference on Corporate Social Responsibility. Employ- ovych, V.: Foreign investment and domestic productivity: identifying ment and industrial relations: promoting responsible business conduct in knowledge spillovers and competition effects. NBER Working Paper No. a Globalising economy 23–24 June 2008, OECD Conference Centre, Paris, 23643, August (2017) France (2008b) Fosfuri, A., Motta, M., Ronde, T.: Foreign direct investment and spillovers Pettit, B., Lyons, C.J.: Incarceration and the legitimate labor market: examin- through workers’ mobility. J. Int. Econ. 53(1), 205–222 (2001) ing age-graded effects on employment and wages. Law Soc. Rev. 43(4), Görg, H., Strobl, E.: Spillovers from foreign firms through worker mobility: an 725–756 (2009) empirical investigation. Scand. J. Econ. 107(4), 693–709 (2005) Poole, J.P.: Knowledge transfers from multinational to domestic firms: evidence Gorodnichenko, Y., Svejnar, J., Terrell, K.: When does FDI have positive spillo- from worker mobility. Rev. Econ. Stat. 95(2), 393–406 (2013) vers? Evidence from 17 transition market economies. J. Comp. Econ. Smarzynska-Javorcik, B.: Does foreign direct investment increase the pro- 42, 954–969 (2014) ductivity of domestic firms? In search of spillovers through backward Grogger, J.: The effect of arrests on the employment and earnings of young linkages. Am. Econ. Rev. 94(3), 605–627 (2004) men. Q. J. Econ. 110(1), 51–71 (1995) Stoyanov, A., Zubanov, N.: Productivity spillovers across firms through worker Guimaraes, P., Portugal, P.: A simple feasible alternative procedure to estimate mobility. Am. Econ. J. Appl. Econ. 4(April), 168–198 (2012). https ://doi. models with high-dimensional fixed effects. Stata J. 10(4), 628–649 (2010) org/10.1257/app.4.2.168 Heyman, F., Sjöholm, F., Tingvall, P.: Multinationals, cross-border acquisitions Suyanto, R.S., Bloch, H.: Which firms benefit from foreign direct investment? and wage dispersion. Can. J. Econ. 44(2), 627–650 (2007) Empirical evidence from Indonesian manufacturing. J. Asian Econ. 33, Jenkins, S.P.: Easy estimation methods for discrete-time duration models. 16–29 (2014) Oxford Bull. Econ. Stat. 57(1), 129–136 (1995) UNECE: Economic Growth and Foreign Direct Investment in the Transition Jurajda, Š, Stančík, J.: Foreign ownership and corporate performance: the Economies. Economic Survey of Europe, 2001 No. 1, United Nations Czech Republic at EU entry. Financ. úvěr-Czech J. Econ. Financ. 62(4), Economic Commission for Europe, Geneva (2001) 306–324 (2012) Keller, W., Yeaple, S.R.: Multinational enterprises, international trade, and tech- Publisher’s Note nology diffusion: a firm-level analysis of the productivity effects of foreign Springer Nature remains neutral with regard to jurisdictional claims in pub- competition in the United States. Rev. Econ. Stat. 91(4), 821–831 (2009) lished maps and institutional affiliations. Konings, J.: The effects of foreign direct investment on domestic firms. Econ. Transit. 9(3), 619–633 (2001) Kosová, R.: Do foreign firms crowd out domestic firms? Evidence from the Czech Republic. Rev. Econ. Stat. 92(4), 861–881 (2010) LaLonde, R.J., Cho, R.: The impact of incarceration in state prison on the employment prospects of women. J. Quant. Criminol. 24(3), 243–267 (2008) Lazear, E.P.: Firm-specific human capital: a skill-weights approach. J. Political Econ. 117(5), 914–940 (2009) Lipsey, R.E., Sjöholm, F.: FDI and wage spillovers in Indonesian manufacturing. Rev. World Econ. 140, 321–332 (2004) Malchow-Moller, N., Markusen, J., Schjerning B.: Foreign firms, domestic work - ers. NBER Working Paper, No. 13001, Cambridge, Massachusetts (2007)

Journal

Journal for Labour Market ResearchSpringer Journals

Published: Feb 15, 2021

There are no references for this article.