Reassessing the foreign ownership wage premium in Germany

• Author: Hartmut Egger,Stefan Kornitzky,Elke Jahn
• DOI: http://doi.org/10.1111/twec.12910
• Published date: 01 February 2020
• Date: 01 February 2020

302

|

World Econ. 2020;43:302–325.

wileyonlinelibrary.com/journal/twec

Received: 24 June 2019

|

Revised: 23 October 2019

|

Accepted: 8 November 2019

DOI: 10.1111/twec.12910

ORIGINAL ARTICLE

Reassessing the foreign ownership wage premium in

Germany

HartmutEgger1

|

ElkeJahn2

|

StefanKornitzky3

1CESifo, GEP and IfW, University of Bayreuth, Bayreuth, Germany

2IAB Nuremberg, University of Bayreuth, Bayreuth, Germany and IZA

3University of Bayreuth, Bayreuth, Germany

Funding information

Deutsche Forschungsgemeinschaft, Grant/Award Number: 322072184

KEYWORDS

difference-in-difference estimation, foreign ownership wage premium, impact versus adjustment effects, propensity-score

matching

1

|

INTRODUCTION

Empirical evidence reported by Aitken, Harrison, and Lipsey (1996) that foreign direct investment

leads to higher wage payments has sparked a lot of interest among economists and has started a new

strand of empirical research in international trade, documenting the existence of a foreign ownership

wage premium (Girma, Greenaway, & Wakelin, 2001; Lipsey & Sjöholm, 2004; Velde & Morrissey,

2003). With access to detailed firm-level data, the causal relationship between foreign ownership and

wage premia has become an important topic in this strand of research over the last decade (Balsvik,

2011; Girma & Görg, 2007; Heyman, Sjöholm, & Tingvall, 2007). This paper provides new evidence

for a foreign ownership wage premium in Germany.

We contribute to the literature in two important ways. First, we estimate the foreign ownership

wage premium in Germany, using a longitudinal linked employer–employee data set that covers an

observation window spanning the years 2003–14. This long time span not only allows us to separate

the impact effect of foreign takeover in the period of ownership change from lagged adjustment ef-

fects in subsequent years, but also to estimate the foreign ownership wage premium by skill groups.

Second, we test three channels put forward by previous research, which give rise to a foreign owner-

ship wage premium. In particular, we consider rent appropriation by managers, technology protection

and training on new technology as distinct arguments justifying the existence of a foreign ownership

wage premium. Finally, we control for an export platform motive of foreign takeover to ensure that our

estimates do not accidentally pick up an exporter wage premium.

This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution

and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes.

© 2019 The Authors. The World Economy published by John Wiley & Sons Ltd

|

303

EGGER Et al.

To estimate a causal effect of foreign takeover on wages paid by German establishments, we com-

bine propensity-score matching with a difference-in-difference estimator. We associate the probability

of a worker in our sample to be treated—defined by staying in an establishment acquired by a foreign

investor—with the product of two probabilities: the probability that the employer is target of a foreign

takeover and the probability of the worker to remain employed in the same establishment. As both,

worker and establishment characteristics, matter for the treatment, we follow Martins (2004) and

Hijzen, Martins, Schank, and Upward (2013) and combine establishment information with data on

worker characteristics in the propensity-score matching. To construct our control group, we match to

each treated worker the nearest observational neighbour from establishments that are not subject to

foreign takeover.

Using a difference-in-difference approach, we then shed light on the effect of treatment on wages

distinguishing between an impact effect in the year after takeover and a lagged adjustment effect aris-

ing in the second and the third year after takeover. We use data from the year prior to acquisition to de-

termine the control group and build our analysis on a four-year window around the takeover. Because

we rely on takeover events from different years, choosing the same four-year observation window for

all acquisitions is important (see Egger, Egger, & Greenaway, 2008). In our baseline specification,

we estimate an average wage premium from foreign takeover of 4.0 log points, which further grows

to 6.3 log points 3years after ownership change. When analysing the impact of ownership change

for different skill groups, we find that the wage premium is larger for high-skilled than for low- and

medium-skilled workers. We also provide evidence that foreign takeover exerts a lagged adjustment

effect particularly on the wages of medium- and high-skilled workers.

To shed light on the economic mechanisms that can explain a foreign ownership wage premium,

we rely on insights from previous research and distinguish three arguments. First, managers receive

a higher remuneration after a successful acquisition (see Heyman, Sjöholm, & Tingvall, 2011). If

rent appropriation by managers plays a role, we would expect to find a transitory increase in manager

wages. Second, as a result of knowledge transfer by the investor, foreign takeover may lead to the use

of new technology requiring training of workers (see Fosfuri, Motta, & Rønde, 2001; Görg, Strobl, &

Walsh, 2007). If training of workers requires time, we would expect to observe a lagged adjustment

of wages. Third, foreign-owned plants may have an incentive to protect their technology by reducing

knowledge dissipation through worker turnover (see Glass & Saggi, 2002). Workers should thus re-

ceive wage premia as long as foreign-owned plants possess a technology advantage.

Using information on manager status and product as well as process innovation, we analyse these

different explanations empirically and find support for the relevance of all three channels. Our results

also indicate that none of the considered channels can simultaneously explain the immediate wage

premium in the year after takeover and the lagged adjustment effect in consecutive periods. Because

previous research suggests that an important motive for foreign investment can be market access in

nearby countries, we also account for the initial export status of acquired plants to control for an export

platform motive. We thus ensure that the estimated wage premia do not incorrectly pick up an exporter

wage premium due to an expansion of trade with Eastern Europe after the millennium (cf. Dauth,

Findeisen, & Suedekum, 2014).1 Since the estimated wage premium from foreign takeover does not

differ for initial exporters and non-exporters, this concern seems not to be justified.

1 The platform argument has been put forward by Motta and Norman (1996), Yeaple (2003), Ekholm et al. (2007), and Neary

(2009) as an important motive for foreign investment. It has been used to explain the finding that foreign investment

increased in times of falling trade costs. Bernard and Bradford Jensen (1999, 1995) were the first providing evidence that

exporters pay higher wages than non-exporters. The existence of an exporter wage premium has been empirically confirmed

by Schank et al. (2007) and Wagner (2007).